SEPA
United States
Environmental Protection
Agency
Economic Analysis
for the
Final Revised Total Coliform Rule
-------
Office of Water (4606M)
EPA 815-R-12-004
www, epa. gov/ safewater
September 2012
-------
Contents
Appendices vi
Exhibits vii
List of Acronyms xiii
Health Risk Reduction and Cost Analysis xvi
Executive Summary ES-1
ES.l Need for the Rule ES-1
ES.2 Consideration of Regulatory Options ES-4
ES.3 Summary of the RTCR Requirements ES-5
ES.4 Systems Subject to the RTCR ES-8
ES.5 National Benefits and Costs of the RTCR ES-8
ES.5.1 Derivation of Benefits ES-10
ES.5.2 Derivation of Costs ES-19
ES.6 Projected Impacts on Household Costs ES-20
ES.7 Comparison of Benefits and Costs, and Regulatory Options of the RTCR ES-22
ES.8 Conclusion ES-27
1 Introduction 1-1
1.1 Summary of the Revised Total Coliform Rule (RTCR) 1-1
1.2 Document Organization 1-4
1.3 Calculations and Citations 1-5
2 Statement of Need for the Rule 2-1
2.1 Introduction 2-1
2.1.1 Description of the Issue 2-1
2.2 Public Health Concerns, Fecal Contamination, and Waterborne Pathogens 2-3
2.2.1 Rule Objectives and Public Health Concerns 2-3
2.2.2 Total Coliforms as Indicators of Treatment Effectiveness and Integrity of the
Distribution System 2-9
2.2.3 Sanitary Defects 2-9
2.2.4 Occurrence of Fecal Contamination and/or Waterborne Pathogens 2-9
2.3 Statutory Authority for Promulgating the Rule 2-11
2.4 Economic Rationale 2-11
3 Consideration of Regulatory Options 3-1
3.1 Introducti on 3-1
3.2 Total Coliform Rule/Distribution System Advisory Committee 3-1
3.3 Regulatory Options Considered 3-2
3.3.1 Comparative Summary of Regulatory Options 3-5
3.4 Final Rule Requirements 3-13
Economic Analysis for the Final RTCR i September 2012
-------
4 Baseline Conditions 4-1
4.1 Introduction 4-1
4.1.1 Background and Purpose 4-1
4.1.2 Chapter Organization 4-1
4.2 Data Sources 4-2
4.2.1 Background on SDWIS/FED Data 4-2
4.2.2 Background on 2005 Six-Year Review 2 Data 4-6
4.2.3 Background on Other Data and Information Used 4-9
4.3 Baseline Profile 4-11
4.3.1 Pre-GWR Baseline 4-12
4.3.2 Post-GWR Baseline 4-15
4.3.3 Baseline Population Served 4-17
4.3.4 Baseline Water Quality 4-19
4.4 Sensitive Sub-Populations 4-24
4.5 Summary of Baseline Assumptions 4-25
5 Occurrence and Predictive Model 5-1
5.1 Introducti on 5-1
5.2 Modeling of Current Total Coliform and E. coli Occurrence for Systems
Serving <4,100 People 5-3
5.2.1 Di stributi onal Model and Notati on 5-3
5.2.2 Data Reduction 5-3
5.2.3 Basic Subsets of Systems 5-4
5.2.4 Estimation Methodology 5-6
5.2.5 Results 5-9
5.3 Predictive Modeling of Occurrence for Systems Serving Up to 4,100 People 5-15
5.3.1 Summary of GWR factors and timing affecting the 1989 TCR and RTCR 5-15
5.3.2 Summary of Predictive Model 5-18
5.3.3 Predictive Model Results 5-27
5.4 Occurrence Analysis for Systems Serving More Than 4,100 People 5-54
5.4.1 Model (for Systems Serving >4,100 People) 5-55
5.4.2 Model Results (for Systems Serving >4,100 People) 5-56
5.4.3 Model Uncertainty (for Systems Serving >4,100 People) 5-56
5.5 Summary of Key Drivers for Benefit and Cost Analyses Output from the
Predictive Model 5-59
6 Benefits Analysis 6-1
6.1 Introduction 6-1
6.2 Qualitative Benefits Analyses 6-1
6.2.1 Implementation Activities 6-2
Economic Analysis for the Final RTCR ii September 2012
-------
6.2.2 Routine Monitoring 6-2
6.2.3 Repeat Monitoring 6-5
6.2.4 Additional Routine Monitoring 6-6
6.2.5 Annual Site Inspections 6-7
6.2.6 Assessments 6-8
6.2.7 Corrective Actions 6-9
6.2.8 Public Notification 6-9
6.2.9 Summary of Qualitative Benefits Analyses 6-16
6.3 Assessment of Predictive Analysis Results 6-16
6.3.1 Assessment of Predictive Analysis Results for Systems Serving <4,100 People
and Systems Serving >4,100 People 6-18
6.4 Uncertainty and Sensitivity Analyses 6-28
6.5 Other Potential Benefits 6-32
6.5.1 Increased System Knowledge 6-32
6.5.2 Accelerated Infrastructure Repair/Replacement 6-32
6.5.3 Reduction in Averting Behavior 6-33
6.5.4 Reduction of Co-Occurring and Other Contaminants 6-33
6.5.5 Reduction in Outbreak Risk and Response Costs 6-33
7 Cost Analysis 7-1
7.1 Introducti on 7-1
7.2 General Cost Assumptions and Methodology 7-1
7.2.1 Labor Rates 7-2
7.2.2 TCR Monitoring Costs per Sample 7-3
7.2.3 Technology Unit Costs and Compliance Forecasts 7-3
7.2.4 Cost Model 7-4
7.2.5 Modeled Variability and Uncertainty in National Costs 7-4
7.3 Projecting and Discounting National Costs 7-5
7.4 Derivation of Costs for PWSs and States 7-6
7.4.1 Rule Implementation and Annual Administration 7-6
7.4.2 Revise Sample Siting Plan 7-9
7.4.3 Monitoring 7-12
7.4.4 Annual Site Visits 7-19
7.4.5 Assessments 7-19
7.4.6 Corrective Actions 7-28
7.4.7 Public Notification 7-33
7.4.8 Uncertainty in Unit Costs 7-38
7.5 Household Costs 7-38
7.6 Non-quantified Costs 7-41
Economic Analysis for the Final RTCR
Hi
September 2012
-------
7.7 Uncertainty Analysis 7-41
7.8 Comparison of Total and Net Annualized Costs for All Regulatory Options 7-45
8 Economic Impact Analysis 8-1
8.1 Introducti on 8-1
8.2 Executive Order 12866: Regulatory Planning and Review 8-2
8.3 Paperwork Reduction Act 8-2
8.4 The Regulatory Flexibility Act 8-6
8.5 Unfunded Mandates Reform Act 8-10
8.6 Executive Order 13132: Federalism 8-11
8.7 Executive Order 13175: Consultation and Coordination with Indian Tribal
Governments 8-12
8.8 Executive Order 13045: Protection of Children from Environmental Health
Risks and Safety Risks 8-13
8.9 Executive Order 13211: Action Concerning Regulations That Significantly
Affect Energy Supply, Distribution, or Use 8-13
8.10 National Technology Transfer and Advancement Act 8-14
8.11 Executive Order 12898: Federal Actions to Address Environmental Justice in
Minority Populations and Low-Income Populations 8-15
8.12 Consultations with the Science Advisory Board, National Drinking Water
Advisory Council, and the Secretary of Health and Human Services as
Required by Section 1412 (d) and (e) of the SDWA 8-15
8.13 Consideration of Impacts on Sensitive Subpopulations as Required by Section
1412(b)(3)(c)(i)(V) of the 1996 Amendments to the Safe Drinking Water Act
(SDWA) 8-16
8.14 Effect of Compliance with the RTCR on the Technical, Financial, and
Managerial Capacity of Public Water Systems as Required by Section
1420(d)(3) of SDWA 8-17
8.14.1 Requirements of the RTCR 8-18
8.14.2 Systems Subject to the RTCR 8-18
8.14.3 Impact of the RTCR on System Capacity 8-19
8.14.4 Derivation of RTCR Scores 8-21
8.14.5 Small Water Systems (Those Serving 10,000 or Fewer People) 8-22
8.14.6 Large Water Systems (Those Serving Greater Than 10,000 People) 8-23
9 Comparison of Benefits and Costs 9-1
9.1 National Benefits and Costs of the RTCR Considered in Comparison to the
1989 Total Coliform Rule and Alternative Option 9-1
9.1.1 National Benefits of the Regulatory Options Considered 9-3
9.1.2 National Cost Summary 9-18
9.1.3 Comparison of National Benefits and Costs of the Regulatory Options
Considered 9-26
Economic Analysis for the Final RTCR
iv
September 2012
-------
9.2 Effect of Uncertainties and Non-quantified Benefit/Cost Estimates on the
Estimation of National Benefits and Costs 9-27
9.2.1 Summary of Major Uncertainties in EA Analyses 9-27
9.2.2 Summary of Non-quantified Costs and Benefits 9-29
9.3 Comparison of the Regulatory Options Considered 9-30
9.3.1 Net Change in Costs and Benefits 9-30
9.3.2 Cost Effectiveness Measures 9-31
9.3.3 Break-Even Analysis 9-33
9.3.4 Summary of Conclusions 9-39
10 References 10-1
Economic Analysis for the Final RTCR
v
September 2012
-------
Appendices
Appendix A: Detailed Predictive Model Results
Appendix B: Graphs of Predicted Occurrence Over Time
Appendix C: Detailed Cost Model Results
Appendix D: Detailed Compliance Forecast and Unit Costs Estimates
Appendix E: Break-Even Analysis
Appendix F: Alternative Strategies for Combining Basic Subsets
Appendix G: Evaluation of Representativeness of Six-Year Review 2 Data
Appendix H: Analysis of Repeat Sample Records from the Six-Year Review 2 Data
Appendix I: Supporting Information for Regulatory Flexibility Act Screening Analysis
Economic Analysis for the Final RTCR
vi
September 2012
-------
Exhibits
Executive Summary
Exhibit ES.l Implementation Schedule ES-8
Exhibit ES.2 Modeled TC+ Occurrence in Transient Noncommunity Ground Water
Systems Serving < 4,100 ES-12
Exhibit ES.3 Potential Changes in Risk under the RTCR and Alternative Option
Relative to the 1989 TCR ES-13
Exhibit ES.4 Predicted Annual Number of Acute (1989 TCR) or E. coli MCL (RTCR and
Alternative Option) Violations by Regulatory Option and
System Type (Year 9) ES-18
Exhibit ES.5 Comparison of Total and Net Change from 1989 TCR in Annualized
Present Value Costs ($Millions, 2007$) ES-20
Exhibit ES.6 Summary of Net Annual Per-Household Costs for the RTCR (2007$) ES-22
Exhibit ES.7 Estimated Annual Breakeven Threshold for Avoided Cases of
STEC 0157 ES-24
Exhibit ES.8 Estimated Annual Breakeven Threshold for Avoided Cases of
Salmonella ES-25
Exhibit ES.9 Total Net Annual Cost Per Corrective Action Implemented under
the RTCR and Alternative Option, Annualized Using 3% and 7%
Discount Rates ($2007) ES-26
Exhibit ES. 10 Annualized Net Change in Costs Per Corrective Action (CA)
Implemented for All PWSs under the RTCR and Alternative Option
(SVlillions. 2007$) ES-27
Exhibit ES. 11 Annualized Net Change in Costs Per CA Implemented for TNCWSs
(Serving <100 people) under the RTCR and Alternative Option
(SVlillions. 2007$) ES-27
1. Introduction
Exhibit 1.1 RTCR Monitoring Frequency Requirements 1-3
2. Statement of Need for the Rule
3. Consideration of Regulatory Options
Exhibit 3.1 Comparison of RTCR Regulatory Options 3-6
4. Baseline Conditions
Exhibit 4.1 Pre-GWR Baseline Number of GW Systems 4-13
Exhibit 4.2 Baseline Number of SW Systems 4-13
Exhibit 4.3 Pre-GWR Baseline Population Served by GW Systems 4-14
Exhibit 4.4 Percent Distribution of GW System Monitoring Frequencies by
PWS Size and Type for 1989 TCR 4-15
Economic Analysis for the Final RTCR
vii
September 2012
-------
Exhibit 4.5 Baseline Number of GW Systems Post-Implementation of the
GWR (Model Year 5) 4-17
Exhibit 4.6 Baseline Population Served by GW Systems Post-Implementation
of the GWR (Model Year 5) 4-18
Exhibit 4.7 Baseline Population Served by SW Systems 4-18
Exhibit 4.8 Household Baseline 4-19
Exhibit 4.9 Total Coliform and E. coli Percent Positive by System Size and Type 4-21
Exhibit 4.10 Baseline Number of 1989 TCR Violations by System Size and
Type (2005) 4-23
Exhibit 4.11 Number of PWSs with Violations by System Type (2001-2007) 4-24
Exhibit 4.12 Estimates of Sensitive Subpopulations in the United States 4-25
Exhibit 4.13 Summary of Baseline Assumptions Influencing RTCR Estimates 4-26
5. Occurrence and Predictive Model
Exhibit 5.1 Basic Classifications of PWSs Used for Occurrence Modeling 5-5
Exhibit 5.2 Directed Graph of Model Used for RTTC Occurrence 5-7
Exhibit 5.3 Maximum Likelihood a and b Parameter Estimates for RTTC and RPTC 5-10
Exhibit 5.4 Maximum Likelihood a and b Parameter Estimates for RTEC and RPEC 5-11
Exhibit 5.5 a and P Parameter Estimates for RTTC and RPTC 5-13
Exhibit 5.6 a and P Parameter Estimates for RTEC and RPEC 5-14
Exhibit 5.7 Simulated Impact of the RTCR and Alternative Option on Systems
Serving <4,100 People—Surface Water Systems 5-22
Exhibit 5.8 Simulated Impact of the RTCR and Alternative Option on Systems
Serving <4,100 People—Ground Water Systems 5-23
Exhibit 5.9a Percent of GW Systems Assumed to be on Monthly, Quarterly, and
Annual (M/Q/A) Monitoring by System Category under 1989
TCR (Baseline)—Initial Estimates (Post-GWR Implementation) 5-24
Exhibit 5.9b Percent of GW Systems Predicted to be on M/Q/A Monitoring by
System Category under RTCR—Adjusted Estimates
(Post-RTCR Implementation) 5-25
Exhibit 5.9c Percent of GW Systems on M/Q/A Monitoring by System
Category under Alternative Option—Adjusted Estimates
(Post-RTCR Implementation) 5-26
Exhibit 5.10 Ground Water Community Water System Model Output Cumulative
Endpoints 5-30
Exhibit 5.11 Ground Water Nontransient Noncommunity Water System Model Output
Cumulative Endpoints 5-30
Exhibit 5.12 Ground Water Transient Noncommunity Water System Model Output
Cumulative Endpoints 5-31
Exhibit 5.13 Surface Water Community Water System Model Output Cumulative
Endpoints 5-31
Exhibit 5.14 Surface Water Nontransient Noncommunity Water System Model Output
Cumulative Endpoints 5-32
Exhibit 5.15 Surface Water Transient Noncommunity Water System Model Output
Cumulative Endpoints 5-32
Economic Analysis for the Final RTCR
September 2012
-------
Exhibit 5.16 Ground Water Community Water System (Serving < 4,100)
TC Occurrence 5-33
Exhibit 5.17 Ground Water Nontransient Noncommunity Water System
(Serving < 4,100) TC Occurrence 5-34
Exhibit 5.18 Ground Water Transient Noncommunity Water System
(Serving < 4,100) TC Occurrence 5-35
Exhibit 5.19 Surface Water Community Water System (Serving < 4,100)
TC Occurrence 5-36
Exhibit 5.20 Surface Water Nontransient Noncommunity Water System
(Serving < 4,100) TC Occurrence 5-37
Exhibit 5.21 Surface Water Transient Noncommunity Water System
(Serving < 4,100) TC Occurrence 5-38
Exhibit 5.22a Summary of Model Parameters Influencing 1989 TCR, RTCR and
Alternative Option 5-40
Exhibit 5.22b Summary of Model Parameters Influencing RTCR and
Alternative Option Only 5-43
Exhibit 5.23 Sensitivity Analysis Assumptions for Frequency and Effectiveness of
Corrective Actions following Level 1 or 2 Assessments 5-48
Exhibit 5.24a Cumulative Effect of Alternative Assumptions for Corrective Action
(CA) Effectiveness and Duration on RTCR Model Results
(Nondisinfecting TNCWS Serving <500 People over 25 Years) 5-48
Exhibit 5.24b Cumulative Effect of 50 percent Assumption for Corrective Action
Implementation Rate on RTCR Model Results (Nondisinfecting
TNCWS Serving <500 People over 25 Years) 5-50
Exhibit 5.25 Effect of Seasonality on Occurrence Analysis Endpoints 5-51
Exhibit 5.26 Comparison of TC+ Occurrence Predicted as a 25-Year Annual
Average under the 1989 TCR with 2005 Six-Year Review Data 5-53
Exhibit 5.27 Comparison of SDWIS Data for Non-acute and Acute Violations with
Predictive Model Annual Results for the 1989 TCR 5-54
Exhibit 5.28 Results for Systems Serving more than 4,100 People—1989 TCR 5-57
Exhibit 5.29 Results for Systems Serving more than 4,100 People—RTCR 5-58
Exhibit 5.30 Results for Systems Serving more than 4,100 People—Alternative Option 5-59
6. Benefits Analysis
Exhibit 6.1 Potential Changes in Risk under the RCTR and Alternative Option
Relative to the 1989 TCR 6-13
Exhibit 6.2 Predicted Outcomes (25-Year Period of Analysis) under the 1989 TCR 6-22
Exhibit 6.3 Predicted Outcomes (25-Year Period of Analysis) for the RTCR 6-23
Exhibit 6.4 Predicted Outcomes (25-Year Analysis Period) for the Alternative Option 6-24
Exhibit 6.5 Predicted Average Annual Acute Violations by Regulatory
Option and System Type 6-25
Exhibit 6.6 Predicted Change in Average Annual Acute Violations by
Regulatory Option and System Type 6-26
Exhibit 6.7 Relative Impacts Analysis for TNCWS Serving <101 People
for the Range of Corrective Action and Sampling Regimens Predicted 6-30
Economic Analysis for the Final RTCR
ix
September 2012
-------
7. Cost Analysis
Exhibit 7.1 Labor Rates by PWS Size (2007$) 7-2
Exhibit 7.2 Monitoring Costs per Sample (2007$) 7-3
Exhibit 7.3 Net Change in PWS Unit Burden and Cost Estimates for Rule
Implementation (2007$) 7-7
Exhibit 7.4 Net Change in State Unit Burden and Cost Estimates for Rule
Implementation (2007$) 7-8
Exhibit 7.5 Annualized Cost Estimates for Rule Implementation
(SVlillions. 2007$) 7-9
Exhibit 7.6 Net Change in PWS and State Burden and Cost Estimates to
Revise Sample Siting Plans (2007$) 7-11
Exhibit 7.7 Annualized Cost Estimates to Revise (PWSs) and Review (States)
Sample Siting Plans ($Millions, 2007$) 7-12
Exhibit 7.8 Summary of Monitoring Requirements Under the 1989 TCR, RTCR,
and Alternative Option 7-13
Exhibit 7.9 Cumulative Number of Samples over 25-Year Period of Analysis 7-17
Exhibit 7.10 Annualized PWS and State Cost Estimates for Monitoring Costs
(SVlillions. 2007$) 7-18
Exhibit 7.11 PWS Unit Costs Estimates for Assessment Activities
(1989 TCR) (2007$) 7-22
Exhibit 7.12 PWS Unit Costs Estimates for Level 1 and Level 2 Assessments
(RTCR and Alternative Option) (2007$) 7-23
Exhibit 7.13 Number of Level 1 and Level 2 Assessments over the 25-Year
Compliance Period 7-24
Exhibit 7.14 State Unit Cost Estimates for Review of Level 1 and Level 2
Assessments under the 1989 TCR, RTCR, and Alternative
Option (2007$) 7-27
Exhibit 7.15 Annualized PWS and State Cost Estimates for Level 1 and Level 2
Assessments ($Millions, 2007$) 7-28
Exhibit 7.16a Compliance Forecast for Corrective Actions based on Level 1 and
Level 2 Assessments 7-30
Exhibit 7.16b Detailed PWS Compliance Forecast for Corrective Actions based on
Level 1 and Level 2 Assessments 7-31
Exhibit 7.17 Net Change in PWS and State Unit Costs Estimates for Reporting and
Recordkeeping for Corrective Actions (2007$) 7-32
Exhibit 7.18 Annualized PWS and State Cost Estimates for Corrective Actions
based on Level 1 and Level 2 Assessments ($Millions, 2007$) 7-33
Exhibit 7.19 PWS Unit Cost Estimates for Public Notification (2007$) 7-35
Exhibit 7.20 State Unit Costs Estimates for Public Notification (1989 TCR, RTCR,
Alternative Option) 7-36
Exhibit 7.21 Number of Tier 1 and Tier 2 Public Notifications over the 25-Year
Compliance Period 7-37
Exhibit 7.22 Annualized PWS and State Cost Estimates for Public Notification
(SVlillions. 2007$) 7-38
Economic Analysis for the Final RTCR x September 2012
-------
Exhibit 7.23 Summary of Net Annual Per-Household Costs for the RTCR (2007$) 7-40
Exhibit 7.24 Sensitivity Analysis—Annualized Net Change in Costs based on
Changes in Compliance Forecast ($Millions, 2007$) 7-43
Exhibit 7.25 Cost Uncertainty Summary 7-44
Exhibit 7.26 Comparison of Total and Net Change from 1989 TCR in
Annualized Present Value Costs ($Millions, 2007$) 7-45
Exhibit 7.27 Comparison of Total and Net Change in Annualized Present
Value Costs by Rule Component ($Millions, 2007$) 7-46
Exhibit 7.28 Total and Net Change in Annualized Costs to PWSs by
PWS Size and Type (SVlillions. 2007$) 7-47
Exhibit 7.29 Total and Net Change in Annualized Per PWS Costs by
PWS Size and Type (2007$) 7-48
8. Economic Impact Analysis
Exhibit 8.1 Average Annual Net Change Burden and Costs for the RTCR
Information Collection Request 8-5
Exhibit 8.2 RTCR—Average Annualized Revenue by System Size and Percent of
Systems with Costs Exceeding One Percent and Three Percent of
Revenue (Three Percent Discount Rate) 8-7
Exhibit 8.3 RTCR—Average Costs per System and as Percentage of Revenue
(2007$) 8-8
Exhibit 8.4 RTCR—Annualized Net Rule Costs Predicted for Small Entities
(PWSs serving <10,000) by Rule Component Using Three
Percent and Seven Percent Discount Rates (2007$) 8-9
Exhibit 8.5 Estimated Impact of the RTCR on Small Systems' Technical,
Managerial, and Financial Capacity 8-20
Exhibit 8.6 Estimated Impact of the RTCR on Large Systems' Technical,
Managerial, and Financial Capacity 8-21
9. Comparison of Benefits and Costs
Exhibit 9.1a Ground Water (GW) Transient Noncommunity Water Systems
(Serving < 4,100) TC Occurrence 9-5
Exhibit 9. lb Potential Changes in Risk under the RTCR and Alternative Option
Relative to the 1989 TCR 9-7
Exhibit 9.2 Estimates of Non-Acute Violations (TCR) and Level 1 Assessment
Triggers (RTCR and Alternative Option) 9-13
Exhibit 9.3 Estimates of Acute Violations (TCR) and E. coli MCL Violations
(RTCR and Alternative Option) 9-14
Exhibit 9.4 Estimates of Corrective Actions 9-15
Exhibit 9.5 Discounted Estimates of Non-Acute Violations (TCR) and Level 1
Assessment Triggers (RTCR and Alternative Option) (3% Discount Rate) 9-16
Exhibit 9.6 Discounted Estimates of Acute Violations (TCR) and E. coli
MCL Violations (RTCR and Alternative Option) (3% Discount Rate) 9-17
Exhibit 9.7 Discounted Estimates of Corrective Actions (3% Discount Rate) 9-18
Economic Analysis for the Final RTCR
xi
September 2012
-------
Exhibit 9.8 Net Costs to PWSs (2007$) 9-21
Exhibit 9.9 Discounted Net Costs to PWSs Over Time (3% Discount Rate) (2007$) 9-22
Exhibit 9.10 Discounted Net Costs to PWSs (7% Discount Rate) (2007$) 9-23
Exhibit 9.11 Comparison of Total and Net Change in Annualized Present
Value Costs (SVlillions. 2007$) 9-24
Exhibit 9.12 Comparison of Total and Net Change in Annualized Present Value
Costs by Rule Component ($Millions, 2007$) 9-25
Exhibit 9.13 Total and Net Change in Annualized Costs to PWSs by PWS Size
and Type ($Millions, 2007$) 9-26
Exhibit 9.14 Net Change in Annualized Present Value Costs ($Millions, 2007$)
and Benefits (Number of Level 2 Corrective Actions) 9-31
Exhibit 9.15 Total Net Annual Cost per Corrective Action Implemented under
RTCR and Alternative Option, Annualized Using 3% and 7%
Discount Rates ($2007) 9-32
Exhibit 9.16 Incremental Rule Cost per Corrective Action (CA) Implemented for
All PWSs under RTCR and Alternative Option ($Millions, 2007$) 9-33
Exhibit 9.17 Incremental Rule Cost per CA Implemented for GW TNCWSs
(Serving <100 people) under RTCR and Alternative Option
(SVlillions. 2007$) 9-33
Exhibit 9.18 Average Estimated Value per STEC 0157 Case Avoided (2007$) 9-36
Exhibit 9.19 Average Estimated Value per Salmonella Case Avoided (2007$) 9-37
Exhibit 9.20 Estimated Annual Break-Even Threshold for Avoided Cases
of STEC 0157 9-37
Exhibit 9.21 Estimated Annual Break-Even Threshold for Avoided Cases
of Salmonella 9-38
10. References
Economic Analysis for the Final RTCR
xii
September 2012
-------
List of Acronyms
A/CA
Assessment and corrective action if necessary
AGI
Acute Gastrointestinal Illness
AIDS
Acquired Immune Deficiency Syndrome
AIP
Agreement in Principle
ASDWA
Association of State Drinking Water Administrators
BMPs
Best Management Practices
CAs
Corrective Actions
CCR
Consumer Confidence Report
CDC
Centers for Disease Control and Prevention
CEA
Cost Effectiveness Analysis
COI
Cost of Illness
CWS
Community Water System
DBP
Disinfection Byproduct
DBPR
Disinfection Byproducts Rule
DisGW
Disinfected Ground Water
DisinfSW
Disinfected Surface Water
DV
Data Verification
EA
Economic Analysis
EC
E. coli bacteria
EC+
E. co//-positive
ECI
Employee Cost Index
EO
Executive Order
EPA
United States Environmental Protection Agency
ERS
Economic Research Service
FAC
Federal Advisory Committee
FACA
Federal Advisory Committee Act
FC
Fecal coliform bacteria
FC+
Fecal Coliform-positive
FDA
U.S. Food and Drug Administration
FR
Federal Register
GW
Ground Water
GWR
Ground Water Rule
GWUDI
Ground Water Under the Direct Influence of Surface Water
HAA5
Five Haloacetic Acids
HAV
Hepatitis A Virus
HEV
Hepatitis E Virus
HHS
U.S. Department of Health and Human Services
HUS
Hemolytic Uremic Syndrome
ICR
Information Collection Rule
ICR
Information Collection Request
IDSEs
Initial Distribution System Evaluations
IESWTR
Interim Enhanced Surface Water Treatment Rule
KRPEC
Number of repeat TC-positive samples assayed that tested positive for EC
KRPTC
Number positive repeat samples during year
Economic Analysis for the Final RTCR
xiii
September 2012
-------
KRTEC
Number of routine TC-positive samples assayed that tested positive
for EC
KRTTC
Number positive routine samples during year
LRAAs
Locational Running Annual Averages
LT1ESWTR
Long Term 1 Enhanced Surface Water Treatment Rule
LT2ESWTR
Long Term 2 Enhanced Surface Water Treatment Rule
M/DBP
Microbial/Disinfection By-Product
M/Q/A
Monthly, Quarterly, and Annual
MAC
Mycobacterium avium Complex
MCL
Maximum Contaminant Level
MCLG
Maximum Contaminant Level Goal
MCMC
Bayesian Markov Chain Monte Carlo
MLEs
Maximum Likelihood Estimates
MRDLGs
Maximum Residual Disinfectant Level Goals
MRDLs
Maximum Residual Disinfectant Levels
NAICS
North American Industry Classification System
NCWS
Non-Community Water System
NDPWR
National Primary Drinking Water Regulation
NondisGW
Nondisinfected ground water
NPDWRs
National Primary Drinking Water Regulations
NRPEC
Number of repeat TC-positive samples that were assayed for EC
NRPTC
Total number repeat samples assayed
NRTEC
Number of routine TC-positive samples that were assayed for EC
NRTTC
Number routine samples assayed
NTNC
Nontransient Noncommunity
NTNCWS
Nontransient Noncommunity Water System
NTTAA
National Technology Transfer and Advancement Act
NTUs
Nephelometric Turbidity Units
O&M
Operations and Maintenance
OMB
Office of Management and Budget
PAM
Primary Amoebic Meningioencephalitis
PN
Public Notification
POU
Point of Use
pRPEC
Repeat probability that a TC-positive repeat sample will also test EC-
positive
pRPTC
Probability that a repeat sample will test positive
pRTEC
Probability that a TC-positive routine sample will also test EC-positive
pRTTC
Probability that a routine sample will test positive
PWSs
Public Water Systems
RAA
Running Annual Average
RFA
Regulatory Flexibility Act
RPEC
Repeat coli
RPTC
Repeat TC
RTCR
Revised Total Coliform Rule
RTEC
Routine coli
RTTC
Routine TC
Economic Analysis for the Final RTCR
xiv
September 2012
-------
SAB Science Advisory Board
SBA Small Business Administration
SBREFA Small Business Regulatory Enforcement Fairness Act
SDWA Safe Drinking Water Act
SDWIS Safe Drinking Water Information System
SDWIS/FED Safe Drinking Water Information System Federal Version
Six Year Review Six Year Review of National Primary Drinking Water Regulations
SS Sanitary survey
STEC Shiga toxin-producing E. coli
SW Surface Water
SWTR Surface Water Treatment Rule
TC Total Coliform bacteria
TC+ Total Coliform-positive
TCR Total Coliform Rule
TCRDSAC Total Coliform Rule/Distribution System Advisory Committee
TMF Technical, Managerial, and Financial
TNC Transient Noncommunity
TNCWS Transient Noncommunity Water Systems
TTHM Total Trihalomethane
TWG Technical Work Group
UMRA Unfunded Mandates Reform Act
UV Ultraviolet light
VSL Value of a Statistical Life
Economic Analysis for the Final RTCR
xv
September 2012
-------
Health Risk Reduction and Cost Analysis
Under the Safe Drinking Water Act (SDWA) Amendments of 1996, when promulgating a
national primary drinking water regulation that includes a maximum contaminant level (MCL),
the U.S. Environmental Protection Agency (EPA) must conduct a health risk reduction and cost
analysis (HRRCA). A HRRCA addresses seven requirements, all of which are addressed in this
economic analysis (EA) for the final Revised Total Coliform Rule (RTCR).
HRRCA Crosswalk Summary
HRRCA Requirement
Addressed in Economic Analysis
Quantifiable and nonquantifiable health risk
reduction benefits
Chapter 6 (all sections)
Chapter 8 (sections 8.8)
Chapter 9 (sections 9.1.1, 9.2.2, and 9.3)
Quantifiable and nonquantifiable health risk
reduction benefits from co-occurring
contaminants
Chapter 6 (section 6.5.1, 6.5.4)
Chapter 9 (section 9.2.2)
Quantifiable and nonquantifiable costs
Chapter 7 (all sections)
Chapter 8 (sections 8.3-8.5, and 8.14)
Chapter 9 (sections 9.1.2, 9.1.3, 9.2.2, and 9.3)
Incremental costs and benefits associated
with regulatory options
Chapter 6 (sections 6.2 and 6.3)
Chapter 7 (sections 7.4, 7.5, and 7.8)
Chapter 9 (sections 9.1 and 9.3)
Effects of the contaminants on the general
population and sensitive subpopulations
Chapter 2 (section 2.2)
Chapter 6 (all sections)
Chapter 8 (section 8.13)
Chapter 9 (sections 9.1.1, 9.2.2, and 9.3)
Increased health risk that may occur as a
result of compliance
Chapter 6 (section 6.2)
Chapter 9 (section 9.1.1)
Other relevant factors (quality and
uncertainty of information)
Chapter 4 (sections 4.2 and 4.3)
Chapter 5 (section 5.3.3.1)
Chapter 6 (section 6.4)
Chapter 7 (section 7.7)
Chapter 9 (section 9.2.1)
Economic Analysis for the Final RTCR
xvi
September 2012
-------
Executive Summary
This economic analysis (EA) presents the evaluation of the benefits and costs of the
Revised Total Coliform Rule (RTCR). The analysis is performed in compliance with Executive
Order 12866, Regulatory Planning and Review (58 Federal Register (FR) 51735, September
1993), which requires the United States Environmental Protection Agency (EPA or Agency) to
estimate the economic impact of rules that have an annual effect on the economy of over $100
million. The Order further requires agencies to make the analysis for an "economically
significant" rule available to the public in conjunction with publication of the proposed rule.
Although EPA's analysis of the RTCR has determined that its annual costs are below this
threshold, EPA has chosen to publish a complete EA for this rule.
The revisions to the 1989 Total Coliform Rule (1989 TCR) are in accordance with the
Safe Drinking Water Act (SDWA) as amended, which requires EPA to review and revise, as
appropriate, each national primary drinking water regulation (NPDWR) promulgated under the
SDWA at least every six years. In the Six-Year Review determination published in July 2003,
EPA gave notice of its intent to review the 1989 TCR. EPA has since developed the proposed
RTCR in collaboration with states, other interested stakeholders, and the Total Coliform
Rule/Distribution System Advisory Committee (TCRDSAC), and then developed the final rule
after assessing the comments received on the proposal. The Agency's primary reasons for
revising the 1989 TCR are implementation-related issues. The RTCR offers a meaningful
opportunity for greater public health protection against fecal contamination and waterborne
pathogens in the distribution systems of public water systems (PWSs) beyond the 1989 TCR. As
with the 1989 TCR, the RTCR applies to all PWSs.
ES.l Need for the Rule
EPA promulgated the 1989 TCR to decrease the risk of waterborne illness. Among all
SDWA rules promulgated for preventing waterborne illness, only the 1989 TCR applies to all
PWSs, making the rule an essential component of the multi-barrier approach in public health
protection against endemic (and epidemic) disease. The objectives of the 1989 TCR are: (1) to
evaluate the effectiveness of treatment, (2) to determine the integrity of the distribution system,
and (3) to signal the possible presence of fecal contamination.
In recent years, the number of violations under the 1989 TCR has remained relatively
steady, as shown in Exhibit 4.11 and discussed in Appendix G of the RTCR EA. EPA believes
that this is reflective of a steady state among PWSs complying with the 1989 TCR;
improvements likely to occur under that rule have largely been achieved. In outlining
recommendations for further reductions in occurrence, in September 2008 EPA and the
TCRDSAC developed the Agreement in Principle (AIP), which has become the basis for the
structure of the RTCR. EPA published a proposed RTCR which was consistent with the
recommendations in the AIP. As a result of public comments received some provisions of the
RTCR were changed from the proposal. However, the final RTCR remains largely consistent
with the recommendations in the AIP.
In combination with the other SDWA rules, the RTCR will better address the 1989 TCR
objectives and enhance the multi-barrier approach to protecting public health, especially with
Economic Analysis for the Final RTCR
ES-1
September 2012
-------
respect to small ground water PWSs. The RTCR maintains the three objectives of the 1989 TCR,
but better addresses these objectives by requiring systems that may be vulnerable to fecal
contamination (as indicated by their monitoring results) to do an assessment, to identify whether
any sanitary defect(s) is (are) present, and to correct the defects. Therefore, greater public health
protection is anticipated under the RTCR compared to the 1989 TCR because of its more
preventive approach to identifying and fixing problems that affect or may affect public health.
The key provisions of the rule are summarized in Section ES.3.
Public health concerns, fecal contamination, and waterborne pathogens
SDWA section 1412(b)(9) requires that any revision to an NPDWR "shall maintain, or
provide greater, protection of the health of persons." The RTCR aims to increase public health
protection through the reduction of potential pathways of entry for fecal contamination into the
distribution system. Since these potential pathways represent vulnerabilities in the distribution
system whereby fecal contamination and/or waterborne pathogens, including bacteria, viruses,
and parasitic protozoa could possibly enter the system, the reduction of these pathways in
general should lead to reduced exposure and associated risk from the contaminants. Fecal
contamination and waterborne pathogens can cause a variety of illnesses, including acute
gastrointestinal illness (AGI) with diarrhea, abdominal discomfort, nausea, vomiting, and other
symptoms. Most AGI cases are of short duration and result in mild illness. Other more severe
illnesses caused by waterborne pathogens include hemolytic uremic syndrome (HUS) (kidney
failure), hepatitis, and bloody diarrhea (WHO, 2004). Chronic disease such as irritable bowel
syndrome, reduced kidney function, hypertension and reactive arthritis can result from infection
by a waterborne agent (Clark et al., 2008).
When humans are exposed to and infected by an enteric pathogen, the pathogen becomes
capable of reproducing in the gastrointestinal tract. As a result, healthy humans shed pathogens
in their feces for a period ranging from days to weeks. This shedding of pathogens often occurs
in the absence of any signs of clinical illness. Regardless of whether a pathogen causes clinical
illness in the person who sheds it in his or her feces, the pathogen being shed may infect other
people directly by person-to-person spread, contact with contaminated surfaces, and other means
which are referred to as secondary spread. As a result, pathogens that are initially waterborne
may subsequently infect other people through a variety of routes (WHO, 2004). Sensitive
subpopulations are at greater risk from waterborne disease than the general population (Gerba et
al., 1996).
Indicators
Total coliforms (TC) are a group of closely related bacteria that, with few exceptions, are
not harmful to humans. Coliforms are abundant in the feces of warm-blooded animals, but can
also be found in aquatic environments, in soil, and on vegetation. Coliform bacteria may be
transported to surface water by runoff or to ground water by infiltration. TC bacteria are common
in ambient water and may be injured by environmental stresses such as lack of nutrients, and
water treatments such as chlorine disinfection, in a manner similar to most bacterial pathogens
and many viral enteric pathogens (including fecal pathogens). EPA considers TC to be a useful
indicator that a potential pathway exists through which fecal contamination can enter the
distribution system. The absence (versus the presence) of TC in the distribution system indicates
Economic Analysis for the Final RTCR
ES-2
September 2012
-------
a reduced likelihood that fecal contamination and/or waterborne pathogens are occurring in the
distribution system.
Under the 1989 TCR, each total coliform-positive (TC+) sample is assayed for either
fecal coliforms oris. coli. Fecal coliform bacteria are a subgroup of TCs that traditionally have
been associated with fecal contamination. Since the promulgation of the 1989 TCR, more
information and understanding of the suitability of fecal coliform and E. coli as indicators have
become available. Study has shown that the fecal coliform assay is imprecise and too often
captures bacteria that do not originate in the human or mammal gut (Edberg et al., 2000). On the
other hand, E. coli is a more restricted group of coliform bacteria that almost always originate in
the human or animal gut (Edberg et al., 2000). Thus, E. coli is a better indicator of fecal
contamination than fecal coliforms.
Presence of fecal contamination
EPA believes that E. coli is a meaningful indicator for fecal contamination and the
potential presence of associated pathogen occurrence. Fecal contamination is a very general term
that includes all of the organisms found in feces, both pathogenic and nonpathogenic. Fecal
contamination can occur in drinking water both through use of contaminated source water as
well as direct intrusion of fecal contamination into the drinking water distribution system.
Biofilms in distribution systems may harbor waterborne bacterial pathogens and accumulate
enteric viruses and parasitic protozoa (Skraber et al., 2005; Helmi et al., 2008). Waterborne
pathogens in biofilms may have entered the distribution system as fecal contamination from
humans or animals.
Co-occurrence of indicators and waterborne pathogens is difficult to measure. The
analytical methods approved by EPA to assay for E. coli do not specifically identify most of the
pathogenic E. coli strains. There are at least 700 recognized E. coli strains (Kaper et al., 2004).
About 10 percent of recognized E. coli strains are pathogenic to humans (Feng, 1995; Hussein,
2007; Kaper et al., 2004). Pathogenic E. coli include E. coli 0157:H7, which is the primary
cause of HUS in the United States (Rangel et al., 2005). The US Centers for Disease Control and
Prevention (CDC) estimates that there are 73,000 cases of illness each year in the US due to E.
coli 0157:H7 (Mead et al., 1999). The CDC estimates that about 15 percent of all reportedE.
coli 0157:H7 cases are due to water contamination (Rangel et al., 2005). Active surveillance by
CDC shows that 6.3 percent of E. coli 0157:H7 cases progress to HUS (Griffin and Tauxe,
1991; Gould et al., 2009) and about 12 percent of HUS cases result in death within four years
(Garg et al., 2003). About 4 to 15 percent of cases are transmitted within households by
secondary transmission (Parry and Salmon, 1998).
Because EPA-approved standard methods for E. coli do not typically identify the
presence of the pathogenic E. coli strains, an E. co/z'-positive monitoring result is an indicator of
fecal contamination but is not necessarily a measure of waterborne pathogen occurrence.
Specialized assays and methods are used to identify waterborne pathogens, including pathogenic
E. coli.
One notable exception is the data reported by Cooley et al. (2007), which showed high
concentrations of pathogenic E. coli strains in samples containing high concentrations of fecal
Economic Analysis for the Final RTCR
ES-3
September 2012
-------
indicator E. coli. These data are from streams and other poor quality surface waters surrounding
California spinach fields associated with the 2006 E. coli 0157:H7 foodborne outbreak. Data
equivalent to these samples are not available from drinking water samples collected under the
TCR.
Waterborne disease outbreaks
The CDC defines a waterborne disease outbreak as occurring when at least two persons
(or one with amoebic meningoencephalitis) experience a similar illness after ingesting a specific
drinking water contaminated with pathogens (or chemicals) (Kramer et al., 1996). The CDC
maintains a database on waterborne disease outbreaks in the United States. The database is based
upon responses to a voluntary and confidential survey form that is completed by state and local
public health officials.
The National Research Council strongly suggests that the number of identified and
reported outbreaks in the CDC database for surface and ground waters represents a small
percentage of actual number of waterborne disease outbreaks (NRC, 1997; Bennett et al., 1987;
Hopkins et al., 1985 for Colorado data). Underreporting occurs because most waterborne
outbreaks in community water systems are not recognized until a sizable proportion of the
population is ill (Perz et al., 1998; Craun, 1996), perhaps 1 percent to 2 percent of the population
(Craun, 1996).
EPA drinking water regulations are designed to protect against endemic waterborne
disease and to minimize waterborne outbreaks. In contrast to outbreaks, endemic disease refers
to the persistent low to moderate level or the unusual ongoing occurrence of illness in a given
population or geographic area (Craun, et al. 2006).
ES.2 Consideration of Regulatory Options
EPA evaluated the following three regulatory options as part of the regulatory
development process: (1) 1989 TCR, (2) RTCR, and (3) Alternative option. EPA discusses the
three regulatory options briefly in this executive summary and in greater detail in Chapter 3.
The first regulatory option, the 1989 TCR, reflects EPA's understanding of how the 1989
TCR (USEPA, 1989, 54 FR 27544, June 29, 1989) is currently being implemented. That is, the
1989 TCR is assumed to include "status quo" PWS and state implementation practices. The
second regulatory option, which is the preferred regulatory option, is the RTCR. The RTCR is a
revision of the 1989 TCR based on the recommendations of the advisory committee. The
provisions of the preferred regulatory option are based on the AIP and are described in detail in
section III of the preamble of the rule. The third regulatory option, the Alternative option,
parallels the RTCR in most ways but includes variations of some of the provisions that were
discussed by the advisory committee before consensus was reached on the AIP. Under the
Alternative option, at the compliance date all PWSs are required to sample monthly for an initial
period until they meet the eligibility criteria for reduced monitoring. EPA assumes that eligibility
for reduced monitoring is determined during the next sanitary survey following the RTCR
compliance date (corresponding to year 11 of the model runs presented later on in this EA). This
more stringent approach differs from the RTCR that allows PWSs to continue to monitor at their
Economic Analysis for the Final RTCR
ES-4
September 2012
-------
current frequencies (with an additional annual site visit or voluntary Level 2 assessment
requirement for PWSs wishing to remain on annual monitoring) until they are triggered into an
increased sampling frequency. Under the Alternative option, no PWSs are allowed to reduce
monitoring to an annual basis. EPA defined the Alternative option this way and included it in
this EA to assess the relative impacts of a more stringent rule and to better understand the
balance between costs and public health protection.
To understand the relative impacts of the options, EPA gathered available data and
information to develop and provide input into an occurrence and predictive model. EPA
estimated both baseline conditions and changes to these conditions anticipated to occur over time
as a result of these revised rule options. The analysis is described in more detail in the remainder
of this EA.
ES.3 Summary of the RTCR Requirements
The RTCR maintains and strengthens the objectives of the 1989 TCR and is largely
consistent with the recommendations in the AIP. The objectives are: (1) to evaluate the
effectiveness of treatment, (2) to determine the integrity of the distribution system, and (3) to
signal the possible presence of fecal contamination. The revised rule better addresses these
objectives by requiring systems that may be vulnerable to fecal contamination (as indicated by
their monitoring results) to do an assessment, to identify whether any sanitary defect(s) is (are)
present, and to correct the defects. Therefore, greater public health protection is anticipated under
the RTCR compared to the 1989 TCR because of its more preventive approach to identifying and
fixing problems that affect or may affect public health. The following is an overview of the key
provisions of the RTCR:
MCLG andMCL for E. coli and coliform treatment technique for protection against potential
fecal contamination
The RTCR establishes a maximum contaminant level goal (MCLG) and maximum
contaminant level (MCL) for is. coli. Under the RTCR there is no longer an MCL violation for
multiple TC detections. The RTCR takes a preventive approach to protecting public health by
establishing a coliform treatment technique for protection against potential fecal contamination.
The treatment technique uses both TC and E. coli monitoring results to start an evaluation
process that, where necessary, will require the PWS to conduct follow-up and corrective action
that could prevent further incidences of contamination and exposure to fecal contamination
and/or waterborne pathogens. See section III.B of the RTCR preamble (USEPA, 2010c) for a
detailed discussion on the MCLG, MCL, and coliform treatment technique requirements.
Monitoring
As with the 1989 TCR, PWSs will continue to monitor for TC and E. coli according to a
sample siting plan and schedule specific to the system.
Sample siting plans under the RTCR must continue to be representative of the water
throughout the distribution system. Under the RTCR, systems will have the flexibility to propose
repeat sample locations that will best verify and determine the extent of potential contamination
of the distribution system rather than having to sample within five connections upstream and
Economic Analysis for the Final RTCR
ES-5
September 2012
-------
downstream of the TC+ sample location. For some systems, most likely those that have limited
or no distribution systems, the repeat samples may satisfy the requirement for source water
samples. State approval is required because this constitutes a reduction in monitoring (no
separate triggered source water samples), relative to requiring separate samples for compliance
with the two rules. EPA believes that this reduction in monitoring is appropriate only if the state
determines that the dual purpose sample provides public health protection equivalent to that
provided by separate repeat and source water samples. On the other hand, EPA believes that dual
purpose samples may not be appropriate for systems with extensive distribution systems because
the reduction in monitoring (i.e., one less repeat sample in a distribution system that extends far
from the source water sample site) may not provide public health protection equivalent to taking
separate samples.
As with the 1989 TCR, the RTCR allows reduced monitoring for some small ground
water systems. The RTCR is expected to improve public health protection compared to the 1989
TCR by requiring small ground water systems that are on or wish to move to reduced monitoring
to meet certain eligibility criteria. Examples of the criteria include a sanitary survey showing that
the system is free of sanitary defects, a clean compliance history for a minimum of 12 months,
and a recurring annual site visit by the state and/or a voluntary Level 2 assessment by a party
approved by the state, or meeting criteria established by the state.
For small ground water systems, the RTCR requires increased monitoring for higher-risk
systems that meet certain criteria such as unacceptable compliance history under the RTCR. The
RTCR specifies conditions under which systems will no longer be eligible for reduced
monitoring and will therefore be required to return to routine monitoring or to monitor at an
increased frequency.
The RTCR requires systems on a quarterly or annual monitoring frequency (applicable
only to ground water systems serving 1,000 or fewer people) to conduct additional routine
monitoring the month following one or more TC+ samples. Under the RTCR, systems must
collect at least three routine samples during the next month, unless the state waives the additional
routine monitoring. This is a reduction in the required number of additional routine samples from
the 1989 TCR, which requires at least five routine samples in the month following a TC+ sample
for all systems serving 4,100 or fewer people.
The 1989 TCR requires all systems serving 1,000 or fewer people to collect at least four
repeat samples while PWSs serving greater than 1,000 people to collect three repeat samples.
The revised rule requires three repeat samples after a routine TC+ sample, regardless of the
system type and size.
Sections III.C and III.D of the RTCR preamble provide detailed discussions of the
routine monitoring and repeat sampling requirements of the RTCR.
Seasonal systems
The RTCR establishes special monitoring requirements for seasonal systems for the first
time. Seasonal systems represent a special case in that the shutdown and start-up of these water
systems present additional opportunities for contamination to enter or spread through the
Economic Analysis for the Final RTCR
ES-6
September 2012
-------
distribution system. Seasonal systems must demonstrate completion of a state-approved start-up
procedure. In addition, they must designate the time period(s) for monitoring based on site-
specific considerations (such as during periods of highest demand or highest vulnerability to
contamination) in their state-reviewed sample siting plan. Section Ill.C.l.f of the RTCR
preamble provides a detailed discussion of seasonal systems.
Assessment and corrective action
As part of a treatment technique, all PWSs are required to assess their systems when
monitoring results show that the system may be vulnerable to contamination. Systems must
conduct a simple self-assessment (Level 1) or a more detailed assessment (Level 2) depending on
the level of concern raised by the results of the indicator sampling. The system is responsible for
correcting any sanitary defect(s) found through either a Level 1 or Level 2 assessment. Section
III.E of the RTCR preamble provides more discussion of the treatment technique requirement of
the RTCR.
Violations and public notification
The RTCR establishes an E. coli MCL violation, a treatment technique violation, a
monitoring violation, and a reporting violation. Public notification (PN) is required for each type
of violation, depending on the degree of potential public health concern consistent with EPA's
current PN requirements under 40 CFR part 141, subpart Q. The RTCR also modifies the PN and
consumer confidence report language to reflect the construct of the revised rule and the role of
TC as an indicator of a potential pathway for the contamination of the distribution system.
Sections III.F and III.G of the RTCR preamble provide detailed discussions of violations and PN
under the RTCR.
Transition to the RTCR
The RTCR allows all systems to transition to the revised rule at their 1989 TCR
monitoring frequency, including systems on reduced monitoring under the 1989 TCR. States will
then evaluate whether the system is on an appropriate monitoring schedule by performing a
special monitoring evaluation during each sanitary survey to review the status of the system,
including the distribution system. The first such evaluation will be conducted during the first
scheduled sanitary survey after the effective date of the rule; a system may remain on its 1989
TCR monitoring schedule until this time unless it is triggered into more frequent monitoring.
After its first evaluation, the state may allow the system to remain on its 1989 TCR monitoring
schedule as long as the system meets the conditions for doing so. Initial grandfathering of
monitoring frequencies reduces state burden by not requiring the state to determine appropriate
monitoring frequency at the same time the state is working to adopt primacy, develop policies,
and train their own staff and the PWSs in the state.
Economic Analysis for the Final RTCR
ES-7
September 2012
-------
Exhibit ES. 1 shows the implementation schedule, from year 1 through year 10, for the
RTCR. PWSs will begin sampling under the RTCR in the fourth year after RTCR promulgation.1
ES.4 Systems Subject to the RTCR
The RTCR will apply to all PWSs in the United States. The baseline inventory for the EA
is derived from EPA's Safe Drinking Water Information System (SDWIS), fourth quarter 2007
data (USEPA, 2007b), which contains information reported by primacy agencies on the
approximately 155,000 active PWSs. The systems are subdivided by water source (ground water
or surface water), type (community water systems (CWS), nontransient noncommunity water
systems (NTNCWS), and transient noncommunity water systems (TNCWS)), and size (nine size
categories ranging from those serving 100 or fewer people to those serving more than 1 million
people).
EPA developed the baseline water quality used in the RTCR EA from the Six-Year
Review 2 data collected between 1998 and 2005, which include coliform monitoring data
voluntarily provided to EPA by 37 primacy agencies (35 states and 2 tribes). The database
consists of over nine million 1989 TCR records collected between 1998 and 2005. In addition,
EPA incorporated information from the GWR, the U.S. Census, and the 1989 TCR to develop
assumptions for the model.
EPA used the available data and the requirements of the regulatory options considered to
develop a model that predicts the number of systems in each category that experience TC+ and
E. coli+ assays over the 25-year period of analysis. The results of the occurrence model and cost
model indicate the number of systems that conduct the various activities under three regulatory
options (1989 TCR, RTCR, and Alternative option). These results are discussed in terms of
benefits of the current rule (changes in E. coli MCL violations incurred and corrective actions to
be implemented) and rule costs in Section ES.5 below.
ES.5 National Benefits and Costs of the RTCR
The consensus resulting from TCRDSAC deliberations was that an RTCR consistent with
the AIP would achieve a net risk reduction compared to the 1989 TCR. The committee applied
best professional judgment in determining that the increased protection provided by the new
requirements for implementing focused assessments and appropriate corrective actions would
more than offset any potential increase in risk introduced by the reduction in samples or other
changes resulting from the RTCR.
1 Chapter 4 of the RTCR EA describes the baseline schedule under which systems would begin sampling following
the RTCR effective date; Chapter 5 describes how the baseline schedule is adjusted, based on acute and non-acute
violations incurred during an initial assessment period, to determine the steady state distribution of systems on
monthly, quarterly, and annual sampling.
2 SDWIS/Federal Version (SD WIS/FED) is a database created by EPA containing data submitted by States and
regions regarding inventory as well as compliance with SDWA regulations.
Economic Analysis for the Final RTCR
ES-8
September 2012
-------
Exhibit ES.1 Implementation Schedule
Year
Year 1
Year 2
Year 3
Year 4
Year 5
Year 6
Year 7
Year 8
Year 9
Year 10
State
Implementation
Revising Sample Siting
Plan
Routine Monitoring, Additional Routine Monitoring, Repeat Monitoring
Annual Site Visits
Level 1 and Level 2 Assessments
Correction Actions Based on Level 1 and Level 2 Assessments
Public Notification
PWS
Implementation
Revising Sample Siting
Plan
Routine Monitoring, Additional Routine Monitoring, Repeat Monitoring
Annual Site Visits
Level 1 and Level 2 Assessments
Correction Actions Based on Level 1 and Level 2 Assessments
Public Notification
Source: Final Information Collection Request for the Revised Total Coliform Rule, Figure 4.1.
Note: Activities occurring in Year 10 continue throughout the remaining years of analysis
Economic Analysis for the Final RTCR
ES-9
September 2012
-------
Based on limitations in available data (described further in Chapter 6, Section 6.4), EPA
determined that benefits could not be calculated in terms of avoided cases of (or costs related to)
morbidity or mortality. Since E. coli is an indicator of fecal contamination, EPA assumed that a
decrease in E. coli occurrence in the distribution system would be associated with a decrease in
fecal contamination in the distribution system. In general, this decrease in fecal contamination
should reduce the potential risk to human health for PWS customers. Thus, any reduction in E.
coli occurrence is considered a benefit of the RTCR. EPA's qualitative analysis in Section 6.2 of
this EA considers the overall changes in occurrence of contaminant indicators under the RTCR
as compared to the 1989 TCR, and predicts a net decrease in risk. As discussed in Chapter 6, this
reduction in risk is an indicator of the benefits of the RTCR.
Sections ES.5.1 and ES.5.2 summarize the qualitative benefits analysis and the methods
used to derive the costs of the rule, respectively. Section ES.5.1 describes the qualitative analyses
of net changes in TC and E. coli occurrence anticipated to occur under the RTCR compared to
the 1989 TCR. As found in Section ES.5.2, EPA's national cost estimates include cost to
implement the rule; revise sample siting plans; conduct routine monitoring, additional routine
monitoring, and repeat monitoring; perform Level 1 and Level 2 assessments; implement
corrective actions; and provide PN in the case of violations. Estimates for present value and
annualized national costs are presented using a 3 percent and 7 percent discount rate. Chapters 6
(benefits), 7 (costs), and 9 (comparison of benefits and costs), as well as the appendices, provide
a more detailed discussion of all the analyses discussed in the sections below.
Changes in risk associated with RTCR activities are characterized by their anticipated
effects on potential pathways of contamination into PWS, as indicated by their effects on such
pathways for TC IE. coli. These activities are considered under each rule component presented in
Exhibit ES.3.
ES.5.1 Derivation of Benefits
In promulgating the RTCR, EPA expects to further reduce the risk of contamination of
public drinking water from the current baseline risk under the 1989 TCR. The RTCR and
Alternative option considered during development of this rule and analyzed as part of this EA
(see Chapters 6, 7, and 9) are designed to achieve this reduction while maintaining public health
protection in a cost-effective manner.
The EA examines the benefits in terms of tradeoffs between compliance with the 1989
TCR and the other options considered (RTCR and Alternative option). Because there are
insufficient data reporting the co-occurrence in a single sample of fecal indicator E. coli and
pathogenic organisms and because the available fecal indicator E. coli data from the Six-Year
Review 2 dataset were limited to presence-absence data, EPA was unable to quantify health
benefits for the RTCR. EPA used several methods to qualitatively evaluate the benefits of the
3
There is much discussion among economists of the proper social discount rate to use for policy analysis.
For RTCR cost analyses, calculations are made using two social discount rates (3 and 7 percent) thought to best
represent current policy evaluation methodologies. Historically, the use of 3 percent is based on rates of return on
relatively risk-free investments, as described in the Guidelines for Preparing Economic Analyses (USEPA, 2000b).
The rate of 7 percent is a recommendation of the Office of Management and Budget (OMB) as an estimate of
"before-tax rate of return to incremental private investment" (OMB, 1996).
Economic Analysis for the Final RTCR
ES-10
September 2012
-------
RTCR and Alternative option. The qualitative evaluation uses both the judgment of EPA as
informed by the TCRDSAC deliberations as well as quantitative estimates of changes in TC
occurrence and counts of systems implementing corrective actions. The evaluation characterizes,
in relative terms, the reduction in risk for each regulatory option as compared to baseline
conditions.
As presented in the description of the baseline (see Chapter 4), the percentages of
samples that are positive TC and E. coli are generally higher for PWSs serving 4,100 or fewer
people than those serving more than 4,100 people. PWSs with higher TC and E. coli occurrence
are more likely to be triggered into assessments and corrective action. As discussed previously,
EPA believes that the assessments and corrective action will lead to a decrease in TC and E. coli
occurrence. Because the PWSs serving 4,100 or fewer people have a higher initial E. coli
occurrence and will be triggered into more assessments and corrective actions than larger PWSs,
the increase in benefits for these small systems will likely be more evident as compared to the
larger systems. In particular, model results suggest that customers of small ground water
TNCWSs serving 100 or fewer people, which constitute approximately 40 percent of PWSs,
experience the most improvement in water quality under the RTCR. That is, the occurrence of E.
coli is predicted to decrease more for these systems that for other systems types.
When revising an existing drinking water regulation, one of the main concerns is to
ensure that backsliding on water quality and public health protection does not occur. SDWA
requires that EPA at least maintain or improve public health protection for any rule revision.
EPA believes that the RTCR is more stringent than the 1989 TCR with regard to protecting
public health. The basis for this perspective is provided in chapters 6 and 9 of this EA.
Risk reduction for the RTCR is characterized by the activities performed that are
presumed to reduce risk of exposing the public to contaminated water. These activities are
considered under each rule component presented in Exhibit ES.3.
More frequent monitoring has the potential to decrease the risk of contamination in PWSs
based on an enhanced ability to diagnose and mitigate system issues in a more timely fashion.
Conversely, less frequent monitoring has the potential to increase risk. Real-time continuous
sampling would mitigate the most risk possible based on sampling schedule; however, it would
cost prohibitively more than the periodic sampling practiced under the 1989 TCR and included in
the RTCR and the Alternative option. EPA's objective in revising the sampling schedules
included in the RTCR and Alternative option was to find an appropriate balance between the
factors of risk mitigation and cost management.
Under the RTCR and Alternative option, the reduction in the number of repeat samples
and additional routine samples for some PWSs has the potential to contribute to increased risk
for PWS customers. However, this increase in risk is expected to be more than offset by potential
decreases in risk from increased routine samples and the addition of the assessments and
corrective action provisions that find and fix problems indicated by monitoring. Exhibit ES.2
illustrates the predicted reduced frequency at which TC occur subsequent to the implementation
of the RTCR and Alternative option. As discussed previously, the RTCR uses TC and E. coli
occurrence as an indicator of potential pathways for possible contamination to enter the
distribution system. Exhibit ES.2 illustrates the combined effects on TC occurrence resulting
Economic Analysis for the Final RTCR
ES-11
September 2012
-------
from changes in monitoring and the effects of assessments and corrective actions for the different
rule options illustrated. The relative trends indicated in Exhibit ES.2 for small transient
noncommunity ground water systems also pertain to other PWS categories as illustrated in
Chapter 5. EPA chose to include the characterization for this set of TNCWSs because they
represent the system category of largest influence on the national impacts.
Exhibit ES.2 Modeled TC+ Occurrence in Transient Noncommunity Ground Water
Systems Serving < 4,100
0.045
0.040
> 0.035
C0
>
(/>
< 0.030
I 0.025
i
'<75
£ 0.020
4-
o
£
O
¦4-J
1989 TCR no GVv'R
1989 TCR
0.015
RTCR
O
A3
Alt Option
i-
LL.
0.010
0.005
0.000
0
5
10
15
20
25
30
35
Years
Notes:
1) Six Year 2005 TC+ occurrence is representative of all GWTNCWS. The rate presented may
underestimate the occurrence for systems serving 25-4,100 individuals.
2) Graph shows the 30-year modeled period discussed in Ch. 5. Model years 3-27 represent the 25-year
period of analysis for this EA. Model year 11 begins the steady state, during which systems that qualified for
reduced monitoring are now sampling on their reduced schedules. The criteria and timing of this monitoring
adjustment is discussed in Section 5.3.2.2 ofthis EA.
3) The results represented by the curves for 1989 TCR and RTCR and Alternative options all incorporate the
effects of the GWR.
Economic Analysis for the Final RTCR
ES-12
September 2012
-------
Exhibit ES.3 Potential Changes in Risk under the RTCR and Alternative Option Relative to the 1989 TCR
Rule Component
Factors Leading to a Potential
Increase in Risk
Factors Leading to a Potential Decrease in
Risk
Overall Predicted Change in
Risk
RTCR
Alternative
option
RTCR
Alternative
option
RTCR
Alternative
option
Implementation
Activities
None
None
None
None
No change
No change
Routine
Monitoring
(Including
Reduced
Monitoring)
None
None
Increased stringency
in requirements to
qualify for reduced
monitoring along with
requirement to return
to baseline monitoring
upon loss of these
criteria is expected to
result in decreased
risk (i.e., PWSs that
qualify for reduced
monitoring will be
better operated, PWSs
that no longer qualify
will monitor more
frequently).
PWSs all monitor monthly
in the first few years of
implementation of the
RTCR, which is an
increase in sampling
frequency for systems
that monitor quarterly or
annually under the 1989
TCR. After the first few
years, systems may
reduce to quarterly, but
none may reduce to
annual monitoring,
creating a decrease in risk
for systems on annual
monitoring under the 1989
TCR.
Decrease
Decrease
Repeat
Monitoring
Required repeat
samples reduced
from 4 to 3 for
systems serving
<1,000 people
Required repeat
samples reduced
from 4 to 3 for
systems serving
<1,000 people
None
None
Increase
Increase
Additional
Routine
Monitoring
Additional routine
samples are no
longer required for
PWSs monitoring
monthly.
Additional routine
samples are no
longer required
for PWSs
monitoring
monthly.
None
None
Increase
Increase
Economic Analysis for the Final RTCR
ES-13
September 2012
-------
Rule Component
Factors Leading to a Potential
Increase in Risk
Factors Leading to a Potential Decrease in
Risk
Overall Predicted Change in
Risk
RTCR
Alternative
option
RTCR
Alternative
option
RTCR
Alternative
option
GW PWSs serving
<1,000 people
would reduce
additional routine
samples from 5 to
3.
GW PWSs
serving <1,000
people would
reduce additional
routine samples
from 5 to 3.
Annual Site Visits
None (only states
currently
performing annual
site visits are
expected to
continue)
Annual
monitoring is not
permitted under
the Alternative
option so annual
site visits will no
longer be
conducted.
None (only states
currently performing
annual site visits are
expected to continue)
None
No change
Increase
Assessments
None
None
Mandatory
assessments are a
new requirement.
Mandatory assessments
are a new requirement.
Decrease
Decrease
Corrective
Actions
None
None
Mandatory corrective
actions are a new
requirement.
Mandatory corrective
actions are a new
requirement.
Decrease
Decrease
Public
Notification—
Monthly/Non-
Acute MCL
Violations
Reduction in
available public
information
Possible PWS
complacency
Reduction in
available public
information
Possible PWS
complacency
Less confusion (PN
more in line with
potential health risks)
PWS resources used
more efficiently
Less confusion (PN more
in line with potential
health risks)
PWS resources used
more efficiently
Unknown
Unknown
Public
Notification—
Monitoring and
Reporting
Violations
None
None
Improved focus of
required PN for rule
aspects with potential
adverse health
consequences,
notably E. coli MCL
Improved focus of
required PN for rule
aspects with potential
adverse health
consequences, notably E.
coli MCL violations and
Decrease
Decrease
Economic Analysis for the Final RTCR
ES-14
September 2012
-------
Rule Component
Factors Leading to a Potential
Increase in Risk
Factors Leading to a Potential Decrease in
Risk
Overall Predicted Change in
Risk
RTCR
Alternative
option
RTCR
Alternative
option
RTCR
Alternative
option
violations and failure
to conduct
assessments and
corrective actions, will
motivate PWSs to
conduct sampling and
other treatment
technique
requirements.
failure to conduct
assessments and
corrective actions, will
motivate PWSs to
conduct sampling and
other treatment technique
requirements.
Overall
Decrease
Decrease
Note:
1) Detailed discussion of the rationale for determinations of potential risk for each rule component is presented in Ch. 6 (Section 6.2) of this EA. Implementation
activities consist of administrative activities by PWSs and states to implement the rule.
2) Assessment of potential changes in risk for monitoring components is an overall assessment. Potential changes (or static state) of risk for particular system
sizes and types differ according to individual regulatory requirements and are discussed in Section 6.2. Chapter 3 provides a detailed description of the
regulatory components for all three regulatory options, and the preamble to the RTCR provides additional discussion of the TCRDSAC process and the rationale
underlying the structure of the regulatory options considered.
Economic Analysis for the Final RTCR
ES-15
September 2012
-------
The effect that the changes to PN requirements under the 1989 TCR for monthly/non-
acute MCL violations have on risk is difficult to predict. Some factors, such as reduction in
available public information and possible PWS complacency, lead to a potential increase in risk
and other factors, such as less confusion (PN more in line with potential health risks) and PWSs
resources used more efficiently, lead to a potential decrease, as discussed in Chapter 9. This
change to PN is addressing a key concern expressed by various stakeholders in the advisory
committee and during the Six-Year Review 1 comment solicitation process. By eliminating the
requirement and replacing it with assessment and corrective action requirements, the Agency
expects less public confusion, more effective use of resources, increased transparency, and
increased public health protection. Other rule components are expected to have a negligible
effect on risk. However, the overall effect of the RTCR is expected to be a further reduction in
risk from the current baseline risk under the 1989 TCR. Chapter 6 presents a detailed discussion
of the potential influence on health risk for each rule component.
Although a qualitative analysis is the primary method employed for analyzing potential
changes in risk, some quantitative measures are considered in the overall assessment of the
RTCR, as presented in Chapter 9 of the RTCR EA. In particular, EPA developed a model to
describe TC and E. coli occurrence in water systems for the 1989 TCR (baseline for this EA) and
the RTCR and Alternative option. The model generates estimates of reduced TC and E. coli
occurrence based on requirements of the RTCR and Alternative option to perform assessments
and corrective actions not explicitly required in the 1989 TCR. In addition, the model takes into
account in the baseline for the 1989 TCR the reductions attributable to the implementation of the
GWR, which was effective as of December 2009. The predictive occurrence model and analyses
results are discussed further in Chapters 5, 6, and 9.
Exhibit ES.4 presents the annual number of acute violations (1989 TCR) and E. coli
MCL violations (RTCR and Alternative option) as predicted by the RTCR occurrence model in
Year 9. By Year 9, PWSs that are expected to meet the criteria for reduced monitoring begin
reduced monitoring, and the distribution of PWSs that monitor monthly, quarterly, and annually
is assumed to remain relatively constant. EPA did not quantify changes in violation or trigger
rates for systems serving more than 4,100 people among the 1989 TCR, RTCR, and Alternative
option because of: (1) limited Six-Year Review data to characterize these systems, (2) the
essentially unchanged monitoring requirements across options for these systems, and (3) the
level of effort already occurring to implement the 1989 TCR.
The estimates of E. coli MCL violations have two major drivers: the total number of
samples taken over time (including routine, additional routine, and repeat) and the impact of
corrective actions taken. When looking at the comparisons between the 1989 TCR with the
RTCR across all PWSs, the overall impact of the total numbers of samples taken is negligible
because the total number of samples predicted to be taken throughout the period of analysis is
almost the same (approximately 82 million samples) under both the 1989 TCR and the RTCR.
For the Alternative option, the analysis predicts that 88 million total samples will be taken over
the period of analysis.
The changes in the steady state estimates of annual E. coli MCL violations from the 1989
TCR to the RTCR and Alternative option are shown in Exhibit 6.5. The steady state in the model
refers to the period beginning in years 7 (CWSs) and 9 (NCWSs) following promulgation, after
Economic Analysis for the Final RTCR
ES-16
September 2012
-------
the proportions of systems sampling on monthly, quarterly, or annual regimens are adjusted
following a period of assessment.4 Systems that qualify for reduced monitoring will begin their
new regimens in years 7 and 9 after promulgation, respectively, for CWSs and NCWSs. The
annual number of acute violations (1989 TCR) and E. coli MCL violations (RTCR and
Alternative option) shown in Exhibit ES.4 for small systems are from the predictive model and
reflect the estimates for Year 9 of the 25-year period of analysis.
The steady state reductions in the number of annual E. coli MCL violations found under
the RTCR and Alternative option primarily reflect the benefits of corrective actions under these
two options in preventing many of the E. coli MCL violations that would otherwise occur over
this period. Under the RTCR, reductions (or no increases) are predicted for all PWS sizes and
types, while under the Alternative option, reductions (or no increases) are predicted for all PWSs
except TNCWSs serving <1,000 people. For these small TNCWSs under the Alternative option,
increased monitoring is expected to lead to an overall increase in annual E. coli MCL violations
(and is also the driver of the greater total number of annual E. coli MCL violations predicted).
The interplay of the reducing effect on E. coli MCL violations that required corrective actions
induce and the increasing effect from reductions in some monitoring requirements are explored
in Section 6.4 of this EA. The step-wise analysis presented in Section 6.4 shows howis. coli
MCL violations could increase under the Alternative option and concludes that, on balance, more
E. coli MCL events are prevented than are missed.
4 The effective date of the RTCR occurs after 3 years of implementation, at the start of year 4 post-promulgation.
For CWSs, years 4-6 post-promulgation are the period of assessment for considering potential to move to reduced
monitoring for systems; for NCWSs, years 4-8 are the period of assessment.
Economic Analysis for the Final RTCR
ES-17
September 2012
-------
Exhibit ES.4 Predicted Annual Number of Acute (1989 TCR) or E. coli MCL
(RTCR and Alternative Option) Violations by Regulatory Option and System Type
(Year 9)1
PWS Size
Number of
Alternative
(Population
Systems
1989 TCR
RTCR
Option
Served)
A
B
C
D
Community Water Systems (CWSs) - SW
<100
1,170
7
5
5
101 -500
2,150
7
8
8
501-1,000
1,173
5
4
4
1,001-4,100
2,938
8
7
7
4,101-33,000
3,164
9
9
9
33,001-96,000
720
3
3
3
96,001-500,000
308
1
1
1
500,001-1 Mllion
31
> 1 Million
17
Totals
11,671
39
36
36
Community Water Systems (CWSs) - GW
<100
11,938
58
41
40
101 -500
13,892
40
35
36
501-1,000
4,467
13
7
8
1,001-4,100
6,443
26
15
15
4,101-33,000
3,156
12
12
12
33,001-96,000
335
2
2
2
96,001-500,000
63
0
0
0
500,001-1 Mllion
4
> 1 Million
3
Totals
40,301
151
113
114
Nontransient Noncommunity Water Systems (NTNCWSs) - SW
<100
250
2
1
1
101 -500
253
2
1
1
501-1,000
88
0
0
0
1,001-4,100
72
1
1
1
4,101-33,000
22
33,001-96,000
2
96,001-500,000
1
500,001-1 Mllion
> 1 Million
Totals
688
5
3
3
Nontransient Noncommunity Water Systems (NTNCWSs) - GW
<100
8,826
43
25
22
101 -500
6,613
19
13
16
501-1,000
1,718
4
3
4
1,001-4,100
812
8
4
4
4,101-33,000
70
0
0
0
33,001-96,000
2
96,001-500,000
500,001-1 Mllion
> 1 Million
Totals
18,041
74
46
47
Transient Noncommunity Water Systems (TNCWSs) - SW
<100
1,339
18
11
11
101 -500
497
7
3
3
501-1,000
88
1
1
1
1,001-4,100
67
2
1
1
4,101-33,000
18
33,001-96,000
96,001-500,000
500,001-1 Mllion
> 1 Million
1
Totals
2,010
29
16
16
Transient Noncommunity Water Systems (TNCWSs) - GW
<100
60,200
339
237
279
101 -500
19,275
96
69
95
501-1,000
1,963
11
9
8
1,001-4,100
617
8
5
5
4,101-33,000
67
0
0
0
33,001-96,000
2
96,001-500,000
1
500,001-1 Mllion
1
> 1 Million
Totals
82,126
454
321
388
Grand Total
154,837
751
535
605
Source: Output from RTCR models (described in Sections 5.3 and 5.4 of this EA).
Notes:
1) For modeling purposes, additional routine sample counts include regular routine
samples taken in the same month.
2) Detail may not add due to independent rounding.
Economic Analysis for the Final RTCR
ES-18
September 2012
-------
ES.5.2 Derivation of Costs
Costs are estimated for different PWS types and size categories (nine size categories are
used based on population served) using unit costs obtained from the advisory committee
technical work group and vendors and presented in the Technology and Cost Document for the
Final Revised Total Coliform Rule (USEPA, 2010d). Cost analyses for PWSs include estimates
to implement the rule; to revise sample siting plans; to conduct routine monitoring, additional
routine monitoring, and repeat monitoring; to perform Level 1 and Level 2 assessments and
implement corrective actions; and to provide PN in the case of violations. State cost analyses
include estimates of the labor burdens that states would incur, including staff training on RTCR
requirements, conducting annual administration, reviewing monitoring reports, reviewing and
approving corrective action plans, and for recordkeeping. Chapter 7 of this EA provides detailed
discussion on the underlying cost-buildup for each rule component included in the cost model.
National costs are estimated using a cost model specifically developed for the RTCR. The
model builds on the occurrence model described in Chapter 5. Within the modeling structure,
costs for PWSs serving more than 4,100 retail customers are analyzed differently from smaller
PWSs to capture differing baseline structures and to account for differences in available
occurrence data as described in Chapter 4 of this EA. The resulting national cost estimates for
the RTCR regulatory options considered are shown in Exhibit ES.5. To evaluate the impact of
costs under the RTCR, emphasis in this EA is placed on the net changes in costs of the RTCR or
Alternative option compared to the 1989 TCR (i.e., incremental costs over the 1989 TCR), also
shown in Exhibit ES.5.
As noted throughout the RTCR EA, there is variability among many of the input
parameters to the cost model and several rule compliance assumptions based on PWS size and
type (e.g., population served, labor rates, TC hit rates, and occurrence distributions are different
for different sizes and types of PWSs). However, there is insufficient information to fully
characterize the distribution of variability (i.e., calculating confidence bounds) within each of
these PWS classifications on a national scale; therefore, EPA uses mean values for these input
parameters.
EPA also recognizes that there is uncertainty in the national cost estimates. Many of the
uncertainties have the same impact on both the 1989 TCR and RTCR and Alternative options
(e.g., baseline assumptions and effects of GWR implementation). Because the EA analyses focus
on net changes between the 1989 TCR and the RTCR and Alternative option, these common
sources of uncertainty do not impact conclusions based on the net change analyses. For
assumptions that are major drivers of the analyses and differ between the 1989 TCR and RTCR
and Alternative option (e.g., corrective action compliance forecast), EPA has evaluated
uncertainty and performed sensitivity analyses to qualitatively and quantitatively characterize the
potential impacts of alternative input parameters. Chapters 4 and 5 present a comprehensive
discussion of factors contributing uncertainty to the occurrence analysis. Chapter 5 presents a
sensitivity analysis pertaining to the predictive occurrence model results, which also impact the
cost calculations. Section 7.7 of the EA discusses uncertainty and provides sensitivity analysis
results as they specifically pertain to the cost analyses.
Economic Analysis for the Final RTCR
ES-19
September 2012
-------
Exhibit ES.5 Comparison of Total and Net Change from 1989 TCR in Annualized
Present Value Costs ($Millions, 2007$)
PWSs
State
Total
PWSs
State
Total
3% Discount Rate
7% Discount Rate
1989 TCR-Total
$ 185
$ 0.9
$ 186
$ 178
$ 0.9
$ 179
RTCR-Total
$ 199
$ 1.1
$ 200
$ 192
$ 1.3
$ 193
RTCR - Net Change
$ 14
$ 0.1
$ 14
$ 14
$ 0.4
$ 14
RTCR - Percent Change
8%
16%
8%
8%
48%
8%
Alternative Option - Total
$ 214
$ 1.2
$ 216
$ 209
$ 1.5
$ 210
Alternative Option - Net Change
$ 29
$ 0.3
$ 30
$ 31
$ 0.6
$ 32
Alternative Option - Percent Change
16%
34%
16%
17%
69%
18%
Source: RTCR cost model.
Notes:
1) Detail may not add due to independent rounding.
2) Annualized costs for the state are greater using the 7% than 3% discount rate for the RTCR and Alternative
option. This occurs because the costs under the RTCR and Alternative option are frontloaded in the 25-year time
period due to the all-monthly sampling requirement (which doesn't occur underthe 1989 TCR). Discounting a
given stream using 3% and 7% will always result in a higher present value using 3%; annualization of an
identical value (any value) using 3% and 7% would result in a higher value under 7%. Depending on how costs
accrue over the period, and how long the period is, one effect will be stronger than the other. Generally, the
discounting effect is stronger, resulting in the pattern commonly seen where the 3% annualized amount is
greater than the 7% one. However, in cases like the Alternative option where the costs accrue faster early in the
time period, the annualization effect can more than compensate for the discounting effect, resulting in a higher
annualized value under 7% compared to 3%.
ES.6 Projected Impacts on Household Costs
The household cost analysis considers the impact that the costs incurred by CWSs have
on the households they serve. This analysis considers the potential increase in a household's
annual water bill if a CWS passed the entire cost increase resulting from the rule on to their
customers. This analysis is a tool to gauge potential impacts and should not be construed as a
precise estimate of potential changes to household water bills. State costs and costs to TNCWSs
and NTNCWSs are not included in this analysis since their costs are not typically passed through
directly to households. Exhibit ES.6 presents the mean expected increases in annual household
costs for all CWSs, including those systems that do not have to take corrective action. Exhibit
ES.6 also presents the same information for CWSs that must take corrective action. Household
costs tend to decrease as system size increases, due mainly to the economies of scale for the
corrective actions.
The first section of Exhibit ES.6 presents net costs per household under the RTCR and
Alternative option for all rule components spread across all CWSs. In this scenario, comparison
to the 1989 TCR shows a cost savings for households in the largest size category. For those
households that are expected to see a cost increase, the average annual water bill would be
expected to increase by less than ten cents on average.
While the average increase in annual household water bills to implement the RTCR is
less than a dollar, customers served by a small CWS that have to take corrective actions as a
result of the rule would incur slightly larger increases in their water bills. The subsequent
sections of the exhibit present net costs per household for three different subsets of CWSs
(CWSs that perform assessments but no corrective actions, CWSs that do perform corrective
Economic Analysis for the Final RTCR
ES-20
September 2012
-------
actions, and CWSs that do not perform assessments or corrective actions). As shown in the
second section of Exhibit ES.6, approximately 67% of households are served by CWSs that
would perform assessments but would not perform corrective actions (because no sanitary
defects are found). The 9% of households belonging to CWSs that would perform corrective
actions would experience an increase in annual net household costs of less than $1 on average for
CWSs serving >4,100 people to approximately $26 for CWSs serving 100 people or fewer.
The final section of Exhibit ES.6 presents the 24% of households belonging to CWSs that
would not perform assessments or corrective actions. Households of this category would
experience an increase in cost savings when compared to those performing corrective actions,
and a decrease in cost savings when compared to those performing assessments but no corrective
actions. This decrease in costs savings is because no PN costs are associated with systems not
performing assessments. Overall, the main driver of additional household costs under the RTCR
is whether or not additional corrective actions are performed.
Economic Analysis for the Final RTCR
ES-21
September 2012
-------
Exhibit ES.6 Summary of Net Annual Per-Household Costs for the RTCR (2007$)
PWS Size
(Population
Served)
Number of
Households
(RTCR)
Number of
Households
(Alternative Option)
3% Discount Rate
7% Discount Rate
RTCR Net
RTCR Net Cost per
Household
Alternative
Option Net
Alternative
Option Net
Cost per
Household
RTCR Net
RTCR Net
Cost per
Household
Alternative
Option Net
Alternative
Option Net
Cost per
Household
A
B
C
D=C/A
E
^E/B
G
H=G/A
I
J=l/B
All Community Water Systems (CWSs)
<100
307,243
307,243
$ 111,694
$ 0.364
$ 225,693
$ 0.735
$ 207,140
$ 0.674
$ 348,782
$ 1.135
101-500
1,589,510
1,589,510
$ 371,004
$ 0.233
$ 464,207
$ 0.292
$ 471,664
$ 0.297
$ 598,244
$ 0.376
501-1,000
1,624,853
1,624,853
$ 46,687
$ 0.029
$ 87,294
$ 0.054
$ 96,473
$ 0.059
$ 148,274
$ 0.091
1,001-4,100
7,816,592
7,816,592
$ 371,294
$ 0.048
$ 371,294
$ 0.048
$ 418,935
$ 0.054
$ 418,935
$ 0.054
4,101-33,000
27,997,647
27,997,647
$ 2,385,056
$ 0.085
$ 2,385,056
$ 0.085
$ 2,083,266
$ 0.074
$ 2,083,266
$ 0.074
33,001-96,000
21,933,438
21,933,438
$ 1,532,410
$ 0.070
$ 1,532,410
$ 0.070
$ 1,273,202
$ 0.058
$ 1,273,202
$ 0.058
96,001-500,000
26,770,609
26,770,609
$ 1,479,280
$ 0.055
$ 1,479,280
$ 0.055
$ 1,214,316
$ 0.045
$ 1,214,316
$ 0.045
500,001-1 Million
9,764,979
9,764,979
$ 157,138
$ 0.016
$ 157,138
$ 0.016
$ 125,671
$ 0.013
$ 125,671
$ 0.013
> 1 Million
16,309,853
16,309,853
$ (2,223)
$ (0.000)
$ (2,223)
$ (0.000)
$ (1,479)
$ (0.000)
$ (1,479)
$ (0.000)
Total
114,114,724
114,114,724
$ 6,452,342
$ 0.057
$ 6,700,151
$ 0.059
$ 5,889,190
$ 0.052
$ 6,209,211
$ 0.054
Community Water Systems (CWSs) performing Level 1/Level 2 Assessments (and no Corrective Actions
<100
125,340
124,920
$ (185,923)
$ (1.483)
$ (149,283)
$ (1.195)
$ (125,037)
$ (0.998)
$ (75,885)
$ (0.607)
101-500
460,577
464,568
$ (137,193)
$ (0.298)
$ (103,168)
$ (0.222)
$ (82,657)
$ (0.179)
$ (40,057)
$ (0.086)
501-1,000
394,643
401,009
$ (133,453)
$ (0.338)
$ (123,890)
$ (0.309)
$ (111,147)
$ (0.282)
$ (98,598)
$ (0.246)
1,001-4,100
2,341,578
2,341,578
$ (262,222)
$ (0.112)
$ (262,222)
$ (0.112)
$ (213,247)
$ (0.091)
$ (213,247)
$ (0.091)
4,101-33,000
24,827,588
24,827,588
$ (195,108)
$ (0.008)
$ (195,108)
$ (0.008)
$ (76,420)
$ (0.003)
$ (76,420)
$ (0.003)
33,001-96,000
19,232,570
19,232,570
$ (173,238)
$ (0.009)
$ (173,238)
$ (0.009)
$ (142,573)
$ (0.007)
$ (142,573)
$ (0.007)
96,001-500,000
23,912,325
23,912,325
$ (146,682)
$ (0.006)
$ (146,682)
$ (0.006)
$ (131,967)
$ (0.006)
$ (131,967)
$ (0.006)
500,001-1 Million
5,524,188
5,524,188
$ (40,160)
$ (0.007)
$ (40,160)
$ (0.007)
$ (36,993)
$ (0.007)
$ (36,993)
$ (0.007)
> 1 Million
-
-
$
$
$
$
$
$
$
$
Total
76,818,809
76,828,746
$ (1,273,980)
$ (0.017)
$ (1,193,752)
$ (0.016)
$ (920,040)
$ (0.012)
$ (815,739)
$ (0.011)
Community Water Systems (CWSs) performing Corrective Actions
<100
13,927
13,880
$ 365,576
$ 26.250
$ 388,703
$ 28.004
$ 330,235
$ 23.712
$ 353,002
$ 25.432
101-500
51,175
51,619
$ 516,102
$ 10.085
$ 508,146
$ 9.844
$ 452,342
$ 8.839
$ 450,070
$ 8.719
501-1,000
43,849
44,557
$ 178,205
$ 4.064
$ 181,189
$ 4.066
$ 155,596
$ 3.548
$ 158,538
$ 3.558
1,001-4,100
260,175
260,175
$ 590,719
$ 2.270
$ 590,719
$ 2.270
$ 512,568
$ 1.970
$ 512,568
$ 1.970
4,101-33,000
3,170,059
3,170,059
$ 2,605,551
$ 0.822
$ 2,605,551
$ 0.822
$ 2,198,499
$ 0.694
$ 2,198,499
$ 0.694
33,001-96,000
2,700,868
2,700,868
$ 1,709,801
$ 0.633
$ 1,709,801
$ 0.633
$ 1,424,758
$ 0.528
$ 1,424,758
$ 0.528
96,001-500,000
2,858,284
2,858,284
$ 1,625,742
$ 0.569
$ 1,625,742
$ 0.569
$ 1,346,704
$ 0.471
$ 1,346,704
$ 0.471
500,001-1 Million
613,799
613,799
$ 195,076
$ 0.318
$ 195,076
$ 0.318
$ 161,185
$ 0.263
$ 161,185
$ 0.263
> 1 Million
-
-
$
$
$
$
$
$
$
$
Total
9,712,136
9,713,240
$ 7,786,773
$ 0.802
$ 7,804,928
$ 0.804
$ 6,581,888
$ 0.678
$ 6,605,324
$ 0.680
Community Water Systems (CWSs) not performing Level 1/Level 2 Assessments, or Corrective Actions
<100
167,976
168,442
(67,959)
$ (0.405)
(13,728)
$ (0.081)
1,943
$ 0.012
71,665
$ 0.425
101-500
1,077,758
1,073,324
(7,905)
$ (0.007)
59,230
$ 0.055
101,979
$ 0.095
188,231
$ 0.175
501-1,000
1,186,361
1,179,288
1,936
$ 0.002
29,995
$ 0.025
52,024
$ 0.044
88,334
$ 0.075
1,001-4,100
5,214,839
5,214,839
42,797
$ 0.008
42,797
$ 0.008
119,614
$ 0.023
119,614
$ 0.023
4,101-33,000
-
-
-
$
-
$
-
$
-
$
33,001-96,000
-
-
-
$
-
$
-
$
-
$
96,001-500,000
-
-
-
$
-
$
-
$
-
$
500,001-1 Million
3,626,992
3,626,992
-
$
-
$
-
$
-
$
> 1 Million
16,309,853
16,309,853
-
$
-
$
-
$
-
$
Total
27,583,779
27,572,738
$ (31,132)
$ (0.001)
$ 118,294
$ 0.004
$ 275,560
$ 0.010
$ 467,844
$ 0.017
Sources: (C), (E), (G), (I) Exhibit 7.28.
ES.7 Comparison of Benefits and Costs, and Regulatory Options of the RTCR
As required by the SDWA, EPA has determined that the benefits of the RTCR justify the
costs. In making this determination, EPA considered quantified and non-quantified benefits and
costs as well as the other components of the HRRCA outlined in section 1412 (b)(3)(C) of the
SDWA.
Additionally, EPA used several other techniques to compare benefits and costs including
a break-even analysis and a cost effectiveness analysis. The break-even analysis (see Chapter 9
of the RTCR EA) was conducted using two example pathogens responsible for some (unknown)
proportion of waterborne illness in the United States. The analysis shows that a relatively small
number of cases (based on STEC 0157 or Salmonella) would need to be avoided for the rule to
break-even with the best estimates of net costs.
Economic Analysis for the Final RTCR
ES-22
September 2012
-------
Under the RTCR, approximately two deaths would need to be avoided annually using a
3% discount rate based on consideration of the bacterial pathogen STEC 0157. Alternatively,
approximately 3,000 or 8,000 non-fatal cases, using the enhanced or traditional benefits
valuations approaches,5 respectively, would need to be avoided to break even with the RTCR
costs. As expected based on its costs, the Alternative option would require that a higher number
of cases be avoided annually for that option to break even (between 100% and 113% more than
the RTCR under the enhanced and traditional approaches, respectively).
As Exhibit ES.8 shows, approximately 2 deaths would need to be avoided annually from
a Salmonella infection for the rule to break even. The estimated number of non-fatal Salmonella
cases that would need to be avoided to break even is approximately 10,000 or 68,000 cases under
the enhanced and traditional benefits valuations approaches, respectively. As expected based on
its costs, the Alternative option would require that a higher number of Salmonella cases be
avoided annually for that option to break even (approximately 110% more than the RTCR under
either the enhanced or traditional approaches).
As the discussion presented in Chapter 2 of this EA describes, disease and deaths are
attributable to drinking water contamination across the country. The CDC estimates that there are
73,000 cases of illness each year in the U.S. due to STEC 0157 (Mead et al., 1999). The CDC
has found that 15% of outbreak cases of STEC 0157 are waterborne (Rangel et al., 2005); if that
rate applies as well to endemic cases, then approximately 11,000 cases would be due to
waterborne exposure to STEC 0157 and could be mitigated. Based on that assumption, up to
11,000 cases annually may be avoided by measures that mitigate pathways of contamination into
PWSs for just this one potential contaminant. That amount of reduction in terms of STEC 0157
would be enough for the rule to break even based on either the enhanced or traditional cost of
illness approaches. If more than one contaminant was reduced or prevented from occurring in
PWSs, the rule would be that much more likely to break even. Avoided illness and death from
secondary transmission of infection could also be significant. Additionally, fewer cases would
need to be avoided if all the benefits of the rule as described in chapter 6 were included in the
analysis. If increased assessments and corrective actions result in a level of knowledge of the
system that enables earlier mitigation of potential pathways of contamination, then additional
illnesses or deaths may be avoided beyond those suggested by the decrease in acute events that
the model predicts.
Chapter 9 of the RTCR EA has a complete discussion of the break-even analysis and how
costs per case were calculated.
5 Both traditional and enhanced COI approaches count the value of the direct medical costs and of time lost that
would been spent working for a wage, but differ in their assessment of the value of time lost that would be spent in
nonmarket work (e.g. housework, yardwork, and raising children) and leisure (e.g. recreation, family time, and
sleep). They also differ in their valuation of (other) disutility, which encompasses a range of factors of well-being,
including both inconvenience and any pain and suffering. A complete discussion of the traditional and enhanced
COI approaches can be found in Appendix E.
Economic Analysis for the Final RTCR
ES-23
September 2012
-------
Exhibit ES.7 Estimated Annual Breakeven Threshold for Avoided Cases of STEC
0157
COI
Methodology
Discount
Rate
RTCR
Alternative Option
Non-fatal
cases only
Fatal
cases
only1
Non-fatal
cases only
Fatal
cases
only1
Traditional
COI
3%
8,000
1.6
17,000
3.4
7%
8,000
1.6
18,000
3.6
Enhanced COI
3%
3,000
1.6
6,000
3.4
7%
3,000
1.6
6,000
3.6
Calculations for fatal cases include the non-fatal COI component for the underlying
illness prior to death.
Notes:
1 )The number of cases needed to reach break-even threshold calculated by dividing
the net change in costs for the RTCR (Exhibit 9.11) by the average estimated value
of avoided cases (Exhibit 9.18). Threshold estimates based on a weighted average
of the cost of "all cases" that includes both fatal and non-fatal cases is shown in
Appendix E. STEC 0157 is only an example of a pathogenic endpoint that could
have been used for this analysis. Use of additional pathogenic contaminants in
addition to this single endpoint would result in lower threshold values. Detail may
not add due to independent rounding.
2)The break-even threshold is higher using a 7% discount rate than a 3% discount
rate under the Alternative option. This result is consistent with the annualized costs
of the Alternative option being higher using the 7% discount rate, which is caused
by the frontloading of costs in the period of analysis, as explained further in Chapter
7 of the EA.
Economic Analysis for the Final RTCR
ES-24
September 2012
-------
Exhibit ES.8 Estimated Annual Breakeven Threshold for Avoided Cases of
Salmonella
COI
Methodology
Discount
Rate
RTCR
Alternative Option
Non-fatal
cases only
Fatal
cases
only1
Non-fatal
cases only
Fatal
cases
only1
Traditional
COI
3%
68,000
1.6
141,000
3.4
7%
68,000
1.6
151,000
3.6
Enhanced COI
3%
10,000
1.6
21,000
3.4
7%
10,000
1.6
23,000
3.6
Note: Calculations for fatal cases include the non-fatal COI component for the
underlying illness prior to death.
Notes:
1 )The number of cases needed to reach break-even threshold calculated by dividing
the net change in costs for the RTCR (Exhibit 9.11) by the average estimated value
of avoided cases (Exhibit 9.19). Threshold estimates based on a weighted average of
the cost of "all cases" that includes both fatal and non-fatal cases is shown in
Appendix E. Salmonella is only an example of a pathogenic endpointthat could have
been used for this analysis. Use of additional pathogenic contaminants in addition to
this single endpoint would result in lower threshold values. Detail may not add due to
independent rounding.
2)The break-even threshold is higher using a 7% discount rate than a 3% discount rate
under the Alternative option. This result is consistent with the annualized costs of the
Alternative option being higher using the 7% discount rate, which is caused by the
frontloading of costs in the period of analysis, as explained further in Chapter 7 of the
EA.
Economic Analysis for the Final RTCR
ES-25
September 2012
-------
Cost effectiveness is another way of examining the benefits and costs of the rule. ES.9
shows the cost of the rule per corrective action implemented. The cost effectiveness analysis, as
with the net benefits, is limited because EPA was able to only partially quantify and monetize the
benefits of the RTCR. The RTCR achieves the lowest cost per corrective action implemented
among the options considered.
Exhibit ES.9 Total Net Annual Cost Per Corrective Action Implemented under the
RTCR and Alternative Option, Annualized Using 3% and 7% Discount Rates
($Millions, $2007)
Regulatory Option
3% Discount Rate
7% Discount Rate
RTCR - Net Change
$ 14.3
$ 14.2
RTCR - Incremental Number of Corrective Actions (L1 & L2)
616
594
RTCR - Cost Effectiveness Analysis1
$ 0.02
$ 0.02
Alternative Option - Net Change
$ 29.6
$ 31.7
Alternative Option - Incremental Number of Corrective Actions (L1 & L2)
808
819
Alternative Option - Cost Effectiveness Analysis1
$ 0.04
$ 0.04
Notes:
1CEA = (Net Change)/(lncremental Number of corrective actions).
2Detailed cost information is provided in Appendix C.
EPA also considered the net cost effectiveness of the RTCR as compared to the
Alternative option to determine the additional benefit associated with that portion of cost for the
Alternative option that exceeds the cost of the RTCR. Exhibit ES.9 shows that in terms of net
rule cost for all PWSs, the RTCR has a far lower unit cost per unit benefit than the Alternative
option. EPA further considered the group of 60,200 TNCWSs using GW and serving <100
people, which is the largest subset of systems by size and type, and is expected to bear the
highest burden under the RTCR. Exhibit ES. 10 shows that in terms of net rule cost, the cost
effectiveness of the RTCR far exceeds that of the Alternative option using either a 3% or 7%
discount rate. The two net cost analyses (ES.9 and ES. 10) together indicate that the RTCR is
significantly more cost effective than the Alternative option for the most burdened subset of
systems and for all the PWSs together. Additional information about this analysis and other
methods used to compare benefits and costs can be found in Chapter 9 of the RTCR EA.
Economic Analysis for the Final RTCR
ES-26
September 2012
-------
Exhibit ES.10 Annualized Net Change in Costs Per Corrective Action (CA)
Implemented for All PWSs under the RTCR and Alternative Option ($Millions,
2007$)
Regulatory Option
3% Discount Rate
7% Discount Rate
RTCR - Net Change
$ 14.30
$ 14.17
RTCR - Incremental Number of Corrective Actions (L1 & L2)1
615.82
593.71
RTCR - Incremental Cost per Corrective Action ($)
$ 0.02
$ 0.02
Alternative Option - Net Change (over RTCR)2
$ 15.29
$ 17.52
Alternative Option - Incremental Number of Corrective Actions (L1 & L2) (over RTCR)2
192.23
224.93
Alternative Option - Incremental Cost per Corrective Action ($)
$ 0.08
$ 0.08
Notes:
1 Exhibit includes the number of corrective actions predicted by the RTCR occurrence model to be implemented in
addition to those implemented under the 1989 TCR.
2Add net values for Alternative option to net values for RTCR to calculate total net values of Alternative option over
the 1989 TCR.
3Detailed cost information is provided in Appendix C.
Exhibit ES.11 Annualized Net Change in Costs per CA Implemented for TNCWSs
(Serving <100 people) under the RTCR and Alternative Option ($Millions, 2007$)
Regulatory Option
3% Discount Rate
7% Discount Rate
RTCR - Net Change
$ 5.3
$ 5.3
RTCR - Incremental Corrective Actions (L1 & L2)1
287
275
RTCR - Incremental Cost per Corrective Action ($)
$ 0.02
$ 0.02
Alternative Option - Net Change (over RTCR)2
CD
4^
$ 10.8
Alternative Option - Incremental Corrective Actions (L1 & L2) (over RTCR)2
132
155
Alternative Option - Incremental Cost per Corrective Action ($)
$ 0.07
$ 0.07
Notes:
1 Derived by dividing incremental rule costs applicable to TNCWS <100 by the incremental number of corrective
actions to be implemented under the RTCR (relative to the 1989 TCR) and the Alternative option (relative to the
RTCR) by TNCWSs using GW.
2Add net values for Alternative option to net values for the RTCR to calculate total net values of Alternative option
over the 1989 TCR.
3Detailed cost information is provided in Appendix C.
ES.8 Conclusion
The RTCR aims to further reduce occurrence of TC and E. coli (and any waterborne
pathogens that may co-occur) beyond the reductions achieved under the 1989 TCR. EPA's goal
is to increase public health protection from potential fecal contamination and/or waterborne
pathogen exposure, while continuing to pursue the objectives of the 1989 TCR. Although EPA's
analysis of the RTCR has determined that its annual costs are most likely below the threshold
stated in Executive Order 12866 of $100 million, EPA has chosen to publish a complete EA for
this rule.
EPA has developed the final RTCR to be consistent with the majority of the AIP
recommendations, with consideration of the comments received on the proposed rule, because it
will not allow for backsliding from current levels of protection of public health and will likely
Economic Analysis for the Final RTCR
ES-27
September 2012
-------
reduce risk from waterborne disease further while costing relatively little on a net basis. This
conclusion is consistent with the conclusion of the TCRDSAC regarding the likely impact of the
rule provisions. As required by the SDWA, EPA has determined that the benefits of the RTCR
justify the costs. In making this determination, EPA considered quantified and non-quantified
benefits and costs as well as the other components of the HRRCA outlined in section 1412
(b)(3)(C) of the SDWA.
As shown in the uncertainty analysis presented in Chapter 6 (section 6.4), reducing repeat
samples and additional samples both have the potential to increase risk for some PWSs (see also
sections III.C and III.D of the RTCR preamble for discussions on the additional routine sample
and repeat sample provisions respectively). However, this increase in risk is expected to be more
than offset by potential decreases in risk from increased routine monitoring and the addition of
the assessments and corrective action provisions that find and fix problems indicated by
monitoring. That is, based on the analyses presented in this EA, EPA concludes that the effect of
the corrective actions outweighs that of reduced sampling; both the RTCR and the Alternative
option are expected to result in a net decrease in risk as compared to the 1989 TCR. Although the
Alternative option decreases risk to a greater degree than the RTCR, this additional reduction is
achieved at a higher cost, both in absolute terms and in terms of cost-effectiveness. The
estimated net cost of the RTCR is not only small relative to the 1989 TCR (approximately $14
million annually using either a 3% or 7% discount rate), but it is also small compared to net cost
increase of the Alternative option relative to the 1989 TCR (approximately $30-$32 million
using a 3% and 7% discount rate, respectively). This cost differential is especially important
considering the potential concentration of impacts on the smallest TNCWSs and the potential
frontloading of costs during the first few years of implementation under the Alternative option.
In addition, no backsliding in overall risk is predicted.
The analyses performed as part of this EA support the collective judgment and consensus
of the TCRDSAC that the RTCR requirements provide for effective and efficient revisions to
1989 TCR regulatory requirements.
Economic Analysis for the Final RTCR
ES-28
September 2012
-------
1 Introduction
This economic analysis (EA) presents the evaluation of the benefits and costs of the
Revised Total Coliform Rule (RTCR). The analysis is performed in compliance with Executive
Order 12866, Regulatory Planning and Review (58 FR 51735), which requires the United States
Environmental Protection Agency (EPA) to estimate the economic impact of rules that have an
annual effect on the economy of over $100 million and make that analysis available to the public
in conjunction with publication of the final rule. Although EPA's analysis of the RTCR has
determined that its annual costs are most likely below this threshold, EPA has chosen to publish
a complete EA for this rule.
EPA developed the RTCR in collaboration with states, other interested stakeholders, and
the Total Coliform Rule/Distribution System Advisory Committee (TCRDSAC). In July 2003, as
part of the Agency's Six Year Review of existing national primary drinking water regulations,
EPA published its intent to revise the 1989 Total Coliform Rule (TCR). The Agency's primary
reasons for revising the 1989 TCR are implementation-related issues. The RTCR offers a
meaningful opportunity for greater public health protection against waterborne pathogens in the
distribution systems of public water systems (PWSs) beyond the 1989 TCR.
This chapter provides a summary of the RTCR in Section 1.1. Section 1.2 outlines the
organization of this EA, and Section 1.3 provides information regarding supporting calculations
and citations.
1.1 Summary of the Revised Total Coliform Rule (RTCR)
The RTCR applies to all community and noncommunity PWSs. The RTCR takes a
proactive approach to protect public health, maintaining a maximum contaminant level goal
(MCLG) and maximum contaminant level (MCL) for E. coli and using both E. coli and total
coliform (TC) monitoring to establish a framework for PWSs to assess for sanitary defects and to
correct them as appropriate.
Like the requirements of the 1989 TCR, the requirements of the RTCR ensure that PWSs
address the following objectives:
• Evaluate the effectiveness of treatment;
• Determine the integrity of the distribution system; and
• Signal the possible presence of fecal contamination.
To accomplish the above objectives, the RTCR uses TC as an indicator to start an
evaluation process that, where necessary, requires the PWS to correct sanitary defects, defined in
the RTCR as:
A defect that could provide a pathway of entry for microbial contamination into the
distribution system or that is indicative of a failure or imminent failure in a barrier that is
already in place.
Economic Analysis for the Final RTCR
1-1
September 2012
-------
Under the RTCR, E. coli remains a regulated contaminant with an MCLG of zero and an
MCL of zero. All fecal coliform provisions (including the MCLG and MCL) are removed in the
RTCR. There is no longer an MCL for TC: instead, TC is used as an indicator as part of a
coliform treatment technique.
For all PWSs serving >4,100 people, monitoring requirements under the RTCR remain
essentially unchanged from 1989 TCR requirements. However, PWSs serving <4,100 people
may experience changes in their required monitoring schemes depending on specific size, type,
and source water categorizations. Changes for these PWSs may be for routine, additional routine
or repeat monitoring as summarized in Exhibit 1.1 below.
Under the RTCR, PWSs must complete Level 1 or Level 2 assessments to identify the
presence of "sanitary defects" and defects in distribution system coliform monitoring practices if
sampling results in one of the following triggers listed below.
Level 1 Assessment Triggers
• For PWSs taking 40 or more samples per month, the PWS exceeds 5.0 percent TC-
positive samples for the month; or
• For PWSs taking fewer than 40 samples per month, the PWS has two or more TC-
positive samples in the same month; or
• Failure to take every required repeat sample after a single routine TC-positive
sample.
Level 2 Assessment Triggers
• The PWS has an E. coli violation (See Section III.B. 1 a of the RTCR preamble for a
description of what constitutes an E. coli MCL violation.); or
• The PWS has a second Level 1 treatment trigger within a rolling 12-month period,
unless the first Level 1 treatment trigger was based on exceeding the allowable
number of TC+ samples, the state has determined a likely reason for the TC+
samples that caused the initial Level 1 treatment trigger, and the state establishes
that the system has fully corrected the problem;
• For PWSs with approved reduced annual monitoring, a Level 1 treatment technique
trigger in two consecutive years.
Economic Analysis for the Final RTCR
1-2
September 2012
-------
Exhibit 1.1 RTCR Monitoring Frequency Requirements
PWS Type
Source
Population Served
Monitoring Requirements
Routine Monitoring1
Noncommunity
(NCWS)
Ground water
< 1,000 people
Quarterly monitoring forTC2
Additional requirements to qualify for reduced
monitoring apply3
Community
(CWS)
Ground water
< 1,000 people
Monthly monitoring forTC
Additional requirements to qualify for reduced
monitoring apply4
All
(CWS & NCWS)
Surface water
All
Same requirements as under 1989 TCR
(Monthly monitoring forTC)
Ground water
>1,000 people
Same requirements as under 1989 TCR
(Monthly monitoring forTC)
Additional Routine Monitoring
All
(CWS & NCWS)
Ground water
< 1,000 people
Additional routine monitoring no longer
required if monitoring monthly
Number of additional routine samples required
reduced from five to three if monitoring
quarterly or annually
Ground water
1,101-4,100 people
Additional routine monitoring no longer
required
Surface water
<4,100 people
Additional routine monitoring no longer
required
Ground and
surface water
>4,100 people
Same requirements as under 1989 TCR
(Additional routine monitoring not required)
Repeat Monitoring
All
(CWS & NCWS)
Ground and
surface water
<1,000 people
Number of repeat samples required reduced
from four to three
Ground and
surface water
>1,000 people
Same requirements as under 1989 TCR
(Three repeat samples required)
monitoring.
2 PWSs may be increased to monthly monitoring based on compliance with rule. Examples include system with Level
2 assessment violation, an E. coli MCL violation, an RTCR treatment technique violation, and /or two monitoring
violations in a rolling 12-month period.
3 PWSs may qualify for annual monitoring based on compliance record, the existence of enhancements to water
system barriers to contamination, and completion of an initial and recurring annual site visit conducted by the state or
an annual voluntary Level 2 assessment by a party approved by the state.
4 PWSs may qualify for quarterly monitoring based on compliance record and either the existence of enhancements
to water system barriers to contamination or the completion of an annual site visit conducted by the state or an
annual voluntary Level 2 assessment by a party approved by the state.
Economic Analysis for the Final RTCR
1-3
September 2012
-------
The PWS shall be responsible for correcting sanitary defects found through either Level 1
or Level 2 assessments as defined above by implementing appropriate corrective actions. Failure
to complete a required assessment or take necessary corrective action constitutes a Treatment
Technique violation under the RTCR.
The RTCR requires public notification for three events that are tiered as follows:
• Tier 1. E. coli MCL violation
• Tier 2. Treatment technique violation
• Tier 3. Routine monitoring and reporting violations
The RTCR also provides the primacy agency with the discretion to reduce the monitoring
frequency for well-operated ground water PWSs serving <1,000 people, if the PWS can
demonstrate that it meets the criteria for reduced monitoring stipulated in the rule. The specific
criteria to qualify for reduced monitoring are described in Chapter 3.
Ground Water Rule (GWR) implementation began in December of 2009 prior to the
transition from the 1989 TCR to the RTCR. Compliance with the GWR requirements, including
sanitary surveys and site visits, can be used to help determine the level of monitoring required
for ground water PWSs serving <1,000 people.
1.2 Document Organization
The remainder of this EA is organized into the following chapters:
• Chapter 2 summarizes the technical, regulatory, and public health issues addressed
by the RTCR. It also explains the statutory authority for the RTCR and the
economic rationale for the regulatory approach.
• Chapter 3 reviews various regulatory options that EPA considered during the
development of the rule and presents the rationale for selecting the RTCR
requirements.
• Chapter 4 characterizes baseline conditions that exist (including PWS inventory,
treatment, and water quality data) before PWSs make changes to meet the RTCR
requirements.
• Chapter 5 summarizes TC and EC occurrence analysis, providing a description of
the occurrence model, the sources used, and the limitations or constraints to the
analysis based on the nature of the data provided by those sources.
• Chapter 6 presents the relative risk evaluation performed to estimate the potential
benefits of the RTCR relative to the baseline and other regulatory options
considered.
Economic Analysis for the Final RTCR
1-4
September 2012
-------
• Chapter 7 presents an estimate of the net costs of implementing the RTCR to
industry, households, and states. The net costs of all regulatory options considered
are compared.
• Chapter 8 discusses distributional analyses performed to evaluate the effects of the
rule on different segments of the population, and considers various executive orders
and requirements, including the Regulatory Flexibility Act and Unfunded Mandates
Reform Act.
• Chapter 9 compares the benefits and costs of the RTCR to evaluate the potential
net benefits and cost-effectiveness. The results are discussed and compared to other
regulatory options considered.
• Chapter 10 includes a detailed list of references.
1.3 Calculations and Citations
This EA involves detailed and complex analyses, and the following are provided to help
the reader:
• The detailed reference section provided in Chapter 10.
• Appendices containing supporting spreadsheets and analyses:
- Appendix A—Detailed Predictive Model Results
- Appendix B—Graphs of Predicted Hit Rates Over Time
- Appendix C—Detailed Cost Model Results
- Appendix D—Detailed Compliance Forecast and Unit Costs Estimates
- Appendix E—Supporting information for Value of Statistical Life and
Cost of Illness estimates used in break-even analyses
- Appendix F—Detailed analysis of potential uncertainty and variability in
occurrence estimates
- Appendix G—Evaluation of representativeness of Six Year Review data
- Appendix H—Analysis of potential impacts of reduced repeat sampling
- Appendix I—Supporting Information for Regulatory Flexibility Act
Screening Analysis
• Exhibits. Most tabular exhibits include a row that provides the formulas used to
compute the contents of each column.
• Sources for values used if they were not calculated within the exhibits.
Economic Analysis for the Final RTCR
1-5
September 2012
-------
• Supporting electronic file outputs (i.e., RTCR occurrence and cost model output).
• Flowcharts that illustrate methodologies of analyses as well as RTCR requirements.
Economic Analysis for the Final RTCR
1-6
September 2012
-------
2 Statement of Need for the Rule
2.1 Introduction
The United States Environmental Protection Agency (EPA) is revising the 1989 Total
Coliform Rule (TCR) to better protect public health. EPA noticed its intent to revise the 1989
TCR as part of its Six-Year Review determination published in 2003 (USEPA 2003b, 68 FR
42907, July 18, 2003). In July 2007 EPA convened a Total Coliform Rule/Distribution System
Federal Advisory Committee (TCRDSAC or advisory committee) charged with evaluating how
well the objectives of the TCR are met and recommending possible revisions to the rule. The
TCRDSAC completed its analysis and provided recommendations for revising the 1989 TCR in
an Agreement in Principle (AIP) (USEPA, 2009, 74 FR 1683, January 13, 2009) that was signed
by advisory committee members in September 2008. In July 2010, EPA proposed a rule that was
consistent with the recommendations in the AIP (see Chapter 3 of this economic analysis (EA)
for more details on the TCRDSAC). The final RTCR provisions were created by also taking into
consideration comments received on the RTCR proposal.
The RTCR maintains and strengthens the objectives of the 1989 TCR. The objectives are:
(1) to evaluate the effectiveness of treatment, (2) to determine the integrity of the distribution
system, and (3) to signal the possible presence of fecal contamination. The RTCR better
addresses these objectives by requiring systems that may be vulnerable to fecal contamination
(as indicated by their monitoring results) to do an assessment, to identify whether any sanitary
defect(s) is (are) present, and to correct the defects. Therefore, greater public health protection is
anticipated under the RTCR compared to the 1989 TCR because of its more preventive approach
to identifying and fixing problems that affect or may affect public health (see Chapter 3 of this
EA for more details on the provisions of the RTCR).
This chapter summarizes the technical, regulatory, and public health issues addressed by
the final rule. It also explains the statutory authority for the final RTCR and the economic
rationale for choosing a regulatory approach rather than non-regulatory options.
2.1.1 Description of the Issue
EPA is required to review each existing national primary drinking water regulation
(NPDWR) every six years. In 2003, EPA completed its review of the 1989 TCR and 68
NPDWRs for chemicals that were established prior to 1997. The purpose of the review was to
identify current health risk assessments, changes in technology, and other factors that would
provide a health or technological basis to support a regulatory revision that will maintain or
improve public health protection. In the 2003 announcement of the completion of the Six Year
Review, EPA provided public notice of its intent to revise the 1989 TCR (USEPA 2003b, 68 FR
42907, July 18, 2003). Implementation-related issues are the primary reason for EPA's decision
to revise the 1989 TCR. Since promulgation of the 1989 TCR, EPA has received comments from
a number of stakeholders suggesting modifications to reduce the burden of implementing the
1989 TCR. These comments included, among others, suggestions to modify the 1989 TCR's
monitoring requirements (e.g., with regards to sampling locations and repeat samples) (USEPA
2002b, 67 FR 19030, April 17, 2002 and USEPA 2003b, 68 FR 42907, July 18, 2003).
Economic Analysis for the Final RTCR
2-1
September 2012
-------
The 1989 TCR was promulgated to decrease the risk of waterborne illness. Among all
rules promulgated for preventing waterborne illness, only the 1989 TCR applies to all PWSs, so
the rule is an essential component of the multiple barrier approach in public health protection
against endemic and epidemic disease. However, improvements to the 1989 TCR construct and
requirements can better protect public health as shown in the following discussions.
In recent years, violation rates under the 1989 TCR have remained relatively steady (see
Chapter 4, Exhibit 4.11 of this EA). EPA believes that this is reflective of a steady state among
PWSs complying with the 1989 TCR, suggesting the possibility that improvements likely to
occur under the 1989 TCR have largely been achieved. Potential exposure to waterborne
pathogens continues to be a health risk. Within the United States, disease is caused by a
relatively small variety of bacterial, viral, and parasitic protozoan pathogens in drinking water.
Nevertheless, the number of potential waterborne pathogens and the number of potential
waterborne illnesses are significant. Humans can be exposed via fecal-oral, inhalation or dermal
exposure, with fecal-oral exposure being the most significant and prevalent pathway.
The 1989 TCR established a maximum contaminant level goal (MCLG) and a maximum
contaminant level (MCL) for total coliforms (TC) (including fecal coliforms (FC) and E. coli).
Under the 1989 TCR, a TC MCL violation required public notification. However, many of the
organisms detected by TC and FC methods are not of fecal origin and do not have any direct
public health implication. Information has also become available since promulgation of the 1989
TCR that indicates that measurement of FC sometimes detects organisms that may not have any
connection to fecal contamination (Edberg et al., 2000). Attributing greater public health
significance to the presence of TC or FC could result in public confusion. On the other hand,
EPA continues to believe that E. coli is a meaningful indicator of fecal contamination and of the
potential presence of associated pathogens (see Section 2.2.4.1 of this chapter).
The RTCR aims to maintain and strengthen the objectives of the 1989 TCR more
effectively and efficiently. E. coli remains a regulated contaminant with a defined MCLG and
MCL while TC (including FC) no longer has an MCLG and an MCL. Instead, the final RTCR
uses TC as an indicator as part of a coliform treatment technique. The treatment technique uses
both TC and E. coli monitoring results to start an evaluation process that, where necessary, will
require the PWS to conduct assessments and corrective actions. The final rule also takes into
account the capacity of small systems (especially for PWSs serving 1,000 or fewer people) and
primacy agencies to effectively implement the final rule requirements.
Modeling results show that the assessment and corrective action provisions of the RTCR
will result in a decrease in E. coli occurrence (see Exhibits 5.16-5.21) which is anticipated to
lead to a decrease in the potential exposure of the public to fecal contamination and its associated
pathogens. In general, this decrease in fecal contamination should reduce the potential risk to
human health of PWS customers. Thus, any reduction in E. coli occurrence is considered a
benefit of the RTCR. Also, since fecal contamination can contain waterborne pathogens
including bacteria, viruses, and parasitic protozoa, in general a reduction in fecal contamination
should also reduce the potential risk from these other contaminants.
Economic Analysis for the Final RTCR
2-2
September 2012
-------
2.2 Public Health Concerns, Fecal Contamination, and Waterborne Pathogens
2.2.1 Rule Objectives and Public Health Concerns
The RTCR aims to increase public health protection through a reduction of potential
pathways of entry for fecal contamination into the distribution system. Since these potential
pathways represent vulnerabilities in the distribution system whereby fecal contamination and/or
waterborne pathogens, including bacteria, viruses, and parasitic protozoa can enter the system,
the reduction of these pathways in general should lead to reduced exposure and associated risk
from these contaminants. Waterborne pathogens that can be associated with fecal contamination
may cause a variety of illnesses, including acute gastrointestinal illness (AGI) with diarrhea,
abdominal discomfort, nausea, vomiting, and other symptoms. Most AGI cases are of short
duration and result in mild illness. Consuming water contaminated with waterborne pathogens
can also result in acute illnesses such as hepatitis, hemolytic uremic syndrome (kidney failure),
and bloody diarrhea; milder, acute illnesses such as conjunctivitis; and severe chronic illnesses
such as diabetes and dilated cardiomyopathy. Other examples of potential chronic diseases
resulting from infection by a waterborne agent include irritable bowel syndrome, hypertension,
and reactive arthritis; chronic illnesses are generally costly to treat.6
Sensitive subpopulations are at greater risk from waterborne disease than the general
population. These sensitive subpopulations include children (especially the very young); the
elderly; the malnourished; pregnant women; chronically ill patients (e.g., those with diabetes or
cystic fibrosis); and a broad category of those with compromised immune systems, such as
Acquired Immunodeficiency Syndrome (AIDS) patients, those with autoimmune disorders (e.g.,
rheumatoid arthritis, lupus erythematosus, and multiple sclerosis), organ transplant recipients,
and those receiving chemotherapy (Rose, 1997). Sensitive subpopulations represent almost 20
percent of the population in the United States (Gerba et al., 1996). The severity and duration of
illness is often greater in sensitive subpopulations than in healthy individuals, and may
occasionally result in death.
When humans are exposed to and infected by an enteric pathogen (pathogens that most
commonly occur in the gut of humans and other animals), the pathogen becomes capable of
reproducing in the gastrointestinal tract. As a result, healthy humans shed pathogens in their
feces for a period ranging from days to weeks. This shedding of pathogens often occurs in the
absence of any signs of clinical illness. Regardless of whether a pathogen causes clinical illness
in the person who sheds it in his or her feces, the pathogen being shed may infect other people
directly by person-to-person spread, contact with contaminated surfaces, and other means which
are referred to as secondary spread. As a result, waterborne pathogens that are initially
waterborne may subsequently infect other people through a variety of routes (WHO, 2004).
Waterborne pathogens include pathogens of both fecal and non-fecal origin. Non-fecal
pathogens are found in the soil and soil water interface and include Legionella, Naegleria fowleri
6 Lifetime costs associated with a new case of diabetes, for example, assuming an average illness duration of 30
years, are estimated at $255,833 using a three percent discount rate and $161,967 using a seven percent discount rate
(year 2007 dollars). For dilated cardiomyopathy, the lifetime (21-year average) cost is $68,870 (seven percent
discount rate, year 2007 dollars).
Economic Analysis for the Final RTCR
2-3
September 2012
-------
and probably Helicobacter pylori, or are components of the distribution system environment and
can include Mycobacterium avium complex (MAC). Examples of common fecal bacterial
pathogens include pathogenic E. coli, Salmonella, Shigella, and Campylobacter jejuni. Some
waterborne bacterial pathogens cause disease by rapid growth and dissemination (e.g.,
Salmonella) while others primarily cause disease via toxin production (e.g., Shigella, E. coli
0157, Campylobacter jejuni). Campylobacter jejuni, E. coli, and Salmonella have a host range
that includes both animals and humans; Shigella is associated only with humans (Geldreich,
1996). Unlike viruses, bacteria are able to reproduce outside of the host.
Because a positive TC assay is an indicator of ambient water entry into or growth within
the distribution system, it represents an indication of potential for both fecal and non-fecal
contamination. Because E. coli originates in the gut of warm-blooded animals, it is an indication
of fecal contamination. Together, the TC IE. coli indicator system can be used in monitoring to
identify the potential for fecal contamination. Since fecal contamination can contain waterborne
pathogens including bacteria, viruses, and parasitic protozoa, in general a reduction in fecal
contamination should also reduce the potential risk from these other contaminants.
The group of bacteria known as E. coli contains both pathogenic and non-pathogenic
isolates. The methods approved by EPA to assay for E. coli do not identify most of the
pathogenic E. coli strains. The most dangerous E. coli bacteria contain the gene for producing
shiga-toxins. E. coli 0157:H7 is the most widespread shiga-toxin producing E. coli but at least
81 serotypes have been identified (Prager et al., 2005). Release of toxins in the body can result in
kidney failure, shock and death in otherwise healthy individuals, especially small children.
Typically, kidney failure occurs in 2-7 percent of illnesses. Death or end-stage renal disease
occurs in about 12 percent of patients four years after diarrhea-associated kidney failure (Garg et
al., 2003). Twenty five percent of kidney failure survivors demonstrate long-term renal
consequences (Garg et al., 2003). For patients with moderate and severe gastroenteritis caused by
E. coli, long-term study shows that they have an increased risk of irritable bowel syndrome,
hypertension and reduced kidney function (Garg et al., 2005).
The Centers for Disease Control and Prevention (CDC) indicates that drinking water is
responsible for a portion of the 73,000 illnesses each year from E. coli 0157:H7 (Mead et al.,
1999). The CDC estimates that about 15 percent of all reported E. coli 0157:H7 cases are due to
water contamination (Rangel et al., 2005). E. coli 0157:H7 is the primary cause of hemolytic
uremic syndrome (HUS) in the United States (Rangel et al., 2005). Active surveillance by CDC
shows that 6.3 percent of E. coli 0157:H7 cases progress to HUS (Griffin and Tauxe, 1991;
Gould et al., 2009) and about 12 percent of HUS cases result in death within four years (Garg et
al., 2003). About 4 to 15 percent of STEC 0157 cases are transmitted within households by
secondary transmission (Parry and Salmon, 1998).
Ground water outbreaks due to E. coli 0157:H7 are prominent because of the fatal
outcomes associated with those outbreaks. In Walkerton, Ontario, 6 individuals died and 27
developed kidney failure from ground water contaminated with E. coli 0157:H7 (and
Campylobacter jejuni) (Health Canada, 2000). The outbreak coincided with a reduction in
chlorination treatment coincident with a large fecal contamination event. In Washington County,
NY, two individuals (including an otherwise healthy two year old child) died from E. coli
0157:H7 contamination of a county fair water supply system. (The fairground's water supply
Economic Analysis for the Final RTCR
2-4
September 2012
-------
system was not recognized as a PWS because it was used for less than 60 days each year.) Four
individuals died in Cabool, MO due to E. coli 0157:H7 (Swerdlow et al., 1992) in an outbreak
that was likely due to source water contamination.
Although manure is often considered to be the source of shiga-toxin producing E. coli,
these bacteria have also been isolated from municipal sewage (Holler et al., 1999). E. coli
0157:H7 was found to survive on a pasture surface for almost 4 months. About 4-15 percent of
cases are acquired via secondary transmission (Parry and Salmon, 1998). In addition to the shiga-
toxin producing E. coli, there are a substantial number of other E. coli bacteria that are
pathogenic, mostly through production of other toxins. Little data are available on the hazard
associated with waterborne transmission for most of the pathogenic E. coli other than E. coli
0157:H7.
As the result of a seven-year-long study of individuals infected or ill during the
Walkerton, Ontario waterborne disease outbreak due to E. coli 0157.HI and Campylobacter
jejuni, researchers noted an increased incidence in chronic disease sequellae. Within a few years
after the outbreak, irritable bowel syndrome, hypertension, reactive arthritis and reduced kidney
function increased 33-38 percent (Clark et al., 2008).
Shigella bacteria are distinct because they are often associated with bloody diarrhea
(bacillary dysentery). The enterohemorraghic E. coli bacteria acquired the capability to produce
toxins by exchanging plasmids with Shigella (Murray et al., 2007). Thus, Shigella often also
causes kidney failure and chronic kidney disease. Shigella contamination only results from
human fecal contamination and thus it is probably less common than E. coli contamination,
which has both human and animal sources.
Campylobacter bacteria (like Salmonella) are very common contaminants of food and
water, and result in a large number of illnesses. Campylobacter is commonly associated with
animal manure, especially cow and chicken manure. More deaths may be associated with
Campylobacter and Salmonella than with viruses. Uniquely, Campylobacter is often associated
with Guillain-Barre paralysis that can last for weeks or months. About one paralysis case occurs
for every 1,000 cases of campylobacteriosis (Altekruse et al., 1999). About 20 percent of
paralysis patients are left with some disability and approximately 5 percent die.
Campylobacterosis is also associated with Reiter syndrome (reactive arthritis). Approximately 1
percent of patients with camplybacterosis have arthritis onset in one or more joints (especially
the knee) in the 7 to 10 days after diarrheal onset (Altekruse et al., 1999). Arcobacter (now a
separate genus from Campylobacter) was responsible for a ground water outbreak at a camp in
Coeur d'Alene, Idaho (McMillan, 1996).
Salmonella causes typhoid fever, once a common and dangerous waterborne disease.
Typhoid is no longer a problem in the United States, and in recent years, Salmonella has become
increasingly less common as a common source outbreak agent while campylobacterosis
outbreaks have correspondingly increased. The reasons for this change are unclear. Seven deaths
that occurred due to Salmonella contamination in a ground water PWS in Gideon, Missouri were
due to bird entry into a storage tank (Angulo et al., 1997). Salmonella resulted in a very large
outbreak in a ground water utility in Riverside, California during the 1960s (16,000 illnesses, 70
hospitalizations and 3 deaths) prior to the promulgation of the 1989 TCR (Boring et al., 1971). A
Economic Analysis for the Final RTCR
2-5
September 2012
-------
2008 waterborne outbreak in Alamosa, Colorado was due to Salmonella. Salmonella-positive
samples were collected from the distribution system but no E. coli was identified in other
distribution samples collected on the same day (CDPHE, 2009).
Legionella are opportunistic bacterial pathogens that colonize water distributions
systems. An estimated 8,000-18,000 cases of Legionnaires disease and Pontiac fever occur in the
U.S. each year due to Legionella (CDC, 2011). Twenty-one of 48 known species are able to
infect humans.
Helicobacter pylori is often associated with ground water (Hegarty et al.,1999; Rolle-
Kampzczyk et al., 2004) and is known to cause gastric ulcers. However, Helicobacter is not
culturable and so occurrence data are fairly uncertain. Improved hygiene and water treatment
have together reduced the number of ulcers caused by this organism over the last few decades
but it is impossible to quantify that decrease.
Most of the waterborne bacterial pathogens cause gastrointestinal illness, but some can
cause other severe illnesses as well. For example, Legionella causes Legionnaires Disease, a
form of pneumonia that has a fatality rate of about 15 percent. It can also cause Pontiac Fever,
which is a milder respiratory infection form of Legionnaires Disease. Several strains of E. coli
can cause severe disease, including kidney failure. Some bacterial pathogens are opportunistic
(i.e., they are only infectious in the presence of another, preexisting condition or weakness).
Opportunistic pathogens usually cause illness only in immunocompromised persons or in other
sensitive subpopulations, such as the very young or the elderly. Other pathogens, such as
Salmonella, Shigella, and Campylobacter jejuni, are not entirely opportunistic but result in
certain diseases with greater frequency and severity in immunocompromised persons (Framm
and Soave, 1997).
Examples of common viral pathogens include norovirus and the enteroviruses
(coxsackievirus, echovirus, poliovirus, enterovirus 70 and 71). Viruses can cause a spectrum of
mild to severe clinical illness, including paralytic disease and death. There is considerable
information that Type 1 diabetes may be associated with enterovirus infection, including
infection with coxsackievirus and echoviruses (Maria et al., 2005; Vreugdenhil et al., 2000).
Dilated cardiomyopathy can follow myocarditis caused by echovirus and other enterovirus
infection.
Mild enteroviral illness includes nonspecific febrile illness, respiratory illness,
photophobia or sensitivity, stiff neck, and gastrointestinal illness. Aseptic meningitis may or may
not require a doctor's visit, but more severe illnesses such as viral encephalitis, myocarditis, and
non-polio flaccid paralysis are likely to require hospitalization. Most likely to be hospitalized are
infants less than 3 months old with non-specific febrile illnesses that require treatment to rule out
and expectantly treat serious bacterial illness. Dilated cardiomyopathy can follow myocarditis
caused by echovirus and other enterovirus infection.
The health effects of norovirus illness include acute onset of nausea, vomiting, abdominal
cramps and diarrhea. Vomiting is more prevalent among children. However, many adults also
experience vomiting, as well as diarrhea. Constitutional symptoms (e.g., headache, fever, chills,
and myalgia) are frequently reported. Although rare, severe dehydration caused by norovirus
Economic Analysis for the Final RTCR
2-6
September 2012
-------
gastroenteritis can be fatal, with this outcome occurring among susceptible persons (e.g., older
persons with debilitating health conditions). No long-term sequelae of norovirus infection have
been reported (CDC, 2001a). Duration of illness is typically 12-60 hours.
Hepatitis A (HAV) virus is the only waterborne virus that is reportable to CDC. About
28,000 HAV cases are reported to CDC each year; however, that number is expected to decline
with time because children are currently vaccinated for HAV in high risk states, and newer
recommendations are for increased vaccination coverage. Mead et al. (1999) estimate about
83,000 HAV cases each year (nearly three times the number of cases reported yearly to CDC),
with a hospitalization rate of 13 percent and a mortality rate of 0.3 percent. Because HAV is
more severe as an adult disease, an aging U.S. population may have greater disease burden.
Hepatitis E (HEV) virus is another fecal oral virus that is potentially waterborne.
Serology and case histories of individual patients indicate that HEV is established as endemic
within the United States (Tsang et al., 2000). However, the data suggest that only one or a few
percent of the population has been infected. Unlike HAV, no waterborne outbreaks of HEV have
occurred in the United Sates, although they have occurred in China and Somalia. The disease is
severe, with up to 20 percent mortality among pregnant women in developing countries.
The adenoviruses are a large group of viruses that produce diverse symptoms. Two
adenovirus serotypes, adenovirus 40 and 41, produce primarily enteric symptoms, but several
other adenoviruses are also capable of producing such symptoms. Additionally, some
adenoviruses cause conjunctivitis. Most significantly, adenoviruses caused a fatal outcome in
otherwise healthy young males in military settings (CDC, 2001b). All adenoviruses, no matter
the infection site and characteristic illness, are shed copiously through the gut. Thus, the
adenoviruses are fecal-oral viruses (Carter, 2005) potentially transmissible via water to infection
sites in the gut or, for the respiratory adenoviruses, in the lung, via waterborne aerosol drops.
Most children are exposed to astroviruses at an early age. Astrovirus is commonly
acquired in child care settings and causes mild disease in children. However, a small percentage
of exposed children may suffer more significant health effects and may require in-patient care.
The disease burden in older children and adult populations is underestimated because the disease
is mild (Carter, 2005). Astroviruses are shed in stool in large numbers. In France, prospective
epidemiology studies have implicated water as a route of astrovirus infection (Gofti-Laroche, et
al., 2003).
The CDC has determined that the incidence of rotavirus diarrhea can reach 0.30
episodes/child/year by age two, with a cumulative incidence approaching 0.80 episodes/child by
age five (Glass et al., 1996). Hospitalizations for rotavirus diarrhea are most common in children
6 months to 3 years of age (Parashar et al., 1998), while self-limiting norovirus infections are
prevalent in school-age children (LeBaron et al., 1990). Although deaths from infectious diarrhea
have generally declined among U.S. children since 1965 because of re-hydration therapy,
newborn children, especially infants born prematurely, remain at risk of death from severe
diarrheal illness (Kilgore et al., 1995). Common strain rotaviruses typically affect young
children, particularly those less than 3 years old, but other strains (Gl, G2 and G9) have been
found to be common in adults and the elderly in nursing homes, and found to be responsible for
more severe illness in children (Griffin et al., 2000). In 2006, the U.S. Food and Drug
Economic Analysis for the Final RTCR
2-7
September 2012
-------
Administration (FDA) approved a new pentavalent vaccine for rotavirus (Vesikari et al., 2006) in
children. The vaccine protects against the most common rotaviruses now found in the U.S.
(serotypes Gl, G2, G3, G4, and G9) but does not protect against all rotaviruses.
Reovirus is commonly found co-occurring with the enteroviruses in PWS wells. Carducci
et al. (2002) found that in some cases enterovirus detection was limited because reovirus
reproduction was so highly favored. Reovirus is more closely related to rotavirus and thus has
some similar characteristics. Reovirus is now recognized as a human pathogen in children (Tyler
et al., 2004).
Cryptosporidium is of particular concern to EPA because, unlike pathogens such as
viruses and bacteria, Cryptosporidium oocysts are resistant to inactivation by many common
disinfection methods. Since the oocysts are especially resistant to chlorine disinfection, simply
increasing existing chlorination dosage levels or contact time above those most commonly
practiced in the United States is not effective. Emerging disinfectant-resistant pathogens, such as
Microsporidia, Cyclospora, and Toxoplasma, are also a concern for similar reasons.
Cryptosporidiosis is a protozoal infection that usually causes 7 tol4 days of diarrhea,
possibly accompanied by low-grade fever, nausea, and abdominal cramps in individuals with
healthy immune systems (Juranek, 1998). It is caused by the ingestion of infectious
Cryptosporidium oocysts, which are readily carried in water. The most common source of
oocysts in water is the feces of infected hosts (Perz et al.,1998; Rose, 1997). Although
cryptosporidiosis often occurs through ingestion of contaminated food or water, it may also
result from direct or indirect contact with infected people or animals (Casemore, 1990; Juranek,
1998; Rose, 1997). Infected humans and other animals excrete oocysts, which can then be
transmitted to others. Cryptosporidiosis can also cause non-gastrointestinal symptoms, such as
eye and joint pain, headaches, dizziness, and fatigue (Hunter et al., 2004). There is no treatment
that can eliminate a Cryptosporidium infection, and only a few antiparasite or antimicrobial
agents have shown even a slight ability to reduce a patient's parasite load (Guerrant, 1997).
In some occurrences, cryptosporidiosis can be fatal, particularly among subpopulations
such as AIDS patients, the elderly with other underlying illnesses, and other immuno-
compromised individuals. In a Cryptosporidium outbreak in Milwaukee in 1994, which had an
estimated 403,000 cases of illness (Kramer et al., 1996), 54 people died who had
cryptosporidiosis listed on their death certificate. Of those 54 people, 46 also had AIDS listed as
an underlying cause of death (Hoxie et al., 1997).
Giardiasis is a protozoal infection that usually causes diarrhea, possibly accompanied by
fever, nausea and abdominal cramps in individuals with healthy immune systems. It is caused by
ingestion of Giardia cysts.
In 2003, two 5-year-old boys living in the same water service area near Phoenix, AZ died
in the same week from Primary Amoebic Meningioencephalitis (PAM) (Marciano-Cabral et al.,
2003). Both boys lived in homes supplied by untreated PWS wells. Atypically, the wells in that
area provide water at elevated temperatures as a result of the elevated geothermal gradient in the
subsurface. Seventeen samples taken from the boys homes were positive for Naegleria fowleri,
and N. fowleri was also responsible for the boys' deaths from PAM. It is likely that the elevated
Economic Analysis for the Final RTCR
2-8
September 2012
-------
ground water temperature provided a suitable habitat for N. fowleri colonization and growth
either in the aquifer, the well, the distribution system, or the household plumbing.
2.2.2 Total Coliforms as Indicators of Treatment Effectiveness and Integrity of the
Distribution System
TC is a group of closely related bacteria that, with a few exceptions, are not harmful to
humans. Many bacteria in the TC group are soil bacteria but some, such as E. coli, originate in
the gut of warm-blooded animals. Coliform bacteria may be transported to surface water by
runoff or to ground water by infiltration. TCs are common in ambient water and may be injured
by environmental stresses such as lack of nutrients, and water treatment, such as chlorine
disinfection, in a manner similar to most bacterial pathogens and many viral enteric pathogens
(including fecal pathogens). EPA considers TC to be a useful indicator of a potential pathway
through which fecal contamination can enter the distribution system. The absence (versus the
presence) of TC in the distribution system indicates a reduced likelihood that fecal contamination
and/or waterborne pathogens are occurring in the distribution system.
Under the 1989 TCR, each TC-positive sample is assayed for either FC oris. coli. Fecal
coliform bacteria are a subgroup of TC that traditionally has been associated with fecal
contamination. Since the promulgation of the TCR, more information and understanding of the
suitability of FC and E. coli as indicators have become available. Study has shown that the FC
assay is imprecise and too often captures bacteria that do not originate in the human or mammal
gut (Edberg et al., 2000). On the other hand, E. coli is a more restricted group of coliform
bacteria that almost always originate in the human or animal gut (Edberg et al., 2000). Thus, E.
coli is a better indicator of fecal contamination than FC.
2.2.3 Sanitary Defects
As part of the RTCR, TC is used as an indicator to start an evaluation process that, where
necessary, will require the PWS to correct sanitary defects, which are defined as "a defect that
could provide a pathway of entry for microbial contamination into the distribution system or that
is indicative of a failure or imminent failure in a barrier that is already in place." "Sanitary
defect" is a term specific to the TCR assessment and correction provisions. Sanitary defects are
not intended to be linked directly to the "significant deficiencies" under the SWTR and GWR,
although some problems could meet either definition. The RTCR is not intended to limit the
existing authorities of primacy agencies under other regulations. Primacy agencies may allow for
integrated assessments meeting the requirements of multiple rules, where appropriate (see
Chapter 3 of this EA for more details on the provisions of the RTCR).
2.2.4 Occurrence of Fecal Contamination and/or Waterborne Pathogens
2.2.4.1 Presence of Fecal Contamination
Fecal contamination is a very general term that includes all of the organisms found in
feces, both pathogenic and nonpathogenic. Fecal contamination can occur in drinking water both
through use of contaminated source water as well as direct intrusion into the drinking water
distribution system. Biofilms in distribution systems may harbor waterborne pathogens and
accumulate enteric viruses and parasitic protozoa (Skraber et al., 2005; Helmi et al., 2008).
Economic Analysis for the Final RTCR
2-9
September 2012
-------
Waterborne pathogens harbored and accumulated in biofilms may have originally entered the
distribution system as fecal contamination from humans or animals.
EPA believes that E. coli is a meaningful indicator for fecal contamination and the
potential presence of associated pathogen occurrence. However, co-occurrence of indicators and
waterborne pathogens is difficult to measure. The analytical methods approved by EPA to assay
for E. coli do not specifically identify most of the pathogenic E. coli strains. There are at least
700 recognized E. coli strains. About 10 percent of the recognized E. coli strains are pathogenic
to humans (Feng, 1995; Hussein, 2007; Kaper et al., 2004). Because EPA-approved standard
methods for E. coli do not typically identify the presence of the pathogenic E. coli strains, an E.
6'o//-positive monitoring result is an indicator of fecal contamination but is not necessarily a
definitive measure of waterborne pathogen occurrence. Specialized assays and methods must be
used to identify waterborne pathogens, including pathogenic E. coli. These specialized assays
and methods are generally considered too costly and time consuming to be effective for routine
monitoring of drinking water distribution systems.
One notable exception is the data reported by Cooley et al. (2007), which showed high
concentrations of pathogenic E. coli strains in samples containing high concentrations of fecal
indicator E. coli. These data are from streams and other poor quality surface waters surrounding
California spinach fields associated with the recent E. coli 0157:H7 foodborne outbreak. Data
equivalent to these are not available from drinking water samples collected under the 1989 TCR.
2.2.4.2 Waterborne Disease Outbreaks
The CDC defines a waterborne disease outbreak as occurring when at least two persons
experience a similar illness, or one person is ill with amoebic meningioencephalitis, after the
ingestion of drinking water or following exposure to water used for recreational purposes, in
cases where the epidemiologic evidence implicates water as the probable source of the illness
(Kramer et al., 1996). The CDC maintains a database on waterborne disease outbreaks in the
United States. The database is based upon responses to a voluntary and confidential survey form
that is completed by state and local public health officials.
The National Research Council strongly suggests that the number of identified and
reported outbreaks in the CDC database for surface and ground waters represents only a small
percentage of the actual number of waterborne disease outbreaks (NRC, 1997; Bennett et al.,
1987, Hopkins et al., 1985 for Colorado data). Underreporting occurs because most waterborne
outbreaks in community water systems are not recognized until a sizable proportion of the
population is ill (Perz et al., 1998; Craun, 1996), perhaps 1 to 2 percent of the population (Craun,
1996).
EPA drinking water regulations are designed to protect against endemic waterborne
disease and to minimize waterborne outbreaks. Endemic waterborne disease may be defined as
any waterborne disease not associated with an outbreak.
Economic Analysis for the Final RTCR
2-10
September 2012
-------
2.3 Statutory Authority for Promulgating the Rule
As mentioned previously, SDWA requires EPA to review and revise, as appropriate, each
existing NPDWR at least once every six years (SDWA section 1412(b)(9), 42 U.S.C. § 300g-
1(b)(9)). Section 1412(b)(9) of SDWA states:
The Administrator shall, not less often than every 6 years, review and revise, as
appropriate, each national primary drinking water regulation promulgated under this title.
Any revision of a national primary drinking water regulation shall be promulgated in
accordance with this section, except that each revision shall maintain, or provide for
greater, protection of the health of persons.
In 2003, EPA completed its review of the 1989 TCR and 68 NPDWRs for chemicals that
were established prior to 1997 (USEPA 2002b, 67 FR 19030, April 17, 2002). In the Six-Year
Review determination published in July 2003, EPA stated its intent to revise the 1989 TCR
(USEPA 2003b, 68 FR 42907, July 18, 2003).
2.4 Economic Rationale
As a revision to an existing regulation, the primary goal of the RTCR is to achieve the
objectives of the 1989 TCR more effectively and efficiently, taking into account the changes in
regulatory framework for implementing the SDWA over the past 20 years and experience with
the 1989 TCR since promulgation. In this context, the overall economic rationale behind the
promulgation of the RTCR is the same as the rationale for the promulgation of the 1989 TCR.
This section addresses the economic rationale for choosing a regulatory approach under the
RTCR (and 1989 TCR) rather than non-regulatory options.
An economic rationale for the rule is required by Executive Order 12866, Regulatory
Planning and Review (58 FR 51735, October 1993), which states:
"[E]ach agency shall identify the problem that it intends to address (including, where
applicable, the failures of the private markets or public institutions that warrant new
agency action) as well as assess the significance of that problem." (Section 1, b(l))
In addition, Office of Management and Budget (OMB) guidance, dated January 11, 1996,
states that "in order to establish the need for the proposed action, the analysis should discuss
whether the problem constitutes a significant market failure" (OMB, 1996).
In a perfectly competitive market, prices and quantities are determined solely by the
aggregated decisions of buyers and sellers. Such a market occurs when many producers of a
product are selling to many buyers, and where both producers and consumers have perfect
information on the characteristics and prices of each firm's products. Barriers to entry in the
industry cannot exist, and individual buyers and sellers must be "price takers" (i.e., their
individual decisions cannot affect the price). Several properties of the public water supply do not
satisfy the conditions for a perfectly competitive market and thus lead to market failures that
require regulation.
Economic Analysis for the Final RTCR
2-11
September 2012
-------
Many PWSs are natural monopolies. A natural monopoly exists when it is impossible for
more than one firm in each area to recover the costs of production and survive. There are high
fixed costs associated with reservoirs and wells, transmission and distribution systems, treatment
plants, and other facilities. For other potential suppliers to enter the market, they would have to
provide the same extensive infrastructure to realize similar economies of scale and be
competitive. A splitting of the market with increased fixed costs (e.g., two supplier networks in a
single market) usually makes this situation unprofitable. The result is a market suitable for a
single supplier and hostile to alternative suppliers. In such natural monopolies, suppliers have
fewer incentives for providing quality services or maintaining competitive prices. In these
situations, governments often intervene to help protect the public interest.
For example, because PWSs are legal, as well as natural, monopolies, they are often
subject to price controls, if not outright public ownership. While customers may demand
improvements in water quality, the regulatory structure may not facilitate the transmission of that
demand to the water supplier or allow the supplier to raise its price to recover the cost of the
improvements. If consumers do not believe that their drinking water is safe enough, they cannot
simply switch to another water utility. Other options for obtaining safe drinking water (e.g.,
buying bottled water or installing point of use filtration) most often represent a higher water cost
to consumers than the purchase from PWSs. Therefore, the water supplier may have little
incentive to improve water quality.
The public may also not understand the health and safety issues associated with poor
drinking water quality. Understanding the health risks posed by trace quantities of drinking water
contaminants involves analysis and synthesis of complex toxicological and health sciences data.
Therefore, the public may not be aware of the risks it faces. EPA has implemented a Consumer
Confidence Reports (CCR) Rule (USEPA, 1998b, 63 FR 44512, August 1998) that makes water
quality information more easily available to consumers. This rule requires CWSs to publish an
annual report on local drinking water quality. Consumers, however, still have to analyze this
information for its health risk implications. Furthermore, even if informed consumers are able to
engage PWSs in a dialogue about health issues, the transaction costs of such interaction
(measured in personal time and monetary outlays) present another significant impediment to
consumer expression of risk reduction preferences.
SDWA regulations are intended to provide a level of protection from exposure to
drinking water contaminants. The regulations set minimum performance requirements to protect
consumers from exposure to contaminants. SDWA regulations are not intended to restructure
market mechanisms or to establish competition in supply; rather, they establish the level of
service to be provided that best reflects public preference for safety. The federal regulations
reduce the high information and transaction costs by acting on behalf of consumers in balancing
risk reduction and the social costs of achieving this risk reduction.
Economic Analysis for the Final RTCR
2-12
September 2012
-------
3 Consideration of Regulatory Options
3.1 Introduction
This chapter describes the regulatory options considered during the development of the
Revised Total Coliform Rule (RTCR) and evaluated as part of this economic analysis (EA).
3.2 Total Coliform Rule/Distribution System Advisory Committee
In the July 2003 NPDWR review decision, EPA provided public notice of its intent to
revise the 1989 TCR. In 2007, EPA decided to establish a committee under the Federal Advisory
Committee Act (FACA) called the Total Coliform Rule/Distribution System Advisory
Committee (TCRDSAC) (USEPA, 2007a). The advisory committee was charged with
developing an Agreement in Principle (AIP) 7 with recommendations for the revision of the 1989
TCR. The advisory committee was to also provide EPA with recommendations needed to
understand and address possible public health impacts from potential degradation of drinking
water quality in distribution systems. Specifically, the major objectives of the advisory
committee were to provide advice and recommendations on:
• Revisions to the 1989 TCR that would improve implementation while maintaining
or improving public health protection and distribution system water quality.
Examples of the issues the TCRDSAC considered include: the 1989 TCR
monitoring framework, sanitary survey provisions, definition of maximum
contaminant level (MCL) violations and potential follow-up corrective actions, and
public notification of violations.
• What data should be collected, research conducted, and/or risk management
strategies evaluated to better project distribution system contaminant occurrence
and associated public health risks in the distribution systems. This was intended to
"initiate a process for addressing cross connection control and backflow prevention
requirements and consider additional distribution system requirements related to
significant health risks" as recommended by the M/DBP Federal Advisory
Committee.
EPA held a series of 13 meetings of the advisory committee between July 2007 and
September 2008. The 15 committee members, along with other stakeholders, discussed options
for revising the 1989 TCR and began the process of developing the RTCR by discussing the
purpose, efficacy and applicability of the 1989 TCR, as well as data collection and research
needed to better understand potential distribution system risks. The advisory committee formed a
Technical Work Group (TWG) to provide data analysis and information to inform the advisory
committee's discussions. The advisory committee also discussed the relevant provisions of the
Ground Water Rule (GWR) and discussed the extent to which other rules contribute to the
objectives of the 1989 TCR. All advisory committee members signed the AIP in September
2008. All of the recommendations of the AIP, on which all the members of the committee
7 Under the proposed rule, the preferred regulatory option was called Agreement in Principle (AIP). Under the final
rule, the preferred regulatory option is called the RTCR.
Economic Analysis for the Final RTCR
3-1
September 2012
-------
agreed, are found in the signed AIP (USEPA 2009, 74 FR 1683, January 13, 2009). The AIP and
details about the advisory committee can be found at EPA's website at:
http://water.epa.gov/lawsregs/rulesregs/sdwa/tcr/regulation revisions tcrdsac.cfm.
3.3 Regulatory Options Considered
EPA evaluated three regulatory options as part of the development of the RTCR—the
1989 TCR, the RTCR, and the Alternative option. The following discussion provides an
overview of the three regulatory options considered followed by a detailed table comparing
specific regulatory components. A detailed evaluation and comparison of the benefits and costs
of each of the three regulatory options for the RTCR is found in later chapters of this EA.
Baseline Option: 1989 Total Coliform Rule
The baseline option for the EA is the 1989 TCR as written, with adjustments made to
reflect the requirements of the GWR that PWSs must comply with beginning December 1, 2009.
Analysis of this option is primarily based on data collected under the 1989 TCR, as described in
Chapter 4 (Baseline Conditions) of this EA.
The 1989 TCR requires public water systems (PWSs) to monitor routinely for total
coliforms (TC) and fecal coliforms (FC) oris. coli. Monitoring requirements are based on system
type, population served by the PWS and source water type. Each PWS must sample according to
a written sample siting plan. Plans are subject to state review and revision.
The 1989 TCR specifies a maximum contaminant level goal (MCLG) of zero for TC
(including FC and E. coli) as well as an MCL where compliance is based on the presence or
absence of TC and/or FC oris, coli in the samples. A PWS is required to take repeat samples
following a total coliform-positive (TC+) sample and to test the TC+ sample for E. coli. A
monthly MCL violation is triggered if: (1) a PWS collecting fewer than 40 samples per month
has >1 TC+ sample per month; or (2) a PWS collecting at least 40 samples per month has >5.0%
TC+ samples per month. An acute MCL violation is triggered if any PWS has any fecal
coliform- oris, coli -positive (FC+ / EC+) repeat sample or has a FC+ or EC+ routine sample
followed by a TC+ repeat sample. A PWS must demonstrate compliance with the MCL for TC
each month it serves water to the public (or each calendar month that sampling occurs for PWSs
on reduced monitoring). MCL violations must be reported to the state no later than the end of the
next business day after the PWS learns of the violation. The public must also be notified
depending on the severity of the MCL violation (within 30 days for monthly MCL violations and
within 24 hours for acute MCL violations).
RTCR
The RTCR is designed to trigger PWSs that exceed specified levels of TC/E. coli in their
finished water to do an assessment, to identify whether sanitary defects are present, and to
correct such defects accordingly. EPA believes that the RTCR is an improvement over the 1989
TCR framework because it takes a more preventive approach to identifying and fixing problems
with public health implications. Under the RTCR, EPA establishes an MCLG of zero for is. co//',
an MCL for is. coli based on TC and is. coli monitoring results, and a coliform treatment
technique for protection against potential fecal contamination.
Economic Analysis for the Final RTCR
3-2
September 2012
-------
On July 14, 2010, EPA proposed a RTCR that had the same elements and effect as the
recommendations in the advisory committee AIP (USEPA 2010b, 75 FR 40926, July 14, 2010).
The final RTCR was developed based on the proposed rule and the comments received on it. The
RTCR maintains the 1989 TCR requirement for all PWSs to monitor for TC and E. coli
according to a sample siting plan and schedule specific to the PWS. PWSs are required to take
repeat samples and to test for E. coli following a TC+ sample.
Under the RTCR, the PWS monitoring frequencies take into account the unique
characteristics of various PWS types and sizes. Small, well-operated PWSs using ground water
may be able to reduce monitoring frequencies by meeting specific criteria, thus reducing their
monitoring and reporting burden. Criteria for reduced monitoring includes monitoring results
that reflect a clean compliance history for a minimum of 12 months as well as preventive
practices designed to continue to maintain the integrity of the distribution system.
The RTCR requires PWSs to complete either Level 1 or Level 2 assessments following
certain triggers (see Exhibit 3.1). The purpose of Level 1 and 2 assessments is to identify the
presence of sanitary defects and deficiencies in distribution system coliform monitoring
practices. The minimum elements of both Level 1 and 2 assessments include review and
identification of:
• Atypical events that may affect distributed water quality or indicate that distributed
water quality was impaired;
• Changes in distribution system maintenance and operation that may affect
distributed water quality (including water storage);
• Source and treatment considerations that bear on distributed water quality, where
appropriate (e.g., whether a ground water system is disinfected)
• Existing water quality monitoring data; and
• Inadequacies in sample sites, sampling protocol, and sample processing.
A Level 1 assessment (completed by the PWS) consists of a simple examination of the
system and relevant operational practices. The PWS must complete the assessment form and
submit it to the state for review within 30 days after determination that the PWS has exceeded
the Level 1 trigger. The completed assessment form must include assessments conducted, all
sanitary defects identified (or indicate if no sanitary defects were found), corrective actions
completed, and a timetable for any corrective actions not already completed. Upon completion
and submission of the assessment form by the PWS, the state will determine if the PWS has
identified a likely cause for the Level 1 trigger and determine whether the PWS has corrected the
problem, or has included a schedule acceptable to the state for correcting the problem. Whether
or not the system has identified any sanitary defects or a likely cause for the Level 1 trigger, the
state may determine whether or not the Level 1 assessment is sufficient, and if it is not the state
must discuss its concerns with the system. The state may require revisions to the assessment after
the consultation. If the state requires revision after consultation, the PWS must submit a revised
assessment form to the state on an agreed-upon schedule not to exceed 30 days from the date of
consultation. Upon completion and submission of the assessment form by the PWS, the state
Economic Analysis for the Final RTCR
3-3
September 2012
-------
must determine if the PWS has identified a likely cause for the Level 1 trigger and, if so,
establish that the PWS has corrected the problem, or has included a schedule acceptable to the
state for correcting the problem.
The Level 2 assessment must be conducted either by the state, a third party approved by
the state, or the PWS where the PWS has the staff or management with the required certification
or qualifications specified by the state. A Level 2 assessment is a more detailed examination of
the system's monitoring and operational practices than a Level 1 assessment. As with the Level 1
assessment, the completed assessment form for a Level 2 assessment must also be submitted to
the state for review within 30 days after determination that the PWS has exceeded the Level 2
trigger. The PWS must also indicate in the completed assessment form the sanitary defects
detected (or indicate if no sanitary defects were found), corrective actions completed, and a
timetable for any corrective actions not already completed. Upon completion and submission of
the assessment form by the PWS, the state will determine if the PWS has identified a likely cause
for the Level 2 trigger and determine whether the PWS has corrected the problem, or has
included a schedule acceptable to the state for correcting the problem. Whether or not the system
has identified any sanitary defects or a likely cause for the Level 2 trigger, the state may
determine whether or not the Level 2 assessment is sufficient, and if it is not the state must
discuss its concerns with the system. The state may require revisions to the assessment after the
consultation. If the state requires revision after consultation, the PWS must submit a revised
assessment form to the state on an agreed-upon schedule not to exceed 30 days from the date of
consultation. Upon completion and submission of the assessment form by the PWS, the state
must determine if the PWS has identified a likely cause for the Level 2 trigger and, if so,
establish that the PWS has corrected the problem, or has included a schedule acceptable to the
state for correcting the problem.
The RTCR requires PWSs to correct sanitary defects found through either a Level 1 or
Level 2 assessment. The 1989 TCR does not require PWSs that have TC MCL violations to
perform corrective actions. For corrections that are not completed by the time the PWS submits
the completed assessment form to the state, the PWS must complete the corrective action(s) on a
schedule determined by the state in consultation with the PWS. The PWS must notify the state
when each scheduled corrective action is completed.
The RTCR specifies violations corresponding to different degrees of potential public
health concern and public notification. A violation of an E. coli MCL occurs when:
• A routine sample is TC+ and one of its associated repeat samples is EC+; or
• A routine sample is EC+ and one of its associated repeat samples is TC+ or EC+; or
• A system fails to test for E. coli when any repeat sample is TC+; or
• A system fails to take all required repeat samples following a routine sample that is
EC+.
An E. coli MCL violation requires a Tier 1 public notification. A coliform treatment
technique violation occurs when the PWS exceeds a treatment technique trigger (see Exhibit 3.1)
and then fails to conduct the required assessment or corrective action within the required
Economic Analysis for the Final RTCR
3-4
September 2012
-------
timeframes. A coliform treatment technique violation requires a Tier 2 public notification. If the
system does conduct the assessment and satisfies the requirements of the coliform treatment
technique (including corrective action when a sanitary defect is identified), no public notification
is required. A monitoring violation occurs when a PWS fails (1) to take every required routine or
additional routine sample in a compliance period, or (2) to analyze for E. coli following a TC+
sample. A reporting violation occurs when (1) a PWS properly conducts monitoring or an
assessment but fails to submit a monitoring report or a completed assessment form, or (2) a PWS
fails to notify the state following an EC+ sample. Both monitoring violations and reporting
violations require a Tier 3 public notification, but community water systems can use their annual
consumer confidence report to complete the notification.
The provisions of the RTCR consider the implication and linkages to other rules
promulgated by EPA under SDWA (e.g., GWR, Surface Water Treatment Rules (SWTRs), and
the Disinfectant/ Disinfection Byproduct (DBP) Rules (USEPA, 1998c)). Compliance activities
for other rules, such as the sanitary survey, are used under the RTCR as criteria to assess the
integrity of the PWS and to implement the reduced monitoring provisions in a cost-effective
manner. By January 2015 ground water PWSs will have completed at least one round of the
sanitary survey component of the GWR. Therefore, sanitary survey findings can be used by
states in the determination of the monitoring frequency.
Alternative Option: All PWSs Initially Sample for TC/EC on Monthly Basis
The third option is the Alternative option which parallels the RTCR in most ways but
includes variations of some of the provisions that were discussed by the advisory committee
before consensus was reached on the RTCR. Under the Alternative option, at the compliance
date, all PWSs would be required to sample monthly for an initial period until they meet the
eligibility criteria for reduced monitoring. This more stringent approach differs from the RTCR.
Under the RTCR PWSs are allowed to continue to monitor at their current frequencies (with an
additional annual site visit / voluntary Level 2 assessment for PWSs wishing to remain on annual
monitoring) until they are triggered into an increased sampling frequency. Under the Alternative
option, however, no PWSs would be allowed to reduce monitoring to an annual basis. EPA
defined the Alternative option this way and included it in the economic analysis to assess the
relative impacts of a more stringent rule and to better understand the balance between costs and
public health protection.
3.3.1 Comparative Summary of Regulatory Options
Exhibit 3.1 below summarizes the components of each regulatory option considered for
the RTCR.
Economic Analysis for the Final RTCR
3-5
September 2012
-------
Exhibit 3.1 Comparison of RTCR Regulatory Options
Element
1989 TCR
RTCR
Alternative option
Rule construct
• MCLG of zero forTC (including
FC and EC).
• TC monthly MCL based on the
number/percent of TC+ samples
in a month.
• FC/EC acute MCL based on
FC+/EC+ samples.
• Public Notification required for
MCL violations, including the TC
monthly MCL violation based only
on TC occurrence.
• No MCLG/MCL for TC, FC no longer used.
• EC MCLG of zero, and an EC MCL and a coliform treatment
technique based on TC and/or EC results.
• Assessment (and corrective action if necessary) required if
PWS has a coliform treatment technique trigger.
• Public Notification not required for only TC occurrence.
Same as RTCR.
Transition to the
New Rule
N/A
• PWSs continue on their 1989 TCR monitoring schedule
provided they meet criteria.
• Noncommunity Water Systems (NCWSs) on quarterly/annual
monitoring remain on that schedule unless/until they have an
event that triggers increased monitoring.
• Community Water Systems (CWSs) on reduced monitoring
remain on that schedule unless/until they have an event that
triggers return to routine monitoring.
• Monitoring schedules will be evaluated during the "special
monitoring evaluation" conducted by the state as part of the
periodic sanitary survey to determine if the monitoring
frequency is appropriate. The first such evaluation must be
conducted during the first scheduled sanitary survey after the
effective date of the rule; a system may remain on its 1989
TCR monitoring schedule until this time unless it is triggered
into more frequent monitoring.
No transition. PWSs
start new requirements
as soon as the rule is
effective.
Economic Analysis for the Final RTCR
3-6
September 2012
-------
Element
1989 TCR
RTCR
Alternative option
Routine
Monitoring
• 1 sample per quarter for NCWS
<1,000 Ground Water (GW).
• 1 sample per month for NCWS
<1,000 Surface Water (SW) and
all CWS <1,000.
• For all PWS >1,000, the number
of samples per month is based
on population served and
specified in the rule.
• 1 sample per month for seasonal NCWS <1,000 (GW)8 with
criteria to qualify for reduced monitoring. Seasonal systems
must also demonstrate completion of a state-approved start-
up procedure.
• For all others, same as 1989 TCR, with more explicit criteria
to qualify for reduced monitoring (see "Reduced Monitoring" in
this exhibit).
A minimum frequency
of monthly monitoring
for all PWSs.
Reduced
Monitoring
• NCWS <1,000 (GW) can reduce
to 1 sample per year if system is
free of sanitary defects.
• CWS <1,000 (GW) can reduce to
1 sample per quarter if no history
of TC contamination, no sanitary
defects, and protected GW
source.
• No other systems are eligible for
reduced monitoring.
• NCWS < 1,000 (GW)—same as in 1989 TCR, but more
criteria to qualify and remain on reduced annual monitoring,
o Most recent sanitary survey shows that system is free of
sanitary defects,9 has a protected water source, and meets
approved construction standards;
o Clean compliance history for a minimum of 12 months;
o Level 2 assessment by party approved by state within the
last 12 months and correction of all identified sanitary
defects. System must also have an annual site visit every
year thereafter to remain on annual monitoring.
• Seasonal NCWS < 1,000 (GW) can be eligible for reduced
monitoring by having an approved sample site plan that
designates the time period for monitoring based on site-
specific considerations (e.g., during periods of highest
demand or highest vulnerability to contamination). The
system must collect compliance samples during this time.
o For reduced quarterly monitoring the seasonal system must
have a sanitary survey or site visit or Level 2 assessment
within last 12 months; a protected water source; a clean
• NCWS < 1,000 (GW)
(including seasonal
syste ms)—red u ced
quarterly monitoring
(criteria same as
RTCR). Reduced
annual monitoring is
not allowed.
• CWS < 1,000
(GW)—same as
RTCR.
• No other systems
are eligible for
reduced monitoring.
8 A seasonal system is a non-community water system that is not operated as a public water system on a year-round basis and starts up and shuts down the system
at the beginning and end of each operating season (§141.851 of the RTCR).
9 The RTCR defines a sanitary defect as "a defect that could provide a pathway of entry for microbial contamination into the distribution system or that is
indicative of a failure or imminent failure in a barrier that is already in place"
Economic Analysis for the Final RTCR
3-7
September 2012
-------
Element
1989 TCR
RTCR
Alternative option
compliance history for a minimum of 12 months, and be
free of sanitary defects,
o To reduce to 1 sample per year, the seasonal system must
meet the criteria specified above for quarterly monitoring
and have in place or adopt one or more additional
enhancements to barriers to contamination (cross
connection control, certified operator, meet disinfection
criteria, maintenance of at least 4-log removal or
inactivation of viruses, other equivalent enhancements).
• CWS < 1,000 (GW)—same as in 1989 TCR, but more criteria
to qualify and remain on reduced quarterly monitoring.
o State certified operator;
o Most recent sanitary survey shows that system is free of
sanitary defects (or has an approved plan and schedule to
correct them), has a protected water source, and meets
approved construction standards;
o Clean compliance history for a minimum of 12 months;
o Meets at least one of the following criteria: annual site visit
by the state or a voluntary Level 2 assessment by a party
approved by the state or meeting criteria established by the
state; cross connection control; meet disinfection criteria;
maintenance of at least 4-log removal or inactivation of
viruses; other equivalent enhancements to water systems
as approved by the state.
• No other systems are eligible for reduced monitoring.
Increased
Monitoring
(NCWS) and
Return to Routine
Monitoring (CWS)
• N/A (none specified)—NCWS.
• N/A (none specified)—CWS
• NCWS < 1,000 (GW only) increases from quarterly or annual
monitoring to monthly monitoring if one of the following
occurs:
o Triggered Level 2 assessment or a 2nd Level 1 assessment
in 12 months;
o EC MCL violation;
o Treatment technique violation;
o Two monitoring violations within 12 months if on quarterly
Same as RTCR with
the exception that
reduced annual
monitoring is not
allowed under this
option.
Economic Analysis for the Final RTCR
3-8
September 2012
-------
Element
1989 TCR
RTCR
Alternative option
monitoring; or
o One monitoring violation and one Level 1 assessment
within 12 months if on quarterly monitoring.
• NCWS < 1,000 (GW only) increases from quarterly
monitoring to monthly monitoring if it has one monitoring
violation and one Level 1 assessment within 12 months.
• NCWS < 1,000 (GW only) increases from annual monitoring
to quarterly monitoring if it has one monitoring violation.
• CWS < 1,000 (GW only) on quarterly monitoring return to
monthly monitoring based on the first four criteria that trigger
a NCWS < 1,000 (GW only) to increased its monitoring to
monthly, as listed above.
Return to
Reduced
Monitoring (After
Being Triggered
to Increased
Monitoring)
N/A
• NCWS < 1,000 (GW) must meet the following criteria to return
to routine quarterly monitoring after being triggered to
increased monitoring:
o Within the last 12 months, system must have completed a
sanitary survey or a site visit by the state or a voluntary
Level 2 assessment, must be free of sanitary defects, and
must have a protected water source; and
o Clean compliance history for a minimum of 12 months.
• NCWS < 1,000 (GW) must meet the following criteria to return
to reduced annual monitoring in addition to meeting the
criteria for returning to routine quarterly monitoring:
o An annual site visit by the state or a voluntary Level 2
assessment and correction of all identified sanitary defects;
and
o Adoption of one or more additional enhancements to the
water system barriers to contamination (cross connection
control, certified operator, meet disinfection criteria,
maintenance of at least 4-log removal or inactivation of
viruses, other equivalent enhancements).
• CWS < 1,000 (GW) must meet the same criteria for qualifying
for reduced quarterly monitoring.
Same as RTCR with
the exception that
reduced annual
monitoring is not
allowed under this
option.
Economic Analysis for the Final RTCR
3-9
September 2012
-------
Element
1989 TCR
RTCR
Alternative option
Repeat
Monitoring
• PWS serving >1,000 people must
take 3 repeat samples after a
TC+ sample.
• PWS serving <1,000 must take 4
repeat samples after a TC+
sample.
o For ground water systems, 1
sample can be a source
water sample to comply with
the GWR.
• All PWSs must take 3 repeat samples after a TC+ sample.
o PWS < 1,000, the number of repeat samples is reduced
from 4 samples to 3.
• GW PWS must still take an additional source sample to
comply with the GWR.
• For GW PWS < 1,000, a single sample can meet both the
triggered source water requirements of the GWR and the
repeat sample requirements of the RTCR, but only if the state
approves the use of the single sample to meet both rule
requirements and the use of EC as the fecal indicator.
Same as RTCR.
Additional
Routine
Monitoring
PWS taking <5 routine samples per
month (PWS serving < 4,100) must
take at least 5 routine samples the
month after a TC+ sample.
• For PWS taking samples less than monthly, the number of
samples required after a TC+ is reduced from 5 to 3.
• For PWS taking at least 1 sample per month, the additional
routine sample requirement is eliminated (they take their
usual number of samples the following month).
Same as RTCR.
Sample Siting
Plan
• Sampling must occur at sites
representative of water quality in
the distribution system.
• Sample siting plans are subject to
state review/ revision.
• Special purpose samples are not
used for determining compliance.
• Take at least one repeat sample
within 5 connections up- and
downstream of the TC+ site.
Same as 1989 TCR except:
• Specifically allows for dedicated sampling stations.
• Provides more flexibility for systems in determining the
locations for taking repeat samples. For systems that want to
establish repeat sampling locations other than the within-five-
connections-upstream-and-downstream of the TC-positive
sample, the system must submit the sample siting plan for
review and the state may modify the sampling locations as
needed.
• States may allow entry point sampling at ground water
systems if the overall plan remains representative of water
quality in the distribution system.
Same as RTCR.
Assessment
N/A—none required under the
1989 TCR.
• The PWS must conduct a Level 1 (self-assessment) if it
exceeds any of the following triggers:
o For systems taking >40 samples per month, the PWS
Same as RTCR.
Economic Analysis for the Final RTCR
3-10
September 2012
-------
Element
1989 TCR
RTCR
Alternative option
exceeds 5.0% TC+ samples for the month; or
o For systems taking <40 samples per month, the PWS has >
2 TC+ samples for the month; or
o The PWS fails to take every required repeat sample after
any single routine TC+ sample.
• The PWS must ensure that a Level 2 assessment is
conducted either by the state or a state-approved 3rd party
(including qualified PWS employee) if it exceeds any of the
following triggers:
o The PWS has an E. coli MCL violation.
o The PWS has a second Level 1 trigger within a rolling 12-
month period, or in 2 consecutive years for systems on
annual monitoring.
• Assessment results and description of corrective action taken
will be submitted to the state within 30 days.
Corrective Action
N/A—none required under the
1989 TCR.
• System must correct all sanitary defects found in the
assessment.
• State must be satisfied with the assessment.
• For corrections not completed by the time the assessment
form is submitted, the systems must be in compliance with a
state determined schedule and must notify the state when
completed.
Same as RTCR.
Economic Analysis for the Final RTCR
3-11
September 2012
-------
Element
1989 TCR
RTCR
Alternative option
Violations and
Public
Notification
• EC/FC MCL violation—acute
violation, Tier 1 Public
Notification.
• Monthly TC MCL violation—Tier
2 Public Notification.
• Monitoring or reporting
violation—Tier 3 Public
Notification.
• PWS must notify state regarding
single EC+/FC+ result.
• EC MCL violation—Tier 1 Public Notification. Failure to take
repeat samples following an EC+ is also an EC MCL violation.
• Monthly TC MCL violation is dropped—conditions that trigger
a monthly TC MCL violation under the TCR trigger
assessment and corrective action instead.
• Coliform treatment technique violation occurs when a PWS
fails to conduct required assessment and corrective action—
Tier 2 Public Notification.
• Monitoring and reporting violations will be tracked separately
—Tier 3 Public Notification.
Same as RTCR.
Economic Analysis for the Final RTCR
3-12
September 2012
-------
3.4 Final Rule Requirements
Based on the recommendations in the AIP, as well as the comments received on the
proposed RTCR during the public comment period, EPA selected the RTCR as the preferred
regulatory option. The RTCR was developed by the EPA with the primary goal of achieving the
objectives of the 1989 TCR more effectively and efficiently, taking into account the changes in
regulatory framework for implementing the SDWA over the past 20 years and experience with
the TCR since it was promulgated in 1989. Additional discussion supporting the selection of the
RTCR as the preferred regulatory option in terms of benefits and costs is provided in Chapters 6
(benefits), 7 (costs), and 9 (benefits and costs). For a detailed discussion of the requirements of
the RTCR, see section III of the RTCR preamble and subpart Y of 40 CFR part 141.
Economic Analysis for the Final RTCR
3-13
September 2012
-------
4 Baseline Conditions
4.1 Introduction
A primary step in the development of the economic analysis (EA) for the Revised Total
Coliform Rule (RTCR) was to estimate the baseline conditions. The U.S. Environmental
Protection Agency (EPA) used available data to develop an occurrence and predictive model for
public water systems (PWSs) serving 4,100 people or fewer based primarily on the 2005 Six-
Year Review 2 data to predict total coliform (TC) and E. coli occurrence, Level 1 and Level 2
assessment triggers, corrective actions, and violations, both in the baseline and over time. EPA
developed another, simpler, predictive model for PWSs serving more than 4,100 people based
primarily on the 2005 Safe Drinking Water Information System/Federal Version (SDWIS/FED)
violation data (USEPA, 2005) that predicts triggers, assessments, corrective actions, and
violations both in the baseline and over time, but not TC and E. coli occurrence. Because five
years of Ground Water Rule (GWR) implementation prior to the effective date of the RTCR are
expected to cause changes to ground water systems, the baseline conditions that EPA developed
for the EA account for the effects of the GWR, as described in this chapter and Chapter 5. The
remainder of this chapter describes the data sources used to develop the baselines and how the
data were used. The resulting estimate of baseline conditions serves as a reference point for
understanding net impacts of the RTCR, as discussed in Chapters 6, 7, and 9 of this EA.
4.1.1 Background and Purpose
The baseline analysis is a characterization of the water industry and its current operations
in effect before systems make changes to meet requirements of the RTCR. This chapter presents
estimates of the number of systems, the population affected, water quality measures, and
subpopulations affected. The data collected for this profile serves as the baseline in the RTCR
EA. The development of the baseline analysis consisted of the following steps:
• Compilation of a profile of water systems—identifying and collecting information
on all PWSs.
• Characterization of current monitoring requirements, including testing
requirements, sample siting plans, routine and repeat sampling requirements, and
sanitary surveys.
• Characterization of current PWS compliance information, including acute and
monthly maximum contaminant level (MCL) violations, triggered monitoring, and
public notification reporting.
4.1.2 Chapter Organization
The remainder of this document is organized into four general sections:
• Section 4.2 describes general information on data sources.
• Section 4.3 describes the baseline conditions.
Economic Analysis for the Final RTCR
4-1
September 2012
-------
• Section 4.4 estimates baseline impact on sensitive populations.
• Section 4.5 presents a summary of assumptions used to develop baseline estimates,
associated uncertainty and variability, and potential impacts on the results of the
RTCR EA.
4.2 Data Sources
Several primary data sources were used for the EA to characterize the RTCR baseline to
create system and population baselines, including the SDWIS/FED (USEPA, 2007b), the 2005
Six-Year Review of National Primary Drinking Water Regulations (Six-Year Review 2), the
GWR EA, and the U.S. Census. Each of these data sources is further explained in Sections 4.2.1-
4.2.3 below.
4.2.1 Background on SDWIS/FED Data
SDWIS/FED10 is EPA's national regulatory compliance database for the drinking water
program and is the main source of PWS inventory and violation data for the RTCR baseline.
SDWIS/FED contains information on each of the approximately 155,000 active PWSs as
reported by primacy agencies,11 EPA Regions, and EPA headquarters personnel. SDWIS/FED
does not include sample results, but does include identification of MCL violations and
12
monitoring and reporting violations (both routine and repeat, and minor and major). It also
contains information to characterize the U.S. inventory of PWSs, namely: system name and
location; retail population served; source water type (i.e., ground water (GW), surface water
(SW), or ground water under the direct influence of surface water (GWUDI)); whether or not
systems disinfect the water; and PWS type, as described below.
10 Further information on SDWIS/FED can be found on EPA's website:
http://www.epa.gov/safewater/databases/sdwis/index.html.
11 States and Indian Tribes are given primary enforcement responsibility (primacy) for regulations pertaining to
PWSs in their State/jurisdiction if they meet certain requirements specified under the primacy regulations at
40CFR142, Subpart B.
12
Under the 1989 Total Coliform Rule, violation categories are defined as follows (category names in bold). A non-
acute (monthly) MCL violation occurs when >1 routine and/or repeat sample are total coliform-positive (TC+) for
PWSs that take fewer than 40 samples monthly, or greater than 5.0% of monthly samples are TC+ for PWSs that
take at least 40 samples. An acute MCL violation is triggered if a PWS has a fecal coliform-positive (FC+) or E.
coli-positive repeat sample following a TC+ or has a FC+/EC+ routine sample followed by a TC+ repeat sample. A
monitoring violation occurs when a PWS 1) does not satisfy the sample siting plan requirement; 2) does not sample
in accordance with its required schedule; 3) does not analyze a TC+ routine sample for FC or E. coli\ 4) does not
collect repeat samples within 24 hours and analyze for TC following a routine TC+; or 5) does not test repeat
samples that are TC+ for FC or E. coli. If a PWS fails to conduct some of the required routine samples in a
compliance period, the PWS incurs a minor routine monitoring violation. Equivalently, a PWS that fails to
conduct some or all of its required repeat samples following a TC+ routine sample incurs a minor repeat
monitoring violation or a major repeat monitoring violation, respectively. Reporting violations occur when a
PWS does not report to the state or to the public according to the schedule for each violation type. For acute MCL
violations, a PWS must report to the state by end of day and to the public within 24 hours (by posting, hand delivery,
or mass media subject to primacy agency approval). For monthly MCL violations, a PWS must report to the state by
end of next day, and to the public within 30 days by mail, hand delivery, or other methods approved by the primacy
agency. For monitoring violations, a PWS must notify the public with 1 year in accordance with general public
notification requirements approved by the primacy agency.
Economic Analysis for the Final RTCR
4-2
September 2012
-------
EPA defines a PWS as a system that provides water for human consumption through
pipes or other constructed conveyances if such a system has at least 15 service connections or
regularly serves an average of at least 25 individuals per day for at least 60 days per year. PWSs
are categorized as follows:
• Community water systems (CWS) are PWSs that supply water to the same
population year-round.
• Noncommunity water systems (NCWS) are PWSs that supply water to a varying
population or one that is served less than year-round. They are sub-categorized as
follows:
- Nontransient noncommunity water systems (NTNCWS) are PWSs that are
not CWSs and that regularly supply water to at least 25 of the same people
at least six months per year, for example, schools.
- Transient noncommunity water systems (TNCWS) are NCWSs that
provide water in places such as gas stations or campgrounds where people
do not remain for long periods of time.
Additionally, PWS are analyzed in this EA according to categories defined by the number
of people they serve. The following nine categories of populations served by PWSs are used
throughout this document:
• <100
• 101-500
• 501-1,000
• 1,001-4,100
• 4,101-33,000
• 33,001-96,000
• 96,001-500,000
• 501,001-1 Million
• Over 1 Million
These population categories are mostly consistent with those analyzed for other rules and
consider the distinctions in cost and system operation that are meaningful when considering the
economic effects of rule requirements. In particular, under the 1989 TCR, PWSs serving more
than 1,000 people all test for TC monthly; PWSs serving fewer than 1,000 people may monitor
quarterly or even annually and (assuming that they collect fewer than 5 samples per month) must
have a sanitary survey every 5 years. The 33,000 threshold is significant as it corresponds with
sampling requirements (at least 40 samples per month are taken) that affects how compliance
with the rule is calculated. (PWSs serving >33,000 persons must take 40 or more samples in a
month, and may trigger an MCL violation based on the percentage of total coliform-positive
(TC+) samples, in this case, >5.0 percent.) Finally, the size category thresholds of 1,000; 4,100;
33,000; and 96,000 are consistent with the size categories used in the 1989 TCR to determine the
monitoring regimen.
Economic Analysis for the Final RTCR
4-3
September 2012
-------
4.2.1.1 Data Used from SDWIS/FED
To create the system and population baseline, EPA obtained the most current PWS
inventory data by downloading data from the 4th quarter of 2007 from SDWIS/FED (USEPA,
2007b). These data represented all current, active PWSs and the population served by these
systems at the time of the download. This information comprises the PWS inventory baseline for
this EA, and is summarized in Section 4.3.
EPA also used MCL violations data from SDWIS/FED to validate model predictions for
systems serving <4,100 people as described in Section 5.3.3.2, and to predict E. coli (acute)
MCL violations (1989 TCR, RTCR, and Alternative option), TC (non-acute or monthly) MCL
violations (1989 TCR) and Level 1 and Level 2 assessment triggers (RTCR and Alternative
option) for systems serving more than 4,100 people, as described in Section 4.3.4.2.
4.2.1.2 Data Cleaning
Data obtained from SDWIS/FED comprised PWS inventory data for the 4th quarter of
2007 and violations data from 3rd quarter 2005. The 4th quarter inventory data included the
number of each type of PWS and the retail population served by each system for all active
PWSs, and was used from SDWIS/FED without manipulation for the purpose of this analysis.
For the SDWIS/FED violations dataset, EPA excluded some data representing OH, tribes, and
territories as described in Section 4.3.4.2. The data set taken from the 2005 Six-Year Review 2
database was cleaned as described in Section 4.2.2 of this EA.
4.2.1.3 Representativeness and Quality of SDWIS/FED Data
As noted above, SDWIS/FED is the source of PWS inventory data, and includes
information on all of the approximately 155,000 PWSs to the extent that such data was entered
by primacy agencies, EPA Regions, and EPA headquarters personnel. In 1998, EPA began a
major effort to assess the quality of the drinking water data in SDWIS, including performance of
a data quality assessment that has been published triennially since 2002. The most recent report
from this periodic assessment, the 2006 Drinking Water Data Reliability Analysis and Action
Plan (2006 report) (USEPA, 2008a), evaluated data for 2002-2004 and found the PWS inventory
data quality to be high (87 percent) and the 1989 TCR violations data quality to be moderately
high (81 percent). For the PWS inventory data used in this analysis, EPA found specifically that:
".. .The overall data quality of the eight inventory (water system identification)
parameters assessed was 87%. In other words, 87% of PWSs from data verification (DV)
13
states between 2002 and 2004 had consistent data for all eight inventory data elements
between their state files and SDWIS/FED database, or 13% of PWSs had at least one data
element reported with a discrepancy. The highest discrepancy rate was for the
administrative contact address element."
13 EPA routinely evaluates state programs by conducting DV audits, which evaluate the accuracy of a state's
decisions regarding PWS compliance with SDWA regulations and the accuracy with which inventory and
compliance data is reported (entered or uploaded into) SDWIS/FED.
Economic Analysis for the Final RTCR
4-4
September 2012
-------
The report's assessment indicates a high degree of completeness and accuracy in
inventory data submitted by states to SDWIS/FED, and suggests that the information is largely
representative of the PWSs in the United States. Although the RTCR EA uses inventory data
from 2007, the assessment above applies to the most recent period analyzed in this report (2002-
2004). EPA believes that the inventory data from SDWIS/FED in 2007 is also likely to be
representative of the approximately 155,000 PWSs and that there is low uncertainty in the data
with respect to numbers of systems, source water type, size classification, and disinfection
status.14
For the 1989 TCR MCL violations data, the overall data quality estimate was 81 percent,
based on a rating of 84 percent for completeness of data and 97 percent accuracy of the data in
SDWIS/FED as compared to data observed in states' databases during the verification process.15
As described in the report, the 84 percent completeness assessment indicates that for
approximately 16 percent of those systems found by EPA to have 1989 TCR MCL violations,
states did not report the data in SDWIS/FED. Additionally, data completeness for monitoring
and reporting violations as a whole for all Safe Drinking Water Act (SDWA) rules in
SDWIS/FED is low at 29 percent, although this metric specific to the 1989 TCR is not provided
in the report. Low compliance with monitoring and reporting may occur if systems would rather
incur a Monitoring/Reporting violation rather than risk an MCL violation by sampling. These
factors together suggest that violation rates would likely be higher if systems were fully in
compliance with monitoring and reporting requirements, and if states were fully reporting those
results to SDWIS/FED. Despite this potential downward bias in the MCL violations data, the
Total Coliform Rule Distribution System Advisory Committee (TCRDSAC) was still able to
note systematic trends in the data (across time and among PWS types and sizes) and informed its
decisions and recommendations based on key observations:
• Most systems with violations were NCWSs.
• Lower violation rates were observed among larger systems.
• Higher violation rates were observed among GW systems than SW systems.
• Lower violation rates were observed among disinfecting GW systems than other
GW systems.
14 In particular, in the 2006 report the disinfection status (disinfecting, nondisinfecting, or status unknown) inventory
item was assessed at 99.3%, indicating that less than 1% of systems from DV states analyzed between 2002 and
2004 had a discrepancy in this information between State files and SDWIS/FED. For the purpose of the RTCR EA,
systems with status listed as "unknown" were included in the nondisinfecting group.
15 The Overall Data Quality Estimate in SDWIS/FED measures how many noncompliance determinations are
correctly reported in SDWIS/FED among all noncompliance determinations (that are either violations or false-
positive violations). This quantity is estimated based on the violations confirmed by EPA and correctly reported to
SDWIS/FED out of all violations found by EPA or in the state files and SDWIS/FED. When the false positive rate is
0%, this measure is the product of Completeness and Accuracy. The False Positive rate of the violation data in
SDWIS/FED describes how much of the reported violation data in SDWIS are, in fact, false violations, expressed as
a percentage. Further description can be found in the 2006 Drinking Water Data Reliability Analysis and Action
Plan (2006 report) (USEPA, 2008a).
Economic Analysis for the Final RTCR
4-5
September 2012
-------
An additional concern with the representativeness of the violations data in SDWIS/FED
that would not be accounted for in the data quality estimate from the 2006 report is the potential
bias introduced by oversampling. In a recent paper by (Bennear et al., 2009), the authors attempt
to determine how often oversampling is an effort to "sample out" of a potential violation versus
just over compliance—taking additional samples to increase the PWS's diagnostic power is
assessing the quality of the water. The authors suggested that strategic oversampling was
occurring in the one state included in their analysis16 and suggested that some systems were
reducing their probability of incurring monthly violations (on the basis of TC positives as a
percent of total monthly samples taken) by increasing the total number of samples taken.
Although the study was not broad enough in scope to warrant adjusting the data used from
SDWIS/FED, the issue is included in the discussion of net benefits in Chapter 9 of this EA.
4.2.2 Background on 2005 Six-Year Review 2 Data
Through an information collection request (USEPA, 2006b), EPA requested that states
voluntarily submit monitoring data (sample results) that were collected between January 1998
and December 2005 that were available electronically for specified chemical, radiological, and
microbiological contaminants. This request included data collected in compliance with the 1989
TCR regarding the presence/absence of TC, E. coli, and/or fecal coliforms (FC) and any
disinfection data collected at 1989 TCR monitoring sites. (SW systems are required to monitor
for the presence of a disinfectant residual when collecting coliform samples in the distribution
system.)
These data are an important component in supporting EPA's second Six-Year Review of
national primary drinking water regulations (NPDWRs). EPA encouraged each primacy agency
to submit its contaminant occurrence information because these data directly contribute to EPA's
understanding of national contaminant occurrence, populations exposed to regulated
contaminants, and exposure reductions associated with the current regulations.
EPA requested the 1989 TCR monitoring results with the intent of conducting analyses
and developing models to assess the potential impacts of changes to the rule. Models of the
occurrence of TC and E. coli were developed using the 1989 TCR data to assess revisions to
monitoring requirements and compliance decisions for different types of PWSs as suggested by
the stakeholders during the TCRDSAC process, described further in Chapter 3 of this EA.
For the RTCR EA, EPA used the 2005 Six Year Review data to develop the key
parameters used in the RTCR predictive model to forecast water quality (in terms of TC and E.
coli occurrence) under the 1989 TCR, the RTCR, and the Alternative option. Additionally, 2005
Six Year Review data was used to inform EPA's assumptions regarding the proportion of GW
systems serving 1,000 or fewer people that sample monthly, quarterly, or annually (presented in
Exhibit 4.4).
16 Bennear et al. found that violations in the state of Massachusetts may have been approximately 30% greater had
the oversampling not taken place.
Economic Analysis for the Final RTCR
4-6
September 2012
-------
4.2.2.1 2005 Six-Year Review 2 Data Used
As described in the Total Coliform Rule Compliance Monitoring Data Quality and
Completion Report (USEPA, 2010e) (Data Quality Report) prepared during the TCRDSAC
process prior to this EA, the 2005 Six-Year Review 2 data is based on coliform monitoring data
voluntarily provided to EPA by 37 primacy agencies (35 states and 2 tribes). The data consist of
over nine million TCR records collected between 1998 and 2005. The data elements include the
following:
• PWS information—system type, population, source water type.
• Sampling types—routine, repeat, special, unknown.
• Sampling locations—entry point to the distribution system, distribution system, and
for repeat samples, original location, downstream, upstream, and other.
• Methods—for testing for TC and E. coli.
• Monitoring results—presence/absence for TC, FC, and E. coli.
EPA prepared a tool that states using the SDWIS/State database system could use to
easily extract data from their databases and submit it to EPA. The data extraction tool was a
query that collected the data elements requested in the letter to the states. The states that used this
extraction tool are categorized as Tier 1 states. States submitting data in formats not compatible
with the EPA extraction tool are categorized as Tier 2 states. States that submitted incomplete or
problematic data are categorized as Tier 3, and states that did not provide data are categorized as
Tier 4 (e.g., California and Pennsylvania). The complete lists of states included in each of the
Tiers 1-4 are provided in the Data Quality Report.
EPA used 2005 data exclusively in developing this EA. This decision was made for
several reasons:
• The 2005 data, being the most recent in the Six-Year Review, were judged to be
more representative of present conditions than the less recent data, especially in
terms of the percent of TC records that were positive.
• There were fewer data in the years before 2005, especially 2001 and earlier, and
therefore data in earlier years were judged to be too sparse to be comparable with
the 2005 data. The difference in the base of PWSs studied year-to-year might have
introduced an unknown bias.
• As noted earlier, the 2005 data had more records than other years, and enough to
represent the full spectrum of PWSs within states that provided the data.
• Using only a single year of data was beneficial in that it simplified the analysis to
include a fairly static set of PWSs, and did not require interpretation of the meaning
of differences in occurrence data observed across multiple years. Understanding
these types of differences would involve analysis of changing environmental
Economic Analysis for the Final RTCR
4-7
September 2012
-------
factors, program participation and administration factors, and possibly other factors
before the data could be used as a baseline in the model. Additionally, a full year
(and not less) of observed data was required for consistency with the basic unit of
time in occurrence estimates, which is one year; and a period of 12 consecutive
months is assumed to capture seasonal variation normally experienced in water
quality.
• A single year of data would allow for use of a single database to store all the data
for easier analysis.
4.2.2.2 Data Cleaning of 2005 Six-Year Review 2 Data
The Data Quality Report (USEPA, 2010e) describes how TCR monitoring data were
obtained, evaluated, and modified where necessary to make the database internally consistent
and usable for analysis. Exhibit 2.1 in the Data Quality Report provides a complete list of states
or territories that submitted data and a description of the use of these data.
To determine whether a PWS's sampling data were complete in the 2005 Six-Year
Review 2 dataset, a representative month for a system was defined as having at least 50 percent
of the number of samples expected based on the size of system and 1989 TCR requirements.
Using this 50 percent criterion, months that were not representative of a PWS were excluded;
this cleaning resulted in a decrease of only 2 percent of submitted data, and is unlikely to have
skewed occurrence results, as shown by Exhibits 3.1 and 3.2, respectively, of the Data Quality
Report. EPA determined that a 50 percent criterion systematically eliminated systems that were
only reporting positive results or otherwise not reporting many of its samples, while retaining
data for systems that had increased from one size category to the next (and were therefore
expected to have, for example, 15 samples per month instead of 10). The TCRDSAC agreed with
this approach, which is explained further in Section 3.2 of the Data Quality Report.
4.2.2.3 Representativeness of 2005 Six-Year Review 2 Data
In the course of cleaning the 2005 Six-Year Review 2 data, EPA examined data from the
10 largest PWSs in each state and found, by matching with SDWIS inventory information, that
many larger systems were not included in the data submitted by the states. For the largest
systems that are represented in the 2005 Six-Year Review 2 data, EPA also compared the
expected number of samples based on population served (i.e., the 1989 TCR requires a system
serving a million people to sample 300 times per month) to the number of samples in the actual
data to evaluate completeness. In some cases it was found that only data from the state laboratory
were included. Since many large PWSs have their own laboratory and do not use the state
laboratories, data for these large systems were not available. In many cases, states handled large
PWSs with negative samples in a special way, and those data were not captured by the data
extraction tool. The end result was that large PWSs had a lower percentage of samples included
in the data than did smaller systems; therefore, data for systems serving more than 4,100 were
not used from this database. Instead, as described in Section 4.3.4.2, EPA used SDWIS/FED
violations data to model changes in the number of corrective actions to be implemented and the
net change in costs under the RTCR (i.e., incremental costs over the 1989 TCR) for PWSs
serving greater than 4,100 people.
Economic Analysis for the Final RTCR
4-8
September 2012
-------
Based on the number of smaller PWSs (serving fewer than 4,101 people) in states whose
data were excluded for data quality/quantity issues or for non-submittal of data, approximately
39 percent of systems were excluded from the analysis. Specifically, the following states or
territories submitted data but were excluded because of issues with the quality or quantity of data
submitted: AL, American Somoa, DC, ME, MD, NJ, SC, SD, TN, and the Navajo Nation.
Additionally, the following states/territories did not submit data: CA, Guam, HI, LA, MA, MS,
Northern Mariana Islands, PA, Puerto Rico, TX, Virgin Islands, and WA. The remaining 70
percent of PWSs' data were used to simulate occurrence of TC and E. coli under the 1989 TCR,
the RTCR and the Alternative option and to develop parameters for use in the RTCR predictive
model, described in Chapter 5 of this EA.
Although approximately 61 percent of the systems are represented in the analysis, only an
estimated 32 percent of the population served by PWSs is represented due to the omission of
some large population states (i.e., CA, PA, and TX).17 However, since EPA is not predicting
changes in occurrence for the largest systems (i.e., the majority of the population is served by
very large CWSs) the underrepresentation of population served in the data is not expected to
have a significant impact on the analyses performed.
4.2.3 Background on Other Data and Information Used
SDWIS/FED data and 2005 Six-Year Review 2 data represent the largest portion of
information used for the RTCR baseline analysis. In addition, EPA incorporated information
from the GWR, the U.S. Census, and the 1989 TCR to develop assumptions for the model. EPA
also relied upon the knowledge and experience of stakeholders representing industry, states,
small systems, and the public to inform the process throughout development of the RTCR.
4.2.3.1 1989 TCR
Under the 1989 TCR, PWSs experienced reduced occurrence of fecal indicator
contamination after the effective date as a result of increased sampling, reporting, and to some
degree corrective actions. In recent years under the 1989 TCR, the occurrence of fecal indicators
has reached a steady state, as reflected in SDWIS/FED violation data (see Section 4.3.4.2 for
discussion). Additional reductions in occurrence are expected prior to RTCR implementation that
reflect water quality improvements due to implementation of the GWR. Because the water
quality data used in this analysis is the Six-Year Review 2 data for 2005, adjustments are made
to the 2005 data in the modeling to reflect this GWR impact. Further reductions in occurrence of
fecal indicators under the RTCR can then be properly modeled based on more focused
assessments and corrective actions in response to TC+ oris, co/z'-positive (EC+) samples.
4.2.3.2 GWR data
The GWR was promulgated to address microbial contamination of ground water sources
used to supply drinking water and, specifically, to address the concern about potential adverse
health risks that may be associated with fecal contamination. Before the GWR was promulgated,
17 As described further in the Data Quality Report, submission of the data was voluntary; CA, PA, and TX did not
submit data.
Economic Analysis for the Final RTCR
4-9
September 2012
-------
there were no federal regulations to require filtration or disinfection of ground water sources to
remove microbial contaminants. The GWR requirements include sanitary surveys, triggered
monitoring in response to TCR samples testing positive for contaminant indicators, source water
18
monitoring for indicators of fecal contamination such as E. coli, and corrective actions that may
include disinfection.
Both the GWR and the RTCR seek to decrease the level of fecal contamination in
drinking water. Since the RTCR applies to all PWSs including systems that use ground water
sources, the implementation of the GWR (requiring compliance beginning December 1, 2009) in
advance of the RTCR is expected to modify the TCR baseline (see Sections 4.3.1-4.3.2) taken
from the 2007 SDWIS/FED (4th quarter) and 2005 Six Year Review data (see Sections 4.2.1 and
4.2.2 for discussion of SDWIS/FED and 2005 Six Year Review data, respectively). The number
of disinfecting and nondisinfecting public ground water systems prior to the start of GWR
implementation (based on 4th quarter 2007 SDWIS/FED data) is 54,469 and 85,999, respectively.
By October 2012 when the GWR is fully implemented, EPA predicts that the number of systems
disinfecting will have increased from the 2007 4th quarter baseline. The RTCR occurrence model
incorporates the estimated change in the number of systems with disinfection into this EA
baseline (Section 5.3.1).
The RTCR occurrence model (described in further detail in Section 5.3 of this EA)
accounts for GWR implementation in the following ways:
• The baseline number of disinfecting systems in the 4th quarter 2007 data of
SDWIS/FED will be increased to account for the additional systems that are
estimated to begin disinfecting under the GWR.
• Sanitary surveys, as required by the GWR, will be phased in for applicable PWSs in
the model (over five years for systems on a five year sanitary survey cycle; and over
three years for systems on a three year cycle), resulting in reduced occurrence for
these systems indefinitely to represent that systems are correcting deficiencies
identified in the survey. Further detail on this component of the model is provided
in Chapter 5 of this EA.
• GWR compliance monitoring will be incorporated with an additional decrease in
probability of occurrence to represent anticipated efforts by systems to identify and
correct problems.
• The baseline estimates will account for the decreased levels of TC occurrence and
fecal contamination achieved under the GWR. This decrease will be reflected in
model output for the following RTCR rule components:
- Repeat monitoring
- Additional routine monitoring
- Triggers for level 1 and 2 assessments
18 Standridge, J. (2008) discusses the use of E. coli as an indicator of drinking water quality.
Economic Analysis for the Final RTCR
4-10
September 2012
-------
- Corrective actions
- Public notification
• Under the GWR, if a PWS experiences a TC+ while conducting TCR monitoring,
the system must sample the source water for the presence of a fecal indicator (E.
coli, enterococci, or coliphage, as determined by the state). GW systems that have a
positive initial fecal indicator sample from the source are required to take five
repeat fecal indicator samples if they are not required by the state to take corrective
action. Under the RTCR, PWSs having less than 4-log treatment and serving 1,000
people or fewer would only be required to take 3 repeat samples. To meet GWR
requirements, these systems would need to take one additional sample at the source.
• Under the RTCR, PWSs having less than 4-log treatment and serving 1,000 people
or fewer would be required to take 3 repeat samples after a TC+ routine sample as
opposed to the 4 repeat samples required under the 1989 TCR. To meet GWR
requirements, these systems would need to take one additional sample (i.e., a fourth
repeat sample) at the source. Under the RTCR, the PWS may, with written state
approval, take one of its repeat RTCR samples at the location required for triggered
source water monitoring under the GWR if the PW S demonstrates that the repeat
sample sites in the sample siting plan remain representative of water quality in the
distribution system.
4.2.3.3 U.S. Census Data
The baseline analysis incorporates data from the U.S. Census that are publically available
and include surveys of the U.S. population, economics, industry, and geography. Census 2000
data were used in the baseline analyses to estimate the sensitive subpopulations in the U.S. who
may be more susceptible to illness as a result of poor drinking water quality, as explained in
Section 4.4 of this EA. Additionally, the Census 2008 Annual Social and Economic Supplement
was used in the estimates of household size as described in Section 4.3.3.
4.3 Baseline Profile
The RTCR applies to all PWSs, regardless of their size, water source (GW or SW), or
type (CWS, NTNCWS, or TNCWS). This section estimates the baseline number of PWSs and
the size of the population subject to the RTCR in each of these subcategories. EPA used 4th
quarter 2007 data from SDWIS to develop this inventory baseline because at the time when the
TCRDSAC began discussing the RTCR and its analysis, 2007 data were the most recent
inventory data available. As described in Section 4.2, the best available occurrence data for
smaller PWSs were collected in 2005. For larger systems (serving >4,100 people) as Section
4.3.4.2 describes, EPA used SDWIS/FED violations data for modeling benefits and costs for
PWSs serving more than 4,100 people because the 2005 Six-Year Review 2 dataset (containing
sampling data) for larger systems was not as robust. Third quarter 2005 SDWIS/FED data is used
for the violation baseline since EPA compiles violation data in the 3rd quarter every year (as
opposed to the 4th quarter freeze used for annual compilation inventory data) and was the most
recently available data at the time that this analysis was begun. Use of the 2005 data was deemed
Economic Analysis for the Final RTCR
4-11
September 2012
-------
to be equivalent to an average of multiple years due to the steady numbers of violations seen in
the data over the past several years.
The baseline described in the following sections is used as a reference point for informing
net costs and benefits as described in Chapters 6 and 7 of this EA.
4.3.1 Pre-GWR Baseline
Estimates of the number of GW systems subject to the RTCR are presented in Exhibit 4.1
below. These numbers reflect SDWIS/FED inventory from the fourth quarter of 2007, prior to
implementation of the GWR. Ground water system inventories indicate disinfection status as
disinfecting or unreported; since reporting is voluntary, the status of a system that did not report
disinfection is actually unknown, although EPA assumes in this EA that the system is not
disinfecting. The number of disinfecting PWSs includes those achieving less than 4-log
disinfection, which matches the categorization of systems used to estimate the underlying
distributions of occurrence of TC and E. coli modeled, as described in Chapter 5.
Exhibit 4.2 presents the number of SW systems subject to the RTCR, also derived from
the SDWIS/FED database. GWR implementation does not affect the disinfection status of SW
systems.
Exhibit 4.3 presents the population associated with each PWS category of system size
and type prior to incorporating the effects of the GWR into the model; this is the pre-GWR
baseline affected population. After incorporating GWR effects, some systems shift into the
disinfecting category and some of the population shifts accordingly.
Lastly, Exhibit 4.4 presents the distribution of PWSs by frequency of sampling—
monthly, quarterly, or annually—under the 1989 TCR for those systems that qualify for reduced
monitoring (all other systems are required to sample monthly). This is the baseline distribution of
sampling frequencies used for analyses in this EA. This distribution was derived as part of the
work done by the Technical Workgroup (TWG) supporting the TCRDSAC, and validated by the
Association of State Drinking Water Administrators (ASDWA). Workgroup members analyzed
2005 Six Year Review and other readily available state data to derive the estimates shown in
Exhibit 4.4.
Economic Analysis for the Final RTCR
4-12
September 2012
-------
Exhibit 4.1 Pre-GWR Baseline Number of GW Systems
Number of GW PWSs (Pre-GWR)
PWS Size
cws
NTNCWS
TNCWS
Disinfecting
Non-Disinfecting
Disinfecting
Non-Disinfecting
Disinfecting
Non-Disinfecting
<100
6,132
5,806
2,907
5,919
13,558
46,642
101-500
9,260
4,632
2,753
3,860
5,341
13,934
501-1,000
3,502
965
868
850
673
1,290
1,001-4,100
5,405
1,038
542
270
269
348
4,101-33,000
2,798
358
56
14
27
40
33,001-96,000
307
28
2
-
-
2
96,001-500,000
62
1
-
-
-
1
500,001-1 Million
4
-
-
-
-
1
> 1 Million
3
-
-
-
-
-
Total
27,473
12,828
7,128
10,913
19,868
62,258
Combined Total
40,301
18,041
82,126
Source: Data extracted from SDWIS/FED PWS Inventory, 2007 4th Quarter Data based on listed disinfection status.
PWSs listed as "unknown" disinfection status in SDWIS/FED are counted as non-disinfecting.
Exhibit 4.2 Baseline Number of SW Systems
PWS Size
Number of SW PWSs
cws
NTNCWS
TNCWS
<100
1,170
250
1,339
101-500
2,150
253
497
501-1,000
1,173
88
88
1,001-4,100
2,938
72
67
4,101-33,000
3,164
22
18
33,001-96,000
720
2
-
96,001-500,000
308
1
-
500,001-1 Million
31
-
-
> 1 Million
17
-
1
Total
11,671
688
2,010
Source: Data extracted from SDWIS/FED PWS Inventory, 2007 4th Quarter
Data. SWcount includes GWUDI systems.
Economic Analysis for the Final RTCR
4-13
September 2012
-------
Exhibit 4.3 Pre-GWR Baseline Population Served by GW Systems
PWS Size
Population Served by GW PWSs (Pre-GWR)
CWS
NTNCWS
TNCWS
Disinfecting
Non-Disinfecting
Disinfecting
Non-Disinfecting
Disinfecting
Non-Disinfecting
<100
387,558
342,243
168,482
326,342
645,949
2,139,529
101-500
2,375,507
1,085,555
708,424
948,500
1,231,077
3,005,703
501-1,000
2,570,662
703,562
630,071
600,643
516,707
956,781
1,001-4,100
11,307,740
2,074,139
991,971
456,985
449,757
599,171
4,101-33,000
29,346,057
3,627,365
418,368
122,865
194,136
353,290
33,001-96,000
15,587,186
1,527,861
89,405
-
-
119,700
96,001-500,000
9,935,500
107,323
-
-
-
100,000
500,001-1 Million
2,670,841
-
-
-
-
725,000
> 1 Million
4,389,948
-
-
-
-
-
Total
78,570,999
9,468,048
3,006,721
2,455,335
3,037,626
7,999,174
Combined Total
88,039,047
5,462,056
11,036,800
Source: Data extracted from SDWIS/FED PWS Inventory, 2007 4th Quarter Data based on listed disinfection status.
Populations for PWSs listed as "unknown" disinfection status in SDWIS/FED are counted as non-disinfecting.
Economic Analysis for the Final RTCR
4-14
September 2012
-------
Exhibit 4.4 Percent Distribution of GW System Monitoring Frequencies by PWS
Size and Type for 1989 TCR
Size
Number of
Systems
Monthly
Quarterly
Annual
Community Water Systems (CWSs), Disinfecting
<100
6,132
86.6%
13.4%
0.0%
101-1000
12,762
88.5%
11.5%
0.0%
1001-4100
5,405
100.0%
0.0%
0.0%
Community Water Systems (CWSs), Non-Disinfecting
<100
5,806
86.6%
13.4%
0.0%
101-1000
5,597
88.6%
11.4%
0.0%
1001-4100
1,038
100.0%
0.0%
0.0%
Nontransient, Noncommunity Water Systems (NTNCWSs), Disinfecting
<100
2,904
19.3%
64.7%
16.0%
101-1000
3,621
18.5%
66.7%
14.7%
1001-4100
542
100.0%
0.0%
0.0%
Nontransient, Noncommunity Water Systems (NTNCWSs), Non-Disinfecting
<100
5,913
19.3%
64.7%
16.0%
101-1000
4,710
18.5%
66.7%
14.8%
1001-4100
270
100.0%
0.0%
0.0%
Transient, Noncommunity Water Systems (TNCWSs), Disinfecting
<100
13,558
4.8%
62.9%
32.3%
101-1000
6,014
7.9%
66.9%
25.2%
1001-4100
269
100.0%
0.0%
0.0%
Transient, Noncommunity Water Systems (TNCWSs), Non-Disinfecting
<100
46,642
4.8%
62.9%
32.3%
101-1000
15,224
7.8%
66.8%
25.3%
1001-4100
348
100.0%
0.0%
0.0%
Source: Based on EPAand TWG analysis of 2005 Six-Year data and individual state
statutes during the TCRDSAC and review byASDWA
Note: All other system sizes and types are required to sample monthly.
4.3.2 Post-GWR Baseline
PWSs must comply with requirements of the GWR beginning December 1, 2009, which
will be approximately five years prior to the effective date of the RTCR. To account for the
impact of the GWR on baseline occurrence for those PWSs using a GW source, EPA performed
a number of adjustments to the 2005 data prior to its use in this EA.
Using 2007 SDWIS inventory data (as explained in Section 4.3.1) as a pre-GWR
inventory baseline, EPA applied probabilities of Psampie and Pweii19 to determine the probability of
a positive sample for a fecal indicator in nondisinfecting PWSs. A fraction of those systems with
19 The term "Pwell" refers to the probability that a randomly selected well across the United States will ever test
positive in its source water for a virus or indicator species, such as TC; "Psample" is the probability that given a
contaminated well, a random sample at the well will test positive for a virus or indicator. The GWR EA (November
2006) (USEPA, 2006a), and in particular the Baseline Chapter of that document, provides a detailed explanation of
the analysis used to generate these PweU and Psampie estimates.
Economic Analysis for the Final RTCR
4-15
September 2012
-------
a positive sample is expected to move to a disinfecting status in the five years prior to RTCR
implementation. This modification ensured that systems in the analysis that disinfect drew from
the probability distribution for TC+ and EC+ for disinfecting systems, which is different from
the distribution for those that do not disinfect. The model continues to apply Pweii and Psampie
throughout the RTCR period of analysis (the 25 years following promulgation). The application
of this model parameter, as it is applied in the context of other aspects of the predictive
occurrence model, is explained in further detail in Chapter 5 of this EA. Exhibit 4.5 shows the
baseline inventory of GW systems by disinfection status in model year 5, which is the year
immediately prior to the effective date of the RTCR. The estimates in Exhibit 4.5 of
nondisinfecting PWSs include a number of systems that selected a corrective action other than
converting to a disinfecting system in response to fecal contamination at the source.
As explained in Chapter 5 of this EA, a number of PWSs will undertake sanitary surveys
(for all GW systems) and compliance monitoring (for disinfecting GW systems) prior to
implementation of the RTCR; although the costs are not considered in this EA (they were
considered in the GWR EA), the systems are assigned a reduced occurrence, which is reflected
in the RTCR baseline. The predictive model incorporates the contribution of sanitary survey
results to the qualification of systems for reduced monitoring by assuming a 10 percent reduction
in the number of systems finding a TC+, which is applied to a portion of systems each year and
retained throughout the period of analysis. For NCWSs, the reduction is applied to one-fifth of
systems per year for the five years of GWR implementation prior to the RTCR effective date; for
CWSs, it is applied to one-third of systems in each of the first three years of the GWR
implementation period. Similarly, the system incorporates the effect of compliance monitoring
by disinfecting GW systems by applying an additional 10 percent reduction for systems that are
disinfecting as of the RTCR effective date, and additionally to those determined in the model to
select disinfection as a corrective action; the compliance monitoring reduction in occurrence is
assumed to apply in all remaining years of the analysis once a system qualifies for the reduction.
As described in Section 5.3.2.2 of this EA, model output was generated for a 30-year
period to encompass 5 years of the effects of GWR implementation and 25 years of the effects of
RTCR implementation. Chapter 5 presents this output for the 30-year period. For the purpose of
the benefit-cost analysis, years 3 through 27 of this period are included in the calculations and
presented in the results shown in Chapter 6 (Benefits Analysis), Chapter 7 (Cost Analysis), and
Chapter 9 (Comparison of Benefits and Costs). Initially in the modeled period, all systems under
the 1989 TCR and the RTCR are assumed to sample based on the sampling regimen applicable
to their system type, size, type of water source, and whether or not the system disinfects (Exhibit
20
4.4). Under the Alternative option, all systems are assumed to sample monthly for the first five
years after the RTCR effective date. Following an assessment period as described in Section
5.3.2.2, monitoring frequencies are adjusted from the baseline for GW systems serving fewer
than 1,000 people. All other systems remain on monthly monitoring.
The estimates for distribution of PWSs across the M/Q/A monitoring schedules as
adjusted from the baseline distribution are presented in Chapter 5 for the 1989 TCR, RTCR, and
Alternative option (Exhibits 5.9a-5.9c).
20 Based on EPA and TWG analysis of Six-Year data and individual state statutes during the TCRDSAC and review
by ASDWA.
Economic Analysis for the Final RTCR
4-16
September 2012
-------
Exhibit 4.5 Baseline Number of GW Systems Post-Implementation of the GWR
(Model Year 5)
PWS Size
Number of GW PWSs (Post-GWR)
CWS
NTNCWS
TNCWS
Disinfecting
Non-Disinfecting
Disinfecting
Non-Disinfecting
Disinfecting
Non-Disinfecting
<100
6,190
5,748
2,938
5,888
13,753
46,447
101 -500
9,311
4,581
2,776
3,837
5,451
13,824
501-1,000
3,512
955
873
845
684
1,279
1,001-4,100
5,422
1,021
547
265
274
343
4,101-33,000
2,798
358
56
14
27
40
33,001-96,000
307
28
2
-
-
2
96,001-500,000
62
1
-
-
-
1
500,001-1 Million
4
-
-
-
-
1
> 1 Million
3
-
-
-
-
-
Total
27,610
12,691
7,191
10,850
20,189
61,937
Combined Total
40,301
18,041
82,126
Source: Estimates calculated based on the proportion of PWSs changing disinfection status, which is an output from the predictive occurrence
model, as described in Chapter 5 of this EA.
4.3.3 Baseline Population Served
PWS population characteristics are important to this analysis for determining the number
of people, both prior to and following implementation of the GWR, for which risk changes under
each component of the RTCR, as discussed in Chapter 6 of this EA. These population estimates
are based on SDWIS/FED 2007 4th quarter data as described in Section 4.2.1, and stratified by
PWS size, type, whether the system disinfects or not and whether the system uses GW or SW
sources. The estimates represent the number of systems which disinfect water and their
populations served; both estimates are adjusted upward based on predicted impacts of the GWR
in the five years prior to the effective date of RTCR requirements, as shown in Exhibits 4.6 and
4.7 below. (See Exhibits 4.1 and 4.3 for PWS counts and population served by disinfection status
prior to effects of GWR.)
Economic Analysis for the Final RTCR
4-17
September 2012
-------
Exhibit 4.6 Baseline Population Served by GW Systems Post-Implementation of
the GWR (Model Year 5)
PWS Size
Population Served by Post-GWR PWSs
cws
NTNCWS
TNCWS
Disinfecting
Non-Disinfecting
Disinfecting
Non-Disinfecting
Disinfecting
Non-Disinfecting
<100
390,986
338,815
170,177
324,647
654,897
2,130,581
101 - 500
2,387,462
1,073,600
713,960
942,964
1,254,805
2,981,975
501-1,000
2,578,316
695,908
633,744
596,970
524,599
948,889
1,001-4,100
11,341,403
2,040,476
999,913
449,043
458,242
590,686
4,101-33,000
29,346,057
3,627,365
418,368
122,865
194,136
353,290
33,001-96,000
15,587,186
1,527,861
89,405
-
-
119,700
96,001-500,000
9,935,500
107,323
-
-
-
100,000
500,001-1 Million
2,670,841
-
-
-
-
725,000
> 1 Million
4,389,948
-
-
-
-
-
Total
78,627,698
9,411,349
3,025,567
2,436,489
3,086,679
7,950,121
Combined Total
88,039,047
5,462,056
11,036,800
Note:
Estimates calculated based on the proportion of PWSs changing disinfection status, which is an output from the predictive
occurrence model, as described in Chapter 5 ofthis EA
Exhibit 4.7 Baseline Population Served by SW Systems
PWS Size
Population Served by SW PWSs
cws
NTNCWS
TNCWS
<100
56,740
13,297
57,454
101-500
608,084
68,083
116,529
501-1,000
885,400
67,958
70,147
1,001-4,100
6,628,597
140,168
143,347
4,101-33,000
38,700,554
174,408
147,423
33,001-96,000
39,034,554
121,446
-
96,001-500,000
58,489,936
203,000
-
500,001-1 Million
22,327,506
-
-
> 1 Million
37,363,275
-
2,000,000
Total
204,094,646
788,360
2,534,900
Source: Data extracted from SDWIS/FED PWS Inventory, 2007 4th Quarter Data.
Population figures for SWsystems include GWUDI systems.
Number of Households Served
Because PWS costs are often passed onto customers in the form of water rate increases,
the RTCR EA also includes analyses to assess the impact of the rule provisions at a household
level. The number of households served by CWSs expected to be subject to the RTCR is
estimated by dividing the population for each PWS size category (Exhibits 4.6 and 4.7 above) by
the average number of people per household, which was estimated as 2.56 for the year 2007
Economic Analysis for the Final RTCR
4-18
September 2012
-------
(U.S. Census Bureau, 2008). Exhibit 4.8 below shows the number of households served by
ground water and surface water systems, respectively.
Exhibit 4.8 Household Baseline
PWS Size
Number of Households
GW CWSs
SW CWSs
Disinfecting
No n-Disinfecting
Disinfecting
<100
151,390
133,689
22,164
101-500
927,932
424,045
237,533
501-1,000
1,004,165
274,829
345,859
1,001-4,100
4,417,086
810,211
2,589,296
4,101-33,000
11,463,304
1,416,939
15,117,404
33,001-96,000
6,088,745
596,821
15,247,873
96,001-500,000
3,881,055
41,923
22,847,631
500,001-1 Million
1,043,297
-
8,721,682
> 1 Million
1,714,823
-
14,595,029
Total
30,691,796
3,698,456
79,724,471
Sources: U.S. Census Bureau, Current Population Survey, 2008 Annual Social and
Economic Supplement; SDWIS/FED PWS Inventory, 2007 4th Quarter Data.
4.3.4 Baseline Water Quality
The following sections provide an overview of summary statistics relating to baseline
water quality. The source data from which these summary statistics are derived form the basis of
further analysis in the RTCR occurrence and risk assessment models as described in later
chapters of the EA.
4.3.4.1 Percent of TC and EC-Positive Samples Based on 2005 Six-Year Review 2 Data
Exhibit 4.9 below shows the percent of TC+ and EC+, samples based on PWS type and
size. As described in Section 4.2.2.2, the 2005 Six-Year Review 2 data was cleaned using a
criterion that a given system-month of data should include a minimum of 50 percent of the
expected number of samples based on the system's population served and water system category.
The "TC+ samples" column was calculated by taking the total number of routine TC+
samples and dividing by the number of routine TC samples. For small PWSs, additional routine
TC samples in the month following a TC+ were included in the denominator. To calculate the
EC+ rate, the total number of EC+ samples was divided by the total number of TC- samples21
plus the number of TC+ samples that were tested for E. coli. This EC+ computation did not
include additional routine samples taken in the month following a TC+ because a significant
21 EPA assumes that a PWS will not test for EC if the TC assay is negative.
Economic Analysis for the Final RTCR
4-19
September 2012
-------
number of PWSs did not provide data on EC+ oris, co/z'-negative samples, and some systems
tested for FC and not E. coli.
Economic Analysis for the Final RTCR
4-20
September 2012
-------
Exhibit 4.9 Total Coliform and E. coli Percent Positive by System Size and Type
PWS
Type
Source
Water
Population
Served
TC
(# Samples)
TC
(+ Samples)
TC
(% Positive)
EC
(# Samples)1
EC
(+ Samples)
EC
(% Positive)2
<100
93,105
2,479
2.66%
1,172
72
0.08%
101 -500
125,490
2,500
1.99%
1,639
61
0.05%
501-1,000
48,265
736
1.52%
483
20
0.04%
GW
1,001-4,100
110,391
1,176
1.07%
732
21
0.02%
4,101-33,000
183,721
877
0.48%
458
22
0.01%
33,001-100,000
96,361
214
0.22%
44
2
0.00%
>100,000
64,965
289
0.44%
34
1
0.00%
Total GW
722,298
8,271
1.15%
4,562
199
0.03%
cws
<100
6,735
95
1.41%
64
6
0.09%
101 -500
19,716
227
1.15%
159
10
0.05%
501-1,000
12,828
90
0.70%
70
7
0.05%
SW
1,001-4,100
55,310
314
0.57%
233
17
0.03%
4,101-33,000
175,758
525
0.30%
399
41
0.02%
33,001-100,000
112,894
157
0.14%
106
5
0.00%
>100,000
112,143
235
0.21%
99
2
0.00%
Total SW
495,384
1,643
0.33%
1,130
88
0.02%
GW & SW
Total CWS
1,217,682
9,914
0.81%
5,692
287
0.02%
<100
163,730
7,820
4.78%
5,820
316
0.20%
101 -500
52,891
2,418
4.57%
1,869
99
0.19%
GW
501-1,000
6,952
299
4.30%
217
4
0.06%
>1,000
7,062
143
2.02%
85
2
0.03%
Total GW
230,635
10,680
4.63%
7,991
421
0.18%
TNCWS
<100
6,723
150
2.23%
141
17
0.25%
101 -500
2,854
75
2.63%
69
13
0.46%
SW
501-1,000
523
19
3.63%
19
-
0.00%
>1,000
988
6
0.61%
37
-
0.00%
Total SW
11,088
250
2.25%
266
30
0.27%
GW & SW
Total TNCWS
241,723
10,930
4.52%
8,257
451
0.19%
<100
46,505
1,476
3.17%
1,061
34
0.07%
101 -500
33,084
893
2.70%
628
19
0.06%
GW
501-1,000
9,531
166
1.74%
103
2
0.02%
>1,000
13,138
177
1.35%
103
5
0.04%
Total GW
102,258
2,712
2.65%
1,895
60
0.06%
NTNCWS
<100
1,668
32
1.92%
30
4
0.24%
101 -500
2,304
9
0.39%
9
2
0.09%
SW
501-1,000
932
6
0.64%
5
-
0.00%
>1,000
1,316
1
0.08%
1
-
0.00%
Total SW
6,220
48
0.77%
45
6
0.10%
GW & SW
Total NTNCWS
108,478
2,760
2.54%
1,940
66
0.06%
Source: Derived using 2005 Six-Year Review 2 Data, which was filtered by including a month of data for a given system (system-month)
only if it represented at least 50% of the expected number of samples for a month based on the system's size. The Total Coliform Rule
Compliance Monitoring Data Quality and Completion Report (USEPA, 201 Oe) includes a detailed description of this data cleaning
process.
Notes:
1) Number of samples that were specifically tested for E. coli. The denominator of the E. coli percent positive calculation includes this
number plus the number of total coliform negative samples (TC #samples - TC +samples).
2) Percent EC+ was calculated as (# EC+ samples)/(# EC samples).
Economic Analysis for the Final RTCR
4-21
September 2012
-------
4.3.4.2 Violation Levels Based on SDWIS/FED Data
Non-acute violations are triggered under the 1989 TCR by water quality violations. They
are defined as >1 routine or repeat monitoring sample testing positive for TC in a given
compliance period (month, quarter, or year) for PWSs serving fewer than 33,001 people, or by
>5.0 percent of samples being positive for systems taking >40 samples per month (typically
those systems serving greater than 33,000 people). Acute violations are triggered when any PWS
has a repeat sample with an EC+ or FC-positive (FC+) or has an FC+ or EC+ routine sample
followed by a TC+ repeat sample in a given month.
Exhibit 4.10 below presents an assessment of water quality prior to the RTCR
promulgation in terms of the number of acute and non-acute violations incurred by PWSs. The
number of violations from this data is directly input into the cost model for PWSs serving more
than 4,100 people to derive estimates of the number of assessments and corrective actions that
will be undertaken under the RTCR. As noted in Exhibit 4.10, the data used is from 2005 3rd
22
quarter. Exhibit 4.11 presents the seven years of PWS violation data evaluated by the TWG.
Appendix G presents further detail on this assessment, including queries used in SDWIS/FED
and the resulting datasets downloaded.
In addition to the acute and non-acute system violation data, Exhibit 4.11 also presents
data on the numbers of PWSs with monitoring and reporting violations (minor and major for
routine and repeat monitoring and reporting). Although PWS monitoring and reporting violation
data is not used directly in the quantitative analyses performed for the EA, the TCRDSAC did
consider the data in its deliberations on the impact of the regulation. In particular, the TCRDSAC
determined that the revisions to the RTCR may significantly reduce the high numbers of
monitoring and reporting violations by: a) reducing the numbers of additional routine and repeat
samples required; b) providing more flexibility in development of sampling site plans; and c)
adding consequences for missed monitoring (possible triggering to increased monitoring for
missed samples).
The data verification process was used to review individual states' data in comparison to
SDWIS/FED data and to compare violations rates across states. This process revealed many
differences between states in their implementation of the 1989 TCR, and allowed EPA to
identify and review outliers in the national database. Specifically, EPA did not include violation
data from Ohio, U.S. territories, or tribal PWSs in the summaries presented in Exhibits 4.10 and
4.11. Review of the data verification information revealed that Ohio's broad interpretation of
what constitutes a violation led to abnormally high violation counts compared (on average) to
other states. For U.S. territories and tribal systems, it was established that unique environmental
factors and operating conditions contribute to abnormally high TC+ results and associated
violations. Thus, inclusion of these results would skew the national averages used for analysis,
and were therefore excluded.
22 Exhibit 4.11 presents PWS counts (PWSs with at least one violation during the year) to be consistent with the
metric evaluated by the TCRDSAC TWG. Exhibit 4.10 presents total violation counts, which are used as inputs to
the predictive modeling for the EA.
Economic Analysis for the Final RTCR
4-22
September 2012
-------
Exhibit 4.10 Baseline Number of 1989 TCR Violations by System Size and Type
(2005)
Number of Violations in
Number of Violations in
Total Number of
Total Number of
GW PWSs
SW PWSs
GW PWSs
SW PWSs
Non -Acute
Acute
Non-Acute
Acute
(without OH & PR)
(without OH & PR)
A
B
C
D
E
F
CWSs
<100
905
52
16
3
11,709
1,133
101-500
809
34
50
7
13,508
2,067
501-1,000
203
13
16
3
4,299
1,112
1,001-3,300
272
8
55
7
5,490
2,406
3,301-10,000
171
8
75
3
2,628
2,015
10,001-50,000
125
8
78
4
1,247
1,752
50,001-100,000
11
2
5
4
147
362
100,001-250,000
1
1
3
1
250,001-500,000
-
-
-
-
61
306
500,001-1 Million
-
-
1
-
> 1 Million
-
-
-
-
3
15
Total CWSs
2,497
126
299
32
39,092
11,168
NTNCWSs
<100
514
34
7
2
8,392
249
101-500
346
20
4
-
6,294
246
501-1,000
57
6
2
-
1,622
87
1,001-3,300
58
4
-
-
726
66
3,301-10,000
9
2
1
-
102
23
10,001-50,000
1
-
-
-
11
4
50,001-100,000
-
-
-
-
-
1
100,001-250,000
-
-
-
-
-
250,001-500,000
-
-
-
-
-
1
500,001-1 Million
-
-
-
-
-
> 1 Million
-
-
-
-
-
-
Total NTNCWSs
985
66
14
2
17,147
677
TNCWSs
<100
2,665
278
19
5
58,396
1,334
101-500
833
76
11
1
18,184
492
501-1,000
133
11
4
-
1,868
85
1,001-3,300
58
2
1
-
578
57
3,301-10,000
5
-
1
-
77
22
10,001-50,000
-
-
-
-
11
3
50,001-100,000
-
-
-
-
3
-
100,001-250,000
-
-
-
-
250,001-500,000
-
-
-
-
500,001-1 Million
-
-
-
-
1
-
> 1 Million
-
-
-
-
-
1
Total TNCWSs
3,694
367
36
6
79,118
1,994
Sources:
(A) - (D) Acute/Non-Acute Violations from SDWIS/FED annual data for period ending 3rd quarter 2001 - 2007 (only 2005 data is
presented in this exhibit). OH, U.S. territories, tribal PWS data excluded.
(E), (F) 2007 4th Quarter SDWIS Data, Total Number of PWSs.
Economic Analysis for the Final RTCR
4-23
September 2012
-------
Exhibit 4.11 Number of PWSs with Violations by System Type (2001-2007)
PWS Type
Year
2001
2002
2003
2004
2005
2006
2007
Acute MCL Violations
cws
143
144
185
171
151
171
171
NTNCWS
51
53
70
58
65
68
45
TNCWS
261
278
322
351
349
361
295
All
455
475
577
580
565
600
511
Non^Acute MCL Violations
CWS
2,074
2,110
2,204
2,314
2,196
2,095
1,996
NTNCWS
601
679
725
750
753
735
655
TNCWS
2,707
2,934
3,036
3,132
3,039
3,244
3,209
All
5,382
5,723
5,965
6,196
5,988
6,074
5,860
Major Monitoring and Reporting Violations (Routine and Repeat)
CWS
3,312
3,327
3,900
3,924
3,760
3,659
3,415
NTNCWS
1,503
1,561
1,676
1,679
1,588
1,468
1,377
TNCWS
10,360
10,531
11,230
11,043
10,426
10,630
10,389
All
15,175
15,419
16,806
16,646
15,774
15,757
15,181
Minor Monitoring and Reporting Violations (Routine and Repeat)
CWS
1,218
1,257
1,389
1,445
1,302
1,285
1,244
NTNCWS
187
182
200
233
194
214
183
TNCWS
701
842
883
903
875
843
839
All
2,106
2,281
2,472
2,581
2,371
2,342
2,266
Note: PWSs counts are of systems that had at least one violation during the year.
Source: SDWIS/FED annual data for period ending 3rd quarter 2001 - 2007. OH, U.S. territories, tribal PWS
data excluded.
4.4 Sensitive Sub-Populations
Under the 1996 Amendments to the Safe Drinking Water Act (SDWA) (PL 104-182),
EPA must analyze health impacts of rulemaking on sensitive subpopulations. Sensitive
populations include "infants, children, pregnant women, the elderly, individuals with a history of
serious illness, and other subpopulations that are identified as likely to be at a greater risk of
adverse health effects due to exposure to contaminants in drinking water than the general
population." (1996 Amendments to the Safe Drinking Water Act).
Exhibit 4.12 presents EPA's estimates of the number of U.S. individuals who are at
increased risk of developing more severe symptoms from illnesses caused by waterborne
pathogens. Persons suffering from certain diseases and/or conditions are believed to be sensitive
to microbes and chemicals in drinking water. These subgroups of the population include
pregnant women, the very young, the elderly, and the immunocompromised. In total, these
subgroups represent approximately 20 percent of the current population of the U.S.
Economic Analysis for the Final RTCR
4-24
September 2012
-------
Exhibit 4.12 Estimates of Sensitive Subpopulations in the United States
Sensitive
Population
Individuals
Approximate
Percent of
U.S.
Population1
Citation/Notes
Pregnant women and neonates
Pregnant Women
6,240,000
2.2
Vital and Health Statistics, CDC (Ventura et
al., 2000)
Neonates (under one
month)2
317,137
0.1
U.S. 2000 Census (US Census Bureau,
2001a)
Age-based sensitive populations
Children (<5 years
old)
19,175,798
6.8
U.S. 2000 Census (US Census Bureau,
2001a)
Elderly (>65 years
old)
34,991,753
12.4
U.S. 2000 Census (US Census Bureau,
2001a)
Compromised immune status
Bone marrow
transplant Recipients
20,000
0.01
National Marrow Donor Program
http://. marrow.org/MEDIA/facts_figures.pdf
AIDS Patients
816,149
0.3
HIV/AIDS Surveillance Report, cases
through 2001 (CDC, 2002)
Organ transplant
recipients
23,143
0.01
U.S. Census Bureau, Statistical Abstract of
the U.S., based on 1998 data (2001b)
Total
61,583,980
21.8
Based on U.S. Census estimate (July 2000).
21/12 of the 2000 census population for age <1 year.
4.5 Summary of Baseline Assumptions
In Section 4.2 (Data Sources), EPA discusses the representativeness and quality of the
data used from the 2005 Six-Year Review 2 data and SDWIS/FED. An additional source of
uncertainty is introduced into the baseline in the monitoring frequencies estimated for systems,
shown in Exhibit 4.4. As described in that section, to arrive at this distribution EPA and the
Technical Workgroup conducted an analysis of 2005 Six-Year Data and individual state statutes,
and had the results validated by ASDWA during the TCRDSAC deliberations.
Exhibit 4.13 below presents a summary of the assumptions made in developing the
RTCR EA baseline. These assumptions may introduce uncertainty into the predictive model and
its outcomes in terms of benefits and costs. However, all of the assumptions influencing the EA
baseline (1989 TCR) do so in the same way as for the RTCR and Alternative option. Therefore,
EPA does not expect the net results of the analyses presented in this EA to be significantly
influenced by the uncertainty in the assumptions applied in developing the RTCR EA baseline.
Economic Analysis for the Final RTCR
4-25
September 2012
-------
Exhibit 4.13 Summary of Baseline Assumptions Influencing RTCR Estimates
Uncertainty
Factor
Current
Assumption
Section
with Full
Discussion
Contributes
Variability,
Uncertainty,
or
Constant?
Most Likely Effect on Benefits1
Most Likely Effect on Costs1
Underestimate
Overestimate
Unknown
Impact
Underestimate
Overestimate
Unknown
Impact
Pre-GWR
baseline
distribution of
systems
across
sampling
frequency
categories
(M/Q/A)
Varies per
system
category as
shown in Ex.
4.4
4.3.1
Variability,
Uncertainty
X
X
Selection of
2005 data from
Six Year
Review 2 for
use as
baseline
Data are
adequately
representative
of U.S. PWSs
and of sufficient
quality.2
4.2.2
Variability,
Uncertainty
X
X
Use of
SDWIS/FED
for inventory
data (2007 Q3)
and violations
data (2001 —
2007 Q3)
Data are
adequately
representative
of U.S. PWSs
and of sufficient
quality.
4.2.1
Variability,
Uncertainty
X
X
Distribution of
systems
across
sampling
frequency
categories
(M/Q/A)
following
Varies per
system
category as
shown in Ex.
5.9a
5.3.2.2
Variability,
Uncertainty
X
X
Economic Analysis for the Final RTCR
4-26
September 2012
-------
Uncertainty
Factor
Current
Assumption
Section
with Full
Discussion
Contributes
Variability,
Uncertainty,
or
Constant?
Most Likely Effect on Benefits1
Most Likely Effect on Costs1
Underestimate
Overestimate
Unknown
Impact
Underestimate
Overestimate
Unknown
Impact
implementation
of GWR
sanitary
surveys prior
to RTCR
implementation
Pwell—the
portion of GW
Systems
having viral
pathogens in
their source
waters
(adopted from
GWR EA)
21.58 percent
5.3.1
Constant,
Uncertainty
X
X
Psample—the
probability that
a random
sample will test
positive for
viral pathogens
given a
contaminated
source water
(adopted from
GWR EA)
Variable drawn
from a beta
distribution with
a range of
alpha and beta
estimates
having a
median value of
5.8 percent and
an expected
value of 12.4
percent.
5.3.1
Variability,
Uncertainty
X
X
Economic Analysis for the Final RTCR
4-27
September 2012
-------
Uncertainty
Factor
Current
Assumption
Section
with Full
Discussion
Contributes
Variability,
Uncertainty,
or
Constant?
Most Likely Effect on Benefits1
Most Likely Effect on Costs1
Underestimate
Overestimate
Unknown
Impact
Underestimate
Overestimate
Unknown
Impact
Reduced
occurrence for
sanitary
surveys
performed
under the
GWR
90 percent of
baseline
occurrence;
applied to equal
number of
systems
annually over
the sanitary
survey cycle
(20 percent
each year of 5
years for
CWSs; 33.3
percent each
year for 3 years
for NCWSs)
5.3.1
Constant,
Uncertainty
X
X
Reduced
occurrence for
GWR
compliance
monitoring
(applies to
subset of GW
Systems that
disinfect).
90 percent of
baseline
occurrence
5.3.1
Constant,
Uncertainty
X
X
Percentage of
nondisinfecting
GW Systems
choosing
disinfection
corrective
action vs. non-
disinfection
Range: High
end is the
percentage of
CWS entry
points
employing
disinfection by
system size,
5.3.1
Constant,
Uncertainty
X
X
Economic Analysis for the Final RTCR
4-28
September 2012
-------
Uncertainty
Factor
Current
Assumption
Section
with Full
Discussion
Contributes
Variability,
Uncertainty,
or
Constant?
Most Likely Effect on Benefits1
Most Likely Effect on Costs1
Underestimate
Overestimate
Unknown
Impact
Underestimate
Overestimate
Unknown
Impact
corrective
action in
response to
source water
quality issue
(adopted from
GWR)
low end
assumed to be
10 percent
based on
discussions
with State
representatives.
Population and
sensitive sub
population
estimates
Varies by
category
4.2.3.3 and
4.4
Variability,
Uncertainty
X
X
Notes:
1) All baseline variables or factors were incorporated into the predictive model in the same way for the 1989 TCR (baseline option) and the other two regulatory options
considered in this EA (RTCR and Alternative Option), therefore EPA expects that any under- or overestimation would affect the baseline and other options similarly,
resulting in no net effect on the results of the analysis.
2) A complete discussion of the data cleaning and assumption used in preparing Six-Year Review data for use in the RTCR baseline is provided in the Total Coliform
Rule Compliance Monitoring Data Quality and Completion Report (USEPA, 201 Oe).
Economic Analysis for the Final RTCR
4-29
September 2012
-------
5 Occurrence and Predictive Model
5.1 Introduction
Based on evaluation of the Safe Drinking Water Information System/Federal Version
(SDWIS/FED) violation data, the U.S. Environmental Protection Agency (EPA) assumes that the
national occurrence of total coliform (TC) and E. coli (EC) has reached a steady state in recent
years under the 1989 Total Coliform Rule (TCR). Cycles of normal deterioration and
repair/replacement appear to be occurring at the individual system level, while the numbers of
violations at the national level remain relatively unchanged. Chapter 4 (Exhibit 4.11) presents the
number of public water systems (PWSs) with TCR violations from 2001-2007, which shows that
national violation rates have remained relatively steady. Revisions to the 1989 TCR would affect
this steady state, likely resulting in a reduction of the underlying occurrence and associated
violations. However, prior to the Revised Total Coliform Rule (RTCR) implementation, the
Ground Water Rule (GWR), which became effective December 2009, would also have an effect
on the steady state. This chapter explains the development of a model that both reproduces the
steady-state occurrence conditions under the 1989 TCR and predicts the effects of both the
RTCR and the GWR in further reducing occurrence from current estimated levels.
The occurrence and predictive model developed for the RTCR was used for PWSs
serving 4,100 people or fewer, and it has two components. Occurrence and predictive modeling
focuses on PWSs serving <4,100 people because (a) there are adequate data to support
development of the occurrence distributions that are needed for the predictive components of the
model, (b) these systems are known to have higher occurrence levels than systems serving
>4,100 people, and (c) these systems, in particular, are subject to changes in the monitoring
requirements under the RTCR. The first component of the model characterizes how the presence
or positive rates of TC and EC detections vary across the population of PWSs serving <4,100
people in the U.S. These rates vary by the type of sample (routine or repeat), by analyte (TC or
EC), by system type (community water system (CWS), nontransient noncommunity water
system (NTNCWS), or transient noncommunity water systems (TNCWS)), and by system size.
EPA determined from the Six-Year Review data23 that systems differ considerably with respect
to the observed occurrence of TC and EC in both routine and repeat sampling. Some of this
variability can be explained by the type, size, water source, and disinfection practices of the
systems. However, even among systems having the same characteristics, some systems were
found to rarely experience positive assays, while others often did experience positive assays. The
second component of the model uses the TC and EC occurrence distributions, together with the
various sampling and response requirements, to simulate a set of nationally-representative
systems within the context of the three regulatory options (1989 TCR, RTCR, and Alternative
option) to predict changes in TC and EC occurrence, triggers, assessments, corrective actions
over time, and violations. The model generates estimates of reduced TC and EC occurrence
based on requirements of the RTCR to perform assessments and corrective actions not explicitly
required in the 1989 TCR. Additionally, the occurrence model takes into account reductions
attributable to the implementation of the GWR.
23 A discussion of the use of data from the Six-Year Review of SDWA regulations is presented in Chapter 4 of this
EA.
Economic Analysis for the Final RTCR
5-1
September 2012
-------
24
The occurrence model produces outputs over a 30-year time horizon that include
national estimates of TC and EC occurrence in PWSs across the U.S., and the resulting Level 1
25
and Level 2 assessments and corrective actions performed. These outputs are used to
characterize reduced exposure to the potential contamination under each regulatory option
considered, as presented in Chapter 6 of this economic analysis (EA), and to estimate the net
change in costs (i.e., incremental costs over the 1989 TCR) for PWSs across the U.S., as
described in Chapter 7 of this EA.
The two major components of the occurrence and predictive model used for systems
serving <4,100 people are described in Sections 5.2 and 5.3 of this chapter.
For systems serving more than 4,100 people, EPA estimated violation and trigger rates
using SDWIS/FED (USEPA, 2005) because the Six-Year Review data for PWSs serving more
than 4,100 people were not as robust as the Six-Year Review data for systems serving 4,100 or
fewer people. EPA did not quantify changes in violation or trigger rates for systems serving more
than 4,100 people among the 1989 TCR, RTCR, and Alternative option for the following
reasons: (1) limited Six-Year Review data to characterize these systems, (2) essentially
unchanged monitoring requirements across the regulatory options for these systems, and (3) level
of effort already occurring to implement the 1989 TCR. A complete description of the methods
for developing the occurrence estimates as described above is found in the remaining sections of
this chapter:
• Section 5.2 describes development of occurrence models to describe the occurrence
of TC and EC for systems serving less than or equal to 4,100 people.
• Section 5.3 describes how the occurrence model developed for systems serving less
than or equal to 4,100 people (Sections 5.3.1-5.3.2) is used to predict occurrence
following assessments and corrective actions (5.3.3) and includes a comprehensive
list of assumptions contributing to uncertainty, a sensitivity analysis, and a
discussion of model validation (5.3.3.1-5.3.3.2).
• Section 5.4 describes how outcomes were predicted for systems serving more than
4,100 people.
24 The occurrence model includes an additional 5 years prior to the promulgation of the RTCR monitoring to
account for effects of the GWR after it is fully implemented, plus 25 years of RTCR analysis, for a total of 30
modeled years.
25 The RTCR and Alternative option require PWSs to complete either a Level 1 or Level 2 assessment following
triggers specified in Exhibit 3.1 of this EA. The purpose of Level 1 and 2 assessments is to identify the presence of
sanitary defects and deficiencies in distribution system coliform monitoring practices, similar to an annual
inspection but more focused on determining the cause of a contamination event. In order to complete a Level 2
assessment the PWS must perform a more detailed examination of the system than for a Level 1 assessment,
including its monitoring and operational practices.
Economic Analysis for the Final RTCR
5-2
September 2012
-------
5.2 Modeling of Current Total Coliform and E. coli Occurrence for Systems Serving
<4,100 People
This section describes the development of parameters used to define variability across
and within groups of similar PWSs with regard to occurrence of TC and EC. Details related to
this section are provided in Appendix F to this EA.
5.2.1 Distributional Model and Notation
The Six-Year Review data demonstrate that, even among systems of similar size and
type, there are some PWSs that rarely find routine samples to be TC-positive and others that
often find TC-positive samples. For example, among the TNCWSs serving <101 people that are
using nondisinfected ground water, 47 systems reported ten or more routine TC positives while
assaying 40 or fewer samples during 2005. Among the same set of TNCWSs with nondisinfected
ground water are 60 that assayed 20 or more routine samples, finding all of them TC-negative.
Clearly, routine TC-positive rates vary considerably within this set of systems. Different
fractions of positives can be found among any basic subset of systems (defined by system type,
water type, and population served, as described in Section 5.2.2 of this EA) and for any type of
assay (both routine and repeat TC and both routine and repeat EC).
Within a group of similar systems, pRTTC;, which is defined as the probability that a
routine sample taken from system i will test positive, is assumed to vary as a beta random
variable. The beta distribution is commonly used to model varying probabilities because it is
limited to values between 0 and 1 and can assume a wide variety of shapes. EPA used a beta
distribution to describe how probabilities of virus-positive samples vary among virus-
contaminated wells in the EA for the Ground Water Rule (USEPA, 2006a).
Samples that test negative for TC are generally not tested further for the presence of EC,
unlike TC-positive samples. The probability that a TC-positive routine sample taken from system
i will also test EC-positive (EC+), pRTEC;, is assumed to vary among similar systems as another
beta random variable, but with parameters that can be quite different from those of the routine
TC samples. Although systems with frequent routine TC positives necessarily generated more
EC assays, the EC+ rates for these systems did not appear to be significantly different than the
EC+ rates of systems with rare TC positives. The variables pRTTC and pRTEC are therefore
modeled as independent beta-distributions.
Similar assumptions are extended to the repeat positive probabilities, pRPTC, and
pRPEC;. Like pRTTC; and pRTEC;, these are assumed to be independently beta-distributed
among sets of similar systems. A process for determining "sets of similar systems" is described
in Sections 5.2.3 and 5.2.4. The process is informed by system-specific data that are reduced
from the sample-specific data described in Chapter 4 of this EA, as discussed in the following
section.
5.2.2 Data Reduction
The occurrence model assumes that positive routine TC measurements for system i result
from some unobserved probability (pRTTC;). This probability may vary from month-to-month
and season-to-season. However, in this model each result, whether positive or negative, is
Economic Analysis for the Final RTCR
5-3
September 2012
-------
assumed to be an independent trial with positive probability pRTTC;. Each of these is called a
Bernoulli trial. The total number of positives observed for system i during a year (KRTTC;) is
therefore a binomial random variable, whose parameters are the system's positive probability
(pRTTC;) and the total number of samples assayed (NRTTCj). Data for the system can be
reduced to the number of positive samples (KRTTC;) and the number of samples assayed
(NRTTC;). The leading character "N" designates the number of samples tested and "K"
designates the number positive.
Similarly, the routine EC data for system i can be reduced to the number of routine TC-
positive samples that were assayed for EC (NRTEC,) and, of those, the number testing positive
(KRTECi). For many systems, both NRTEC and KRTEC will be zero because they encountered
no routine TC positives during 2005 (KRTTC = 0), and EPA assumes that in the case of a TC-
negative, no EC assay is performed.
Finally, the data for repeat samples for system i can be reduced to four numbers: NRPTC,
KRPTCi, NRPECi, and KRPECi. Again, the leading characters, "N" and "K", designate numbers
of samples assayed and numbers positive, respectively. "RP" designates that these are repeat
samples, while "TC" and "EC" designate assays for TC and EC, respectively. For many systems,
all four of these numbers are zero because no routine samples were TC positive.
It is important to note that only TC-positive samples are assayed for EC. The positive
probabilities for EC are therefore conditional, that is, they apply only to samples that have tested
TC positive. All TC-negative samples are assumed to be EC negative. The overall
(unconditional) probability that a routine sample from system i will be EC+ is the product
pRTTCi*pRTEC; and the overall (unconditional) probability that a repeat sample from the
system will be EC+ is the product pRPTCi*pRPECi.
The reduced dataset includes eight integer values (for NRTTC, KRTTC, NRTEC,
KRTEC, NRPTC, KRPTC, NRPEC, KRPEC) for each of approximately 93,000 systems.
Although none of the NRTTC is zero (all systems assayed at least one routine TC sample during
2005), the great majority of the other values are zeros.
5.2.3 Basic Subsets of Systems
The 92,747 systems serving <4,100 people in the Six-Year Review dataset (which
represent the approximately 147,000 total systems that exist in this size range) can be classified
by system type, source water type and disinfection status, and population served. The three types
of systems are CWS, NTNCWS and TNCWS. The three source water types are surface water
(SW), which is always disinfected, nondisinfected ground water (NondisGW) and disinfected
ground water (DisGW). Three important size categories are systems serving <100 people,
systems serving 101 to 1,000 people, and systems serving 1,001 to 4,100 people. Systems
serving more than 4,100 people are not included here, for reasons discussed in Chapter 4 of this
EA and earlier in this chapter. In total, there are 27 possible categories that each PWS may
belong to (3 system types x 3 source water types x 3 population served sizes). Each PWS may
only be in one of the 27 categories. Exhibit 5.1 lists the 27 basic subsets and corresponding
numbers of systems in the dataset.
Economic Analysis for the Final RTCR
5-4
September 2012
-------
Exhibit 5.1 Basic Classifications of PWSs Used for Occurrence Modeling
System Type
Water Type
Population
Served Group
Number of
Systems in 6-
Year Review
Data
cws
SW
<100
484
cws
sw
101-1,000
2,034
cws
SW
1,001-4,100
1,699
cws
DisGW
<100
3,662
cws
DisGW
101-1,000
7,678
cws
DisGW
1,001-4,100
2,707
cws
NondisGW
<100
3,788
cws
NondisGW
101-1,000
3,755
cws
NondisGW
1,001-4,100
644
NTNCWS
SW
<100
114
NTNCWS
SW
101-1,000
172
NTNCWS
SW
1,001-4,100
33
NTNCWS
DisGW
<100
1,577
NTNCWS
DisGW
101-1,000
1,780
NTNCWS
DisGW
1,001-4,100
252
NTNCWS
NondisGW
<100
4,026
NTNCWS
NondisGW
101-1,000
3,264
NTNCWS
NondisGW
1,001-4,100
152
TNCWS
SW
<100
838
TNCWS
SW
101-1,000
358
TNCWS
SW
1,001-4,100
30
TNCWS
DisGW
<100
8,112
TNCWS
DisGW
101-1,000
3,007
TNCWS
DisGW
1,001-4,100
119
TNCWS
NondisGW
<100
32,028
TNCWS
NondisGW
101-1,000
10,217
TNCWS
NondisGW
1,001-4,100
217
TOTAL
92,747
Source: Derived using 2005 Six-Year Review 2 Data.
Note: WaterType includes disinfection status, as informed by information
from SDWIS/FED. Systems with no indication of disinfection status in
SDWIS/FED were assumed to not disinfect (Nondis).
Economic Analysis for the Final RTCR
5-5
September 2012
-------
5.2.4 Estimation Methodology
The methodology described in this section applies to all four kinds of
measurements, routine TC (RTTC), routine EC (RTEC), repeat TC (RPTC) and repeat EC
(RPEC). However, to simplify the presentation, the description will be expressed only in terms
of RTTC. Results presented in Section 5.2.5 show that the final system classifications used to
model TC (both routine and repeat) are quite different from the final subsets used to model EC
occurrence.
For any individual subset of systems from Exhibit 5.1 or grouping of these subsets, two
methods were used to estimate parameters of the beta-distributed RTTC: maximum likelihood
estimation and Bayesian Markov Chain Monte Carlo (MCMC) sampling. The MCMC samples
are used to test whether two basic subsets have equal average positive rates and to check the
maximum likelihood estimates (MLEs). This check was conducted to ensure that the algorithm
used to find the MLE had not stopped running too soon.
Both the MLE and MCMC methodologies require computation of the likelihood function,
L(data | a,P), where a and P are parameters of beta-distributed pRTTC. For a specific system in
a set of similar systems, the number of positive RTTC samples is a binomial random variable
with parameters NRTTC and pRTTC, which are the numbers of RTTC samples assayed by the
system and the system's unobserved probability of a positive, respectively. The probability of a
positive, pRTTC, is a beta random variable, whose parameters are a and y3. Exhibit 5.2 is a
directed graph showing the model structure. At the top are the two high-level parameters {a and
P) and arrows from them showing their influence on the distribution of the system-specific
parameters, the positive probabilities or pRTTCs. Finally, the exhibit shows how the pRTTCs,
together with the numbers of samples assayed (NRTTCs), influenced the numbers estimated to
be positive (KRTTCs).
Economic Analysis for the Final RTCR
5-6
September 2012
-------
Exhibit 5.2 Directed Graph of Model Used for RTTC Occurrence
KRTTCj
NRTTC,
KRTTC,
NRTTC,
KRTTCn
NRTTCn
This same model could have been equally well represented with all of the arrows pointing
up to show the direction of inference, starting with the data (NRTTCs and KRTTCs), which
inform estimates of the unobserved positive probabilities (pRTTCs), and then those informing
estimates of the high level parameters (a and [3).
5.2.4.1 Maximum Likelihood Estimation
From this point through the end of this chapter, the approach for estimating parameters
for routine TC occurrence (RTTC) is the same approach used for repeat TC (RPTC) and both
routine and repeat EC occurrence (RTEC and RPEC). To simplify the notation, the designation
of measurement type (e.g., RTTC) will be dropped. For example, KRTTC is reduced to K,
NRTTC is reduced to N, and pRTTC is reduced to p. Subscript i is used to denote a particular
system in a basic subset of systems.
For a single system in a subset, if pi were known, the likelihood of the system's RTTC
data (Ki positives observed among Ni routine samples) would be given by the binomial
probability mass function:
dbinom (K;, N;, p;) =
N-!
KjKNj —Kj)!
Economic Analysis for the Final RTCR 5-7 September 2012
-------
But pi is not known, nor can it be observed explicitly. It is, therefore, treated as a beta
random variable. The expected value of the likelihood function is obtained by integrating the
product of the beta probability density function and the binomial probability mass function. The
beta probability density function is:
r(a + |3) / \a-l / \p-l
r(a)-r(P)
And the expected likelihood is:
-l
K.!
Ni! Ra + p) a-l+K; p-i+Ni-K;
•(P) (1 " P) dp
r(a)T(P)
Because both N; and K, are constant (observed data), the function only depends on the
factors involving p, a, and (3, so the factorial terms in the above may be ignored when seeking to
maximize the likelihood function. This simplifies the computation considerably, and the
resulting integral has a closed-form solution, which is:
l~(a + P)-I~(a + K.j-r^p + N. - Kjj Beta(a + K.,p + N. - K.j
r(a)T(P)-r^a + p + N.j Beta(a,p)
The total likelihood for a set of systems is the product of the system-specific likelihoods:
Beta|a + K., P + N. - K.j
Lkelihood(a, P) = | j -
Beta (a + P)
In order to find maximum likelihood estimates of parameters a and (3, some additional
steps were taken:
1. Log likelihood, rather than likelihood, was computed to avoid the problems of
numerical overflow and underflow. Because the log transformation is positive and
monotonic, parameters found to maximize the log likelihood function will also
maximize the untransformed likelihood function:
f Beta(a + K.,P + N. - K.V
LogLikelihood(a,P) = -
Beta(a + P)
i
2. Parameters of the Beta distribution were expressed in terms of the mean (a=a/(a+|3))
and dispersion (b=l/sqrt(a+|3)). In terms of these new parameters, the log likelihood
function is:
Economic Analysis for the Final RTCR
5-8
September 2012
-------
f f
Beta
LogLikelihood(a,b) =
In
~~ + Kj,——¦
,2 1 , 2
+ N. -K.
i i
A A
Beta
a 1 - a^i
— +
2 , 2
3. Bayesian MCMC estimation, as described in Appendix F was used to check the
MLEs and to support decisions about combining (pooling) different categories or
subsets of systems.
4. Where subsets of systems were found to have insufficient data to support maximum
likelihood estimation, engineering judgment was used to decide if combining with
other subsets made sense, for the purpose of parameter estimation. Rationale and
strategies for this are described in Appendix F.
5.2.5 Results
Appendix F of this EA describes how systems and their data were combined, or pooled,
for the purpose ofRTTC, RTEC, RPTC, andRPEC modeling. Section 5.2.5.1 and 5.2.5.2 report
the resulting MLEs in terms of a, b and a, /?.
RTTC data were addressed first because: a) so much more data were available for RTTC
than for the other three measurement types (RTEC, RPTC, and RPEC); and b) summary statistics
suggested that systematic differences due to system size and other factors were larger for TC
than for EC measurements.
When evaluating data for any one type of measurement, system size was evaluated first
because it was expected to be important for a number of reasons, primarily: a) larger systems
have more resources and are more likely to employ full-time professional operators than smaller
systems; and b) having more resources, the larger systems may provide better treatment and
distribution systems. System type and water type were evaluated after system size, but the order
in which these were evaluated does not appear to have been important.
5.2.5.1 Maximum Likelihood Estimates (MLEs)
Exhibit 5.3 provides MLEs for routine and repeat TC, while Exhibit 5.4 provides MLEs
for routine and repeat EC. Note that both of these exhibits present the a and b parameters as
defined in Section 5.2.4.1.
Economic Analysis for the Final RTCR
5-9
September 2012
-------
Exhibit 5.3 Maximum Likelihood a and b Parameter Estimates for RTTC and RPTC
Size
No. Systems
a RTTC
bRTTC
a RPTC
bRPTC
Community Water Systems (CWSs
, Surface Water (SW), Disinfecting
<100
484
0.01502
0.2090
0.06464
0.6107
101-1,000
2,034
0.009458
0.1490
0.06464
0.6107
1,001-4,100
1,699
0.005864
0.1275
0.03430
0.2721
Community Water Systems (CWSs
, Ground Water (GW), Disinfecting
<100
3,662
0.01592
0.2243
0.09415
0.8493
101-1,000
7,678
0.01243
0.2142
0.09415
0.8493
1,001-4,100
2,707
0.007767
0.1841
0.05221
0.4916
Community Water Systems (CWSs
, Ground Water (GW), Non-Disinfecting
<100
3,788
0.03147
0.2680
0.1898
0.8407
101-1,000
3,755
0.02690
0.2840
0.1898
0.8407
1,001-4,100
644
0.02690
0.2840
0.1312
0.6938
Nontransient, Noncommunity Water Systems (NTNCWSs), Surface Water (SW), Disinfecting
<100
114
0.01584
0.3576
0.1064
1.078
101-1,000
172
0.01127
0.2458
0.1064
1.078
1,001-4,100
33
0.01127
0.2458
0.1064
1.078
Nontransient, Noncommunity Water Systems (NTNCWSs), Ground Water (GW), Disinfecting
<100
1,577
0.01584
0.3576
0.1064
1.078
101-1,000
1,780
0.01127
0.2458
0.1064
1.078
1,001-4,100
252
0.01127
0.2458
0.1064
1.078
Nontransient, Noncommunity Water Systems (NTNCWSs), Ground Water (GW), Non-Disinfecting
<100
4,026
0.03658
0.4044
0.2575
0.9319
101-1,000
3,264
0.02690
0.2840
0.2575
0.9319
1,001-4,100
152
0.02690
0.2840
0.2575
0.9319
Transient, Noncommunity Water Systems (TNCWSs), Surface Water (SW), Disinfecting
<100
838
0.02312
0.2992
0.1439
1.108
101-1,000
358
0.02312
0.2992
0.1439
1.108
1,001-4,100
30
0.02312
0.2992
0.1439
1.108
Transient, Noncommunity Water Systems (TNCWSs), Ground Water (GW), Disinfecting
<100
8,112
0.02312
0.2992
0.1439
1.108
101-1,000
3,007
0.02312
0.2992
0.1439
1.108
1,001-4,100
119
0.02312
0.2992
0.1439
1.108
Transient, Noncommunity Water Systems (TNCWSs), Ground Water (GW), Non-Disinfecting
<100
32,028
0.04758
0.3990
0.2788
1.068
101-1,000
10,217
0.04758
0.3990
0.2487
0.9359
1,001-4,100
217
0.02467
0.3121
0.1694
0.4052
Note: Disinfection status ("Disinf?") is from information downloaded from SDWIS/FED. Systems with no indication of
disinfection status in SDWIS/FED are assumed to not disinfect.
Economic Analysis for the Final RTCR
5-10
September 2012
-------
Exhibit 5.4 Maximum Likelihood a and b Parameter Estimates for RTEC and RPEC
Size
No. Systems
a RTEC
bRTEC
a RPEC
bRPEC
Community Water Systems (CWSs
, Surface Water (SW), Disinfecting
<100
484
0.1291
0.6287
0.05175
1.153
101-1,000
2,034
0.1291
0.6287
0.05175
1.153
1,001-4,100
1,699
0.07514
0.8767
0.05175
1.153
Community Water Systems (CWSs
, Ground Water (GW), Disinfecting
<100
3,662
0.05449
0.6387
0.03950
1.168
101-1,000
7,678
0.05449
0.6387
0.03950
1.168
1,001-4,100
2,707
0.05449
0.6387
0.03950
1.168
Community Water Systems (CWSs
, Ground Water (GW), Non-Disinfecting
<100
3,788
0.04223
0.8129
0.03569
1.059
101-1,000
3,755
0.04223
0.8129
0.03569
1.059
1,001-4,100
644
0.04223
0.8129
0.03569
1.059
Nontransient, Noncommunity Water Systems (NTNCWSs), Surface Water (SW), Disinfecting
<100
114
0.1291
0.6287
0.05175
1.153
101-1,000
172
0.1291
0.6287
0.05175
1.153
1,001-4,100
33
0.1291
0.6287
0.05175
1.153
Nontransient, Noncommunity Water Systems (NTNCWSs), Ground Water (GW), Disinfecting
<100
1,577
0.08599
0.8796
0.07063
1.196
101-1,000
1,780
0.08599
0.8796
0.07063
1.196
1,001-4,100
252
0.08599
0.8796
0.07063
1.196
Nontransient, Noncommunity Water Systems (NTNCWSs), Ground Water (GW), Non-Disinfecting
<100
4,026
0.04223
0.8129
0.03569
1.059
101-1,000
3,264
0.04223
0.8129
0.03569
1.059
1,001-4,100
152
0.04223
0.8129
0.03569
1.059
Transient, Noncommunity Water Systems (TNCWSs), Surface Water (SW), Disinfecting
<100
838
0.1291
0.6287
0.05175
1.153
101-1,000
358
0.1291
0.6287
0.05175
1.153
1,001-4,100
30
0.1291
0.6287
0.05175
1.153
Transient, Noncommunity Water Systems (TNCWSs), Ground Water (GW), Disinfecting
<100
8,112
0.08599
0.8796
0.07063
1.196
101-1,000
3,007
0.08599
0.8796
0.07063
1.196
1,001-4,100
119
0.08599
0.8796
0.07063
1.196
Transient, Noncommunity Water Systems (TNCWSs), Ground Water (GW), Non-Disinfecting
<100
32,028
0.04223
0.8129
0.03569
1.059
101-1,000
10,217
0.04223
0.8129
0.03569
1.059
1,001-4,100
217
0.04223
0.8129
0.03569
1.059
Note: Disinfection status ("Disinf?") is from information downloaded from SDWIS/FED. Systems with no indication of
disinfection status in SDWIS/FED are assumed to not disinfect.
Economic Analysis for the Final RTCR
5-11
September 2012
-------
5.2.5.2 Derivation of a, ff from Maximum Likelihood Estimates
Using the estimates of a and b presented in Exhibits 5.3 and 5.4, EPA derived a, /?
estimates of the variation in PWSs within groups of systems, and incorporated these into the
occurrence model as described in Section 5.2.4.1.
These parameters are derived from a and b as follows:
a = a / b2
|3 = (1- a) / b2
Exhibits 5.5 and 5.6 present the estimates of a, corresponding to estimates of a and b in
exhibits 5.3 and 5.4.
Economic Analysis for the Final RTCR
5-12
September 2012
-------
Exhibit 5.5 a and /? Parameter Estimates for RTTC and RPTC
Size
No. Systems
aRTTC
(3RTTC
aRPTC
(3RPTC
Community Water Systems (CWSs
, Surface Water (SW), Disinfecting
<100
484
0.3438
22.54
0.173
2.508
101-1,000
2,034
0.4262
44.63
0.173
2.508
1,001-4,100
1,699
0.3608
61.17
0.463
13.04
Community Water Systems (CWSs
, Ground Water (GW), Disinfecting
<100
3,662
0.3166
19.57
0.1305
1.256
101-1,000
7,678
0.2709
21.53
0.1305
1.256
1,001-4,100
2,707
0.2291
29.26
0.2160
3.922
Community Water Systems (CWSs
, Ground Water (GW), Non-Disinfecting
<100
3,788
0.4381
13.49
0.2685
1.146
101-1,000
3,755
0.3336
12.07
0.2685
1.146
1,001-4,100
644
0.3336
12.07
0.2726
1.805
Nontransient, Noncommunity Water Systems (NTNCWSs), Surface Water (SW), Disinfecting
<100
114
0.1239
7.697
0.092
0.7697
101-1,000
172
0.1866
16.37
0.092
0.7697
1,001-4,100
33
0.1866
16.37
0.092
0.7697
Nontransient, Noncommunity Water Systems (NTNCWSs), Ground Water (GW), Disinfecting
<100
1,577
0.1239
7.697
0.09164
0.7697
101-1,000
1,780
0.1866
16.37
0.09164
0.7697
1,001-4,100
252
0.1866
16.37
0.09164
0.7697
Nontransient, Noncommunity Water Systems (NTNCWSs), Ground Water (GW), Non-Disinfecting
<100
4,026
0.2236
5.890
0.2965
0.8550
101-1,000
3,264
0.3336
12.07
0.2965
0.8550
1,001-4,100
152
0.3336
12.07
0.2965
0.8550
Transient, Noncommunity Water Systems (TNCWSs), Surface Water (SW), Disinfecting
<100
838
0.2584
10.92
0.117
0.6972
101-1,000
358
0.2584
10.92
0.117
0.6972
1,001-4,100
30
0.2584
10.92
0.117
0.6972
Transient, Noncommunity Water Systems (TNCWSs), Ground Water (GW), Disinfecting
<100
8,112
0.2584
10.92
0.1172
0.6972
101-1,000
3,007
0.2584
10.92
0.1172
0.6972
1,001-4,100
119
0.2584
10.92
0.1172
0.6972
Transient, Noncommunity Water Systems (TNCWSs), Ground Water (GW), Non-Disinfecting
<100
32,028
0.2989
5.983
0.2444
0.6323
101-1,000
10,217
0.2989
5.983
0.2839
0.8577
1,001-4,100
217
0.2532
10.01
1.03175
5.059
Note: Disinfection status ("Disinf?") is from information downloaded from SDWIS/FED. Systems with no indication of
disinfection status in SDWIS/FED are assumed to not disinfect.
Economic Analysis for the Final RTCR
5-13
September 2012
-------
Exhibit 5.6 a and /? Parameter Estimates for RTEC and RPEC
Size
No. Systems
aRTEC
(3RTEC
aRPEC
(3RPEC
Community Water Systems (CWSs
, Surface Water (SW), Disinfecting
<100
484
0.3266
2.203
0.03891
0.7129
101-1,000
2,034
0.3266
2.203
0.03891
0.7129
1,001-4,100
1,699
0.0978
1.203
0.03891
0.7129
Community Water Systems (CWSs
, Ground Water (GW), Disinfecting
<100
3,662
0.1336
2.318
0.02896
0.7041
101-1,000
7,678
0.1336
2.318
0.02896
0.7041
1,001-4,100
2,707
0.1336
2.318
0.02896
0.7041
Community Water Systems (CWSs
, Ground Water (GW), Non-Disinfecting
<100
3,788
0.0639
1.449
0.03184
0.8602
101-1,000
3,755
0.0639
1.449
0.03184
0.8602
1,001-4,100
644
0.0639
1.449
0.03184
0.8602
Nontransient, Noncommunity Water Systems (NTNCWSs), Surface Water (SW), Disinfecting
<100
114
0.3266
2.203
0.03891
0.7129
101-1,000
172
0.3266
2.203
0.03891
0.7129
1,001-4,100
33
0.3266
2.203
0.03891
0.7129
Nontransient, Noncommunity Water Systems (NTNCWSs), Ground Water (GW), Disinfecting
<100
1,577
0.1112
1.181
0.04936
0.6496
101-1,000
1,780
0.1112
1.181
0.04936
0.6496
1,001-4,100
252
0.1112
1.181
0.04936
0.6496
Nontransient, Noncommunity Water Systems (NTNCWSs), Ground Water (GW), Non-Disinfecting
<100
4,026
0.0639
1.449
0.03184
0.8602
101-1,000
3,264
0.0639
1.449
0.03184
0.8602
1,001-4,100
152
0.0639
1.449
0.03184
0.8602
Transient, Noncommunity Water Systems (TNCWSs), Surface Water (SW), Disinfecting
<100
838
0.3266
2.203
0.03891
0.7129
101-1,000
358
0.3266
2.203
0.03891
0.7129
1,001-4,100
30
0.3266
2.203
0.03891
0.7129
Transient, Noncommunity Water Systems (TNCWSs), Ground Water (GW), Disinfecting
<100
8,112
0.1112
1.181
0.04936
0.6496
101-1,000
3,007
0.1112
1.181
0.04936
0.6496
1,001-4,100
119
0.1112
1.181
0.04936
0.6496
Transient, Noncommunity Water Systems (TNCWSs), Ground Water (GW), Non-Disinfecting
<100
32,028
0.0639
1.449
0.03184
0.8602
101-1,000
10,217
0.0639
1.449
0.03184
0.8602
1,001-4,100
217
0.0639
1.449
0.03184
0.8602
Note: Disinfection status ("Disinf?") is from information downloaded from SDWIS/FED. Systems with no indication of
disinfection status in SDWIS/FED are assumed to not disinfect.
Economic Analysis for the Final RTCR
5-14
September 2012
-------
5.3 Predictive Modeling of Occurrence for Systems Serving Up to 4,100 People
For the purpose of understanding both the potential to reduce exposure to potential
contamination and for estimating the net change in costs that may be incurred under the RTCR,
EPA developed a model to predict the changes in TC and EC occurrence over the modeled
period with respect to requirements of the RTCR regulatory options. This predictive occurrence
model takes into account the requirements of existing regulations (1989 TCR and GWR) and
develops estimates of changes in occurrence (as well as the frequency of the new Level 1 and
Level 2 assessments and corrective actions) based on requirements of the regulatory options
considered under the RTCR. The resulting estimates are used to generate cost estimates that can
be compared across the three regulatory options26—the 1989 TCR, the RTCR, and the
Alternative option, as described in Chapter 7 of this EA. Potential benefits are compared in terms
of changes in the level of occurrence associated with each regulatory option, as presented in
Chapter 6 of this EA.
The occurrence model focuses on a 30-year period beginning with full implementation of
the GWR and covering a 25-year period of analysis from the anticipated RTCR effective date of
2015. The predictive model is implemented as a Monte Carlo simulation and currently uses
10,000 iterations (simulated systems) within each of the 27 type, source and size categories to
predict the effects of changes to the current rule over the modeling period.
The PWSs included in this model are those serving <4,100 people primarily because
these are the systems that will experience the major changes in monitoring frequency and other
requirements that will affect their TC and EC occurrence. For systems serving more than 4,100
people, which will not have significant monitoring frequency changes, a separate analysis was
performed based on historical violations rates instead of occurrence rates, as explained in Section
5.4 of this EA. The characteristics of all systems are described in Section 4.3 of this EA.
5.3.1 Summary of GWR factors and timing affecting the 1989 TCR and RTCR
The triggered and compliance monitoring requirements of the GWR must be
implemented beginning December 2009, which will be approximately five years prior to the
anticipated compliance date of the RTCR. Sanitary survey requirements must be implemented by
states by December 2012 for most CWSs and not until December 2014 for higher performing
CWSs and noncommunity water systems (NCWSs). To account for the impact of the GWR
requirements on baseline occurrence for those PWSs using a ground water source (and serving
<4,100 people), the model performs a number of adjustments to the existing data prior to its
application in the model as a baseline.
26 The three regulatory options considered in the RTCR EA (1989 TCR, RTCR, and Alternative option) are
described in detail in Chapter 3 of this EA. Briefly, the RTCR and Alternative option both differ from the 1989 TCR
primarily in the requirements for Level 1 and Level 2 assessments and corrective actions and for reduced monitoring
eligibility. The RTCR allows a continuation of current monitoring frequency while the Alternative option requires
systems to sample monthly for an initial period. Both allow reduced monitoring for qualified systems, but only the
RTCR allows annual monitoring. Systems on annual monitoring would require an annual site visit or an annual
voluntary Level 2 assessment.
Economic Analysis for the Final RTCR
5-15
September 2012
-------
Using 2005 SDWIS data (as explained in Section 4.3.1 of this EA) as a pre-GWR
27
inventory baseline, EPA applies probabilities of Psampie and Pweii to determine the number of
nondisinfecting systems that are expected to move to a disinfecting status in the five years prior
to RTCR implementation. EPA derived Psampie and Pweii28 in the GWR EA (USEPA, 2006a); Pweii
is a fixed value of 21.58 percent, and Psampie is drawn from a beta distribution with a range of
alpha and beta estimates having a median value for each system of 5.8 percent and an expected
value of 12.4 percent. Incorporation of these factors into the model ensures that the corrective
actions anticipated to be performed under the GWR are reflected in reduced occurrence for
systems implementing them. In the predictive model, simulated nondisinfecting ground water
systems having a TC-positive (TC+) take one or more source water samples per the GWR
triggered monitoring requirement and test the samples for fecal indicator presence (EC is
29
assumed to be the fecal indicator used). The probability of observing an EC-positive (EC+) is
based on the product of Pweii and the Psampie value drawn for that system. If positive, a second
assessment is done to determine if that system implements disinfection or a nondisinfection
corrective action based on the estimated proportion choosing those options as defined in the
30
GWR EA for the various types and sizes of systems.
The simulation model keeps track of systems that begin as nondisinfecting ground water
systems at the start of the analysis period and elect disinfection at some point during the 30-year
period from the effective date of the GWR (December 2009) through the end of the analysis
period for the RTCR. The inventory for disinfecting and nondisinfecting ground water systems
are then adjusted accordingly for subsequent stages of the analysis.
Exhibit 4.5 in Chapter 4 of this EA presents the baseline inventory of ground water
systems (GW systems) by disinfection status at the beginning of RTCR implementation. Those
GW systems achieving <4 log disinfection must sample at their source when they have a TC+
sample in their distribution system monitoring under the 1989 TCR and the RTCR. Sampling
costs were accounted for in the GWR EA (USEPA, 2006a).
Throughout the modeling period, GW systems that do not currently disinfect or add
disinfection may have EC+ source water samples and implement a nondisinfection corrective
action. Modeling in the GWR EA did not incorporate estimates of the effectiveness of
implementing GWR requirements because the GWR was only effective as of December 2009.
Absent this data, EPA applied best professional judgment in assuming that these corrective
27 The term "Pweu" refers to the probability that a randomly selected well across the United States will test positive
for a virus or fecal indicator species, such as EC, in its source water; "Psampie" is the probability that given a
contaminated well, a random sample at the well will eventually test positive for a virus or fecal indicator. The GWR
EA (USEPA, 2006a), and in particular the Baseline Chapter of that document, provide a detailed explanation of the
analysis used to generate these Pwen and Psampie estimates.
28 Under the GWR, GW systems with less than 4 log of treatment for viruses must sample their source for a fecal
indicator (e.g., E. coli) when they incur a TC+ under the 1989 TCR (or RTCR). If sample water is fecal indicator
positive, the system may be required to implement remedial actions at the source, one of which achieves disinfection
to 4 log for viruses.
29 Edberg, R. (2000) discusses the use of E. coli as an indicator of drinking water quality.
30 Costs for compliance with the GWR were considered in the GWR EA (USEPA, 2006a) and are not included in
the RTCR EA.
Economic Analysis for the Final RTCR
5-16
September 2012
-------
actions would have approximately the same effectiveness as a Level 2 corrective action since
they are similar in nature to an acute violation under the 1989 TCR. As shown in Exhibit 5.8, if a
nondisinfecting system performs a nondisinfecting corrective action, the model assumes that
occurrence is 0 percent for the remainder of that year and 2 full years after that. As a result of
this assumption, baseline occurrence will be significantly reduced as a function of movement
from nondisinfecting to disinfecting status. The modeling then assumes the P values are only 25
percent of baseline values for an additional five years. The model assumes the system returns to
baseline P values after that (approximately) seven year period.
Additionally, the model incorporates the effect of SSs performed under the GWR as of
January 2010, which is estimated by EPA to reduce TC and EC occurrence by 10 percent for the
remainder of the period of analysis. For TNCWSs, which are on a 5-year SS cycle under the
GWR, this reduction is initially applied to 20 percent of qualifying systems for each of the 5
years from 2010 to 2015. For CWSs, which are on a 3-year cycle, this reduction is initially
applied to 1/3 of systems annually from 2010 to 2012. EPA considers the estimate of 10 percent
in reduction of occurrence to be a conservative estimate of the effectiveness of the SS provision
of the GWR based on best professional judgment of the TCRDSAC. However, absent empirical
information to assume otherwise and because the factor is similarly incorporated by the model
into the estimates of occurrence for all three regulatory options, EPA assumes that this
conservatism does not significantly affect the net results for the RTCR or Alternative option,
which are calculated by comparison to the 1989 TCR (baseline).
A third GWR requirement that the model incorporates is compliance monitoring by the
subset of GW systems that disinfect, for which the model assigns a 10 percent reduction in
occurrence in effect throughout the period of analysis. Again, this GWR effectiveness estimate
may be conservative, but EPA assumes that because it is incorporated by the model into
occurrence estimates for all three regulatory options, the net effect on results of the RTCR and
Alternative option (in comparison to the 1989 TCR) is unlikely to be significant. Taking into
account the effect of both SS and compliance reductions, the disinfecting GW systems are
expected to ultimately have TC and EC occurrence rates that are 81 percent of the values derived
from the occurrence distributions (from 90 percent for sanitary surveys x 90 percent for
compliance monitoring).
A fourth and last GWR requirement that the model incorporates concerns the repeat
number of samples following a TC positive. For GW systems serving <1,000 people, the GWR
allows the PWS to use of one of the four repeat samples required under the 1989 TCR following
a TC+ sample to meet the requirement for source water testing under the GWR. By comparison,
under the RTCR, Alternative option, and the 1989 TCR scenarios for systems serving >1,000
people, only 3 repeat samples are required following a TC+ and in these cases an additional,
separate sample is used for the GWR source water fecal indicator assay.
As mentioned above with regard to SS and disinfecting system GWR compliance, the
actual reductions in occurrence that will result from implementation of GWR requirements may
differ from those predicted based on assumptions used in this model. EPA believes this would
affect neither the net analysis nor how the RTCR or Alternative option compare to each other,
since both would be affected similarly. Section 5.3.3.1 describes the various sources of
uncertainty and variability considered in this EA and presents an analysis of key variables.
Economic Analysis for the Final RTCR
5-17
September 2012
-------
5.3.2 Summary of Predictive Model
As the following Sections 5.3.2.1-5.3.2.2 describe, EPA modeled the expected trend in
occurrence of TC and EC+ assays over a pre-RTCR monitoring period of 5 years (to capture
applicable GWR effects), and a subsequent 25-year period of analysis for the three regulatory
options: continuation of the 1989 TCR, RTCR, and Alternative option.31 EPA sought to capture
the changes in occurrence from the implementation of the GWR requirements, and in turn, the
implementation of requirements of the RTCR and the Alternative option. Along with changes in
TC and EC occurrence, the model predicts behavioral changes: the number of Level 1 and Level
2 assessments (and associated number of Level 1 or Level 2 corrective actions) to be performed
and further resulting adjustments to occurrence, and changes in sampling regimens as systems
qualify for reduced monitoring requirements.
5.3.2.1 Approach to Estimating Parameters
As described in Section 5.2 of this chapter, EPA derived Beta distributions, characterized
by their a, /? parameters, based on analyses of the variation in TC and EC+ assays that occurred
both between and within 27 groupings of systems (as defined in Section 5.2.3); the resulting a, (J>
parameter estimates are presented in Exhibits 5.5 and 5.6. Based on an initial analysis, the 27
groups were condensed into the groups presented in Appendix F of this EA. The final Beta
distributions were based on these groupings to make the best use of the available data, combining
groups of systems when their occurrence levels were similar to avoid having small sample size in
any of the final groups.
5.3.2.2 Description of Predictive Model
Model output in this chapter (Exhibits 5.10-5.21) was generated for a 30-year period to
encompass 5 years of the effects of GWR implementation and 25 years of the effects of RTCR
implementation. For the purpose of the benefit-cost analysis, years 3 through 27 of this period
are included in the calculations and presented in the results shown in Chapter 6 (Benefits
Analysis), Chapter 7 (Cost Analysis), and Chapter 9 (Comparison of Benefits and Costs).
Initially in the modeled period, all systems under the 1989 TCR and the RTCR are assumed to
sample based on the sampling regimen applicable to their system type, size, type of water source,
and whether or not the system disinfects on the effective date of the RTCR.32 Under the
Alternative option, all systems are assumed to sample monthly for the first five years after the
RTCR effective date.
For the RTCR, the model adjusts these monitoring frequencies for GW systems serving
fewer than 1,000 people in years 9 and 11 of the 30-year period, or following 3 and 5 years after
of implementation of the RTCR, respectively, for CWSs and NCWSs. The adjustments are made
based on the acute and non-acute violations modeled during these 3- and 5-year periods,33 and
31 Footnote 4 briefly describes the three regulatory options, which are described in detail in Chapter 3 of this EA.
32 Based on EPA and TWG analysis of Six-Year data and individual state statutes during the TCRDSAC and review
by ASDWA.
33 Systems modeled to incur one or more acute or non-acute violations in this period are assumed to be on monthly
monitoring. For the RTCR, the remaining systems (incurring no acute or non-acute violations) are distributed
Economic Analysis for the Final RTCR
5-18
September 2012
-------
the period of assessment coincides with the sanitary survey for CWSs and NCWSs. The adjusted
monitoring frequencies then apply from years 9-30 and years 11-30 of the analysis for CWSs
and NCWSs, respectively.
For the Alternative option, the model adjusts the monitoring frequencies from the all-
monthly regimen based on the acute and non-acute violations34 predicted during the assessment
period, which is the first five years after the RTCR effective date, during which all systems
sample monthly. The model applies the schedule thus adjusted in years 11-30 of the analysis.
The timing for the adjustments made to the monitoring frequency schedules is consistent
with the length of time of the respective sanitary survey cycles. Although EPA uses the sanitary
cycle to define the period of assessment for updating the schedule, the results of the surveys are
not predicted in this model and are not directly incorporated into the calculations in any way.
Assumptions for the percentage of systems on monthly, quarterly, or annual monitoring upon
RTCR implementation are shown in Exhibit 5.9a-5.9c.
Using the monitoring frequencies described above, the model applies the baseline
occurrence associated with each category of PWS at the start of each modeled year (i.e., the level
of occurrence in 2005 under the 1989 TCR updated to reflect GWR implementation, as described
in Section 5.3.1), as described in Exhibits 5.7 and 5.8 for SW and GW systems, respectively. The
TC and EC occurrence (both from routine and repeat samples) are modeled by combining the
Beta distributions described previously with the binomial distribution to simulate TC+ and EC+
results (i.e., successes) given the number of samples (i.e., trials) each month.
For each simulated system in the model, a probability (P) is drawn randomly for each of
the RTTC, RPTC, RTEC, and RPEC types of samples from their corresponding Beta
distributions. These P values are used in the binomial distribution, together with the number of
samples taken each month, to predict the number of TC+ and EC+ results. In the absence of any
corrective actions being performed under the RTCR, those P values remain the same for that
system for the entire modeling period. If TC or EC occurrence is found that results in a specific
corrective action under the RTCR, temporary reductions in these P values are made as described
further below. In addition, when a system has a TC+ in a particular month that does not result in
a specific corrective action, it is assumed (for both the 1989 TCR and the RTCR) that the RTTC
rate in the following month will change to reflect either some worsening of the problem or an
improvement based on some other actions taken that are not part of the specific corrective
actions being considered. This was carried out by selecting at random a different RTTC P value
between quarterly and annual monitoring according to the percentage distribution from the initial M/Q/A estimates
prior to GWR implementation (Exhibit 4.4). Constraints to these adjustments include: 1) the final estimates of the
proportion of disinfecting CWSs on monthly monitoring should not decrease from the current baseline estimate; and
2) for the disinfecting NCWSs, there are caps on the percent that move to annual monitoring under the RTCR
(TNCWS <500: 13.8 percent, TNCWS 501-1,000: 8.6 percent, NTNCWS <500: 7.3 percent, NTNCWS 501-1,000:
3.0 percent).
34 Systems incurring one or more acute or non-acute violations in this period are assumed to be on monthly
monitoring. For the Alternative option, all of the remaining systems (incurring no acute or non-acute violations)
move to quarterly monitoring.
Economic Analysis for the Final RTCR
5-19
September 2012
-------
in those months following a TC+ for which explicit corrective actions were not initiated. In the
second month following a TC+ that does not result in a corrective action, the RTTC rate reverts
to the rate applicable in the month in which the TC+ occurred. EPA determined that this
approach provided overall results for the 1989 TCR that approximated what was observed in the
6-Year Review data and SDWIS data as discussed in the Section 5.3.3.2 (model validation)
below.
Based on the frequency of sampling and the occurrence probabilities for each system, the
model determines what percentage of these systems will be required to take an action (i.e.,
perform a Level 1 or Level 2 assessment in response to exceeding the threshold of multiple TC
positive assays or one EC+ assay, and, in some cases, implement a corrective action).
For the percentage of systems predicted to require a Level 1 assessment following the
requisite number of TC positives, the model assumes that 10 percent will find and address the
exact source of the problem under the RTCR and Alternative option. This represents a net
increase over the number expected to implement such a corrective action under the 1989 TCR.
The model assumes that systems that successfully identify and correct the problem will not have
a positive assay for the remainder of that year and also 1 additional full year after the assessment,
after which the probability of a TC+ or EC+ will be 50 percent that of the baseline P value for 3
35
additional years. After this period of reduced occurrence, the model assumes that occurrence
probabilities for this subset of systems return to those that were initially assigned. Because the
Level 1 and Level 2 assessments are a new requirement, empirical data on their effectiveness in
determining an appropriate corrective action is not available. Absent this evidence, EPA has
applied best professional judgment in choosing what it believes to be a significant but
conservative success rate—10 percent of Level 1 assessments will result in determining the root
of the problem and effectively implementing an appropriate corrective action. The assumptions
about the success rates of assessments and corrective actions, and the length of time before and
likelihood of subsequent positive samples, are assumptions made for modeling purposes and do
not reflect the expectation for any particular system.
Similarly, based on observing the requisite number of TC+ and/or EC+ results for
systems predicted to require a Level 2 assessment under the RTCR and Alternative option, a 10
percent increase in the number of systems from the 1989 TCR are assumed to implement a
corrective action that addresses the cause of the problem. For those systems, the model assumes
that occurrence is 0 percent for the remainder of that year and then for 2 full years after that. The
modeling then assumes the P values are only 25 percent of baseline values for an additional 5
years. The model assumes the system returns to baseline P values after that (approximately) 7-
year period.
For nondisinfecting GW systems, a percentage of positive samples is estimated (from the
PWeii and Psample parameters) to be related to a source water problem. All nondisinfecting GW
35Using best professional judgment informed by deliberations within the TCRDSAC, EPA assumes that some
systems will have 0 percent occurrence for some number of years longer than one year, and some will have
occurrence levels that are >0 percent within the first year. The assumption used in the analysis (systems will have 1
year with 0 percent occurrence, plus 2 to 3 years of reduced occurrence from their baseline occurrence) is meant to
represent an average system that experiences occurrence that is between these two scenarios.
Economic Analysis for the Final RTCR
5-20
September 2012
-------
systems having source water EC+ results36 will implement disinfection or nondisinfecting
corrective action. Those adding disinfection essentially get removed from the nondisinfecting
category and are reassigned to the disinfecting category for the remainder of the modeling
period. The portion of systems in this subset that are assumed to implement a nondisinfecting
corrective action37 are assumed to experience the same reduced occurrence levels as those
addressing Level 2 Assessments (i.e., 0 percent occurrence for 2 years, then occurrence at 25
percent of baseline for an additional 5 years). As Chapter 7 of this EA explains, no costs are
attributed to the RTCR for any corrective action related to source water EC occurrence since
such costs have already been considered in the GWR EA (USEPA, 2006a).
Outputs generated from this predictive simulation model (for PWSs serving 4,100 or
fewer people) include the following for each modeled year: occurrence of TC+ and EC+ assays,
the number of Level 1 and Level 2 assessments conducted, and the number of Level 1 and Level
2 corrective actions implemented. Based on these model outputs and the criteria for reduced
monitoring, the model also provides estimates of the number of systems on monthly, quarterly,
and annual sampling regimes (and by implication the number of routine, repeat, and additional
routine samples taken annually) under the RTCR and Alternative option. These requirements are
detailed in Chapter 3 of this EA.
The results of the RTCR predictive occurrence model are presented in the following
section, Section 5.3.3.
36 The model determines source water occurrence separately based on the subset of nondisinfecting GW systems, as
described in Section 5.3.1 (2nd paragraph) of this chapter.
37 Estimates of the number of systems implementing a nondisinfection corrective action are taken from Exhibit
6.21b ("Estimated Distribution of Source Water Contamination Corrective Actions") of the GWR EA (USEPA,
2006a). The GWR EA estimated the number of systems choosing a nondisinfection corrective action based on a
range; the high end was the percentage of CWS entry points employing disinfection at that time by system size, and
the low end was assumed to be 10 percent based on discussions with state representatives.
Economic Analysis for the Final RTCR
5-21
September 2012
-------
Exhibit 5.7 Simulated Impact of the RTCR and Alternative Option on Systems
Serving <4,100 People—Surface Water Systems
Decision Tree for Occurrence Analysis
Surface Water Systems
System samples in accordance with designated sampling
scheme at baseline occurrence probability
Level 2
Level 1 ?
Does systems
oes system
xfirid problem?
ind problem?
Is
the system
required to
take an
action?
No
Yes
System does not get a
positive sample for 2 years
after corrective action
System samples in
accordance with designated
sampling scheme at reduced
occurrence probability (25%
of baseline) for 5 years
System does not get a
positive sample for 1 year
after assessment
System samples in
accordance with designated
sampling scheme at reduced
occurrence probability (50%
of baselinei for 3 years
Evaluate eligibility for
reduced monitoring
' Is the system
required to take an action
during the period of reduced
occurrence following a
corrective action?
Mote 1: For level 1 and Level 2 assessments, systems are
estimated to find the exact source of th# problem 10% of the
time under the RTCR and Alternative Option
Economic Analysis for the Final RTCR
5-22
September 2012
-------
Exhibit 5.8 Simulated Impact of the RTCR and Alternative Option on Systems
Serving <4,100 People—Ground Water Systems
Decision Tree for Occurrence Analysis
Ground Water Systems
System samples in accordance with designated sampling
Is
the system
required to
take an
action?
Does
the system move to
disinfection as a
s the action
related to source
water quality?
orrective action?
Baseline occurrence
probability will be
significantly reduced as
a function of movement
from non-disinfecting to
disinfecting status
Level 2
Level 1?
Perform non-
disinfection
corrective action
oes system
oes system
ind problem?7
incf problem?.
Evaluate
eligibility for
reduced
monitoring"
System does not
get a positive
sample for 2 years
after corrective
action
System does not
get a positive
sample for 2 years
after corrective
action
System does riot
get a positive
sample for 1 year
after assessment
System samples in
accordance with designated
sampling scheme at reduced
occurrence probability {25%
of baseline) for 5 years
System samples in
accordance with designated
sampling scheme at reduced
occurrence probability (2.5%
of baseline) for
System samples in
accordance with designated
sampling scheme at reduced
occurrence probability (50%
for
s the system
required to take an action
during the period of reduced
occurrence following a
corrective action?
Not© 1 No costs will be attributed to any corrective action related to source
water quality These costs are attributable to the GWR
Note 2 For Level 1 and Level 2 assessments, systems are estimated to find
the exact source of the problem 10% of the time under the RTCR and
Alternative Option
Note 3 For modeling purposes, adjustments to a new steady state distribution
between monthly, quarterly, and annua! monitoring regimens are made once sn
the period of analysis - after 5 years fee noncommunity water systems and 3
years tor community water systems This time frame is meant to correspond
with a full sanitary survey cyete for these systems based on the Agency's
assumption that States may rant to coordinate their evaluation of PWS
eligibility for reduced monitoring with the sanitary survey cycle.
Economic Analysis for the Final RTCR
5-23
September 2012
-------
Exhibit 5.9a Percent of GW Systems Assumed to be on Monthly, Quarterly, and
Annual (M/Q/A) Monitoring by System Category under 1989 TCR (Baseline)—
Initial Estimates (Post-GWR Implementation)
Size
Number of
Systems
Monthly
Quarterly
Annual
Community Water Systems (CWSs), Disinfecting
<100
6,132
86.6%
13.4%
0.0%
101-1,000
12,762
88.5%
11.5%
0.0%
1,001-4,100
5,405
100.0%
0.0%
0.0%
Community Water Systems (CWSs), Non-Disinfecting
<100
5,806
86.6%
13.4%
0.0%
101-1,000
5,597
88.6%
11.4%
0.0%
1,001-4,100
1,038
100.0%
0.0%
0.0%
Nontransient, Noncommunity Water Systems (NTNCWSs), Disinfecting
<100
2,904
19.3%
64.7%
16.0%
101-1,000
3,621
18.5%
66.7%
14.7%
1,001-4,100
542
100.0%
0.0%
0.0%
Nontransient, Noncommunity Water Systems (NTNCWSs), Non-Disinfecting
<100
5,913
19.3%
64.7%
16.0%
101-1,000
4,710
18.5%
66.7%
14.8%
1,001-4,100
270
100.0%
0.0%
0.0%
Transient, Noncommunity Water Systems (TNCWSs), Disinfecting
<100
13,558
4.8%
62.9%
32.3%
101-1,000
6,014
7.9%
66.9%
25.2%
1,001-4,100
269
100.0%
0.0%
0.0%
Transient, Noncommunity Water Systems (TNCWSs), Non-Disinfecting
<100
46,642
4.8%
62.9%
32.3%
101-1,000
15,224
7.8%
66.8%
25.3%
1,001-4,100
348
100.0%
0.0%
0.0%
Notes:
1) EPAassumes that the frequencies noted in this exhibit forthe 1989 TCR would also
apply underthe RTCR in the first 10 years ofthe analysis period. Underthe Alternative
option, all systems would monitor monthly for the first 5 years after the RTCR effective
date.
2) Estimates forthe size categories presented in this exhibitare produced from weighted
averages from the following size categories (and corresponding estimates of proportion
sampling monthly, quarterly, or annually): <100; 101-500; 501-1,000; 1,001-2,500; 2,501 -
3,300; and 3,301-4,100. WaterType includes disinfection status, as informed by
information from SDWIS/FED.
3) Systems with no indication of disinfection ("Disinf?") status in SDWIS/FED were
assumed to not disinfect.
4) Some figures may notadd due to rounding in the model.
Economic Analysis for the Final RTCR
5-24
September 2012
-------
Exhibit 5.9b Percent of GW Systems Predicted to be on M/Q/A Monitoring by
System Category under RTCR—Adjusted Estimates (Post-RTCR Implementation)
Size
Number of
Systems
Monthly
Quarterly
Annual
Community Water Systems (CWSs), Disinfecting
<100
6,132
86.6%
13.4%
0.0%
101-1,000
12,762
88.5%
11.5%
0.0%
1,001-4,100
5,405
100.0%
0.0%
0.0%
Community Water Systems (CWSs), Non-Disinfecting
<100
5,806
86.6%
13.4%
0.0%
101-1,000
5,597
88.6%
11.4%
0.0%
1,001-4,100
1,038
100.0%
0.0%
0.0%
Nontransient, Noncommunity Water Systems (NTNCWSs), Disinfecting
<100
2,907
2.8%
89.9%
7.3%
101-1,000
3,621
2.6%
91.2%
6.3%
1,001-4,100
542
100.0%
0.0%
0.0%
Nontransient, Noncommunity Water Systems (NTNCWSs), Non-Disinfecting
<100
5,919
14.7%
78.0%
7.3%
101-1,000
4,710
13.5%
79.9%
6.5%
1,001-4,100
270
100.0%
0.0%
0.0%
Transient, Noncommunity Water Systems (TNCWSs), Disinfecting
<100
13,558
5.2%
81.0%
13.7%
101-1,000
6,014
5.1%
81.7%
13.2%
1,001-4,100
269
100.0%
0.0%
0.0%
Transient, Noncommunity Water Systems (TNCWSs), Non-Disinfecting
<100
46,642
16.9%
71.0%
12.0%
101-1,000
15,224
16.7%
71.2%
12.1%
1,001-4,100
348
100.0%
0.0%
0.0%
Notes:
1) M/Q/A percentage estimates are based on model predictions of acute and non-acute
violations from years 6-10 (NCWSs) and 6-8 (CWSs) of the 30-year modeled period.
Systems incurring 1 or more acutes or non-acutes in this period are assumed to be on
monthly monitoring for years 11 - 25 ofthe analysis; the remaining systems are
distributed between quarterly and annual monitoring according to the percentage
distribution from the initial M/Q/A estimates (Exhibit 5.9a). Other constraints applyto the
determination ofthe M/Q/Adistribution ofsystems as described in Section 5.3.2 above.
2) Estimates forthe size categories presented in this exhibitare produced from weighted
averages from the following size categories (and corresponding estimates of proportion
sampling monthly, quarterly, or annually): <100; 101-500; 501-1,000; 1,001-2,500; 2,501 -
3,300; and 3,301-4,100.
3) Systems with no indication of disinfection ("Disinf?") status in SDWIS/FED were
assumed to not disinfect.
4) Some figures may notadd due to rounding in the model.
Economic Analysis for the Final RTCR
5-25
September 2012
-------
Exhibit 5.9c Percent of GW Systems on M/Q/A Monitoring by System Category
under Alternative option—Adjusted Estimates (Post-RTCR Implementation)
Size
Number of
Systems
Monthly
Quarterly
Annual
Community Water Systems (CWSs), Disinfecting
<100
6,132
86.6%
13.4%
0.0%
101-1,000
12,762
88.5%
11.5%
0.0%
1,001-4,100
5,405
100.0%
0.0%
0.0%
Community Water Systems (CWSs), Non-Disinfecting
<100
5,806
86.6%
13.4%
0.0%
101-1,000
5,597
88.6%
11.4%
0.0%
1,001-4,100
1,038
100.0%
0.0%
0.0%
Nontransient, Noncommunity Water Systems (NTNCWSs), Disinfecting
<100
2,907
4.4%
95.6%
0.0%
101-1,000
3,621
4.5%
95.5%
0.0%
1,001-4,100
542
100.0%
0.0%
0.0%
Nontransient, Noncommunity Water Systems (NTNCWSs), Non-Disinfecting
<100
5,919
20.8%
79.2%
0.0%
101-1,000
4,710
21.5%
78.5%
0.0%
1,001-4,100
270
100.0%
0.0%
0.0%
Transient, Noncommunity Water Systems (TNCWSs), Disinfecting
<100
13,558
9.9%
90.1%
0.0%
101-1,000
6,014
10.1%
89.9%
0.0%
1,001-4,100
269
100.0%
0.0%
0.0%
Transient, Noncommunity Water Systems (TNCWSs), Non-Disinfecting
<100
46,642
25.8%
74.2%
0.0%
101-1,000
15,224
25.4%
74.6%
0.0%
1,001-4,100
348
100.0%
0.0%
0.0%
Notes:
1) M/Q/A percentage estimates are based on model predictions from years 6-10
(NCWSs) and 6-8 (CWSs) ofthe period of analysis. Systems incurring 1 or more
acutes or non-acutes in this period are assumed to be on monthly monitoring for years
11 - 25 ofthe analysis; the remaining systems are assumed to be on quarterly
monitoring. Other constraints applyto the determination ofthe M/Q/Adistribution of
systems as described in Section 5.3.2 above. This is a modeling assumption that
reflects an expected steadystate of some systems that will regain and others that will
lose reduced monitoring post RTCR implementation.
2) Estimates forthe size categories presented in this exhibitare produced from weighted
averages from the following size categories (and corresponding estimates of proportion
sampling monthly, quarterly, or annually): <100; 101-500; 501-1,000; 1,001-2,500; 2,501 -
3,300; and 3,301-4,100.
3) Systems with no indication of disinfection ("Disinf?") status in SDWIS/FED were
assumed to not disinfect.
4) Some figures may notadd due to rounding in the model.
Economic Analysis for the Final RTCR
5-26
September 2012
-------
5.3.3 Predictive Model Results
As noted above, the predictive model produces output that includes national estimates of
sampling and occurrence in PWSs across the United States, the resulting Level 1 and Level 2
assessments, corrective actions resulting from Level 1 and Level 2 assessments, and reductions
in sampling and occurrence for these systems in accordance with requirements under the 1989
TCR, the RTCR, and the Alternative option. In Exhibits 5.10 through 5.15, output includes the
model years 1-30 described in Section 5.3.2.2. Years 1-5 are summed for the 1989 TCR only
and represent the period prior to the start of monitoring under the final RTCR, including 5 years
after the GWR effective date. Year 5 in the analysis (also for just the 1989 TCR) is also shown
separately because it represents the baseline year, reflecting conditions just prior to the start of
monitoring under the RTCR. Model years 6 through 30 are summed for the 1989 TCR, the
RTCR, and the Alternative option because they represent the 25-year period of analysis
following promulgation of the RTCR.
For all types of PWSs, the model output is presented for the 1989 TCR, the RTCR, and
the Alternative option. The 1989 TCR with the GWR effects incorporated serves as the
appropriate baseline for comparisons with the RTCR.
EPA's general expectations for TC occurrence for GW systems shown in Exhibits 5.16—
5.18 are as follows:
1. Years 1-5 show a decline in the TC+ rates due to GWR implementation effects.
These include expected reductions in RTTC rates for all GW systems related to SSs,
for disinfecting GW systems related to the compliance monitoring, and for
nondisinfecting GW systems due to corrective actions taken following a source water
EC+ sample.
2. Years 6-30 reflect additional reductions (beyond GWR effects) in TC+ rates due to
the RTCR or Alternative option. The RTCR and Alternative option occurrence rates
are generally lower than those for the 1989 TCR with the GWR primarily because of
the additional corrective actions (Level 1 and Level 2). In some subgroups, some
counts are higher under the Alternative option than under the 1989 TCR due to the
increased sampling e.g., routine TC+ and routine EC+ results for GW NTNCWSs
(Ex. 5.11). The balance of the two effects of reduced sampling (reduced additional
and repeat samples) and the additional assessments and corrective actions, which
move the counts in opposite directions, is explored in the sensitivity analysis in
Section 5.3.3.1 of this chapter.
3. In Years 6 through 10 and years 11-30 differences in TC+ rates between the RTCR
and Alternative option are due to different proportions of systems that are performing
monthly, quarterly, or annual monitoring. In general, more monitoring (e.g., more
systems doing monthly than quarterly, or more systems doing quarterly than
annual)—in conjunction with the corrective actions—will push the TC occurrence
levels down (more samples are expected to be taken and fewer samples are likely to
be positive).
Economic Analysis for the Final RTCR
5-27
September 2012
-------
a. In years 6-10, the Alternative option has a lower TC+ rate than the RTCR
because all systems under the Alternative option begin the 5-year period on
monthly sampling. The increased sampling and resulting increase in corrective
actions performed will together reduce the occurrence rate.
b. By year 11, the systems are sampling according to their "steady state"
regimens, when all of the systems that qualified for reduced monitoring during
their respective periods of assessment (as described in Section 5.3.2.2) are
assigned to their reduced regimens in the model. For the RTCR, the
proportions of systems on M/Q/A sampling is identical to those under the
1989 TCR; under the Alternative option, however, systems that would qualify
under the 1989 TCR for annual monitoring can only reduce to quarterly
monitoring.
The following paragraphs describe TC occurrence in PWSs serving less than or equal to
4,100 people (as shown in Exhibits 5.16 through 5.21). See Appendix B for detailed breakouts of
TC occurrence by population category.
In Exhibit 5.16 (GW CWSs), the graph shows the expected drop in TC rates for years 1-5
as the GWR is implemented prior to RTCR. There are only very small differences between the
RTCR and Alternative option, and these are not observable in this graph. A comparison of the
estimates in Exhibit 5.10 clarifies these small differences. Most of these systems are already on
monthly sampling, so there is no significant difference seen between the RTCR and Alternative
option in years 6-8 for those few systems not on monthly sampling and it is assumed that the
monthly/quarterly distribution after this period returns to that under the 1989 TCR. Therefore,
the TC+ rates over the 25-year period for these systems show no difference between the RTCR
and Alternative option.
In Exhibit 5.17 (GW NTNCWSs), the graph also shows the expected drop in TC+ rates
for years 1-5 as the GWR is implemented prior to the RTCR. For years 6-10 the occurrence
rates decrease under the RTCR and the Alternative option due to the implementation of
corrective actions, and the rates for the Alternative option are lower than for the RTCR since all
systems are on monthly sampling under the Alternative option for this period (having the effect
of reducing the occurrence level as noted above). However, from year 11 to 30, the M/Q/A
distributions for these two are essentially the same and therefore the TC occurrence levels are the
same over this period.
In Exhibit 5.18 (GW TNCWS), the graph also shows the expected drop in TC+ rates for
years 1-5 as the GWR is implemented prior to the RTCR. For years 6-10, the TC rates drop
further as the corrective actions implemented in response to Level 1 and Level 2 assessments are
begun for both the RTCR and Alternative option. TC occurrence is markedly lower during this
period for the Alternative option where all systems are on monthly monitoring, whereas the
majority of systems under the RTCR are on quarterly monitoring. Beginning in Year 11, there is
a shift under the RTCR towards more systems on monthly monitoring than under the 1989 TCR,
and more systems on quarterly monitoring than under the 1989 TCR, and hence the overall TC+
rate drops. For the Alternative option, beginning in Year 11, a substantial portion of systems
move from the all monthly sampling during years 6-10 to quarterly sampling (but none on
Economic Analysis for the Final RTCR
5-28
September 2012
-------
annual sampling). Because there are slightly more systems under the Alternative option that end
up on monthly and quarterly monitoring than under the RTCR for years 11-30, the TC+ rate for
the RTCR is slightly higher than for the Alternative option.
EPA's general expectations for TC+ rates for SW systems shown in Exhibits 5.19-5.21
are as follows:
1. The 1989 TCR is provided as the baseline for comparison with the RTCR (the GWR
does not apply to SW systems and therefore its effect is not included in these graphs).
2. Because all SW systems of all sizes do monthly monitoring (no systems are on
quarterly or annual monitoring), the applicable provisions of the RTCR are identical
for the RTCR and Alternative option.
3. The reductions in TC+ rates shown by the RTCR relative to the 1989 TCR therefore
reflect solely the effects of the corrective actions implemented in response to Level 1
and Level 2 assessments.
In Exhibit 5.19 (SW CWSs), the graph shows little difference in TC+ rates between the
1989 TCR and RTCR. This is primarily because the current TC+ rate is very low for these
systems, and therefore there are few corrective actions implemented over the 25-year period.
In Exhibit 5.20 (SW NTNCWSs), the graph shows that there is a slightly greater
difference between the 1989 TCR and the RTCR for the SW NTNCWSs than for the SW CWSs
shown in Exhibit 5.19. This is because the baseline TC+ rate is slightly higher and, therefore,
more corrective actions are performed for SW NTNCWSs than SW CWSs.
In Exhibit 5.21 (SW TNCWSs), the graph indicates that this group exhibits the most
significant difference between the 1989 TCR and RTCR since the baseline TC+ rate for these
systems is approximately twice that of the SW NTNCWSs and SW CWSs.
A comparison of results for SW versus GW systems reveals a pattern of lower occurrence
for SW systems; this is explained by the uniform requirement that SW systems disinfect. Many
GW systems are not required to disinfect, and the result is a level of TC occurrence that is
periodically elevated in some systems, raising the overall average TC occurrence in GW system
categories. This difference is larger for the GW NCWSs than for the GW CWSs based on the
relatively large proportion of GW NCWSs that do not disinfect. The difference is largest among
the TNCWSs, a category in which the largest proportion of GW systems do not disinfect.
Complete results from the model (by year and system size) and additional graphs are
shown in Appendix B. The implications of these results in terms of changes in the level of risk
associated with contamination of PWS water supplies are discussed in Chapter 6 of this EA.
Model results are used to calculate national estimates of the net change in costs for PWSs in
Chapter 7 of this EA.
Economic Analysis for the Final RTCR
5-29
September 2012
-------
Exhibit 5.10 Ground Water Community Water System Model Output Cumulative Endpoints
GWR
Additional
Source
Routine
Repeat
Routine
Repeat
GWR
Non-
Level 1
Level 2
GWR, Non-
GWR,
Regulatory
Routine
Routine
Repeat
Water
TC+
TC+
E. co//+
E. co//+
E. co//+
Acute
Acute
Level 1
Level 2
Corrective
Corrective
Disinfecting
Disinfecting
Period
Scenario
Samples
Samples
Samples
Samples
Samples
Samples
Samples
Samples
Samples
Events
Events
Assessments
Assessments
Actions
Actions
Systems1
Systems2
Sum Years
6-30
1989 TCR No GWR
12,712,859
1,052,957
816,079
0
215,787
111,640
10,267
4,344
0
55,639
5,773
34,556
30,657
0
0
0
0
1989 TCR
12,761,616
864,605
588,252
93,331
176,379
66,797
7,236
2,106
872
39,769
3,335
26,947
18,545
0
0
714
158
RTCR
12,984,469
19,027
488,523
83,140
162,841
48,053
6,438
1,347
816
30,271
2,362
22,112
11,667
2,220
1,155
665
151
Alternate Option
13,073,297
16,686
490,683
83,299
163,561
47,829
6,562
1,595
781
29,955
2,511
21,917
11,660
2,242
1,175
622
160
Sum Years
1-5
1989 TCR No GWR
2,542,182
254,662
197,228
0
52,150
27,053
2,471
1,033
0
13,486
1,391
8,410
7,393
0
0
0
0
1989 TCR
2,550,295
180,676
122,391
19,799
36,791
14,208
1,532
460
312
8,358
711
5,542
4,103
0
0
203
109
Year 5
1989 TCR No GWR
508,401
42,520
32,969
0
8,719
4,613
402
176
0
2,259
240
1,415
1,241
0
0
0
0
1989 TCR
510,266
35,345
24,060
3,864
7,231
2,655
297
70
48
1,605
130
1,107
731
0
0
31
17
Source: RTCR predictive model.
Note: Years 1 - 5 in the model reflect incorporation of GWR effects (explained in Section 5.3.3) into pre-GWR baseline data; Year 5 in the model reflects baseline conditions immediately prior to the effective date of the RTCR; and Years 6-30 represent a
period of 25 years after monitoring begins under the RTCR.
1Number of Non-Disinfecting GWR Systems with an E. coli+ remaining as Non-Disinfecting Systems.
2Number of Non-Disinfecting GWR Systems with an E. coli+ changing to Disinfecting Systems.
Exhibit 5.11 Ground Water Nontransient Noncommunity Water System Model Output Cumulative Endpoints
GWR
Additional
Source
Routine
Repeat
Routine
Repeat
GWR
Non-
Level 1
Level 2
GWR, Non-
GWR,
Regulatory
Routine
Routine
Repeat
Water
TC+
TC+
E. co//+
E. co//+
E. co//+
Acute
Acute
Level 1
Level 2
Corrective
Corrective
Disinfecting
Disinfecting
Period
Scenario
Samples
Samples
Samples
Samples
Samples
Samples
Samples
Samples
Samples
Events
Events
Assessments
Assessments
Actions
Actions
Systems1
Systems2
Sum Years
6-30
1989 TCR No GWR
2,661,114
328,889
264,574
0
68,232
59,455
3,484
2,338
0
24,614
2,723
14,759
14,021
0
0
0
0
1989 TCR
2,668,723
271,756
178,204
42,339
55,965
34,716
2,534
1,269
446
17,922
1,778
11,658
8,991
0
0
351
95
RTCR
2,564,440
90,884
145,401
37,384
48,467
25,164
2,115
841
435
13,642
1,213
10,157
5,088
1,015
499
350
85
Alternate Option
3,345,137
71,729
185,814
48,181
61,938
31,758
2,628
1,089
496
17,403
1,547
11,996
7,585
1,194
761
389
107
Sum Years
1-5
1989 TCR No GWR
532,239
66,134
53,191
0
13,714
11,904
693
446
0
4,925
531
2,956
2,794
0
0
0
0
1989 TCR
533,205
58,599
38,489
9,192
12,094
7,867
574
288
150
3,924
403
2,477
2,084
0
0
99
52
Year 5
1989 TCR No GWR
106,453
13,330
10,711
0
2,760
2,425
137
89
0
1,001
102
603
555
0
0
0
0
1989 TCR
106,725
10,980
7,219
1,717
2,267
1,416
95
48
23
729
68
473
362
0
0
16
6
Source: RTCR predictive model.
Note: Years 1 - 5 in the model reflect incorporation of GWR effects (explained in Section 5.3.3) into pre-GWR baseline data; Year 5 in the model reflects baseline conditions immediately prior to the effective date of the RTCR; and Years 6-30 represent a
period of 25 years after monitoring begins under the RTCR.
1Number of Non-Disinfecting GWR Systems with an E. coli+ remaining as Non-Disinfecting Systems.
2Number of Non-Disinfecting GWR Systems with an E. coli+ changing to Disinfecting Systems.
Economic Analysis for the Final RTCR
5-30
September 2012
-------
Exhibit 5.12 Ground Water Transient Noncommunity Water System Model Output Cumulative Endpoints
Period
Regulatory
Scenario
Routine
Samples
Additional
Routine
Samples
Repeat
Samples
GWR
Source
Water
Samples
Routine
TC+
Samples
Repeat
TC+
Samples
Routine
E. co//+
Samples
Repeat
E. co//+
Samples
GWR
E. co//+
Samples
Non-
Acute
Events
Acute
Events
Level 1
Assessments
Level 2
Assessments
Level 1
Corrective
Actions
Level 2
Corrective
Actions
GWR, Non-
Disinfecting
Systems1
GWR,
Disinfecting
Systems2
Sum Years
6-30
1989TCR No GWR
7,508,835
1,766,220
1,489,330
0
374,523
373,849
17,881
14,195
0
143,238
15,757
87,899
79,745
0
0
0
0
1989 TCR
7,524,239
1,458,180
966,919
253,032
305,814
213,958
12,859
7,416
2,651
105,646
10,326
70,031
51,552
0
0
2,134
517
RTCR
9,780,374
674,328
981,521
283,807
327,174
191,401
13,121
5,792
2,993
99,961
8,421
71,817
39,668
7,188
4,039
2,455
538
Alternate Option
14,651,904
528,324
1,379,719
398,642
459,906
258,975
18,354
7,963
3,622
135,781
11,760
87,063
66,458
8,498
6,564
3,012
610
Sum Years
1-5
1989 TCR No GWR
1,501,831
350,417
295,149
0
74,227
73,576
3,577
2,709
0
28,251
3,150
17,357
15,692
0
0
0
0
1989 TCR
1,503,848
311,414
206,844
54,227
65,441
47,969
2,873
1,791
942
22,651
2,360
14,769
11,528
0
0
610
332
Year 5
1989 TCR No GWR
300,376
70,103
59,023
0
14,844
14,766
686
612
0
5,627
610
3,545
3,008
0
0
0
0
1989 TCR
300,923
59,747
39,565
10,447
12,536
8,813
506
331
129
4,269
429
2,821
2,101
0
0
87
42
Source: RTCR predictive model.
Note: Years 1 - 5 in the model reflect incorporation of GWR effects (explained in Section 5.3.3) into pre-GWR baseline data; Year 5 in the model reflects baseline conditions immediately prior to the effective date of the RTCR; and Years 6-30 represent a
period of 25 years after monitoring begins under the RTCR.
1 Number of Non-Disinfecting GWR Systems with an E. coli+ remaining as Non-Disinfecting Systems.
2Number of Non-Disinfecting GWR Systems with an E. coli+ changing to Disinfecting Systems.
Exhibit 5.13 Surface Water Community Water System Model Output Cumulative Endpoints
GWR
Additional
Source
Routine
Repeat
Routine
Repeat
GWR
Non-
Level 1
Level 2
GWR, Non-
GWR,
Regulatory
Routine
Routine
Repeat
Water
TC+
TC+
E. co//+
E. co//+
E. co// +
Acute
Acute
Level 1
Level 2
Corrective
Corrective
Disinfecting
Disinfecting
Period
Scenario
Samples
Samples
Samples
Samples
Samples
Samples
Samples
Samples
Samples
Events
Events
Assessments
Assessments
Actions
Actions
Systems1
Systems2
1989 TCR
3,516,224
137,063
98,139
0
27,729
5,244
2,909
293
0
3,322
666
2,829
1,237
0
0
0
0
Sum Years
RTCR
3,565,200
0
81,187
0
27,062
3,860
2,866
180
0
2,755
502
2,415
901
243
93
0
0
6-30
Alternate
Option
3,565,200
0
81,187
0
27,062
3,860
2,866
180
0
2,755
502
2,415
901
243
93
0
0
Sum Years
1-5
1989 TCR
703,236
27,458
19,651
0
5,550
1,066
584
57
0
675
137
573
255
0
0
0
0
Year 5
1989 TCR
140,622
5,559
3,956
0
1,117
224
119
13
0
138
29
115
55
0
0
0
0
Source: RTCR predictive model.
Note: Years 1 - 5 in the model reflect incorporation of GWR effects (explained in Section 5.3.3) into pre-GWR baseline data; Year 5 in the model reflects baseline conditions immediately prior to the effective date of the RTCR; and Years 6-30
represent a period of 25 years after monitoring begins under the RTCR.
1Number of Non-Disinfecting GWR Systems with an E. coli+ remaining as Non-Disinfecting Systems.
2Number of Non-Disinfecting GWR Systems with an E. coli+ changing to Disinfecting Systems.
Economic Analysis for the Final RTCR
5-31
September 2012
-------
Exhibit 5.14 Surface Water Nontransient Noncommunity Water System Model Output Cumulative Endpoints
GWR
Additional
Source
Routine
Repeat
Routine
Repeat
GWR
Non-
Level 1
Level 2
GWR, Non-
GWR,
Regulatory
Routine
Routine
Repeat
Water
TC+
TC+
E. co//'+
E. co//+
E. co// +
Acute
Acute
Level 1
Level 2
Corrective
Corrective
Disinfecting
Disinfecting
Period
Scenario
Samples
Samples
Samples
Samples
Samples
Samples
Samples
Samples
Samples
Events
Events
Assessments
Assessments
Actions
Actions
Systems1
Systems2
1989 TCR
222,494
13,728
10,670
0
2,796
1,127
360
51
0
458
105
292
294
0
0
0
0
Sum Years
RTCR
225,900
0
8,183
0
2,728
675
350
37
0
330
76
232
187
24
19
0
0
6-30
Alternative
Option
225,900
0
8,183
0
2,728
675
350
37
0
330
76
232
187
24
19
0
0
Sum Years
1-5
1989 TCR
44,500
2,740
2,131
0
558
221
72
10
0
90
21
57
59
0
0
0
0
Year 5
1989 TCR
8,898
554
429
0
113
43
15
2
0
17
4
11
12
0
0
0
0
Source: RTCR predictive model.
Note: Years 1 - 5 in the model reflect incorporation of GWR effects (explained in Section 5.3.3) into pre-GWR baseline data; Year 5 in the model reflects baseline conditions immediately prior to the effective date of the RTCR; and Years 6-30
representa period of 25 years after monitoring begins underthe RTCR.
1Number of Non-Disinfecting GWR Systems with an E. coli+ remaining as Non-Disinfecting Systems.
2Number of Non-Disinfecting GWR Systems with an E. coli+ changing to Disinfecting Systems.
Exhibit 5.15 Surface Water Transient Noncommunity Water System Model Output Cumulative Endpoints
Period
Regulatory
Scenario
Routine
Samples
Additional
Routine
Samples
Repeat
Samples
GWR
Source
Water
Samples
Routine
TC+
Samples
Repeat
TC+
Samples
Routine
E. co//'+
Samples
Repeat
E. co//+
Samples
GWR
E. co// +
Samples
Non-
Acute
Events
Acute
Events
Level 1
Assessments
Level 2
Assessments
Level 1
Corrective
Actions
Level 2
Corrective
Actions
GWR, Non-
Disinfecting
Systems1
GWR,
Disinfecting
Systems2
Sum Years
6-30
1989 TCR
609,514
71,280
57,424
0
14,612
8,365
1,855
398
0
3,148
765
1,893
2,214
0
0
0
0
RTCR
625,200
0
39,905
0
13,302
4,311
1,679
218
0
2,037
437
1,357
1,200
130
118
0
0
Alternative
Option
625,200
0
39,905
0
13,302
4,311
1,679
218
0
2,037
437
1,357
1,200
130
118
0
0
Sum Years
1-5
1989 TCR
121,900
14,265
11,476
0
2,920
1,671
382
81
0
628
152
370
449
0
0
0
0
Year 5
1989 TCR
24,373
2,891
2,314
0
589
331
77
17
0
124
30
74
88
0
0
0
0
Source: RTCR predictive model.
Note: Years 1 - 5 in the model reflect incorporation of GWR effects (explained in Section 5.3.3) into pre-GWR baseline data; Year 5 in the model reflects baseline conditions immediately prior to the effective date of the RTCR; and Years 6-30
representa period of 25 years after monitoring begins underthe RTCR.
1Number of Non-Disinfecting GWR Systems with an E. coli+ remaining as Non-Disinfecting Systems.
2Number of Non-Disinfecting GWR Systems with an E. coli+ changing to Disinfecting Systems.
Economic Analysis for the Final RTCR
5-32
September 2012
-------
Exhibit 5.16 Ground Water Community Water System
(Serving < 4,100) TC Occurrence
0.018
0.016
ST 0.014
0.012
0.010
0.008
1989 TCR no GWR
0.006
RTCR
Alt Option
0.004
Ll_
0.002
0.000
0
5
10
15
20
25
30
35
Years
Source: RTCR Predictive Model
Notes:
1) 1989 TCR No GWR: Represents the predicted TC occurrence, absent the effects of GWR
implementation, under the 1989 TCR requirements.
2) 1989 TCR: Represents the projected TC occurrence under the 1989 TCR requirements, including the
effects of GWR implementation.
3) RTCR: Represents the projected TC occurrence under RTCR requirements, including the effects of
GWR implementation.
4) Alt Option: Represents the projected TC occurrence under the Alternative option requirements,
including the effects of GWR implementation.
5) Steady state: By Year 11, all systems that qualify for reduced monitoring during their assessment period
(Section 5.3.2.2) have moved to their reduced sampling schedule under the RTCR and Alternative
option. This period in the model (from years 11—30) reflects a steady state with regard to proportions of
systems sampling on monthly, quarterly, or annual regimens.
Economic Analysis for the Final RTCR
5-33
September 2012
-------
Exhibit 5.17 Ground Water Nontransient Noncommunity Water System
(Serving < 4,100) TC Occurrence
0.025
0.020
(ft
>
C0
(/>
(/>
<
^ 0.015
g
w
o
0.
ii-
O
c
o
0.010
1989 TCR no GWR
O
(U
L.
LL
RTCR
Alt Option
0.005
0.000
0
5
10
15
20
25
30
35
Years
Source: RTCR Predictive Model
Notes:
1) 1989 TCR No GWR: Represents the predicted TC occurrence, absent the effects of GWR
implementation, under the 1989 TCR requirements.
2) 1989 TCR: Represents the projected TC occurrence under the 1989 TCR requirements, including the
effects of GWR implementation.
3) RTCR: Represents the projected TC occurrence under RTCR requirements, including the effects of
GWR implementation.
4) Alt Option: Represents the projected TC occurrence under the Alternative option requirements, including
the effects of GWR implementation.
5) Steady state: By Year 11, all systems that qualify for reduced monitoring during their assessment period
(Section 5.3.2.2) have moved to their reduced sampling schedule under the RTCR and Alternative
option. This period in the model (from years 11—30) reflects a steady state with regard to proportions of
systems sampling on monthly, quarterly, or annual regimens.
Economic Analysis for the Final RTCR
5-34
September 2012
-------
Exhibit 5.18 Ground Water Transient Noncommunity Water System
(Serving < 4,100) TC Occurrence
0.045
0.040
> 0.035
)
J 0.030
i-
01
> 0.025
>
£¦ 0.020
H-
o
c
o
TCR no GWR
TCR
0.015
RTCR
O
(U
Alt Option
LL
0.010
0.005
0.000
0
5
10
15
20
25
30
35
Years
Source: RTCR Predictive Model
Notes:
1) 1989 TCR No GWR: Represents the predicted TC occurrence, absent the effects of GWR
implementation, under the 1989 TCR requirements.
2) 1989 TCR: Represents the projected TC occurrence under the 1989 TCR requirements, including the
effects of GWR implementation.
3) RTCR: Represents the projected TC occurrence under RTCR requirements, including the effects of
GWR implementation.
4) Alt Option: Represents the projected TC occurrence under the Alternative option requirements,
including the effects of GWR implementation.
5) Steady state: By Year 11, all systems that qualify for reduced monitoring during their assessment period
(Section 5.3.2.2) have moved to their reduced sampling schedule under the RTCR and Alternative
option. This period in the model (from years 11—30) reflects a steady state with regard to proportions of
systems sampling on monthly, quarterly, or annual regimens.
Economic Analysis for the Final RTCR
5-35
September 2012
-------
Exhibit 5.19 Surface Water Community Water System
(Serving < 4,100) TC Occurrence
0.008
0.007
>
>
<2 0.006
%
0.005
"55 0.004
o
° 0.003
0
1 0.002
i-
LL
0.001
0.000
10
15 20
Years
25
-1989 TCR
-RTCR and Alt
Option
30
35
Source: RTCR Predictive Model
Notes:
1) 1989 TCR: Represents the projected TC occurrence under the 1989 TCR requirements.
2) RTCR and Alt Option: Represents the projected TC occurrence under the RTCR and Alternative option,
which have the same requirements for SW systems, resulting in the same TC occurrence under the two
options.
Economic Analysis for the Final RTCR
5-36
September 2012
-------
Exhibit 5.20 Surface Water Nontransient Noncommunity Water System
(Serving < 4,100) TC Occurrence
0.014
0.012
w
>
TO
(0
(0
<
O
%
0.010
0.008
t/>
o
a.
o
c
o
0.006
¦S 0.004
-1989 TCR
-RTCRand Alt
Option
0.002
0.000
I
5
10
15
20
25
30
35
Years
Source: RTCR Predictive Model
Notes:
1) 1989 TCR: Represents the projected TC occurrence under the 1989 TCR requirements, which
compose the baseline regulatory option for the RTCR EA.
2) RTCR and Alt Option: Represents the projected TC occurrence under the RTCR and Alternative option,
which have the same requirements for SW systems, resulting in the same TC occurrence under the two
options.
Economic Analysis for the Final RTCR
5-37
September 2012
-------
Exhibit 5.21 Surface Water Transient Noncommunity Water System
(Serving < 4,100) TC Occurrence
w
>1
-------
5.3.3.1 Uncertainty and Sensitivity Analysis in Occurrence Modeling
There are two primary sources of uncertainty in the RTCR occurrence modeling:
uncertainty related to the baseline occurrence model due to the limited data on which the model
is based, and uncertainty regarding the frequency and effectiveness of corrective actions and
their effects on subsequent TC and EC occurrence.
Data limitations in part derive from a lack of observed data on the effects of the
requirements of the GWR, for which implementation began as of December 2009. In the absence
of this data, EPA made assumptions regarding the effectiveness of each of the relevant
requirements, as described in Section 5.3.1 of this chapter. Each of these assumptions affects the
modeled 1989 TCR, which is the baseline for this EA. The predictive model uses this same
baseline (1989 TCR) in determining the net influence of both the RTCR and Alternative option
requirements on occurrence. Therefore, although the actual reductions in occurrence resulting
from GWR implementation may differ from the assumptions used in this model, this difference
would likely not affect the net results of the RTCR and Alternative option relative to the 1989
TCR, or their performance relative to each other. Other data limitations are discussed in Section
4.2 of this EA.
Many other assumptions incorporated into the predictive model also have no net effect on
results. More generally, assumptions that contribute to uncertainty in the predictive model results
can be separated into two categories: those that have no net effect on results and those that may
bias results downward or upward. Exhibit 5.22a below presents assumptions that influence the
baseline (1989 TCR) and the RTCR and Alternative option in the same way, and therefore are
expected to have no significant effect on net results. Exhibit 5.22b presents those assumptions
that may have a significant effect on net results because they influence only the RTCR and
Alternative option, but not the 1989 TCR. Exhibits 5.22a and 5.22b also include EPA's best
estimate of whether the assumption would tend to overestimate, underestimate, or have an
unknown impact.
Based on a lack of empirical data, EPA did not include in the model an estimate for the
number of systems that might be required to return to more frequent monitoring after having
been operating on a reduced schedule. A result of more frequent monitoring when combined
with assessments and corrective actions as required under the RTCR (or Alternative option), as
described elsewhere in this chapter, is expected to be an increase in TC and EC hits, and a
corresponding increase in assessments and corrective actions, resulting in improved water
quality. EPA expects that the result of not accounting for movement of systems back to a more
rigorous monitoring schedule will tend to underestimate benefits and costs for the RTCR and
Alternative option relative to the 1989 TCR. It may have more of an impact on net results for the
RTCR than for the Alternative option because the RTCR is expected to have a larger proportion
of systems on more reduced monitoring.
Economic Analysis for the Final RTCR
5-39
September 2012
-------
Exhibit 5.22a Summary of Model Parameters Influencing 1989 TCR, RTCR, and Alternative Option
Model Parameter
Current
Assumption
Section with
Full
Discussion
Contributes
Variability,
Uncertainty,
or Constant?
Most Likely Effect on Benefits1
Most Likely Effect on Costs1
Underestimate
Overestimate
Unknown
Impact
Underestimate
Overestimate
Unknown
Impact
Distribution of systems
across sampling
frequency categories
(M/Q/A) following
implementation of GWR
sanitary surveys prior to
RTCR implementation
Varies per system
category as shown
in Ex. 5.9a
5.3.2.2
Variability,
Uncertainty
X
X
Alpha, beta parameters
developed as model
inputs describing
distribution of occurrence
based on Six Year
Review 2 Data
(conducted in 2005)
Varies per system
category as shown
in Ex. 5.5-5.6
Appendix F
Variability,
Uncertainty
X
X
Pweii—the portion of
ground water systems
having viral pathogens in
their source waters
(adopted from GWR EA)
21.58 percent
5.3.1
Constant,
Uncertainty
X
X
Psampie—the probability
that a random sample will
test positive for viral
pathogens given a
contaminated source
water (adopted from
GWR EA)
Variable drawn from
a beta distribution
with a range of alpha
and beta estimates
having a median
value of 5.8 percent
and an expected
value of 12.4
percent.
5.3.1
Variability,
Uncertainty
X
X
Economic Analysis for the Final RTCR
5-40
September 2012
-------
Model Parameter
Current
Assumption
Section with
Full
Discussion
Contributes
Variability,
Uncertainty,
or Constant?
Most Likely Effect on Benefits1
Most Likely Effect on Costs1
Underestimate
Overestimate
Unknown
Impact
Underestimate
Overestimate
Unknown
Impact
Reduced occurrence for
sanitary surveys
performed under the
GWR.
90 percent of
baseline occurrence;
applied to equal
number of systems
annually over the
sanitary survey cycle
(20 percent each
year of 5 years for
CWSs; 33.3 percent
each year for 3
years for NCWSs)
5.3.1
Constant,
Uncertainty
X
X
Reduced occurrence for
GWR compliance
monitoring (applies to
subset of GW systems
that disinfect).
90 percent of
baseline occurrence
5.3.1
Constant,
Uncertainty
X
X
Percentage of
nondisinfecting GW
systems choosing
disinfection corrective
action vs. nondisinfection
corrective action in
response to source water
quality issue (adopted
from GWR)
Range: High end is
the percentage of
CWS entry points
employing
disinfection by
system size, low end
assumed to be 10
percent based on
discussions with
state
representatives.
5.3.1
Constant,
Uncertainty
X
X
Seasonal variation in TC
occurrence.
TC occurrence is
constant over the
course of the year.
5.3.3.1
Variability,
Uncertainty
X
X
Economic Analysis for the Final RTCR
5-41
September 2012
-------
Model Parameter
Current
Assumption
Section with
Full
Discussion
Contributes
Variability,
Uncertainty,
or Constant?
Most Likely Effect on Benefits1
Most Likely Effect on Costs1
Underestimate
Overestimate
Unknown
Impact
Underestimate
Overestimate
Unknown
Impact
RTTC rate in the months
following TC+ not
resulting in corrective
action
RTTC rate will
change to reflect
either worsening of
the problem or
improvement based
on actions taken
(other than the
corrective actions
being considered).
This was modeled
by selecting at
random a different
pRTTC value.
5.3.2.2
Variability,
Uncertainty
X
X
Note: For variables or factors that were incorporated into the predictive model in the same way for the 1989 TCR (baseline option) and the other two regulatory options
considered in this EA (RTCR and Alternative Option), EPA expects that any under- or overestimation would affect the baseline and other options equally, resulting in no net
effect on the results of the analysis.
Economic Analysis for the Final RTCR
5-42
September 2012
-------
Exhibit 5.22b Summary of Model Parameters Influencing RTCR and Alternative option Only
Model Parameter
Current
Assumption
Section with
Full
Discussion
Contributes
Variability,
Uncertainty,
or Constant?
Most Likely Effect on Benefits1
Most Likely Effect on Costs1
Underestimate
Overestimate
Unknown
Impact
Underestimate
Overestimate
Unknown
Impact
Percentage of systems
implementing an Effective
corrective action after a
Level 1 or Level 2
Assessment
10 percent
5.3.3.1
Constant,
Uncertainty
X
X
Occurrence immediately
following corrective action
implementation for period
of 3 or 5 years (Level 1
and 2, respectively) (or
nondisinfecting GW
systems choosing a
nondisinfection corrective
action)
0 percent
5.3.3.1
Constant,
Uncertainty
X
X
Duration of initial phase of
reduced occurrence (0
percent) following
corrective action
implementation
Remainder of year
+ 1 or 2 years
(Level 1 and 2,
respectively)
5.3.3.1
Constant,
Uncertainty
X
X
Occurrence in second
phase following corrective
action implementation
(after initial phase of 0
percent occurrence)
50 percent or 25
percent of baseline
occurrence
(Level 1 and 2,
respectively)
5.3.3.1
Constant,
Uncertainty
X
X
Duration of reduced
occurrence in second
phase following corrective
action implementation
3 or 5 years
(Level 1 and 2,
respectively)
5.3.3.1
Constant,
Uncertainty
X
X
Economic Analysis for the Final RTCR
5-43
September 2012
-------
Model Parameter
Current
Assumption
Section with
Full
Discussion
Contributes
Variability,
Uncertainty,
or Constant?
Most Likely Effect on Benefits1
Most Likely Effect on Costs1
Underestimate
Overestimate
Unknown
Impact
Underestimate
Overestimate
Unknown
Impact
Duration of reduced
occurrence (0 percent) in
initial phase following
corrective action
implementation (applies
to nondisinfecting GW
systems choosing
nondisinfecting corrective
action)
Remainder of year
+ 2 years
5.3.3.1
Constant,
Uncertainty
X
X
Occurrence in second
phase following corrective
action implementation
(after initial phase of 0
percent occurrence)
(applies to nondisinfecting
GW systems choosing
nondisinfecting corrective
action)
25 percent of
baseline
occurrence
5.3.1
Constant,
Uncertainty
X
X
Duration of reduced
occurrence in second
phase following corrective
action implementation
(applies to nondisinfecting
GW systems choosing
nondisinfecting corrective
action)
5 years
5.3.2.2
Constant,
Uncertainty
X
X
Economic Analysis for the Final RTCR
5-44
September 2012
-------
Model Parameter
Current
Assumption
Section with
Full
Discussion
Contributes
Variability,
Uncertainty,
or Constant?
Most Likely Effect on Benefits1
Most Likely Effect on Costs1
Underestimate
Overestimate
Unknown
Impact
Underestimate
Overestimate
Unknown
Impact
For systems serving more
than 4,100 people:
Application of a rate that
varies by type of system
and water source but is
held constant through
time (does not change
over the 25 years of
analysis)
These groups incur
the same number
of violations
annually throughout
the analysis (i.e.,
no reduction in
violations in the
years immediately
following corrective
action). The
number of
violations is used to
calculate the
number of Level 1
and Level 2
assessments and
corrective actions
by PWS category
that systems will
implement under
the 1989 TCR,
RTCR, and
Alternative option.
5.4.3
Variability,
Uncertainty
X
X
Note: For variables and factors that affect the options considered (RTCR and Alternative option) but do not have a parallel influence on the 1989 TCR (baseline), EPA expects
that any under- or over-estimation will have an effect as noted in the above exhibit on the net results for the RTCR or Alternative option.
Economic Analysis for the Final RTCR
5-45
September 2012
-------
Key Factors Driving Analysis Uncertainty
The remainder of this section focuses on those assumptions identified in Exhibit 5.22b as
contributing the most uncertainty to the net results of the analysis: the frequency and
effectiveness of corrective actions (i.e., percent reduction in occurrence and duration of reduction
period resulting from corrective actions) applied in the predictive model. These are the key
drivers in determining net results, that is, the difference between the RTCR or Alternative option
and the 1989 TCR.
As described previously in Section 5.3, EPA incorporated the following assumptions for
the frequency and effectiveness of the corrective actions that systems will implement following a
Level 1 or a Level 2 assessment:
Level 1: Following a Level 1 Assessment in a particular year, 10 percent of the systems
performing the assessment will implement corrective actions that will reduce the TC and EC
occurrence levels to 0 percent for the remainder of that year and for 1 full year after that. Then,
for an additional 3 full years, the TC and EC occurrence levels for each of these systems will be
reduced to 50 percent of their initial values. After that, the TC and EC occurrence levels return to
their initial values.
Level 2: Following a Level 2 Assessment in a particular year, 10 percent of the systems
performing the assessment will implement corrective actions that will reduce the TC and EC
occurrence levels to 0 percent for the remainder of that year and for 2 full years after that. Then,
for an additional 5 full years, the TC and EC occurrence levels for each of these systems will be
reduced to 25 percent of their initial values. After that, the TC and EC occurrence levels return to
their initial values.
Also, for the nondisinfecting GW systems that are found in the simulation to have an
EC+ in their source water but do not move to disinfection, the assumptions for the effectiveness
of the nondisinfecting corrective actions are identical to those above for Level 2 corrective
actions. (However, note that all nondisinfecting systems that discover EC+ source water that do
not go to disinfection will implement these corrective actions (under GWR), not just 10 percent
as assumed for systems that already disinfect.)
To assess the influence of these assumptions on the model results, EPA ran the model
with alternative assumptions reflecting corrective actions that are less effective (50 percent of the
assumption of efficacy in the primary analysis in this EA) and corrective actions that are more
effective (2x the assumption of efficacy in the primary analysis). The assumptions made in the
primary analysis, and the alternative assumptions made in this sensitivity analysis, do not
represent a sum of the effectiveness of all individual corrective actions, but indicate the average
frequency across all systems with which a given system will correctly diagnose and effectively
implement a corrective action, mitigating the source of the problem. The alternative assumptions
assume this frequency will be half or twice that used in the primary analysis. Using these
alternative assumptions, key model outputs that can serve as proxy indicators of the costs and
benefits of the RTCR were compared with those same outputs from the main model assumptions
for the RTCR.
Economic Analysis for the Final RTCR
5-46
September 2012
-------
For the less effective corrective actions, the alternative assumptions for Level 1 and Level
2 are:
Level 1: Following a Level 1 Assessment in a particular year, 5 percent of the systems
performing the assessment will implement corrective actions that reduce the TC and EC
occurrence levels to 0 percent for the remainder of that year only. Then, for an additional 2 full
years, the TC and EC occurrence levels for each of these systems will be reduced to 50 percent
of their initial values. After that, the TC and EC occurrence levels return to their initial values.
Level 2: Following a Level 2 Assessment in a particular year, 5 percent of the systems
performing the assessment will implement a corrective action that will reduce the TC and EC
occurrence levels to 0 percent for the remainder of that year and for 1 full year after that. Then,
for an additional 3 full years, the TC and EC occurrence levels for that system will be reduced to
25 percent of their initial values. After that, the TC and EC occurrence levels return to their
initial values. (These alternative Level 2 durations were also applied to the nondisinfecting GWR
corrective actions.)
For the more effective corrective actions, the alternative assumptions for Level 1 and
Level 2 are:
Level 1: Following a Level 1 Assessment in a particular year, 20 percent of the systems
performing the assessment will implement corrective actions that will reduce the TC and EC
occurrence levels to 0 percent for the remainder of that year and for two full years after that.
Then, for an additional 6 full years, the TC and EC occurrence levels for that system will be
reduced to 50 percent of their initial values. After that, the TC and EC occurrence levels return to
their initial values.
Level 2: Following a Level 2 Assessment in a particular year, 20 percent of the systems
performing the assessment will implement corrective actions that will reduce the TC and EC
occurrence levels to 0 percent for the remainder of that year and for 4 full years after that. Then,
for an additional 10 full years, the TC and EC occurrence levels for that system will be reduced
to 25 percent of their initial values. After that, the TC and EC occurrence levels return to their
initial values. (Again, these alternative Level 2 durations were also applied to the nondisinfecting
GWR corrective actions.)
A summary of the assumptions used in these alternative sensitivity analyses is provided
in Exhibit 5.23. The analyses were run in the predictive model for the RTCR for the
approximately 60,000 nondisinfecting GW TNCWSs serving <500 people. This subgroup
contains the largest number of systems and has on average the largest number of violations of all
subgroups. This subgroup may experience the largest change in activity under the RTCR
because, unlike the larger systems that have more resources at their disposal, these systems are
not generally going beyond requirements of the 1989 TCR, and generally have less frequent
requirements for sampling under the 1989 TCR than the larger systems. To conduct the
sensitivity analysis described in this section, EPA applied to the predictive model the alternative
assumptions that are presented in Exhibit 5.23 along with assumptions applied in the primary
analysis for comparison.
Economic Analysis for the Final RTCR
5-47
September 2012
-------
Exhibit 5.23 Sensitivity Analysis Assumptions for Frequency and Effectiveness of
Corrective Actions following Level 1 or 2 Assessments
Type of
Assessment
% Performing
Corrective
Action after a
Level 1 or Level
21 Assessment
Reduced hit
rate
Immediately
after Corrective
Action
Period of time
for reduced hit
rate2
Additional
period of
reduced hit rate
(yrs)
Reduction from
initial hit rate in
additional
reduced period
Predictive Model
Level 1
10%
0%
Remainder + 1
3
50%
Level 2
10%
0%
Remainder + 2
5
25%
Sensitivity 1 -
Less Effective
Level 1
5%
0%
Remainder only
2
50%
Level 2
5%
0%
Remainder + 1
3
25%
Sensitivity 2 -
More effective
Level 1
20%
0%
Remainder + 2
6
50%
Level 2
20%
0%
Remainder + 4
10
25%
Notes:
1. Level 2 assumptions for the predictive model and sensitivity analyses are also applied to 100% of nondisinfecting ground water systems
that incur an EC+ in the source water and implement a nondisinfecting corrective action.
2. "Remainder" refers to the balance of the yearfollowing implementation of a corrective action.
Exhibit 5.24a Cumulative Effect of Alternative Assumptions for Corrective Action
(CA) Effectiveness and Duration on RTCR Model Results
(Nondisinfecting TNCWS Serving <500 People over 25 Years)
Corrective
Actions
Assumption
Model Output
RTTC+
RTTC+
as rate
RPTC+
RTEC+
RPEC+
GWR EC +
Nonacutes
Acutes
L1 Assmt
L2 Assmt
L1 CA
L2 CA
Low CAs1
300,209
3.81%
204,124
11,585
6,613
3,364
104,161
9,162
68,941
48,658
3,539
2,395
RTCR2
277,776
3.55%
174,200
9,937
5,038
2,906
90,389
7,170
64,275
36,090
6,448
3,671
High CAs3
240,248
3.11%
124,822
8,271
3,942
2,240
67,662
5,352
53,823
20,710
10,664
3,918
50% Option4
229,270
2.97%
110,839
7,887
3,225
2,641
61,424
4,534
53,896
12,506
27,227
6,288
% Change from RTCR
Low CAs1
8.1%
7.42%
17.2%
16.6%
31.2%
15.8%
15.2%
27.8%
7.3%
34.8%
-45.1%
-34.7%
High CAs3
-13.5%
-12.43%
-28.3%
-16.8%
-21.8%
-22.9%
-25.1%
-25.4%
-16.3%
-42.6%
65.4%
6.7%
50% Option4
-17.5%
-16.22%
-36.4%
-20.6%
-36.0%
-9.1%
-32.0%
-36.8%
-16.1%
-65.3%
322.3%
71.3%
Relative Change
Low CAs1
1.08
1.08
1.17
1.17
1.31
1.16
1.15
1.28
1.07
1.35
0.55
0.65
High CAs3
0.86
0.83
0.72
0.83
0.78
0.77
0.75
0.75
0.84
0.57
1.65
1.07
50% Option4
0.83
0.77
0.64
0.79
0.64
0.91
0.68
0.63
0.84
0.35
4.22
1.71
Notes:
1. "LowCAs" uses half of the primary analysis estimate, or 5% incremental increase in CAs beyond the 1989 TCR (50% decrease in the assumption).
2. "RTCR" refers to the RTCR under assumptions used in the primary analysis, which includes an assumption that 10% more CAs addressing the root cause of a TC/EC
event will be implemented than under the 1989 TCR (0% change in the assumption).
3. "High CAs" doubles the primary analysis CAs estimate from 10% to 20% (100% increase in the assumption).
4. "50% option" uses a CAs estimate of 50% instead of the primary analysis assumption of 10% (500% increase in the assumption).
Cumulative effects shown in this exhibit are based on application in the predictive occurrence model of the alternative assumptions for frequency and effectiveness of
corrective actions following Level 1 or 2 Assessments, as presented in Ex. 5.23. The effectiveness of corrective actions is based both on the effectiveness of the Level 1 or
Level 2 assessments performed in diagnosing the correct system deficiency(ies) and upon the effectiveness of implementation of the intended corrective action.
Economic Analysis for the Final RTCR
5-48
September 2012
-------
As expected, the results of these runs indicate that the less frequent and less effective
corrective action assumptions result in more TC+ and EC+ samples and having more non-acute
and acute events than the main assumptions. In addition, more LI and L2 assessments are
conducted in response to those additional events, but fewer corrective actions are performed.
Conversely, the more frequent and more effective corrective action assumptions lead to
fewer TC+, EC+, non-acute, and acute events occurring and more LI and L2 corrective actions
being performed.
The fewer number of TC+, EC+, non-acute and acute events occurring with the more
effective corrective action assumptions relative to the main assumptions can be viewed as events
that are prevented from occurring as a result of those more effective actions relative to the main
assumptions. Similarly, the increased number of those events occurring with the less effective
corrective actions can be viewed as those that would not be prevented relative to the main
assumptions.
It is important to note that this sensitivity analysis is not intended as a rigorous,
quantitative comparison of the regulatory options, but rather as a general indicator of the
magnitude and direction of change in these outputs relative to inputs. Broadly speaking, since the
sensitivity analysis input assumptions reflect changes that are approximately one half and two
times the main analysis input assumptions, sensitivity analysis outputs that are substantially less
than half or significantly greater than two times the main analysis outputs could be considered to
be very sensitive to the changes in the input assumptions. Sensitivity analysis outputs that are
between one half and twice the main analysis outputs can be considered to be much less sensitive
to changes in the input assumptions. Note, if the ratio of the outputs from the sensitivity analysis
to the main analysis is approximately one, it suggests that the outputs of the sensitivity analysis
are not sensitive to changes in the input assumptions.
As indicated in Exhibit 5.24b, all of the ratios for outputs fall within the 0.5 to 2.0 range,
indicating that these outputs do not appear to be highly sensitive to these alternative assumptions
for corrective action effectiveness.
The outputs that could serve as indicators of the sensitivity of the benefits to these
assumptions are the numbers of TC+ and EC+ results predicted as well as the number of non-
acute and acute events predicted. These outputs have factors that generally fall well within the
0.5 to 2.0 range, indicating that these changes in the inputs assumptions appear to result in
relatively small changes in these outputs. That is, the less effective corrective action assumptions
do not appear to result in missing a disproportionately greater number of these events, and the
more effective corrective actions appear to result in finding a disproportionately greater number
of these events.
The outputs that could serve as indicators of the sensitivity of the costs on these
assumptions are the Level 1 and Level 2 assessments and the Level 1 and Level 2 corrective
actions. The assessments have ratios that, like those for the outputs like TC+ and EC+, are
generally well within the 0.5 to 2.0 range. The ratios for the corrective actions are also within
this range but tend to be close to the 0.5 and 2.0 values. The input assumptions used for the
number of corrective actions done is directly proportional to the number of assessments
Economic Analysis for the Final RTCR
5-49
September 2012
-------
performed, so these changes would be expected to be more directly linear (closer to 0.5 or 2.0).
For example, using the more effective corrective action assumptions, the number of TC+
samples predicted is reduced, as are the number of Level 1 assessments that are related to TC+
events. The ratio for Level 1 assessments under the more effective corrective action assumptions
is 0.84, indicating that under these assumptions systems will only need to perform about 85
percent of the Level 1 assessments over the 25-year period (that is, the others are prevented by
the more effective actions). However, this set of alternative assumptions includes a factor that 20
percent of those doing Level 1 assessments go on to do Level 1 corrective actions (which is
twice the 10 percent assumption in the main model). Therefore, we see that almost twice as many
of these Level 1 corrective actions are performed as under the main model. More specifically, the
ratio for Level 1 corrective actions is 1.65 which is approximately two times the 0.84 ratio noted
for Level 1 assessment predicted.
An additional analysis was run to test the model's sensitivity to the primary analysis
assumption that 10 percent more corrective actions addressing the root cause of a TC/EC event
will be implemented than under the 1989 TCR. The additional sensitivity analysis investigates
the implications of a much greater rate of corrective action implementation—50 percent rather
than 10 percent, or a 400 percent increase from the primary analysis assumption. All other
parameters in this sensitivity test, including the effectiveness and duration of the corrective
actions, are identical to those of the RTCR primary analysis.
Exhibit 5.24b Cumulative Effect of 50 percent Assumption for Corrective Action
Implementation Rate on RTCR Model Results
(Nondisinfecting TNCWS Serving <500 People over 25 Years)
Corrective
Model Output
Actions
RTTC+
Assumption
RTTC+
as rate
RPTC+
RTEC+
RPEC+
GWR EC +
Nonacutes
Acutes
L1 Assmt
L2 Assmt
L1 CA
L2 CA
RTCR1
277,776
3.55%
174,200
9,937
5,038
2,906
90,389
7,170
64,275
36,090
6,448
3,671
50% Option2
229,270
2.97%
110,839
7,887
3,225
2,641
61,424
4,534
53,896
12,506
27,227
6,288
% Change from RTCR
50% Option2
-17.5%
-16.22%
-36.4%
-20.6%
-36.0%
-9.1%
-32.0%
-36.8%
-16.1%
-65.3%
322.3%
71.3%
Relative Change
50% Option2
0.83
0.77
0.64
0.79
0.64
0.91
0.68
0.63
0.84
0.35
4.22
1.71
Notes:
1. "RTCR" refers to the RTCR under assumptions used in the primary analysis, which includes an assumption that 10% more CAs addressing the root cause of a TC/EC event will
be implemented than under the 1989 TCR (0% change in the assumption).
2. "50% option" uses a CAs estimate of 50% instead of the primary analysis assumption of 10% (500% increase in the assumption).
As might be expected, this modification causes rates of Level 1 and Level 2 corrective
actions to increase. This in turn decreases the RTTC+ rate and other levels of occurrence,
including the number of Level 1 and Level 2 assessments. It is interesting to note that the rate of
Level 1 corrective actions increases much more dramatically than the rate of Level 2 corrective
actions. One possible explanation for this is that with the higher corrective action implementation
rate, more systems move to a reduced TC occurrence level during the first years of RTCR.
Thereafter these systems are more likely to trigger Level 1 assessments than Level 2
assessments, causing the ratio of Level 1 corrective actions to Level 2 corrective actions to shift
substantially from that predicted in the primary analysis.
Since this analysis involves changing a parameter by a factor of five, changes in model
output greater than a factor of five would indicate a high degree of sensitivity to the parameter.
Economic Analysis for the Final RTCR
5-50
September 2012
-------
In the table it is evident that no changes on that scale occur. Nonetheless, the analysis does
indicate that an increase in the rate of implementation of corrective actions would reduce the
long-run incidence of TC/EC events and that it would increase the number of Level 1 corrective
actions more dramatically than it would increase the number of Level 2 corrective actions.
Additional Factors Assessed
As described in the introduction to this section, there were assumptions in addition to the
frequency and effectiveness of corrective actions that EPA considered might influence the
analysis. These assumptions were categorized as those that were unlikely to have an impact on
net results (5.22a), and those that affected the RTCR and Alternative option differently than the
1989 TCR (5.22b), thus having the potential to impact net results.
Because the effect of seasonality on occurrence is commonly a concern when using
annual average occurrence rates in an analysis, and because RTTC occurrence rates were found
to be twice as high in the summer and fall as in the winter and spring, EPA ran a sensitivity
analysis on the influence of seasonal trends. This analysis was performed on the subset including
transient nondisinfecting ground water systems serving fewer than 101 people because of its
relatively large number of systems and high occurrence rate. For simplicity, the sensitivity
analysis was limited to only the RTCR.
EPA performed the sensitivity analysis through two runs of the RTCR model. The first
run established a baseline, and the second multiplied the assigned baseline pRTTC values by a
factor of 0.67 during December-May and by a factor of 1.33 during June-November. (No
seasonality factors were applied for systems on annual sampling.)
Summary results in Table 5.25 show expected yearly averages for selected occurrence
metrics during the first 25 years under full implementation of the RTCR. Overall the differences
between the model results are small. Also, they are not consistently higher for one model (e.g.,
when accounting for seasonality, total TC occurrence is slightly lower but the number of acutes
is slightly higher). These variations are likely within the range of Monte Carlo error. Therefore,
EPA determined that not incorporating seasonality into the occurrence model is an acceptable
simplification.
Exhibit 5.25 Effect of Seasonality on Occurrence Analysis Endpoints
RTCR Metric
No Seasonality
Seasonality
Total TC Positives
8,560
8,556
Total Level 1 Triggers
2,824
2,784
Total Acutes1
225
240
L1 Corrective Actions
201
194
L2 Corrective Actions
113
104
Notes:
Estimates represent a yearly average during the first 25 years
underfull implementation of the RTCR.
1. Includes those Level 2 Triggers resulting from an EC+
(does not include those resulting from a second Level 1
trigger occurring within a 12-month rolling period).
Economic Analysis for the Final RTCR
5-51
September 2012
-------
5.3.3.2 Model Validation
There are a limited number of ways in which outputs from the predictive model for the
systems serving <4,100 people can be compared to observed data. EPA has identified two types
of analyses where comparisons can be made. One of these is to compare the model's predictions
of average annual TC and EC occurrence levels for various types and sizes of systems under the
1989 TCR against the observations from the 2005 Six-Year Review data used to parameterize
the model. The other is to compare the model's predictions for the 1989 TCR against those
reported in SDWIS/FED.
It is necessary to keep in mind that the primary purpose of the modeling effort for the
systems serving <4,100 people was not to exactly match observed results but to provide a
framework for comparing relative changes across regulatory options. Nevertheless, it is
important that, to the extent possible, the model outputs for conditions that can be checked
against observed data compare reasonably well. This provides assurance that the model is
operating in a manner that is a reasonable simulation of how the 1989 TCR currently operates
and provides some measure of confidence that the relative changes seen for the regulatory
options are meaningful.
Exhibit 5.26 provides a comparison of the model results for average annual TC+ assays
under the 1989 TCR with those observed in the 2005 Six-Year Review data.
Economic Analysis for the Final RTCR
5-52
September 2012
-------
Exhibit 5.26 Comparison of TC+ Occurrence Predicted as a 25-Year Annual
Average under the 1989 TCR with 2005 Six-Year Review Data
Six Year review
Predictive Model
Population
TC
EC
TC
EC
Served
(% Positive)
(% Positive)
(% Positive)
(% Positive)
Community Water Systems (CWSs) - SW
<100
1.41%
0.09%
1.43%
0.18%
101-500
1.15%
0.05%
0.92%
0.12%
501-1,000
0.70%
0.05%
0.95%
0.13%
1,001-4,100
0.57%
0.03%
0.57%
0.04%
Community Water Systems (CWSs) - GW
<100
2.66%
0.08%
2.23%
0.10%
101-500
1.99%
0.05%
1.69%
0.08%
501-1,000
1.52%
0.04%
1.49%
0.07%
1,001-4,100
1.07%
0.02%
0.99%
0.05%
Nontransient, Noncommunity Water Systems (NTNCWSs) - SW
<100
1.92%
0.24%
1.43%
0.18%
101-500
0.39%
0.09%
1.08%
0.14%
501-1,000
0.64%
0.00%
1.06%
0.14%
1,001-4,100
0.08%
0.00%
1.03%
0.12%
Nontransient, Noncommunity Water Systems (NTNCWSs) - GW
<100
3.17%
0.07%
2.91%
0.14%
101-500
2.70%
0.06%
2.00%
0.10%
501-1,000
1.74%
0.02%
1.86%
0.11%
1,001-4,100
1.35%
0.04%
1.49%
0.09%
Transient, Noncommunity Water Systems (TNCWSs) - SW
<100
2.23%
0.25%
2.15%
0.27%
101-500
2.63%
0.46%
2.18%
0.29%
501-1,000
3.63%
0.00%
2.17%
0.29%
1,001-4,100
0.61%
0.00%
2.04%
0.27%
Transient, Noncommunity Water Systems (TNCWSs) - GW
<100
4.78%
0.20%
4.17%
0.19%
101-500
4.57%
0.19%
4.04%
0.20%
501-1,000
4.30%
0.06%
3.86%
0.20%
1,001-4,100
2.02%
0.03%
2.09%
0.12%
Source: Derived using 2005 Six-Year Review 2 Data. 1989 TCR data from model output.
The comparison of the predictive model output with the Six-Year data for TC and EC
occurrence levels (based on number of TC routine samples taken) shows a reasonable
concordance for all types and sizes of systems. With respect to the differences, the model does
not appear to be either systematically overestimating or underestimating the occurrence levels
compared with the Six-Year Review data (except where the Six-Year Review data show 0
percent observed, the model does show a low level of occurrence).
Exhibit 5.27 provides a comparison of the model results of the average annual non-acute
and acute violations for the 1989 TCR with the same metric from SDWIS data for 2005 3rd
Economic Analysis for the Final RTCR
5-53
September 2012
-------
rd
quarter. (Violations data are downloaded from SDWIS on an annual basis in 3 quarter only, and
are not disaggregated by month.)
Exhibit 5.27 Comparison of SDWIS Data for Non-acute and Acute Violations with
Predictive Model Annual Results for the 1989 TCR
Non-Acute Violations
Acute Violations
GW
SW
GW
SW
System
Predictive
Predictive
Predictive
Predictive
Size
SDWIS
Model
SDWIS
Model
SDWIS
Model
SDWIS
Model
CWSs
<100
905
955
16
31
52
97
3
7
101-500
809
733
50
34
34
81
7
8
501-1,000
203
191
16
19
13
19
3
5
1,001-4,100
338
347
83
49
12
35
6
7
Total
2,255
2,226
165
133
111
231
19
27
NTNCWSs
<100
514
560
7
7
34
60
2
2
101-500
346
265
4
6
20
30
-
1
501-1,000
57
61
2
2
6
7
-
0
1,001-4,100
62
98
1
3
6
12
-
1
Total
979
985
14
18
66
109
2
4
TNCWSs
<100
2,665
4,109
19
81
278
445
5
19
101-500
833
1,371
11
31
76
156
1
8
501-1,000
133
144
4
5
11
16
-
1
1,001-4,100
58
105
2
9
2
13
-
2
Total
3,689
5,730
36
126
367
630
6
31
Grand Total
6,923
8,940
215
277
544
970
27
61
Source: SDWIS 2005 Q3 download and occurrence and predictive model results.
Note: Predictive Model reflects annual average results for model years 16 to 40 without inclusion of the GWR.
The comparison of the predictive model output with the SDWIS violations data also
shows a reasonable concordance for all types and sizes of systems. EPA believes there is some
under-counting of actual violations in SDWIS due to monitoring and reporting violations, which
the model does not incorporate. Therefore, EPA generally expects that the model would generate
results that are higher than the observed SDWIS data. While a comparison reveals that some of
the model estimates are lower than the SDWIS statistics, the grand totals for the model are
consistently higher than those observed in SDWIS.
5.4 Occurrence Analysis for Systems Serving More Than 4,100 People
Systems serving populations greater than 4,100 are similar in many ways. In particular,
these systems generally have the resources and are managed in ways that lead to generally high
expectations for the integrity of the distribution systems and the ability of the systems to identify
and correct problems. Some of these similarities may include:
Economic Analysis for the Final RTCR
5-54
September 2012
-------
• Most of these systems operate with certified operators, are staffed at all operating
times, and usually operate continuously.
• Most of these systems have operators for whom producing water is their primary
activity, and most are owned by communities rather than run as ancillary
businesses.
• As a result of the professional operations, these systems have a much lower
incidence of monitoring and reporting violations.
• Many of these systems have their own laboratory for analyzing samples.
• Many more of these systems disinfect their water.
• These systems take at least five samples per month. The largest of these systems
take hundreds of samples per month. There is no requirement for "additional next
month samples" as there is for systems serving <4,100 people because systems
serving >4,100 people already take at least five samples each month.
• The monitoring frequency changes under the RTCR mostly do not affect systems
serving more than 4,100 people; all systems take at least five samples per month.
For systems serving more than 4,100 people, EPA assumes that occurrence may change
based on the extra distribution system awareness created by: a) applying the Level 1 and Level 2
assessments in lieu of prior assessments that may, in some cases, have been less structured; and
b) reporting the assessment results. Therefore, EPA developed a simple model to predict the
effects of the RTCR and Alternative option on PWSs and to compare those data to the baseline
data predicted for the 1989 TCR. EPA did not quantify changes in violation or trigger rates for
systems serving more than 4,100 people among the 1989 TCR, the RTCR, and Alternative
option because of: (1) limited Six-Year Review data to characterize these systems; (2) the
essentially unchanged monitoring requirements across options for these systems; and (3) the
level of effort already occurring to implement the 1989 TCR.
5.4.1 Model (for Systems Serving >4,100 People)
As input for this model, EPA first considered 2005 data for the 1989 TCR on sampling
and positive assays that were compiled under the Six-Year Review, as was incorporated into the
model for systems serving <4,100 people. However, as discussed in Section 4.2.2.3, detailed
sampling results from the Six Year Review data were not representative of the universe of
systems serving >4,100 people. EPA instead found that violations data were adequately
representative of the universe of systems serving >4,100 people. This input to the model, based
on 2005 SDWIS data (see Exhibit 4.10), was used to predict violations that systems serving more
than 4,100 people will incur in the 25-year period of RTCR analysis, and the behavioral
responses (Level 1 and Level 2 assessments and corrective actions) that would result.
The model uses the same stratification as the model for systems serving <4,100 people—
by system type (CWS, TNCWS, or NTNCWS) and by water source (SW or GW). For each
system in a given category the model applies the 2005 SDWIS violations rate for that category.
Economic Analysis for the Final RTCR
5-55
September 2012
-------
This rate is applied in each year of the analysis, that is, the violations rate is a constant in the
model. For estimating the effects on violations under the RTCR and Alternative option, the
model estimates that 10 percent of the assessments resulting from violations will result in
corrective actions, the same value applied in the analysis of the systems serving <4,100 people.
This 10 percent represents the net increase in efficacy of addressing the root cause of system
contamination under the RTCR and Alternative option compared to the 1989 TCR.
Repeat samples are reduced in the model under the RTCR and Alternative option from
the 1989 TCR. They are calculated by applying the ratios of repeat samples to routine samples
from systems serving 1,001-4,100 people to the systems serving more than 4,100 people.
The model output, shown in Section 5.4.2, includes predictions of the annual number of
violations (non-acute = Level 1, acute = Level 2), and the number of Level 1 and Level 2
assessments and corrective actions to be implemented.
5.4.2 Model Results (for Systems Serving >4,100 People)
Exhibits 5.28 through 5.30 present model results for systems serving more than 4,100
people by size category.
5.4.3 Model Uncertainty (for Systems Serving >4,100 People)
As explained in Section 5.4.1 of this chapter, EPA does not expect systems serving
>4,100 people to experience changes in routine monitoring or repeat sample regimes under the
regulatory options; furthermore, all systems in this size range are on monthly sampling.
Therefore, the model for systems serving >4,100 people was relatively simple compared to that
for the systems serving <4,100 people, and EPA did not develop an uncertainty analysis for this
group of systems.
Economic Analysis for the Final RTCR
5-56
September 2012
-------
Exhibit 5.28 Results for Systems Serving more than 4,100 People—1989 TCR
FWS Size
(Population
Served)
Number of
Routine
Monitoring
Samples
Num ber of
Additional
Routine
Monitoring
Sam pies
Nu m be r of
Repeat
Monitoring
Sam pies
PWSs
Performing
Additional
Annual Site
Inspections
Num ber of
Non-Acute
Violation
Assessments
(Single
Violations)
Number of
Corrective
Actions
(based on
Single Non-
Acute
Violation
Assessment)
Number of
Acute
Violation
Assessments
Num ber of
Non-Acute
Violation
Assessments
(Multiple
Violations)
Num ber of
Corrective
Actions
(based on
Acute and
Multiple Non-
Acute Violation
Assessments)
A
B
C
D
E
F
G
H
I
Com m unity Water System s (CWSs) - SW
4,101-33,000
10,636,296
186,729
2,152
197
33,001-96,000
11,058,960
194,149
534
56
96,001-500,000
10,190,400
178,901
233
24
500,001-1 Million
2,019,600
35,456
22
> 1 Million
1,686,960
29,616
Com m unity Water Systems (CWSs) - GW
4,101-33,000
9,145,224
230,201
4,545
263
33,001-96,000
4,884,000
122,938
656
53
96,001-500,000
1,945,680
48,976
129
10
500,001-1 Million
253,440
6,380
> 1 Million
269,280
6,778
Nontransient Noncom m unity Water Systems (NTNCWSs) - SW
4,101-33,000
50,424
1,628
5
33,001-96,000
34,320
1,108
96,001-500,000
31,680
1,023
500,001-1 Million
> 1 Million
Nontransient Noncom m unity Water Systems (NTNCWSs) - GW
4,101-33,000
153,648
5,936
123
9
33,001-96,000
23,760
918
4
96,001-500,000
500,001-1 Million
> 1 Million
Transient Noncom m unity Water Systems (TNCWSs) - SW
4,101-33,000
40,656
8
33,001-96,000
96,001-500,000
500,001-1 Million
> 1 Million
102,960
Transient Noncom m unity Water Systems (TNCWSs) - GW
4,101-33,000
156,288
8,909
116
4
33,001-96,000
34,320
1,956
96,001-500,000
26,400
1,505
500,001-1 Million
63,360
3,612
> 1 Million
Source: Appendix A, Exhibit A. 1 .z.
Note: For modeling purposes, EPA estimated only the net change in the number of corrective actions performed under the RTCR and Alternative option
compared to the 1989 TCR. Because only the net change in the number of corrective actions is estimated, no additional corrective actions are modeled for the
1989 TCR (it is assumed that PWSs are already performing some corrective actions under the 1989 TCR).
Economic Analysis for the Final RTCR
5-57
September 2012
-------
Exhibit 5.29 Results for Systems Serving more than 4,100 People—RTCR
PWS Size
(Population
Served)
Num ber of
Routine
Monitoring
Sam pies
Number of
Additional
Routine
Monitoring
Sam pies
Num ber of
Repeat
Monitoring
Sam pies
PWSs
Perform ing
Additional
Annual Site
Inspections
Num ber of
Level 1
Assessm ents
Number of
Corrective
Actions (based
on Level 1
Assessments)
Number of
Level 2
Assessments
(based on
Acute
Violations)
Num ber of
Level 2
Assessments
(based on Non>
Acute
Violations)
Number of
Corrective
Actions (based
on Level 2
Assessments)
A
B
C
D
E
F
G
H
I
Com m unity Water Systems (CWSs) - SW
4,101-33,000
10,636,296
-
181,661
.
2,152
215
197
.
20
33,001-96,000
11,058,960
-
188,880
-
534
53
56
-
6
96,001-500,000
10,190,400
-
174,046
.
233
23
24
.
2
500,001-1 Million
2,019,600
-
34,493
.
22
2
-
.
-
> 1 Million
1,686,960
-
28,812
-
-
-
-
-
-
Com m unity Water Systems (CWSs) - GW
4,101-33,000
9,145,224
-
217,321
.
4,545
454
263
.
26
33,001-96,000
4,884,000
-
116,060
-
656
66
53
-
5
96,001-500,000
1,945,680
-
46,236
.
129
13
10
.
1
500,001-1 Million
253,440
-
6,023
.
-
-
-
.
-
> 1 Million
269,280
-
6,399
-
-
-
-
-
-
Nontransient Noncom m unity Water Systems (NTNCWSs) - SW
4,101-33,000
50,424
-
1,448
.
5
0
-
.
-
33,001-96,000
34,320
-
985
-
-
-
-
-
-
96,001-500,000
31,680
-
910
.
-
-
-
.
-
500,001-1 Million
-
-
-
.
-
-
-
.
-
> 1 Million
-
-
-
-
-
-
-
-
-
Nontransient Noncom m unity Water Systems (NTNCWSs) - GW
4,101-33,000
153,648
-
5,157
.
123
12
9
.
1
33,001-96,000
23,760
-
797
.
4
0
-
.
-
96,001-500,000
-
-
-
.
-
-
-
.
-
500,001-1 Million
-
-
-
.
-
-
-
.
-
> 1 Million
-
-
-
-
-
-
-
-
-
Transient Noncom m unity Water Systems (TNCWSs) - SW
4,101-33,000
40,656
-
2,225
.
8
1
-
.
-
33,001-96,000
-
-
-
.
-
-
-
.
-
96,001-500,000
-
-
-
.
-
-
-
.
-
500,001-1 Million
-
-
-
.
-
-
-
.
-
> 1 Million
102,960
-
5,636
-
-
-
-
-
-
Transient Noncom m unity Water Systems (TNCWSs) - GW
4,101-33,000
156,288
-
7,188
.
116
12
4
.
0
33,001-96,000
34,320
-
1,578
.
-
-
-
.
-
96,001-500,000
26,400
-
1,214
.
-
-
-
.
-
500,001-1 Million
63,360
-
2,914
.
-
-
-
.
-
> 1 Million
-
-
-
-
-
-
-
-
-
Source: Appendix A, Exhibit A2.z.
Note: Estimates of the number of assessments and corrective actions are net increases in activity predicted to occur under the RTCR relative to the 1989 TCR;
estimates of zero reflect that no additional such activity occurs as compared to baseline (the 1989 TCR).
Economic Analysis for the Final RTCR
5-58
September 2012
-------
Exhibit 5.30 Results for Systems Serving more than 4,100 People—Alternative
Option
PWS Size
(Population
Served)
Number of
Routine
Monitoring
Sam pies
Num ber of
Additional
Routine
Monitoring
Sam pies
Num ber of
Re pe at
Monitoring
Sam pies
PWSs
Perform ing
Additional
Annual Site
Inspections
Num ber of
Level 1
Assessm ents
Number of
Corrective
Actions (based
on Level 1
Assessm ents)
Num ber of
Level 2
Assessments
(based on
Acute
Violations)
Num ber of
Level 2
Assessm ents
(based on Non>
Acute
Violations)
Num ber of
Corrective
Actions (based
on Level 2
Assessments)
A
B
C
D
E
F
G
H
I
Com m unity Water System s (CWSs) - SW
4,101-33,000
10,636,296
-
181,661
.
2,152
215
197
.
20
33,001-96,000
11,058,960
-
188,880
.
534
53
56
.
6
96,001-500,000
10,190,400
-
174,046
-
233
23
24
-
2
500,001-1 Million
2,019,600
-
34,493
-
22
2
-
-
-
> 1 Million
1,686,960
-
28,812
-
-
-
-
-
-
Com m unity Water System s (CWSs) - GW
4,101-33,000
9,145,224
-
217,321
.
4,545
454
263
.
26
33,001-96,000
4,884,000
-
116,060
.
656
66
53
.
5
96,001-500,000
1,945,680
-
46,236
.
129
13
10
.
1
500,001-1 Million
253,440
-
6,023
-
-
-
-
-
-
> 1 Million
269,280
-
6,399
-
-
-
-
-
-
Nontransient Noncom m unity Water Systems (NTNCWSs) - SW
4,101-33,000
50,424
-
1,448
-
5
0
-
-
-
33,001-96,000
34,320
-
985
.
-
-
-
.
-
96,001-500,000
31,680
-
910
.
-
-
-
.
-
500,001-1 Million
-
-
-
.
-
-
-
.
-
> 1 Million
-
-
-
-
-
-
-
-
-
Nontransient Noncom m unity Water Systems (NTNCWSs) - GW
4,101-33,000
153,648
-
5,157
-
123
12
9
-
1
33,001-96,000
23,760
-
797
-
4
0
-
-
-
96,001-500,000
-
-
-
-
-
-
-
-
-
500,001-1 Million
-
-
-
.
-
-
-
.
-
> 1 Million
-
-
-
-
-
-
-
-
-
Transient Noncom m unity Water Systems (TNCWSs) - SW
4,101-33,000
40,656
-
2,225
.
8
1
-
.
-
33,001-96,000
-
-
-
.
-
-
-
.
-
96,001-500,000
-
-
-
-
-
-
-
-
-
500,001-1 Million
-
-
-
-
-
-
-
-
-
> 1 Million
102,960
-
5,636
-
-
-
-
-
-
Transient Noncom m unity Water Systems (TNCWSs) - GW
4,101-33,000
156,288
-
7,188
.
116
12
4
.
0
33,001-96,000
34,320
-
1,578
.
-
-
-
.
-
96,001-500,000
26,400
-
1,214
.
-
-
-
.
-
500,001-1 Million
63,360
-
2,914
-
-
-
-
-
-
> 1 Million
-
-
-
-
-
-
-
-
-
Source: AppendixA, Exhibit A3.z.
Note: Estimates of the number of assessments and corrective actions are net increases in activity predicted to occur under the Alternative option relative to the
1989 TCR; estimates of zero reflect that no additional such activity occurs as compared to baseline (the 1989 TCR).
5.5 Summary of Key Drivers for Benefit and Cost Analyses Output from the Predictive
Model
As presented in Exhibits 5.10-5.15, the model for systems serving <4,100 people
produces estimates of the number of samples, the number of Level 1 and Level 2 assessments,
and the number of corrective actions implemented as a result of Level 1 and Level 2 assessments
under each regulatory option. For the RTCR and Alternative option, the model calculates the
corrective actions performed as a net increase from those performed under the 1989 TCR. Those
corrective actions performed under the 1989 TCR are described in Chapter 7 in the discussion of
1989 TCR costs. The net change in each of the model outputs, in particular sampling regimens
and implementation of assessments and corrective actions relative to the 1989 TCR and under
the RTCR and Alternative option provides the basis for estimating the benefits in the rule,
Economic Analysis for the Final RTCR
5-59
September 2012
-------
described in Chapter 6 of this EA. Similarly, the cost model applies unit costs to the activities
described by these outputs to produce net costs for each regulatory option considered, as
described in Chapter 7.
As described in the uncertainty discussion and sensitivity analysis presented in Section
5.3.3.1, these key drivers represent the most significant source of uncertainty in the analysis.
However, the sensitivity analysis showed that a doubling in the implementation and effectiveness
assumptions for corrective actions (the proportional increase in the number of effective
corrective actions implemented and the duration and extent of reduced occurrence) implemented
under the RTCR would result in only approximately a 25 percent decrease in acute events.
Conversely, a 50 percent reduction in corrective action implementation and effectiveness was
estimated to induce less than a 28 percent increase in acute events. These results are expected to
apply similarly to all categories of systems. Based on these results, EPA concludes that changing
the assumptions within the likely range (a doubling or halving of current assumptions) would not
have a significant effect on the conclusions drawn from this EA. Similarly, varying the
assumptions by size and type of system within a likely range would not be expected to have a
significant effect on the results.
For systems serving >4,100 people, the net change in costs is estimated for the additional
reporting requirements under the rule, as well as the additional corrective actions implemented
for these systems. However, as described in Section 5.4, the sampling regimens are not expected
to change for these systems, and they likely would experience minimal changes in risk and cost,
as described in Chapter 6 and Chapter 7 of this EA, respectively.
Economic Analysis for the Final RTCR
5-60
September 2012
-------
6 Benefits Analysis
6.1 Introduction
This chapter considers the overall change in risk of contamination to public water
systems (PWSs), as indicated by the presence of total coliforms (TC) and E. coli, associated with
compliance with the 1989 Total Coliform Rule (1989 TCR) and with the Revised Total Coliform
Rule (RTCR) and the Alternative option. Since E. coli is an indicator of fecal contamination
(Edberg et al., 2000), EPA assumes that a decrease in E. coli occurrence in the distribution
system would be associated with a decrease in fecal contamination in the distribution system. In
general, this decrease in fecal contamination should reduce the potential risk to human health for
PWS customers. Thus, any reduction in E. coli occurrence is considered a benefit of the RTCR.
Since fecal contamination may contain waterborne pathogens including bacteria, viruses, and
parasitic protozoa, in general, a reduction in fecal contamination should reduce the risk from all
of these contaminants.
Based on limitations in available data as described further in Section 6.4 of this chapter,
EPA determined that benefits could not be calculated in terms of avoided costs or other
quantified benefits related to avoided morbidity or mortality. Therefore, this economic analysis
(EA) focuses on a qualitative analysis of risk, which is supported by quantitative analysis of
predicted outcomes for each regulatory option. The quantitative discussion focuses on net
changes in occurrence of contaminant indicators under both regulatory options (RTCR and
Alternative option) considered as compared to the 1989 TCR option. The qualitative analysis
considers the direction of anticipated changes in risk related to changes in sampling and
corrective action regimens under each regulatory option (Section 6.2). EPA considered the
results of the qualitative and quantified assessments to determine how the 1989 TCR compares to
the RTCR and Alternative option in terms of overall change in risk to the population served by
PWSs across the United States. The remainder of this chapter is organized into four sections, as
follows:
• Section 6.2 presents qualitative benefits analyses.
• Section 6.3 presents an assessment of the predictive analysis results.
• Section 6.4 presents uncertainty and sensitivity analyses.
• Section 6.5 discusses other potential benefits.
6.2 Qualitative Benefits Analyses
When revising an existing drinking water regulation, one of the main concerns is to
ensure that backsliding on water quality and public health protection does not occur. The Safe
Drinking Water Act (SDWA) states that any revision to a regulation must "maintain, or provide
for greater protection for the health of persons." Risk reduction for the RTCR is characterized by
the activities performed that are presumed to reduce risk of exposing the public to contaminated
water. These activities are considered under each rule component discussed in Sections 6.2.1-
6.2.8 and summarized in Figure 6.1 below. The discussion in Section 6.2.9 considers how
Economic Analysis for the Final RTCR
6-1
September 2012
-------
expected overall increases in the risk for some rule components are offset by expected overall
decreases resulting from other components, and provides a net assessment of the direction in
change of risk for the regulatory options considered.
The qualitative analysis uses the best professional judgment of EPA as informed by Total
38
Coliform Rule Distribution System Advisory Committee (TCRDSAC) deliberations, as well as
certain quantitative estimates, to predict the directional change in risk for each rule component of
the RTCR and the Alternative option. Quantitative estimates considered include the changes in
TC occurrence and counts of systems conducting assessments and implementing corrective
actions shown in Exhibits 6.2-6.6. Exhibit 6.1 presents a summary of this evaluation for the
RTCR and Alternative option as compared to the 1989 TCR. The qualitative analysis discusses
the influence of the individual rule components under each regulatory option considered on the
occurrence of TC-positive (TC+) and E. co/z'-positive (EC+) samples. Since a dose-response
relationship between exposure to the fecal indicator E. coli and adverse health effects from
waterborne pathogens that can be present with fecal contamination is not available, the resulting
risk to human health is discussed only in terms of the anticipated change in direction of risk.
6.2.1 Implementation Activities
Rule implementation activities are expected to be similar under the RTCR and
Alternative option. These activities are primarily administrative in nature and include items such
as reading and understanding the rule, training, and development of reporting and recordkeeping
protocols. Because of the similarities in expected implementation activities under the regulatory
options, they are not expected to have an observably different effect on overall risk relative to the
1989 TCR. Both PWSs and states would incur additional burden and costs for transitioning to
operations under the RTCR or Alternative option requirements. Because the activities undertaken
to make the transition are primarily administrative, the additional activities are not expected to
have any direct impact on risk.
6.2.2 Routine Monitoring
EPA expects that more frequent monitoring would decrease the risk of contamination in
PWSs based on an enhanced ability to diagnose and mitigate system issues in a more timely
fashion. Conversely, EPA assumes that a less frequent monitoring schedule would result in
increased risk. Real-time continuous sampling would mitigate the most risk possible based on
sampling schedule; however, it would cost prohibitively more than the periodic sampling
practiced under the 1989 TCR and included in the RTCR and the Alternative option. EPA's
objective in proposing the sampling schedules included in the RTCR and Alternative option was
to find an appropriate balance between the factors of risk mitigation and cost management.
1989 TCR
The 1989 TCR requires a large portion (approximately 40 percent) of water systems to
perform monthly monitoring for TC. The remaining systems are eligible for reduced monitoring,
as follows.
38 TCRDSAC deliberations are described in Chapter 3 of this EA.
Economic Analysis for the Final RTCR
6-2
September 2012
-------
• GW noncommunity water systems (GW NCWSs) serving <1,000 people are
required to monitor quarterly. States may allow annual monitoring if a sanitary
survey conducted in the past 5 years shows the system is supplied solely by
protected GW sources and is free of sanitary defects.
• GW community water systems (GW CWSs) serving <1,000 people may monitor
quarterly if they have no history of TC contamination in their current
configurations, and if a sanitary survey conducted in the past 5 years shows that the
system is supplied solely by a protected GW source and has no sanitary defects.
RTCR
The eligibility requirements for GW systems to qualify for reduced monitoring under the
RTCR are more stringent than under the 1989 TCR, leading to fewer PWSs qualifying for
reduced monitoring and therefore a higher total number of routine samples being taken over the
25-year period of analysis. The primacy agency has the discretion to reduce monitoring
frequency for well-operated GW NCWSs and GW CWSs serving <1,000 people. Under the
RTCR, eligibility for reduced monitoring for NCWSs and CWSs, respectively, is provided only
for the following types of PWSs, as discussed as follows.
NCWSs
To be eligible to qualify for and remain on annual monitoring after the compliance
effective date, GW NCWSs serving <1,000 people must meet each of the following criteria:
• The most recent sanitary survey shows the system is free of sanitary defects and has
a protected water source and meets approved construction standards;
• The system must have a clean (1989 TCR or RTCR) compliance history (no MCL
violations, Level 1 triggers, Level 2 triggers, treatment technique violations or
monitoring violations) for a minimum of 12 months;
• An annual site visit (recurring) by the primacy agency within the last 12 months and
correction of all identified sanitary defects. A voluntary Level 2 assessment by a
party approved by the primacy agency may be substituted for the primacy agency
annual site visit; and
• The primacy agency should encourage additional enhancements to the barriers
protecting the distribution system from contamination. These measures could
include but are not limited to the following:
- Cross connection control, as approved by the primacy agency;
- An operator certified by an appropriate primacy agency certification
program; regular visits by a circuit rider may substitute for this
requirement for some PW S types as permitted by the state;
Economic Analysis for the Final RTCR
6-3
September 2012
-------
- Continuous disinfection entering the distribution system and a residual in
the distribution system in accordance with criteria specified by the
primacy agency; and
- Maintenance of at least a 4-log inactivation of viruses each day of the
month based on daily monitoring as specified in the Ground Water Rule
(GWR) (with allowance for a 4-hour exception).
- Other equivalent enhancements to water system barriers as approved by
the primacy agency.
CWSs
To be eligible to change from monthly to quarterly reduced monitoring after the
compliance date, GW CWSs serving <1,000 people must be in compliance with any state-
certified operator provisions and meet each of the following criteria:
• The most recent sanitary survey shows the system is free of sanitary defects (or has
an approved plan and schedule to correct them), has a protected water source, and
meets approved construction standards;
• The system must have a clean (1989 TCR or RTCR) compliance history (no MCL
violations, Level 1 or Level 2 triggers, treatment technique violations or monitoring
violations) for a minimum of 12 months; and
• Meet at least one of the following criteria:
- An annual site visit by the primacy agency or a voluntary Level 2
assessment by a party approved by the primacy agency and correction of
all identified sanitary defects (or an approved plan and schedule to correct
them); or
- A cross connection control program, as approved by the primacy agency;
or
- Continuous disinfection entering the distribution system and a residual in
the distribution system in accordance with criteria specified by the
primacy agency; or
- At least 4-log inactivation of viruses each day of the month based on daily
monitoring as specified in the GWR (with allowance for a 4-hour
exception); or
- Other equivalent enhancements to water systems as approved by the
primacy agency.
Based on the additional protection provided by the more stringent criteria to qualify for
reduced monitoring, the RTCR is expected to reduce risk for GW systems that qualify for
Economic Analysis for the Final RTCR
6-4
September 2012
-------
reduced monitoring compared to the eligibility requirements under the 1989 TCR. In addition,
those systems that no longer qualify for reduced monitoring are expected to have a reduced risk
as a function of the increased numbers of samples taken.
Seasonal systems (NCWSs that are not operated as a PWS on a year-round basis and start
up and shut down the system at the beginning and end of each operating season) that sample on a
reduced schedule under the 1989 TCR may retain their current schedules unless a state
determines that more frequent sampling is appropriate. For seasonal systems required to move to
an increased monitoring schedule, risk would decrease as a function of the increase in
monitoring. In addition, more explicit procedural requirements related to monitoring under the
RTCR (i.e., samples must be taken during the period of highest vulnerability or peak usage)
would also be expected to reduce risk for seasonal systems.
Alternative option
Under the Alternative option, all systems must initially collect TC samples monthly
regardless of size or type (i.e., water source), which is more frequent for those systems on
reduced monitoring schedules under the 1989 TCR or RTCR. Over time, in the second year of
rule implementation and beyond, the Alternative option would allow some GW systems to
reduce to quarterly monitoring if the systems meet the same qualifications required for quarterly
monitoring under the RTCR. Reduced monitoring on an annual schedule is not allowed under the
Alternative option, creating a further increase in samples taken over the 1989 TCR and RTCR.
Overall, the more frequent monitoring requirements (i.e., all PWSs monitor monthly in the first
few years after promulgation and no annual reduced monitoring is allowed) would reduce risk as
compared to the 1989 TCR to a greater extent than under the RTCR.
6.2.3 Repeat Monitoring
Under the 1989 TCR, PWSs serving <1,000 people take four repeat samples at and within
five service connections upstream and downstream of the initial TC+ occurrence event over the
course of 24 hours following the event. Three repeat samples are required for PWSs serving
>1,000 people, including one sample at the site of the initial TC+ and two additional samples
within five service connections up or downstream of that site.
Under the RTCR and Alternative option, PWSs serving <1,000 people are only required
to take three repeat samples. The reduction in the number of repeat samples required could
increase risk for any individual sampling event. However, the effect of the reduced sampling is
expected to be minor. Analysis of the data (see Appendix H for the complete analysis) shows
that for all PWSs serving 1,000 people or fewer, two or more of the repeat samples are positive
in 75 percent of those instances in which there are any positive repeat samples. For those 75
percent of instances, reducing the number of repeat samples from four to three would have no
effect on the number of systems that would be triggered to conduct an assessment of the system
under the RTCR. In these cases, at least one of the remaining repeat samples would still be TC+,
and only one positive repeat sample is required to trigger an assessment at these PWSs.
For 25 percent of the cases in which repeat samples are TC+ under the 1989 TCR, only
one sample will be positive. EPA estimates that if the number of repeat samples were reduced
Economic Analysis for the Final RTCR
6-5
September 2012
-------
from four to three, any positive samples occurring at three of the four original locations (75
percent) would still be detected. Thus, if the number of required repeat samples were reduced
from four to three, the number of situations triggering an assessment for the system would be 94
percent (0.75 + (0.75*0.25) = 0.94) of the number of assessments triggered by taking four repeat
samples.
Although dropping the required number of repeat samples from four to three means that
some fraction of triggered assessments may be missed, representing an increase in risk, EPA
believes that the other provisions of the RTCR as described in this chapter compensate for that
change and that, taken as a whole, the provisions of the RTCR provide for greater protection of
public health.
6.2.4 Additional Routine Monitoring
Under the 1989 TCR, PWSs serving <4,100 people must conduct additional routine
monitoring in the month following a TC+ sample. All systems must collect and test a minimum
of five samples in the month following the TC+ sample, unless the primacy agency finds that
additional sampling is unnecessary or the primacy agency determines the cause of the TC+
sample and establishes that the system has corrected or would correct the problem.
Under the RTCR, EPA will retain the requirement of taking additional routine samples
the month following a TC+ sample for systems on quarterly or annual monitoring. Under the
RTCR, a system that has a Level 1 trigger must conduct an assessment, and if a problem is
found, the system must take corrective action. Under such circumstances, the advisory
committee believed that additional samples collected the following month are appropriate to help
recognize the problem if it still persists. Without the provision of additional monitoring, systems
on annual or quarterly monitoring would not take any samples the following month. Systems
having a Level 2 assessment are triggered into a monthly monitoring and therefore have less
need for additional routine monitoring to indicate if a problem persists.
For systems required to take the additional routine samples the following month (i.e.,
systems on quarterly or annual monitoring), the RTCR changes the requirement from taking a
total of five routine samples to a requirement of just three routine samples. The advisory
committee recognized that it is appropriate to drop from five to three samples the following
month to reduce monitoring costs while still maintaining a substantial likelihood of identifying a
problem if a problem persists. EPA recognizes that a reduction in the number of samples taken
could also mean a reduction in the number of positive samples found. However, the reduction in
the number of additional routine samples in conjunction with the new assessment and corrective
action provisions of the RTCR (discussed in sections III.E.2 and III.E.3 of the RTCR preamble
(USEPA 2010c)) leads to a rule that is ultimately more protective of public health (i.e., more E.
coli MCL violations being prevented) and improvement in water quality (i.e., decrease in the
TC+ and EC+ hit rates observed as shown by EA occurrence modeling results).
For systems taking at least one sample monthly, the advisory committee recommended
no additional routine samples for the following reason. Taking no additional routine samples the
following month substantially reduces monitoring costs. The assessment and corrective action
provisions will give systems the ability to identify and prevent the occurrence of problems. EA
Economic Analysis for the Final RTCR
6-6
September 2012
-------
modeling results show that although there is a decrease in the number of E. coli MCL violations
found with the decrease in the number of additional routine samples taken (i.e., going from five
samples to one during the month following a TC+ sample), the assessment and corrective action
provisions leads to more E. coli MCL violations being prevented compared to the 1989 TCR (see
Exhibit 6.7 for more details).
In addition, whenever a TC+ sample occurs during routine sampling, there is also a
requirement to conduct repeat sampling to clarify if potential pathways to contamination persist.
For systems serving 1,000 people or fewer, if a repeat sample is TC+, at least a Level 1
assessment will be triggered. If a sanitary defect is found, the system is required to correct the
sanitary defect. The absence of any repeat positive sample provides some indication that the
problem is not persisting. For systems on monthly monitoring, these two conditions mitigate the
need for additional routine sampling for the following month.
Although the changes to the additional routine monitoring provisions mean that some
fraction of triggers may be missed representing an increase in risk, EPA believes that the other
provisions of the RTCR as described in this chapter compensate for that change and that, taken
as a whole, the provisions of the RTCR provide for greater protection of public health.
6.2.5 Annual Site Inspections
The 1989 TCR does not include any requirements for annual site inspections. However,
based on discussions with stakeholders, some states do perform annual site visits for any systems
on an annual sampling schedule.
Under the RTCR, GW NCWSs serving <1,000 people must, within one year of the
compliance effective date, have an initial (and annually thereafter) visit by the state or an annual
voluntary Level 2 assessment by a party approved by the state to remain on an annual monitoring
schedule. Because of the cost differential between conducting annual site inspections and the
alternative (quarterly monitoring), EPA has estimated that only those states that already
voluntarily conduct annual site inspections under the 1989 TCR would do so under the RTCR.
Therefore, no risk reduction is expected for these systems (and thus overall) from this rule
component under the RTCR.
The Alternative option does not allow systems to reduce to a frequency of annual
monitoring and therefore does not include an annual site inspection requirement. However, based
on discussions with stakeholders, those states that currently conduct annual site assessments
under the 1989 TCR may no longer have the resources to continue the inspections and conduct
quarterly monitoring under the Alternative option. For NCWSs on annual monitoring, the
TCRDSAC believed that requiring a system to have an annual site visit or a Level 2 assessment
provides at least an equivalent level of diagnosis of problems and vulnerabilities that might exist
as compared to quarterly monitoring without an annual site visit. This tradeoff between annual
monitoring with site inspections and three additional routine samples (i.e., quarterly monitoring
with no site inspections) would potentially result in an increased risk for the Alternative option
compared to the RTCR.
Economic Analysis for the Final RTCR
6-7
September 2012
-------
6.2.6 Assessments
Under the 1989 TCR there is no explicit "assessment" required when TC+ results are
observed. However, systems are required to notify the public and state under different scenarios.
Specifically, aPWS must:
• Report any acute MCL violation to the state no later than the end of the business
after the system learns of a violation;
• Notify the public within 24 hours of an acute MCL violation.
• Notify the public within 30 days of a monthly/non-acute MCL violation;
39
Under the RTCR and Alternative option, PWSs are required to perform and submit a
Level 1 assessment if:
• Systems taking 40 or more samples per month have more than 5.0 percent TC+
samples;
• Systems taking less than 40 samples per month have two or more TC+ samples in
one month; or
• A system fails to take all required repeat samples after a single TC+ sample
A more detailed examination of the system, including its monitoring and operational
practices (a Level 2 assessment), is required if a system has:
• An E. coli MCL violation;
• A second Level 1 treatment technique trigger within a rolling 12-month period,
unless the first Level 1 treatment technique trigger was based on exceeding the
allowable number of TC+ samples, the State has determined a likely reason for the
TC+ samples that caused the initial Level 1 treatment technique trigger, and the
State establishes that the system has fully corrected the problem; or
• For PWSs with approved reduced annual monitoring, the system has a Level 1
treatment technique trigger in two consecutive years.
Mandatory assessments are a new requirement under the RTCR and Alternative option,
and also represent an increased focus on problem solving from the less defined investigations (if
any) conducted under the 1989 TCR. Because of the more explicit requirements of the
assessments, it is expected that more problems would be identified and resolved. As a result, the
risk relative to the 1989 TCR is assumed to decrease.
39 Requirements for the 1989 TCR, RTCR, and Alternative option are described in detail in Chapter 3 of this EA.
Economic Analysis for the Final RTCR
6-8
September 2012
-------
6.2.7 Corrective Actions
Corrective actions are not explicitly required under the 1989 TCR. However, responsible,
well-operated systems do perform corrective actions based on investigations they perform (see
Section 6.2.6) in response to positive samples.
Under the RTCR and Alternative option, PWSs must correct any sanitary defects found
during either a Level 1 or Level 2 assessment. For corrections not completed by the time of
submission of the assessment form, systems must complete the corrective action(s) on a schedule
determined by the state in consultation with the system. Systems are required to notify the state
when they have completed each corrective action. Failure to implement a corrective action is
considered a treatment technique violation, subject to public notification (PN).
EPA does not have data on the existing rates at which corrective actions are completed
under the 1989 TCR, so an assumption of the net change in the percentage of assessments
resulting in corrective action (10 percent)40 is made for the RTCR and Alternative option as part
of this EA and subsequent evaluation of changes in risk. Overall, increased protection provided
by this net increase in corrective action as a result of requiring systems to implement a correction
action through an enforceable mechanism would reduce risk under both the RTCR and
Alternative option.
6.2.8 Public Notification
Monthly/Non-acute MCL Violations
The 1989 TCR requires PN within 30 days of a monthly/non-acute MCL violation or 24
hours of an acute MCL violation.
Both the RTCR and Alternative option would require:
• Tier 1 PN within 24 hours of an E. coli MCL violation;
• Tier 2 PN within 30 days of a treatment technique violation; and;
• Tier 3 PN within a year following either a routine monitoring violation or a
reporting violation.
However, under both the RTCR and Alternative option, the PN requirements for
monthly/non-acute MCL violations would no longer be required (the RTCR and Alternative
options do not include monthly/non-acute MCL violations). Since monthly/non-acute violations
account for a large number of violations under the 1989 TCR, there is expected to be a large
decrease in the number of notices presented to the public. If it is assumed that such notices
provide information that aid in risk avoidance by water customers, risk may increase as a result
of reducing this PN requirement to the extent that a risk exists. Because PWSs are no longer
40 The 10 percent assumption is based on EPA discussions with stakeholders regarding experiences with
implementing the 1989 TCR and the expected impact of RTCR. A sensitivity analysis evaluating alternative
assumptions was conducted. Results of these analyses are discussed in Chapter 5 (section 5.3.3.1).
Economic Analysis for the Final RTCR
6-9
September 2012
-------
required to send out notices for monthly/non-acute MCL violations or the monitoring results41
that triggered them under the 1989 TCR, there is also the potential that some PWSs may become
less responsive in addressing TC hits or preventing them from occurring in the first place.
The TCRDSAC evaluation of the 1989 TCR PN also concluded that the numbers of
monthly/non-acute MCL violation notices that go out to the public are confusing. Unlike acute
MCL violations associated with positive E. coli samples, monthly/non-acute MCL violations
may have no connection to a direct health risk because TC+ samples do not indicate a direct
threat to public health. To the extent that a high number of notices issued for monthly/non-acute
MCL violations result in false alarms, consumers may not give appropriate attention to a notice
of an acute MCL violation that indicates a greater potential risk to public health. Therefore, the
EPA concluded, using best professional judgment informed by the TCRDSAC's evaluation that
risk may decrease through the elimination of the PN requirement for monthly/non-acute MCL
violations under the RTCR and Alternative option. Additionally, resources used to issue high
numbers of monthly/non-acute MCL violations and the time spent responding to customer
inquiries about the violations may be better employed on other PWS issues which could result in
further reduced risk.
As discussed above, the influences on risk of eliminating the PN requirements for
monthly/non-acute MCL violations may move risk in both directions. A decrease in the overall
information received by consumers may result in reduced use of averting behaviors when they
become necessary to avoid potentially contaminated drinking water, thus increasing potential
risk. Conversely, too frequent notice of information that may not be directly related to a health
risk may reduce confidence in the PWS, resulting in averting behavior that is not necessary, or
may cause confusion or indifference that may result in consumers not taking averting actions
when appropriate (i.e., ignoring an acute violation notice). Thus focusing on fewer, yet more
"serious" notices may result in a decrease in potential risk. Risk may also decrease as a result of
PWSs being able to better employ resources currently used on notice issuance and follow-up.
EPA also considered the effect of Tier 2 PN requirements for treatment technique
violations, which allow for up to a 30-day time delay between incurrence of the violation and
notification to the public. This time delay could equate to some increased risk for the public
relative to a scenario where an immediate notification was required. However, the change in risk
relative to the 1989 TCR, which does not require assessments or corrective actions, is a decrease
in risk. In allowing PWSs the flexibility to take up to a month to correct a sanitary defect before
being required to issue PN, EPA is attempting to balance the potential public health benefits
associated with having prompt notification for all treatment technique violations with the
potential inefficiency and extra costs related to cases where the same level of expedience may
not be necessary. Tier 1 PN is still required within 24 hours to address situations where an
imminent health threat occurs.
In summary, EPA assumes that there would not be an overall increase in risk by changing
PN requirements for monthly/non-acute MCL violations under the RTCR, but given the
41 Under the 1989 TCR, for systems taking fewer than 40 samples per month, a non-acute violation occurs if 2 or
more samples are TC+; for systems taking 40 or more samples per month, a non-acute violation is triggered by >5
percent of TC+ samples.
Economic Analysis for the Final RTCR
6-10
September 2012
-------
contradictions in potential risk reductions the potential change in risk is best characterized as
"unknown."
Monitoring and Reporting Violations
The TCRDSAC also evaluated the effects of revised PN requirements for monitoring and
reporting violations under the RTCR and Alternative option and concluded that significant
reductions in monitoring and reporting violations and associated costs may be realized through
the regulatory framework of the RTCR. For the 1989 TCR, monitoring and reporting violations
result in Tier 3 PN. Under the RTCR and Alternative option, monitoring violations are separate
from the reporting violations, and the PN requirements for failing to take some types of samples
are more stringent as compared to the monitoring and reporting violations under the 1989 TCR.
For repeat sampling, a PWS that fails to take every required repeat sample following a routine
EC+ sample must:
• Initiate Tier 1 PN within 24 hours;
• Initiate consultation with the state no later than 24 hours after learning of the
violation, to determine additional PN requirements, if any;
• Perform a level 2 assessment/corrective action: and
• Increase to a minimum of monthly monitoring.
For a PWS that fails to take every required repeat sample after any single TC+ sample,
the PWS must perform a level 1 assessment/corrective action. Failure to perform a required
assessment and/or corrective action results in a treatment technique violation (Tier 2) and a
minimum of monthly monitoring.
For routine and additional routine monitoring, a PWS that does not take every required
routine sample, or every required additional routine sample, in a compliance period is still
subject to Tier 3 PN; noncompliance with sampling requirements does not necessarily increase
the likelihood that a PWS will be contaminated with fecal contamination and/or a waterborne
pathogen and so does not present a direct or immediate public health risk. However, if the PWS
has monitoring violations in 2 of 4 quarters (for systems on quarterly monitoring) or misses its
required annual sample (for systems on annual monitoring), the PWS must revert to monitoring
no less than monthly.
Overall, the added PN stringency for monitoring violations is expected to decrease
potential risk under the RTCR and Alternative option as PWSs opt to perform required sampling
to avoid transition to increased monitoring requirements or other additional actions.
Summary Exhibit
The component discussions in Sections 6.2.1 through 6.2.8 above address the individual
effects under each rule component on the various system types and sizes. The terms "increase,"
"decrease," and "no change" in Exhibit 6.1 indicate the direction of change in risk under the
RTCR and Alternative option relative to the 1989 TCR. Risk may change for some system sizes
Economic Analysis for the Final RTCR
6-11
September 2012
-------
or types and not for others under a given rule component. In such cases, Exhibit 6.1 reflects the
overall change in risk direction and does not necessarily apply to all types and sizes of systems.
Economic Analysis for the Final RTCR
6-12
September 2012
-------
Exhibit 6.1 Potential Changes in Risk under the RTCR and Alternative Option Relative to the 1989 TCR
Rule Component
Factors Leading to a Potential
Increase in Risk
Factors Leading to a Potential Decrease in
Risk
Overall Predicted Change in
Risk
RTCR
Alternative
option
RTCR
Alternative
option
RTCR
Alternative
option
Implementation
Activities
None
None
None
None
No change
No change
Routine
Monitoring
(Including
Reduced
Monitoring)
None
None
Increased stringency
in requirements to
qualify for reduced
monitoring along with
requirement to return
to baseline monitoring
upon loss of these
criteria is expected to
result in decreased
risk (i.e., fewer PWSs
will qualify and
therefore more will
monitor more
frequently).
PWSs all monitor monthly
in the first few years of
implementation of the
RTCR, which is an
increase in sampling
frequency for systems
that monitor quarterly or
annually under the 1989
TCR. After the first few
years, systems may
reduce to quarterly, but
none may reduce to
annual monitoring,
creating a decrease in risk
for systems on annual
monitoring under the 1989
TCR.
Decrease
Decrease
Repeat
Monitoring
Required repeat
samples reduced
from 4 to 3 for
systems serving
<1,000 people
Required repeat
samples reduced
from 4 to 3 for
systems serving
<1,000 people
None
None
Increase
Increase
Economic Analysis for the Final RTCR
6-13
September 2012
-------
Rule Component
Factors Leading to a Potential
Increase in Risk
Factors Leading to a Potential Decrease in
Risk
Overall Predicted Change in
Risk
RTCR
Alternative
option
RTCR
Alternative
option
RTCR
Alternative
option
Additional
Routine
Monitoring
Additional routine
samples are no
longer required for
PWSs monitoring
monthly.
GW PWSs serving
<1,000 people
would reduce
additional routine
samples from 5 to
3.
Additional routine
samples are no
longer required
for PWSs
monitoring
monthly.
GW PWSs
serving <1,000
people would
reduce additional
routine samples
from 5 to 3.
None
None
Increase
Increase
Annual Site Visits
None (only states
currently
performing annual
site visits are
expected to
continue)
Annual
monitoring is not
permitted under
the Alternative
option, so the
protective benefit
of the annual site
visit is lost.
None (only states
currently performing
annual site visits are
expected to continue)
None
No change
Increase
Assessments
None
None
Mandatory
assessments are a
new requirement.
Mandatory assessments
are a new requirement.
Decrease
Decrease
Corrective
Actions
None
None
Mandatory corrective
actions are a new
requirement.
Mandatory corrective
actions are a new
requirement.
Decrease
Decrease
Economic Analysis for the Final RTCR
6-14
September 2012
-------
Rule Component
Factors Leading to a Potential
Increase in Risk
Factors Leading to a Potential Decrease in
Risk
Overall Predicted Change in
Risk
RTCR
Alternative
option
RTCR
Alternative
option
RTCR
Alternative
option
Public Notification
—Monthly/Non-
Acute MCL
Violations
Reduction in
available public
information
Possible PWS
complacency
Reduction in
available public
information
Possible PWS
complacency
Less confusion (PN
more in line with
potential health risks)
PWS resources used
more efficiently
Less confusion (PN more
in line with potential
health risks)
PWS resources used
more efficiently
Unknown
Unknown
Public Notification
—Monitoring and
Reporting
Violations
None
None
Increased stringency
of PNs motivates
PWSs to conduct
required sampling.
Increased stringency of
PNs motivates PWSs to
conduct required
sampling.
Decrease
Decrease
Overall
Decrease
Decrease
Note:
1) Detailed discussion ofthe rationale for determinations of potential risk for each rule component is presented in Ch. 6 (Sections 6.2.1-6.2.8 above) of this EA.
Implementation activities consist of administrative activities by PWSs and states to implement the rule.
2) Assessment of potential changes in risk for monitoring components is an overall assessment. Potential changes (or static state) of risk for particular system
sizes and types differ according to individual rule requirements and are discussed in Sections 6.2.1-6.2.8 above. Chapter 3 provides a detailed description of
the rule components for all three regulatory options, and the Preamble to the RTCR provides additional discussion ofthe TCRDSAC process and the rationale
underlying the structure ofthe regulatory options considered.
Economic Analysis for the Final RTCR
6-15
September 2012
-------
6.2.9 Summary of Qualitative Benefits Analyses
The discussions above describe relative risks in terms of individual components of the
regulatory options. Based on the discussions presented above, EPA has used best professional
judgment informed by the TCRDSAC to qualitatively estimate the potential changes in public
health for each regulatory option, as compared with the 1989 TCR. These assessments were
made with contributions from a wide range of drinking water experts, including public health
scientists, engineers, administrators, and regulatory experts. The overall change in risk relative to
the 1989 TCR is a result of the complex interactions of all rule components. As the discussion
above shows, under repeat and additional routine monitoring provisions for the RTCR and
Alternative option, there is a potential for increased risk for PWS customers because TC
monitoring frequency may be reduced for some PWSs. However, this increase is expected to be
more than offset by potential decreases in risk from increased routine monitoring for some PWSs
due to more stringent criteria to qualify and stay on reduced monitoring and the addition of the
assessments and corrective action provisions that will find and fix problems identified by
monitoring.
The consensus opinion resulting from the TCRDSAC deliberations was that the RTCR
would achieve a net risk reduction compared to the 1989 TCR. The committee applied best
professional judgment in determining that the increased protection provided by the new
requirements for implementing focused assessments and implementing appropriate corrective
actions would more than offset any potential increase in risk introduced by the reduction in
samples and other changes resulting from the RTCR. To present the potential for further
reduction in risk due to the increased numbers of samples taken, especially in the first several
years of implementation, the committee considered monthly monitoring for all systems, similar
42
to that required under the Alternative option. However, the additional burden of requiring all
PWSs to initially monitor on a monthly basis (regardless of PWS size or type) and limiting
reduced monitoring to quarterly (disallowing annual monitoring) would fall disproportionately
on small systems based on their proportionately large increase in activity under the RTCR.
6.3 Assessment of Predictive Analysis Results
Based on discussions in and information developed for the TCRDSAC meetings
(described in Chapter 3 of this EA), EPA anticipated prior to beginning this EA that the RTCR
would not be a significant rule in terms of costs (i.e., less than $100 million annually). However,
EPA considered the feasibility of performing a traditional risk assessment that would produce
quantified estimates of costs, benefits, and net costs and benefits, as outlined in section 1412
(b)(3)(C) of the Safe Drinking Water Act (SDWA). For this type of analysis, the minimum
information requirements comprise data on contaminant occurrence; exposure rates in the
population defined for the various pathways (i.e., water consumption, inhalation, and dermal
contact); potential health effects associated with exposure to contaminated water; and a dose-
response relationship. A quantified benefits analysis would use this information to estimate the
number of avoided cases of morbidity or mortality associated with the rule, which would then be
valued in terms of saved lives and preserved quality of life and work capacity.
42 Additional costs under the Alternative option are discussed Chapter 7 of this EA.
Economic Analysis for the Final RTCR
6-16
September 2012
-------
For the RTCR EA, the available information includes data on the incidence of TC+
samples collected under the 1989 TCR over the 6-year period from 1998-2005, from which EPA
selected one complete year (2005) to use for analyses in this EA, as explained in Section 4.2.2 of
this EA. For the year 2005, EPA also has data on TC+ samples that subsequently test positive for
E. coli (using EPA-approved standard methods available at that time). No other national
occurrence data are available.
As discussed in Chapter 2 of this EA, the presence of E. coli is an indication that the
water contains fecal contamination. SomeE. coli strains (or serotypes) such as E. coli 0157:H7,
for example, are pathogens. However, EPA recognizes that the EPA-approved standard methods
available for E. coli do not typically identify the presence of the pathogenic E. coli strains, such
as E. coli 0157:H7. Thus, E. coli occurrence, as used in this EA, serves as an indication of fecal
contamination but not necessarily pathogenic contamination. E. coli occurrence does not confer
any significant quantitative information about the likelihood of health effects (e.g., acute
gastrointestinal illness or chronic illnesses as described in Chapter 2 of this EA) from consuming
drinking water contaminated with fecal indicator organisms.
There are few data reporting the co-occurrence in a single sample of fecal indicator E.
coli (assayed using EPA-approved standard methods) and pathogenic E. coli strains. One notable
exception are the data reported by Cooley et al. (2007), which showed high concentrations of
pathogenic E. coli strains in samples containing high concentrations of fecal indictor E. coli.
These data are from streams and other poor quality surface waters surrounding California
spinach fields associated with an E. coli 0157:H7 foodborne outbreak. Data equivalent to these
are not available from drinking water samples collected under the 1989 TCR.
Absent any definitive data on co-occurrence of fecal indicator (E. coli) and pathogenic E.
coli, EPA did not estimate the cases of morbidity or mortality avoided. Instead, EPA estimated
changes in occurrence and the resulting changes in assessments and corrective actions performed
for systems serving <4,100 people. For systems serving >4,100 people, EPA applied the 2007
violations data to estimate the increase in effective corrective actions implemented. Discussion
of reductions in risk, then, considers the change in occurrence and corrective actions
implemented for systems serving <4,100 and the changes in corrective actions implemented for
systems serving >4,100.43
For all systems, EPA also estimated the behavioral response of the regulated community
based on projected occurrence rates or violations under the RTCR and Alternative option,
including the frequency of Level 1 or Level 2 assessments and the type and number of corrective
actions implemented by PWSs to address the problems identified. EPA expects that the effects of
these changes on risk will be varied, as described in Section 6.3.1 below.
43 The rationale for using different metrics as proxies for risk reduction in systems serving <4,100 people and those
serving >4,100 is explained in Chapters 4 and 5 of this EA.
Economic Analysis for the Final RTCR
6-17
September 2012
-------
6.3.1 Assessment of Predictive Analysis Results for Systems Serving <4,100 People and
Systems Serving >4,100 People
Because the PWSs serving 4,100 people or fewer have a higher initial E. coli occurrence
and will be triggered into more assessments and corrective actions than PWSs serving >4,100
people, the increase in benefits for systems serving <4,100 people will be more evident as
compared to the systems serving >4,100 people. In particular, model results suggest that
customers of GW TNCWSs serving 100 or fewer people, which constitute approximately 40
percent of PWSs, experience the most improvement in water quality under the RTCR. That is,
the occurrence of E. coli is predicted to decrease more for these systems that for other systems
types.
6.3.1.1 Systems Serving <4,100 People
For systems serving <4,100 people, EPA developed a model to simulate regulatory
responses for a 30-year period of analysis, including years 1-5 in which GWR44 is in effect,
years 3-5 in which the RTCR is being implemented, and years 6-30 during which RTCR
requirements are also in effect. The 30-year modeled time period includes a total of 28 years of
RTCR impacts post-promulgation. Although this EA considers benefits and costs for only 25
years post-promulgation, the 30-year period was presented in Chapter 5 for broader
consideration of the appropriateness of the model in terms of its results. The 25 years post-
promulgation that are considered in this EA encompass years 3-27 of the 30-year modeled
period. A complete description of the model is provided in Chapter 5 of this EA, and data
sources used are described in Chapter 4.
Output from the system simulation model for systems serving <4,100 people includes the
following estimates for each year of analysis:
• Samples taken (routine, additional routine, and repeat);
• Number of positive results (TC+ and/or EC+)
• Level 1 and Level 2 assessments conducted based on non-acute violations;
• Level 2 assessments conducted based on acute violations; and
• Corrective actions based on Level 1 assessments and Level 2 assessments.
The results of analyses in terms of expected changes in hit rates (positive samples/totals
samples taken) over time are presented in chapter 5 (Section 5.3.3) and Appendix B. Predicted
TC+ hit rate results reflect the overall increase in water quality expected over time under the
RTCR and Alternative option. Exhibits 6.2-6.4 present a summary of the additional endpoints
listed above for the 25-year period of analysis following rule promulgation.
44 Reductions in EC+ occurrence or violations are only attributable to the RTCR if they result from requirements of
the RTCR regulatory options; reductions resulting from the GWR are not attributable to the RTCR and are not
considered further in this chapter. As described in sections 4.2, 4.3, and 5.3 of this EA, GWR effects are
incorporated to adjust source data to the appropriate baseline for this EA.
Economic Analysis for the Final RTCR
6-18
September 2012
-------
As discussed earlier in this section, because a dose-response relationship between EC+
occurrence (i.e., acute violations) and illness is not available, EPA is focusing on changes in EC+
occurrence. EPA assumes that reduced occurrence in PWSs should generally correlate with a
reduction in risk of contamination to PWS drinking water. Therefore, the acute violation rates
predicted under the RTCR and Alternative option (Exhibits 6.3-6.4) represent a level of risk of
contamination that is reduced from that predicted under the 1989 TCR (Exhibit 6.2). The
numbers of predicted acute violations have two major drivers. First, improvements in water
quality are predicted to result in fewer acute violations. Second, the monitoring frequency
impacts the number of acute violations found, regardless of water quality. Both decreases in
acute violations attributable to water quality improvements and increases due to additional
diagnostic ability of more samples taken result in reduced risk. The combination of these two
influences in the predictive model means that the risk is expected to be lower even though the
number of predicted acute violations may actually be higher or lower than under the 1989 TCR
as reflected in exhibits 6.2-6.4. This concept is discussed in the uncertainty analysis in Section
6.4 of this chapter.
The changes in the steady state estimates of annual acute violations from the 1989 TCR
to the RTCR and Alternative option are shown in Exhibit 6.5 and 6.6 in absolute numbers and as
a percentage change from the 1989 TCR, respectively. The steady state in the model refers to the
period beginning in years 7 (CWSs) and 9 (NCWSs) following promulgation, after the
proportions of systems sampling on monthly, quarterly, or annual regimens are adjusted
following a period of assessment.45 systems that qualify for reduced monitoring will begin their
new regimens in years 7 and 9 after promulgation, respectively, for CWSs and NCWSs. The
estimates shown in Exhibit 6.5 for systems serving <4,100 people are from the predictive model
and reflect the average annual estimates for the 25-year period of analysis, which includes 3
years of initial implementation followed by 22 years in which new rule requirements are also in
effect. To accurately reflect the average annual results under this schedule, the model output for
the 22 years of additional activity following the 3 years of initial implementation is divided by
22, rather than the entire 25 years, since that would distort the results downward. The steady
state reductions in the number of annual acute violations found under the RTCR and Alternative
option primarily reflect the benefits of corrective actions under these two options in preventing
many of the acute violations that would otherwise occur over this period.
These results show that under the RTCR, no subgroups are predicted to experience an
increase in annual acute violations. While most categories of systems/sizes would experience a
decrease in predicted numbers of violations under the Alternative option, four categories would
45 The effective date of the RTCR occurs after 3 years of initial implementation, that is, at the start of year 4 post-
promulgation. For CWSs, years 4-6 post-promulgation are the period of assessment for potential to move to reduced
monitoring; for NCWSs, years 4-8 post-promulgation are the period of assessment for reduced monitoring. Under
the Alternative option, this period of assessment occurs in years 4-8 post-promulgation for all PWSs. In the
aggregate, the steady state regarding reduced monitoring schedules under the Alternative option begins in year 9
post-promulgation. Chapter 5 provides a description of the criteria used to estimate the steady percent of systems
qualifying for reduced monitoring under the RTCR and Alternative option. In this chapter, these transitions are
presented on the 30-year scale of the modeling period, in which years 1-3 are the baseline occurrence for the 1989
TCR with the GWR effects and are pre-promulgation. Exhibits 5.16-5.21 show that the steady state begins in model
year 9 (CWSs) and model year 11 (NCWSs). This modeling is not performed for systems operating under the 1989
TCR; they are assumed to remain on the initial schedule presented in Chapter 4 (baseline).
Economic Analysis for the Final RTCR
6-19
September 2012
-------
actually experience an increase (a positive percent value in Exhibit 6.6): GW NTNCWSs serving
501-1,000 people and GW TNCWSs serving <100, 101-500, and 501-1,000 people. It is
important to note that in two of those cases, the absolute increases in annual acute violations
predicted under either option is very small (<1 annual acute violation as shown in Exhibit 6.5),
but translate into observable percentage changes in Exhibit 6.6. The two categories with an
increase of greater than one annual acute violation are the GW TNCWSs systems serving <100
people and those serving 101-500 people, for which increases of approximately 52 and 16
annual acute violations, respectively, are predicted. For these TNCWSs under the Alternative
option, increased monitoring is expected to lead to an overall increase in annual acute violations
(and is also the driver of the greater total number of annual acute violations predicted).
As discussed earlier, a decrease in acute violations may be caused by an improvement in
water quality, which in this model would result from an increased number of effective corrective
actions being implemented (i.e., occurrence events are "prevented"). An increase in acute
violations can be attributable to a decrease in water quality or a PWS improving its ability to
detect more issues (e.g., through more sampling). Alternatively, a decrease in acute violations
could be caused by a decreased ability in PWSs to detect the occurrence of TC/EC because of a
reduced sampling schedule, resulting in undetected or "missed" occurrence events. Section 6.4 of
this chapter presents a stepwise analysis to discern the relative significance of the effects of
reduced additional routine samples and increased corrective action efficacy as shown by
predictions of "prevented", "found", and "missed" acute violations under the RTCR.
6.3.1.2 Systems Serving >4,100 People
The number of acute and non-acute MCL violations for a given group (based on system
type and water source) of PWSs serving >4,100 people was estimated using 2005 Safe Drinking
Water Information System/Federal Version (SDWIS/FED) data (USEPA, 2005). These estimates
were used to determine the number of Level 1 and Level 2 assessments and associated corrective
actions triggered for systems serving >4,100 in the period of analysis. EPA made a simplifying
assumption that for PWSs serving >4,100 people the number of annual assessment triggers
would remain constant throughout the 22 years following the 3 years of initial implementation of
the RTCR. As assumed in the occurrence model for systems serving <4,100 people, this analysis
also assumes that systems responding to a Level 2 assessment trigger (i.e., an acute MCL
violation under the 1989 TCR) would identify and specifically address the cause of the
contamination at a rate increased by 10 percent under the RTCR and Alternative option
compared to the 1989 TCR.46
Exhibits 6.2-6.4, described in Section 6.3.1.1 for systems serving <4,100 people, also
include the number of activities (assessments and corrective actions) EPA expects PWSs to
implement under each of the regulatory options considered for systems serving >4,100. Any
reduction in risk is estimated to derive from the additional corrective actions predicted per the
explanation above. Exhibit 6.5 also reflects that EPA is not quantifying any potential change in
46The 10 percent assumption is based on EPA discussions with stakeholders regarding experiences with
implementing the 1989 TCR and the expected impact of RTCR. A sensitivity analysis evaluating alternative
assumptions was conducted. Results of these analyses are discussed in Chapter 5 (section 5.3.3.1).
Economic Analysis for the Final RTCR
6-20
September 2012
-------
the number of annual acute violations for PWSs serving >4,100 people during the period. This
assumption simplifies the analysis and is consistent with EPA's understanding based on
TCRDSAC deliberations and best professional judgment that systems serving >4,100 are at a
relatively steady state with regard to operations. These systems have been in a position relative to
systems serving <4,100 people to diagnose more and address the cause of acute violations more,
and they have likely had more resources to apply in maintaining and updating their systems on a
regular basis. Therefore, EPA does not believe that systems serving >4,100 will make many
changes based on RTCR implementation, or that they will experience large changes in their
occurrence rates for TC and E. coli after RTCR promulgation.
Economic Analysis for the Final RTCR
6-21
September 2012
-------
Exhibit 6.2 Predicted Outcomes (25-Year Period of Analysis) under the 1989
TCR
PWS Size
(Population
Served)
PWSs
Subject
to 1989
TCR
Number of
Routine
Monitoring
Samples
Number of
Additional
Routine
Monitoring
Samples
Number of
Repeat
Monitoring
Sam pies
Number of
Non-Acute
Violation
Assessments
(Single
Violations)
Number of
Corrective
Actions
(based on
Single Non-
Acute Violation
Assessment)
Number of
Acute
Violation
Assessments
Number of
Non-Acute
Violation
Assessments
(Multiple
Violations)
Number of
Corrective
Actions
(based on
Acute and
Multiple Non-
Acute Violation
Assessments)
Tier 2 PNs
A
B
C
D
E
F
G
H
I
J=EK3+H
Community Water Systems (CWSs) - SW
<100
1,170
304,247
23,167
18,698
525
157
184
865
101 - 500
2,150
562,198
27,009
21,684
649
167
111
927
501-1,000
1,173
306,605
15,334
12,299
361
102
63
526
1,001-4,100
2,938
1,921,237
55,132
33,729
954
162
149
1,265
4,101-33,000
3,164
10,636,296
186,729
2,152
197
2,349
33,001-96,000
720
11,058,960
194,149
534
56
590
96,001-500,000
308
10,190,400
178,901
233
24
257
500,001-1 Million
31
2,019,600
35,456
22
22
> 1 Million
17
1,686,960
29,616
Totals
11,671
38,686,502
120,642
711,259
5,429
865
507
6,801
Community Water Systems (CWSs) - GW
<100
11,938
2,815,951
286,073
194,462
9,772
1,141
5,383
16,295
101 - 500
13,892
3,344,578
243,895
171,252
8,169
1,025
4,214
13,408
501-1,000
4,467
1,072,202
70,803
51,673
2,250
284
1,050
3,584
1,001-4,100
6,443
3,997,293
160,710
100,618
3,545
477
2,808
6,831
4,101-33,000
3,156
9,145,224
230,201
4,545
263
4,807
33,001-96,000
335
4,884,000
122,938
656
53
709
96,001-500,000
63
1,945,680
48,976
129
10
139
500,001-1 Million
4
253,440
6,380
> 1 Million
3
269,280
6,778
Totals
40,301
27,727,648
761,481
933,279
29,066
3,253
13,455
45,773
Nontransient Noncommunity Water Systems (NTNCWSs) - SW
<100
250
65,018
4,910
3,991
98
35
77
210
101 - 500
253
66,045
3,735
3,011
88
29
40
157
501-1,000
88
22,976
1,278
1,029
30
9
13
52
1,001-4,100
72
41,759
2,142
1,348
42
19
37
98
4,101-33,000
22
50,424
1,628
5
5
33,001-96,000
2
34,320
1,108
96,001-500,000
1
31,680
1,023
500,001-1 Million
> 1 Million
Totals
688
312,223
12,065
13,138
262
93
167
522
Nontransient Noncommunity Water Systems (NTNCWSs) - GW
<100
8,826
971,538
128,775
84,992
5,581
856
3,829
10,267
101 - 500
6,613
725,785
66,525
43,597
3,130
447
1,273
4,849
501-1,000
1,718
190,649
16,037
10,680
744
95
298
1,136
1,001-4,100
812
460,470
28,214
17,790
818
169
974
1,961
4,101-33,000
70
153,648
5,936
123
9
132
33,001-96,000
2
23,760
918
4
4
96,001-500,000
500,001-1 Million
> 1 Million
Totals
18,041
2,525,850
239,551
163,913
10,400
1,577
6,373
18,350
Transient Noncommunity Water Systems (TNCWSs) - SW
<100
1,339
345,401
40,475
33,065
1,093
430
780
2,302
101 - 500
497
128,156
15,261
12,454
410
170
320
900
501-1,000
88
22,691
2,704
2,207
67
28
47
142
1,001-4,100
67
40,151
4,155
2,707
92
50
128
270
4,101-33,000
18
40,656
8
8
33,001-96,000
96,001-500,000
500,001-1 Million
> 1 Million
1
102,960
Totals
2,010
680,015
62,596
50,434
1,670
677
1,275
3,622
Transient Noncommunity Water Systems (TNCWSs) - GW
<100
60,200
4,493,808
905,554
600,315
44,730
6,649
25,425
76,805
101 - 500
19,275
1,614,924
316,238
210,714
14,530
2,089
8,864
25,483
501-1,000
1,963
177,264
32,730
22,064
1,477
221
896
2,595
1,001-4,100
617
335,283
29,957
19,113
927
186
1,138
2,251
4,101-33,000
67
156,288
8,909
116
4
120
33,001-96,000
2
34,320
1,956
96,001-500,000
1
26,400
1,505
500,001-1 Million
1
63,360
3,612
> 1 Million
Totals
82,126
6,901,647
1,284,478
868,188
61,780
9,149
36,324
107,253
Grand Total
154,837
76,833,885
2,480,814
2,740,210
108,608
15,613
58,102
182,322
Source:
1) Appendix A, Exhibit A. 1.z.
2) Redicted outcomes for systems serving <4,100 are output from the occurrence trodel detailed in Ch. 5 of this Eft; predicted outcomes for systems serving >4,100 people are from the
larger systems rrodel based on 2007 SDl/VIS violations data.
Notes:
1) For modeling purposes, EPA estimated only the net change in the nurrber of corrective actions performed under the RTCR and Alternative option compared to the 1989 TCR. Because
only the net change in the number of corrective actions is estimated, no additional corrective actions are rmdeled for the 1989 TCR (it is assumed that RA/Ss are already performing some
corrective actions under the 1989 TCR).
2) Results differ slightly from those presented in &(. 5.10 -5.15 because they are capturing slightly different time periods of the 30 modeled years. For completeness in discussing the
simulation model, Chapter 5 exhibits show 30 years of results, beginning w ith 5 years of GWR in effect, the last 3 of w hich also include initial irrplementation of the RTCR, follow ed by 25
years of RTCR in effect. Alternatively, Ex. 6.2 - 6.4 include the modeled period that encompasses years 3 - 27 of the 30-year modeling period, including the 3 years of RTCR initial
implementation and the 22 years that follow of RTCR in effect. During the 3 years of RTCR initial irrplementation, under either the RTCR or Alternative option, systems are still w orking
under the requirements of the 1989 TCR and GWR while incurring costs for the initial irrplementation activities of the new rule requirements.
Economic Analysis for the Final RTCR
6-22
September 2012
-------
Exhibit 6.3 Predicted Outcomes (25-Year Period of Analysis) for the RTCR
Number of
Number of
Number of
Number of
Level 2
Level 2
Number of
Level 1 and
Number of
Additional
Number of
Corrective
Assessments
Assessments
Corrective
Level 2 (based
PWSs
Routine
Routine
Repeat
Number of
Actions (based
(based on
(based on non
Actions (based
on non-acute
PWS Size
Subject
Monitoring
Monitoring
Monitoring
Level 1
on Level 1
Acute
acute
on Level 2
triggers)
to RTCR
Samples
Samples
Samples
Assessments
Assessments)
Violations)
triggers)
Assessments)
Assessments
Served)
A
B
C
D
E
F
G
H
1
J^-HH
Community Water Systems (CWSs) - SW
<100
1,170
308,880
13,764
400
36
100
102
21
501
101-500
2,150
567,600
15,660
539
56
119
75
20
615
501-1,000
1,173
309,672
8,708
277
27
75
40
12
317
1,001-4,100
2,938
1,951,224
33,326
920
95
146
132
29
1,052
4,101-33,000
3,164
10,636,296
181,661
2,152
215
197
20
2,152
33,001-96,000
720
11,058,960
188,880
534
53
56
6
534
96,001-500,000
308
10,190,400
174,046
233
23
24
2
233
500,001-1 Mllion
31
2,019,600
34,493
22
2
22
> 1 Mllion
17
1,686,960
28,812
Totals
11,671
38,729,592
679,350
5,076
507
717
349
110
5,425
Community Water Systems (CWSs) - GW
<100
11,938
2,870,075
8,760
156,897
8,004
791
853
3,523
398
11,527
101 - 500
13,892
3,391,200
6,127
136,906
6,502
669
696
2,399
335
8,901
501-1,000
4,467
1,085,730
1,844
39,659
1,780
168
188
626
85
2,406
1,001-4,100
6,443
4,079,328
96,939
3,208
318
342
1,705
206
4,913
4,101-33,000
3,156
9,145,224
217,321
4,545
454
263
26
4,545
33,001-96,000
335
4,884,000
116,060
656
66
53
5
656
96,001-500,000
63
1,945,680
46,236
129
13
10
1
129
500,001-1 Mllion
4
253,440
6,023
> 1 Mllion
3
269,280
6,399
Totals
40,301
27,923,956
16,731
822,439
24,824
2,480
2,405
8,253
1,056
33,077
Nontransient Noncomm unity Water Systems (NTNCWSs)
SW
<100
250
66,000
3,040
75
8
28
41
7
116
101 - 500
253
66,792
2,169
69
7
19
24
4
93
501-1,000
88
23,232
756
24
2
6
9
2
33
1,001-4,100
72
42,768
1,228
37
4
13
23
4
59
4,101-33,000
22
50,424
1,448
5
0
5
33,001-96,000
2
34,320
985
96,001-500,000
1
31,680
910
500,001-1 Mllion
> 1 Mllion
Totals
688
315,216
10,536
209
22
67
98
17
306
Nontransient Noncomm unity Water Systems (NTNCWSs)
GW
<100
8,826
932,025
48,142
68,123
4,797
493
559
2,010
254
6,807
101 - 500
6,613
678,688
25,630
35,860
2,794
271
315
757
107
3,552
501-1,000
1,718
180,145
6,166
8,601
675
66
79
168
24
843
1,001-4,100
812
473,352
15,887
690
68
114
530
65
1,221
4,101-33,000
70
153,648
5,157
123
12
9
1
123
33,001-96,000
2
23,760
797
4
0
4
96,001-500,000
500,001-1 Mllion
> 1 Mllion
Totals
18,041
2,441,617
79,938
134,426
9,084
912
1,077
3,466
450
12,550
Transient Noncomm unity Water Systems (TNCWSs) -SW
<100
1,339
353,496
23,122
796
76
250
425
66
1,221
101-500
497
131,208
8,192
278
27
90
154
25
432
501-1,000
88
23,232
1,533
50
5
17
25
4
75
1,001-4,100
67
42,240
2,312
73
7
29
69
10
142
4,101-33,000
18
40,656
2,225
8
1
8
33,001-96,000
96,001-500,000
500,001-1 Mllion
> 1 Mllion
1
102,960
5,636
Totals
2,010
693,792
43,020
1,204
116
386
674
105
1,878
Transient Noncomm unity Water Systems (TNCWSs) -GW
<100
60,200
6,076,163
446,166
631,105
47,190
4,755
5,477
20,628
2,593
67,818
101-500
19,275
1,940,946
135,822
194,697
13,780
1,363
1,608
5,694
799
19,474
501-1,000
1,963
206,130
14,078
20,078
1,396
143
177
585
76
1,982
1,001-4,100
617
348,480
16,027
773
77
117
638
76
1,412
4,101-33,000
67
156,288
7,188
116
12
4
0
116
33,001-96,000
2
34,320
1,578
96,001-500,000
1
26,400
1,214
500,001-1 Mllion
1
63,360
2,914
> 1 Mllion
Totals
82,126
8,852,088
596,065
874,801
63,256
6,349
7,383
27,546
3,544
90,801
Grand Total
154,837
78,956,260
692,734
2,564,572
103,653
10,386
12,035
40,385
5,282
144,038
Source:
1) Appendix A, Exhibit A ,2.z.
2) Predicted outcomes for systems serving 24,100 are output from the occurrence model detailed in Ch. 5 of this EA; predicted outcomes for systems serving >4,100 people are from the
larger systems model based on 2007 SDWIS violations data.
Notes:
1) Estimates of the number of assessments and corrective actions are net increases in activity predicted to occur under the RTCR relative to the 1989 TCR; estimates of "zero" reflect that no
additional such activity occurs as compared to baseline (the 1989 TCR).
2) Results differ slightly from those presented in Ex. 5.10 - 5.15 because they are capturing slightly different time periods of the 30 modeled years. For completeness in discussing the
simulation model, Chapter 5 exhibits show 30 years of results, beginning with 5 years of GWR in effect, the last 3 of which also include initial implementation of the RTCR, followed by 25
years of RTCR in effect. Alternatively, Ex. 6.2 - 6.4 include the modeled period that encompasses years 3 - 27 of the 30-year modeling period, including the 3 years of RTCR initial
implementation and the 22 years that follow of RTCR in effect. During the 3 years of RTCR initial implementation, under either the RTCR or Alternative option, systems are still working under the
requirements of the 1989 TCR and GWR while incurring costs for the initial implementation activities of the new rule requirements.
Economic Analysis for the Final RTCR
6-23
September 2012
-------
Exhibit 6.4 Predicted Outcomes (25-Year Analysis Period) for the Alternative
Option
Number of
Number of
Number of
Number of
Level 2
Level 2
Number of
Level 1 and
PWSs
Number of
Additional
Number of
Corrective
Assessments
Assessments
Corrective
Level 2(based
Subject to
Routine
Routine
Repeat
Number of
Actions (based
(based on
(based on non
Actions (based
on non-acute
PWS Size
Alternative
Monitoring
Monitoring
Monitoring
Level 1
on Level 1
Acute
acute
on Level 2
triggers)
Option
Samples
Samples
Samples
Assessments
Assessments)
Violations)
triqqers)
Assessments)
Assessments
Served)
A
B
C
D
E
F
G
H
I
J=B-H
Community Water Systems (CWSs) - SW
<100
1,170
308,880
13,764
400
36
100
102
21
501
101 - 500
2,150
567,600
15,660
539
56
119
75
20
615
501-1,000
1,173
309,672
8,708
277
27
75
40
12
317
1,001-4,100
2,938
1,951,224
33,326
920
95
146
132
29
1,052
4,101-33,000
3,164
10,636,296
181,661
2,152
215
197
20
2,152
33,001-96,000
720
11,058,960
188,880
534
53
56
6
534
96,001-500,000
308
10,190,400
174,046
233
23
24
2
233
500,001-1 Million
31
2,019,600
34,493
22
2
22
> 1 Million
17
1,686,960
28,812
Totals
11,671
38,729,592
679,350
5,076
507
717
349
110
5,425
Community Water Systems (CWSs) - GW
<100
11,938
2,908,469
7,545
158,439
7,871
812
926
3,272
432
11,143
101 - 500
13,892
3,428,876
5,264
137,959
6,495
667
747
2,543
322
9,038
501-1,000
4,467
1,098,488
1,616
39,580
1,772
174
203
607
83
2,379
1,001-4,100
6,443
4,079,328
96,939
3,208
318
342
1,705
206
4,913
4,101-33,000
3,156
9,145,224
217,321
4,545
454
263
26
4,545
33,001-96,000
335
4,884,000
116,060
656
66
53
5
656
96,001-500,000
63
1,945,680
46,236
129
13
10
1
129
500,001-1 Million
4
253,440
6,023
> 1 Million
3
269,280
6,399
Totals
40,301
28,012,784
14,425
824,956
24,675
2,504
2,544
8,127
1,077
32,802
Nontransient Noncommunitv Water Systems (NTNCWSs) - SW
<100
250
66,000
3,040
75
8
28
41
7
116
101 - 500
253
66,792
2,169
69
7
19
24
4
93
501-1,000
88
23,232
756
24
2
6
9
2
33
1,001-4,100
72
42,768
1,228
37
4
13
23
4
59
4,101-33,000
22
50,424
1,448
5
0
5
33,001-96,000
2
34,320
985
96,001-500,000
1
31,680
910
500,001-1 Million
> 1 Million
Totals
688
315,216
10,536
209
22
67
98
17
306
Nontransient Noncommunity Water Systems (NTNCWSs) - GW
<100
8,826
1,314,175
36,965
91,416
5,673
562
723
3,390
399
9,063
101 - 500
6,613
976,627
19,382
48,269
3,551
356
446
1,333
183
4,884
501-1,000
1,718
249,760
4,802
11,817
814
81
99
298
41
1,112
1,001-4,100
812
473,352
15,887
690
68
114
530
65
1,221
4,101-33,000
70
153,648
5,157
123
12
9
1
123
33,001-96,000
2
23,760
797
4
0
4
96,001-500,000
500,001-1 Million
> 1 Million
Totals
18,041
3,191,322
61,149
173,343
10,855
1,080
1,393
5,551
689
16,406
Transient Noncommunity Water Systems (TNCWSs) - SW
<100
1,339
353,496
23,122
796
76
250
425
66
1,221
101 - 500
497
131,208
8,192
278
27
90
154
25
432
501-1,000
88
23,232
1,533
50
5
17
25
4
75
1,001-4,100
67
42,240
2,312
73
7
29
69
10
142
4,101-33,000
18
40,656
2,225
8
1
8
33,001-96,000
96,001-500,000
500,001-1 Million
> 1 Million
1
102,960
5,636
Totals
2,010
693,792
43,020
1,204
116
386
674
105
1,878
Transient Noncommunity Water Systems (TNCWSs) - GW
<100
60,200
9,524,123
333,524
912,589
57,597
5,737
7,796
37,532
4,450
95,129
101 - 500
19,275
3,021,771
104,732
282,740
17,358
1,616
2,441
10,924
1,337
28,282
501-1,000
1,963
304,534
10,412
27,932
1,661
163
230
1,015
123
2,676
1,001-4,100
617
348,480
16,027
773
77
117
638
76
1,412
4,101-33,000
67
156,288
7,188
116
12
4
0
116
33,001-96,000
2
34,320
1,578
96,001-500,000
1
26,400
1,214
500,001-1 Million
1
63,360
2,914
> 1 Million
Totals
82,126
13,479,275
448,667
1,252,181
77,506
7,605
10,589
50,109
5,986
127,615
Grand Total
154,837
84,421,981
524,241
2,983,387
119,526
11,834
15,695
64,908
7,983
184,433
Source:
1) Appendix A, Bchibit A.3.z.
2) Predicted outcomes for systems serving <4,100 are output from the occurrence model detailed in Ch. 5 of this EA; predicted outcomes for systems serving >4,100 people are from the larger
systems rrndel based on 2007 SDl/VIS violations data.
Notes:
1) Estimates of the nurrber of assessments and corrective actions are net increases in activity predicted to occur under the Alternative option relative to the 1989 TCR; estimates of "zero"
reflect that no additional such activity occurs as compared to baseline (the 1989 TCR).
2) Results differ slightly from those presented in Ex. 5.10 -5.15 because they are capturing slightly different time periods of the 30 modeled years. For completeness in discussing the simulation
model, Chapter 5 exhibits show 30 years of results, beginning w ith 5 years of GWR in effect, the last 3 of w hich also include initial implementation of the RTCR follow ed by 25 years of RTCR in
effect. Alternatively, &(. 6.2 - 6.4 include the modeled period that encompasses years 3 - 27 of the 30-year modeling period, including the 3 years of RTCR initial implementation and the 22 years
that follow of RTCR in effect. During the 3 years of RTCR initial irrplementation, under either the RTCR or Alternative option, systems are still working under the requirements of the 1989 TCR and
GWR while incurring costs for the initial implementation activities of the new rule requirements.
Economic Analysis for the Final RTCR
6-24
September 2012
-------
Exhibit 6.5 Predicted Average Annual Acute Violations by Regulatory Option and
System Type1
PWS Size
(Population
Number of
Alternative
Served)
Systems
1989 TCR
RTCR
Option
Community Water Systems (CWSs
-SW
<100
1,170
7
5
5
101-500
2,150
8
5
5
501-1,000
1,173
5
3
3
1,001-4,100
2,938.0
7.4
6.6
6.6
4,101-33,000
3,164
9
9
9
33,001-96,000
720
3
3
3
96,001-500,000
308
1
1
1
500,001-1 Million
31
0
0
0
> 1 Million
17
0
0
0
Totals
11,671
39
33
33
Community Water Systems (CWSs
-GW
<100
11,938
52
39
42
101-500
13,892
47
32
34
501-1,000
4,467
13
9
9
1,001-4,100
6,443
22
16
16
4,101-33,000
3,156
12
12
12
33,001-96,000
335
2
2
2
96,001-500,000
63
0
0
0
500,001-1 Million
4
0
0
0
> 1 Million
3
0
0
0
Totals
40,301
148
109
116
Nontransient Noncommunity Wate
r Systems (NTNCWSs) - SW
<100
250
2
1
1
101-500
253
1.3
0.9
0.9
501-1,000
88
0.4
0.3
0.3
1,001-4,100
72
0.9
0.6
0.6
4,101-33,000
22
0
0
0
33,001-96,000
2
0
0
0
96,001-500,000
1
0
0
0
500,001-1 Million
0
0
0
0
> 1 Million
0
0
0
0
Totals
688
4
3
3
Nontransient Noncommunity Wate
r Systems (NTNCWSs) - GW
<100
8,826
39
25
33
101-500
6,613
20.31
14.31
20.29
501-1,000
1,718
4.31
3.61
4.51
1,001-4,100
812
8
5
5
4,101-33,000
70
0
0
0
33,001-96,000
2
0
0
0
96,001-500,000
0
0
0
0
500,001-1 Million
0
0
0
0
> 1 Million
0
0
0
0
Totals
18,041
72
49
63
Transient Noncommunity Water S^
^sterns (TNCWSs) - SW
<100
1,339
20
11
11
101-500
497
8
4
4
501-1,000
88.00
1.25
0.79
0.79
1,001-4,100
67.0
2.3
1.3
1.3
4,101-33,000
18
0
0
0
33,001-96,000
0
0
0
0
96,001-500,000
0
0
0
0
500,001-1 Million
0
0
0
0
> 1 Million
1
0
0
0
Totals
2,010
31
18
18
Transient Noncommunity Water S^
^sterns (TNCWSs) - GW
<100
60,200
302
249
354
101-500
19,275
94.93
73.11
110.96
501-1,000
1,963
10.07
8.04
10.47
1,001-4,100
617
8
5
5
4,101-33,000
67
0
0
0
33,001-96,000
2
0
0
0
96,001-500,000
1
0
0
0
500,001-1 Million
1
0
0
0
> 1 Million
0
0
0
0
Totals
82,126
416
336
481
Grand Total
154,837
710
547
713
Source:
Output from RTCR models as described in Sections 5.3 and 5.4 of this EAfor
systems serving <4,100 people and systems serving >4,100 people.
Note:
Monitoring activates under RTCR begin in year 4 after rule promulgation. Therefore
to represent the average annual acute violations occurring over the 25-year modeled
time period, the total number of violations is divided by22 years where monitoring
activities are conducted.
Economic Analysis for the Final RTCR
6-25
September 2012
-------
Exhibit 6.6 Predicted Change in Average Annual Acute Violations by Regulatory
Option and System Type
PWS Size
(Population
Served)
RTCR
Alternative Option
Community Water Systems (CWSs) - SW
<100
-36%
-36%
101-500
-29%
-29%
501-1,000
-27%
-27%
1,001-4,100
-10%
-10%
Totals
-17%
-17%
Community Water Systems (CWSs) - GW
<100
-25%
-19%
101-500
-32%
-27%
501-1,000
-34%
-29%
1,001-4,100
-28%
-28%
Totals
-26%
-22%
Nontransient Noncommunity Water Systems (NTNCWSs) - SW
<100
-20%
-20%
101-500
-35%
-35%
501-1,000
-32%
-32%
1,001-4,100
-30%
-30%
Totals
-28%
-28%
Nontransient Noncommunity Water Systems (NTNCWSs) - GW
<100
-35%
-15%
101-500
-30%
0%
501-1,000
-16%
5%
1,001-4,100
-32%
-32%
Totals
-32%
-12%
Transient Noncommunity Water Systems (TNCWSs) - SW
<100
-42%
-42%
101-500
-47%
-47%
501-1,000
-37%
-37%
1,001-4,100
-43%
-43%
Totals
-43%
-43%
Transient Noncommunity Water Systems (TNCWSs) - GW
<100
-18%
17%
101-500
-23%
17%
501-1,000
-20%
4%
1,001-4,100
-37%
-37%
Totals
-19%
16%
Grand Total
-23%
1%
Source:
Exhibit 6.5. Percentages may not match results calculated directly from
exhibit 6.5 due to rounding in exhibit 6.5.
Notes:
1) Monitoring activates under RTCR begin in year 4 after rule
promulgation. Therefore to represent the average annual acute violations
occurring over the 25-year modeled time period, the total number of
violations is divided by22 years where monitoring activities are conducted.
2) Negative changes indicate reductions in the numberof acute violations
under either option in comparison to the 1989 TCR; positive changes
indicate increases in acute violations. As described further in Section 6.4,
a net increase in acute violations is caused by increased diagnostic power
from increased sampling.
Systems serving >4,100 were omitted from this table because there were
no changes predicted in the numberofviolations to be incurred.
Economic Analysis for the Final RTCR
6-26
September 2012
-------
6.3.1.3 Overall Assessment of Predictive Analysis Results
For PWSs serving <4,100 people, EPA expects risk to decrease under the RTCR and
Alternative option as compared to the 1989 TCR. As shown by the year by year detail of the
output provided in Appendix A of this EA and summarized in Exhibits 5.16-5.21, risk is further
reduced for PWSs under the Alternative option relative to baseline (the 1989 TCR) for the first 5
years after the effective date as compared to the RTCR relative to baseline. This additional
reduction in risk under the Alternative option is driven by the increase in additional sampling
that would occur while all systems were still on monthly sampling. Over time, some of these
systems could qualify to move from monthly to quarterly under either the RTCR or the
Alternative option, or from quarterly to annual sampling under the RTCR. Unlike the RTCR, no
systems under the Alternative option would be able to sample annually. However, additional
costs are incurred under the Alternative option for this increased monitoring, especially for the
systems most impacted by the RTCR (TNCWSs serving <500 people). Chapter 7 presents the
full discussion of costs associated with implementation of the regulatory options considered.
When considering the period beginning in Year 9 after RTCR promulgation through the
end of the period of analysis (model year 11 in Exhibits 5.16-5.21), the RTCR and Alternative
option generally have similar estimates of occurrence. Year 9 after RTCR promulgation
represents the first year of the steady state for monitoring regimens, when all PWSs that qualified
for reduced monitoring are following their new regimens. However, for some categories and
sizes of PWSs the RTCR actually has a lower rate of occurrence than the Alternative option.
This may occur because states under the RTCR may have more resources available to perform
the annual site visits at more PWSs based on states needing fewer resources for activities related
to monitoring (tracking, compliance) than under the Alternative option. Appendix B includes
graphs of predicted occurrence for each of the size and PWS categories considered in this
analysis.
EPA does not expect changes in risk for the PWSs serving >4,100 people under the
RTCR to be as significant as they are for the systems serving <4,100 people. Systems serving
>4,100 people are starting from a smaller baseline level of occurrence than systems serving
<4,100 people, and have a correspondingly lower level of triggered assessment and correction
action activity. This suggests that such small percentage increases in these activities will result in
relatively less change in risk, although the expected change is still a reduction in risk.
Additionally, monitoring requirements for PWSs serving >4,100 people would remain
essentially unchanged under either the RTCR or the Alternative option as compared to the 1989
TCR. Thus the observed overall net increase in benefits (and costs) for PWSs serving >4,100
people is driven by the requirements to conduct assessments and to correct any sanitary defects
that are found. The increase of 10% from baseline in effective corrective actions implemented
applies to systems serving >4,100 people just as it does to systems serving <4,100 people;
however, as shown in Chapter 4 (Exhibit 4.10), the systems serving >4,100 people have a much
smaller level of violations than systems serving <4,100 people, based on reasons discussed in
Chapter 5 (Section 5.4). Therefore, increases in corrective actions will be less evident in absolute
(not percentage) terms.
Economic Analysis for the Final RTCR
6-27
September 2012
-------
6.4 Uncertainty and Sensitivity Analyses
Key sources of uncertainty in the estimates provided in this EA include the data used to
develop baseline estimates and the assumptions made regarding model input variables, as
described in Chapters 4 and 5 of this EA, respectively. The quality and representativeness of the
data used in this EA are discussed in Section 4.2, while uncertainty in model input variables is
discussed in Section 5.3.3.1. A summary of assumptions made in developing the baseline and
input parameters to the predictive model that contribute uncertainty to the analysis are included
in Chapters 4 and 5 (Exhibits 4.13 and 5.22a-b). By definition, those uncertainty factors that are
incorporated into the baseline will have a similar effect on the 1989 TCR, the RTCR and
Alternative option; therefore, EPA believes that they will not significantly affect the net results
of the EA. These assumptions are shown in Ex. 5.22a, and include the estimates representing
GWR effects. Although EPA believes its GWR efficacy estimates are conservative, the estimates
are applied in similar fashion to the baseline (1989 TCR) and RTCR and Alternative option;
therefore, any bias that could be introduced is essentially canceled out in the net analysis. By
contrast, the assumptions listed in Ex. 5.22b include those assumptions that will affect only the
RTCR and Alternative option, such as the assumed increase in the number of effective corrective
actions implemented as compared to the 1989 TCR. These types of assumptions are not canceled
out in the net analysis, and are expected to have some effect on net results. Therefore, EPA has
identified the key drivers of the analysis among this type of assumption and performed a
sensitivity analysis on their values (Chapter 5, Section 5.3.3.1).
This discussion focuses on the provisions of the RTCR and Alternative option for
conducting corrective actions based on the results of Level 1 and Level 2 assessments performed,
and the reductions in repeat samples following TC+ samples and in the additional routine
samples required in the month following a TC+. The analyses performed and presented here are
intended to provide insight into the overall impact of these two changes in rule provisions under
the RTCR, each of which moves risk (as defined in this EA) in opposite directions. Unlike the
uncertainty analysis presented in Chapter 5 (Occurrence and Predictive Model), this analysis
does not consider variations on the assumptions related to corrective actions, but instead tests the
relative impact of changes in the sampling regimen given the assumptions for corrective actions
applied in the primary analysis of this EA (and summarized in Section 5.3).
The primary benefit of the RTCR is a potential reduction in exposure to microbial
contaminants from drinking water provided by PWSs. Two features of the RTCR are expected to
influence the exposure reduction.
The first feature is the requirement to perform corrective actions based on the results of
the Level 1 and Level 2 assessments performed in response to their individual triggers.
Implementation of additional corrective actions beyond the level currently implemented under
the 1989 TCR will reduce exposure to microbial contamination both by addressing the
immediate problem identified by the Level 1 or Level 2 assessment and by preventing some
additional future exposures to fecal contamination.
The second feature, which primarily affects the smaller systems, is the reduction in the
numbers of additional routine and repeat samples that systems are required to take whenever
routine samples are found to be TC+. This reduction in sampling may contribute to increased
Economic Analysis for the Final RTCR
6-28
September 2012
-------
exposure and risk because fewer samples provide fewer opportunities to identify and address TC
and E. coli exposures.
The analyses that were performed used the predictive model with the occurrence input
parameters for the nondisinfecting GW TNCWS serving <101 people. This set of systems was
chosen because: 1) it contains the largest number of systems of the 27 occurrence sets (46,642
systems in this set); 2) it is subject to all of the changes in sampling requirements; and 3) it has
the highest routine TC hit rate so that effects on this set of systems are likely to be more
pronounced (and therefore more clearly observable) than other sets. The analysis was carried out
by running simulations of 10,000 systems each for monthly, quarterly, or annual monitoring
schedules.
There were five sampling and corrective action "regimens" considered in the model. As
shown in Exhibit 6.7, Regimens 1 through 3 assume a routine sampling regimen equivalent to
the 1989 TCR (i.e., 1 regular routine sample and a minimum of five routine samples in the month
following a TC+). Regimen 1 is the only one of the five regimens that assumes there are no
corrective actions (in addition to those already conducted under the 1989 TCR); Regimens 2
through 5 all assume corrective actions are conducted in accordance with requirements under the
RTCR. Regimens 3 through 5 decrease the repeat samples from four to three. Regimen 4 also
decreases the minimum number of routine samples in the month following a TC+ from five to
three; Regimen 5 uses the sampling requirements of the RTCR where systems doing monthly
sampling need only take their one regular routine sample in a month following a TC+, while for
those that monitor quarterly and annually, the requirement remains for a minimum of three
additional routine samples in the month following a TC+.
In this analysis, Regimen 1 describes the 1989 TCR and Regimen 5 describes the RTCR.
Going from Regimen 1 to Regimen 2, where corrective actions are brought in without any
change in sampling, provides key insights to the benefits (reduction in exposure) derived from
the corrective action aspect of the RTCR. Going from Regimen 2 through Regimen 5 provides
some insight into how much of those corrective action benefits might be foregone because of the
small reductions in the number of TC samples required.
The metric used for comparing the relative impacts of these five regimens was the
number of acute violations based on EC+ assays (referred to as "acutes" throughout this section)
because this was considered to be the most relevant measure of the potential microbial health
risk-based benefits of the RTCR. The numbers presented are the average annual numbers of
acute violations.
It is important to note that in order to isolate the effects of the RTCR corrective actions
and sampling changes, this analysis excluded any effect of the GWR. To further understand the
effects of sampling regimen, this analysis also presents the results separately for monthly,
quarterly, and annual sampling since one of the changes in the RTCR applies only to those on
monthly sampling. Exhibit 6.7 provides a summary of the results of this analysis.
Economic Analysis for the Final RTCR
6-29
September 2012
-------
Exhibit 6.7 Relative Impacts Analysis for TNCWS Serving <101 People
for the Range of Corrective Action and Sampling Regimens Predicted
Regimen
Minimum Samples in
Month
Following a TC+
Repeat
Samples
L1 & L2
CAs?
Average Total
TC Samples per Year
Average Acutes Per Year
Found
Average Acutes Per Year
Prevented
Average Acutes Per Year
Missed
M
Q
A
M
Q
A
M
Q
A
M
Q
A
M
Q
A
1
5
5
5
4
No
168,252
60,980
15,317
279
106
25
2
5
5
5
4
Yes
160,394
59,693
15,103
198
92
26
80
14
-1
0
0
0
3
5
5
5
3
Yes
156,779
57,851
14,894
172
79
25
70
12
-1
37
14
2
4
3
3
3
3
Yes
151,500
55,411
14,031
189
82
21
77
13
-1
13
11
5
5
1
3
3
3
Yes
146,129
175
71
32
Ratios of Prevented to Found:
0.41
0.15
-0.05
Key: (CA) = correction action; L1 and L2 = Level 1 and Level 2; M= monthly; Q = quarterly and A = annual.
Definitions: Acutes "found" are acute violations the model predicts will be identified; acutes "prevented" are those avoided by implementation of CAs; acutes "missed" are those found under the
sampling regimen of the 1989 TCR option (Regimen 1) that were not found under the reduced sampling of Regimens 3, 4, and 5 (calculated by subtracting acutes found under Regimens 3, 4,
and 5 from those found under Regimen 1).
Notes:
Results shown are averages based on simulations of 10,000 systems each for monthly quarterly, and annual monitoring using occurrence inputs for nondisinfecting GW TNCWS serving <101
people.
Assumptions for L1 and L2 CAefficacyand resulting duration of reduced occurrence are those used in the primaryanalysis.
For the purpose of isolating the relative impacts of changes in monitoring regimens and implementation of L1 and L2 CAs, no GWR effects (including GWR corrective actions) are modeled in
these runs.
Ratios of acutes "prevented" to those "found" are based on Regimen 2 results, which reflect the 1989 TCR option sampling regimen but includes implementation of L1 and L2 CAs.
Regimen 1, representing the 1989 TCR, provides a baseline against which the other
regimens can be compared. Although arguably more acute violations could be found if more
routine and repeat samples were taken, the numbers of acute violations shown here as "found"
(monthly = 279, quarterly = 106, and annual = 25) represent the maximum number that can be
found given the amount of sampling done. This finding is based on analysis using 10,000
simulated systems each for monthly, quarterly, and annual; the sampling scheme under the 1989
TCR; and no additional corrective actions being performed.
Regimen 2 uses the exact same sampling scheme as Regimen 1, but includes the
performance of Level 1 and Level 2 corrective actions consistent with the RTCR. Here there is a
reduction observed in the number of acutes found for monthly and quarterly sampling although
not for annual sampling. Since there is no change in the sampling requirements, differences (and
similarities) between Regimen 1 and Regimen 2 numbers are due to: a) the implementation of
corrective actions under Regimen 2; and b) and random variation that is endemic to the Monte
Carlo simulation.
For monthly sampling, which requires a large number of TC samples per year, Monte
Carlo variation is relatively small with respect to the effect of the parameter adjustments.
Approximately 80 acute violations occur under monthly sampling for Regimen 1 (279 acutes)
that do not occur under Regimen 2 (198 acutes) as the result of corrective actions. Quarterly and
annual sampling regimens require far fewer TC samples per year, and therefore the Monte Carlo
variation may obscure the salient data trends in those model results. For instance, the increase in
the number of acutes under annual sampling under Regimen 1 from 25 to 26 under Regimen 2,
which is not consistent with the trend seen in the larger sample size in the simulation of monthly
systems, is likely to be reflective of stochastic noise from the Monte Carlo simulation performed
rather than of the impact of corrective actions. For this reason most of the discussion that follows
on the trends across these regimens focuses on the monthly sampling results.
Regimen 3 reflects the reduction in repeat sampling requirements of four to three
following a TC+. As would be expected, this also results in a small reduction in the number of
acutes found. For example, in the monthly sampling group this falls from 198 to 172. Using an
Economic Analysis for the Final RTCR
6-30
September 2012
-------
assumption that the ratio of acutes prevented to acutes found in Regimen 2 apply to the other
regimens as well, then results for the monthly sampling group (where 80/198 = 0.41 for Regimen
2) indicate that 70 acutes are still prevented under Regimen 3. The sum of those found under
Regimen 3 (172) and those prevented under Regimen 3 (70) for the group of systems on monthly
sampling is 242. This implies, then, that as a result of the reduction in repeat sampling from four
to three, a total of 37 acutes are missed under Regimen 3. 279 acutes are found under the 1989
TCR sampling regimen (Regimen 1); therefore, the number of acutes missed is calculated as 279
- (172+70) = 37. That is, these 37 acute violations are neither found nor prevented under
Regimen 3, but are assumed to occur unobserved.
Regimen 4 reduces the number of routines in a month following a TC+ from five to three,
but retains the three repeats and the corrective action requirements. The number of acutes now
"missed" changes to 13, 11, and 5 respectively for those on monthly, quarterly, and annual
sampling.
Regimen 5, which represents the RTCR, excludes any additional routine samples in the
month following a TC+ for those on monthly sampling. Here the number of "missed" acutes for
those on monthly increases from 13 to 32. There are no changes for those on quarterly or annual
sampling relative to Regimen 4 since systems sampling quarterly and annually are treated
identically in both scenarios.
In addition to displaying the changes in the number of annual average acutes for these
five regimens, Exhibit 6.7 also shows the average annual number of TC samples (regular routine,
additional routines and repeat samples) taken. The number of samples declines considerably
across the regimens due both to the reduction in actual sampling requirements and to reduced
number of additional routines and repeats that need to be taken because of the prevention of TC
occurrence by the corrective actions taken.
This analysis points first and foremost to the highly positive public health benefits of
including the corrective actions as part of the RTCR (based on the modest assumptions regarding
their effectiveness used in the occurrence analysis detail in Chapter 5 of this EA). This is seen
not only in the number of acutes that are found for which corrective actions may be performed,
but also—and perhaps more importantly—in the large number of additional acutes that are
prevented from occurring again in those systems as a result of the corrective actions.
While some of these potential benefits are "missed" as a result of the reductions in
additional routine and repeat sampling requirements, these numbers (shown in Exhibit 6.7) seem
small when compared to the numbers found and prevented. This is particularly important when
considering the potential cost savings from reducing the number of TC samples that are taken, as
discussed in Chapter 9 (Net Benefits) of this EA.
The group of 61,539 TNCWSs serving <100 currently has, and will continue to have, a
vast majority of systems on either quarterly or annual sampling in the steady state of the analysis
period, beginning in approximately Year 9 after RTCR promulgation. Since quarterly and annual
estimates of acutes are less stable, predictions for this specific subset of systems are less reliable.
Nonetheless, since approximately 40% of systems affected by the RTCR are on monthly
sampling, the trends in the monthly numbers cited above should be indicative of trends for other
Economic Analysis for the Final RTCR
6-31
September 2012
-------
types and sizes of systems included in this model (those serving <4,100 people). That is, relative
to the 1989 TCR, the inclusion of the Level 1 and Level 2 corrective actions will always result in
more benefits accruing from corrective actions being performed and acute violations being
avoided than the relatively small reductions in sample numbers will result in "missing" acute
violations.
Because all of the other system types have lower overall TC and E. coli hit rates than the
nondisinfecting GW TNCWS serving <101 people, on a per system basis their missed acute
violations would be fewer than those estimated for the set of systems addressed in this analysis.
Furthermore, for the systems serving 1,001 to 4,100 people where the RTCR change in the
minimum number of samples to be taken in the month following a TC+ is smaller, the number of
missed acute violations would be less than that seen in this analysis for the systems serving <101
people. For example, for systems serving 3,301 to 4,100 (which are all on monthly sampling),
the minimum number of next month samples is reduced from five to four compared with a
reduction from five to one for the systems serving fewer than 1,001 people that perform monthly
sampling. Thus this analysis presents an estimate that is near the higher end of the range for
various system types in expected missed acute violations due to reduced sampling requirements
under the RTCR. In spite of this conservative estimate, the effect of the RTCR requirements for
corrective actions more than balances this effect with a larger change in risk in the opposite
(reducing) direction, as shown by the relatively large number of prevented to missed acute
violations.
6.5 Other Potential Benefits
A number of other benefits may accrue to PWSs and their populations served that are not
included in the qualitative relative risk comparison or reductions in occurrence discussed in
Sections 6.1-6.4 (above), they are described in the following sections.
6.5.1 Increased System Knowledge
By requiring additional assessments focused on isolating and identifying system
problems in response to TC+ or EC+ samples, the RTCR will increase the likelihood that PWS
operators, in particular those of systems triggered to conduct assessments, will develop further
general understanding of system operations and potential issues. This heightened familiarity with
the system may increase preventive maintenance, or may increase the efficiency with which
future problems are identified, decreasing risk to the PWS population served in both cases.
Delaying system component replacement costs or avoiding an increase in treatment costs may
also result from this increased knowledge and result in cost savings for some communities.
6.5.2 Accelerated Infrastructure Repair/Replacement
As described in 6.5.1 (above), the increased familiarity of operators with their systems
may encourage an increase in preventive maintenance, preempting some potential contamination
issues and decreasing risk for the PWS population served. Some systems may see additional non-
quantified benefits associated with the acceleration of their capital replacement fund investments
in response to early identification of impending problems with large capital components.
Although such capital investment would have occurred anyway, earlier investment may ensure
Economic Analysis for the Final RTCR
6-32
September 2012
-------
that problems are addressed in a preventive manner and may preclude some decrease in
protection that may have occurred otherwise. At the very least, the increased operator awareness
should reduce the occurrence of unplanned capital expenditures in any given year.
6.5.3 Reduction in Averting Behavior
As the risk of contamination is reduced in PWSs over time following RTCR
promulgation, EPA expects that the public will eventually become aware of increased quality,
and consistency of quality, of their water. This may lead to PWS customers becoming
increasingly comfortable with drinking tap water and cause them to exercise less averting
behavior (e.g., drinking bottled water or using point of use filtration devices). Such averting
behaviors are costly relative to consuming tap water.
6.5.4 Reduction of Co-Occurring and Other Contaminants
In addition to the specific E. coli endpoint targeted by the RTCR, there are many
potentially co-occurring and emerging pathogens (such as viruses, parasitic protozoa and/or other
bacteria) that may be avoided as part of any avoided contamination event. To the extent that E.
coli co-occurs with pathogens sufficiently in abundance to result in health effects, the RTCR
offers the potential for additional morbidity and mortality prevention.
Potential benefits from the RTCR include avoidance of a full range of health effects,
including acute and chronic illness, endemic and epidemic disease, associated outbreaks and
death that may occur from the consumption of fecally contaminated drinking water. Also, since
fecal contamination may contain waterborne pathogens including bacteria, viruses, and parasitic
protozoa, in general, a reduction in fecal contamination should also reduce the risk from these
other contaminants.
Systems may choose corrective actions that also address other drinking water
contaminants. For example, correcting for a pathway of potential contamination into the
distribution system can mitigate a variety of potential contaminants. Due to a lack of
contamination co-occurrence data that quantify the effect that treatment corrective action may
have on contamination entering through distribution system pathways, EPA has not quantified
such potential benefits.
6.5.5 Reduction in Outbreak Risk and Response Costs
Besides reducing the endemic risk of illnesses from waterborne pathogens, the RTCR
would reduce the likelihood of major outbreaks from occurring. These avoided illnesses and
other costs are not estimated or included in the RTCR analyses and would be difficult to
quantify. The economic value of reducing the risk of outbreaks could be quite high when the
magnitude of potential costs is considered. The Agency was unable to quantify or monetize the
cost associated with acute and chronic illnesses or death acquired from consuming water
contaminated with waterborne pathogens because cases avoided could not be calculated, as
described in Section 6.2. Examples of potential illnesses associated with ingestion of waterborne
pathogens and their potential costs are described in Chapter 2 of this EA.
Economic Analysis for the Final RTCR
6-33
September 2012
-------
Other types of costs associated with outbreaks include spending by local, state, and
national public health agencies; emergency corrective actions by utilities; and possible legal
costs if liability is a factor. Affected water systems and local governments may incur costs
through provisions of alternative water supplies and issuing customer water use warnings and
health alerts. Commercial establishments (e.g., restaurants) and their customers may incur costs
due to interrupted and lost service. Local businesses, institutions, and households may incur costs
associated with undertaking averting and defensive actions. Thus, to the extent that the RTCR
reduces the likelihood of waterborne disease outbreaks, avoided response costs are potentially
numerous and significant. For example, an analysis of the economic impacts of a waterborne
disease outbreak in Walkerton, Ontario (population 5,000) estimated the economic impact
(excluding estimates of the value of a statistical life for seven deaths and intangible costs for
47
illness-related suffering) to be over $45.9 million in 2007 Canadian dollars (approximately 42.8
48
million 2007 U.S. dollars) (Livernois, 2002). Note that some of these costs were incurred by
individuals and businesses in neighboring communities. The author believed that this was a
conservative estimate.
47 Households and businesses in the town of Walkerton were unable to use municipal water for eight months
following the contamination event. The response involved a massive effort at all levels of government in terms of
public health response and investigation of the cause. The paper asserts that the impact of the crisis extended beyond
Walkerton to nearby towns and the countryside, resulting in economic costs that the author expects were at a
minimum equal to the costs presented in the paper (of which only the non-medical costs are included here).
48 Updated from $43 million in Canadian dollars, $32 million U.S. dollars in 2000. Costs updated from 2000 dollars
to 2007 dollars using the Canadian core CPI (= 95.71667 109.8167 = 1.14731) and the 2007 exchange rate
(1.0734 Canadian dollars /1.0 U.S. dollars).
Economic Analysis for the Final RTCR
6-34
September 2012
-------
7 Cost Analysis
7.1 Introduction
This chapter presents estimates of the total national and household costs of the Revised
Total Coliform Rule (RTCR). To estimate the national costs of the RTCR, the United States
Environmental Protection Agency (EPA or Agency) calculated the net change in costs (i.e.,
incremental costs over the 1989 Total Coliform Rule (1989 TCR)) of rule components associated
with state49 practices and system activities required under the revised rule. The remainder of this
chapter provides detailed discussion of the methodology used and results from the cost analyses
and is organized as follows:
• Section 7.2 describes the cost model and general costing and compliance
assumptions used to estimate national costs of the RTCR.
• Section 7.3 describes the methodology of projecting costs over a 25-year period
(discounted at 3 and 7 percent, respectively) according to the RTCR compliance
schedule, estimating the present value of each cost, and annualizing each over a 25-
year period.
• Section 7.4 describes the methodology for developing costs for all rule activities.
• Section 7.5 presents household cost estimates.
• Section 7.6 presents a discussion of non-quantified costs.
• Section 7.7 presents a discussion of uncertainties in cost estimates.
• Section 7.8 presents a comparison of cost estimates for all regulatory options.
7.2 General Cost Assumptions and Methodology
The RTCR cost model builds on the baseline data, occurrence analysis, and benefits
model results described in Chapters 4-6. Based on these analyses, the annual and cumulative
numbers of public water systems (PWSs) that would be required to comply with each rule
component of the RTCR over the 25-year compliance period are provided in Appendix A
(Exhibits 6.2-6.4 provide a summary of this information). In general, the numbers of PWSs
presented in Appendix A are multiplied by the unit cost assumptions described in this chapter to
calculate total annual costs.
There are also several general costing assumptions that are unique to the costing process
and are used as inputs to the cost model. The derivation of these inputs is discussed in detail
below.
49 The term "state" in the context of this chapter refers to any state or other primacy agency that has oversight
authority for drinking water programs.
Economic Analysis for the Final RTCR
7-1
September 2012
-------
7.2.1 Labor Rates
For costing purposes, EPA estimates the labor needs and hourly labor rates of PWSs and
states. EPA recognizes that there may be significant variation in labor rates across all PWSs.
However, for purposes of this economic analysis (EA), and to implement national policy, EPA
uses national-level estimates from Labor Costs for National Drinking Water Rules (USEPA,
2003a) (as used in the Ground Water Rule Economic Analysis (USEPA, 2006a)). For the RTCR
cost analyses, these labor rates were inflated to 2007$ using the Employee Cost Index (ECI), and
weighted based on the PWS size categories used in the RTCR EA. To account for the general
composition of staff at PWSs of smaller sizes (e.g., PWSs serving 3,300 or fewer), EPA uses
only the technical rate. For PWSs serving more than 3,300 people, EPA uses a ratio of 80 percent
technical labor to 20 percent managerial labor to arrive at a labor cost, or weighted labor rate.
The actual ratio between technical and managerial rates employed may vary by PWS and among
the different compliance activities under the RTCR. However, for simplicity, the 80/20 ratio is
used as a general assumption for costing purposes in this EA. A full description of the derivation
of the labor rates used is provided in the Technology and Cost Document for the Final Revised
Total Coliform Rule (USEPA, 2010d). The weighted labor rates ($2007) are shown in Exhibit
7.1.
Exhibit 7.1 Labor Rates by PWS Size (2007$)
PWS Size (population served)
Weighted Labor Rate ($/hour)
< 100
$ 25.10
101 -500
$ 27.03
501 - 1,000
$ 28.96
1,001 -4,100
$ 29.73
4,101 -33,000
$ 36.00
33,001 - 96,000
$ 36.39
96,001 - 500,000
$ 41.01
500,001-1 Million
$ 41.01
> 1 Million
$ 41.01
Note: Labor rates for each size category are assumed to be the same
regardless of system type (CWS, NTNCWS, and TNCWS).
Source: Final RTCR T&C Document
For states, the administrative and field engineer labor rates from the 2001 State Drinking
Water Needs Analysis (ASDWA, 2001) are used in the RTCR EA (as used in the Ground Water
Rule (GWR) EA (USEPA, 2006a)). These rates include a 60 percent overhead rate and were
inflated to 2007$ using the ECI. EPA recognizes that there may be significant variation in labor
rates across all states. The state labor rates in 2007$ are $39.22 for an administrative state
employee and $43.58 for a state field engineer. EPA assumes that the state field engineer would
conduct annual site visits,50 and the administrative state employee would work with PWSs on all
remaining aspects of the RTCR. Because this separation between field engineer and
50 Because of the high cost for an annual site visit by a state, for this analysis, EPA assumes that no states would
choose to conduct annual site visits unless they already do so under the 1989 TCR. See section 7.4.4 for additional
information.
Economic Analysis for the Final RTCR
7-2
September 2012
-------
administrative employee is used, the 80/20 weighting ratio between technical and managerial
rates is not used to develop state costs.
7.2.2 TCR Monitoring Costs per Sample
A cost per sample is associated with distribution system monitoring. For the purpose of
this cost analysis, PWSs would perform total coliform (TC) monitoring, supplemented by E. coli
analyses as required. EPA estimated the sample analysis cost for both in-house and commercial
laboratory analysis. The weighted unit costs for monitoring provided in Exhibit 7.2 are based on
the percentage of PWSs conducting in-house and commercial laboratory analysis based on
conversations of the Total Coliform Rule Distribution System Advisory Committee (TCRDSAC)
Technical Workgroup (TWG). For in-house sample analysis, the estimated burden includes
sample collection and analysis and also accounts for operations and maintenance (O&M) costs
such as equipment and maintenance. For commercial laboratory analysis, the estimated burden
includes sample collection, shipping and delivery, and the laboratory analysis fee. These
estimates reflect a national average; however, individual PWSs may realize collection burden
that is either less than or greater than this average depending on the locations of sampling points
in a particular PWS.
Rates may vary due to regional variations in laboratory fees, the number of samples
processed (quantity discounts), and laboratory capacity. As shown in Exhibit 7.2, the cost per
sample decreases as more samples are taken, and as PWSs take advantage of savings from bulk
shipping. A full description of the derivation of the monitoring costs per sample is provided in
the Technology and Cost Document for the Final Revised Total Coliform Rule (USEPA, 2010d).
Exhibit 7.2 Monitoring Costs per Sample (2007$)
PWS Size
(population served)
Numbers of Samples Taken and Delivered at the Same Time
1
2
3
4
> 5
< 100
$ 50.54
$ 42.64
$ 40.01
$ 38.69
$ 37.90
101 -500
$ 51.55
$ 43.63
$ 40.99
$ 39.67
$ 38.87
501 - 1,000
$ 59.81
$ 51.86
$ 49.21
$ 47.88
$ 47.09
1,001 -4,100
$ 60.40
$ 52.45
$ 49.79
$ 48.47
$ 47.67
4,101 -33,000
$ 65.26
$ 57.23
$ 54.55
$ 53.21
$ 52.40
33,001 - 96,000
$ 60.57
$ 56.55
$ 55.21
$ 54.54
$ 54.14
96,001 - 500,000
$ 72.38
$ 71.57
$ 71.30
$ 71.17
$ 71.09
500,001-1 Million
$ 72.38
$ 71.57
$ 71.30
$ 71.17
$ 71.09
> 1 Million
$ 72.38
$ 71.57
$ 71.30
$ 71.17
$ 71.09
Note: Per sample monitoring costs for each size category are assumed to be the same regardless
of PWS type (CWS, NTNCWS, and TNCWS).
Source: Final RTCR T&C Document
7.2.3 Technology Unit Costs and Compliance Forecasts
EPA has assumed that PWSs may use a variety of existing best management practices
(BMPs) and technologies to address distribution system deficiencies discovered during Level 1
Economic Analysis for the Final RTCR
7-3
September 2012
-------
and Level 2 assessments. These BMPs and technologies include a combination of flushing
programs, training of personnel to collect samples, replacing valves/pipes/hydrants/meters,
installing new water mains, modifying operation of storage facilities, booster disinfection,
physical security devices, etc. For a full list of technologies and BMPs that are anticipated to be
used to meet rule requirements, see Appendix D. EPA estimated unit costs for these various
components using equipment price lists and quotes, costs associated with BMPs from PWSs,
engineering cost data sources (e.g., Means, 1998), consultations with the TWG supporting the
TCRDSAC Federal Advisory Committee (FAC), and other relevant assumptions used in
economic analyses performed for existing drinking water rules (e.g., GWR). Detailed
explanations of the unit cost derivations for these BMPs and technologies are presented in the
Technology and Cost Document for the Final Revised Total Coliform Rule (USEPA, 2010d).
Compliance forecasts (or technology selection forecasts) are estimates of which
technologies PWSs undergoing corrective action would use. Section 7.4.6 provides details on
compliance forecasts for PWSs performing corrective actions based on Level 1 and Level 2
assessments.
7.2.4 Cost Model
National costs are estimated using a cost model specifically developed for the RTCR. The
model builds on the occurrence model described in Chapter 5. Within the modeling structure,
costs for PWSs serving >4,100 retail customers are analyzed differently from smaller PWSs to
capture differing baseline structures and to account for differences in available occurrence data
as described in Chapter 4.
PWS costs are estimated for different PWS types and size categories (nine size categories
are used based on population served, consistent with the Technology and Cost Document for the
Final Revised Total Coliform Rule (USEPA, 2010d)). PWS cost analyses include estimates to
implement the rule; to revise sample siting plans; to conduct routine monitoring, additional
routine monitoring, and repeat monitoring; to perform Level 1 and Level 2 assessments and
implement corrective actions; and to provide public notification (PN). State cost analyses include
estimates of the labor burdens that states would face, including staff training on RTCR
requirements and conducting annual administration, reviewing monitoring reports, reviewing
assessments, reviewing and approving corrective action plans, and recordkeeping. Section 7.4
provides detailed discussion on the underlying cost-buildup for each rule component analyzed
within the cost model.
7.2.5 Modeled Variability and Uncertainty in National Costs
As noted throughout this EA, there is variability among many of the input parameters to
the RTCR cost model and several rule compliance assumptions based on PWS size and type
(e.g., population served, labor rates, TC hit rates, and occurrence distributions are different for
different sizes and types of PWSs). However, there is insufficient information to fully
characterize the distribution of variability (i.e., calculating confidence bounds) within each of
these PWS size and type classifications on a national scale; therefore, EPA uses mean values for
these input parameters.
Economic Analysis for the Final RTCR
7-4
September 2012
-------
EPA also recognizes that there is uncertainty in the national cost estimates, much of
which has the same impact on the modeled results of both the 1989 TCR and RTCR (e.g.,
baseline assumptions and effects of GWR implementation). Because the EA analyses focus on
net changes between the 1989 TCR and the RTCR, these common sources of uncertainty cancel
each other out in the net change analyses. For assumptions that are major drivers of the analysis
and differ between the 1989 TCR and RTCR (e.g., corrective action compliance forecast), EPA
has evaluated uncertainty and performed sensitivity analyses to qualitatively and quantitatively
characterize the potential impacts of alternative input parameters. Chapter 5 discusses
uncertainty and presents sensitivity analyses pertaining to the predictive occurrence model
results, which also impact the cost calculations. Section 7.7 discusses uncertainty and provides
sensitivity analysis results as they specifically pertain to the cost analyses.
7.3 Projecting and Discounting National Costs
Costs must be expressed in common units so they can be added together to calculate total
annual costs and compared to benefits. For the RTCR, the performance of activities varies over
time in response to regulatory requirements and monitoring results. To compare the values of
performing these activities, the year or years in which all costs are expended must be determined
and the costs must be calculated as a net present value. For the purposes of this EA, one-time and
yearly costs were projected over a 25-year time period to allow comparison with other drinking
water regulations using the same analysis period. The net present values of costs are calculated
using discount rates of 3 and 7 percent based on EPA policy and guidance from the Office of
Information and Regulatory Affairs of the Office of Management and Budget (OMB).51 A
summary of the steps used in making adjustments to the national-level costs presented in this EA
is as follows:
• Estimate all costs (noncorrective action, corrective action, and state) over a 25-year
time horizon based on the rule implementation schedule.
• Calculate total net present value costs using 3 and 7 percent discount rates.
• Annualize the costs over 25 years using the same discount rates.
Appendix C presents step-by-step results for the projection and discounting of costs to
show how yearly costs for each rule component are accounted for by the cost model for
community water systems (CWSs), nontransient noncommunity water systems (NTNCWSs),
transient noncommunity water systems (TNCWSs), and states. Exhibits C.l through C.9 show
the nominal costs projected over the rule schedule and the present value of each cost calculated
to the expected year of rule implementation for the 1989 TCR. Exhibits C. 10 through C.45 show
the results for the RTCR and the Alternative option.
51 The choice of an appropriate discount rate is a complex and controversial issue among economists and policy
makers. Therefore, the Agency compares streams of future national level costs and benefits using two alternative
discount rates, 3 and 7 percent. The underlying logic for each discount rate can be found in Guidelines for Preparing
Economic Analyses (USEPA, 2000).
Economic Analysis for the Final RTCR
7-5
September 2012
-------
7.4 Derivation of Costs for PWSs and States
This section presents the methodology and unit costs used to derive national costs for
PWSs and states to perform 1989 TCR- and RTCR-related activities. Chapter 3 contains detailed
summaries of the activities under the 1989 TCR, RTCR, and Alternative option. The following
subsections provide a brief summary of each activity and the assumptions used to estimate the
burden and costs attributable to both PWSs and states for:
7.4.1 Rule Implementation and Annual Administration
7.4.2 Revise Sample Siting Plan
7.4.3 Monitoring
7.4.4 Annual Site Visits
7.4.5 Assessments
7.4.6 Corrective Actions
7.4.7 Public Notification
This chapter uses information from the baseline analysis in Chapter 4 as a starting point
for analysis of PWSs subject to each rule requirement. Additional baseline information and
detailed intermediate model outputs are provided in Appendix A. There are also 57 states and
primacy agencies that would incur costs as a result of the rule.
7.4.1 Rule Implementation and Annual Administration
PWSs
Under the RTCR and Alternative option, all PWSs subject to the rule would incur one-
time costs that include time for staff to read the rule and become familiar with its provisions and
to train employees on rule requirements. No additional implementation burden or costs are
incurred by PWSs to implement the 1989 TCR, as these PWSs have already performed
implementation and are continuing to perform annual administration activities under the 1989
TCR. Under the RTCR and Alternative option, all PWSs subject to the RTCR would perform
additional or transitional implementation activities. The labor rates presented in Section 7.2.1 are
used along with estimates of labor hours to generate estimated implementation costs for all
PWSs. Based on previous experience with rule implementation and consistent with estimates
used in the GWR EA, EPA estimates that PWSs would require a total of 4 hours to read and
understand the rule, and a total of 8 hours to plan and mobilize (i.e., assign appropriate personnel
and resources to carry out rule activities).
Economic Analysis for the Final RTCR
7-6
September 2012
-------
Exhibit 7.3 Net Change in PWS Unit Burden and Cost Estimates for Rule
Implementation (2007$)
Read and
Understand
Planning and
PWS Size
Labor Cost
Rule
Mobilization
(Population
(per hour)
(hours/system)
(hours/system)
Unit Cost
Served)
A
B
C
D=A*(B+C)
Community Water Systems (CWSs)
<100
$ 25.10
4.0
8.0
$ 301.20
101-500
$ 27.03
4.0
8.0
$ 324.36
501-1,000
$ 28.96
4.0
8.0
$ 347.52
1,001-4,100
$ 29.73
4.0
8.0
$ 356.76
4,101-33,000
$ 36.00
4.0
8.0
$ 432.00
33,001-96,000
$ 36.39
4.0
8.0
$ 436.68
96,001-500,000
$ 41.01
4.0
8.0
$ 492.12
500,001-1 Million
$ 41.01
4.0
8.0
$ 492.12
> 1 Million
$ 41.01
4.0
8.0
$ 492.12
Nontransient Noncommunity Water Systems (NTNCWSs)
<100
$ 25.10
4.0
8.0
$ 301.20
101-500
$ 27.03
4.0
8.0
$ 324.36
501-1,000
$ 28.96
4.0
8.0
$ 347.52
1,001-4,100
$ 29.73
4.0
8.0
$ 356.76
4,101-33,000
$ 36.00
4.0
8.0
$ 432.00
33,001-96,000
$ 36.39
4.0
8.0
$ 436.68
96,001-500,000
$ 41.01
4.0
8.0
$ 492.12
500,001-1 Million
$ 41.01
4.0
8.0
$ 492.12
> 1 Million
$ 41.01
4.0
8.0
$ 492.12
Transient Noncomm unity Water Systems (TNCWSs
<100
$ 25.10
4.0
8.0
$ 301.20
101-500
$ 27.03
4.0
8.0
$ 324.36
501-1,000
$ 28.96
4.0
8.0
$ 347.52
1,001-4,100
$ 29.73
4.0
8.0
$ 356.76
4,101-33,000
$ 36.00
4.0
8.0
$ 432.00
33,001-96,000
$ 36.39
4.0
8.0
$ 436.68
96,001-500,000
$ 41.01
4.0
8.0
$ 492.12
500,001-1 Million
$ 41.01
4.0
8.0
$ 492.12
> 1 Million
$ 41.01
4.0
8.0
$ 492.12
Notes:
FWS burden and cost estimates for implementation activities are assumed to be identical under the
RTCR and Alternative Option.
Sources:
(A) Labor rates for FWSs from Exhibit 7.1.
(B), (C) EPA estimates based on best professional judgment.
Economic Analysis for the Final RTCR
7-7
September 2012
-------
States
States would incur administrative costs to implement the RTCR. These implementation
costs are not directly required by specific provisions of the RTCR but are necessary for states to
ensure that the provisions of the RTCR are properly carried out. States would need to allocate
time for their staff to establish and maintain the programs necessary to comply with the RTCR,
including developing and adopting state regulations and modifying data management systems to
track new required PWS reports to the states. As a one-time burden covered under the 520 hours
allocated to modifying data management systems, each state would modify their data
management system to be able to track the changes in monitoring regimes. Note, on average no
on-going annual cost is assumed for modifying data management because the tracking system
would already be in place and oversight would be accounted under state review of sampling
results, assessments, etc. Time requirements for a variety of state agency activities and responses
are estimated in this EA. Exhibit 7.4 lists the activities required to revise the program following
promulgation of the RTCR along with their respective costs and burden. Because time
requirements for implementation and annual administration activities vary among state agencies,
EPA recognizes that the burden and cost estimates presented in Exhibit 7.4 may be an over- or
underestimate for some states.
Exhibit 7.4 Net Change in State Unit Burden and Cost Estimates for Rule
Implementation (2007$)
Labor Cost
(per hour)
Hours
FTEs
Cost
Compliance Activity
A
B
C=B/2,080
D=A*B
Read and Understand Rule
$
39.22
15
0.01
$
588
Regulation Adoption and Program Development
$
39.22
260
0.13
$
10,197
Initial Laboratory Certification
$
39.22
-
-
$
-
Modify Data Management Systems
$
39.22
520
0.25
$
20,393
PWS Training and Technical Assistance
$
39.22
520
0.25
$
20,393
Staff Training
$
39.22
130
0.06
$
5,098
Per State Total
1,445
$
56,670
National Totals (57 States/Primacy Agencies)
82,365
$
3,230,201
Notes:
Detail may not add due to independent rounding.
State burden and cost estimates for Implementation activities are assumed to be identical under the RTCR and Alternative
Option.
Sources:
(A) Labor rate for state employee from Section 7.2.1.
(B) Labor hours for start-up activities are based on GWR estimates. Because the RTCR is a revision of the 1989 TCR, one
fourth of the State unit start up burden from GWR is used in the RTCR.
(C) Full-time equivalent (FTE) assumes individual working 40 hours per week, 52 weeks per year.
In addition to these one-time costs, states would use resources to continue administrative
activities. On an annual basis, states must coordinate with their particular EPA regional office to
be certain that the state's program is consistent with federal requirements. States would also
continue to train state and PWS staffs, maintain laboratories' certifications, and report PWS
compliance information to the Safe Drinking Water Information System (SDWIS). However,
Economic Analysis for the Final RTCR
7-8
September 2012
-------
based on discussions with stakeholders, once the initial implementation activities are complete,
the annual burden (on average) for general administrative tasks for the RTCR would not be any
higher than the burden incurred under the 1989 TCR requirements. In some cases, the general
administrative burden for the 1989 TCR may actually decrease as PWSs experience better
performance, and thus require less state interaction, under the RTCR. Therefore, no continuing
annual administrative costs are estimated for the EA.
States would also spend time responding to specific requirements under the RTCR (i.e.,
review assessment reports, consult with PWSs, etc.). In these cases, the state costs are estimated
under the costing for that particular rule requirement.
Implementation Net Cost Summary
Because EPA does not anticipate early implementation of the RTCR, EPA expects that
implementation activities would take place in years 1 through 3 of the 25-year compliance period
before PWSs begin monitoring activities. Annualized cost estimates for PWSs and states to
perform implementation activities are estimated by multiplying the number of PWSs or states
required to comply with the RTCR (i.e., all PWSs) by the unit costs presented in Exhibits 7.3 and
7.4. Total and net change in annualized present value cost estimates for PWSs and states to
perform implementation activities under the 1989 TCR, RTCR, and Alternative option are
presented in Exhibit 7.5.
Exhibit 7.5 Annualized Cost Estimates for Rule Implementation ($Millions, 2007$)
PWSs
State
Total
PWSs
State
Total
3% Discount Rate
7% Discount Rate
1989 TCR-Total
$
$
$
$
$
$
RTCR-Total
$ 2.77
$ 0.18
$ 2.95
$ 4.00
$ 0.26
$ 4.26
RTCR - Net Change
$ 2.77
$ 0.18
$ 2.95
$ 4.00
$ 0.26
$ 4.26
Alternative Option - Total
$ 2.77
$ 0.18
$ 2.95
$ 4.00
$ 0.26
$ 4.26
Alternative Option - Net Change
$ 2.77
$ 0.18
$ 2.95
$ 4.00
$ 0.26
$ 4.26
Notes:
1) Detail may not add due to independent rounding.
2) FWS and state burden and cost estimates for irrplerrentation activities are assurred to be identical under the RTCR and Alternative Option.
Source: Final RTCR Cost Model.
7.4.2 Revise Sample Siting Plan
PWSs
Under the RTCR and Alternative option, all PWSs subject to the RTCR would incur one-
time costs to revise existing sample siting plans to identify sampling locations and collection
schedules that are representative of water throughout the distribution system. System sample
siting plans must include routine and repeat sample sites and any sampling points necessary to
meet GWR requirements. Under the TCR, no additional burden or costs are expected to be
incurred by PWSs to revise sample siting plans, as these PWSs are already collecting TC
samples in accordance with a written sample siting plan. The labor rates presented in Section
7.2.1 are used along with estimates of labor hours to generate sample siting plan costs for all
Economic Analysis for the Final RTCR
7-9
September 2012
-------
PWSs. Based on previous experience, EPA estimates that PWSs would require 2-8 hours for
revising their sample siting plan, depending on PWS size. Estimates of PWS unit costs to revise
sample siting plans are presented in Exhibit 7.6.
States
Under the RTCR and Alternative option, states are expected to incur one-time costs to
review sample siting plans and recommend any revisions to PWSs. Under the 1989 TCR, no
additional burden or costs are incurred by states to review sample siting plans, as these PWSs'
sample siting plans have already been reviewed and approved. State costs are based on the
number of PWSs submitting revised sample siting plans to PWSs each year. The state labor rate
presented in Section 7.2.1, the number of PWSs in each PWS size category required to revise
sample siting plans, and estimates of labor hours are used to generate sample siting plan costs
incurred by states. Based on previous experience, EPA estimates that states would require 1-4
hours to review revised sample siting plans and provide any necessary revisions to PWSs,
depending on PWS size. Estimates of state unit costs to revise sample siting plans are presented
in Exhibit 7.6.
Economic Analysis for the Final RTCR
7-10
September 2012
-------
Exhibit 7.6 Net Change in PWS and State Burden and Cost Estimates to Revise
Sample Siting Plans (2007$)
PWS Size
(Population
Served)
PWSs
States
PWS Labor Cost
(per hour)
Revise Sample
Siting Plan
(hours/system)
Unit Cost
State Labor
Cost (per hour)
Review and
Revise Sample
Siting Plan
(hours/system)
Unit Cost
A
B
C=A*B
D
E
F=D*E
Community Water Systems (CWSs)
<100
$ 25.10
2.0
$ 50.20
$ 39.22
1.0
$ 39.22
101-500
$ 27.03
2.0
$ 54.06
$ 39.22
1.0
$ 39.22
501-1,000
$ 28.96
4.0
$ 115.84
$ 39.22
2.0
$ 78.44
1,001-4,100
$ 29.73
4.0
$ 118.92
$ 39.22
2.0
$ 78.44
4,101-33,000
$ 36.00
6.0
$ 216.00
$ 39.22
3.0
$ 117.65
33,001-96,000
$ 36.39
8.0
$ 291.12
$ 39.22
4.0
$ 156.87
96,001-500,000
$ 41.01
8.0
$ 328.08
$ 39.22
4.0
$ 156.87
500,001-1 Million
$ 41.01
8.0
$ 328.08
$ 39.22
4.0
$ 156.87
> 1 Million
$ 41.01
8.0
$ 328.08
$ 39.22
4.0
$ 156.87
Nontransient Noncommunity Water Systems (NTNCWSs)
<100
$ 25.10
2.0
$ 50.20
$ 39.22
1.0
$ 39.22
101-500
$ 27.03
2.0
$ 54.06
$ 39.22
1.0
$ 39.22
501-1,000
$ 28.96
4.0
$ 115.84
$ 39.22
2.0
$ 78.44
1,001-4,100
$ 29.73
4.0
$ 118.92
$ 39.22
2.0
$ 78.44
4,101-33,000
$ 36.00
6.0
$ 216.00
$ 39.22
3.0
$ 117.65
33,001-96,000
$ 36.39
8.0
$ 291.12
$ 39.22
4.0
$ 156.87
96,001-500,000
$ 41.01
8.0
$ 328.08
$ 39.22
4.0
$ 156.87
500,001-1 Million
$ 41.01
8.0
$ 328.08
$ 39.22
4.0
$ 156.87
> 1 Million
$ 41.01
8.0
$ 328.08
$ 39.22
4.0
$ 156.87
Transient Noncommunity Water Systems (TNCWSs)
<100
$ 25.10
2.0
$ 50.20
$ 39.22
1.0
$ 39.22
101-500
$ 27.03
2.0
$ 54.06
$ 39.22
1.0
$ 39.22
501-1,000
$ 28.96
4.0
$ 115.84
$ 39.22
2.0
$ 78.44
1,001-4,100
$ 29.73
4.0
$ 118.92
$ 39.22
2.0
$ 78.44
4,101-33,000
$ 36.00
6.0
$ 216.00
$ 39.22
3.0
$ 117.65
33,001-96,000
$ 36.39
8.0
$ 291.12
$ 39.22
4.0
$ 156.87
96,001-500,000
$ 41.01
8.0
$ 328.08
$ 39.22
4.0
$ 156.87
500,001-1 Million
$ 41.01
8.0
$ 328.08
$ 39.22
4.0
$ 156.87
> 1 Million
$ 41.01
8.0
$ 328.08
$ 39.22
4.0
$ 156.87
Notes:
RA/S and state burden and cost estimates to review and revise sample siting plans are assumed to be identical under the RTCR and
Alternative Option.
Sources:
(A) Labor rates for systems from Exhibit 7.1.
(B) RA/S labor hours to review and revise sample siting plans reflect EPA estimate.
(D) Labor rates for state employee from Section 7.2.1.
(E) State labor hours to review and revise sample siting plans reflect EPA estimate.
Economic Analysis for the Final RTCR
7-11
September 2012
-------
Sample Siting Plan Net Cost Summary
PWSs are expected to revise sample siting plans before monitoring begins. For modeling
purposes costs are split between years 2 and 3 of the 25-year compliance period (monitoring is
required starting in year 4). Total and net change in annualized present value cost estimates for
PWSs to revise sample siting plans and states to review the revised sample siting plans (and
consult with PWSs if necessary) under the 1989 TCR, RTCR, and Alternative option are
presented in Exhibit 7.7.
Exhibit 7.7 Annualized Cost Estimates to Revise (PWSs) and Review (States)
Sample Siting Plans ($Millions, 2007$)
PWSs
State
Total
PWSs
State
Total
3% Discount Rate
7% Discount Rate
1989 TCR-Total
$
$
$
$
$
$
RTCR-Total
$ 0.59
$ 0.42
$ 1.01
$ 0.84
$ 0.59
$ 1.42
RTCR - Net Change
$ 0.59
$ 0.42
$ 1.01
$ 0.84
$ 0.59
$ 1.42
Alternative Option - Total
$ 0.59
$ 0.42
$ 1.01
$ 0.84
$ 0.59
$ 1.42
Alternative Option - Net Change
$ 0.59
$ 0.42
$ 1.01
$ 0.84
$ 0.59
$ 1.42
Note: Detail may not add due to independent rounding.
Source: Final RTCR Cost Model.
7.4.3 Monitoring
PWSs
Monitoring costs for PWSs are calculated by multiplying the total numbers of routine,
additional routine, and repeat samples required under the 1989 TCR, RTCR, and Alternative
option (Exhibit 7.9) by the monitoring costs per sample presented in Exhibit 7.2.
EPA assumed that the numbers of systems on monthly, quarterly, and annual monitoring
remain unchanged at the rule effective date for a continuation of the 1989 TCR. For RTCR, EPA
assumed that only systems that received an annual site visit under the 1989 TCR would continue
on annual monitoring; systems that would no longer qualify for annual monitoring under the
RTCR were assumed to revert to baseline quarterly monitoring. (See section 7.4.4 for more
information regarding annual site visits.) Under the Alternative option, all PWSs, regardless of
size or type, start at monthly monitoring at the rule effective date. These differences in
monitoring requirements between the 1989 TCR and the RTCR and Alternative option, which
are presented in more detail in Exhibit 7.8 below, drive the differences in monitoring costs
between the options. Chapters 4 (Exhibit 4.4) and 5 (Exhibit 5.9a-5.9c) show the distribution of
monitoring frequencies prior to and after rule implementation, that are used to further inform the
analysis. The effects of the differences in monitoring regimes on net changes in costs are
described in further detail following the exhibit.
Economic Analysis for the Final RTCR
7-12
September 2012
-------
Exhibit 7.8 Summary of Monitoring Requirements Under the 1989 TCR, RTCR, and Alternative Option
Monitoring
Requirement
1989 TCR
RTCR
Alternative Option
Default Routine
Monitoring Frequency
The default TC monitoring frequency for
ground water Noncommunity Water Systems
(NCWSs) serving <1,000 people is quarterly
The default TC monitoring frequency for all
other PWSs is monthly for TC
The default TC monitoring frequency for non-
seasonal ground water NCWSs serving <1,000
people is quarterly
The default TC monitoring frequency for all other
PWSs is monthly
PWSs would be permitted to transition to the
RTCR at their current TC monitoring
frequencies52
The default TC monitoring frequency for all PWSs
is monthly
All PWSs would start on monthly TC monitoring
Red uced/lncreased
Routine Monitoring
Frequency
Ground water NCWSs serving <1,000
people can reduce to annual monitoring if no
sanitary defects and served only by
protected GW sources
Ground water CWSs serving <1,000 can
reduce to quarterly monitoring if no history of
TC+ in current configuration, no sanitary
defects, and served only by protected GW
sources
All other PWSs are ineligible for reduced
monitoring
PWSs not meeting criteria for reduced
monitoring return to default monitoring (no
increased monitoring provision under 1989
TCR)
Ground water NCWSs serving <1,000 people can
reduce to annual monitoring if no sanitary
defects, clean compliance history for minimum of
12 months, and annual site visit or Level 2
assessment, and correction of all identified
sanitary defects
Ground water CWSs serving <1,000 people can
reduce to quarterly monitoring if no sanitary
defects, clean compliance history for a minimum
of 12 months, and at least one of the following:
1) Annual site visit or voluntary Level 2
assessment and correction of all identified
sanitary defects;
2) An approved cross connection control
program;
3) Continuous disinfection & a residual;
4) 4-log inactivation of viruses daily as perGWR
(4-hr exception allowed); or
5) Other equivalent measures as approved by
the primacy agency.
All other PWSs are ineligible for reduced
Ground water NCWSs serving <1,000 people can
reduce to quarterly monitoring if no sanitary
defects, clean compliance history for minimum of
12 months, and annual site visit or Level 2
assessment, and correction of all identified sanitary
defects.
Ground water CWSs serving <1,000 people can
reduce to quarterly monitoring if no sanitary
defects, clean compliance history for a minimum of
12 months, and at least one of the following:
1) Annual site visit or voluntary Level 2 assessment
and correction of all identified sanitary defects;
2) An approved cross connection control program;
3) Continuous disinfection & a residual;
4) 4-log inactivation of viruses daily as per GWR (4
hour exception allowed); or
5) Other equivalent measures as approved by the
primacy agency.
All other PWSs are ineligible for reduced
52 In order for PWSs to be able to transition to the RTCR at their current TC monitoring frequencies, an annual site visit or voluntary Level 2 assessment would
be needed in the first year for PWSs on annual monitoring, including those transitioning.
Economic Analysis for the Final RTCR
7-13
September 2012
-------
Monitoring
Requirement
1989 TCR
RTCR
Alternative Option
monitoring
Ground water PWSs serving <1,000 people on
quarterly or annual monitoring that experience
any of the following events will be required to
begin monthly monitoring:
1) System triggers a Level 2 assessment (or a
2nd Level 1 assessment in a rolling 12-month
period);
2) System has an E. coli Maximum Contaminant
Level (MCL) violation;
3) System has an RTCR treatment technique
violation (either Level 1 or 2); or
4) System has two monitoring violations in a
rolling 12-month period.
5) NCWSs serving <1,000 increase from
quarterly to monthly monitoring if system has one
monitoring violation and one Level 1 assessment
within 12 months.
6) NCWSs serving <1,000 increase from annual
to quarterly monitoring if system has one
monitoring violation.
monitoring
Ground water PWSs serving <1,000 people on
quarterly monitoring that experience any of the
following events will be required to begin monthly
monitoring (i.e., return to default monitoring
frequency):
1) System triggers a Level 2 assessment (or a 2nd
Level 1 assessment in a rolling 12-month period);
2) System has an E. coli MCL violation;
3) System has an RTCR treatment technique
violation (either Level 1 or 2); or
4) System has two monitoring violations in a rolling
12-month period.
5) NCWSs serving <1,000 increase from quarterly
to monthly monitoring if system has one monitoring
violation and one Level 1 assessment within 12
months.
6) NCWSs serving <1,000 increase from annual to
quarterly monitoring if system has one monitoring
violation.
Additional Routine
Monitoring Frequency
All PWSs serving <4,100 people must take
at least 5 samples in the month following a
TC+ unless state performs site visit and
deems additional sampling unnecessary OR
determines the reason for the TC+ and PWS
has or will correct problem.
Not required for PWSs serving >4,100
people
Ground water NCWSs serving <1,000 people
and monitoring quarterly or annually must take at
least 3 samples in the month following TC+
Ground water CWSs serving <1,000 people and
monitoring quarterly must take at least 3 samples
in the month following TC+
Not required for PWSs monitoring monthly
Ground water PWSs (NCWS and CWS) serving
<1,000 people and monitoring quarterly must take
at least 3 samples in the month following TC+
Not required for PWSs monitoring monthly
Repeat Monitoring
Frequency
All PWSs serving <1,000 people must take 4
repeat TC samples
All PWSs serving >1,000 people must take 3
repeat TC samples
All PWSs must take 3 repeat TC samples
All PWSs must take 3 repeat TC samples
Economic Analysis for the Final RTCR
7-14
September 2012
-------
Routine Monitoring
Under the RTCR, the increased stringency to qualify for reduced monitoring results in
more routine samples being taken over time (fewer PWSs are on reduced monitoring) for most
PWS sizes and types. The only PWSs predicted to have a decrease in routine monitoring under
the RTCR are ground water NTNCWSs serving <1,000 people due to more PWSs monitoring
quarterly than under the 1989 TCR. For the Alternative option, this effect would be combined
with the requirement that all PWSs start the implementation period on monthly monitoring,
resulting in more routine samples being taken for all PWS sizes and types. The Alternative
option also prohibits annual monitoring leading to an even greater increase in the number of
routine samples. The resulting increases in total national costs due to increased monitoring on the
national level are reflected in the routine monitoring costs shown in Exhibit 7.10.
53
Additional Routine Monitoring
The overall reductions in additional routine samples required under the RTCR and
Alternative option result in reduced costs (Exhibit 7.10). Under the RTCR and Alternative
option, additional routine monitoring is no longer required for systems that monitor at least
monthly, and when additional routine monitoring is required, the number of samples required is
reduced from five to three. Cost reductions are greater under the Alternative option than under
the RTCR because all PWSs start on monthly monitoring and are not required to take additional
routine samples during that period.
Repeat Monitoring
Under the RTCR and Alternative option, all PWSs are only required to take three repeat
samples. However, EPA assumes that ground water PWSs treating to less than 4-log would still
take an additional source water sample to comply with the GWR (no change in cost).
Additionally, the number of repeat samples taken is a function of the number of regular and
additional routine samples taken, which in turn affects the number of TC+ samples found (i.e.,
the more samples taken, the greater chance of finding a TC+). Thus, the overall increase in
routine sampling under the RTCR and Alternative option would result in more repeat samples
while the decreases in additional routine samples under both of these options would lead to fewer
repeat samples.
In most PWS size and type categories, the large reductions in additional routine
monitoring samples is the major driver, leading to decreases in repeat samples. Under the RTCR,
only ground water TNCWSs serving <100 people are predicted to see an increase repeat
sampling due to the large increase in routine sampling within this category. Under the
Alternative option, ground water NTNCWSs and TNCWSs serving <1,000 people are predicted
to have increased repeat monitoring as a function of large increases in routine monitoring. The
overall effect on the national level is a reduction in the number (and cost) of required repeat
samples under the RTCR and an increase under the Alternative option compared to the 1989
TCR (Exhibit 7.10).
53 For modeling purposes and discussion throughout the cost chapter, regular routine monitoring samples taken in
the month following a TC+ are included in the additional routine monitoring sample counts.
Economic Analysis for the Final RTCR
7-15
September 2012
-------
Cumulative Monitoring Summary
Exhibit 7.9 summarizes the total number of samples taken by PWS size and category for
routine, additional, and repeat monitoring under the 1989 TCR, RTCR, and Alternative option
over the entire 25-year period of analysis. Appendix A presents additional information on
samples taken for each individual year during the analysis period.
States
Under the 1989 TCR, states are estimated to incur a monthly 15-minute burden to review
each PWS's sample results. This estimate reflects the methodology used to calculate reporting
and recordkeeping burden under the 1989 TCR in the Information Collection Request for the
Microbial Rules (USEPA, 2008b). Because the existing methodology is calculated on a per PWS
basis and the total number of PWSs is the same for cost modeling under the 1989 TCR, RTCR,
and Alternative option, the net change in costs for reviewing monitoring results is estimated to be
zero for the RTCR and Alternative option. Specific actions by states related to positive samples
are accounted for under the actions required in response to those samples. Maintenance of
sample results in SDWIS is accounted for under general implementation and administrative
activities, which are discussed in Section 7.4.1.
Monitoring Net Cost Summary
Total and net change in annualized present value cost estimates for PWSs and states to
perform monitoring under the 1989 TCR, RTCR, and Alternative option are presented in Exhibit
7.10. All monitoring is modeled to begin in year 4 of the 25-year analysis period.
The overall estimated increase in monitoring costs seen under the RTCR is driven by
increases in routine monitoring due to stricter requirements to qualify for reduced monitoring.
However, this is mostly offset by reductions in additional routine and repeat monitoring required
under the revised regulation. For the Alternative option, the requirement for all PWSs to sample
on a monthly basis at the beginning of rule implementation results in a large cost differential that
is only partially offset by reduced costs due to reductions in additional routine monitoring
requirements. Although not shown in Exhibit 7.10, costs for individual PWS categories (size and
type) are expected to move in the same direction relative to the numbers of samples taken as
reflected in Exhibit 7.9. Exceptions to the general trends on the national level are discussed in the
subsections describing routine, additional routine, and repeat sampling above.
Economic Analysis for the Final RTCR
7-16
September 2012
-------
Exhibit 7.9 Cumulative Number of Samples over 25-Year Period of Analysis
1989 TCR
RTCR
Alternative Option
Additional
Additional
Additional
Routine
Routine
Repeat
Routine
Routine
Repeat
Routine
Routine
Repeat
PWS Size
Monitoring
Monitoring
Monitoring
Monitoring
Monitoring
Monitoring
Monitoring
Monitoring
Monitoring
(Population
Samples
Samples
Samples
Samples
Samples
Samples
Samples
Samples
Samples
Served)
A
B
C
D
E
F
G
H
I
Community Water Systems (CWSs) - SW
<100
304,247
23,167
18,698
308,880
-
13,764
308,880
-
13,764
101-500
562,198
27,009
21,684
567,600
-
15,660
567,600
.
15,660
501-1,000
306,605
15,334
12,299
309,672
-
8,708
309,672
.
8,708
1,001-4,100
1,921,237
55,132
33,729
1,951,224
-
33,326
1,951,224
.
33,326
4,101-33,000
10,636,296
-
186,729
10,636,296
-
181,661
10,636,296
.
181,661
33,001-96,000
11,058,960
-
194,149
11,058,960
-
188,880
11,058,960
.
188,880
96,001-500,000
10,190,400
-
178,901
10,190,400
-
174,046
10,190,400
.
174,046
500,001-1 Million
2,019,600
-
35,456
2,019,600
-
34,493
2,019,600
.
34,493
> 1 Million
1,686,960
-
29,616
1,686,960
-
28,812
1,686,960
.
28,812
Total
38,686,502
120,642
711,259
38,729,592
-
679,350
38,729,592
.
679,350
Community Water Systems (CWSs) - GW
<100
2,815,951
286,073
194,462
2,870,075
8,760
156,897
2,908,469
7,545
158,439
101-500
3,344,578
243,895
171,252
3,391,200
6,127
136,906
3,428,876
5,264
137,959
501-1,000
1,072,202
70,803
51,673
1,085,730
1,844
39,659
1,098,488
1,616
39,580
1,001-4,100
3,997,293
160,710
100,618
4,079,328
-
96,939
4,079,328
-
96,939
4,101-33,000
9,145,224
-
230,201
9,145,224
-
217,321
9,145,224
-
217,321
33,001-96,000
4,884,000
-
122,938
4,884,000
-
116,060
4,884,000
-
116,060
96,001-500,000
1,945,680
-
48,976
1,945,680
-
46,236
1,945,680
-
46,236
500,001-1 Million
253,440
-
6,380
253,440
-
6,023
253,440
-
6,023
> 1 Million
269,280
-
6,778
269,280
-
6,399
269,280
-
6,399
Total
27,727,648
761,481
933,279
27,923,956
16,731
822,439
28,012,784
14,425
824,956
Nontransient Noncommunity Water Systems (NTNCWSs) - SW
<100
65,018
4,910
3,991
66,000
-
3,040
66,000
-
3,040
101-500
66,045
3,735
3,011
66,792
-
2,169
66,792
-
2,169
501-1,000
22,976
1,278
1,029
23,232
-
756
23,232
-
756
1,001-4,100
41,759
2,142
1,348
42,768
-
1,228
42,768
-
1,228
4,101-33,000
50,424
-
1,628
50,424
-
1,448
50,424
-
1,448
33,001-96,000
34,320
-
1,108
34,320
-
985
34,320
-
985
96,001-500,000
31,680
-
1,023
31,680
-
910
31,680
-
910
500,001-1 Million
-
-
-
-
-
-
-
-
-
> 1 Million
-
-
-
-
-
-
-
-
-
Total
312,223
12,065
13,138
315,216
-
10,536
315,216
-
10,536
Nontransient Noncommunity Water Systems (NTNCWSs) - GW
<100
971,538
128,775
84,992
932,025
48,142
68,123
1,314,175
36,965
91,416
101-500
725,785
66,525
43,597
678,688
25,630
35,860
976,627
19,382
48,269
501-1,000
190,649
16,037
10,680
180,145
6,166
8,601
249,760
4,802
11,817
1,001-4,100
460,470
28,214
17,790
473,352
-
15,887
473,352
-
15,887
4,101-33,000
153,648
-
5,936
153,648
-
5,157
153,648
-
5,157
33,001-96,000
23,760
-
918
23,760
-
797
23,760
-
797
96,001-500,000
-
-
-
-
-
-
-
-
-
500,001-1 Million
-
-
-
-
-
-
-
-
-
> 1 Million
-
-
-
-
-
-
-
-
-
Total
2,525,850
239,551
163,913
2,441,617
79,938
134,426
3,191,322
61,149
173,343
Transient Noncommunity Water Systems (TNCWSs) - SW
<100
345,401
40,475
33,065
353,496
-
23,122
353,496
-
23,122
101-500
128,156
15,261
12,454
131,208
-
8,192
131,208
-
8,192
501-1,000
22,691
2,704
2,207
23,232
-
1,533
23,232
-
1,533
1,001-4,100
40,151
4,155
2,707
42,240
-
2,312
42,240
-
2,312
4,101-33,000
40,656
-
-
40,656
-
2,225
40,656
-
2,225
33,001-96,000
-
-
-
-
-
-
-
-
-
96,001-500,000
-
-
-
-
-
-
-
-
-
500,001-1 Million
-
-
-
-
-
-
-
-
-
> 1 Million
102,960
-
-
102,960
-
5,636
102,960
-
5,636
Total
680,015
62,596
50,434
693,792
-
43,020
693,792
-
43,020
Transient Noncommunity Water Systems (TNCWSs) - GW
<100
4,493,808
905,554
600,315
6,076,163
446,166
631,105
9,524,123
333,524
912,589
101-500
1,614,924
316,238
210,714
1,940,946
135,822
194,697
3,021,771
104,732
282,740
501-1,000
177,264
32,730
22,064
206,130
14,078
20,078
304,534
10,412
27,932
1,001-4,100
335,283
29,957
19,113
348,480
-
16,027
348,480
-
16,027
4,101-33,000
156,288
-
8,909
156,288
-
7,188
156,288
-
7,188
33,001-96,000
34,320
-
1,956
34,320
-
1,578
34,320
-
1,578
96,001-500,000
26,400
-
1,505
26,400
-
1,214
26,400
-
1,214
500,001-1 Million
63,360
-
3,612
63,360
-
2,914
63,360
-
2,914
> 1 Million
-
-
-
-
-
-
-
-
-
Total
6,901,647
1,284,478
868,188
8,852,088
596,065
874,801
13,479,275
448,667
1,252,181
Grand Total
76,833,885
2,480,814
2,740,210
78,956,260
692,734
2,564,572
84,421,981
524,241
2,983,387
Note: (B), (E), (H) For modeling purposes, additional routine sample counts include regular routine samples taken in the same month.
Source: Appendix A - Total FVVS Counts (A.1z, A.2z, A.3z)
Economic Analysis for the Final RTCR
7-17
September 2012
-------
Exhibit 7.10 Annualized PWS and State Cost Estimates for Monitoring Costs
($Millions, 2007$)
PWSs
State
Total
PWSs
State
Total
3% Discount Rate
7% Discount Rate
Routine Monitoring
1989 TCR-Total
$ 170.59
$
$ 170.59
$ 163.94
$
$ 163.94
RTCR - Total
$ 174.71
$
$ 174.71
$ 167.74
$
$ 167.74
RTCR - Net Change
$ 4.12
$
$ 4.12
$ 3.80
$
$ 3.80
RTCR - Percent Change
2.42%
-
2.42%
2.32%
-
2.32%
Alternative Option - Total
$ 187.50
$
$ 187.50
$ 182.48
$
$ 182.48
Alternative Option - Net Change
$ 16.91
$
$ 16.91
$ 18.54
$
$ 18.54
Alternative Option - Percent Change
9.91%
-
9.91%
11.31%
-
11.31%
Additional Routine Monitoring
1989 TCR-Total
$ 3.87
$
$ 3.87
$ 3.72
$
$ 3.72
RTCR - Total
$ 1.12
$
$ 1.12
$ 1.09
$
$ 1.09
RTCR - Net Change
$ (2.75)
$
$ (2.75)
$ (2.63)
$
$ (2.63)
RTCR - Percent Change
-71.11%
-
-71.11%
-70.75%
-
-70.75%
Alternative Option - Total
$ 0.78
$
$ 0.78
$ 0.66
$
$ 0.66
Alternative Option - Net Change
$ (3.10)
$
$ (3.10)
$ (3.06)
$
$ (3.06)
Alternative Option - Percent Change
-79.98%
-
-79.98%
-82.29%
-
-82.29%
Repeat Monitoring
1989 TCR-Total
$ 5.11
$
$ 5.11
$ 4.92
$
$ 4.92
RTCR - Total
$ 4.88
$
$ 4.88
$ 4.70
$
$ 4.70
RTCR - Net Change
$ (0.23)
$
$ (0.23)
$ (0.22)
$
$ (0.22)
RTCR - Percent Change
-4.53%
-
-4.53%
-4.43%
-
-4.43%
Alternative Option - Total
$ 5.66
$
$ 5.66
$ 5.59
$
$ 5.59
Alternative Option - Net Change
$ 0.54
$
$ 0.54
$ 0.67
$
$ 0.67
Alternative Option - Percent Change
10.61%
-
10.61%
13.71%
-
13.71%
Total
1989 TCR-Total
$ 179.57
$
$ 179.57
$ 172.57
$
$ 172.57
RTCR - Total
$ 180.71
$
$ 180.71
$ 173.52
$
$ 173.52
RTCR - Net Change
$ 1.14
$
$ 1.14
$ 0.95
$
$ 0.95
RTCR - Percent Change
0.63%
-
0.63%
0.55%
-
0.55%
Alternative Option - Total
$ 193.93
$
$ 193.93
$ 188.72
$
$ 188.72
Alternative Option - Net Change
$ 14.36
$
$ 14.36
$ 16.15
$
$ 16.15
Alternative Option - Percent Change
7.99%
-
7.99%
9.36%
-
9.36%
Notes:
1) Detail may not add due to independent rounding.
2) For modeling purposes, additional routine sample counts include regular routine samples taken in the same month.
3) State costs are premised on a per system basis. State costs for monitoring are expected to be identical under the 1989 TCR, RTCR,
and Alternative Option, and are therefore not included in the total costs.
Source: Final RTCR Cost Model.
Economic Analysis for the Final RTCR
7-18
September 2012
-------
7.4.4 Annual Site Visits
Under the RTCR, any PWS on an annual monitoring schedule would be required to also
have an annual site visit conducted by the state or state-designated third party. A voluntary Level
2 assessment can also satisfy the annual site visit requirement. In many cases a sanitary survey
performed during the same year can also be used to satisfy this requirement.54 Although similar
annual site visits are not currently required under the 1989 TCR, discussions with states during
the TCRDSAC proceedings revealed that some do, in fact, conduct such inspections for PWSs
on annual monitoring schedules. Because of the high cost for an annual site visit by a state, for
this analysis, EPA assumes that no states would choose to conduct annual site visits unless they
already do so under the 1989 TCR. Similarly, because of the high cost of the voluntary Level 2
assessment relative to other options that PWSs have to comply with the rule (such as quarterly
monitoring) it was assumed that any PWS that is currently on annual monitoring, but not
receiving an annual site visit, would opt not to conduct a voluntary Level 2 assessment.
Therefore, for overall costing purposes, no net change in state or PWS costs are assumed for
annual monitoring site visits under the RTCR or Alternative option.
7.4.5 Assessments
PWSs
Level 1 Assessments
Under the RTCR and Alternative option, all PWSs experiencing a Level 1 trigger must
complete a Level 1 assessment of the PWS. A Level 1 trigger under the RTCR and Alternative
option is defined as:
• For PWSs taking >40 samples per month, TC+ exceeds 5.0% for a given month;
• For PWSs taking <40 samples per month, two or more TC+ in a month; or
• Failure to take all required repeat samples after a single TC+ sample.
The 1989 TCR does not require a specific assessment to be performed in response to
events comparable to the Level 1 triggers described above (i.e., non-acute violations). However,
PWSs do perform some level of activity similar to a Level 1 assessment in response to
violations. This effort is taken into account in the cost model to accurately assess the net cost of
changes attributable to the RTCR.
A Level 1 assessment is an evaluation to identify the possible presence of sanitary
defects, defects in distribution system coliform monitoring practices, and (when possible) the
likely reason that the system triggered the assessment. It is conducted by the system operator or
54 In some instances, the performance of an assessment (especially a Level 2 assessment) may overlap with a
scheduled sanitary survey. To the extent that the requirements for performing a sanitary survey may be satisfied as
part of the assessment process, PWSs and states may realize a cost savings compared to performing a separate
sanitary survey. This potential for indirect cost savings is not captured in the cost model, resulting in a potential
overestimate of costs stemming from RTCR implementation.
Economic Analysis for the Final RTCR
7-19
September 2012
-------
owner. Minimum elements include review and identification of atypical events that could affect
distributed water quality or indicate that distributed water quality was impaired; changes in
distribution system maintenance and operation that could affect distributed water quality
(including water storage); source and treatment considerations that bear on distributed water
quality, where appropriate (e.g., whether a ground water system is disinfected); existing water
quality monitoring data; and inadequacies in sample sites, sampling protocol, and sample
processing. The state may tailor specific assessment elements to the size and type of the system.
Systems may tailor their assessment activities based on the characteristics of the distribution
system (consistent with any state directives). Additionally, as part of the Level 1 assessment,
PWSs would be required to submit to the state a form identifying sanitary defects detected,
corrective actions completed, and a timetable for any corrective actions not already completed.
Additional detail on the requirements of a Level 1 assessment and the derivation of associated
unit burden (labor hours) for performance of the assessment is provided in the Technology and
Cost Document for the Final Revised Total Coliform Rule (USEPA, 2010d).
Level 2 Assessments
Under the RTCR and Alternative option, all PWSs experiencing a Level 2 trigger must
complete a Level 2 assessment of the PWS. A Level 2 trigger under the RTCR and Alternative
option is defined as:
• An E. coli MCL violation; or
• A second Level 1 treatment technique trigger, within a rolling 12-month period,
unless the first Level 1 treatment technique trigger was based on exceeding the
allowable number of TC+ samples, the state has determined a likely reason for the
TC+ samples that caused the initial Level 1 treatment technique trigger, and the
state establishes that the system has fully corrected the problem; or
• For PWSs with approved reduced annual monitoring, a Level 1 trigger in two
consecutive years.
As with Level 1 assessments, the 1989 TCR does not require a specific assessment to be
performed in response to events comparable to the Level 2 triggers described above (i.e., acute
violations), but PWSs do currently perform some level of activity similar to a Level 2 assessment
in response to acute violations. These actions are taken into account in the cost model to properly
assess the net cost of changes attributable to the RTCR regulatory options.
A Level 2 assessment (or comparable assessment under the 1989 TCR) would be more
involved than a Level 1 assessment. A Level 2 assessment would be an evaluation to identify the
possible presence of sanitary defects, defects in distribution system coliform monitoring
practices, and (when possible) the likely reason that the system triggered the assessment. A Level
2 assessment provides a more detailed examination of the system (including the system's
monitoring and operational practices) than does a Level 1 assessment through the use of more
comprehensive investigation and review of available information, additional internal and
external resources, and other relevant practices. It is conducted by an individual approved by the
state, which may include the system operator. Minimum elements include review and
Economic Analysis for the Final RTCR
7-20
September 2012
-------
identification of atypical events that could affect distributed water quality or indicate that
distributed water quality was impaired; changes in distribution system maintenance and
operation that could affect distributed water quality (including water storage); source and
treatment considerations that bear on distributed water quality, where appropriate (e.g., whether a
ground water system is disinfected); existing water quality monitoring data; and inadequacies in
sample sites, sampling protocol, and sample processing. The state may tailor specific assessment
elements to the size and type of the system. The state may also direct expedited actions or
additional actions in the case of an E. coli MCL violation. Systems may tailor their assessment
activities based on the characteristics of the distribution system (consistent with any state
directives). Additionally, as part of the Level 2 assessment, PWSs must submit to the state a
form identifying sanitary defects detected, corrective actions completed, and a timetable for
completion of any corrective actions that not already completed. Additional detail on the
derivation of associated unit burden (labor hours) for performance of the assessment is provided
in the Technology and Cost Document for the Final Revised Total Coliform Rule (USEPA,
2010d).
The labor rates presented in Section 7.2.1 are used along with estimates of labor hours as
presented in the Technology and Cost Document for the Final Revised Total Coliform Rule
(USEPA, 2010d) to generate Level 1 and Level 2 assessment unit costs by PWS size and type.
Labor hours provided are assumed to include time for reporting and recordkeeping activities.
Estimates of PWS unit costs for Level 1 and Level 2 assessments are presented in Exhibits 7.11
and 7.12. Additionally, the numbers of Level 1 and level 2 assessments over the 25-year
compliance period (used to calculate total costs) are presented in Exhibit 7.13.
Economic Analysis for the Final RTCR
7-21
September 2012
-------
Exhibit 7.11 PWS Unit Costs Estimates for Assessment Activities
(1989 TCR) (2007$)
PWS Size
(Population
Served)
Labor Cost
(per hour)
Activities Similar to Level 1 Assessments
Activities Similar to Level 2 Assessments
Non-Acute
Violations (single
violation) (hours)
Unit Cost
Acute
Violations
(hours)
Unit Cost
Non-Acute
Violations
(multiple
violations)
(hours)
Unit Cost
A
B
C=A*B
D
E=A*D
F
G=A*F
Community Water Systems (CWSs) - SW
<100
$ 25.10
11.0
$ 276.10
14.0
$ 351.40
14.0
$ 351.40
101-500
$ 27.03
11.0
$ 297.33
14.0
$ 378.42
14.0
$ 378.42
501-1,000
$ 28.96
13.0
$ 376.48
15.0
$ 434.40
15.0
$ 434.40
1,001-4,100
$ 29.73
22.0
$ 654.06
29.0
$ 862.17
29.0
$ 862.17
4,101-33,000
$ 36.00
30.0
$ 1,080.00
36.0
$ 1,296.00
36.0
$ 1,296.00
33,001-96,000
$ 36.39
59.0
$ 2,147.01
75.0
$ 2,729.25
75.0
$ 2,729.25
96,001-500,000
$ 41.01
108.0
$ 4,429.08
117.0
$ 4,798.17
117.0
$ 4,798.17
500,001-1 Million
$ 41.01
108.0
$ 4,429.08
117.0
$ 4,798.17
117.0
$ 4,798.17
> 1 Million
$ 41.01
108.0
$ 4,429.08
117.0
$ 4,798.17
117.0
$ 4,798.17
Community Water Systems (CWSs) - GW
<100
$ 25.10
11.0
$ 276.10
14.0
$ 351.40
14.0
$ 351.40
101-500
$ 27.03
11.0
$ 297.33
14.0
$ 378.42
14.0
$ 378.42
501-1,000
$ 28.96
13.0
$ 376.48
15.0
$ 434.40
15.0
$ 434.40
1,001-4,100
$ 29.73
22.0
$ 654.06
29.0
$ 862.17
29.0
$ 862.17
4,101-33,000
$ 36.00
30.0
$ 1,080.00
36.0
$ 1,296.00
36.0
$ 1,296.00
33,001-96,000
$ 36.39
59.0
$ 2,147.01
75.0
$ 2,729.25
75.0
$ 2,729.25
96,001-500,000
$ 41.01
108.0
$ 4,429.08
117.0
$ 4,798.17
117.0
$ 4,798.17
500,001-1 Million
$ 41.01
108.0
$ 4,429.08
117.0
$ 4,798.17
117.0
$ 4,798.17
> 1 Million
$ 41.01
108.0
$ 4,429.08
117.0
$ 4,798.17
117.0
$ 4,798.17
Nontransient Noncommunity Water Systems (NTNCWSs) - SW
<100
$ 25.10
4.0
$ 100.40
6.0
$ 150.60
6.0
$ 150.60
101-500
$ 27.03
4.0
$ 108.12
6.0
$ 162.18
6.0
$ 162.18
501-1,000
$ 28.96
4.0
$ 115.84
6.0
$ 173.76
6.0
$ 173.76
1,001-4,100
$ 29.73
4.0
$ 118.92
6.0
$ 178.38
6.0
$ 178.38
4,101-33,000
$ 36.00
30.0
$ 1,080.00
36.0
$ 1,296.00
36.0
$ 1,296.00
33,001-96,000
$ 36.39
59.0
$ 2,147.01
75.0
$ 2,729.25
75.0
$ 2,729.25
96,001-500,000
$ 41.01
108.0
$ 4,429.08
117.0
$ 4,798.17
117.0
$ 4,798.17
500,001-1 Million
$ 41.01
108.0
$ 4,429.08
117.0
$ 4,798.17
117.0
$ 4,798.17
> 1 Million
$ 41.01
108.0
$ 4,429.08
117.0
$ 4,798.17
117.0
$ 4,798.17
Nontransient Noncommunity Water Systems (NTNCWSs) - GW
<100
$ 25.10
4.0
$ 100.40
6.0
$ 150.60
6.0
$ 150.60
101-500
$ 27.03
4.0
$ 108.12
6.0
$ 162.18
6.0
$ 162.18
501-1,000
$ 28.96
4.0
$ 115.84
6.0
$ 173.76
6.0
$ 173.76
1,001-4,100
$ 29.73
4.0
$ 118.92
6.0
$ 178.38
6.0
$ 178.38
4,101-33,000
$ 36.00
30.0
$ 1,080.00
36.0
$ 1,296.00
36.0
$ 1,296.00
33,001-96,000
$ 36.39
59.0
$ 2,147.01
75.0
$ 2,729.25
75.0
$ 2,729.25
96,001-500,000
$ 41.01
108.0
$ 4,429.08
117.0
$ 4,798.17
117.0
$ 4,798.17
500,001-1 Million
$ 41.01
108.0
$ 4,429.08
117.0
$ 4,798.17
117.0
$ 4,798.17
> 1 Million
$ 41.01
108.0
$ 4,429.08
117.0
$ 4,798.17
117.0
$ 4,798.17
Transient Noncommunity Water Systems (TNCWSs) - SW
<100
$ 25.10
4.0
$ 100.40
6.0
$ 150.60
6.0
$ 150.60
101-500
$ 27.03
4.0
$ 108.12
6.0
$ 162.18
6.0
$ 162.18
501-1,000
$ 28.96
4.0
$ 115.84
6.0
$ 173.76
6.0
$ 173.76
1,001-4,100
$ 29.73
4.0
$ 118.92
6.0
$ 178.38
6.0
$ 178.38
4,101-33,000
$ 36.00
30.0
$ 1,080.00
36.0
$ 1,296.00
36.0
$ 1,296.00
33,001-96,000
$ 36.39
59.0
$ 2,147.01
75.0
$ 2,729.25
75.0
$ 2,729.25
96,001-500,000
$ 41.01
108.0
$ 4,429.08
117.0
$ 4,798.17
117.0
$ 4,798.17
500,001-1 Million
$ 41.01
108.0
$ 4,429.08
117.0
$ 4,798.17
117.0
$ 4,798.17
> 1 Million
$ 41.01
108.0
$ 4,429.08
117.0
$ 4,798.17
117.0
$ 4,798.17
Transient Noncommunity Water Systems (TNCWSs) - GW
<100
$ 25.10
4.0
$ 100.40
6.0
$ 150.60
6.0
$ 150.60
101-500
$ 27.03
4.0
$ 108.12
6.0
$ 162.18
6.0
$ 162.18
501-1,000
$ 28.96
4.0
$ 115.84
6.0
$ 173.76
6.0
$ 173.76
1,001-4,100
$ 29.73
4.0
$ 118.92
6.0
$ 178.38
6.0
$ 178.38
4,101-33,000
$ 36.00
30.0
$ 1,080.00
36.0
$ 1,296.00
36.0
$ 1,296.00
33,001-96,000
$ 36.39
59.0
$ 2,147.01
75.0
$ 2,729.25
75.0
$ 2,729.25
96,001-500,000
$ 41.01
108.0
$ 4,429.08
117.0
$ 4,798.17
117.0
$ 4,798.17
500,001-1 Million
$ 41.01
108.0
$ 4,429.08
117.0
$ 4,798.17
117.0
$ 4,798.17
> 1 Million
$ 41.01
108.0
$ 4,429.08
117.0
$ 4,798.17
117.0
$ 4,798.17
Note:
(F) EPA assumes that the burden incurred by operators to assess their FWSs follow ing a second non-acute violation is equal to the burden incurred by an
assessment follow ing an acute violation.
Sources:
(A) Labor rates for FWSs from Bchibit 7.1.
(B), (D), (F) Final RTCR T&C Document.
Economic Analysis for the Final RTCR
7-22
September 2012
-------
Exhibit 7.12 PWS Unit Costs Estimates for Level 1 and Level 2 Assessments
(RTCR and Alternative Option) (2007$)
Level 1 Assessments
Level 2 Assessments
Non-Acute
Level 2 Triggers
Trigger
Acute
(triggered by
PWS Size
Labor Cost
(single trigger)
Violations
multiple Level 1s)
(Population
(per hour)
(hours)
Unit Cost
(hours)
Unit Cost
(hours)
Unit Cost
Served)
A
B
C=A*B
D
E=A*D
F
G=A*F
Community Water Systems (CWSs) -SW
<100
$ 25.10
19.0
$ 476.90
23.0
$ 577.30
22.0
$ 552.20
101-500
$ 27.03
19.0
$ 513.57
23.0
$ 621.69
22.0
$ 594.66
501-1,000
$ 28.96
20.0
$ 579.20
24.0
$ 695.04
23.0
$ 666.08
1,001-4,100
$ 29.73
31.0
$ 921.63
48.0
$ 1,427.04
46.0
$ 1,367.58
4,101-33,000
$ 36.00
41.0
$ 1,476.00
71.0
$ 2,556.00
69.0
$ 2,484.00
33,001-96,000
$ 36.39
68.0
$ 2,474.52
121.0
$ 4,403.19
116.0
$ 4,221.24
96,001-500,000
$ 41.01
159.0
$ 6,520.59
252.0
$ 10,334.52
238.0
$ 9,760.38
500,001-1 Million
$ 41.01
159.0
$ 6,520.59
252.0
$ 10,334.52
238.0
$ 9,760.38
> 1 Million
$ 41.01
159.0
$ 6,520.59
252.0
$ 10,334.52
238.0
$ 9,760.38
Community Water Systems (CWSs) -GW
<100
$ 25.10
19.0
$ 476.90
23.0
$ 577.30
22.0
$ 552.20
101-500
$ 27.03
19.0
$ 513.57
23.0
$ 621.69
22.0
$ 594.66
501-1,000
$ 28.96
20.0
$ 579.20
24.0
$ 695.04
23.0
$ 666.08
1,001-4,100
$ 29.73
31.0
$ 921.63
48.0
$ 1,427.04
46.0
$ 1,367.58
4,101-33,000
$ 36.00
41.0
$ 1,476.00
71.0
$ 2,556.00
69.0
$ 2,484.00
33,001-96,000
$ 36.39
68.0
$ 2,474.52
121.0
$ 4,403.19
116.0
$ 4,221.24
96,001-500,000
$ 41.01
159.0
$ 6,520.59
252.0
$ 10,334.52
238.0
$ 9,760.38
500,001-1 Million
$ 41.01
159.0
$ 6,520.59
252.0
$ 10,334.52
238.0
$ 9,760.38
> 1 Million
$ 41.01
159.0
$ 6,520.59
252.0
$ 10,334.52
238.0
$ 9,760.38
Nontransient Noncommunity Water Systems (NTNCWSs) - SW
<100
$ 25.10
7.0
$ 175.70
21.0
$ 527.10
9.0
$ 225.90
101-500
$ 27.03
7.0
$ 189.21
21.0
$ 567.63
9.0
$ 243.27
501-1,000
$ 28.96
7.0
$ 202.72
21.0
$ 608.16
9.0
$ 260.64
1,001-4,100
$ 29.73
8.0
$ 237.84
29.0
$ 862.17
10.0
$ 297.30
4,101-33,000
$ 36.00
41.0
$ 1,476.00
71.0
$ 2,556.00
69.0
$ 2,484.00
33,001-96,000
$ 36.39
68.0
$ 2,474.52
121.0
$ 4,403.19
116.0
$ 4,221.24
96,001-500,000
$ 41.01
159.0
$ 6,520.59
252.0
$ 10,334.52
238.0
$ 9,760.38
500,001-1 Million
$ 41.01
159.0
$ 6,520.59
252.0
$ 10,334.52
238.0
$ 9,760.38
> 1 Million
$ 41.01
159.0
$ 6,520.59
252.0
$ 10,334.52
238.0
$ 9,760.38
Nontransient Noncommunity Water Systems (NTNCWSs) - GW
<100
$ 25.10
7.0
$ 175.70
21.0
$ 527.10
9.0
$ 225.90
101-500
$ 27.03
7.0
$ 189.21
21.0
$ 567.63
9.0
$ 243.27
501-1,000
$ 28.96
7.0
$ 202.72
21.0
$ 608.16
9.0
$ 260.64
1,001-4,100
$ 29.73
8.0
$ 237.84
29.0
$ 862.17
10.0
$ 297.30
4,101-33,000
$ 36.00
41.0
$ 1,476.00
71.0
$ 2,556.00
69.0
$ 2,484.00
33,001-96,000
$ 36.39
68.0
$ 2,474.52
121.0
$ 4,403.19
116.0
$ 4,221.24
96,001-500,000
$ 41.01
159.0
$ 6,520.59
252.0
$ 10,334.52
238.0
$ 9,760.38
500,001-1 Million
$ 41.01
159.0
$ 6,520.59
252.0
$ 10,334.52
238.0
$ 9,760.38
> 1 Million
$ 41.01
159.0
$ 6,520.59
252.0
$ 10,334.52
238.0
$ 9,760.38
Transient Noncommunity Water Systems (TNCWSs) - SW
<100
$ 25.10
7.0
$ 175.70
21.0
$ 527.10
9.0
$ 225.90
101-500
$ 27.03
7.0
$ 189.21
21.0
$ 567.63
9.0
$ 243.27
501-1,000
$ 28.96
7.0
$ 202.72
21.0
$ 608.16
9.0
$ 260.64
1,001-4,100
$ 29.73
8.0
$ 237.84
29.0
$ 862.17
10.0
$ 297.30
4,101-33,000
$ 36.00
41.0
$ 1,476.00
71.0
$ 2,556.00
69.0
$ 2,484.00
33,001-96,000
$ 36.39
68.0
$ 2,474.52
121.0
$ 4,403.19
116.0
$ 4,221.24
96,001-500,000
$ 41.01
159.0
$ 6,520.59
252.0
$ 10,334.52
238.0
$ 9,760.38
500,001-1 Million
$ 41.01
159.0
$ 6,520.59
252.0
$ 10,334.52
238.0
$ 9,760.38
> 1 Million
$ 41.01
159.0
$ 6,520.59
252.0
$ 10,334.52
238.0
$ 9,760.38
Transient Noncommunity Water Systems (TNCWSs) - GW
<100
$ 25.10
7.0
$ 175.70
21.0
$ 527.10
9.0
$ 225.90
101-500
$ 27.03
7.0
$ 189.21
21.0
$ 567.63
9.0
$ 243.27
501-1,000
$ 28.96
7.0
$ 202.72
21.0
$ 608.16
9.0
$ 260.64
1,001-4,100
$ 29.73
8.0
$ 237.84
29.0
$ 862.17
10.0
$ 297.30
4,101-33,000
$ 36.00
41.0
$ 1,476.00
71.0
$ 2,556.00
69.0
$ 2,484.00
33,001-96,000
$ 36.39
68.0
$ 2,474.52
121.0
$ 4,403.19
116.0
$ 4,221.24
96,001-500,000
$ 41.01
159.0
$ 6,520.59
252.0
$ 10,334.52
238.0
$ 9,760.38
500,001-1 Million
$ 41.01
159.0
$ 6,520.59
252.0
$ 10,334.52
238.0
$ 9,760.38
> 1 Million
$ 41.01
159.0
$ 6,520.59
252.0
$ 10,334.52
238.0
$ 9,760.38
Sources:
(A) Labor rates for FW3s from Exhibit 7.1.
(B), (D), (F) Rnal RTCRT&C Document.
Economic Analysis for the Final RTCR
7-23
September 2012
-------
Exhibit 7.13 Number of Level 1 and Level 2 Assessments over the 25-Year
Compliance Period
1989 TCR
RTCR
Alternative Option
Activities
Similar to
Activities Similar to
Level 1
Level 2
Level 1
Level 2
Level 1
Level 2
Assessments
Assessments
Assessments
Assessments
Assessments
Assessments
Level 2
Level 2
Non-Acute
Non-Acute
Non-Acute
Triggers
Non-Acute
Triggers
Violations
Violations
Trigger
(triggered
Trigger
(triggered
PWS Size
(single
Acute
(multiple
(single
Acute
by multiple
(single
Acute
by multiple
(Population
violation)
Violations
violations)
trigger)
Violations
Level 1s)
trigger)
Violations
Level 1s)
Served)
A
B
C
D
E
F
G
H
I
Community Water Systems (CWSs)
¦SW
<100
525
157
184
400
100
102
400
100
102
101-500
649
167
111
539
119
75
539
119
75
501-1,000
361
102
63
277
75
40
277
75
40
1,001-4,100
954
162
149
920
146
132
920
146
132
4,101-33,000
2,152
197
2,152
197
2,152
197
33,001-96,000
534
56
534
56
534
56
96,001-500,000
233
24
233
24
233
24
500,001-1 Million
22
22
22
> 1 Million
Total
5,429
865
507
5,076
717
349
5,076
717
349
Community Water Systems (CWSs)
-GW
<100
9,772
1,141
5,383
8,004
853
3,523
7,871
926
3,272
101-500
8,169
1,025
4,214
6,502
696
2,399
6,495
747
2,543
501-1,000
2,250
284
1,050
1,780
188
626
1,772
203
607
1,001-4,100
3,545
477
2,808
3,208
342
1,705
3,208
342
1,705
4,101-33,000
4,545
263
4,545
263
4,545
263
33,001-96,000
656
53
656
53
656
53
96,001-500,000
129
10
129
10
129
10
500,001-1 Million
> 1 Million
Total
29,066
3,253
13,455
24,824
2,405
8,253
24,675
2,544
8,127
Nontransient Noncommunitv Water Systems (NTNCWSs) - SW
<100
98
35
77
75
28
41
75
28
41
101-500
88
29
40
69
19
24
69
19
24
501-1,000
30
9
13
24
6
9
24
6
9
1,001-4,100
42
19
37
37
13
23
37
13
23
4,101-33,000
5
5
5
33,001-96,000
96,001-500,000
500,001-1 Million
> 1 Million
Total
262
93
167
209
67
98
209
67
98
Nontransient Noncommunitv Water Systems (NTNCWSs) - GW
<100
5,581
856
3,829
4,797
559
2,010
5,673
723
3,390
101-500
3,130
447
1,273
2,794
315
757
3,551
446
1,333
501-1,000
744
95
298
675
79
168
814
99
298
1,001-4,100
818
169
974
690
114
530
690
114
530
4,101-33,000
123
9
123
9
123
9
33,001-96,000
4
4
4
96,001-500,000
500,001-1 Million
> 1 Million
Total
10,400
1,577
6,373
9,084
1,077
3,466
10,855
1,393
5,551
Transient Noncomm unity Water Systems (TNCWSs) - SW
<100
1,093
430
780
796
250
425
796
250
425
101-500
410
170
320
278
90
154
278
90
154
501-1,000
67
28
47
50
17
25
50
17
25
1,001-4,100
92
50
128
73
29
69
73
29
69
4,101-33,000
8
8
8
33,001-96,000
96,001-500,000
500,001-1 Million
> 1 Million
Total
1,670
677
1,275
1,204
386
674
1,204
386
674
Transient Noncomm unity Water Systems (TNCWSs) - GW
<100
44,730
6,649
25,425
47,190
5,477
20,628
57,597
7,796
37,532
101-500
14,530
2,089
8,864
13,780
1,608
5,694
17,358
2,441
10,924
501-1,000
1,477
221
896
1,396
177
585
1,661
230
1,015
1,001-4,100
927
186
1,138
773
117
638
773
117
638
4,101-33,000
116
4
116
4
116
4
33,001-96,000
96,001-500,000
500,001-1 Million
> 1 Million
Total
61,780
9,149
36,324
63,256
7,383
27,546
77,506
10,589
50,109
Q~and Total
108,608
15,613
58,102
103,653
12,035
40,385
119,526
15,695
64,908
Notes:
1) Detail may not add due to rounding.
2) Assessments conducted under the 1989 TCRare not labeled as Level 1 and Level 2 assessments; however they are sinilar to the assessments under the
RTCR and Alternative option.
Sources:
(A), (B), (Q-A.1.Z
(D), (E), (F) - A.2.Z
(G), (H), (I) - A.3.Z
Economic Analysis for the Final RTCR
7-24
September 2012
-------
States
Level 1 and Level 2 Assessments
Under the 1989 TCR, RTCR and Alternative option, states would incur burden to review
the completed assessment forms required to be filed by PWSs (or similar reports required by
states under the 1989 TCR). Although some states may choose to conduct assessments for their
PWSs, EPA does not quantify these costs since this is not a regulatory requirement. State costs
are based on the number of PWSs submitting assessment reports. The state labor rate presented
in Section 7.2.1 and estimates of labor hours are used to generate state Level 1 and Level 2
assessment unit costs. EPA estimates that state burden to review PWS assessment forms would
range from one to eight hours depending on PWS size and type, as well as the level of the
assessment. This burden includes any time required to consult with the PWS about the
assessment report. Estimates of state unit costs for Level 1 and Level 2 assessments are presented
in Exhibit 7.14.
Assessment Net Cost Summary
Annualized cost estimates for Level 1 and Level 2 assessments under the 1989 TCR,
RTCR, and Alternative option are calculated by multiplying the number of assessments
estimated by the predictive modeling (summarized in Exhibit 7.13) by the unit costs presented in
Exhibits 7.11,7.12, and 7.14. Exhibit 7.13 presents the estimated totals of non-acute and acute
MCL violations (1989 TCR) and Level 1 and Level 2 assessments (RTCR and Alternative
option). The model predicts a total of approximately 109,000 single non-acute MCL violations,
58,000 cases of a second non-acute MCL violation, and 16,000 acute MCL violations for the
1989 TCR under which some PWSs do currently engage in some assessment activity which may
or may not meet the RTCR criteria (see section 7.4.5). For the RTCR, the model predicts
approximately 104,000 Level 1 assessments and 52,000 Level 2 assessments. For the Alternative
option, the model predicts approximately 120,000 Level 1 assessments and 81,000 Level 2
assessments. (Appendix A provides a detailed breakout of the number of Level 1 and Level 2
assessments estimated by the occurrence model.) Total and net change in annualized present
value cost estimates for PWSs and states to perform Level 1 and Level 2 assessments under the
1989 TCR, RTCR, and Alternative option are presented in Exhibit 7.15 below.
Under the RTCR, all PWSs are required to conduct assessments of their systems when
they exceed Level 1 or Level 2 treatment technique triggers. While PWSs are not required to
conduct assessments under the 1989 TCR, some PWSs do currently engage in some assessment
activity (which may or may not meet the RTCR criteria) following non-acute and acute MCL
violations. EPA estimates both the costs to PWSs to conduct assessments under the RTCR as
well as the level of effort that PWSs already put towards assessment activities under the 1989
TCR; these estimates are based on the work of the stakeholders in the TWG during the
proceedings of the TCRDSAC. These estimates allowed EPA to determine the average net costs
to conduct assessments under the RTCR. EPA assumes that the numbers of non-acute and acute
MCL violations would remain steady under a continuation of the 1989 TCR (based on review of
SDWIS/FED violation data). Under the RTCR, EPA assumes that the numbers of assessment
triggers decrease over time from the steady state level estimate based on the 1989 TCR to a new
Economic Analysis for the Final RTCR
7-25
September 2012
-------
steady state level, as a result of reduced fecal indicator occurrence associated with the beneficial
effects of requiring assessments and corrective action.
The overall number of assessments increases under the Alternative option. This is a result
of the initial monthly monitoring requirements for all PWSs under this analysis. The modeling
results indicate that the much higher sampling early on after the rule's effective date would result
in more positive samples and associated assessments despite the predicted long term reductions
in occurrence. This increase in total assessments performed, combined with the higher unit cost
of performing assessments compared to existing practices under the 1989 TCR, results in a
higher net cost increase for PWSs under the Alternative option than under the RTCR. For states,
the increase in the number of assessments is estimated to translate directly to a cost increase. The
total net change in cost for the Alternative option is estimated to be positive, and greater than
under the RTCR.
Economic Analysis for the Final RTCR
7-26
September 2012
-------
Exhibit 7.14 State Unit Cost Estimates for Review of Level 1 and Level 2
Assessments under the 1989 TCR, RTCR, and Alternative Option (2007$)
Level 1 Assessments
Level 2 Assessments
Non-Acute
Level 2 Triggers
Trigger
Acute
(triggered by
PWS Size
Labor Cost
(single trigger)
Violations
multiple Level 1s)
(Population
(per hour)
(hours)
Unit Cost
(hours)
J nit Cost
(hours)
Unit Cost
Served)
A
B
C=
A*B
D
E=A*D
F
OA*F
Community Water Systems (CWSs) - SW
<100
39.22
1.0
$
39.22
2.0
78.44
2.0
78.44
101-500
39.22
1.0
$
39.22
2.0
78.44
2.0
78.44
501-1,000
39.22
2.0
$
78.44
4.0
$
156.87
4.0
$
156.87
1,001-4,100
39.22
2.0
$
78.44
4.0
$
156.87
4.0
$
156.87
4,101-33,000
39.22
3.0
$
117.65
6.0
$
235.31
6.0
$
235.31
33,001-96,000
39.22
4.0
$
156.87
8.0
$
313.74
8.0
$
313.74
96,001-500,000
39.22
4.0
$
156.87
8.0
$
313.74
8.0
$
313.74
500,001-1 Million
39.22
4.0
$
156.87
8.0
$
313.74
8.0
$
313.74
> 1 Million
39.22
4.0
$
156.87
8.0
$
313.74
8.0
$
313.74
Community Water Systems (CWSs) - GW
<100
39.22
1.0
$
39.22
2.0
78.44
2.0
78.44
101-500
39.22
1.0
$
39.22
2.0
78.44
2.0
78.44
501-1,000
39.22
2.0
$
78.44
4.0
$
156.87
4.0
$
156.87
1,001-4,100
39.22
2.0
$
78.44
4.0
$
156.87
4.0
$
156.87
4,101-33,000
39.22
3.0
$
117.65
6.0
$
235.31
6.0
$
235.31
33,001-96,000
39.22
4.0
$
156.87
8.0
$
313.74
8.0
$
313.74
96,001-500,000
39.22
4.0
$
156.87
8.0
$
313.74
8.0
$
313.74
500,001-1 Million
39.22
4.0
$
156.87
8.0
$
313.74
8.0
$
313.74
> 1 Million
39.22
4.0
$
156.87
8.0
$
313.74
8.0
$
313.74
Nontransient Noncommunity Water Systems (NTNCWSs)
SW
<100
39.22
1.0
$
39.22
2.0
78.44
2.0
78.44
101-500
39.22
1.0
$
39.22
2.0
78.44
2.0
78.44
501-1,000
39.22
2.0
$
78.44
4.0
$
156.87
4.0
$
156.87
1,001-4,100
39.22
2.0
$
78.44
4.0
$
156.87
4.0
$
156.87
4,101-33,000
39.22
3.0
$
117.65
6.0
$
235.31
6.0
$
235.31
33,001-96,000
39.22
4.0
$
156.87
8.0
$
313.74
8.0
$
313.74
96,001-500,000
39.22
4.0
$
156.87
8.0
$
313.74
8.0
$
313.74
500,001-1 Million
39.22
4.0
$
156.87
8.0
$
313.74
8.0
$
313.74
> 1 Million
39.22
4.0
$
156.87
8.0
$
313.74
8.0
$
313.74
Nontransient Noncommunity Water Systems (NTNCWSs)
GW
<100
39.22
1.0
$
39.22
2.0
78.44
2.0
78.44
101-500
39.22
1.0
$
39.22
2.0
78.44
2.0
78.44
501-1,000
$
39.22
2.0
$
78.44
4.0
$
156.87
4.0
$
156.87
1,001-4,100
$
39.22
2.0
$
78.44
4.0
$
156.87
4.0
$
156.87
4,101-33,000
$
39.22
3.0
$
117.65
6.0
$
235.31
6.0
$
235.31
33,001-96,000
$
39.22
4.0
$
156.87
8.0
$
313.74
8.0
$
313.74
96,001-500,000
$
39.22
4.0
$
156.87
8.0
$
313.74
8.0
$
313.74
500,001-1 Million
$
39.22
4.0
$
156.87
8.0
$
313.74
8.0
$
313.74
> 1 Million
$
39.22
4.0
$
156.87
8.0
$
313.74
8.0
$
313.74
Transient Noncommunity Water Systems (TNCWSs) - SW
<100
$
39.22
1.0
$
39.22
2.0
78.44
2.0
78.44
101-500
$
39.22
1.0
$
39.22
2.0
78.44
2.0
78.44
501-1,000
$
39.22
2.0
$
78.44
4.0
$
156.87
4.0
$
156.87
1,001-4,100
$
39.22
2.0
$
78.44
4.0
$
156.87
4.0
$
156.87
4,101-33,000
$
39.22
3.0
$
117.65
6.0
$
235.31
6.0
$
235.31
33,001-96,000
$
39.22
4.0
$
156.87
8.0
$
313.74
8.0
$
313.74
96,001-500,000
$
39.22
4.0
$
156.87
8.0
$
313.74
8.0
$
313.74
500,001-1 Million
$
39.22
4.0
$
156.87
8.0
$
313.74
8.0
$
313.74
> 1 Million
$
39.22
4.0
$
156.87
8.0
$
313.74
8.0
$
313.74
Transient Noncommunity Water Systems (TNCWSs) - GW
<100
$
39.22
1.0
$
39.22
2.0
78.44
2.0
78.44
101-500
$
39.22
1.0
$
39.22
2.0
78.44
2.0
78.44
501-1,000
$
39.22
2.0
$
78.44
4.0
$
156.87
4.0
$
156.87
1,001-4,100
$
39.22
2.0
$
78.44
4.0
$
156.87
4.0
$
156.87
4,101-33,000
$
39.22
3.0
$
117.65
6.0
$
235.31
6.0
$
235.31
33,001-96,000
$
39.22
4.0
$
156.87
8.0
$
313.74
8.0
$
313.74
96,001-500,000
$
39.22
4.0
$
156.87
8.0
$
313.74
8.0
$
313.74
500,001-1 Million
$
39.22
4.0
$
156.87
8.0
$
313.74
8.0
$
313.74
> 1 Million
$
39.22
4.0
$
156.87
8.0
$
313.74
8.0
$
313.74
Sources:
(A) Labor rates for state employee from Section 7.2.1.
(B), (D), (E) Labor hour assumptions based on best professional judgment.
Economic Analysis for the Final RTCR
7-27
September 2012
-------
Exhibit 7.15 Annualized PWS and State Cost Estimates for Level 1 and Level 2
Assessments ($Millions, 2007$)
PWSs
State
Total
PWSs
State
Total
3% Discount Rate
7% Discount Rate
Level 1 Assessment
1989 TCR-Total
$
1.13
$
0.21
$
1.34
$
1.08
$
0.20
$
1.29
RTCR-Total
$
1.63
$
0.20
$
1.84
$
1.57
$
0.20
$
1.77
RTCR - Net Change
$
0.51
$
(0.01)
$
0.50
$
0.49
$
(0.01)
$
0.48
Alternative Option - Total
$
1.76
$
0.23
$
1.99
$
1.72
$
0.23
$
1.94
Alternative Option - Net Change
$
0.63
$
0.02
$
0.65
$
0.63
$
0.02
$
0.65
Level 2 Assessment
1989 TCR-Total
$
0.70
$
0.26
$
0.96
$
0.68
$
0.25
$
0.92
RTCR-Total
$
0.90
$
0.19
$
1.08
$
0.88
$
0.18
$
1.06
RTCR - Net Change
$
0.20
$
(0.07)
$
0.12
$
0.20
$
(0.07)
$
0.13
Alternative Option - Total
$
1.26
$
0.29
$
1.55
$
1.30
$
0.31
$
1.61
Alternative Option - Net Change
$
0.55
$
0.03
$
0.58
$
0.62
$
0.06
$
0.68
Note: Detail may not add due to independent rounding.
Source: Final RTCR Cost Model.
7.4.6 Corrective Actions
PWSs
Under the RTCR and Alternative option, all PWSs would be required to correct sanitary
defects found through the performance of Level 1 or Level 2 assessments. For modeling
purposes, EPA estimated only the net change in the number of corrective actions performed
under the RTCR and Alternative option55 compared to the 1989 TCR. Based on discussions with
state representatives, EPA estimates that additional corrective actions (on top of those already
being performed under the 1989 TCR) would be performed for only 10% of the assessments
undertaken as a result of the RTCR. Because only the net change in costs is estimated, no
additional costs for corrective actions are modeled for the 1989 TCR (it is assumed that PWSs
are already performing some corrective actions under the 1989 TCR).
To estimate the costs incurred for the correction of sanitary defects, EPA estimated the
percent distribution of PWSs that would perform different types of corrective actions as
presented in the compliance forecast below (Exhibit 7.16a). The compliance forecast presented
below focuses on the higher level categorization of corrective actions anticipated. The categories
of anticipated corrective actions were informed by both EPA judgment and discussions of the
TCRDSAC TWG and are essentially the same as those presented to the advisory committee. For
each of the categories listed, a PWS is assumed to take a specific action that falls under that
general category. Exhibit 7.16b lists the specific corrective actions that fall under the higher level
categorization, along with the percent of PWSs estimated to perform that corrective action based
on level 1 and level 2 assessments. Detailed compliance forecasts showing the specific corrective
actions used in the cost analysis are provided in Appendix D, along with summary tables of the
55 Any corrective actions based on a positive source water sample are assumed to be accounted for under the GWR
and not under the RTCR.
Economic Analysis for the Final RTCR
7-28
September 2012
-------
unit costs used in the analysis. Each corrective action in the detailed compliance forecast is also
assigned a representative unit cost. Detailed descriptions of the derivation of unit costs are
provided in Exhibits 5-1 through 5-47 of the Technology and Cost Document for the Final
Revised Total ColiformRule (USEPA, 2010d).
As shown in the compliance forecast in Exhibit 7.16a, EPA estimates that corrective
actions found through Level 1 assessments would result in corrective actions that focus more on
actions by PWS staff such as flushing or training (columns A and B) than on permanent fixes to
the PWS. This reflects the assumption that Level 1 assessments would generally be less involved
than Level 2 assessments and may not result in finding more complex problems. Corrective
actions taken as a result of Level 2 assessments are expected to find a higher proportion of
structural/technical issues (columns C-K) resulting in material fixes to the PWSs and distribution
system. Consistent with the discussions of the TCRDSAC regarding major structural fixes or
replacements, EPA did not include these major costs in the analysis. Distribution system
appurtenances such as storage tanks generally have a useful life that is accounted for in water
system capital planning and the assessments conducted in response to RTCR triggers could
identify when that useful life has ended but are not solely responsible for the need to correct the
defect.56
It is not anticipated that the overall effectiveness of corrective actions will differ
significantly by the type of corrective action. Some activities, such as spot flushing, may have a
shorter, transient impact on water quality. However, most PWSs are expected to institute
ongoing flushing programs that will have a continuing impact on water quality. Other corrective
actions in the compliance forecast are expected to have ongoing benefits, though the exact
duration may be highly variable depending on the exact conditions at the PWS. The overall
duration of reduced occurrence modeled as a result of corrective actions initiated in response to
either a Level 1 or Level 2 assessment is intended to reflect an average effectiveness of
corrective actions. Uncertainties and associated sensitivity analyses that inform the effect of the
reduced occurrence predicted from corrective actions are discussed further in Section 7.7.
PWSs would also incur reporting and recordkeeping burden to notify the state upon
completion of each corrective action. PWSs may also consult with the state or with outside
parties to determine the appropriate corrective action to be implemented. PWS reporting and
recordkeeping costs (including consultations) are derived by multiplying PWSs labor rates
(Section 7.2.1) by an EPA-estimated labor burden. Exhibit 7.17 presents the estimated unit costs
for this reporting and recordkeeping burden.
56 Additionally, EPA ran two sensitivity analyses to assess the potential impacts of different distributions within the
compliance forecast. Results of the sensitivity analyses are presented in Section 7.7, and indicate that the low bound
estimates of annualized net change in costs at three percent discount rate are approximately $3M for the RTCR and
$ 15M for the Alternative option, and the high bound estimates are approximately $25M for the RTCR and $40M for
the Alternative option. Varying the assumptions about the percentage of corrective actions identified and the
effectiveness of those actions had less than a linear effect on outcomes, and the RTCR continues to be less costly
than the Alternative option under all scenarios modeled.
Economic Analysis for the Final RTCR
7-29
September 2012
-------
States
For each corrective action performed under the RTCR and Alternative option, states
would incur recordkeeping and reporting burden to review and coordinate with PWSs. This
includes burden incurred from any optional consultations states may conduct with PWSs or
outside parties to determine the appropriate corrective action to be implemented. The state labor
rate presented in Section 7.2.1 and estimates of labor hours are used to generate state unit
corrective action costs. Exhibit 7.17 presents the estimated PWS and state reporting and
recordkeeping unit costs (including consultations) by PWS size.
Exhibit 7.16a Compliance Forecast for Corrective Actions based on Level 1 and
Level 2 Assessments
Maintenance
Cross-
Addition or
Development
Replace/Repair
of
connection
Upgrade of
Addition
and
of Distribution
Maintenance
appropriate
Storage
Control and
On-line
of
Implementation
PWS Size
PWS
Sampler
System
of Adequate
Hydraulic
Facility
Booster
Backflow
Monitoring
Security
of an
(Population
Flushing
Training
Components
Pressure
Residence
Maintenance
Disinfection
Prevention
and Control
Measures
Operations Plan
Served)
A
B
C
D
E
F
G
H
I
J
K
Level 1 Compliance Forecast
<100
39%
15%
12%
9%
8%
6%
4%
1%
3%
1%
2%
101-500
39%
15%
12%
9%
8%
6%
4%
1%
3%
1%
2%
501-1,000
39%
15%
12%
9%
8%
6%
4%
1%
3%
1%
2%
1,001-4,100
39%
15%
12%
9%
8%
6%
4%
1%
3%
1%
2%
4,101-33,000
39%
15%
12%
9%
8%
6%
4%
1%
3%
1%
2%
33,001-96,000
39%
15%
12%
9%
8%
6%
4%
1%
3%
1%
2%
96,001-500,000
39%
15%
12%
9%
8%
6%
4%
1%
3%
1%
2%
500,001-1 Million
39%
15%
12%
9%
8%
6%
4%
1%
3%
1%
2%
> 1 Million
39%
15%
12%
9%
8%
6%
4%
1%
3%
1%
2%
Level 2 Compliance Forecast
<100
15%
4%
18%
15%
15%
11%
8%
2%
6%
2%
4%
101-500
15%
4%
18%
15%
15%
11%
8%
2%
6%
2%
4%
501-1,000
15%
4%
18%
15%
15%
11%
8%
2%
6%
2%
4%
1,001-4,100
15%
4%
18%
15%
15%
11%
8%
2%
6%
2%
4%
4,101-33,000
15%
4%
18%
15%
15%
11%
8%
2%
6%
2%
4%
33,001-96,000
15%
4%
18%
15%
15%
11%
8%
2%
6%
2%
4%
96,001-500,000
15%
4%
18%
15%
15%
11%
8%
2%
6%
2%
4%
500,001-1 Million
15%
4%
18%
15%
15%
11%
8%
2%
6%
2%
4%
> 1 Million
15%
4%
18%
15%
15%
11%
8%
2%
6%
2%
4%
Source: (A) - (K) Percent of FWSs performing corrective actions based on Level 1 and Level 2 assessments re
ect EPA estimate.
Economic Analysis for the Final RTCR
7-30
September 2012
-------
Exhibit 7.16b Detailed PWS Compliance Forecast for Corrective Actions based on
Level 1 and Level 2 Assessments
General Corrective Action
Category
Specific Corrective Action
Level 1
Compliance
Forecast
Level 2
Compliance
Forecast
PWS Flushing
Routine Flushing
30%
10%
Spot Flushing
9%
5%
Sampler Training
Operator Training/Certification
15%
4%
Replace/Repair of Distribution
System Components
Replace Valve
2%
3%
Installing a Dedicated Sampling Tap
2%
3%
Replace Iron Pipe
2%
3%
Replace Fittings
2%
3%
Replace Hydrant
2%
3%
Replace Meter
2%
3%
Maintenance of Adequate
Pressure
New Booster Pump Station
2%
3%
Modify/Replace Existing Pumps
2%
3%
Install Variable Frequency Drives
2%
3%
Maintenance of Adequate
Pressure
New Elevated Storage Tank
0%
0%
Install Surge Relief valve
2%
3%
Install Surge Control Tank
1%
3%
Maintenance of appropriate
Hydraulic Residence Time
Install Auto Flushing Devices
2%
3%
Modify Inlet/Outlet Piping
1%
2%
Install Mixing Devices
1%
2%
Loop Dead Ends
1%
2%
Install Appropriate Main Sizes
1%
2%
Modify Storage Operation
1%
2%
Decommission Storage
1%
2%
Storage Facility Maintenance
Inspect & Clean Storage Tanks
2%
3%
Line Storage Tanks
1%
2%
Vent Repair/Replace Vent Screen
1%
2%
Storage Facility Maintenance
Repair/Replace Tank Hatch
1%
2%
Repair Storage Tank
1%
2%
Booster Disinfection
Install Permanent CI Booster Station
1%
2%
Install Temp CI Booster Station
1%
2%
Install Permanent NH2CI Booster Station
1%
2%
Install Temp NH2CI Booster Station
1%
2%
Cross-connection Control and
Backflow Prevention Program
Backflow Prevention Assembly
1%
2%
Addition or Upgrade of On-line
Monitoring and Control
Online CI Monitoring & Programming
1%
2%
Online NH2CI Monitoring & Programming
1%
2%
Online Pressure Monitoring & Programming
1%
2%
Addition of Security Measures
Installation of Additional Security Measures
1%
2%
Development and
Implementation of an
Operations Plan
Develop/lmplement/Maintain Operations
Plan
1%
2%
Operator Training/Certification
1%
2%
Economic Analysis for the Final RTCR
7-31
September 2012
-------
Exhibit 7.17 Net Change in PWS and State Unit Costs Estimates for Reporting and
Recordkeeping for Corrective Actions (2007$)
PWS Size
(Population
Served)
Systems
States
Labor Cost
(per hour)
Corrective Action
Burden (hours/
corrective action)
Unit Cost
Labor Cost
(per hour)
Corrective Action
Burden (hours/
corrective action)
Unit Cost
A
B
C=A*B
D
E
F=D*E
<100
$ 25.10
0.5
$ 12.55
$ 39.22
0.5
$ 19.61
101-500
$ 27.03
0.5
$ 13.52
$ 39.22
0.5
$ 19.61
501-1,000
$ 28.96
0.5
$ 14.48
$ 39.22
0.5
$ 19.61
1,001-4,100
$ 29.73
0.5
$ 14.87
$ 39.22
0.5
$ 19.61
4,101-33,000
$ 36.00
0.5
$ 18.00
$ 39.22
0.5
$ 19.61
33,001-96,000
$ 36.39
0.5
$ 18.20
$ 39.22
0.5
$ 19.61
96,001-500,000
$ 41.01
0.5
$ 20.51
$ 39.22
0.5
$ 19.61
500,001-1 Million
$ 41.01
0.5
$ 20.51
$ 39.22
0.5
$ 19.61
> 1 Million
$ 41.01
0.5
$ 20.51
$ 39.22
0.5
$ 19.61
Notes:
FWS and state burden estimates identical for all R/VS types under 1989 TCR, RTCR, and Alternative Option.
Sources:
(A) Labor rates for R/VSs from Exhibit 7.1.
(B) R/VS labor hours for corrective action recordkeeping/reporting reflect EPA estimate.
(D) Labor rates for state employee from Section 7.2.1.
(E) State labor hours for corrective action recordkeeping/reporting reflect EPA estimate.
Corrective Action Net Cost Summary
Annualized net cost estimates for PWSs and states to perform corrective actions are
estimated by multiplying the number of Level 1 and Level 2 corrective actions estimated by the
predictive model (i.e., 10 percent of Level 1 and Level 2 assessments representing the net
increase in corrective actions found), by the percentages in the compliance forecast and unit
costs of corrective actions and associated reporting and recordkeeping. Total and net change in
annualized present value cost estimates for PWSs and states to perform corrective actions under
the 1989 TCR, RTCR, and Alternative option are presented in Exhibit 7.18.
Because only the net change in corrective actions taken is modeled, no costs are
estimated under the 1989 TCR. The differences in the net change in corrective action costs
between the RTCR and Alternative option are a function different number of assessments
estimated to be performed in the predictive model as discussed in Section 7.4.5 above.
Economic Analysis for the Final RTCR
7-32
September 2012
-------
Exhibit 7.18 Annualized PWS and State Cost Estimates for Corrective Actions
based on Level 1 and Level 2 Assessments ($Millions, 2007$)
PWSs
State
Total
PWSs
State
Total
3% Discount Rate
7% Discount Rate
Corrective Actions based on Level 1 Assessments
1989 TCR-Total
$
$
$
$
$
$
RTCR - Total
$ 9.62
$ 0.01
$ 9.63
$ 8.14
$ 0.01
$ 8.15
Alternative Option - Total
$ 10.01
$ 0.01
$ 10.02
$ 8.52
$ 0.01
$ 8.53
RTCR - Net Change
$ 9.62
$ 0.01
$ 9.63
$ 8.14
$ 0.01
$ 8.15
Alternative Option - Net Change
$ 10.01
$ 0.01
$ 10.02
$ 8.52
$ 0.01
$ 8.53
Corrective Actions based on Level 2 Assessments
1989 TCR-Total
$
$
$
$
$
$
RTCR - Total
$ 2.82
$ 0.00
$ 2.82
$ 2.49
$ 0.00
$ 2.49
Alternative Option - Total
$ 3.78
$ 0.01
$ 3.79
$ 3.57
$ 0.01
$ 3.58
RTCR - Net Change
$ 2.82
$ 0.00
$ 2.82
$ 2.49
$ 0.00
$ 2.49
Alternative Option - Net Change
$ 3.78
$ 0.01
$ 3.79
$ 3.57
$ 0.01
$ 3.58
Note: Detail may not add due to independent rounding.
Source: Final RTCR Cost Model.
7.4.7 Public Notification
PWSs
Tier 1 Public Notification
Acute violations (E. coli MCL violations) would require Tier 1 PN under all regulatory
options (1989 TCR, RTCR, and Alternative option). PWSs with acute violations must report the
violation to the state by the end of the business day after the PWS learns of the violation and
must notify the public within 24 hours.
Tier 2 Public Notification
PWSs with non-acute violations under the 1989 TCR must report the violation to the state
by the end of the business day after the PWS learns of the violation, and must provide Tier 2 PN
within 30 days. PWSs with situations similar to 1989 TCR non-acute violations under the RTCR
and Alternative option are not required to perform PN. These PWSs are required to perform Tier
2 notification for a treatment technique violation (failure to perform Level 1 or 2 assessment if
triggered; failure to correct all sanitary defects found), but because the cost model assumes full
compliance with RTCR requirements, no cost is estimated for these violations. Overall, costs
decrease significantly for Tier 2 PN under the Final Rule for both the RTCR and Alternative
option.
Tier 3 Public Notification
Under the 1989 TCR, RTCR, and Alternative option, Tier 3 PN for monitoring and
reporting violations are assumed to be reported once per year as part of the consumer confidence
report (CCR). Because of the use of the CCR to communicate Tier 3 PN on a yearly basis, no
cost differential between the 1989 TCR and the RTCR and Alternative option is estimated in the
cost model. However, although they are not quantitatively evaluated as part of the EA, the
Economic Analysis for the Final RTCR
7-33
September 2012
-------
TCRDSAC concluded that significant reductions in monitoring and reporting violations may be
realized through the revised regulatory framework of the RTCR. See Section 7.6 for a discussion
of non-quantified costs.
Estimates of PWS unit costs for PN are derived by multiplying PWS labor rates from
Section 7.2.1 and burden hour estimates derived from the Information Collection Request for the
Public Water System Supervision Program (USEPA, 2008c). PWS PN unit cost estimates are
presented in Exhibit 7.19.
Economic Analysis for the Final RTCR
7-34
September 2012
-------
Exhibit 7.19 PWS Unit Cost Estimates for Public Notification (2007$)
Tier 1 (acute)
Tier 2 (non-acute)
PWS Size
(Population
Served)
Labor Cost
(per hour)
Average Number
of Service
Connections per
System
Preparation
(labor) (hours/
violation)
Distribution
(labor)
(hours/
violation)
Distribution
(O&M Cost/
notice)
($/service
connection)
Unit Cost
Preparation
(labor)
(hours/
violation)
Distribution
(labor) (hours/
violation)
Distribution
(O&M Cost/notice)
($/service
connection)
Unit Cost
A
B
C
D
E
F=A*(C+D)+(B*E)
G
H
I
J=A*(G+H)+(B*I)
Community Water Systems (CWSs) - SW
<100
$ 25.10
440
8.5
12.0
$ 0.05
$ 536.55
3.5
9.0
$ 0.05
$ 335.75
101-500
$ 27.03
387
8.5
12.0
$ 0.05
$ 573.44
3.5
9.0
$ 0.05
$ 357.20
501-1,000
$ 28.96
303
8.5
12.0
$
$ 593.68
3.5
30.0
$ 0.26
$ 1,049.05
1,001-4,100
$ 29.73
850
8.5
12.0
$
$ 609.47
3.5
30.0
$ 0.26
$ 1,216.99
4,101-33,000
$ 36.00
4,288
9.2
12.0
$
$ 763.66
3.5
30.0
$ 0.25
$ 2,259.65
33,001-96,000
$ 36.39
17,273
10.0
12.0
$
$ 800.58
3.5
30.0
$ 0.23
$ 5,191.88
96,001-500,000
$ 41.01
56,465
10.0
12.0
$
$ 902.22
3.5
30.0
$ 0.23
$ 14,360.67
500,001-1 Million
$ 41.01
205,609
10.0
12.0
$
$ 902.22
3.5
30.0
$ 0.23
$ 48,663.97
> 1 Million
$ 41.01
448,564
10.0
12.0
$
$ 902.22
3.5
30.0
$ 0.23
$ 104,543.57
Community Water Systems (CWSs) - GW
<100
$ 25.10
41
8.5
12.0
$ 0.05
$ 516.62
3.5
9.0
$ 0.05
$ 315.82
101-500
$ 27.03
99
8.5
12.0
$ 0.05
$ 559.05
3.5
9.0
$ 0.05
$ 342.81
501-1,000
$ 28.96
315
8.5
12.0
$
$ 593.68
3.5
30.0
$ 0.26
$ 1,052.07
1,001-4,100
$ 29.73
756
8.5
12.0
$
$ 609.47
3.5
30.0
$ 0.26
$ 1,192.61
4,101-33,000
$ 36.00
3,495
9.0
12.0
$
$ 757.48
3.5
30.0
$ 0.25
$ 2,076.93
33,001-96,000
$ 36.39
16,366
10.0
12.0
$
$ 800.58
3.5
30.0
$ 0.23
$ 4,983.17
96,001-500,000
$ 41.01
50,564
10.0
12.0
$
$ 902.22
3.5
30.0
$ 0.23
$ 13,003.50
500,001-1 Million
$ 41.01
209,220
10.0
12.0
$
$ 902.22
3.5
30.0
$ 0.23
$ 49,494.49
> 1 Million
$ 41.01
473,641
10.0
12.0
$
$ 902.22
3.5
30.0
$ 0.23
$ 110,311.34
Nontransient Noncomm unity Water Systems (NTNCWSs) - SW
<100
$ 25.10
128
8.5
12.0
$ 0.05
$ 520.94
3.5
9.0
$ 0.05
$ 320.14
101-500
$ 27.03
21
8.5
12.0
$ 0.05
$ 555.16
3.5
9.0
$ 0.05
$ 338.92
501-1,000
$ 28.96
46
8.5
12.0
$
$ 593.68
3.5
9.0
$ 0.05
$ 364.29
1,001-4,100
$ 29.73
47
8.5
12.0
$
$ 609.47
3.5
9.0
$ 0.05
$ 373.97
4,101-33,000
$ 36.00
176
8.7
12.0
$
$ 745.36
3.5
9.0
$ 0.05
$ 458.08
33,001-96,000
$ 36.39
94
10.0
12.0
$
$ 800.58
3.5
9.0
$ 0.02
$ 456.75
96,001-500,000
$ 41.01
2,181
10.0
12.0
$
$ 902.22
3.5
9.0
$ 0.02
$ 556.25
500,001-1 Million
$ 41.01
-
10.0
12.0
$
$ 902.22
3.5
9.0
$ 0.02
$ 512.63
> 1 Million
$ 41.01
-
10.0
12.0
$
$ 902.22
3.5
9.0
$ 0.02
$ 512.63
Nontransient Noncomm unity Water Systems (NTNCWSs) - GW
<100
$ 25.10
4
8.5
12.0
$ 0.05
$ 514.76
3.5
9.0
$ 0.05
$ 313.96
101-500
$ 27.03
8
8.5
12.0
$ 0.05
$ 554.52
3.5
9.0
$ 0.05
$ 338.28
501-1,000
$ 28.96
11
8.5
12.0
$
$ 593.68
3.5
9.0
$ 0.05
$ 362.56
1,001-4,100
$ 29.73
42
8.5
12.0
$
$ 609.47
3.5
9.0
$ 0.05
$ 373.71
4,101-33,000
$ 36.00
130
8.7
12.0
$
$ 744.94
3.5
9.0
$ 0.05
$ 456.02
33,001-96,000
$ 36.39
75
10.0
12.0
$
$ 800.58
3.5
9.0
$ 0.02
$ 456.38
96,001-500,000
$ 41.01
-
10.0
12.0
$
$ 902.22
3.5
9.0
$ 0.02
$ 512.63
500,001-1 Million
$ 41.01
-
10.0
12.0
$
$ 902.22
3.5
9.0
$ 0.02
$ 512.63
> 1 Million
$ 41.01
-
10.0
12.0
$
$ 902.22
3.5
9.0
$ 0.02
$ 512.63
Transient Noncomm unity Water Systems (TNCWSs) - SW
<100
$ 25.10
9
8.5
12.0
$ 0.05
$ 514.99
3.5
9.0
$ 0.05
$ 314.19
101-500
$ 27.03
30
8.5
12.0
$ 0.05
$ 555.62
3.5
9.0
$ 0.05
$ 339.38
501-1,000
$ 28.96
49
8.5
12.0
$
$ 593.68
3.5
9.0
$ 0.05
$ 364.45
1,001-4,100
$ 29.73
58
8.5
12.0
$
$ 609.47
3.5
9.0
$ 0.05
$ 374.50
4,101-33,000
$ 36.00
57
8.8
12.0
$
$ 747.00
3.5
9.0
$ 0.05
$ 452.55
33,001-96,000
$ 36.39
-
10.0
12.0
$
$ 800.58
3.5
9.0
$ 0.02
$ 454.88
96,001-500,000
$ 41.01
-
10.0
12.0
$
$ 902.22
3.5
9.0
$ 0.02
$ 512.63
500,001-1 Million
$ 41.01
-
10.0
12.0
$
$ 902.22
3.5
9.0
$ 0.02
$ 512.63
> 1 Million
$ 41.01
2
10.0
12.0
$
$ 902.22
3.5
9.0
$ 0.02
$ 512.67
Transient Noncomm unity Water Systems (TNCWSs) - GW
<100
$ 25.10
5
8.5
12.0
$ 0.05
$ 514.82
3.5
9.0
$ 0.05
$ 314.02
101-500
$ 27.03
15
8.5
12.0
$ 0.05
$ 554.87
3.5
9.0
$ 0.05
$ 338.63
501-1,000
$ 28.96
30
8.5
12.0
$
$ 593.68
3.5
9.0
$ 0.05
$ 363.51
1,001-4,100
$ 29.73
43
8.5
12.0
$
$ 609.47
3.5
9.0
$ 0.05
$ 373.77
4,101-33,000
$ 36.00
39
8.7
12.0
$
$ 746.87
3.5
9.0
$ 0.05
$ 451.77
33,001-96,000
$ 36.39
14
10.0
12.0
$
$ 800.58
3.5
9.0
$ 0.02
$ 455.15
96,001-500,000
$ 41.01
9
10.0
12.0
$
$ 902.22
3.5
9.0
$ 0.02
$ 512.81
500,001-1 Million
$ 41.01
1
10.0
12.0
$
$ 902.22
3.5
9.0
$ 0.02
$ 512.65
> 1 Million
$ 41.01
-
10.0
12.0
$
$ 902.22
3.5
9.0
$ 0.02
$ 512.63
Notes:
(B) Service connections per system is consistent w ith SDWIS 2007 4th Quarter Freeze data. Data for certain size categories (e.g., <100, 101-500, 501-1,000) may seem counterintuitive. EPA is
investigating the SDWIS database for any data discrepancies.
(F), (J) used to derived 1989 TCR PN costs; (F) used to derive RTCR and Alternative Option PN costs.
Sources:
(A) Labor rates for FWSs from Bchibit 7.1.
(B) SDWIS 2007 4th Quarter Freeze.
(C), (D), (G), (H) Labor hour assumptions based on best professional judgment as carried forward from the Information Collection Request for the Public Water System Supervision Program
(USEPA, 2008c).
(E), (I) Distribution cost assumptions based on best professional judgment as carried forward from the original FN ICR.
Economic Analysis for the Final RTCR
7-35
September 2012
-------
States
Under the 1989 TCR, RTCR, and Alternative option, states would incur recordkeeping
and reporting burden to provide consultation, review the PN certification, and file the report of
the violation. State unit PN costs are based on the state labor rate presented in Section 7.2.1 and
burden hour estimates derived from the Information Collection Request for the Public Water
System Supervision Program (USEPA, 2008c). Estimates of state unit costs for PN are presented
in Exhibit 7.20.
Exhibit 7.20 State Unit Costs Estimates for Public Notification
(1989 TCR, RTCR, Alternative Option)
Labor Cost
(per hour)
Tier 1
Consultation
(labor)
(hours/violation)
Tier 2
Consultation
(labor)
(hours/violation)
Receive/Review
PN Certification
(labor)
(hours/violation)
Hie Reports
(labor)
(hours/violation)
Tier 1
Unit Cost
Tier 2
Unit Cost
A
B
C
D
E
W(BHHE)
G=A*(C+I>E)
$ 39.22
3.0
1.1
0.2
0.1
$ 129.42
$ 54.91
Sources:
(A) Labor rate for state errployee fromSection 7.2.1.
(B), (C), (D), (E) Labor hour assumptions based on best professional judgment as carried forward from the Information Collection Request for
the Public Water System Supervision Program (USEPA, 2008c).
Public Notification Net Cost Summary
Total and net change in annualized present value costs for PN are estimated by
multiplying the model estimates of PWSs with acute (Tier 1 PN) and non-acute (Tier 2 PN)
violations by the PWS and state unit costs for performing PN activities. The RTCR Cost Model
assumes that all violations are addressed following initial PN, and no burden would be incurred
by PWSs or states for repeat notification. Exhibit 7.21 summarizes the total number of Tier 1 and
Tier 2 PNs that would be prepared by PWS size and category for under the 1989 TCR, RTCR,
and Alternative option over the entire 25-year period of analysis. Total and net change in
annualized present value cost estimates for PWSs and states to perform PN under the 1989 TCR,
RTCR, and Alternative option are presented in Exhibit 7.22.
A significant reduction in costs is estimated due to the elimination of Tier 2 PN for TC
occurrence (non-acute violations under the 1989 TCR) under the RTCR and Alternative option.
Because state costs are calculated on a per-violation basis, state costs decline. Under the
Alternative option, some of this cost decrease is offset by additional Tier 1 PN.
Economic Analysis for the Final RTCR
7-36
September 2012
-------
Exhibit 7.21 Number of Tier 1 and Tier 2 Public Notifications over the 25-Year
Compliance Period
Alternative
1989
TCR
RTCR
Option
PWS Size
(Population
Tier 1 PN
Tier 2 PN
Tier 1 PN
Tier 1 PN
Served)
A
B
C
D
Community Water Systems (CWSs) -SW
<100
157
709
100
100
101-500
167
760
119
119
501-1,000
102
424
75
75
1,001-4,100
162
1,103
146
146
4,101-33,000
197
2,152
197
197
33,001-96,000
56
534
56
56
96,001-500,000
24
233
24
24
500,001-1 Million
22
> 1 Million
Total
865
5,936
717
717
Community Water Systems (CWSs) -GW
<100
1,141
15,155
853
926
101-500
1,025
12,383
696
747
501-1,000
284
3,300
188
203
1,001-4,100
477
6,354
342
342
4,101-33,000
263
4,545
263
263
33,001-96,000
53
656
53
53
96,001-500,000
10
129
10
10
500,001-1 Million
> 1 Million
Total
3,253
42,521
2,405
2,544
Nontransient Noncommunity Water Systems (NTNCWSs) - SW
<100
35
175
28
28
101-500
29
128
19
19
501-1,000
9
42
6
6
1,001-4,100
19
79
13
13
4,101-33,000
5
33,001-96,000
96,001-500,000
500,001-1 Million
> 1 Million
Total
93
429
67
67
Nontransient Noncommunity Water Systems (NTNCWSs) - ON
<100
856
9,411
559
723
101-500
447
4,402
315
446
501-1,000
95
1,042
79
99
1,001-4,100
169
1,792
114
114
4,101-33,000
9
123
9
9
33,001-96,000
4
96,001-500,000
500,001-1 Million
> 1 Million
Total
1,577
16,774
1,077
1,393
Transient Noncommunity Water Systems (TNCWSs) - SW
<100
430
1,872
250
250
101-500
170
729
90
90
501-1,000
28
115
17
17
1,001-4,100
50
220
29
29
4,101-33,000
8
33,001-96,000
96,001-500,000
500,001-1 Million
> 1 Million
Total
677
2,945
386
386
Transient Noncommunity Water Systems (TNCWSs) - GW
<100
6,649
70,156
5,477
7,796
101-500
2,089
23,394
1,608
2,441
501-1,000
221
2,373
177
230
1,001-4,100
186
2,065
117
117
4,101-33,000
4
116
4
4
33,001-96,000
96,001-500,000
500,001-1 Million
> 1 Million
Total
9,149
98,104
7,383
10,589
Grand Total
15,613
166,709
12,035
15,695
Note: Detail may not add due to rounding.
Sources:
(A), (B) - A.1 .z
(C)-A.2.Z
(D) - A.3.Z
Economic Analysis for the Final RTCR
7-37
September 2012
-------
Exhibit 7.22 Annualized PWS and State Cost Estimates for Public Notification
($Millions, 2007$)
PWSs
State
Total
PWSs
State
Total
3% Discount Rate
7% Discount Rate
1989 TCR-Total
$ 3.75
$ 0.44
$ 4.19
$ 3.60
$ 0.42
$ 4.02
RTCR - Total
$ 0.26
$ 0.06
$ 0.32
$ 0.25
$ 0.06
$ 0.31
RTCR - Net Change
$ (3.49)
$ (0.38)
$ (3.86)
$ (3.35)
$ (0.36)
$ (3.71)
RTCR - Percent Change
-93%
-86%
-92%
-93%
-86%
-92%
Alternative Option - Total
$ 0.35
$ 0.08
$ 0.43
$ 0.35
$ 0.08
$ 0.44
Alternative Option - Net Change
$ (3.40)
$ (0.36)
$ (3.76)
$ (3.25)
$ (0.34)
$ (3.58)
Alternative Option - Percent Change
-91%
-81%
-90%
-90%
-80%
-89%
Note: Detail may not add due to independent rounding.
Source: Final RTCRCost Model.
7.4.8 Uncertainty in Unit Costs
As stated in Section 7.2.5, EPA recognizes that there are both variability and uncertainty
in unit cost estimates used to develop national costs for the RTCR. Variability is expected in the
actual costs that would be experienced by different PWSs of similar size conducting the same
corrective action. Otherwise similar PWSs may experience different capital and/or O&M costs
due to site-specific factors. Inputs to unit costs such as water quality conditions, labor rates, and
land costs can be highly variable and increase the system-to-system variability in unit costs. In
developing the unit cost estimates, there is insufficient information to fully characterize what the
distribution of this variability would be on a national scale; therefore, EPA uses mean values for
these input parameters.
The unit costs used in this EA are developed as average or representative estimates of
what these unit costs would be nationally, and are not the unit costs of any particular PWS. Some
components of monitoring costs, such as the purchase and wear-and-tear of vehicles, are not
quantified because of either limited data or inability to attribute these costs directly to the RTCR.
Additionally, the PWS and state labor hours for each rule component (specifically for Level 1
and Level 2 assessments) are meant to capture national averages for the purposes of developing
national cost estimates and making comparisons between regulatory options. Thus, the unit costs
presented in this document may over- or underestimate the unit costs of any particular PWS.
Detailed information on the derivation of unit costs for each rule component is provided in the
Technology and Cost Document for the Final Revised Total Coliform Rule (USEPA, 2010d).
7.5 Household Costs
The household cost analysis considers the impact that the costs incurred by CWSs have
on the households they serve. This analysis considers the potential increase in a household's
annual water bill if a CWS passed the entire cost increase resulting from the rule on to their
customers. This analysis is a tool to gauge potential impacts and should not be construed as a
precise estimate of potential changes to household water bills. State costs and costs to TNCWSs
and NTNCWSs are not included in this analysis since their costs are not typically passed through
directly to households.
Economic Analysis for the Final RTCR
7-38
September 2012
-------
To calculate household costs (which are in the units of $ per household per year), the
CWS population subject to the RTCR by PWS size is divided by the average number of people
per household, estimated as 2.56 for the year 2007 (United States Census Bureau, 2008), to
calculate the number of households subject to the RTCR by PWS size. The cost of the rule, by
size category, is then divided by that number of households to determine a per-household cost.
The first section of Exhibit 7.23 presents net costs per household under the RTCR and
Alternative option for all rule components spread across all CWSs. In this scenario, comparison
to the 1989 TCR shows a cost savings for households in the largest size category. For those
households that are expected to see a cost increase, the average annual water bill would be
expected to increase by less than ten cents on average. Although this average cost per household
is very low, customers served by PWSs that incur greater costs to comply with the RTCR would
be higher.
While the average increase in annual household water bills to implement the RTCR is
less than a dollar, customers served by a small CWS that have to take corrective actions as a
result of the rule would incur slightly larger increases in their water bills. The subsequent
sections of the exhibit present net costs per household for different subsections of CWSs (e.g.,
CWSs that perform assessments but no corrective actions, CWSs that do perform corrective
actions, and CWSs that do not perform assessments or corrective actions). As shown in the
second section of Exhibit 7.23, approximately 67% of households belong to CWSs that would
perform assessments but would not perform corrective actions (because no sanitary defects are
found). These households would experience a slight cost savings on an annual basis, due to a
slight reduction in monitoring and PN costs. The 9% of households belonging to CWSs that
would perform corrective actions would experience an increase in annual net household costs of
less than $1 to approximately $26 on an annual basis. The final section of the exhibit presents the
24% of households belonging to CWSs that would not perform assessments or corrective actions.
Households of this category would experience an increase in cost savings when compared to
those performing corrective actions, and a decrease in cost savings when compared to those
performing assessments but no corrective actions. This decrease in costs savings is because no
PN costs are associated with systems not performing assessments. Overall, the main driver of
additional household costs under the RTCR is whether or not additional corrective actions are
performed.
Economic Analysis for the Final RTCR
7-39
September 2012
-------
Exhibit 7.23 Summary of Net Annual Per-Household Costs for the RTCR (2007$)
PWS Size
(Population
Served)
Number of
Households
(RTCR)
Number of
Households
(Alternative Option)
3% Discount Rate
7% Discount Rate
RTCR Net
RTCR Net Cost per
Household
Alternative
Option Net
Alternative
Option Net
Cost per
Household
RTCR Net
RTCR Net
Cost per
Household
Alternative
Option Net
Alternative
Option Net
Cost per
Household
A
B
C
D=C/A
E
^E/B
G
H=G/A
I
J=l/B
All Community Water Systems (CWSs)
<100
307,243
307,243
$ 111,694
$ 0.364
$ 225,693
$ 0.735
$ 207,140
$ 0.674
$ 348,782
$ 1.135
101-500
1,589,510
1,589,510
$ 371,004
$ 0.233
$ 464,207
$ 0.292
$ 471,664
$ 0.297
$ 598,244
$ 0.376
501-1,000
1,624,853
1,624,853
$ 46,687
$ 0.029
$ 87,294
$ 0.054
$ 96,473
$ 0.059
$ 148,274
$ 0.091
1,001-4,100
7,816,592
7,816,592
$ 371,294
$ 0.048
$ 371,294
$ 0.048
$ 418,935
$ 0.054
$ 418,935
$ 0.054
4,101-33,000
27,997,647
27,997,647
$ 2,385,056
$ 0.085
$ 2,385,056
$ 0.085
$ 2,083,266
$ 0.074
$ 2,083,266
$ 0.074
33,001-96,000
21,933,438
21,933,438
$ 1,532,410
$ 0.070
$ 1,532,410
$ 0.070
$ 1,273,202
$ 0.058
$ 1,273,202
$ 0.058
96,001-500,000
26,770,609
26,770,609
$ 1,479,280
$ 0.055
$ 1,479,280
$ 0.055
$ 1,214,316
$ 0.045
$ 1,214,316
$ 0.045
500,001-1 Million
9,764,979
9,764,979
$ 157,138
$ 0.016
$ 157,138
$ 0.016
$ 125,671
$ 0.013
$ 125,671
$ 0.013
> 1 Million
16,309,853
16,309,853
$ (2,223)
$ (0.000)
$ (2,223)
$ (0.000)
$ (1,479)
$ (0.000)
$ (1,479)
$ (0.000)
Total
114,114,724
114,114,724
$ 6,452,342
$ 0.057
$ 6,700,151
$ 0.059
$ 5,889,190
$ 0.052
$ 6,209,211
$ 0.054
Community Water Systems (CWSs) performing Level 1/Level 2 Assessments (and no Corrective Actions
<100
125,340
124,920
$ (185,923)
$ (1.483)
$ (149,283)
$ (1.195)
$ (125,037)
$ (0.998)
$ (75,885)
$ (0.607)
101-500
460,577
464,568
$ (137,193)
$ (0.298)
$ (103,168)
$ (0.222)
$ (82,657)
$ (0.179)
$ (40,057)
$ (0.086)
501-1,000
394,643
401,009
$ (133,453)
$ (0.338)
$ (123,890)
$ (0.309)
$ (111,147)
$ (0.282)
$ (98,598)
$ (0.246)
1,001-4,100
2,341,578
2,341,578
$ (262,222)
$ (0.112)
$ (262,222)
$ (0.112)
$ (213,247)
$ (0.091)
$ (213,247)
$ (0.091)
4,101-33,000
24,827,588
24,827,588
$ (195,108)
$ (0.008)
$ (195,108)
$ (0.008)
$ (76,420)
$ (0.003)
$ (76,420)
$ (0.003)
33,001-96,000
19,232,570
19,232,570
$ (173,238)
$ (0.009)
$ (173,238)
$ (0.009)
$ (142,573)
$ (0.007)
$ (142,573)
$ (0.007)
96,001-500,000
23,912,325
23,912,325
$ (146,682)
$ (0.006)
$ (146,682)
$ (0.006)
$ (131,967)
$ (0.006)
$ (131,967)
$ (0.006)
500,001-1 Million
5,524,188
5,524,188
$ (40,160)
$ (0.007)
$ (40,160)
$ (0.007)
$ (36,993)
$ (0.007)
$ (36,993)
$ (0.007)
> 1 Million
-
-
$
$
$
$
$
$
$
$
Total
76,818,809
76,828,746
$ (1,273,980)
$ (0.017)
$ (1,193,752)
$ (0.016)
$ (920,040)
$ (0.012)
$ (815,739)
$ (0.011)
Community Water Systems (CWSs) performing Corrective Actions
<100
13,927
13,880
$ 365,576
$ 26.250
$ 388,703
$ 28.004
$ 330,235
$ 23.712
$ 353,002
$ 25.432
101-500
51,175
51,619
$ 516,102
$ 10.085
$ 508,146
$ 9.844
$ 452,342
$ 8.839
$ 450,070
$ 8.719
501-1,000
43,849
44,557
$ 178,205
$ 4.064
$ 181,189
$ 4.066
$ 155,596
$ 3.548
$ 158,538
$ 3.558
1,001-4,100
260,175
260,175
$ 590,719
$ 2.270
$ 590,719
$ 2.270
$ 512,568
$ 1.970
$ 512,568
$ 1.970
4,101-33,000
3,170,059
3,170,059
$ 2,605,551
$ 0.822
$ 2,605,551
$ 0.822
$ 2,198,499
$ 0.694
$ 2,198,499
$ 0.694
33,001-96,000
2,700,868
2,700,868
$ 1,709,801
$ 0.633
$ 1,709,801
$ 0.633
$ 1,424,758
$ 0.528
$ 1,424,758
$ 0.528
96,001-500,000
2,858,284
2,858,284
$ 1,625,742
$ 0.569
$ 1,625,742
$ 0.569
$ 1,346,704
$ 0.471
$ 1,346,704
$ 0.471
500,001-1 Million
613,799
613,799
$ 195,076
$ 0.318
$ 195,076
$ 0.318
$ 161,185
$ 0.263
$ 161,185
$ 0.263
> 1 Million
-
-
$
$
$
$
$
$
$
$
Total
9,712,136
9,713,240
$ 7,786,773
$ 0.802
$ 7,804,928
$ 0.804
$ 6,581,888
$ 0.678
$ 6,605,324
$ 0.680
Community Water Systems (CWSs) not performing Level 1/Level 2 Assessments, or Corrective Actions
<100
167,976
168,442
(67,959)
$ (0.405)
(13,728)
$ (0.081)
1,943
$ 0.012
71,665
$ 0.425
101-500
1,077,758
1,073,324
(7,905)
$ (0.007)
59,230
$ 0.055
101,979
$ 0.095
188,231
$ 0.175
501-1,000
1,186,361
1,179,288
1,936
$ 0.002
29,995
$ 0.025
52,024
$ 0.044
88,334
$ 0.075
1,001-4,100
5,214,839
5,214,839
42,797
$ 0.008
42,797
$ 0.008
119,614
$ 0.023
119,614
$ 0.023
4,101-33,000
-
-
-
$
-
$
-
$
-
$
33,001-96,000
-
-
-
$
-
$
-
$
-
$
96,001-500,000
-
-
-
$
-
$
-
$
-
$
500,001-1 Million
3,626,992
3,626,992
-
$
-
$
-
$
-
$
> 1 Million
16,309,853
16,309,853
-
$
-
$
-
$
-
$
Total
27,583,779
27,572,738
$ (31,132)
$ (0.001)
$ 118,294
$ 0.004
$ 275,560
$ 0.010
$ 467,844
$ 0.017
Source:
(C), (E), (G), (I) Bchibit 7.28.
Economic Analysis for the Final RTCR
7-40
September 2012
-------
7.6 Non-quantified Costs
EPA believes that all of the rule elements that are the major drivers of the net changes in
costs of the 1989 TCR have been quantified to the greatest degree possible. However, cost
reductions related to fewer possible monitoring and reporting violations are not specifically
accounted for in the cost analysis, and their exclusion from consideration may result in an
overestimate of net change in cost between the 1989 TCR and the RTCR or Alternative option.
In addition, under the 1989 TCR, RTCR, and Alternative option, Tier 3 PN for
monitoring and reporting violations are assumed to be reported once per year as part of the CCR.
Because of the use of the CCR to communicate Tier 3 PN on a yearly basis, no cost differential
between the 1989 TCR and the RTCR and Alternative option is estimated in the cost model.
However, the advisory committee concluded that significant reductions in monitoring and
reporting violations may be realized through the revised regulatory framework of the RTCR.
These possible reductions have not been quantified. System resources used to process monitoring
violation notices for the CCR and respond to customer inquiries about the notices as well as state
resources to remind systems to take samples, may be reduced if significant reductions are
realized. Exclusion of this potential cost savings may lead to an underestimate of the PN cost
savings under both the RTCR and Alternative option.
Additionally, as an underlying assumption to the costing methodology, EPA has assumed
that all PWSs subject to the RTCR requirements are already complying with the 1989 TCR.
There may be some PWSs that are not in full compliance with the 1989 TCR, and if so,
additional costs would be incurred.
7.7 Uncertainty Analysis
There are two primary sources of uncertainty in the RTCR cost modeling. The first is
related to the underlying estimates of events resulting from the rule revisions as generated from
the occurrence and predictive modeling. The occurrence and predictive modeling (discussed in
detail in Chapter 5) does not explicitly consider uncertainty, and therefore does not generate
confidence intervals on the predicted outcomes. However, EPA evaluated the model inputs to
determine which of the inputs were likely to have a significant effect on the results and subjected
those to further review through sensitivity analyses. In particular, EPA evaluated the impacts of
alternative estimates of the net change percentage of PWSs predicted to take corrective actions in
response to an assessment (10 percent in the model) and the associated effectiveness of those
corrective actions (Section 5.3.3.1 discusses these analyses in detail). The results of these
analyses suggest that changes in the major assumptions about the net change percentage of
corrective actions identified and the effectiveness of those actions have a less than linear effect
on predicted outcomes. When applied to costs estimates they have even less impact.
In the case where PWSs increase the percentage of corrective actions taken in response to
an assessment, costs are expected to increase corresponding to the greater number of corrective
actions taken. Using the increases predicted for the small subset of PWSs analyzed, it is
estimated that corrective actions (and related costs) would increase by 65% for Level 1
corrective actions and 7% for Level 2 corrective actions. This is in response to a doubling of the
net change percentage of corrective actions taken in response to an assessment. However, the
Economic Analysis for the Final RTCR
7-41
September 2012
-------
increased effectiveness of more corrective actions has a dampening effect on other costs that
would offset the increased corrective action costs. For instance, although the numbers of
corrective actions increase due to the higher percentages found during assessments, the actual
number of assessments decreases by 16% for Level 1 and by 43% for Level 2 because fewer
positive samples (ranging from a 14% decrease in routine TC+ to a 28% decrease in repeat TC+)
are predicted to be found. The overall effect of the cost increases and decreases in response to the
major predictive occurrence model uncertainties is expected to be minimal.
The other major area of uncertainty that may affect the resulting cost calculations is the
distribution of corrective actions taken by PWSs (the compliance forecast) in response to finding
a problem during a Level 1 or Level 2 assessment. The compliance forecast presented in Exhibits
7.16a and 7.16b represents EPA's best estimate of a distribution of corrective actions that may be
taken by PWSs to respond to problems identified under the RTCR. Because there is a wide
variation in the unit costs of the actual corrective actions underlying the compliance forecast (see
Appendix D for unit cost detail), EPA ran two sensitivity analyses to assess the potential impacts
of different distributions within the compliance forecast. These two sensitivity runs attempt to
bound the analysis of corrective action costs by generating a low and high cost bound to the
estimates.
Low Bound Estimate
During discussions in the TCRDSAC meetings, several stakeholders suggested that
almost all additional corrective actions taken in response to the RTCR would be in response to
transient contamination events or poor sampling techniques. These are also the least expensive
corrective actions. To examine the effects of greater emphasis on these types of corrective
actions, EPA reran the cost model to reflect 90% selection of either spot flushing or sampler
training as corrective actions. The remaining 10% of costs were distributed across the least costly
corrective actions under each of the other compliance forecast categories. This results in an
approximate 81 percent decrease in total net costs for the RTCR and an approximate 43 percent
decrease in total net costs for the Alternative option, using a 3 percent discount rate. Exhibit 7.24
shows the change in the overall costs with this change.
High Bound Estimate
PWSs may also take actions that result in higher corrective action costs than those
predicted by the current compliance forecast, although this would be a less likely scenario based
on stakeholder discussions. Purely economic considerations also suggest that, given the option,
PWSs would opt for the least costly option to address any issues identified. However, to test a
potential high end of corrective action costs, EPA ran the cost model with the compliance
forecast set to take the highest cost corrective action in each compliance forecast category. In this
scenario, only 5% of corrective actions were predicted for flushing and sampler training, and
10%) were estimated for each of the other compliance categories, and unit costs were assigned for
the highest unit cost corrective action in each. This results in an increase in total net costs for the
RTCR by approximately a factor of 1.8, and an increase in total net costs by approximately a
factor of 1.5 for the Alternative option. Exhibit 7.24 shows the annualized total and net change
cost estimates for PWSs and states to comply with RTCR under the 1989 TCR, RTCR, and
Alternative options based on low bound and high bound estimates in the compliance forecast.
Economic Analysis for the Final RTCR
7-42
September 2012
-------
Exhibit 7.24 Sensitivity Analysis—Annualized Net Change in Costs based on
Changes in Compliance Forecast ($Millions, 2007$)
PWSs
State
Total
PWSs
State
Total
3% Discount Rate
7% Discount Rate
RTCR Net Change
$ 14.15
$ 0.15
$ 14.30
$ 13.75
$ 0.42
$ 14.17
RTCR Low Bound Net Change
$ 2.61
$ 0.15
$ 2.75
$ 3.91
$ 0.42
$ 4.33
RTCR High Bound Net Change
$ 25.10
$ 0.15
$ 25.25
$ 23.63
$ 0.42
$ 24.05
Alternative Option Net Change
$ 29.29
$ 0.31
$ 29.60
$ 31.09
$ 0.61
$ 31.69
Alternative Option Low Bound Net Change
$ 16.54
$ 0.31
$ 16.84
$ 19.93
$ 0.61
$ 20.54
Alternative Option High Bound Net Change
$ 42.68
$ 0.31
$ 42.99
$ 43.63
$ 0.61
$ 44.24
Note: Detail may not add due to independent rounding.
Source: Final RTCRCost Model.
Overall, EPA recognizes that there is uncertainty in various parts of its estimates that
could result in either an over- or underestimate of the costs as presented in this chapter. Exhibit
7.25 presents a summary of these issues, references the location in the EA where the information
is introduced, and estimates the effects that each may have on national costs. All of the
assumptions influencing the EA baseline (1989 TCR) do so in the same way for the RTCR and
Alternative option. Therefore, EPA does not expect the net results of the analyses presented in
this EA to be significantly influenced by the uncertainty in the assumptions applied in
developing the RTCR cost analysis. EPA has been careful to use the best available data, to
account for uncertainty quantitatively when possible, and to avoid any consistent biases in
assumptions and the use of data.
Economic Analysis for the Final RTCR
7-43
September 2012
-------
Exhibit 7.25 Cost Uncertainty Summary
Uncertainty
Section with Full
Discussion of
Uncertainty
Most Likely Effect of Current Assumptions on Estimate
of National Costs
Underestimate
Overestimate
Unknown Impact
Labor Rate
7.2.1
X
Labor Burden
Estimates
7.4.1-7.4.8
X
Unit Costs
7.4.1-7.4.8
X [note 1]
Percentage of
systems on monthly,
quarterly, and annual
monitoring
frequencies
7.4.3
X
Number of acute
violations
7.4.5
X
Number of systems
performing level 1 and
level 2 assessments
7.4.5
X
Compliance forecasts
for corrective actions
based on level 1 and
level 2 assessments
7.4.6 and 7.7
X [note 2]
Corrective Actions
based on level 1 and
level 2 Assessments
7.4.6 and 7.7
X
Notes:
1) All unit costs (with the exception of those for corrective actions, which are not applied to the 1989 TCR) were
used in the RTCR cost code in the same way for the 1989 TCR, RTCR, and Alternative option. Therefore, EPA
expects that any under- or over-estimation would affect the baseline and other options similarly, resulting in no
significant net effect on the results of the analysis.
2) The compliance forecast represents EPA's best estimate of a distribution of corrective actions that may be
taken by PWSs to respond to problems identified under the RTCR. EPA ran two sensitivity analyses to assess
the potential impacts of different distributions within the compliance forecast (see Section 7.7). These two
sensitivity runs showed that under the low bound and high bound estimates, the net change costs for the RTCR
were smaller than those for the Alternative option.
Economic Analysis for the Final RTCR 7-44 September 2012
-------
7.8 Comparison of Total and Net Annualized Costs for All Regulatory Options
Based on information presented previously in this chapter, EPA developed national cost
estimates for the 1989 TCR, RTCR, and Alternative option. Exhibit 7.26 presents the total57 and
net change in annualized costs to PWSs and states at 3 and 7 percent discount rates. Exhibit 7.27
presents the total and net change in annualized costs for the 1989 TCR, RTCR, and Alternative
option by rule component at 3 and 7 percent discount rates. Exhibit 7.28 presents the total and
net change in annualized costs for the 1989 TCR, RTCR, and Alternative option by PWS size
and type at 3 and 7 percent discount rates. Exhibit 7.29 presents the costs shown in Exhibit 7.28
on a per-PWS basis. Further discussion of the results follows these exhibits.
Exhibit 7.26 Comparison of Total and Net Change from 1989 TCR in Annualized
Present Value Costs ($Millions, 2007$)
PWSs
State
Total
PWSs
State
Total
3% Discount Rate
7% Discount Rate
1989 TCR - Total
$ 185
$ 0.9
$ 186
$ 178
$ 0.9
$ 179
RTCR-Total
$ 199
$ 1.1
$ 200
$ 192
$ 1.3
$ 193
RTCR - Net Change
$ 14
$ 0.1
$ 14
$ 14
$ 0.4
$ 14
RTCR - Percent Change
8%
16%
8%
8%
48%
8%
Alternative Option - Total
$ 214
$ 1.2
$ 216
$ 209
$ 1.5
$ 210
Alternative Option - Net Change
$ 29
$ 0.3
$ 30
$ 31
$ 0.6
$ 32
Alternative Option - Percent Change
16%
34%
16%
17%
69%
18%
Note: Detail may not add due to independent rounding. Because only the net change in costs of some rule components are considered as
part of the cost analysis, references to "total" costs in this exhibit do not refer to the complete costs for regulatory implementation, but
only to the specific costs considered to calculate net changes in costs.
Source: Final RTCR cost model.
57 Because only the net change in costs of some rule components are considered as part of the cost analysis,
references to total costs in this section do not refer to the complete costs for regulatory implementation, but only to
the specific costs considered to calculate net changes in costs.
Economic Analysis for the Final RTCR
September 2012
7-45
-------
Exhibit 7.27 Comparison of Total and Net Change in Annualized Present Value
Costs by Rule Component ($Millions, 2007$)
PWSs | State | Total
PWSs | State | Total
3% Discount Rate
7% Discount Rate
Rule Im plementation and Annual Adm inistration
1989 TCR-Total
$
$
$
$
$
$
RTCR - Total
$ 2.77
$ 0.18
$ 2.95
$ 4.00
$ 0.26
$ 4.26
RTCR - Net Change
$ 2.77
$ 0.18
$ 2.95
$ 4.00
$ 0.26
$ 4.26
Alternative Option - Total
$ 2.77
$ 0.18
$ 2.95
$ 4.00
$ 0.26
$ 4.26
Alternative Option - Net Change
$ 2.77
$ 0.18
$ 2.95
$ 4.00
$ 0.26
$ 4.26
Revising Sample Siting Plans
1989 TCR-Total
$
$
$
$
$
$
RTCR - Total
$ 0.59
$ 0.42
$ 1.01
$ 0.84
$ 0.59
$ 1.42
RTCR - Net Change
$ 0.59
$ 0.42
$ 1.01
$ 0.84
$ 0.59
$ 1.42
Alternative Option - Total
$ 0.59
$ 0.42
$ 1.01
$ 0.84
$ 0.59
$ 1.42
Alternative Option - Net Change
$ 0.59
$ 0.42
$ 1.01
$ 0.84
$ 0.59
$ 1.42
Routine Monitoring
1989 TCR-Total
$ 170.59
$
$ 170.59
$ 163.94
$
$ 163.94
RTCR - Total
$ 174.71
$
$ 174.71
$ 167.74
$
$ 167.74
RTCR - Net Chanqe
$ 4.12
$
$ 4.12
$ 3.80
$
$ 3.80
Alternative Option - Total
$ 187.50
$
$ 187.50
$ 182.48
$
$ 182.48
Alternative Option - Net Change
$ 16.91
$
$ 16.91
$ 18.54
$
$ 18.54
Additional Routine Monitoring
1989 TCR-Total
$ 3.87
$
$ 3.87
$ 3.72
$
$ 3.72
RTCR - Total
$ 1.12
$
$ 1.12
$ 1.09
$
$ 1.09
RTCR - Net Change
$ (2.75)
$
$ (2.75)
$ (2.63)
$
$ (2.63)
Alternative Option - Total
$ 0.78
$
$ 0.78
$ 0.66
$
$ 0.66
Alternative Option - Net Change
$ (3.10)
$
$ (3.10)
$ (3.06)
$
$ (3.06)
Repeat Monitoring
1989 TCR-Total
$ 5.11
$
$ 5.11
$ 4.92
$
$ 4.92
RTCR - Total
$ 4.88
$
$ 4.88
$ 4.70
$
$ 4.70
RTCR - Net Change
$ (0.23)
$
$ (0.23)
$ (0.22)
$
$ (0.22)
Alternative Option - Total
$ 5.66
$
$ 5.66
$ 5.59
$
$ 5.59
Alternative Option - Net Change
$ 0.54
$
$ 0.54
$ 0.67
$
$ 0.67
Annual Site Visits
1989 TCR-Total
$
$
$
$
$
$
RTCR - Total
$
$
$
$
$
$
RTCR - Net Change
$
$
$
$
$
$
Alternative Option - Total
$
$
$
$
$
$
Alternative Option - Net Change
$
$
$
$
$
$
Level 1 Assessment
1989 TCR-Total
$ 1.13
$ 0.21
$ 1.34
$ 1.08
$ 0.20
$ 1.29
RTCR - Total
$ 1.63
$ 0.20
$ 1.84
$ 1.57
$ 0.20
$ 1.77
RTCR - Net Change
$ 0.51
$ (0.01)
$ 0.50
$ 0.49
$ (0.01)
$ 0.48
Alternative Option - Total
$ 1.76
$ 0.23
$ 1.99
$ 1.72
$ 0.23
$ 1.94
Alternative Option - Net Change
$ 0.63
$ 0.02
$ 0.65
$ 0.63
$ 0.02
$ 0.65
Level 2 Assessment
1989 TCR-Total
$ 0.70
$ 0.26
$ 0.96
$ 0.68
$ 0.25
$ 0.92
RTCR - Total
$ 0.90
$ 0.19
$ 1.08
$ 0.88
$ 0.18
$ 1.06
RTCR - Net Change
$ 0.20
$ (0.07)
$ 0.12
$ 0.20
$ (0.07)
$ 0.13
Alternative Option - Total
$ 1.26
$ 0.29
$ 1.55
$ 1.30
$ 0.31
$ 1.61
Alternative Option - Net Change
$ 0.55
$ 0.03
$ 0.58
$ 0.62
$ 0.06
$ 0.68
Corrective Actions based on Level 1 Assessments
1989 TCR - Total
$
$
$
$
$
$
RTCR-Total
$ 9.62
$ 0.01
$ 9.63
$ 8.14
$ 0.01
$ 8.15
RTCR - Net Change
$ 9.62
$ 0.01
$ 9.63
$ 8.14
$ 0.01
$ 8.15
Alternative Option - Total
$ 10.01
$ 0.01
$ 10.02
$ 8.52
$ 0.01
$ 8.53
Alternative Option - Net Change
$ 10.01
$ 0.01
$ 10.02
$ 8.52
$ 0.01
$ 8.53
Corrective Actions based on Level 2 Assessments
1989 TCR-Total
$
$
$
$
$
$
RTCR - Total
$ 2.82
$ 0.00
$ 2.82
$ 2.49
$ 0.00
$ 2.49
RTCR - Net Change
$ 2.82
$ 0.00
$ 2.82
$ 2.49
$ 0.00
$ 2.49
Alternative Option - Total
$ 3.78
$ 0.01
$ 3.79
$ 3.57
$ 0.01
$ 3.58
Alternative Option - Net Change
$ 3.78
$ 0.01
$ 3.79
$ 3.57
$ 0.01
$ 3.58
Public Notification
1989 TCR-Total
$ 3.75
$ 0.44
$ 4.19
$ 3.60
$ 0.42
$ 4.02
RTCR - Total
$ 0.26
$ 0.06
$ 0.32
$ 0.25
$ 0.06
$ 0.31
RTCR - Net Change
$ (3.49)
$ (0.38)
$ (3.86)
$ (3.35)
$ (0.36)
$ (3.71)
Alternative Option - Total
$ 0.35
$ 0.08
$ 0.43
$ 0.35
$ 0.08
$ 0.44
Alternative Option - Net Change
$ (3.40)
$ (0.36)
$ (3.76)
$ (3.25)
$ (0.34)
$ (3.58)
Notes:
1) Detail may not add due to independent rounding. Because only the incremental costs of some rule components are considered as
part of the cost analysis, references to "total" costs in this exhibit do not refer to the complete costs for regulatory implementation,
but only to the specific costs considered to calculate net changes in costs.
2) For modeling purposes, additional routine sample counts include regular routine samples taken in the same month.
Source: Final RTCRcost model.
Economic Analysis for the Final RTCR
7-46
September 2012
-------
Exhibit 7.28 Total and Net Change in Annualized Costs to PWSs by PWS Size and
Type ($Millions, 2007$)
3% Discount Rate
7%Discount Rate
Alternative
PWS Size
Alternative
Alternative
1989 TCR-
Option -
Alternative
(Population
1989 TCR-Total
RTCR-Total
RTCR-Net
Option - Total
Option - Net
Total
RTCR-Total
RTCR-Net
Total
Option - Net
Served)
A
B
C=B-A
D
E=D-A
F
G
H=G-F
I
J=l-F
Comm unity Water System s (CWSs)
<100
$7.4
$7.5
$0.1
$7.6
$0.2
$7.1
$7.3
$0.2
$7.5
$0.3
101-500
$9.0
$9.4
$0.4
$9.5
$0.5
$8.6
$9.1
$0.5
$9.2
$0.6
501-1,000
$3.7
$3.8
$0.0
$3.8
$0.1
$3.6
$3.7
$0.1
$3.7
$0.1
1,001-4,100
$13.2
$13.6
$0.4
$13.6
$0.4
$12.7
$13.1
$0.4
$13.1
$0.4
4,101-33,000
$42.4
$44.8
$2.4
$44.8
$2.4
$40.7
$42.8
$2.1
$42.8
$2.1
33,001-96,000
$34.9
$36.4
$1.5
$36.4
$1.5
$33.5
$34.8
$1.3
$34.8
$1.3
96,001-500,000
$34.7
$36.2
$1.5
$36.2
$1.5
$33.4
$34.6
$1.2
$34.6
$1.2
500,001-1 Million
$6.5
$6.7
$0.2
$6.7
$0.2
$6.2
$6.4
$0.1
$6.4
$0.1
> 1 Million
$5.6
$5.6
($0.0)
$5.6
($0.0)
$5.3
$5.3
($0.0)
$5.3
($0.0)
Total
$157.4
$163.9
$6.5
$164.1
$6.7
$151.3
$157.2
$5.9
$157.5
$6.2
Nontransient Noncomm unity Water System s (NTNCWSs)
<100
$2.6
$2.7
$0.1
$3.7
$1.1
$2.5
$2.7
$0.2
$3.8
$1.4
101-500
$1.9
$2.0
$0.1
$2.8
$0.9
$1.8
$2.0
$0.2
$2.9
$1.1
501-1,000
$0.6
$0.6
$0.1
$0.9
$0.3
$0.6
$0.6
$0.1
$0.9
$0.3
1,001-4,100
$1.2
$1.3
$0.1
$1.3
$0.1
$1.1
$1.2
$0.1
$1.2
$0.1
4,101-33,000
$0.4
$0.5
$0.1
$0.5
$0.1
$0.4
$0.5
$0.0
$0.5
$0.0
33,001-96,000
$0.1
$0.1
$0.0
$0.1
$0.0
$0.1
$0.1
$0.0
$0.1
$0.0
96,001-500,000
$0.1
$0.1
($0.0)
$0.1
($0.0)
$0.1
$0.1
($0.0)
$0.1
($0.0)
500,001-1 Million
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
> 1 Million
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
Total
$6.9
$7.3
$0.4
$9.3
$2.5
$6.6
$7.2
$0.6
$9.6
$3.0
Transient Noncom m unity Water System s (TNCWSs)
<100
$13.4
$18.7
$5.3
$28.1
$14.7
$12.8
$18.2
$5.3
$28.9
$16.1
101-500
$4.9
$6.5
$1.6
$9.5
$4.7
$4.7
$6.3
$1.6
$9.8
$5.1
501-1,000
$0.6
$0.8
$0.2
$1.2
$0.5
$0.6
$0.8
$0.2
$1.2
$0.6
1,001-4,100
$0.9
$1.0
$0.1
$1.0
$0.1
$0.9
$1.0
$0.1
$1.0
$0.1
4,101-33,000
$0.4
$0.5
$0.1
$0.5
$0.1
$0.4
$0.5
$0.0
$0.5
$0.0
33,001-96,000
$0.1
$0.1
($0.0)
$0.1
($0.0)
$0.1
$0.1
($0.0)
$0.1
($0.0)
96,001-500,000
$0.1
$0.1
($0.0)
$0.1
($0.0)
$0.1
$0.1
($0.0)
$0.1
($0.0)
500,001-1 Million
$0.2
$0.2
($0.0)
$0.2
($0.0)
$0.2
$0.2
($0.0)
$0.2
($0.0)
> 1 Million
$0.3
$0.3
$0.0
$0.3
$0.0
$0.3
$0.3
$0.0
$0.3
$0.0
Total
$20.9
$28.1
$7.3
$41.0
$20.1
$20.1
$27.3
$7.3
$42.0
$21.9
Grand Total
$185.2
$199.3
$14.2
$214.4
$29.3
$177.9
$191.7
$13.8
$209.0
$31.1
Note: Detail may not add due to independent rounding. Because only the incremental costs of some rule components are considered as part of the cost analysis, references to "total" costs
in this exhibit do not refer to the complete costs for regulatory implementation, but only to the specific costs considered to calculate net changes in costs.
Source: Final RTCRcost model.
Economic Analysis for the Final RTCR
7-47
September 2012
-------
Exhibit 7.29 Total and Net Change in Annualized Per PWS Costs by PWS Size and
Type (2007$)
3% Discount Rate
7% Discount Rate
Alternative
Alternative
PWS Size
1989 TCR-
Option -
Alternative
1989TCR-
Option -
Alternative
(Population
Total
RTCR - Total
RTCR - Net
Total
Option - Net
Total
RTCR - Total
RTCR - Net
Total
Option - Net
Served)
A
B
C=B-A
D
&D-A
F
G
H=G-F
1
J=l-F
Community Water Systems (CWSs)
<100
$564
$573
$9
$581
$17
$542
$558
$16
$569
$27
101-500
$561
$584
$23
$590
$29
$539
$568
$29
$576
$37
501-1,000
$663
$671
$8
$679
$15
$637
$654
$17
$664
$26
1,001-4,100
$1,411
$1,451
$40
$1,451
$40
$1,356
$1,401
$45
$1,401
$45
4,101-33,000
$6,707
$7,084
$377
$7,084
$377
$6,445
$6,775
$330
$6,775
$330
33,001-96,000
$33,053
$34,506
$1,453
$34,506
$1,453
$31,764
$32,971
$1,207
$32,971
$1,207
96,001-500,000
$93,620
$97,607
$3,987
$97,607
$3,987
$89,969
$93,242
$3,273
$93,242
$3,273
500,001-1 Million
$185,678
$190,167
$4,490
$190,167
$4,490
$178,437
$182,027
$3,591
$182,027
$3,591
> 1 Million
$277,732
$277,621
($111)
$277,621
($111)
$266,902
$266,828
($74)
$266,828
($74)
Total
$3,029
$3,153
$124
$3,158
$129
$2,911
$3,024
$113
$3,030
$119
Nontransient Noncommunity Water Systems (NTNCWSs)
<100
$284
$295
$11
$406
$122
$273
$293
$20
$423
$150
101-500
$273
$291
$18
$409
$135
$263
$290
$28
$425
$163
501-1,000
$321
$349
$29
$472
$151
$308
$348
$40
$491
$183
1,001-4,100
$1,325
$1,441
$116
$1,441
$116
$1,274
$1,383
$110
$1,383
$110
4,101-33,000
$4,819
$5,382
$563
$5,382
$563
$4,631
$5,124
$493
$5,124
$493
33,001-96,000
$32,003
$33,305
$1,303
$33,305
$1,303
$30,755
$31,852
$1,096
$31,852
$1,096
96,001-500,000
$91,163
$90,898
($265)
$90,898
($265)
$87,609
$87,385
($224)
$87,385
($224)
500,001-1 Million
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
> 1 Million
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
Total
$366
$390
$23
$498
$132
$352
$384
$32
$510
$158
Transient Noncommunity Water Systems (TNCWSs)
<100
$217
$304
$86
$457
$240
$209
$295
$87
$470
$261
101-500
$247
$327
$80
$483
$236
$237
$317
$80
$495
$257
501-1,000
$310
$408
$98
$567
$256
$298
$396
$97
$579
$281
1,001-4,100
$1,364
$1,528
$164
$1,528
$164
$1,311
$1,460
$150
$1,460
$150
4,101-33,000
$5,068
$5,679
$611
$5,679
$611
$4,871
$5,408
$537
$5,408
$537
33,001-96,000
$38,507
$38,155
($352)
$38,155
($352)
$37,008
$36,701
($307)
$36,701
($307)
96,001-500,000
$77,791
$77,043
($748)
$77,043
($748)
$74,762
$74,090
($672)
$74,090
($672)
500,001-1 Million
$186,699
$184,841
($1,858)
$184,841
($1,858)
$179,429
$177,725
($1,704)
$177,725
($1,704)
> 1 Million
$287,010
$302,862
$15,852
$302,862
$15,852
$275,817
$291,208
$15,391
$291,208
$15,391
Total
$248
$334
$86
$487
$239
$238
$325
$86
$499
$260
Grand Total
$1,196
$1,287
$91
$1,385
$189
$1,149
$1,238
$42
$1,350
$154
Note: Detail may not add due to independent rounding. Because only the incremental costs of some rule components are considered as part of the cost analysis,
references to "total" costs in this exhibit do not refer to the complete costs for regulatory implementation, but only to the specific costs considered to calculate net
changes in costs.
Source: Final RTCR cost model.
Economic Analysis for the Final RTCR
7-48
September 2012
-------
Evaluation of Regulatory Options
Exhibit 7.26 summarizes the comparison of total and annualized present value of the net
change difference in cost from the baseline for each of the three regulatory options (1989 TCR,
RTCR, and Alternative option). A continuation of the 1989 TCR would result in no net change in
costs. The net change in mean annualized present value national costs of the RTCR is estimated
to be approximately $14 million using a 3% or 7% discount rate. The net change in mean
annualized present value national costs for the Alternative option are estimated to be
approximately $30 million using a 3% discount rate and $32 million using a 7% discount rate.
The total net change in national annualized present value costs for PWSs serving >4,100
people (approximately $5.6 million at a 3% discount rate) is the same under the RTCR and
Alternative option. This is expected because the provisions for PWSs serving >4,100 are the
same under both the RTCR and Alternative option. Monitoring requirements for PWSs serving
>4,100 people would remain essentially unchanged under either the RTCR or Alternative option.
The observed overall net increase in costs for PWSs serving >4,100 people is driven primarily by
the requirements to conduct assessments and to correct any sanitary defects that are found.
Under the RTCR, PWSs are estimated to incur approximately 97% to 99% of the net
annualized present value costs at 3 and 7 percent discount rates. States are expected to incur the
remaining costs. EPA recognizes that state labor may be more expensive than PWS labor and
that for some states, state costs may be passed onto PWSs. However, state costs are only a very
small percentage of the total net change costs for the RTCR.
Exhibit 7.27 presents the comparison of total and net change in annualized present value
costs by rule component. The exhibit shows that, for the RTCR, corrective action costs are the
most significant contributor to the net increase in costs for PWSs. For the Alternative option,
routine monitoring costs are the most significant contributor to the net increase in costs for
PWSs. Under the RTCR and Alternative option, state costs to review revised sample siting plans
contribute most to the cost increase. For both PWSs and states, a net decrease in costs associated
with PN requirements helps to offset the total net cost increase.
The large difference in net cost increases between the RTCR and the Alternative option is
primarily driven by the increased number of routine samples taken under the Alternative option
in comparison to the RTCR. A larger number of samples are also estimated to result more in
Level 1 and Level 2 triggers and to subsequently require corrective actions based on Level 1 and
Level 2 assessments. Overall, the total net costs of the RTCR are estimated at 45% to 48% of the
net costs of the Alternative option. This cost difference is an important consideration in the
selection of the RTCR over the Alternative option as the preferred option.
Exhibit 7.28 presents the total and net change in annualized costs to PWSs by size and
type for the three regulatory options. No net change in costs would result from a continuation of
the 1989 TCR. Among PWSs serving <4,100 people, the largest increase in net costs would be
incurred by the TNCWSs serving <100 people under either the RTCR ($5.3 million) or
Alternative option ($14.7 million). As shown in Exhibit 7.29, on a per system basis, this
translates to a net annualized present value (3% discounting) increase of approximately $86 per
system under the RTCR and $240 per system under the Alternative option for the TNCWSs
Economic Analysis for the Final RTCR
7-49
September 2012
-------
serving <100 people. Significant impacts are also estimated for CWSs serving greater than 4,100
people, driven by the relatively high number of CWSs in this category, combined with relatively
high unit costs for implementing corrective actions for PWSs of this size. However, PWSs of this
size (and CWSs in general) are expected to be better able to absorb cost increases by passing
costs through to a larger base of customers. Overall, the most important drivers of total net costs
are the numbers of PWSs and the underlying baseline occurrence estimates in a given size
category. Taken together, these influence the numbers of samples and ultimately the numbers of
corrective actions taken, which drive the cost estimates.
Exhibit 7.29 shows the costs from Exhibit 7.28 on a per-PWS basis. On this basis, the
annual impact of the RTCR generally increases with PWS size and the magnitude of the annual
per-PWS net costs do not appear to be prohibitive, even for the most heavily impacted PWS
categories. However, the range of per-PWS costs is expected to be fairly wide and some
individual PWSs may be more heavily impacted.
Economic Analysis for the Final RTCR
7-50
September 2012
-------
8 Economic Impact Analysis
8.1 Introduction
As part of the rulemaking process, the United States Environmental Protection Agency
(EPA) is required to address the direct and indirect burdens that the Revised Total Coliform Rule
(RTCR) may place on certain types of governments, businesses, and populations. This chapter
presents analyses performed by the EPA in accordance with the following 14 federal mandates
and statutory reviews:
1. Executive Order 12866: Regulatory Planning and Review and Executive Order
13563: Improving Regulation and Regulatory Review.
2. Paperwork Reduction Act (the Information Collection Request (ICR) document for
the RTCR contains the complete analysis (USEPA, 2010a)).
3. The Regulatory Flexibility Act (RFA) of 1980, as amended by the Small Business
Regulatory Enforcement Fairness Act (SBREFA) of 1996.
4. Unfunded Mandates Reform Act (UMRA) of 1995.
5. Executive Order 13132: Federalism.
6. Executive Order 13175: Consultation and Coordination with Indian Tribal
Governments.
7. Executive Order 13045: Protection of Children from Environmental Health Risks and
Safety Risks.
8. Executive Order 13211: Action Concerning Regulations That Significantly Affect
Energy Supply, Distribution, or Use.
9. National Technology Transfer and Advancement Act (NTTAA).
10. Executive Order 12898: Federal Action to Address Environmental Justice in Minority
Populations and Low-Income Populations.
11. Consultations with the Science Advisory Board (SAB), National Drinking Water
Advisory Council (NDWAC), and the Secretary of Health and Human Services as
Required by Section 1412 (d) and (e) of the Safe Drinking Water Act (SDWA).
12. Impacts on Sensitive Subpopulations as Required by Section 1412(b)(3)(c)(i)(V) of
the 1996 Amendments to the Safe Drinking Water Act.
13. Effect of Compliance with the RTCR on the Technical, Financial, and Managerial
Capacity of Public Water Systems as Required by Section 1420(d)(3) of SDWA.
Economic Analysis for the Final RTCR
8-1
September 2012
-------
Many of the requirements and executive orders listed above call for an explanation of
why the rule is necessary, the statutory authority for the rule, and the primary objectives that the
rule is intended to achieve (refer to Chapter 2 for more information regarding the objectives of
the rule). Others are designed to assess the financial and health effects of the rule on sensitive,
low-income, and tribal populations as well as on small public water systems (PWSs).
8.2 Executive Order 12866: Regulatory Planning and Review
Under Executive Order 12866 (58 Federal Register (FR) 51735, October 4, 1993), the
Agency must determine whether the regulatory action is significant and therefore subject to
Office of Management and Budget (OMB) review and the requirements of the Executive Order.
The Order defines "significant regulatory action" as one that is likely to result in a rule that may:
• Have an annual effect on the economy of $100 million or more or adversely affect
in a material way the economy, a sector of the economy, productivity, competition,
jobs, the environment, public health or safety, or state, local, or tribal governments
or communities;
• Create a serious inconsistency or otherwise interfere with an action taken or
planned by another agency;
• Materially alter the budgetary impact of entitlement, grants, user fees, or loan
programs or the rights and obligations of recipients thereof; or
• Raise novel legal or policy issues arising out of legal mandates, the President's
priorities, or the principles set forth in the Executive Order.
Under Executive Order 12866, EPA has designated this action as a "significant
regulatory action" because of the legal and policy issues raised. Accordingly, EPA submitted this
action to the OMB for review under Executive Orders 12866 and 13563 (76 FR 3821, January
21, 2011) and any changes made in response to OMB recommendations have been documented
in the docket for this action.
In addition, EPA prepared an analysis of the potential costs and benefits associated with
this action. EPA estimates that the RTCR will have an overall annual impact on PWSs of $14
million and that the impact on small entities (PWSs serving 10,000 people or fewer) will be
$10.0 million-$10.3 million annualized at 3 and 7 percent discount rates, respectively. These
impacts are described in Section 8.4 as well as sections VI and VII.C of the RTCR preamble
(USEPA, 2010c), respectively.
8.3 Paperwork Reduction Act
The information collection requirements for the RTCR have been submitted for approval
to the OMB under the Paperwork Reduction Act (PRA), 44 U.S.C. 3501 et seq. The ICR
document prepared by EPA has been assigned EPA ICR number 1895.06.
The PRA requires EPA to estimate the burden, as defined in 5 CFR 1320.3(b), on PWSs
and state/primacy agencies of complying with the rule. The information collected as a result of
Economic Analysis for the Final RTCR
8-2
September 2012
-------
EPA's efforts toward proposing the RTCR should allow states/primacy agencies and EPA to
determine appropriate requirements for specific systems and evaluate compliance with the
RTCR. Burden is defined at 5 CFR 1320.3(b) and means the total time, effort, and financial
resources required to generate, maintain, retain, disclose, or provide information to or for a
federal agency. The burden includes the time needed to conduct the following state and PWS
activities, as needed:
State activities:
• Read and understand the rule;
• Mobilize (including primacy application), plan, and implement;
• Train PWS and consultant staff;
• Track compliance;
• Analyze and review PWS data;
• Review sample siting plans and recommend any revisions to PWSs;
• Make determinations concerning PWS monitoring requirements;
• Respond to PWSs with positive samples;
• Recordkeeping;
• Review completed assessment forms and consult with the PWS about the
assessment report;
• Review and coordinate with PWSs to determine optimal corrective actions to be
implemented; and
• Provide consultation, review public notification certifications, and file reports of
violations.
PWS activities:
• Read and understand the rule;
• Planning and mobilization activities;
• Revise existing sample siting plans to identify sampling locations and collection
schedules that are representative of water throughout the distribution system;
• Conduct routine, additional routine, and repeat monitoring, and report the results as
required;
Economic Analysis for the Final RTCR
8-3
September 2012
-------
• Complete a Level 1 Assessment if the PWS experiences a Level 1 trigger, and
submit a form to the state to identify sanitary defects detected, corrective actions
completed, and a timetable for any corrective actions not already completed;
• Complete a Level 2 Assessment if the PWS experiences a Level 2 trigger, and
submit a form to the state to identify sanitary defects detected, corrective actions
completed, and a timetable for any corrective actions not already completed;
• Correct sanitary defects found through the performance of Level 1 or Level 2
assessments and report on completion of corrective actions as required;
• Develop and distribute Tier 1 public notices when E. coli MCL violations occur;
• Develop and distribute Tier 2 public notices when the PWSs fail to take corrective
action; and
• Develop and distribute Tier 3 public notices when the PWSs fail to comply with the
monitoring requirements or with mandatory reporting of required information
within the specified timeframe.
For the first three years after publication of the rule in the Federal Register, the major
information requirements apply to 154,894 respondents. The net change in burden associated
with moving from the information requirements of the 1989 TCR to those in the RTCR over the
three years covered by the ICR is 2,518,577 hours, for an average of 839,526 hours per year. The
total net change in costs (i.e., incremental costs over the 1989 TCR) over the three-year clearance
period is $71.3 million, for an average of $23.8 million per year (simple average over three
years). (The ICR estimate is higher than what is presented in the economic analysis (EA) for the
RTCR because in the EA, the upfront costs that occur in the first three years, as well as future
costs, are annualized over a 25-year time horizon.) The average burden per response (i.e., the
amount of time needed for each activity that requires a collection of information) is 5.4 hours;
the average cost per response is $153.4. The collection requirements are mandatory under
SDWA (42 U.S.C. 300j-4 subsections (a)(1)(A) and (a)(1)(B)). Detail on the calculation of the
RTCR information collection burden and costs can be found in the ICR for the RTCR and
Chapter 7 of this EA. A summary of the burdens and costs of the collection is presented in
Exhibit 8.1 below.
Economic Analysis for the Final RTCR
8-4
September 2012
-------
Exhibit 8.1 Average Annual Net Change Burden and Costs for the RTCR
Information Collection Request
Respondent Type
Annual Burden
Hours
Cost
Annual
Responses
Annual Labor
Cost
Annual O&M
Cost
Annual Capital
Cost
Total Annual
Cost
PWSs
747,848
$ 20,171,639
$
$
$ 20,171,639
103,225
States and
Territories
91,678
$ 3,595,421
$
$
$ 3,595,421
51,669
TOTAL
839,526
$ 23,767,060
$
$
$ 23,767,060
154,894
Notes:
1) Detail may not add exactly to total due to independent rounding.
2) "Annual Burden Hours" reflects an annual average for all system sizes over the 3-year ICR period.
Source: ICR for the Final RTCR (USEPA2010b).
An agency may not conduct or sponsor, and a person is not required to respond to, a
collection of information unless it displays a currently valid OMB control number. The OMB
control numbers for EPA's regulations in 40 CFR are listed in 40 CFR part 9.
As part of the Federal Register notice on the proposed RTCR, EPA solicited comments
on this information collection and the estimates in this ICR. EPA solicited comments on specific
aspects of the proposed information collection, as described below:
1. Whether the collection of information is necessary for the proper performance of the
functions of the Agency, including whether the information will have practical utility;
2. Whether the Agency's burden estimate is accurate including the validity of the
methodology and assumptions used;
3. How to enhance the quality, utility, and clarity of the information to be collected; and
4. How to minimize the burden on respondents, including use of appropriate automated
electronic, mechanical, or other technological collection techniques or other forms of
information technology.
EPA did not receive comments that specifically referred to the ICR prepared for the
proposed rule; however, it received several comments (such as the need to increase unit costs) on
the associated EA, as well as the Technology and Cost Document for the Final Revised Total
Coliform Rule (USEPA, 2010d) (which contains many of the unit costs used to estimate costs for
the Cost Analysis and the ICR). For the RTCR, EPA revised many of the unit costs associated
with corrective actions to incorporate those comments; these changes have been incorporated
into this ICR.
EPA's responses to comments received on the proposed rule are available at
http://www.regulations.gov. docket number EPA-HQ-OW-2008-0878.
Economic Analysis for the Final RTCR
8-5
September 2012
-------
In compliance with the PRA (44 USC 3501 et seq.), EPA submitted the ICR for the
RTCR Rule to OMB for review and approval prior to proposal. EPA did not receive any
comments from OMB on the ICR at that time.
8.4 The Regulatory Flexibility Act
The RFA generally requires an agency to prepare a regulatory flexibility analysis of any
rule subject to notice and comment rulemaking requirements under the Administrative Procedure
Act or any other statute unless the agency certifies that the rule will not have a significant
economic impact on a substantial number of small entities. Small entities include small
businesses, small organizations, and small governmental jurisdictions.
The RFA provides default definitions for each type of small entity. Small entities are
defined as: (1) a small business as defined by the Small Business Administration's (SBA)
regulations at 13 CFR 121.201; (2) a small governmental jurisdiction that is a government of a
city, county, town, school district, or special district with a population of less than 50,000; and
(3) a small organization that is any "not-for-profit enterprise which is independently owned and
operated and is not dominant in its field." However, the RFA also authorizes an agency to use
alternative definitions for each category of small entity, "which are appropriate to the activities
of the agency" after proposing the alternative definition(s) in the Federal Register and taking
comment. 5 USC 601(3)-(5). In addition, to establish an alternative small business definition,
agencies must consult with SBA's Chief Counsel for Advocacy.
For purposes of assessing the impacts of the RTCR on small entities, EPA considered
small entities to be PWSs serving 10,000 people or fewer. This is the cutoff level specified by
Congress in the 1996 Amendments to the SDWA for small system flexibility provisions. As
required by the RFA, EPA proposed using this alternative definition in the Federal Register (63
FR 7620, February 13, 1998), requested public comment, consulted with the SBA, and finalized
the alternative definition in the Agency's Consumer Confidence Reports Rule (USEPA, 1998b,
63 FR 44524, August 19, 1998). As stated in that Final Rule, the alternative definition would be
applied for all future drinking water regulations.
After considering the economic impacts of today's rule on small entities, EPA has
certified that the RTCR will not have a significant economic impact on a substantial number of
small entities. The small entities directly regulated by this rule are small PWSs serving 10,000 or
fewer people. These include small community water systems (CWSs), non-transient
noncommunity water systems (NTNCWSs), and transient noncommunity water systems
(TNCWSs), entities such as municipal water systems (publicly and privately owned), and
privately-owned PWSs and for profit businesses where provision of water may be ancillary, such
as mobile home parks, day care centers, churches, schools and homeowner associations. In
evaluating the impact on small PWSs, it has been determined that only 61 of 150,672 small
systems (0.04 percent) will experience an impact of more than one percent of revenues, and that
none of the small systems will experience an impact of three percent or greater of revenue.
Exhibit 8.2 provides a summary of the numbers and percentages of small systems for which costs
exceed one percent of revenues, by size category and system type using a three percent discount
rate. Exhibit 8.3 provides a summary of the average costs and average revenues on a per system
Economic Analysis for the Final RTCR
8-6
September 2012
-------
basis. See Appendix I for the detailed methodology and supporting information for the RFA
screening analysis.
Exhibit 8.2 RTCR - Average Annualized Revenue by System Size and Percent of
Systems with Costs Exceeding One Percent and Three Percent of Revenue (Three
Percent Discount Rate)
System
Type
System
Size
Total
Systems
Average
Revenue/
System
i%of
Revenue
Systems
Exceeding
1%of
Revenue
Percent of
Systems
Exceeding
1%of
Revenue
3% of
Revenue
Systems
Exceeding
3% of
Revenue
Percent of
Systems
Exceeding
3% of
Revenue
A
B
C=B*0.01
D
E=D/A
F=B*0.03
G
H=G/A
CWS
<500
29,150
$ 199,000
$ 2,000
60
0.21%
$ 6,000
0
0.00%
501 -4,100
15,021
$ 1,114,000
$ 11,000
0
0.00%
$ 33,000
0
0.00%
4,101 - 10,000
3,672
$ 4,432,000
$ 44,000
1
0.03%
$ 133,000
0
0.00%
NTNCWS
<500
15,942
$ 3,075,000
$ 31,000
0
0.00%
$ 92,000
0
0.00%
501 -4,100
2,690
$ 11,793,000
$ 118,000
0
0.00%
$ 354,000
0
0.00%
4,101 - 10,000
80
$ 61,663,000
$ 617,000
0
0.00%
$1,850,000
0
0.00%
TNCWS
<500
81,311
$ 1,811,000
$ 18,000
0
0.00%
$ 54,000
0
0.00%
501 -4,100
2,735
$ 3,192,000
$ 32,000
0
0.00%
$ 96,000
0
0.00%
4,101 - 10,000
71
$ 11,280,000
$ 113,000
0
0.00%
$ 338,000
0
0.00%
All
<10,000
150,672
N/A
N/A
61
0.04%
N/A
0
0.00%
Sources: (A) SDWIS 2007 (B) Average CWS revenues from 2006 CWSS survey. Average NTNCWS and TNCWS revenues calculated
based on representative revenues for specific business categorizations (Appendix I provides additional detail.)
(D, G) Number of systems with costs exceeding one percent and three percent of revenue calculated based on distributions of costs for
each rule component by system size compared to the average revenues. As a function of calculations in the cost model, fractional system
counts may be generated. Analyses in the exhibit are based only on whole system counts. (Appendix I provides additional detail.)
Consistent with the rest of the EA, the population break is at 4,100 people rather than 3,300. This population break was based on
TCRDSAC deliberations and consideration that concluded that breaking out the analysis at the 4,100 point would be most informative
because of changes in the rule provisions at the 4,100 people point. (Chapter 4 provides further detail on data selection and size break
outs.)
Economic Analysis for the Final RTCR
8-7
September 2012
-------
Exhibit 8.3 RTCR—Average Costs per System and as Percentage of Revenue
(2007$)
System
Size
Number
of
Systems
Average
Annual
Net Cost/
System
Average
Revenue/
System1
Average Annual
Net Costs as a
Percentage of
Revenue
A
B
C
D=(B/C)*100
<500
126,403
o
CD
$ 1,599,000
0.004%
501 -4,100
20,446
CO
$ 2,797,000
0.002%
4,101 -10,000
3,823
$ 276
$ 5,757,000
0.005%
All PWSs
<10,000
150,672
CD
$ 1,867,000
0.003%
11ncludes water revenues and non-water revenues (e.g., revenues related to the primary
entities that operate a water system to support their business or municipal general revenue
for publicly owned and operated systems).
Although this rule will not have a significant economic impact on a substantial number of
small entities, EPA nonetheless has tried to reduce the impact of this rule on small PWSs.
Provisions in the RTCR that result in reduced costs for many small entities include:
• Reduced routine monitoring for qualifying PWSs serving 1,000 or fewer people.
• Reduced number of repeat samples required for systems serving 1,000 or fewer
people.
• Reduced additional routine monitoring for PWSs serving 4,100 or fewer people.
• Reduced public notification requirements for all systems, including small systems.
For some PWSs, cost savings may be offset in whole or in part by increased costs of
more stringent assessment requirements, stricter rules for qualifying for reduced monitoring, or
performing specific corrective actions. Additionally, seasonal PWSs are subject to increased
monitoring.
A description of activities that small systems perform under the RTCR for each rule
component is provided in Section 7.4 of this EA along with the associated change in cost.
Exhibit 8.4 below provides the distribution of total costs to small entities by rule component.
Underlying these estimates are EPA's assumptions regarding the types of corrective actions that
will be implemented and the number of small PWSs predicted to implement each type. These
assumptions are detailed in Appendix D of this EA.
Economic Analysis for the Final RTCR
8-8
September 2012
-------
Exhibit 8.4 RTCR—Annualized Net Rule Costs Predicted for Small Entities (PWSs
serving <10,000) by Rule Component Using Three Percent and Seven Percent
Discount Rates (2007$)
Rule Component
Net Costs at 3%
Net Costs at 7%
Rule Implementation
$ 2,673,000
$ 3,849,000
Revising Sample Siting Plans
$ 535,000
$ 755,000
Routine Monitoring
$ 4,120,000
$ 3,800,000
Additional Routine Monitoring
$ (2,753,000)
$ (2,633,000)
Repeat Monitoring
$ (178,000)
$ (171,000)
Annual Site Visits
$
$
Level 1 Assessment
$ 417,000
$ 404,000
Level 2 Assessment
$ 172,000
$ 179,000
Correction Actions based on Level 1
Assessments
$ 5,159,000
$ 4,425,000
Correction Actions based on Level 2
Assessments
$ 2,642,000
$ 2,340,000
Public Notification
$ (2,787,000)
$ (2,674,000)
Total
$ 10,000,000
$ 10,273,000
Source: Derived from cost model outputs (AppendixC).
EPA also conducted outreach to small entities and convened a Small Business Advocacy
Review (SBAR) Panel to obtain advice and recommendations from representatives of the small
entities that potentially would be subject to the rule's requirements. EPA consulted with small
entity representatives before and during the review by the Panel. These small entity
representatives included representatives from small water systems of various types and sizes,
representatives from associations that assist and /or advocate for small systems, and federal
agencies that operate small systems. Panel members included representatives from OMB, SBA,
and the EPA Office of Ground Water and Drinking Water. The consultation led to the
development of a report providing recommendations to EPA on how to revise the 1989 TCR to
address small system concerns, which EPA considered in drafting the RTCR (SBAR Panel,
2008). EPA also made presentations to the advisory committee on the recommendations of the
Panel so the advisory committee could consider their recommendations in developing the
Agreement in Principle (AIP).
Consistent with the RFA/SBREFA requirements, the SBAR Panel evaluated the
assembled materials and small-entity comments on issues and prepared a final report to the EPA
Administrator. A copy of the SBAR Panel report is included in the docket for this rule. The rule
is consistent with the SBAR Panel recommendations to use total coliform (TC) as a trigger for
investigation and/or corrective action, to balance monitoring requirements and costs with risk, to
further differentiate requirements based on differences in water systems, to coordinate
requirements with other related rules, and to consider reporting and recordkeeping costs in
estimating burden.
Economic Analysis for the Final RTCR
8-9
September 2012
-------
EPA further reviewed the potential impacts of the RTCR on small entities in terms of the
assumption incorporated into SDWA major rulemakings that households or residential water
58
users are most vulnerable to cost increases (SBAR Panel, 2008). In an Assessment of the
Vulnerability of Noncommunity Water Systems to SDWA Cost Increases (USEPA, 2008d), EPA
considered the burden of SDWA rule costs in comparison to the average revenues of various
categories of noncommunity water systems (NCWSs). All of the NCWS categories reviewed
were found to be less vulnerable to SDWA-related increases than a typical household. The report
notes that in some categories of businesses, costs are more easily passed on to the customer base
than in others. However, in each NCWS category, expenditures on water were found to be a
relatively small percentage of total revenues. Water expenditures (including expenditures for
sewer service and miscellaneous other utilities) totaled less than one percent of total revenues in
nearly all cases, and were not more than 1.3 percent of total revenues for any category. Several
caveats were put forth in this report, including one that considered the potential for
underestimating the impact to golf courses, which were grouped in with other recreational
entities whose use of water was less significant to the core business than the golf courses.
Despite the significant caveats listed, the report strongly suggests that TNCWS and NTNCWS
should not be considered particularly vulnerable to operating cost increases resulting from
SDWA rulemakings.
The consistency of many of the aspects of the RTCR with the recommendations of the
SBAR Panel further support the conclusion of the RTCR EA that the RTCR requirements are
conducive to minimizing net impacts on small entities. Overall, the economic analysis
summarized in Exhibit 8.2 shows that net cost increases from the RTCR over time are low
relative to revenue. Based on this result, EPA certifies that there will not be a significant impact
on a substantial number of small entities under the RTCR.
8.5 Unfunded Mandates Reform Act
The UMRA seeks to protect state, local, and tribal governments from the imposition of
unfunded federal mandates. In addition, the Act seeks to strengthen the partnership between the
federal government and state, local, and tribal governments and ensure that the federal
government covers the costs incurred during compliance with federal mandates.
Title II of the UMRA of 1995, Public Law 104-4, establishes requirements for federal
agencies to assess the effects of their regulatory actions on state, local, and tribal governments
and the private sector. Under section 202 of UMRA, EPA generally must prepare a written
statement, including a cost-benefit analysis, for proposed and final rules with federal mandates
that may result in expenditures to state, local, and tribal governments, in the aggregate, or to the
private sector, of $100 million or more in any one year.
Section 205 of UMRA generally requires EPA to identify and consider a reasonable
number of regulatory options and adopt the least costly, most cost-effective or least burdensome
option that achieves the objectives of the rule. The provisions of section 205 do not apply when
they are inconsistent with applicable law. Moreover, section 205 allows EPA to adopt an option
58 Major SDWA rulemakings generally include a section in the EA that presents an analysis of costs of the rule per
household.
Economic Analysis for the Final RTCR
8-10
September 2012
-------
other that the least costly, most cost-effective or least burdensome option if the Administrator
publishes with the rule and explanation why that option was not adopted.
Before EPA establishes any regulatory requirements that may significantly or uniquely
affect small governments, including tribal governments, it must have developed under section
203 of UMRA a small government agency plan. The plan must provide for notifying potentially
affected small governments, enabling officials of affected small governments to have meaningful
and timely input in the development of EPA regulatory proposals with significant federal
intergovernmental mandates, and informing, educating, and advising small governments on
compliance with the regulatory requirements.
This rule does not contain a federal mandate that may result in expenditures to state,
local, and tribal governments, in the aggregate, or to the private sector, of $100 million or more
in any one year. Expenditures associated with compliance, defined as the net change in costs
beyond the 1989 TCR, will not surpass $100 million in the aggregate in any year. Thus, this rule
is not subject to the requirements of sections 202 and 205 of UMRA.
This rule is also not subject to the requirements of section 203 of UMRA because it
contains no regulatory requirements that might significantly or uniquely affect small
governments. Costs to small entities are generally not significant, as described previously in
Section 8.4 and in section VII.C of the RTCR preamble. The regulatory requirements of the
RTCR are not unique to small governments, as they apply to all PWSs regardless of size.
8.6 Executive Order 13132: Federalism
Executive Order 13132, entitled "Federalism" (64 FR 43255, August 10, 1999), requires
EPA to develop an accountable process to ensure "meaningful and timely input by state and local
officials in the development of regulatory policies that have federalism implications." "Policies
that have federalism implications" is defined in the Executive Order to include regulations that
have "substantial direct effects on the states, on the relationship between the national government
and the states, or on the distribution of power and responsibilities among the various levels of
government."
This action does not have federalism implications. It will not have substantial direct
effects on the states, on the relationship between the national government and the states, or on
the distribution of power and responsibilities among the various levels of government as
specified in Executive Order 13132. The net change in cost for state, local, and tribal
governments in the aggregate is estimated to be approximately $0.2 million and $0.4 million per
year at three percent and seven percent discount rates, respectively. Thus, Executive Order 13132
does not apply to this rule.
Although section 6 of Executive Order 13132 does not apply to the RTCR, EPA
conducted a federalism consultation, consistent with Executive Order 13132, in July 2008. The
consultation included a stakeholder meeting where EPA requested comments on the impacts of
the potential revisions to the 1989 TCR with respect to state, county and local governments. EPA
did not receive any comments in response to this consultation. In addition, the TCRDSAC
included representatives of state, local and tribal governments, and through this process EPA
Economic Analysis for the Final RTCR
8-11
September 2012
-------
consulted with state, local, and tribal government representatives to ensure that their views were
considered when the revisions to the 1989 TCR were developed.
In the spirit of Executive Order 13132 and consistent with EPA policy to promote
communications between EPA and state and local governments, during proposal of the RTCR,
EPA specifically solicited comment on this action from state and local officials. EPA conducted
two stakeholder meetings, in April of 2009 and May of 2010, to solicit stakeholder input as it
developed a proposed rule consistent with the recommendations of the AIP. No specific concerns
or considerations related to Federalism were received.
8.7 Executive Order 13175: Consultation and Coordination with Indian Tribal
Governments
Executive Order 13175, entitled "Consultation and Coordination with Indian Tribal
Governments" (65 FR 67249, November 9, 2000), requires EPA to develop an accountable
process to ensure "meaningful and timely input by tribal officials in the development of
regulatory policies that have tribal implications." The Executive Order defines "policies that
have tribal implications: to include regulations that have "substantial direct effects on one or
more Indian tribes, on the relationship between the federal government and the Indian tribes, or
on the distribution of power and responsibilities between the federal government and Indian
tribes."
Under Executive Order 13175, EPA may not issue a regulation that has tribal
implications, that imposes substantial direct compliance costs, and that is not required by statute,
unless the federal government provides the funds necessary to pay the direct compliance costs
incurred by tribal governments, or EPA consults with tribal officials early in the process of
developing the proposed regulation and develops a tribal summary impact statement.
This action does not have tribal implications, as specified in Executive Order 13175.
Because the requirements of the RTCR are estimated to result in low net cost increases (or in
many cases, net cost savings) compared to the 1989 TCR requirements, the RTCR is not
anticipated to have a negative impact on tribal PWSs.
Although Executive Order 13175 does not apply to this action, EPA consulted with tribal
officials in developing this action. EPA has consulted with tribal governments through the EPA
American Indian Environmental Office, included a representative of the Native American Water
Association on the advisory committee that developed recommendations regarding the proposed
rule and signed the AIP, and has addressed tribal concerns throughout the regulatory
development process, as appropriate. The consultation included participation in three tribal
conference calls (EPA regional tribal call (February 2008), National Indian Workgroup call
(March 2008), and National Tribal Water Conference (March 2008)). EPA requested comments
on the 1989 TCR, requested suggestions for 1989 TCR revisions (March 2008), and presented
possible revisions to the 1989 TCR to the National Tribal Council (April 2008). Furthermore, in
the proposed RTCR, EPA specifically solicited additional comments on this action from tribal
officials. None of these consultations or solicitations identified issues that were particular to
tribal entities. As a result of the tribal consultations and other tribal outreach, EPA has
Economic Analysis for the Final RTCR
8-12
September 2012
-------
determined that the RTCR is not anticipated to have a negative impact on tribal systems. Thus,
Executive Order 13175 does not apply to this action.
8.8 Executive Order 13045: Protection of Children from Environmental Health Risks
and Safety Risks
Executive Order 13045 (62 FR 19885; April 23, 1997) applies to any rule initiated after
April 21, 1998, that (1) is determined to be "economically significant" as defined under
Executive Order 12866; and (2) concerns an environmental, health, or safety risk that EPA has
reason to believe may have a disproportionate effect on children. If the regulatory action meets
both criteria, EPA must evaluate the environmental, health, or safety effects of the planned rule
on children, and explain why the planned regulation is preferable to other potentially effective
and reasonably feasible options considered by EPA.
The RTCR is not subject to Executive Order 13045 because it is not economically
significant as defined in Executive Order 12866. This action's health and risk assessments in
relation to Executive Order 13045 are contained in section VI.K.l of the RTCR preamble and
Chapter 6 of the EA. This EA uses a qualitative approach in assessing the changes in risk
anticipated for each regulatory option relative the baseline (1989 TCR). EPA expects that the
RTCR would provide additional protection, through the additional assessments and corrective
action required, to both children and adults who consume drinking water supplied from PWSs.
EPA also assumes that the benefits of the rule, including reduced health risk, will
disproportionally accrue more to children because young children are more susceptible than
adults to some waterborne illnesses and are more likely to experience more serious effects from
infection. For example, the risk of mortality resulting from diarrhea is often greatest in the very
young and elderly (Rose, 1997; Gerba et al., 1996), and viral and bacterial illnesses often
disproportionately affect children. Any overall benefits of the rule would reduce this mortality
risk for children.
During proposal of the RTCR, the public was invited to submit comments or identify
peer-reviewed studies and data that assess effects of early life exposure to drinking water that
contains fecal contaminants. No additional comments were made, nor were any additional
studies or data identified.
8.9 Executive Order 13211: Action Concerning Regulations That Significantly Affect
Energy Supply, Distribution, or Use
Executive Order 13211, "Actions Concerning Regulations That Significantly Affect
Energy Supply Distribution, or Use" (66 FR 28355, May 22, 2001), provides that agencies shall
prepare and submit to the Administrator of the Office of Information and Regulatory Affairs,
OMB, a Statement of Energy Effects for certain actions identified as "significant energy
actions." Section 4(b) of Executive Order 13211 defines "significant energy actions" as "any
action by an agency (normally published in the Federal Register) that promulgates or is expected
to lead to the promulgation of a final rule or regulation, including notices of inquiry, advance
notices of proposed rulemaking, and notices of proposed rulemaking: (l)(i) that is a significant
regulatory action under Executive Order 12866 or any successor order, and (ii) is likely to have a
significant adverse effect on the supply, distribution, or use of energy; or (2) that is designated by
Economic Analysis for the Final RTCR
8-13
September 2012
-------
the Administrator of the Office of Information and Regulatory Affairs as a significant energy
action."
The RTCR is not a "significant energy action" as defined in Executive Order 13211,
because it is not likely to have a significant adverse effect on the supply, distribution, or use of
energy. This rule is not a significant regulatory action under Executive Order 12866, and has not
been designated by the Administrator of the Office of Information and Regulatory Affairs as a
significant energy action, for the reasons described as follows.
Energy Supply
The RTCR does not regulate power generation, either directly or indirectly, and public
and private PWSs that the RTCR applies to do not, as a general rule, generate power. Further, the
energy cost increases borne by customers of PWSs as a result of the RTCR are a low percentage
of the total cost of water. Therefore, power generation utilities that purchase water as part of their
operations are unlikely to face any significant effects as a result of the RTCR.
Energy Distribution
The RTCR does not regulate any aspect of energy distribution and PWSs that are
regulated by the RTCR already have electrical service. The rule is not expected to increase peak
electricity demand at PWSs. Therefore, EPA assumes that the existing connections are adequate
and that the RTCR has no discernible adverse effect on energy distribution.
Energy Use
Because the RTCR modifies existing regulations, very few PWSs are expected to make
modifications or changes that will alter energy use patterns as a result of this rule. Therefore,
EPA does not expect any noticeable effect on the national levels of power generation in terms of
average and peak loads.
8.10 National Technology Transfer and Advancement Act
Section 12(d) of the NTT AA of 1995, Public Law No. 104-113, 12(d) (15 U.S.C. 272
note), directs EPA to use voluntary consensus standards in its regulatory activities unless to do so
would be inconsistent with applicable law or otherwise impractical. Voluntary consensus
standards are technical standards (e.g., materials specifications, test methods, sampling
procedures, and business practices) that are developed or adopted by voluntary consensus
standards bodies. NTTAA directs EPA to provide Congress, through OMB, explanations when
EPA decides not to use available and applicable voluntary consensus standards.
The RTCR involves technical voluntary consensus standards. Under the RTCR, EPA will
use several analytical methods to monitor for TC and/or /•]. coli as described in Standard
Methods for the Examination of Water and Wastewater (Clesceri et al., 1998; Eaton et al., 2005).
Methods included in Clesceri et al. (1998) and Eaton et al. (2005) are voluntary consensus
standards. The RTCR includes 11 methods that can be used to test for TC, four of which are
described in Clesceri et al. (1998) and Eaton et al. (2005).
Economic Analysis for the Final RTCR
8-14
September 2012
-------
During proposal of the RTCR, EPA welcomed comments on this aspect of the
rulemaking and, specifically, invited the public to identify potentially applicable voluntary
consensus standards and to explain why such standards should be used in this regulation.
8.11 Executive Order 12898: Federal Actions to Address Environmental Justice in
Minority Populations and Low-Income Populations
Executive Order 12898 (59 FR 7629, February 16, 1994) establishes federal executive
policy on environmental justice. Its main provision directs federal agencies, to the greatest extent
practicable and permitted by law, to make environmental justice part of their mission. Agencies
must do this by identifying and addressing, as appropriate, any disproportionately high and
adverse human health or environmental effects of their programs, policies, and activities on
minority populations and low-income populations in the United States.
EPA has determined that the RTCR will not have disproportionately high and adverse
human health or environmental effects on minority or low-income populations because it
increases the level of environmental protection for all affected populations without having any
disproportionately high and adverse human health or environmental effects on any population,
including any minority or low-income population. The RTCR applies uniformly to all PWSs.
Consequently, the RTCR provides health protection equally to all income and minority groups
served by PWSs. The RTCR and other drinking water regulations are expected to have a positive
effect on human health regardless of the social or economic status of a specific population. To
the extent that contaminants in drinking water might be disproportionately high among minority
or low-income populations (which is unknown), the RTCR contributes toward removing those
differences by assuring that all PWSs meet drinking water standards and take appropriate
corrective action whenever appropriate. Thus, the RTCR meets the intent of the federal policy
requiring incorporation of environmental justice into federal agency missions.
8.12 Consultations with the Science Advisory Board, National Drinking Water Advisory
Council, and the Secretary of Health and Human Services as Required by Section
1412 (d) and (e) of the SDWA
In accordance with section 1412(d) and (e) of the SDWA, EPA consulted with the
Science Advisory Board (SAB), National Drinking Water Advisory Council (NDWAC), and the
Secretary of the US Department of Health and Human Services (HHS) on the proposed RTCR.
In addition, EPA consulted again with NDWAC and HHS before promulgation of the final
RTCR.
EPA met with the SAB Drinking Water Committee (DWC) and considered their
recommendations in developing data requirements to better understand the impacts of the RTCR.
In response to the SAB DWC recommendations, EPA also conducted sensitivity analyses, to
explore a wider range of assumptions regarding the percentage of assessments leading to
corrective actions and to demonstrate that using an annual average for occurrence provided
results comparable to varying the occurrence based on the season. In addition, EPA added an
exhibit in the EA that summarizes all significant model parameters and assumptions, their
influence on variability and uncertainty, and their most likely effect on benefits or costs. A copy
of the SAB report (SAB 2010) is available in the docket for the RTCR.
Economic Analysis for the Final RTCR
8-15
September 2012
-------
EPA also consulted with NDWAC and requested comments in the proposed RTCR on
areas of concern that NDWAC members raised during the consultation. EPA requested comment
on the following: (1) whether the RTCR would be easier to implement than the 1989 TCR; and
(2) the costs and benefits of reduced monitoring. EPA considered and addressed the comments it
received in developing the final rule. EPA also considered NDWAC's recommendations in
developing the public notification requirements for the rule.
EPA completed its consultation with HHS, as required by SDWA section 1412(d). In
addition, EPA provided an informational briefing to the Food and Safety Group of the Food and
Drug Administration.
Details about EPA's consultations (both for the proposed and Final Rule) with SAB,
NDWAC, and HHS can be found at Section VII.K of the RTCR preamble.
8.13 Consideration of Impacts on Sensitive Subpopulations as Required by Section
1412(b)(3)(c)(i)(V) of the 1996 Amendments to the Safe Drinking Water Act (SDWA)
EPA is required to seek public comment regarding the effects of contamination
associated with the RTCR on the general population and sensitive subpopulations. Sensitive
subpopulations include "infants, children, pregnant women, the elderly, individuals with a
history of serious illness, or other subpopulations that are identified as likely to be at greater risk
of adverse health effects due to exposure to contaminants in drinking water than the general
population" (SDWA section 1412(b)(3)(C)(i)(V), 42 U.S.C 300g-l(b)(3)(C)(i)(V)).
Pregnant and lactating women may be at an increased risk from pathogens as well as act
as a source of infection for newborns. Infection during pregnancy may also result in the
transmission of infection from the mother to the child in utero, during birth, or shortly thereafter.
Since very young children do not have fully developed immune systems, they are at increased
risk and are particularly difficult to treat.
Infectious diseases are also a major problem for the elderly because immune function
declines with age. As a result, outbreaks of waterborne diseases can be devastating on the elderly
community (e.g., nursing homes) and may increase the possibility of significantly higher
mortality rates in the elderly than in the general population.
Immunocompromised individuals are a growing proportion of the population with the
continued increase in HIV/AIDS, the aging population, and the escalation in organ and tissue
transplantations. Immunocompromised individuals are more susceptible to severe and invasive
infection. These infections are particularly difficult to treat and can result in a significantly
higher mortality than in immunocompetent persons.
It is anticipated that the requirements of the RTCR will help reduce pathways of entry for
fecal contamination and/or waterborne pathogens into the distribution system, thereby reducing
exposure and risk from these contaminants in drinking water to the entire general population.
The RTCR seeks to provide a similar level of drinking water protection to all groups including
sensitive subpopulations, thus meeting the intent of this federal policy. See also section VI.K of
the RTCR preamble.
Economic Analysis for the Final RTCR
8-16
September 2012
-------
8.14 Effect of Compliance with the RTCR on the Technical, Financial, and Managerial
Capacity of Public Water Systems as Required by Section 1420(d)(3) of SDWA
Section 1420(d)(3) of the SDWA, as amended, requires that, in promulgating a National
Primary Drinking Water Regulation, the Administrator shall include an analysis of the likely
effect of compliance with the regulation on the technical, managerial, and financial (TMF)
capacity of PWSs. The following analysis fulfills this statutory obligation by identifying the
incremental impact that the RTCR will have on the TMF capacity of regulated PWSs. Analyses
presented in this document reflect only the impact of new or revised requirements, as established
by the RTCR; the impacts of previously established requirements on system capacity are not
considered.
Overall water system capacity is defined in Guidance on Implementing the Capacity
Development Provisions of the Safe Drinking Water Act Amendments of 1996 (USEPA, 1998a)
as the ability to plan for, achieve, and maintain compliance with applicable drinking water
standards. Capacity encompasses three components: technical, managerial, and financial.
Technical capacity is the operational ability of a water system to meet those SDWA
requirements. Key issues of technical capacity include the following:
• Source Water Adequacy—Does the system have a reliable source of water with
adequate quantity? Is the source generally of good quality and adequately
protected?
• Infrastructure Adequacy—Can the system provide water that meets SDWA
standards? What is the condition of its infrastructure, including wells or source
water intakes, treatment facilities, storage facilities, and distribution system? What
is the infrastructure's life expectancy? Does the system have a capital improvement
plan?
• Technical Knowledge and Implementation—Are the system's operators certified?
Do the operators have sufficient knowledge of applicable standards? Can the
operators effectively implement this technical knowledge? Do the operators
understand the system's technical and operational characteristics? Does the system
have an effective operation and maintenance (O&M) program?
Managerial capacity is the ability of a water system's managers to make financial,
operating, and staffing decisions that enable the system to achieve and maintain compliance with
SDWA requirements. Key issues include:
• Ownership Accountability—Are the owners clearly identified? Can they be held
accountable for the system?
• Staffing and Organization—Are the operators and managers clearly identified? Is
the system properly organized and staffed? Do personnel understand the
management aspects of regulatory requirements and system operations? Do they
have adequate expertise to manage water system operations? Do personnel have the
necessary licenses and certifications?
Economic Analysis for the Final RTCR
8-17
September 2012
-------
• Effective External Linkages—Does the system interact well with customers,
regulators, and other entities? Is the system aware of available external resources,
such as technical and financial assistance?
Financial capacity is a water system's ability to acquire and manage sufficient financial
resources to allow the system to achieve and maintain compliance with SDWA requirements.
Key issues include:
• Revenue Sufficiency—Do revenues cover costs?
• Creditworthiness—Is the system financially healthy? Does it have access to capital
through public or private sources?
• Fiscal Management and Controls—Are adequate books and records maintained?
Are appropriate budgeting, accounting, and financial planning methods used? Does
the system manage its revenues effectively?
8.14.1 Requirements of the RTCR
This capacity analysis is presented only for the RTCR, although EPA took similar
considerations into account in the selection of the RTCR over the other options. This process led
to the incorporation of less expensive rule features for systems having fewer capabilities.
The RTCR establishes requirements that may result in the following activities that
influence the TMF capacity of affected PWSs:
1.
Familiarization with Rule Requirements
2.
Revising Sample Siting Plans
3.
Monitoring
4.
Annual Site Visits
5.
Assessments
6.
Corrective Actions
7.
Public Notification
8.14.2 Systems Subject to the RTCR
The RTCR will apply to all PWSs, including 51,972 CWSs, 18,729 NTNCWSs, and
84,136 TNCWSs—154,837 systems in all (Safe Drinking Water Information System/Federal
Version (SDWIS/FED) 2007 4th quarter data). While most will not, some systems may require
increased TMF capacity to comply with the new requirements, or will need to tailor their
compliance approaches to match their capacities. Refer to section 8.14.4 for a detailed discussion
of the changes in TMF capacity for small and large systems.
Economic Analysis for the Final RTCR
8-18
September 2012
-------
8.14.3 Impact of the RTCR on System Capacity
The estimates presented in Exhibits 8.5 and 8.6 reflect the anticipated impact of the
RTCR on system capacity based on the expected measures that systems will be required to adopt.
The extent of the expected impact of a particular requirement on system capacity is estimated
using a scale of 0-5, where 0 represents a requirement that is not expected to have any impact, 1
represents a requirement that is expected to have a minimal impact, and 5 represents a
requirement that is expected to have a very significant impact on system capacity. Criteria used
to develop the scores and associated impacts are discussed further in section 8.14.4.
These impacts are assessed separately for small systems (those serving less than or equal
to 10,000 persons, see Exhibit 8.5) and for large systems (those serving more than 10,000
persons, see Exhibit 8.6). This distinction is necessary because most large systems will face
fewer challenges in implementing the rule than smaller systems. For both large and small
systems, EPA evaluated the capacity impact of each requirement on those systems affected by
that particular requirement. Because in many cases the requirements only affect a small
percentage of systems, the exhibits also display the number of systems and percentage of
systems (of the subset of small or large systems) estimated to be affected by each specific
requirement.
Economic Analysis for the Final RTCR
8-19
September 2012
-------
Exhibit 8.5 Estimated Impact of the RTCR on Small Systems' Technical,
Managerial, and Financial Capacity
(0 = no impact, 1 = minimal impact, and 5 = very significant impact)
Technical
Capacity
Managerial
Capacity
Financial
Capacity
Requirement/
Activity
Number
& Percent of
Small
Systems
Source Water Adequacy
Infrastructure Adequacy
Technical Knowledge and
Information
Ownership Accountability
Staffing & Organization
Effective External Linkages
Revenue Sufficiency
Credit Worthiness
Fiscal Management &
Controls
Familiarization with
rule requirements
150,672 (100%)
0
0
1
1
1
1
1
1
1
Revise Sample
Siting Plans
150,672 (100%)
1
1
1
1
1
1
1
1
1
Monitoring
150,672 (100%)
0
0
2
0
2
0
2
2
2
Annual Site Visits
0 (0%)
0
0
0
0
0
0
0
0
0
Assessments*
151,456
0
0
2
0
2
0
2
2
2
Corrective Actions*
15,206
3
3
3
3
3
3
3
3
3
Public Notification*
11,701
0
0
0
0
0
0
0
0
0
Source:
Number and percent of systems subject to each rule activity derived from Appendix A, Exhibit A.2.z. Impact on
capacity is determined relative to previous regulations based on the cost and number of systems that require
additional capacity to comply with each requirement, as described in section 8.14.4.
Notes:
Small systems are those serving less than or equal to 10,000 persons.
*For these three requirements the number of assessments, corrective actions, and public notifications over the
25-year period of analysis (not number of systems) is shown. A corresponding percentage, therefore, is not
meaningful for these numbers. Additional information is available in Appendix A, Exhibit A.2.z.
1) To analyze the impact of these requirements on system capacity, the requirements believed to have the
most and the least impact on affected systems were analyzed first. These initial analyses were then used
as the bases against which the relative impact of the remaining requirements was assessed. The impact
estimates developed for each requirement were also compared to those developed for the Long Term 2
Enhanced Surface Water Treatment Rule (LT2ESWTR) and the Stage 2 Disinfectants and Disinfection
Byproducts Rule (Stage 2 DBPR), and the Ground Water Rule (GWR) to ensure cross-rule consistency.
2) The scores presented above represent the worst case scenario; the requirements of this rule are expected
to have less impact on the capacity of most systems affected by each requirement.
Economic Analysis for the Final RTCR
8-20
September 2012
-------
Exhibit 8.6 Estimated Impact of the RTCR on Large Systems' Technical,
Managerial, and Financial Capacity
(0 = no impact, 1 = minimal impact, and 5 = very significant impact)
Technical
Capacity
Managerial
Capacity
Financial
Capacity
Requirement/
Activity
Number
& Percent of
Large
Systems
Source Water Adequacy
Infrastructure Adequacy
Technical Knowledge
and Information
Ownership Accountability
Staffing & Organization
Effective External
Linkages
Revenue Sufficiency
Credit Worthiness
Fiscal Management &
Controls
Familiarization with
rule requirements
4,165 (100%)
0
0
1
1
1
1
1
1
1
Revise Sample Siting
Plans
4,165 (100%)
1
1
1
1
1
1
1
1
1
Monitoring
4,165 (100%)
0
0
1
0
1
0
1
1
1
Annual Site Visits
0 (0%)
0
0
0
0
Assessments*
4,617
0
0
1
0
1
0
1
1
1
Corrective Actions*
462
1
1
1
1
1
1
1
1
1
Public Notification*
334
0
0
0
0
0
0
0
0
0
Source:
Number and percent of systems subject to each rule activity derived from Appendix A, Exhibit A.2.z. Impact on
capacity is determined relative to previous regulations based on the cost and number of systems that require
additional capacity to comply with each requirement, as described in section 8.14.4.
Notes:
Large systems are those serving more than 10,000 persons.
* For these three requirements the number of assessments, corrective actions, and public notifications over the
25-year period of analysis (not number of systems) is shown. A corresponding percentage, therefore, is not
meaningful for these numbers. Additional information is available in Appendix A, Exhibit A.2.z.
1) To analyze the impact of these requirements on system capacity, the requirements believed to have the
most and the least impact on affected systems were analyzed first. These initial analyses were then used as
the bases against which the relative impact of the remaining requirements was assessed. The impact
estimates developed for each requirement were also compared to those developed for the LT2ESWTR,
Stage 2 DBPR, and GWR to ensure cross-rule consistency.
2) The scores presented above represent the worst case scenario; the requirements of this rule are expected
to have less impact on the capacity of most systems affected by each requirement.
8.14.4 Derivation of RTCR Scores
EPA developed a 5-point scoring system to analyze the impact compliance with all new
regulations will have on the TMF capacity of PWSs. For each regulation, it is necessary to
complete the following steps:
1. Determine the type and number of PWSs to which the regulation applies.
Economic Analysis for the Final RTCR
8-21
September 2012
-------
2. List all of the requirements of the regulation.
3. Determine the type and number of PWSs to which each requirement applies.
4. Evaluate the impact of each requirement on the capacity of affected PWSs.
The determination of the universe of affected systems and the evaluation of the capacity
impact of individual requirements requires the use of the cost and technical information
contained in the SDWIS, EAs developed for other rules, ICRs, and other supporting
documentation for the rule. These data sources are also used to develop a qualitative description
of the expected response of affected systems to each requirement.
The overall evaluation of the impact of a requirement on the affected systems, presented
in Exhibits 8.5 and 8.6, is based on the impact each requirement and activity has on nine sub-
categories of capacity—three sub-categories under each of the broader divisions of TMF
capacity. Within these sub-categories, EPA evaluated the costs, number of systems affected, and
complexity of each requirement. After estimating the technical, managerial, and financial
impacts within each sub-category, EPA assigned the scores using best professional judgment.
Costs were considered cumulatively for each requirement and activity for small and large
systems. This score reflects the additional capacity that systems will need to develop to comply
with each requirement. Due to a lack of available information on operating budgets, this analysis
does not include a quantitative component.
To ensure cross-rule consistency, to standardize the assignment of numerical scores, and
to minimize the subjectivity of the scoring system, the requirements for systems under the RTCR
are compared to the requirements of those regulations for which capacity impact analyses have
already been conducted (e.g., Long Term 1 Enhanced Surface Water Treatment Rule
(LT1ESWTR), LT2ESWTR, Stage 2 DBPR, GWR). Similar requirements are assigned similar
impact scores.
8.14.5 Small Water Systems (Those Serving 10,000 or Fewer People)
Small systems will likely face only a small challenge to their technical and managerial
capacity as a result of efforts to familiarize themselves with the monitoring requirements of the
RTCR. Routine and repeat monitoring requirements under the RTCR are essentially the same as
under the 1989 TCR, with more explicit criteria to qualify for reduced monitoring. Therefore,
understanding the RTCR monitoring requirements is not expected to pose many new technical or
managerial capacity issues for small systems.
Small system technical and managerial capacity may be affected by the assessment
requirements of the RTCR. Performing assessments may require the system to evaluate staffing
levels and/or the need to access outside assistance to conduct the assessments in addition to
providing training to ensure that system staff understand how those assessments are to be
performed. Reporting, record-keeping, and data administration requirements will also affect the
managerial capacity of small systems.
Economic Analysis for the Final RTCR
8-22
September 2012
-------
Small systems that are required to take corrective action are expected to experience the
most significant financial challenge since some corrective actions may consist of a large, one-
time capital expenditure to resolve the problem.
8.14.6 Large Water Systems (Those Serving Greater Than 10,000 People)
Large systems will likely not face any significant challenge to their technical and
managerial capacity as a result of efforts to familiarize themselves with the RTCR. Most large
systems are familiar with the 1989 TCR and there are no changes in the basic monitoring
requirements for large systems under the RTCR. They are therefore assumed to already have the
TMF capacity in place for the RTCR.
Only large systems performing assessments and corrective actions would be expected to
face a significant challenge meeting the TMF capacity requirements. However, this requirement
is only necessary when monitoring reveals potential problems, and this is not expected to occur
significantly in large systems above that experienced under the 1989 TCR. Many large systems
already have the TMF capacity to conduct assessments and corrective actions if they are needed.
These systems will be affected less significantly than smaller systems that have to implement
corrective actions because it is recognized that they are typically already implementing similar
assessments and corrective actions when a routine monitoring sample tests positive for fecal
indicators under the 1989 TCR.
Economic Analysis for the Final RTCR
8-23
September 2012
-------
9 Comparison of Benefits and Costs
9.1 National Benefits and Costs of the RTCR Considered in Comparison to the 1989
Total Coliform Rule and Alternative Option
The RTCR is consistent with the majority of the Total Coliform Rule/Distribution System
Advisory Committee (TCRDSAC) Agreement in Principle (AIP). The Final Rule was developed
in consideration of the comments received on the proposed rule and therefore is not identical to
the AIP. Chapter 3 of this EA provides more information about the specific requirements of the
RTCR. The primary discussion of costs and benefits in this chapter focuses on the requirements
of the RTCR in comparison to the 1989 Total Coliform Rule (TCR) and the Alternative option
considered.
The primary benefit of the RTCR is to further reduce the risk of fecal contamination of
public drinking water from the current baseline risk under the 1989 TCR. The indicator of this
benefit that is most feasible to predict is the reduction in E. coli occurrence in public water
systems (PWSs). As a fecal contamination indicator, E. coli can also co-occur with other
pathogenic organisms shed in feces, such as viruses, bacteria, and pathogenic protozoa at a rate
that is not quantified.59 EPA believes that a reduction in E. coli occurrence, together with other
RTCR actions, will result in reduced PWS fecal contamination and will likely facilitate a
consequent reduction in endemic and epidemic waterborne disease in the United States.
The 1989 TCR mandates sampling for total coliforms (TC) and E. coli in PWSs in the
U.S. However, under the 1989 TCR, PWSs that encounter samples testing positive for TC oris.
coli have the option to implement an effective long-term corrective action, a short-term
corrective action, or they can choose not to implement any corrective action. This wide
variability in effectiveness of corrective actions increases the potential for exposure to fecally-
contaminated drinking water when the response to E. coli occurrence is inadequate or
ineffective.
EPA is augmenting the 1989 TCR with, among other components, the requirement for
systems to implement mandatory assessments and corrective actions. PWSs would be required to
conduct either a Level 1 or Level 2 assessment, depending on the severity of a trigger (based on
TC vs. E. coli and the number of triggers). As described in Chapter 3, the Level 1 and 2
assessments require a PWS to more formally pursue the cause of triggers, (similar to non-acute
or acute violations under the 1989 TCR), and to address the problem with an appropriate
corrective action. It is the combination of these two requirements (assessments and corrective
actions) from which most of the net benefits and costs of the RTCR would derive. Because of the
benefits from assessment and corrective action and for the other reasons discussed in this chapter
and in analyses and discussion presented throughout this economic analysis (EA), EPA
concludes that the RTCR will improve public health protection compared to the 1989 TCR.
A significant amount of the costs of the additional activities required under the RTCR
would be offset by the reduced costs from some decreases in monitoring (additional routine and
59 Chapter 2 of this economic analysis provides examples of waterborne pathogens that can occur in PWSs. Also,
Edberg, R. (2000) discusses the use of E. coli as an indicator of drinking water quality.
Economic Analysis for the Final RTCR
9-1
September 2012
-------
repeat) and decreases in public notification (PN). The decreases in monitoring may limit the
ability of PWSs to identify E. coli positive samples at the levels found under the 1989 TCR when
using the same monitoring techniques. However, predictive modeling results show that
significantly more contamination events will be prevented by assessments that lead to additional
corrective actions under the RTCR than would be missed by reduced additional routine and
repeat monitoring for systems serving 4,100 or fewer people.60 Systems serving 4,100 people or
more have no changes in the number of required samples under the RTCR. Some of this
reduction in both monitoring and costs would be offset by the increase in routine monitoring
predicted for systems serving 1,000 or fewer people because fewer systems are likely to qualify
for reduced monitoring based on the more stringent requirements under the RTCR.61
Further cost reductions may be achieved by the revisions to the requirement for a
mandatory sample siting plan under the RTCR, which would increase monitoring efficiency by
having the operator revise existing sample siting plans to identifying repeat sample locations in
addition to including regular sampling sites that are representative of the water throughout the
distribution system. The sample siting plan may also specify repeat sample locations that may, if
approved by the state, satisfy source water sampling requirements as well; this benefit would
likely apply to some of the PWSs with limited or no distribution systems.
The increases in routine monitoring for systems that no longer qualify for reduced
monitoring, along with more efficient monitoring based on approved monitoring plans, are
expected to increase the ability of PWSs to detect E. coli a larger percentage of the time when it
is present. This increase in detection will be partially offset by reductions in additional routine
and repeat monitoring.
The Alternative option considered in this EA would provide benefits equal to or greater
than the RTCR in terms of decreased potential health risks from PWSs delivering contaminated
water to the public. However, monitoring costs under the Alternative option would increase
significantly, particularly for small systems, causing a significant challenge to the effective and
efficient implementation of the Alternative option. The remainder of this chapter provides further
detail on the costs and benefits and how they compare under the 1989 TCR, RTCR, and
Alternative option:
• Section 9.1.1 discusses national benefits.
• Section 9.1.2 discusses national costs.
• Section 9.2 discusses uncertainty and non-quantified benefits.
• Section 9.3 presents a comparison of regulatory options in cost/benefit terms.
60 Chapter 6 of this EA explores the tradeoff in a stepwise analysis of the risk benefit resulting from implementation
of assessments resulting in corrective actions vs. the potential increase in risk due to decreased sampling frequency.
That analysis concludes that for each type of system and size category, the RTCR would avoid a larger number of
acute events than it would fail to diagnose.
61 Chapter 5 (Section 5.3) includes a description of the determination of PWS sampling frequencies that are
applicable beginning in year 6 following RTCR promulgation.
Economic Analysis for the Final RTCR
9-2
September 2012
-------
9.1.1 National Benefits of the Regulatory Options Considered
The benefits analysis performed in this EA and summarized in this section consider the
net changes in TC and E. coli occurrence under the RTCR as compared to the 1989 TCR.62 In
promulgating the RTCR, EPA expects to further reduce the overall risk of contamination of
public drinking water from the current baseline risk under the 1989 TCR. The options considered
during development of this rule and analyzed as part of this EA are designed to achieve this
reduction while maintaining public health protection in a cost-effective manner.
This section examines the benefits in terms of trade-offs between compliance with the
1989 TCR, the RTCR, and Alternative option. Based on limitations in available data (described
further in Chapter 6, Section 6.3), EPA determined that benefits could not be calculated in terms
of avoided cases of (or costs related to) morbidity or mortality. EPA used several methods to
qualitatively evaluate the benefits of the RTCR. The qualitative evaluation uses both the
judgment of EPA as informed by the TCRDSAC deliberations as well as quantitative estimates
of changes in E. coli occurrence and counts of systems implementing corrective actions. The
evaluation characterizes, in relative terms, the reduction in risk for each regulatory option as
compared to baseline conditions.
Since E. coli is an indicator of fecal contamination, EPA assumed that a decrease in E.
coli occurrence in the distribution system would be associated with a decrease in fecal
contamination in the distribution system. In general, this decrease in fecal contamination should
reduce the potential risk to human health for PWS customers. Thus, any reduction in E. coli
occurrence is considered a benefit of the RTCR and Alternative option. Also, since fecal
contamination may contain waterborne pathogens including bacteria, viruses, and parasitic
protozoa, in general, a reduction in fecal contamination should also reduce the risk from these
other contaminants.
As presented in Exhibit 4.9, the percentages of samples that are positive for TC and E.
coli are generally higher for PWSs serving 4,100 or fewer people than those serving more than
4,100 people. PWSs with higher TC and E. coli occurrence are more likely to be triggered into
assessments and corrective action. As discussed previously, EPA believes that the assessments
and corrective action under the RTCR and Alternative option will lead to a decrease in TC and E.
coli occurrence. Because the PWSs serving 4,100 people or fewer have a higher initial E. coli
occurrence and will be triggered into more assessments and corrective actions than larger PWSs,
the increase in benefits for these small systems will be more evident as compared to the larger
systems. In particular, model results suggest that customers of small ground water transient
noncommunity water systems (TNCWSs) serving 100 people or fewer, which constitute
approximately 40% of PWSs, would experience the most obvious benefit under the RTCR. That
is, the occurrence of E. coli is predicted to decrease more for TNCWSs than for other systems
types.
62 The 1989 TCR is the baseline used in the RTCR EA, and consists of the rule components and resulting activities
of the 1989 TCR with effects of the Ground Water Rule (GWR) implementation in 2010 incorporated. Chapter 4 of
this EA presents information used to develop this baseline, and Chapter 5 (Section 5.3.1) describes how the
occurrence and predictive model incorporates the GWR effects.
Economic Analysis for the Final RTCR
9-3
September 2012
-------
As discussed previously, because there was insufficient data on the co-occurrence of E.
coli and waterborne pathogens, it was not possible to quantify the health effects in the human
population served by PWSs. Thus, EPA employed a qualitative approach as the primary method
for analyzing the expected change in exposure to fecal contamination based on each individual
rule component. The qualitative analysis considers the anticipated changes in sampling and
corrective action regimens and estimates the corresponding anticipated reduction or increase in
exposure to fecal contamination. Because the reduced exposure to fecal contamination conveys
inadequate information about pathogen exposure, the analysis in this section is a hazard analysis
rather than a risk analysis. Nevertheless, because the intent is to reduce risk by reducing hazard
(potential exposure to fecal contamination and/or waterborne pathogens), risk is used as the
discussion endpoint. Section 9.1.1.1, which follows, summarizes the results of the qualitative
analysis, and Chapter 6 (Section 6.2) presents this information in more detail.
Although a qualitative analysis was the primary method employed for analyzing
potential changes in risk, quantitative measures were used to support the qualitative conclusions,
providing additional perspective on potential changes in risk. In particular, EPA considered
modeled predictions of the reductions in the numbers of violations/assessment triggers and of the
number of corrective actions to be performed as they may relate to risk. Section 9.1.1.2 discusses
the predicted changes in these outcomes. More detailed discussions of the predictive model and
analyses results are presented in Chapters 5 and 6.
9.1.1.1 Qualitative Comparison—Relative Risks of the RTCR and Alternative Option
Compared to Baseline
When revising an existing drinking water regulation, one of the main concerns is to
ensure that backsliding on water quality and public health protection does not occur. Risk
reduction for the RTCR is characterized by the activities performed that are presumed to reduce
risk of exposing the public to contaminated water. These activities are considered under each
rule component presented in Exhibit 9.1b. Under repeat and additional routine monitoring
provisions for both the RTCR and Alternative option, there is a potential to contribute to
increased risk for some PWS customers because TC monitoring frequency may be reduced for
some PWSs. However, this increase in risk is expected to be more than offset by potential
decreases in risk from increased routine monitoring and the addition of the assessments and
corrective action provisions that will find and fix problems identified by monitoring.
Exhibit 9.1a illustrates the predicted reduced frequency at which TC-positives occur
subsequent to the implementation of the RTCR and Alternative option. TC occurrence is used as
a surrogate for indicating the existence of potential pathways through which waterborne
pathogens may enter a PWS. Exhibit 9.1a illustrates the combined effects on TC occurrence
resulting from changes in monitoring and the effects of assessments and corrective actions for
the different rule options illustrated. The relative trends evident in Exhibit 9.1a for TNCWSs also
pertain to other PWS categories as illustrated in Chapter 5 and Appendix B of the RTCR EA.
EPA chose to include the characterization for TNCWSs because they represent the system
category of largest influence on the national impacts.
Economic Analysis for the Final RTCR
9-4
September 2012
-------
Exhibit 9.1a Ground Water (GW) Transient Noncommunity Water Systems
(Serving < 4,100) TC Occurrence
0.045
0.040
>. 0.035
03
(ft
J 0.030
i-
> 0.025
"(ft
£ 0.020
4-
O
.2 0.015
o
rc
£ 0.010
0.005
0.000
0
Notes:
1) Six Year 2005 TC-positive occurrence is representative of all GW TNCWS. The rate presented may
underestimate the occurrence for systems serving 25-4,100 individuals.
2) Graph shows the 30-year modeled period discussed in Ch. 5. Model years 3-27 represent the 25-year
period of analysis for this EA. Model year 11 begins the steady state, during which systems that
qualified for reduced monitoring are now sampling on their reduced schedules. The criteria and timing
of this monitoring adjustment is discussed in Section 5.3.2.2 ofthis EA.
3) The results represented by the curves for 1989 TCR, RTCR, and Alternative option all incorporate the
effects of the GWR.
The effect that the changes to PN requirements for monthly/non-acute MCL violations
have on risk is difficult to predict. Some factors, such as reduction in available public
information and possible PWS complacency, lead to a potential increase in risk, whereas other
factors, such as less confusion (PN more in line with potential health risks) and PWSs resources
used more efficiently, lead to a potential decrease, as discussed in Exhibit 9.1b. This change to
PN addresses a key concern expressed by various stakeholders in the advisory committee and
during the Six-Year Review 1 comment solicitation process. By revising the PN requirements
and adding assessment and corrective action requirements, the Agency expects less public
confusion, more effective use of resources, and increased transparency. Other rule components
are expected to have a negligible effect on risk. However, the overall effect of the RTCR is
expected to be a further reduction in risk from the current baseline risk under the 1989 TCR.
Economic Analysis for the Final RTCR
9-5
September 2012
-------
Chapter 6 presents a detailed discussion of the potential influence on health risk for each rule
component.
Economic Analysis for the Final RTCR
9-6
September 2012
-------
Exhibit 9.1b Potential Changes in Risk under the RTCR and Alternative Option Relative to the 1989 TCR
Rule Component
Factors Leading to a Potential
Increase in Risk
Factors Leading to a Potential Decrease in
Risk
Overall Predicted Change in
Risk
RTCR
Alternative
option
RTCR
Alternative
option
RTCR
Alternative
option
Implementation
Activities
None
None
None
None
No change
No change
Routine
Monitoring
(Including
Reduced
Monitoring)
None
None
Increased stringency
in requirements to
qualify for reduced
monitoring along with
requirement to return
to baseline monitoring
upon loss of these
criteria is expected to
result in decreased
risk (i.e., PWSs that
qualify for reduced
monitoring will be
better operated, PWSs
that no longer qualify
will monitor more
frequently).
PWSs all monitor monthly
in the first few years of
implementation of the
RTCR, which is an
increase in sampling
frequency for systems
that monitor quarterly or
annually under the 1989
TCR. After the first few
years, systems may
reduce to quarterly, but
none may reduce to
annual monitoring,
creating a decrease in risk
for systems on annual
monitoring under the 1989
TCR.
Decrease
Decrease
Repeat
Monitoring
Required repeat
samples reduced
from 4 to 3 for
systems serving
<1,000 people
Required repeat
samples reduced
from 4 to 3 for
systems serving
<1,000 people
None
None
Increase
Increase
Economic Analysis for the Final RTCR
9-7
September 2012
-------
Rule Component
Factors Leading to a Potential
Increase in Risk
Factors Leading to a Potential Decrease in
Risk
Overall Predicted Change in
Risk
Additional
Routine
Monitoring
Additional routine
samples are no
longer required for
PWSs monitoring
monthly.
Ground water
PWSs serving
<1,000 people
would reduce
additional routine
samples from 5 to
3.
Additional routine
samples are no
longer required
for PWSs
monitoring
monthly.
Ground water
PWSs serving
<1,000 people
would reduce
additional routine
samples from 5 to
3.
None
None
Increase
Increase
Annual Site Visits
None (only states
currently
performing annual
site visits are
expected to
continue)
Annual
monitoring is not
permitted so
annual site visits
will no longer be
conducted.
None (only states
currently performing
annual site visits are
expected to continue)
None
No change
Increase
Assessments
None
None
Mandatory
assessments are a
new requirement.
Mandatory assessments
are a new requirement.
Decrease
Decrease
Corrective
Actions
None
None
Mandatory corrective
actions are a new
requirement.
Mandatory corrective
actions are a new
requirement.
Decrease
Decrease
PN —
Monthly/Non-
Acute MCL
Violations
Reduction in
available public
information
Possible PWS
complacency
Reduction in
available public
information
Possible PWS
complacency
Less confusion (PN
more in line with
potential health risks)
PWS resources used
more efficiently
Less confusion (PN more
in line with potential
health risks)
PWS resources used
more efficiently
Unknown
Unknown
Economic Analysis for the Final RTCR
9-8
September 2012
-------
Rule Component
Factors Leading to a Potential
Increase in Risk
Factors Leading to a Potential Decrease in
Risk
Overall Predicted Change in
Risk
PN—Monitoring
and Reporting
Violations
None
None
Improved focus of
required PN for rule
aspects with potential
adverse health
consequences,
notably E. coli MCL
violations and failure
to conduct
assessments and
corrective actions, will
motivate PWSs to
conduct sampling and
other treatment
technique
requirements.
Improved focus of
required PN for rule
aspects with potential
adverse health
consequences, notably E.
coli MCL violations and
failure to conduct
assessments and
corrective actions, will
motivate PWSs to
conduct sampling and
other treatment technique
requirements.
Decrease
Decrease
Overall
Decrease
Decrease
Note:
1) Detailed discussion of the rationale for determinations of potential risk for each rule component is presented in Ch. 6 (Section 6.2) of this EA. Implementation
activities consist of administrative activities by PWSs and states to implement the rule.
2) Assessment of potential changes in risk for monitoring components is an overall assessment. Potential changes (or static state) of risk for particular system
sizes and types differ according to individual regulatory requirements and are discussed in Section 6.2. Chapter 3 provides a detailed description of the
regulatory components for all three regulatory options, and the preamble to the RTCR provides additional discussion of the TCRDSAC process and the
rationale underlying the structure of the regulatory options considered.
Economic Analysis for the Final RTCR
9-9
September 2012
-------
9.1.1.2 Comparison of Quantified Benefits between 1989 TCR and RTCR and
Alternative Option
The quantified portion of the benefits analysis focuses on several measures that
contribute to the changes in risk expected under the RTCR. Specifically, EPA modeled the
predicted outcomes based on each regulatory option considered—baseline (1989 TCR), the
RTCR, and the Alternative option—in the form of estimates of non-acute violations for the 1989
TCR and assessment triggers for the RTCR and Alternative option; E. coli violations; and the
number of corrective actions implemented under each option.
This section presents a summary of the estimated impacts of the RTCR and Alternative
option in comparison to the 1989 TCR. Evaluation of each of these endpoints informed EPA's
understanding of potential changes to the underlying quality of drinking water. In particular, the
number of corrective actions performed has a strong relationship to potential improvements in
water quality and public health. For a given level of TC and E. coli occurrence, an increase in the
number of corrective actions implemented would lead to improved water quality. However, a
reduction in sampling leads to a reduction in TC and E. coli positives being found, which in turn
leads to a reduction in assessments and corrective actions being implemented. The number of TC
and E. coli positives that are prevented, missed, or found under each regulatory option is
considered in comparison to those predicted under the 1989 TCR results in estimates of annual
non-acute and acute violations (1989 TCR) and assessment triggers (RTCR and Alternative
option). Section 6.4 of this EA presents a step-wise uncertainty analysis of the competing effects
of additional protective activity (assessments and corrective actions) and decreased additional
routine and repeat sampling of the regulatory options compared to the 1989 TCR. The results of
this uncertainty analysis showed that for all categories of systems, more TC and E. coli positives
would be prevented than missed.
For each of the graphs presented in Exhibit 9.2 through Exhibit 9.7 there are two main
model drivers that impact the endpoints depicted: the total number of samples taken over time
(including routine, additional routine, and repeat) and corrective actions taken. When looking at
the comparisons between the 1989 TCR with the RTCR across all PWSs, the overall impact of
the total numbers of samples taken is negligible because the total number of samples predicted to
be taken throughout the period of analysis is almost the same (approximately 82 million samples
under either the 1989 TCR or the RTCR) as shown in Exhibits 6.2-6.4 of this EA. For the
Alternative option, the analysis predicts that approximately 88 million total samples will be taken
over the period of analysis. Based on the relationships of total samples taken between the 1989
TCR, RTCR, and Alternative option, the best way to interpret the graphs presented in this section
is in a step-wise manner.
The first comparison is between the 1989 TCR and the RTCR. Because a similar total
number of samples is taken under the 1989 TCR and RTCR, the major effect seen in the graphs
can be isolated to the effects that implementation of corrective actions will have on underlying
occurrence and how that occurrence influences the endpoint in question. In each graph, this is
depicted by a marked reduction in the endpoint under the RTCR compared to the TCR and is a
reflection of overall better water quality. The second comparison can then be made of the
Alternative option against the RTCR. In each graph, the endpoints for the Alternative option are
above those for the RTCR and represent an additional benefit over the RTCR. This additional
Economic Analysis for the Final RTCR
9-10
September 2012
-------
benefit is primarily a function of the additional diagnostic abilities gained through increased
monitoring under the Alternative option, and is especially prominent in the early years of the
analysis when all systems are required to monitor at least monthly.
More detailed descriptions of each endpoint considered in terms of the evaluation process
described previously are provided in this section as they apply to the individual graphs in
Exhibits 9.2 through 9.7. The graphs shown in this section are presented first in nondiscounted
terms and then based on a discount rate of 3% to reflect the reduced valuation of potential
benefits over time, consistent with the presentation of costs in the section that follows. Graphs of
benefits discounted using 7% discount rates are presented in Appendix B to this EA.
Exhibit 9.2 shows the effect (on average across all PWSs) of the RTCR and the
Alternative option on the annual number of non-acute violations (Level 1 assessment triggers
under the RTCR and Alternative option) over time. The estimated reduction of approximately
1,100 trigger events in moving from the 1989 TCR to the RTCR is a net result from the
following effects described previously in this section: improved water quality (events prevented);
retained diagnostic power (events found), which in turn leads to prevention of some future
events; and reduced diagnostic power (missed events). The amount of this reduction would be
some amount lower if the analysis could account for events where contamination was present but
not diagnosed (missed events). This is why a similar, but smaller, reduction from the 1989 TCR
is seen under the Alternative option. During the first 9 years of sampling under the RTCR and
Alternative option sampling regimens, diagnostic power under the Alternative option is at a
maximum based on the all-monthly sampling requirement. In year 9, following a 5-year period
of assessment,63 all systems that qualified for reduced monitoring begin sampling according to
their new regimens. This is the start of the relatively steady state of the period of analysis, driven
by the static nature of the percentages of systems following monthly, quarterly, or annual
sampling regimens throughout the remainder of the analysis (years 9 through 25). Trigger events
under the Alternative option remain at a higher steady state than under the RTCR based on the
increased diagnostic ability provided by more frequent sampling under the Alternative option.
The additional number of triggers identified by increased sampling under the Alternative option
translates into greater potential benefits than those under the RTCR.
Exhibit 9.3 shows the effect (on average across all PWSs) of the RTCR and the
Alternative option with respect to acute (E. coli) violations found over the 25-year period of
analysis in comparison to the 1989 TCR. The overall reduction in annual acute violations under
the RTCR of approximately 170 events is a measure that should correlate more closely with
expected benefits (i.e., reductions in adverse health outcomes) than non-acute violations because
acute violations are a direct result of measurement of E. coli in water. Again a similar, but
smaller, reduction is seen under the Alternative option after steady state is achieved. This is the
result of two off-setting effects. The true number of steady state violations under the Alternative
option is lower, because there is a greater likelihood that violations will be found and fixed.
However, the additional monitoring leads to a higher percentage of violations being detected.
This second effect outweighs the first, so that the total number of detected violations in the
steady state is higher than for the RTCR, even though the underlying true number of violations is
63 The assessment period is 3 years for CWSs, and CWSs begin sampling on their reduced regimens in year 7 after
promulgation.
Economic Analysis for the Final RTCR
9-11
September 2012
-------
lower. This lower number of true violations means that the Alternative option is more protective
of public health, even though more violations are detected.
Exhibit 9.4 presents estimates over the 25-year period of analysis of the increase in
corrective actions (on average across all PWSs) attributable to the regulatory options considered.
The performance of these additional corrective actions is expected to result in the most direct
benefits under the RTCR. Because only the numbers of net corrective actions estimated under
the RTCR and Alternative option (i.e., corrective actions in addition to those estimated under the
1989 TCR) were modeled, the reference point for comparison to the 1989 TCR is the base (zero)
line in the graph. The RTCR EA assumes that corrective actions are being performed under the
1989 TCR. These are taken into account by assuming only a modest net increase in effective
corrective actions implemented, calculated as 10% of assessments performed under each
regulatory option that are net of those performed under the 1989 TCR.
Exhibit 9.4 indicates that more corrective actions would be implemented under the
Alternative option than under the RTCR. This is driven, again, by the increased diagnostic power
of more sampling and reflects additional potential benefits beyond those gained under the RTCR.
However, not quantified are the numbers of corrective actions that would be identified by the
annual inspections required to qualify for annual monitoring under the RTCR but not under the
Alternative option. Inclusion of the additional corrective actions taken under the RTCR due to
the annual inspection requirement could narrow the difference between the potential benefits of
the Alternative option and the RTCR.
Taken together, Exhibits 9.2-9.4 indicate that the modeled endpoints for the RTCR and
Alternative option predict positive benefits in comparison to the 1989 TCR; in particular, the
Alternative option would capture more benefits than the RTCR. These outcomes are consistent
with the qualitative assessment of the benefits summarized in Section 9.1.1.1.
The assessment based on Exhibits 9.5-9.7 is also consistent with the qualitative analysis
in Section 9.1.1.1. For each of the discounted endpoints presented over time in Exhibits 9.5-9.7,
the graphs show that (on average across all PWSs) the Alternative option would provide more
benefits than the RTCR, and both provide more benefits than the 1989 TCR. The major
difference between the RTCR and Alternative option is the increased monitoring that would be
required under the Alternative option. The increased diagnostic ability of the extra samples taken
under the Alternative option is seen in the large difference in the endpoint counts through the
first several years in the following graphs. Absent this effect, the Alternative option would mirror
the RTCR in the graphs. Thus, even though the predicted endpoints are greater than the 1989
TCR at first, it is due to initially finding more problems through monitoring. This would reflect a
frontloading of benefits under the Alternative option at the beginning of the implementation
period. The benefits, however, would tend to even out over time between the RTCR and
Alternative option as eligible systems qualify for less frequent (quarterly) monitoring under the
Alternative option.
Economic Analysis for the Final RTCR
9-12
September 2012
-------
Exhibit 9.2 Estimates of Non-Acute Violations (TCR) and Level 1 Assessment
Triggers (RTCR and Alternative Option)
a>
O)
O)
m >
£ a>
E 3
8 §
Q>
(/) jj
v> 0
-1
| 8
S j£
03
2" -p
hI
a> q.
§°
TT- —
IS
03 ££
o O
>£
a> —
¦4-»
3
O
<
14,000
13,000
12,000
11,000
10,000
9,000
3,000
7,000
3,000
1989 TCR - Non-Acute Violations
RTCR - Level 1 Assessment Triggers
Alt Option - Level 1 Assessment Triggers
5,000
4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Time (Years)
Source: RTCR occurrence model output.
Notes:
1) The X-axis begins at Year 4 after rule promulgation, which is the first year under full implementation of the
RTCR or Alternative option.
2) The annual rates of non-acute violations (TCR) and Level 1 assessment triggers (RTCR and Alternative
option) as predicted by the model reach a steady state beginning in approximately Year 9, by which time
PWSs that would be expected to meet the criteria for reduced monitoring would have begun it.
3) Non-acute violations/Level 1 assessment triggers are predicted to be lower under the RTCR because,
although both options are preventing contamination events as discussed in Section 6.3 ofthis EA, the
Alternative option is finding more because of its increased sampling and correspondingly higher diagnostic
power.
Economic Analysis for the Final RTCR
9-13
September 2012
-------
Exhibit 9.3 Estimates of Acute Violations (TCR) and E. coli MCL Violations (RTCR
and Alternative Option)
(T3
cz
o
I—
£
if)
c
o
If
5s
O T3
^ a?
° 3
° 8
UJ w
1 C
ra o
C£
O
> o
oo S-
CT) O
T". -4-*
if) <
c
o
4-»
TO
O
>
a>
3
O
<
1,300
1,200
1,100
1,000
900
800
" 700
600
500
400
• 1989 TCR-Acute Violations
-RTCR - E coli MCL Violations
Alt Option - E coli MCL Violations
\ / ^
4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Time (Years)
Source: RTCR occurrence model output.
Notes:
1) X-axis begins at Year 4 after rule promulgation, which is the first year under full implementation of the RTCR or
Alternative option.
2) The annual rates of acute violations (TCR) and E. coli violations (RTCR and Alternative option) as predicted by
the model reach steady state in approximately Year 9, by which time PWSs that would be expected to meet the
criteria for a reduced monitoring schedule would have begun it. Estimates represent the annual number of acute
violations found by each option and the TCR.
3) Acute violations are predicted to be lower under the RTCR because, although both options are preventing
contamination events as discussed in Section 6.3 of this EA, the Alternative option is finding more because of
its increased sampling and correspondingly higher diagnostic power.
Economic Analysis for the Final RTCR
9-14
September 2012
-------
Exhibit 9.4 Estimates of Corrective Actions
>
Q>
3
£
-o
b
c
o
(/>
c
o
+-»
o
<
d>
>
+-»
o
a>
i—
i_
o
O
1,600
1,500
1,400
1,300
1,200
1,100
1,000
900
800
700
600
500
- RTCR - Corrective Actions
Alt Option - Corrective Actions
—i 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1—
4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Time (Years)
Source: RTCR occurrence model output.
Notes:
1) X-axis begins at Year 4 after rule promulgation, which is the first year under full implementation of the RTCR or
Alternative option. The annual rates of corrective actions as predicted by the model reach a steady state
beginning approximately in Year 9, by which time PWSs that would be expected to meet the criteria for reduced
monitoring would have begun it.
2) Includes L1 and L2 corrective actions. All corrective actions performed are in addition to activity under the 1989
TCR, which does not require corrective actions. Therefore the 1989 TCR is not included in this graph.
3) Corrective actions are predicted to be higher under the Alternative option because, although both options are
preventing contamination events as discussed in Section 6.3 of this EA, the Alternative option is finding more
because of its increased sampling and correspondingly higher diagnostic power.
Economic Analysis for the Final RTCR
9-15
September 2012
-------
Exhibit 9.5 Discounted Estimates of Non-Acute Violations (TCR) and Level 1
Assessment Triggers (RTCR and Alternative Option) (3% Discount Rate)
13,000
1989 TCR - Non-Acute Violations
12,000
RTCR - Level 1 Assessment Triggers
11,000
Alt Option - Level 1 Assessment
Triggers
10,000
CO
9,000
8,000
7,000
6,000
CD ¦
CO =
O) o
+¦»
— Q_
5,000
4,000
3,000
2,000
U)
O)
1,000
4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Time (Years)
Source: RTCR occurrence model output.
Notes:
1) X-axis begins at Year 4 after rule promulgation, which is the first year under full implementation of the RTCR or
Alternative option.
2) The annual rates of non-acute violations (TCR) and Level 1 assessment triggers (RTCR and Alternative
option) as predicted by the model reach a steady state beginning in approximately Year 9, by which time
PWSs that would be expected to meet the criteria for reduced monitoring would have begun it.
3) Non-acute violations/Level 1 assessment triggers are predicted to be lower under the RTCR because,
although both options are preventing contamination events as discussed in Section 6.3 ofthis EA, the
Alternative option is finding more because of its increased sampling and correspondingly higher diagnostic
power.
Economic Analysis for the Final RTCR
9-16
September 2012
-------
Exhibit 9.6 Discounted Estimates of Acute Violations (TCR) and E. coli MCL
Violations (RTCR and Alternative Option) (3% Discount Rate)
oz
o
i—
E.
"w 2
is
J2 e
O 3
5 o
o
_l (/>
o Q
O «
O Q)
"jf
I-
re c
— a>
o: (/>
O £
i- a.
(D —
CO -p
2 I
'—' '•*->
(/> Q.
£ °
.2 ~
re <
O T3
> re
a>
4-»
3
o
<
1,200
1,100
1,000
900
800
700
600
500
400
300
200
100
— 1989 TCR - Acute Violations
¦ RTCR - E coli MCL Violations
Alt Option - E. coli MCL Violations
—i 1 1 1 1 1 1 1 1 1 1 i 1 1 1 1 1 1 1 1—
4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Time (Years)
Source: RTCR occurrence model output.
Notes:
1) X-axis begins at Year 4 after rule promulgation, which is the first year under full implementation of the RTCR or
Alternative option.
2) The annual rates of acute violations (TCR) and E. coli violations (RTCR and Alternative option) as predicted by
the model reach steady state in approximately Year 9, by which time PWSs that would be expected to meet the
criteria for reduced monitoring would have begun it.
3) Acute violations are predicted to be lower under the RTCR because, although both options are preventing
contamination events as discussed in Section 6.3 of this EA, the Alternative option is finding more because of
its increased sampling and correspondingly higher diagnostic power.
Economic Analysis for the Final RTCR
9-17
September 2012
-------
Exhibit 9.7 Discounted Estimates of Corrective Actions
(3% Discount Rate)
RTCR - Corrective Actions
Alt Option - Corrective Actions
d>
+-»
to
+-»
c
3
o
o
>
00
oT
c
a)
(/)
(1)
(/)
c
o
+-»
o
<
a>
>
+-»
o
CD
i—
i—
o
O
1,400
1,300
1,200
.100
.000
4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Time (Years)
Source: RTCR occurrence model output.
Notes:
1) X-axis begins at Year 4, which is the first year under full implementation of the RTCR or Alternative option.
2) Includes L1 and L2 corrective actions. The annual rates of corrective actions as predicted by the model reach
a steady state beginning in approximately Year 9, by which time PWSs that would be expected to meet the
criteria for reduced monitoring would have begun it. All corrective actions performed are in addition to activity
under the 1989 TCR, which does not require corrective actions. Therefore the 1989 TCR is not included in this
graph.
3) Corrective actions are predicted to be higher under the Alternative option because, although both options are
preventing contamination events as discussed in Section 6.3 of this EA, the Alternative option is finding more
because of its increased sampling and correspondingly higher diagnostic power.
9.1.2 National Cost Summary
To understand the net impacts of the RTCR on PWSs and states, EPA first used available
data, information, and best professional judgment to characterize how PWSs and states are
currently implementing the 1989 TCR. Then, EPA considered the net change in costs (i.e.,
incremental costs over the 1989 TCR) that would result from implementing the RTCR or
Economic Analysis for the Final RTCR
9-18
September 2012
-------
Alternative option as compared to the costs of continuing with the 1989 TCR. The objective was
to provide the net change in costs resulting from revisions to the 1989 TCR rather than absolute
total costs of implementing the TCR as revised by the RTCR. The exhibits in this section present
this net change in costs several different ways to help analyze the impacts of the RTCR.
Costs are estimated for different PWS types and size categories (nine size categories are
used based on population served) using unit costs obtained from the advisory committee
technical work group and vendors and presented in the Technology and Cost Document for the
Final Revised Total Coliform Rule (USEPA, 2010d). Cost analyses for PWSs include estimates
to implement the rule; to revise sample siting plans; to conduct routine monitoring, additional
routine monitoring, and repeat monitoring; to perform Level 1 and Level 2 assessments and
implement corrective actions; and to provide PN in the case of violations. State cost analyses
include estimates of the labor burdens that states would incur, including staff training on RTCR
requirements and conducting annual administration, reviewing monitoring reports, reviewing and
approving corrective action plans, and for recordkeeping. Chapter 7 of this EA provides detailed
discussion on the underlying cost-buildup for each rule component analyzed within the cost
model.
Considering costs over time, Exhibits 9.8 through 9.10 show that most of the cost
difference between the RTCR and Alternative option will be experienced in the first 8 years after
promulgation. Nevertheless, the costs of the RTCR remain lower than those of the Alternative
option throughout the period of analysis. The lower costs of the RTCR are driven by the
monitoring requirements of the respective options (e.g., RTCR allows annual monitoring for
qualifying PWSs while the Alternative option requires at least quarterly monitoring for all
PWSs).
Exhibit 9.11 presents the total and net change in costs for PWSs and states for the 1989
TCR, the RTCR, and the Alternative option. The estimated net change in costs is relatively small
for PWSs and states on a national level; for the RTCR, total costs would increase from the 1989
TCR by approximately 8%, or $14 million (using a 3 percent discount rate). The Alternative
option, based on the increased sampling regimen, increases total costs from baseline by
approximately 16%, or about $30 million (using a 3 percent discount rate). Given these relatively
low net costs, both the RTCR and Alternative option are below the threshold of a "significant
regulatory action" under Executive Order 12866. The approximate $16 million difference
between the RTCR and the Alternative, and the net cost of each option over the 1989 TCR, are
not particularly large in absolute terms compared to a threshold for economic significance of
$100 million annually.
As shown in Exhibit 9.11, the largest portion of the net cost increase is borne by PWSs,
which incur approximately 99% of the RTCR's net annualized present value costs (using a 3
percent discount rate). States are expected to incur the remaining costs. PWS costs represent only
an 8%> and 16% increase, respectively, for the RTCR and Alternative option, in costs over those
estimated for the 1989 TCR. By comparison, state costs are expected to increase by 16% and
34% under the RTCR and Alternative option, respectively; however, state costs remain relatively
low in absolute terms. The higher net change in PWS costs reflects the potential impact on the
subsets of PWSs (primarily small TNCWSs) that would be most affected by either the RTCR or
the Alternative option. Exhibit 9.13 shows that the greatest cost impact is on the smallest subset
Economic Analysis for the Final RTCR
9-19
September 2012
-------
of TNCWSs, for which the net change is more than double under the Alternative option as
compared to the RTCR. The frontloading of costs (shown in Exhibits 9.8 through 9.10) would
further amplify the impacts on small TNCWSs during the first few years of rule implementation.
Comparison of Costs of the Regulatory Options Considered
EPA estimated the net change in costs of the RTCR and Alternative option as compared
to the baseline (1989 TCR).
Exhibit 9.12 presents the total and net change in costs to PWSs and states by rule
component for each of the regulatory options considered. For PWSs, the main driver of
additional costs under the RTCR and Alternative option would be additional corrective actions
and increases in routine monitoring. Offsetting the cost increase are reductions in costs for
additional routine monitoring and PN.
Under the RTCR and Alternative option, the main drivers of net cost increases for states
are reviews of revised sample siting plans. States are expected to see a reduction in costs for PN,
and additionally for assessment activities under the RTCR. States would, however, experience an
overall estimated increase in net costs of $0.2 million and $0.3 million for the RTCR and
Alternative option, respectively. The additional costs associated with the Alternative option
would result from the increased sampling (and reporting) frequency, and the resulting increase in
assessments and corrective actions.
Exhibit 9.13 presents the total and net change in costs to PWSs by system size and type.
For PWSs, the most significant total net cost impact is for the smallest category of TNCWSs
(serving <100 people). Approximately 38% of the net cost increase under the RTCR is estimated
to fall on these systems. This is an important factor in the evaluation of potential impacts when
comparing regulatory options.
Economic Analysis for the Final RTCR
9-20
September 2012
-------
Exhibit 9.8 Net Costs to PWSs (2007$)
h-
o
o
CM
in
c
o
)
o
O
re
c
$60
$50
$40
$30
$20
$10
CR
Alternative Option
9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Time (Years)
Source: RTCR occurrence model output.
Note: X-axis begins at Year 4, which is the first year of implementation of the RTCR or Alternative option. Steady
state begins in approximately Year 9, by which time PWSs that would be expected to meet the criteria would
have qualified for and would begin reduced monitoring.
Economic Analysis for the Final RTCR
9-21
September 2012
-------
Exhibit 9.9 Discounted Net Costs to PWSs Over Time (3% Discount Rate) (2007$)
c
0)
if)
0)
h-
O
o
CN
if)
c
o
(/)
+-»
o
O
$50.00
$45.00
$40.00
$35.00
$30.00
$25.00
$20.00
$15.00
$10.00
$5.00
- RTCR
- Alternative Option
4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Time (Years)
Source: RTCR cost model.
Note: X-axis begins at Year 4, which is the first year of implementation of the RTCR or Alternative option. Steady
state begins in approximately Year 9, by which time PWSs that would be expected to meet the criteria would have
qualified for and would begin reduced monitoring.
Economic Analysis for the Final RTCR
9-22
September 2012
-------
Exhibit 9.10 Discounted Net Costs to PWSs (7% Discount Rate) (2007$)
d>
+-»
CO
Of
<+-»
c
3
o
o
(/>
0)
>
+-»
c
a)
(/>
CD
h-
o
o
CN
U)
c
o
(/)
+-»
if)
o
O
$45
$40
$35
$30
$25
$20
$15
$10
$5
- RTCR
- Alternative Option
4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Time (Years)
Source: RTCR cost model.
Note: X-axis begins at Year 4, which is the first year of implementation of the RTCR or Alternative option. Steady
state begins in approximately Year 9, by which time PWSs that would be expected to meet the criteria would have
qualified for and would begin reduced monitoring.
Economic Analysis for the Final RTCR
9-23
September 2012
-------
Exhibit 9.11 Comparison of Total and Net Change in Annualized Present Value
Costs ($Millions, 2007$)
PWSs
State
Total
PWSs
State
Total
3% Discount Rate
7% Discount Rate
1989 TCR - Total
$ 185
$ 0.9
$ 186
$ 178
$ 0.9
$ 179
RTCR-Total
$ 199
$ 1.1
$ 200
$ 192
$ 1.3
$ 193
RTCR - Net Change
$ 14
$ 0.1
$ 14
$ 14
$ 0.4
$ 14
RTCR - Percent Change
8%
16%
8%
8%
48%
8%
Alternative Option - Total
$ 214
$ 1.2
$ 216
$ 209
$ 1.5
$ 210
Alternative Option - Net Change
$ 29
$ 0.3
$ 30
$ 31
$ 0.6
$ 32
Alternative Option - Percent Change
16%
34%
16%
17%
69%
18%
Note: Detail may not add due to independent rounding. Because only the net change in costs of some rule components are considered as
part of the cost analysis, references to "total" costs in this exhibit do not refer to the complete costs for regulatory implementation, but
only to the specific costs considered to calculate net changes in costs.
Source: Final RTCRcost model.
Economic Analysis for the Final RTCR
9-24
September 2012
-------
Exhibit 9.12 Comparison of Total and Net Change in Annualized Present Value
Costs by Rule Component ($Millions, 2007$)
PWSs | State | Total
PWSs | State | Total
3% Discount Rate
7% Discount Rate
Rule Implementation and Annual Administration
1989 TCR-Total
$
$
$
$
$
$
RTCR - Total
$ 2.77
$ 0.18
$ 2.95
$ 4.00
$ 0.26
$ 4.26
RTCR-Net Change
$ 2.77
$ 0.18
$ 2.95
$ 4.00
$ 0.26
$ 4.26
Alternative Option - Total
$ 2.77
$ 0.18
$ 2.95
$ 4.00
$ 0.26
$ 4.26
Alternative Option - Net Change
$ 2.77
$ 0.18
$ 2.95
$ 4.00
$ 0.26
$ 4.26
Revising Sample Siting Plans
1989 TCR-Total
$
$
$
$
$
$
RTCR - Total
$ 0.59
$ 0.42
$ 1.01
$ 0.84
$ 0.59
$ 1.42
RTCR-Net Change
$ 0.59
$ 0.42
$ 1.01
$ 0.84
$ 0.59
$ 1.42
Alternative Option - Total
$ 0.59
$ 0.42
$ 1.01
$ 0.84
$ 0.59
$ 1.42
Alternative Option - Net Change
$ 0.59
$ 0.42
$ 1.01
$ 0.84
$ 0.59
$ 1.42
Routine Monitoring
1989 TCR-Total
$ 170.59
$
$ 170.59
$ 163.94
$
$ 163.94
RTCR-Total
$ 174.71
$
$ 174.71
$ 167.74
$
$ 167.74
RTCR - Net Change
$ 4.12
$
$ 4.12
$ 3.80
$
$ 3.80
Alternative Option - Total
$ 187.50
$
$ 187.50
$ 182.48
$
$ 182.48
Alternative Option - Net Change
$ 16.91
$
$ 16.91
$ 18.54
$
$ 18.54
Additional Routine Monitoring
1989 TCR-Total
$ 3.87
$
$ 3.87
$ 3.72
$
$ 3.72
RTCR-Total
$ 1.12
$
$ 1.12
$ 1.09
$
$ 1.09
RTCR - Net Change
$ (2.75)
$
$ (2.75)
$ (2.63)
$
$ (2.63)
Alternative Option - Total
$ 0.78
$
$ 0.78
$ 0.66
$
$ 0.66
Alternative Option - Net Change
$ (3.10)
$
$ (3.10)
$ (3.06)
$
$ (3.06)
Repeat Monitoring
1989 TCR-Total
$ 5.11
$
$ 5.11
$ 4.92
$
$ 4.92
RTCR-Total
$ 4.88
$
$ 4.88
$ 4.70
$
$ 4.70
RTCR - Net Change
$ (0.23)
$
$ (0.23)
$ (0.22)
$
$ (0.22)
Alternative Option - Total
$ 5.66
$
$ 5.66
$ 5.59
$
$ 5.59
Alternative Option - Net Change
$ 0.54
$
$ 0.54
$ 0.67
$
$ 0.67
Annual Site Visits
1989 TCR-Total
$
$
$
$
$
$
RTCR-Total
$
$
$
$
$
$
RTCR - Net Change
$
$
$
$
$
$
Alternative Option - Total
$
$
$
$
$
$
Alternative Option - Net Change
$
$
$
$
$
$
Level 1 Assessment
1989 TCR-Total
$ 1.13
$ 0.21
$ 1.34
$ 1.08
$ 0.20
$ 1.29
RTCR-Total
$ 1.63
$ 0.20
$ 1.84
$ 1.57
$ 0.20
$ 1.77
RTCR - Net Change
$ 0.51
$ (0.01)
$ 0.50
$ 0.49
$ (0.01)
$ 0.48
Alternative Option - Total
$ 1.76
$ 0.23
$ 1.99
$ 1.72
$ 0.23
$ 1.94
Alternative Option - Net Change
$ 0.63
$ 0.02
$ 0.65
$ 0.63
$ 0.02
$ 0.65
Level 2 Assessment
1989 TCR-Total
$ 0.70
$ 0.26
$ 0.96
$ 0.68
$ 0.25
$ 0.92
RTCR-Total
$ 0.90
$ 0.19
$ 1.08
$ 0.88
$ 0.18
$ 1.06
RTCR-Net Chanqe
$ 0.20
$ (0.07)
$ 0.12
$ 0.20
$ (0.07)
$ 0.13
Alternative Option - Total
$ 1.26
$ 0.29
$ 1.55
$ 1.30
$ 0.31
$ 1.61
Alternative Option - Net Change
$ 0.55
$ 0.03
$ 0.58
$ 0.62
$ 0.06
$ 0.68
Corrective Actions based on Level 1 Assessments
1989 TCR-Total
$
$
$
$
$
$
RTCR-Total
$ 9.62
$ 0.01
$ 9.63
$ 8.14
$ 0.01
$ 8.15
RTCR-Net Chanqe
$ 9.62
$ 0.01
$ 9.63
$ 8.14
$ 0.01
$ 8.15
Alternative Option - Total
$ 10.01
$ 0.01
$ 10.02
$ 8.52
$ 0.01
$ 8.53
Alternative Option - Net Change
$ 10.01
$ 0.01
$ 10.02
$ 8.52
$ 0.01
$ 8.53
Corrective Actions based on Level 2 Assessments
1989 TCR-Total
$
$
$
$
$
$
RTCR - Total
$ 2.82
$ 0.00
$ 2.82
$ 2.49
$ 0.00
$ 2.49
RTCR-Net Change
$ 2.82
$ 0.00
$ 2.82
$ 2.49
$ 0.00
$ 2.49
Alternative Option - Total
$ 3.78
$ 0.01
$ 3.79
$ 3.57
$ 0.01
$ 3.58
Alternative Option - Net Change
$ 3.78
$ 0.01
$ 3.79
$ 3.57
$ 0.01
$ 3.58
Public Notification
1989 TCR-Total
$ 3.75
$ 0.44
$ 4.19
$ 3.60
$ 0.42
$ 4.02
RTCR - Total
$ 0.26
$ 0.06
$ 0.32
$ 0.25
$ 0.06
$ 0.31
RTCR - Net Change
$ (3.49)
$ (0.38)
$ (3.86)
$ (3.35)
$ (0.36)
$ (3.71)
Alternative Option - Total
$ 0.35
$ 0.08
$ 0.43
$ 0.35
$ 0.08
$ 0.44
Alternative Option - Net Change
$ (3.40)
$ (0.36)
$ (3.76)
$ (3.25)
$ (0.34)
$ (3.58)
Notes:
1) Detail may not add due to independent rounding. Because only the incremental costs of some rule components are considered as
part of the cost analysis, references to "total" costs in this exhibit do not refer to the complete costs for regulatory implementation,
but only to the specific costs considered to calculate net changes in costs.
2) For modeling purposes, additional routine sample counts include regular routine samples taken in the same month.
Source: Rnal RTCRcost model.
Economic Analysis for the Final RTCR
9-25
September 2012
-------
Exhibit 9.13 Total and Net Change in Annualized Costs to PWSs by PWS Size and
Type ($Millions, 2007$)
3% Discount Rate
7%Discount Rate
Alternative
PWS Size
Alternative
Alternative
1989 TCR-
Option -
Alternative
(Population
1989 TCR-Total
RTCR-Total
RTCR-Net
Option - Total
Option - Net
Total
RTCR-Total
RTCR-Net
Total
Option - Net
Served)
A
B
C=B-A
D
E=D-A
F
G
H=G-F
I
J=l-F
Comm unity Water System s (CWSs)
<100
$7.4
$7.5
$0.1
$7.6
$0.2
$7.1
$7.3
$0.2
$7.5
$0.3
101-500
$9.0
$9.4
$0.4
$9.5
$0.5
$8.6
$9.1
$0.5
$9.2
$0.6
501-1,000
$3.7
$3.8
$0.0
$3.8
$0.1
$3.6
$3.7
$0.1
$3.7
$0.1
1,001-4,100
$13.2
$13.6
$0.4
$13.6
$0.4
$12.7
$13.1
$0.4
$13.1
$0.4
4,101-33,000
$42.4
$44.8
$2.4
$44.8
$2.4
$40.7
$42.8
$2.1
$42.8
$2.1
33,001-96,000
$34.9
$36.4
$1.5
$36.4
$1.5
$33.5
$34.8
$1.3
$34.8
$1.3
96,001-500,000
$34.7
$36.2
$1.5
$36.2
$1.5
$33.4
$34.6
$1.2
$34.6
$1.2
500,001-1 Million
$6.5
$6.7
$0.2
$6.7
$0.2
$6.2
$6.4
$0.1
$6.4
$0.1
> 1 Million
$5.6
$5.6
($0.0)
$5.6
($0.0)
$5.3
$5.3
($0.0)
$5.3
($0.0)
Total
$157.4
$163.9
$6.5
$164.1
$6.7
$151.3
$157.2
$5.9
$157.5
$6.2
Nontransient Noncomm unity Water System s (NTNCWSs)
<100
$2.6
$2.7
$0.1
$3.7
$1.1
$2.5
$2.7
$0.2
$3.8
$1.4
101-500
$1.9
$2.0
$0.1
$2.8
$0.9
$1.8
$2.0
$0.2
$2.9
$1.1
501-1,000
$0.6
$0.6
$0.1
$0.9
$0.3
$0.6
$0.6
$0.1
$0.9
$0.3
1,001-4,100
$1.2
$1.3
$0.1
$1.3
$0.1
$1.1
$1.2
$0.1
$1.2
$0.1
4,101-33,000
$0.4
$0.5
$0.1
$0.5
$0.1
$0.4
$0.5
$0.0
$0.5
$0.0
33,001-96,000
$0.1
$0.1
$0.0
$0.1
$0.0
$0.1
$0.1
$0.0
$0.1
$0.0
96,001-500,000
$0.1
$0.1
($0.0)
$0.1
($0.0)
$0.1
$0.1
($0.0)
$0.1
($0.0)
500,001-1 Million
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
> 1 Million
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
Total
$6.9
$7.3
$0.4
$9.3
$2.5
$6.6
$7.2
$0.6
$9.6
$3.0
Transient Noncom m unity Water System s (TNCWSs)
<100
$13.4
$18.7
$5.3
$28.1
$14.7
$12.8
$18.2
$5.3
$28.9
$16.1
101-500
$4.9
$6.5
$1.6
$9.5
$4.7
$4.7
$6.3
$1.6
$9.8
$5.1
501-1,000
$0.6
$0.8
$0.2
$1.2
$0.5
$0.6
$0.8
$0.2
$1.2
$0.6
1,001-4,100
$0.9
$1.0
$0.1
$1.0
$0.1
$0.9
$1.0
$0.1
$1.0
$0.1
4,101-33,000
$0.4
$0.5
$0.1
$0.5
$0.1
$0.4
$0.5
$0.0
$0.5
$0.0
33,001-96,000
$0.1
$0.1
($0.0)
$0.1
($0.0)
$0.1
$0.1
($0.0)
$0.1
($0.0)
96,001-500,000
$0.1
$0.1
($0.0)
$0.1
($0.0)
$0.1
$0.1
($0.0)
$0.1
($0.0)
500,001-1 Million
$0.2
$0.2
($0.0)
$0.2
($0.0)
$0.2
$0.2
($0.0)
$0.2
($0.0)
> 1 Million
$0.3
$0.3
$0.0
$0.3
$0.0
$0.3
$0.3
$0.0
$0.3
$0.0
Total
$20.9
$28.1
$7.3
$41.0
$20.1
$20.1
$27.3
$7.3
$42.0
$21.9
Grand Total
$185.2
$199.3
$14.2
$214.4
$29.3
$177.9
$191.7
$13.8
$209.0
$31.1
Note: Detail may not add due to independent rounding. Because only the incremental costs of some rule components are considered as part of the cost analysis, references to total costs in
this exhibit do not refer to the complete costs for regulatory implementation, but only to the specific costs considered to calculate net changes in costs.
Source: Final RTCRcost model.
9.1.3 Comparison of National Benefits and Costs of the Regulatory Options Considered
Because the potential benefits of the RTCR and Alternative option have not been
monetized, direct questions of how benefits compare to costs cannot be quantitatively answered
(i.e., whether monetary benefits exceed costs and by how much). However, EPA's collective
evaluation of the results presented in this EA lead to the conclusion that the benefits of the
RTCR will, in fact, exceed costs. Further evaluation of the cost effectiveness of the RTCR in
comparison to the Alternative option considered is provided in Section 9.3. The break-even
analysis described in Section 9.3.2 entails determining how many cases of morbidity and/or
mortality would need to be avoided in order for the benefit of those avoided cases to equal the
cost of the rule option. It does not predict cases of illness avoided as a result of the rule. Instead,
it is included to help provide some context for considering the magnitude of cases of waterborne
illness that would need to be avoided to break-even with rule costs.
Economic Analysis for the Final RTCR
9-26
September 2012
-------
9.2 Effect of Uncertainties and Non-quantified Benefit/Cost Estimates on the Estimation
of National Benefits and Costs
Throughout the development of the analyses in this EA, EPA closely evaluated
uncertainties in data used. EPA presents a discussion and comprehensive list of assumptions that
introduce uncertainty into the analysis in Sections 4.5 and 5.3.3.1 (Exhibits 4.13 and 5.22a-b,
respectively). Certain assumptions are incorporated into the predictive model for all three
regulatory options in a similar fashion (1989 TCR as well as the RTCR and Alternative option).
Because the analysis is based on net changes from the 1989 TCR to the RTCR and Alternative
option, incorporating these uncertainties into the 1989 TCR and the options results in no
significant net effect on the analysis results. As discussed in Chapters 4 and 5, although GWR
assumptions are considered by EPA to be conservative estimates of the effectiveness of the
GWR, they are incorporated into the model in similar fashion for all three regulatory options
considered, and therefore are not expected to significantly affect net results of the EA. However,
assumptions for certain variables identified as key drivers in producing the net results of this EA
(the frequency and effectiveness of corrective actions) affect the RTCR and Alternative option
but not the 1989 TCR. These assumptions are further evaluated in a sensitivity analysis in Ch. 5,
which is summarized in Section 9.2.1.
EPA also assessed non-quantified costs and benefits to estimate how they would affect
the overall conclusions of the analyses performed. Potential impacts of non-quantified costs and
benefits are discussed in Chapter 7 and Chapter 6, respectively, of the EA. Section 9.2.2
summarizes the major non-quantified costs and benefits evaluated.
9.2.1 Summary of Major Uncertainties in EA Analyses
The primary sources of uncertainty in the RTCR are introduced into the analysis in the
occurrence model and cost model. In the occurrence model, uncertainty is introduced in two
types of information. First, data on TCIE. coli occurrence and acute and non-acute violations
data are adapted from the Six-Year Review and Safe Drinking Water Information
System/Federal Version. Second, EPA used best professional judgment and stakeholder input in
choosing the values of model inputs to represent the net increase in effective corrective actions to
be implemented under the RTCR, and the resulting period of reduced occurrence (both in length
of time and percent reduction). The data limitations are discussed in Chapter 4 of this EA. The
influence of selected inputs on the model output was tested in a sensitivity analysis, using
substitute values that were factors of 0.5 and 2.0 of the original values. Model runs based on the
0.5 decrease in original inputs yielded results that decreased by a factor less than 0.5 (i.e., the
results were between 0.5 and 1.0 of the original value). Similarly, model runs based on inputs
that were two times the value of the original inputs yielded results that were less than a multiple
of two times the original output. These results, showing that input variable selection has less than
a one-to-one influence on model output, indicate that the model output is not highly sensitive to
changes in these input variables.
Additionally, uncertainty in the analysis results stemmed from the complexity of
interactions between the effects of the various rule components. The occurrence model produced
outputs that showed a decrease in acute violations under the RTCR and an increase under the
Alternative option. Given that the Alternative option incorporated the same risk reducing
Economic Analysis for the Final RTCR
9-27
September 2012
-------
features (assessments and corrective actions) as the RTCR, EPA considered that the additional
sampling under the Alternative option was having a larger upwards effect (increase) on
occurrence compared to the downward effect of the additional corrective actions. EPA developed
a stepwise sensitivity analysis to test the relative importance of the following competing effects
under the RTCR relative to the 1989 TCR:
• New assessments and corresponding corrective actions.
• Reduced repeat and additional routine monitoring.
The results (Section 6.4) revealed that acute violations under the RTCR as compared to
the 1989 TCR differed as follows:
• Many acute violations were prevented from occurring by implementation of
assessments and corresponding corrective actions.
• Some acute violations were missed by the reduction in repeat and additional routine
sampling.
From these results (described in Section 6.4), EPA concluded that under the RTCR, more
acute violations are prevented than are missed. Additionally, the increase in routine monitoring
under the RTCR finds additional acute violations as compared to the 1989 TCR, which explains
the net increase in the acute violations found as compared to the 1989 TCR. Exhibits 9.2-9.7
show the larger net difference (between the alternative and baseline) seen in levels of acute
violations, non-acute violations, and corrective actions implemented under the RTCR as
compared to the Alternative option.
The sources of uncertainty embedded within the results of the occurrence model are also
incorporated into the cost model. As described in reference to the occurrence model, EPA used
best professional judgment and stakeholder input in choosing 10% as the factor of net increase
over corrective actions implemented under the 1989 TCR. A sensitivity analysis of the effect of
an increase in this factor on costs revealed that as the number of corrective actions increases, a
portion of the corresponding cost is offset by the reduced number of assessments required as
water quality improves (Section 7.7 of this EA).
A potentially significant additional source of uncertainty in costs is introduced in the
distribution of corrective actions selected (i.e., the number of each type implemented) for the
10% additional corrective actions. Because there is a wide variation in the unit costs of the actual
corrective actions underlying the compliance forecast (see Appendix D for unit cost detail), EPA
ran two sensitivity analyses to assess the potential impacts of a high and low cost distribution in
the compliance forecast. The results provide a low bound and a high bound on the original
analysis results: approximately $3 million/$25 million for the RTCR and $17 million/$43 million
for the Alternative option (at 3 percent discounting). This indicates that although the estimates of
the number of systems that would choose each type of corrective action has a significant impact
on costs of the rule, the rule remains a relatively low cost rule under the most conservative
(costly) assumptions.
Economic Analysis for the Final RTCR
9-28
September 2012
-------
9.2.2 Summary of Non-quantified Costs and Benefits
EPA believes that all of the rule elements that are the major drivers of the net change in
costs from the 1989 TCR have been quantified to the greatest degree possible. However, cost
reductions related to fewer monitoring and reporting violations are not specifically accounted for
in the cost analysis, and their exclusion from consideration may result in an overestimate of net
change in cost between the 1989 TCR and the RTCR or Alternative option. Using E. coli as the
endpoint, benefits have been estimated and predictions made about the reduced hazard associated
with reduced exposure to fecal contamination and the potential benefits from corrective actions
that, if effective, will reduce exposure. No attempt was made to monetize these benefits by
predicting infections, illnesses or deaths and to compare these values directly with the monetized
costs because of lack of data.
Because the cost analyses in this EA focus on net changes in costs, consideration of non-
quantified costs can result in either an under- or overestimate of the net costs for the RTCR, as
described in Section 7.7 of this EA. Overall, the non-quantified costs identified would potentially
lower the net cost of the RTCR, and include the following:
• Reduced costs due to overlap between sanitary surveys and required assessments;
and
• Reductions in costs related to less frequent PN for reporting and monitoring
violations.
Additionally, EPA identified a potential cost and benefit increase for the RTCR and
Alternative option related to the baseline assumption that all PWSs are currently meeting
requirements of the TCR. To the degree that PWSs are not currently meeting TCR requirements,
the 1989 TCR costs and benefits are overstated in the RTCR EA, resulting in an underestimate of
the net change in costs and benefits under the RTCR.
EPA identified five additional potential areas of non-quantified benefits:
• Reductions in primary and secondary endemic disease burden, both acute and
chronic, associated with consuming PWS drinking water;
• Reductions in risk of outbreak and associated illness and costs;
• Benefits from increased operator knowledge about their systems (including more
consistent water quality);
• Accelerated infrastructure repair/replacement; and
• Reductions in averting behavior (due to higher consumer confidence in water
quality).
Taken together, these items represent a potential for significant additional health benefits.
For the first four items listed, benefits may be realized directly through the elimination,
reduction, and/or prevention of pathogen occurrence. Reductions in averting behavior may lead
Economic Analysis for the Final RTCR
9-29
September 2012
-------
to benefits related to a more efficient use of resources (i.e., using tap water instead of more
expensive, yet not necessarily safer, alternatives).
9.3 Comparison of the Regulatory Options Considered
The following subsections present further analysis of the RTCR in comparison to the
Alternative option. Comparisons based on the net change in costs and benefits and several
effectiveness measures are presented to inform the selection of the RTCR over the Alternative
option.
9.3.1 Net Change in Costs and Benefits
The net change in costs and benefits are those that are incurred or realized to reduce
pathogen illnesses and deaths from one option to the next more stringent option. Estimates of the
net change in costs and benefits are useful when considering the economic efficiency of different
regulatory options considered by EPA. One goal of this analysis is to identify the regulatory
option where net social benefits are maximized. Net benefits analysis is not possible when
benefits are not monetized as in the case with the RTCR. However, this type of analysis can still
provide information on relative cost-effectiveness of different regulatory options.
For the RTCR only costs were monetized. An indirect proxy for benefits was quantified.
Performance of corrective actions is expected to have an impact that is most directly translatable
into potential health benefits. Therefore, to compare the additional cost increases and associated
net benefits of the RTCR and Alternative option, benefits are presented in terms of corrective
actions performed. Specifically, only Level 2 corrective actions are considered because they are
performed in response to a Level 2 treatment technique trigger, which includes either an E. coli
MCL violation or a second Level 1 treatment technique trigger within a 12-month period.
Corrective actions performed in response to Level 2 triggers are, therefore, considered more
indicative of addressing a public health problem than actions addressing a Level 1 trigger.
To compare the additional cost increases and associated benefits (shown as numbers of
Level 2 corrective actions performed) of the RTCR and Alternative option, analyses of both the
net change in costs and benefits are presented in Exhibit 9.14. Note, the RTCR net change in
costs presented in Exhibits 9.14 to 9.17 represents the additional costs between the 1989 TCR
and RTCR, the Alternative option net change in costs represent the additional costs between the
RTCR and Alternative option.
Exhibit 9.14 shows that the annualized net change in costs for the Alternative option
compared to the RTCR are similar to the annualized net change in costs for the RTCR compared
to the 1989 TCR ($15.3 million vs. $14.3 million, using a 3% discount rate). Consistent with the
lower net change in costs, the net change in benefits in terms of Level 2 corrective actions
performed are fewer under the Alternative option than under the RTCR. The new requirement
under both the RTCR and Alternative option for PWSs to implement assessments and corrective
actions will capture the bulk of improvements to be made in comparison to the 1989 TCR. Also,
the additional sampling required under the Alternative option relative to the RTCR provides an
approximately equal increase in costs but a lower net effectiveness relative to the portion of
incurred costs. This suggests that beyond a certain level of corrective action implementation,
Economic Analysis for the Final RTCR
9-30
September 2012
-------
PWSs would experience diminishing net marginal returns for their investment in correction
actions. The relationship between the net change in costs and benefits is examined further with
respect to cost effectiveness in Section 9.3.2.
Exhibit 9.14 Net Change in Annualized Present Value Costs ($Millions, 2007$) and
Benefits (Number of Level 2 Corrective Actions)
Regulatory Option
Costs
Benefits
(L2 Corrective Actions)
3%
7%
3%
7%
1989 TCR
$ 186.1
$ 178.8
No change3
No change3
RTCR
$ 200.4
$ 193.0
208
202
RTCR- Net Change1
$ 14.3
$ 14.2
208
202
Alternative Option
$ 215.7
$ 210.5
336
355
Alternative Option - Net Change2
$ 15.3
$ 17.5
128
153
Represents the net change of the RTCR over the 1989 TCR.
Represents the net change of the Alternative option over the RTCR. Add net change for Alternative option to net change for RTCR to calculate the total net
change of the Alternative option over the 1989 TCR. Note: The RTCR occurrence model yields the number of corrective actions that are expected to be
implemented in addition to (net of) those already implemented under the 1989 TCR. The model does not incorporate an estimate of the number of corrective
actions implemented per year under the 1989 TCR and does not yield a total for the RTCR and Alternative option that includes the 1989 TCR corrective actions.
Benefits shown include corrective actions based on L2 assessments. Detailed benefits and cost information is provided in Appendices Aand C, respectively,
of the Final RTCR EA(USEPA2009a).
3As explained in section VI.F.2.f of the preamble, for modeling purposes, EPA estimates the net change only in the number of corrective actions performed
under the RTCR and Alternative options compared to the 1989 TCR and thus did not quantify the (non-zero) baseline number of corrective actions performed
underthe 1989 TCR.
9.3.2 Cost Effectiveness Measures
To further differentiate between the RTCR and Alternative option results, EPA
performed analyses to assess the cost effectiveness of each regulatory option. In this section,
EPA provides an estimate of the cost per corrective action performed under the regulatory
options, which is a measure of the cost effectiveness of each rule option. As noted elsewhere in
this chapter, corrective actions are considered a clear enhancement to the efforts to reduce E. coli
(and co-occurring pathogen) occurrence in PWSs, and the most direct benefit in this EA.
Corrective actions are also a main driver of the net increases in costs for both rule options
considered. In addition, the modeling of corrective actions is done only on a net basis relative to
the level performed under the 1989 TCR, thus providing a clear dividing line between impacts
under the 1989 TCR and each rule option. As shown in Exhibit 9.15, the annual net cost of the
rule per corrective action performed beyond the baseline is approximately $0.02 million and
$0.04 million (3% discount rate) for the RTCR and Alternative option, respectively.
Economic Analysis for the Final RTCR
9-31
September 2012
-------
Exhibit 9.15 Total Net Annual Cost Per Corrective Action Implemented under
RTCR and Alternative Option, Annualized Using 3% and 7% Discount Rates
($Millions, 2007$)
Regulatory Option
3% Discount Rate
7% Discount Rate
RTCR - Net Change
$ 14.3
$ 14.2
RTCR - Incremental Number of Corrective Actions (L1 & L2)
616
594
RTCR - Cost Effectiveness Analysis1
$ 0.02
$ 0.02
Alternative Option - Net Change
$ 29.6
$ 31.7
Alternative Option - Incremental Number of Corrective Actions (L1 & L2)
808
819
Alternative Option - Cost Effectiveness Analysis1
$ 0.04
$ 0.04
Notes:
1CEA= (Net Change)/(lncremental Number of corrective actions).
2Detaiied cost information is provided in Appendix C.
EPA also considered the net cost effectiveness of the RTCR for the entire population of
PWSs taken as a whole as compared to the Alternative option to determine the additional benefit
associated with that portion of Alternative option cost beyond the cost of the RTCR. Exhibit 9.16
shows that the RTCR has a far lower marginal cost per corrective action for PWSs than the
Alternative option ($0.02 million for the RTCR vs. $0.08 million for the Alternative option).
This indicates that the additional corrective actions performed under the Alternative option are
much more costly on a unit basis than those performed under the RTCR.
EPA further considered the issue of additional costs under the Alternative option as
compared to the RTCR in terms of the effect on small TNCWSs using GW (serving <100). The
additional net change in costs under the Alternative option could cause substantial additional
burden on small TNCWSs, which incur the largest portion of costs of the rule among all
categories of systems. However, these same TNCWSs would also receive the most benefits from
the RTCR (increased assessments and corrective actions following a trigger). EPA considered
the group of 60,200 TNCWSs using GW (serving <100), which are the largest subset of systems
by size/type, and which are expected to bear the highest burden of all system categories under
the RTCR. Exhibit 9.17 shows that the net cost effectiveness of the RTCR is significantly greater
than of the Alternative option (a cost per corrective action of $0.02 million vs. $0.07 million,
respectively) and both options are more cost effective for this subset (TNCWSs using GW
serving <100) than for the entire population of PWSs taken as a whole. The two net change cost
analyses indicate that, using a 3 percent discount rate to compare incremental benefits and costs,
the RTCR is significantly more cost effective (by a factor of approximately 3.9) than the
Alternative option for the most burdened subset of systems, and is significantly more cost
effective (by a factor of approximately 3.4) than the Alternative option when considering all
PWSs together.
Economic Analysis for the Final RTCR
9-32
September 2012
-------
Exhibit 9.16 Annualized Net Change in Costs per Corrective Action Implemented
for All PWSs under RTCR and Alternative Option ($Millions, 2007$)
Regulatory Option
3% Discount Rate
7% Discount Rate
RTCR - Net Change
$ 14.30
$ 14.17
RTCR - Incremental Number of Corrective Actions (L1 & L2)1
615.82
593.71
RTCR - Incremental Cost per Corrective Action ($)
$ 0.02
$ 0.02
Alternative Option - Net Change (over RTCR)2
$ 15.29
$ 17.52
Alternative Option - Incremental Number of Corrective Actions (L1 & L2) (over RTCR)2
192.23
224.93
Alternative Option - Incremental Cost per Corrective Action ($)
$ 0.08
$ 0.08
Notes:
1Exhibit includes the number of corrective actions predicted by the RTCR occurrence model to be implemented in addition to those
implemented underthe 1989 TCR.
2Add net values for Alternative option to net values for RTCR to calculate total net values of Alternative option over 1989 TCR.
3Detailed cost information is provided in AppendixC.
Exhibit 9.17 Annualized Net Change in Costs per CA Implemented for TNCWSs
(Serving <100 people) under RTCR and Alternative Option ($Millions, 2007$)
Regulatory Option
3% Discount Rate
7% Discount Rate
RTCR - Net Change
CO
LO
CO
LO
RTCR - Incremental Corrective Actions (L1 & L2)1
287
275
RTCR - Incremental Cost per Corrective Action ($)
$ 0.02
$ 0.02
Alternative Option - Net Change (over RTCR)2
CD
$ 10.8
Alternative Option - Incremental Corrective Actions (L1 & L2) (over RTCR)2
132
155
Alternative Option - Incremental Cost per Corrective Action ($)
$ 0.07
$ 0.07
Notes:
1Derived by dividing incremental rule costs applicable to TNCWS <100 bythe incremental numberof corrective actions to be implemented
under the RTCR (relative to baseline) and the Alternative option (relative to the RTCR) by TNCWSs using GW.
2Add net values for Alternative option to net values for RTCR to calculate total net values of Alternative option over 1989 TCR.
3Detailed cost information is provided in AppendixC.
9.3.3 Break-Even Analysis
As described in Chapter 6 of this EA, this analysis does not include a fully quantified risk
assessment because there are insufficient data on the co-occurrence in PWSs of fecal indicator E.
coli and pathogenic strains of E. coli or other bacteria, viruses, and parasitic protozoa that can
cause waterborne disease. Given the absence of E. coli and waterborne pathogen co-occurrence
data to predict health effects in PWSs, EPA developed a break-even analysis to inform the
discussion of whether the benefits justify the costs of the regulation. A break-even analysis
typically compares the rule cost with the estimated value of the medical cases (and related costs
such as time off from work) that would be avoided annually by the rule to determine if the rule
benefits break even with avoided costs. In the case of the EA, where cases avoided cannot be
estimated, the break-even analysis informs the benefit/cost discussion by providing estimates of
the number of fatal and non-fatal cases that the rule would need to avoid in order to break even
with rule costs.
The underlying question this analysis seeks to inform is whether corrective actions
performed under the RTCR will sufficiently reduce potential exposure to fecal contamination
and/or waterborne pathogens to compensate for the net costs of the rule. One result of the
Economic Analysis for the Final RTCR
9-33
September 2012
-------
implementation of the RTCR predicted by the model and presented in Exhibit 6.5 is that some
number of acute events will be prevented.64 As discussed in Chapter 6, sufficient data are not
available to quantify the numbers of illnesses and deaths potentially avoided by preventing such
acute events. However, the presence of E. coli in a water system is an indicator of the presence
of fecal contamination;65 it suggests that a pathway for the introduction of waterborne pathogens
may exist. The total numbers of acute events and potentially affected people are small (most
events are predicted in small NCWSs), but nonetheless there will be a positive impact on public
health by avoiding these events. The magnitude of this impact will depend heavily on the
specific elements underlying any acute event avoided: who and how many in the population are
affected, and the etiologic agent to which exposure is avoided. An event can be endemic to a
single person or small group of persons, or it may reach the level of an outbreak. The cause of an
event may be one or more of a range of bacterial, viral, or parasitic organisms, as described in
Chapter 2 of this EA. Although there may be a range of waterborne pathogens mitigated by
implementation of the RTCR, two potential waterborne pathogens are considered in this analysis
to provide input for developing the break-even analysis: shiga toxin-producing E. coli 0157:H7
(STEC 0157)66 and Salmonella 61
The Centers for Disease Control and Prevention (CDC) estimates that STEC 0157
pathogens from all sources are responsible for approximately 73,000 illnesses in the United
States annually. The CDC has found that 15% of outbreak cases of STEC 0157 are waterborne
(Rangel et al., 2005); if that rate applies as well to endemic cases, then approximately 11,000
cases of STEC 0157 infection would be due to waterborne exposure. For Salmonella infections,
the Economic Research Service (ERS) uses an estimate of 1,397,187 cases annually based on
1999 estimates from the CDC that included an estimate of non-reported illnesses. The 1999
estimate stated that the share of cases due to consumption of contaminated food was 95 percent,
indicating that approximately 70,000 cases were due to other sources. While the portion of these
cases due to waterborne exposure is unknown, some public systems have documented cases of
waterborne Salmonellosis. In March of 2008, Salmonella in the water system of Alamosa,
Colorado caused an outbreak in which hundreds of the 10,000 residents were sickened and one
person died. The deep mountain aquifer from which the water was drawn was confirmed not to
be the source of the bacteria (Berg, 2008), but two storage tanks had several inches of sediment
64Results presented in Exhibit 6.5 of this EA indicate that fewer E. coli positive events would occur under the RTCR
than under the 1989 TCR. These results are consistent with the net improvement in PWS water quality (as indicated
by a decrease in fecal indicator E. coli) suggested by the stepwise analysis presented in Section 6.4 of this EA.
Chapter 6 of the EA presents this information in further detail.
65 Standridge, J. (2008) discusses the use of E. coli as an indicator of drinking water quality.
66 According to the website of the American Academy of Family Physicians
(http://www.aafb.org/afp/200004Q 1/tips/l 1 .html'). "Shiga toxin-producing Escherichia coli is a group of bacteria
strains capable of causing significant human disease. The pathogen is transmitted primarily by food and has become
an important pathogen in industrialized North America. The subgroup enterohemorrhagic E. coli includes the
relatively important serotype 0157:H7, and more than 100 other non-0157 strains."
67 Salmonella, according to the CDC website, is actually a group of bacteria that can cause diarrheal illness in
humans. They are microscopic living creatures that pass from the feces of people or animals to other people or other
animals. Salmonella serotype Typhimurium, which caused the 2008 Alamosa, CO outbreak, and Salmonella
serotype Enteritidis are the most common in the United States. Every year, approximately 40,000 cases of
salmonellosis are reported in the United States; most of these are foodborne, but some, as in the case of Alamosa,
are waterborne. (http://www.cdc.gov/nczved/dfbmd/disease listing/salmonellosis gi.html#2. downloaded 10-6-09)
Economic Analysis for the Final RTCR
9-34
September 2012
-------
and breaches, and one tank had breaches large enough for birds and animals to enter (CDPHE,
2009).
For the break-even analysis, medical cases are classified as non-fatal or fatal, and
analyses are presented for these two classifications.68 The valuations for medical cases used in
the break-even analysis are developed using the ERS Foodborne Illness Calculator69 (ERS
Foodborne Illness Calculator), as described in Appendix E of the RTCR, for the two example
pathogens considered: STEC 0157 and Salmonella. For most ERS calculator inputs, default or
suggested alternative values are used to estimate costs. However, for the valuation of fatal cases
70 71
within the model, EPA used a value of a statistical life (VSL) estimate of $8.8 million to be
72
consistent with EPA methodology for calculating the value of this measure.
To represent a range of estimates, EPA used the ERS calculator to calculate break-even
thresholds for a low- and high-end valuation of cases avoided corresponding to "traditional" and
"enhanced" methodologies for estimating Cost of Illness (COI) values. For the traditional COI
valuation, EPA used the ERS calculator default values for STEC 0157, with the exception of the
VSL applied to fatalities, which EPA updated as shown in Appendix E of the RTCR EA. To
estimate an enhanced COI valuation, EPA applied the basic settings used for the traditional COI
valuation (including the EPA-updated VSL value) but with alternative values for estimated
productivity and disutility losses. The traditional COI, as developed in this analysis based on the
default setting in the ERS calculator, places no value on lost nonmarket work time, while the
enhanced COI method values nonmarket work time based on opportunity costs (equivalent to
market work time lost). The alternative valuation used (a suggested option within the calculator)
is to apply the productivity losses to all lost time for market and non-market work (e.g., for
individuals who stay home to raise children or care for the elderly). The ERS calculator default
68 Evaluation of fatal cases in this analysis includes consideration of the underlying non-fatal component of the
illness prior to death.
69 A full detailed description of the ERS Foodborne Illness Calculator, including detailed discussion of the items
discussed previously and derivation of the underlying values used in calculations can be found online at:
http://www.ers.usda.gov/Data/FoodborneIllness/ecoli_Intro.asp.
70 The VSL represents an estimate of the monetary value of reducing risks of premature death. The VSL, therefore,
is not an estimate of the value of saving a particular individual's life. The value of a "statistical" life represents the
sum of the values placed on small individual risk reductions across an exposed population. For example, if a
regulation were to reduce the risk of premature death by 1/1,000,000 for one million exposed individuals, the
regulation would "save" one statistical life (1,000,000 X 1/1,000,000). If each of the 1,000,000 people were willing
to pay $5 to achieve the risk reduction anticipated from the regulation, the VSL would be $5 million ($5 X
1,000,000). Appendix E of this EA provides further detail of the derivation of the updated VSL estimate.
71 An EPA study characterized the range of possible VSL values as a Weibull distribution with a mean of $4.8
million (1990 price level) based on 26 individual study estimates (USEPA, 1997). This represents the value
recommended for use in benefits analyses in EPA's Guidelines for Preparing Economic Analyses (USEPA, 2000)
and endorsed by the Science Advisory Board (SAB) Arsenic review panel (USEPA, 2001a). For use in the break-
even analysis, adjustments to the base value from the EPA study were made to account for income and price level
changes over time. The mean VSL, after all adjustments were made, was $8.8 million in year 2007 dollars. These
adjustments are explained further in Appendix E of this EA.
72 Use of different VSL estimates corresponding to different government agency policy and methodologies is
specifically suggested by ERS, and EPA's methodology is given as a specific example of an alternate value that may
be used.
Economic Analysis for the Final RTCR
9-35
September 2012
-------
73
value is zero for other forms of disutility caused by the illness, including pain and suffering,
inconvenience, and lost non-work time; this setting is used for the traditional approach in this
analysis. The alternative valuation, used for the enhanced approach in this analysis (also a
suggested option within the calculator), is based on values (greater than zero) used by the Food
and Drug Administration.
Exhibits 9.18 and 9.19 show the values calculated for avoided cases using the ERS
Foodborne Illness Calculator for STEC 0157 and Salmonella, as summarized previously in this
section and described in Appendix E of this EA. In addition to separate calculations based on
traditional and enhanced COI valuation methodologies, estimates are presented separately for
non-fatal and fatal cases.
Using the valuations per case as described, EPA estimated the numbers of cases needed
to break even by dividing the total net cost of the RTCR by the values from Exhibits 9.18 and
9.19. The resulting estimates of break-even thresholds are presented in Exhibit 9.20 and 9.21
separately for non-fatal and fatal cases for STEC 0157 and Salmonella, respectively. Estimates
are presented using both three and seven percent discount rates.
Exhibit 9.18 Average Estimated Value per STEC 0157 Case Avoided (2007$)
Case Description
Value Per Case
Traditional COI
Enhanced COI
Non-fetal cases only
$ 1,756
$ 5,269
Fatal cases only1
$ 8,820,401
$ 8,821,911
1 Calculations for fatal cases include the COI component for the
underlying illness prior to death.
Notes: COI estimates derived using USDA ERS Foodborne Illness
Cost Calculator: STEC 0157:H7. Traditional COI estimates are based
on ERS default values for all inputs except for the VSL estimate, for
which EPA substituted its own estimate of approximately $8.8 million.
Enhanced COI estimates are based on the same values except for the
following changes: productivity losses are applied to all cases rather
than just those employed in the formal market, and FDA disutility
values are applied rather than the ERS default value of 0.
73 Disutility is a specific category defined in the ERS calculator, as mentioned previously, to include: pain and
suffering, inconvenience, and lost non-work time.
Economic Analysis for the Final RTCR
September 2012
9-36
-------
Exhibit 9.19 Average Estimated Value per Salmonella Case Avoided (2007$)
Case Description
Value Per Case
Traditional COI
Enhanced COI
Non-fetal cases only
$ 210
$ 1,397
Fatal cases only1
$ 8,802,998
$ 8,803,551
1 Calculations for fatal cases include the COI component for the underlying
illness prior to death.
Notes: COI estimates derived using USDA ERS Foodborne Illness Cost
Calculator: Salmonella. Traditional COI estimates are based on ERS
default values for all inputs except for the VSL estimate, for which EPA
substituted its own estimate of approximately $8.8 million. Enhanced COI
estimates are based on the same values except for the following changes:
productivity losses are applied to all cases rather than just those employed
in the formal market, and FDA disutility values are applied rather than the
ERS default value of 0.
Exhibit 9.20 Estimated Annual Break-Even Threshold for Avoided Cases of
STEC 0157
COI
Methodology
Discount
Rate
RTCR
Alternative Option
Non-fatal
cases only
Fatal
cases
only1
Non-fatal
cases only
Fatal
cases
only1
Traditional
COI
3%
8,000
1.6
17,000
3.4
7%
8,000
1.6
18,000
3.6
Enhanced COI
3%
3,000
1.6
6,000
3.4
7%
3,000
1.6
6,000
3.6
1 Calculations for fetal cases include the non-fatal COI component for the
underlying illness prior to death.
Notes:
1) The number of cases needed to reach breakeven threshold calculated by
dividing the net change in costs for the RTCR (Exhibit 9.11) by the average
estimated value of avoided cases (Exhibit 9.18). E. coli 0157:H7 illness is
only an example of a pathogenic endpoint that could be used for this analysis.
Use of additional pathogenic contaminants in addition to this single endpoint
would result in lower threshold values.
2) Detail may not add due to independent rounding.
3) The breakeven threshold is higher using a 7% discount rate than a 3%
discount rate under the Alternative option. This result is consistent with the
costs of the Alternative option being higher using the 7% discount rate, which
is caused by the frontloading of costs in the period of analysis, as explained
further in Chapter 7 of the EA.
Economic Analysis for the Final RTCR
9-37
September 2012
-------
Exhibit 9.21 Estimated Annual Break-Even Threshold for Avoided Cases of
Salmonella
COI
Methodology
Discount
Rate
RTCR
Alternative Option
Non-fatal
cases only
Fatal
cases
only1
Non-fatal
cases only
Fatal
cases
only1
Traditional
COI
3%
68,000
1.6
141,000
3.4
7%
68,000
1.6
151,000
3.6
Enhanced COI
3%
10,000
1.6
21,000
3.4
7%
10,000
1.6
23,000
3.6
1 Calculations for fatal cases include the non-fatal COI component for the
underlying illness prior to death.
Notes:
1) The number of cases needed to reach breakeven threshold calculated by
dividing the net change in costs for the RTCR (Exhibit 9.11) by the average
estimated value of avoided cases (Exhibit 9.19). Salmonella is only is only an
example of a pathogenic endpoint that could be used for this analysis. Use of
additional pathogenic contaminants in addition to this single endpoint would
result in lower threshold values.
2) Detail may not add due to independent rounding.
3) The breakeven threshold is higher using a 7% discount rate than a 3%
discount rate under the Alternative option. This result is consistent with the
costs of the Alternative option being higher using the 7% discount rate, which
is caused by the frontloading of costs in the period of analysis, as explained
further in Chapter 7 of the EA.
The estimates shown in Exhibits 9.20 and 9.21 for STEC 0157 and Salmonella,
respectively, represent the number of fatal cases that would need to be avoided annually by the
rule for its benefits and costs to society to be equal, or break even, with rule costs. Based on
either example pathogen, the results show that both the RTCR and Alternative option would
need to avoid a relatively small number of fatal cases annually to break even with the rule costs.
Under the RTCR, approximately two deaths would need to be avoided annually using a
3% discount rate based on consideration of the bacterial pathogen STEC 0157. Alternatively,
approximately 3,000 or 8,000 non-fatal cases, using the enhanced or traditional benefits
valuations approaches,74 respectively, would need to be avoided to break even with the RTCR
costs. As expected based on its costs, the Alternative option would require that a higher number
of cases be avoided annually for that option to break even (between 100% and 113% more than
the RTCR under the enhanced and traditional approaches, respectively).
74 Both traditional and enhanced COI approaches count the value of the direct medical costs and of time lost that
would been spent working for a wage, but differ in their assessment of the value of time lost that would be spent in
nonmarket work (e.g. housework, yardwork, and raising children) and leisure (e.g. recreation, family time, and
sleep). They also differ in their valuation of (other) disutility, which encompasses a range of factors of well being,
including both inconvenience and any pain and suffering. A complete discussion of the traditional and enhanced
COI approaches can be found in Appendix E.
Economic Analysis for the Final RTCR
9-38
September 2012
-------
As Exhibit 9.21 shows, approximately two deaths would need to be avoided annually
from a Salmonella infection for the RTCR to break even. The estimated number of non-fatal
Salmonella cases that would need to be avoided to break even under RTCR is approximately
10,000 or 68,000 cases under the enhanced and traditional benefits valuations approaches,
respectively. As expected based on its costs, the Alternative option would require that a higher
number of Salmonella cases be avoided annually for that option to break even (approximately
110% more than the RTCR under either the enhanced or traditional approaches).
As the discussion presented in Chapter 2 of this EA describes, disease and deaths are
attributable to drinking water contamination across the country. The CDC has found that 15% of
outbreak cases of STEC 0157 are waterborne (Rangel et al., 2005); if that rate applies as well to
endemic cases, then approximately 11,000 cases would be due to waterborne exposure to STEC
0157. The revised rule (RTCR) could mitigate pathways of contamination into PWSs and
prevent some of these endemic cases. For only this one contaminant, if just 3,000 non-fatal cases
under the enhanced COI approach, 8,000 non-fatal cases under the traditional COI approach, or
approximately two fatal cases of STEC 0157 are avoided annually by the RTCR, the rule would
break even. If more than one contaminant was reduced or prevented from occurring in PWSs, the
rule would be that much more likely to break even. Avoided illness and death from secondary
transmission of infection could also be significant, and would increase the likelihood that total
avoided cases reach a break-even threshold. An additional consideration is that a larger number
of avoided cases would be predicted if all the benefits of the rule as described in Chapter 6 could
be quantified in the analysis. If increased assessments and corrective actions result in a level of
system knowledge that enables earlier mitigation of potential pathways of contamination, then
additional illnesses or deaths may be avoided beyond those suggested by the decrease in acute
events that the model predicts. In the case of the outbreak of Salmonella infections in Alamosa,
Colorado, as described previously in this section of the EA, some of the key factors that
contributed to the outbreak include significant levels of sediment and the presence of breaches to
the integrity of the storage tank. These are two types of sanitary defects that could be identified
in an assessment and eliminated by corrective action.
9.3.4 Summary of Conclusions
The preferred regulatory option for this rulemaking is the RTCR. The analyses performed
as part of this EA support the collective judgment and consensus of the advisory committee that
their AIP recommendations, as defined by the RTCR, provide for effective and efficient
revisions to 1989 TCR regulatory requirements. The following is a summary of points that must
be considered when weighing the benefits and costs of the rule options:
• The estimated net cost of the RTCR is not only small relative to the 1989 TCR
(approximately $14 million annually using either a 3% or 7% discount rate), but it
is also small compared to the net cost increase of the Alternative option relative to
the 1989 TCR (approximately $30-$32 million using a 3% and 7% discount rate,
respectively) (exhibit 9.11). This cost differential is especially important
considering the potential concentration of impacts on the smallest TNCWSs
(exhibit 9.13) and the potential frontloading of costs under the Alternative option
(exhibits 9.8-9.10).
Economic Analysis for the Final RTCR
9-39
September 2012
-------
• The benefits are expected to be positive under the RTCR and no backsliding in
overall risk is predicted (Section 9.1.1.1) from the 1989 TCR. The total benefits
under the Alternative option would be greater than under the RTCR (Section
9.1.1.2), but the RTCR is more cost effective both on an average basis and based on
the net change in costs (Exhibits 9.15-9.17).
- Of the various outputs of the occurrence and cost models developed in this
EA, the estimate of corrective actions to be implemented has the clearest
connection to potential benefits. Based on this metric, the RTCR is more
cost effective on average (i.e., costs less per corrective action achieved)
than the Alternative option (Exhibit 9.15) by a factor of approximately 1.6.
- EPA conducted a net cost effectiveness analysis based on the increasing
stringency of the options from the 1989 TCR to the RTCR to the
Alternative option. In comparing the ratio of additional costs and benefits
associated with the RTCR compared to the 1989 TCR and of the
Alternative option compared to the RTCR, EPA found the following
results:
¦ Exhibit 9.16 shows that the net cost effectiveness of the RTCR far
exceeds that of the Alternative option for the entire population of
PWSs when taken as a whole, by a factor of approximately 3.4.
¦ Exhibit 9.17 shows that for TNCWSs the cost effectiveness of the
RTCR far exceeds that of the Alternative option, by a factor of
approximately 3.9.
• Although data limitations precluded estimating the number of fatal or non-fatal
illnesses that would be avoided by implementation of the regulatory options
considered, the break-even analysis (Section 9.3.3) shows that approximately two
deaths would need to be avoided annually for the final rule to break even based on
estimated net costs (exhibit 9.20) using a three percent discount rate. Considering
the average costs of non-fatal illnesses resulting from STEC 0157 infection, the
RTCR would break even if it avoided 3,000 or 8,000 such cases annually based on
the enhanced or traditional cost of illness approach, respectively. These thresholds
under the Alternative option are 100% to 113% more than under the RTCR using
the enhanced and traditional approaches. Under the RTCR, the estimated break-
even threshold for non-fatal Salmonella cases is 10,000 or 68,000 annually for the
enhanced and traditional approaches, respectively, and approximately 110% more
(using either the enhanced or traditional approach) for the Alternative option. These
break-even thresholds would be reduced if more than one type of pathogen is
avoided through implementation of the RTCR, or if cases of secondary transmission
are avoided, both of which are logical extensions of the benefits. Given the large
variety of pathogens that may occur in PWS source water and the relatively low net
costs of the RTCR, EPA believes the RTCR is likely to at least break even.
Economic Analysis for the Final RTCR
9-40
September 2012
-------
• There are several items that are not quantified in the EA that represent a potential
for significant additional benefits (Section 9.2.2). These additional benefits may be
realized due to reductions in co-occurring contaminants, increased systems
knowledge (including more consistent water quality), reductions in outbreak risks,
accelerated infrastructure repair/replacement, and reductions in averting behavior
(due to higher consumer confidence in water quality).
As a result of these considerations, EPA believes that the estimated benefits of the RTCR
as defined by the RTCR will likely exceed the modest increases in annualized national costs and
will be effective in reducing risk to consumers from exposure to fecal contamination in drinking
water.
Economic Analysis for the Final RTCR
9-41
September 2012
-------
10 References
Altekurse, S. F., N.J. Stern, P.I. Fields, andD.L. Swerdlow. 1999. Camplybacter jejuni—An
emerging foodborne pathogen. Emerging Infectious Diseases. 5(10): 1-10.
American Academy of Family Physicians. 2000. Foodborne Disease and Shiga Toxin-Producing
E. coli. April 1. Available online at: http://www.aafp.org/afp/2000Q401/tips/ll.html.
Angulo, F.J., S. Tippen, D. Sharp, B.J. Payne, C. Collier, J. Hill, T.J. Barrett, R.M. Clark, E.
Geldreich, H.D. Donnell, and D.L. Swerdlow. 1997. A community waterborne outbreak of
salmonellosis and the effectiveness of a boil water order. American Journal of Public Health.
87(4):580-584.
Association of State Drinking Water Administrators (ASDWA). 2001. Drinking water program
resource needs assessment. Version 9. November 27, 2001.
Bennear, L., K. Jessoe, and S. Olmstead. 2009. Sampling Out: Regulatory Avoidance and the
Total Coliform Rule. Environmental Science and Technology. 43:5176-5182.
Bennett, J.V., S.D. Holmberg, M.F. Rogers, S.L. Solomon, 1987. Infectious and parasitic
diseases. In: Amler, Closing the Gap: The Burden of Unnecessary Illness, R.W. and H.B. Dull,
(eds.). p. 102-114. New York, NY: Oxford Press.
Berg, Rebecca. 2008. The Alamosa Salmonella Outbreak: A Gumshoe Investigation. Journal of
Environmental Health. 71(2): 54-5 5.
Boring, J.R., W.T. Martin, and L.M. Elliott. 1971. Isolation of Salmonella typhimurium from
municipal water, Riverside, California, 1965. American Journal of Epidemiology. 93:49-54.
Carducci A., L. Caniania, R. Moscatelli, B. Casini, E. Rovini, F. Mazzoni, A. Giuninti, and M.
Verani.2002. Interference between enterovirus and reovirus as a limiting factor in environmental
virus detection. Letters in Applied Microbiology. 34:110-113.
Carter, M.J. 2005. A review: Enterically infecting viruses: pathogenicity, transmission and
significance for food and waterborne infection. Journal of Applied Microbiology. 98:1354-1380.
Casemore, DP 1990. Epidemiologic aspects of human cryptosporidiosis. Epidemiology and
Infection 104:1-28.
Centers for Disease Control (CDC). 2001a. Norwalk-like viruses, public health consequences
and outbreak management. Morbidity and Mortality Weekly Report. 50(RR-9):18.
Centers for Disease Control (CDC). 2001b. Two fatal cases of adenovirus-related illness in
previously healthy young adults—Illinois, 2000. Morbidity and Mortality Weekly Report.
50(26):553-555.
Centers for Disease Control (CDC). 2002. HIV/AIDS Surveillance Report: U.S. HIV and AIDS
cases reported through December 2001. Vol. 13. No. 2.
Economic Analysis for the Final RTCR
10-1
September 2012
-------
Centers for Disease Control (CDC). 2011. Patient Facts: Learn More about Legionnaires'
disease. June 1. Available online at: http://www.cdc.gov/legionella/patient facts.htm.
Centers for Disease Control and Prevention Division of Foodborne Bacterial and Mycotic
Diseases (DFBMD). Salmonellosis. 2008. Available online at:
http://www.cdc.gov/nczved/dfbmd/disease listing/salmonellosis gi.html#2; downloaded 10-6-
09.
Clark, W. F., J.J Macnab, and J.M. Santrop. 2008. The Walkerton Health Study, 2002-2008:
Final Report to the Ontario Ministry of Health and Long-Term Care. London Health Sciences
Center. University of Western Ontario. Available online at:
www.lhrionhealth.ca/feature/Walkerton.pdf.
Clesceri, L.S., A.E. Greenberg, and A.D.Eaton, (editors). 1998. Standard Methods for the
Examination of Water and Wastewater. Washington, D.C.: American Public Health Association
Publications. 20th Edition.
Colorado Department of Public Health and Environment (CDPHE). 2009. Waterborne
Salmonella Outbreak in Alamosa, Colorado March and April 2008: Outbreak identification,
response, and investigation. Available online at:
http://www.cdphe.state.co.us/wq/DrinkingWater/pdf/AlamosaInvestRpt.pdf.
Cooley, M., D. Carychao, L. Crawford-Miksza, M.T. Jay, C. Myers, C. Rose, C. Keys, J. Farrar,
and R.E. Mandrell. 2007. Incidence and tracking of Escherichia coli 0157:H7 in a major produce
production region in California. PLoS ONE. November 2007. 14;2(1 l):el 159.
Craun, G.F. 1996. "Cases of illness required to initiate a waterborne outbreak investigation."
Unpublished report to SAIC.
Craun, G.F., R.L. Calderon and T.J. Wade. 2006. Assessing waterborne risks: an introduction.
Journal of Water and Health 4(2):3-18.
Eaton, A.D., A.E. Greenberg, E.W. Rice, and L.S. Clesceri (editors). 2005. Standard Methods
for the Examination of Water and Wastewater. Washington, D.C.: American Public Health
Association Publications. 21st Edition.
Edberg, R. 2000. E. coli, the best biological drinking water indicator for public health protection.
Journal of Applied Microbiology. 88: 106S-116S.
ERS Foodborne Illness Cost Calculator: STEC 0157. Available online at:
http ://www. ers .usda. gov/Data/Foodbornelllness/ecoli Intro, asp .Feng. P. 1995. Escherichia coli
serotype 0157:H7: Novel vehicles of infection and emergence of phenotypic variants. Emerging
Infectious Diseases l(2):47-52.
Framm, S.R. and R. Soave. 1997. Agents of diarrhea. The Medical clinics of North America.
81(2):427-447.
Economic Analysis for the Final RTCR
10-2
September 2012
-------
Garg, A.X., R.S. Suri, N. Barrowman, F. Rehman, D. Matsell, M. Patricia Rosas-Arellano, M.
Salvadori,R.B. Haynes and W.F. Clark. 2003. Long-term renal prognosis of diarrhea-associated
hemolyticuremic syndrome: A systematic review, meta-analysis and meta-regression. Journal of
American Medical Association. 290(10): 1360-1370.
Garg, A. X., L. Moist, D. Matsell, H.R. Thiessen-Philbrook, R.B. Haynes, R.S.Suri, M.
Salvadori, J. Ray, and W.F. Clark. 2005. Risk of hypertension and reduced kidney function after
acute gastroenteritis from bacteria-contaminated drinking water. Canadian Medical Association
Journal. 173(3): 1-8.
Geldreich, E.E. 1996. Microbial quality of water supply in distribution systems. Boca Raton, FL:
Lewis Publishers.
Gerba, C.P., J.B. Rose, and C.N. Haas. 1996. Sensitive subpopulations: who is at the greatest
risk? International Journal of Food and Microbiology. 30:113-123.
Glass, R.I., P.E. Kilgore, R.C. Holman, S. Jin, J.C. Smith, P.A. Woods, M.J. Clarke, M.S. Ho,
and J.R. Gentsch. 1996. The epidemiology of rotavirus diarrhea in the United States: surveillance
and estimates of disease burden. Journal of Infectious Diseases. 174(Suppl 1): S5—S11.
Gofti-Laroche, L., B. Gratacap-Cavallier, D. Demanse, O. Genoulaz, J.M. Seigncurin, and D.
Zmirou. 2003. Are waterborne astrovirus implicated in acute digestive morbidity (EMIRA
Study)? Journal of Clinical Virology. 27:74-82.
Gould, L. H., L. Demma, T. F. Jones, S. Hurd, D.J. Vugia, K. Smith, B. Shiferaw, S.Segler, A.
Palmer, S. Zansky, P. M. Griffin, and the Emerging Infections Program FoodNet Working
Group. 2009. Hemolytic Uremic Syndrome and Death in Persons with Escherichia coli 0157:H7
Infection, Foodborne Diseases Active Surveillance Network Sites, 2000-2006. Clinical
Infectious Diseases. 49:1480-1485.
Griffin, D.D., C. D. Kirkwood, U. D. Parashar, P. A. Woods, J. S. Bresee, R. I. Glass, and J. R.
Gentsch. 2000. National Rotavirus Strain Surveillance System Collaborating Laboratories.
Surveillance of Rotavirus Strains in the United States: Identification of Unusual Strains. Journal
of Clinical Microbiology. 38(7):2784-2787.
Griffin, P. M. and R. V. Tauxe. 1991. The Epidemiology of Infections Caused by Escherichia
coli 0157: H7, Other Enterohemorrhagic E. coli, and the Associated Hemolytic Uremic
Syndrome. Epidemiologic Reviews. 13(1):60—98.
Guerrant, R. L. 1997. Cryptosporidiosis: An Emerging, Highly Infectious Threat. Emerging
Infectious Diseases. 3(1).
Health Canada. 2000. Waterborne outbreak of gastroenteritis associated with a contaminated
municipal water supply, Walkerton, Ontario, May-June 2000. Communicable Disease Report.
26-20.
Hegarty, J.P., M.T. Dowd and K.H. Baker. 1999. Occurrence of helicobacter pylori in surface
water in the United States. Journal of Applied Microbiology. 87:697-701.
Economic Analysis for the Final RTCR
10-3
September 2012
-------
Helmi, K., S. Skraber, C. Gantzer, R. Willame, L. Hoffmann, and H.M. Cauchie. 2008.
Interactions of Cryptosporidium parvum, Giardia lamblia, vaccinal poliovirus type 1, and
bacteriophages phiX174 and MS2 with a drinking water biofilm and a wastewater biofilm.
Applied and Environmental Microbiology. 74:2079-2088.
Holler, C., S. Koschinsky and D. Witthuhn. 1999. Isolation of enterohaemorrhagic Escherichia
coli from municipal sewage. Lancet. 353:2039.
Hopkins, R. S., P. Shillam, B. Gaspard, L. Eisnach, and R. J. Karlin. 1985. Waterborne disease in
Colorado: three years' surveillance and 18 outbreaks. American Journal of Public Health.
75(3):254-257.
Hoxie, N. J., J. P. Davis, J. M. Vergeront, R. D. Nashold and K. A. Blair. 1997.
Cryptosporidiosis-associated mortality following a massive waterborne outbreak in Milwaukee,
Wisconsin. American Journal of Public Health. 87(12):2032-2035.
Hunter P.R., S. Hughes, S. Woodhouse, N. Raj, Q. Syed, R. M. Chalmers, N. Q. Verlander, and
J. Goodacre. 2004. Health Sequelae of Human Cryptosporidiosis in Immunocompetent Patients.
Clinical Infectious Diseases. 39:504-510.
Hussein, H.S. 2007. Prevalence and pathogenicity of Shiga toxin-producing Escherichia coli in
beef cattle and their products. Journal of Animal Science 85(E. Suppl.): E63-E72
Juranek, D. D. 1998. Cryptosporidium and Water: A Public Health Handbook. Atlanta, Georgia:
Working Group on Waterborne Cryptosporidiosis.
Kaper, J.B, J.P. Nataro, H.L.T. Mobley. 2004. Pathogenic Escherichia coli. Nature Reviews
Microbiology 2(2): 123-140
Kilgore, P.E., R.C. Holman, M.J. Clarke, and R.I. Glass. 1995. Trends of diarrheal disease—
associated mortality in U.S. children, 1968 through 1991. Journal of the American Medical
Association.!!4(14): 1143-1148.
Kleckner, N. and J. Neuman. 2000. Update to Recommended Approach to Adjusting WTP
Estimates to Reflect Changes in Real Income. Industrial Economics, Inc. Memorandum to Jim
DeMocker. U.S. Environmental Protection Agency. September 30.
Kramer, M.H., B.L. Herwaldt, R.L. Calderon, and D.D. Juranek. 1996. Surveillance for
waterborne-disease outbreaks—United States, 1993-1994. Morbidity and Mortality Weekly
Report. 45(SS-1): 1—33.
LeBaron, C.W., M.D. Furutan, J.F. Lew, J R. Allen, D.V. Gouvea, D C. Moe, D.S. Monroe.
1990. Viral agents of gastroenteritis; public health importance and outbreak management.
Morbidity and Mortality Weekly Report. 39(RR-5):l-24.
Livernois, J. 2002. The Economic Costs of the Walkerton Water Crisis. The Walkerton Inquiry,
Commissioned Paper 14. Ontario, Canada: Queen's Printer.
Economic Analysis for the Final RTCR
10-4
September 2012
-------
Marciano-Cabral, F., R. MacLean, A. Mensah, L. LaPat-Polasko. 2003. Identification of
Naegleria fowleri in domestic water sources by nested PCR. Applied and Environmental
Microbiology. 69(10):5864-5869.
Maria, H., A. Elshebani, O. Anders, T. Torsten, and F. Gun. 2005. Simultaneous type 1 diabetes
onset in mother and son coincident with an enteroviral infection. Journal of Clinical Virology
33:158-167.
McMillan, S. 1996. Camp Four Echoes outbreak investigation. Unpublished report by the
Panhandle Health District, Environmental Health Division, Coeur d'Alene, Idaho.
Mead, P.S., L. Slutsker, V. Dietz, L.F. McCaig, J.S. Bresee, C. Shapiro, P.M. Griffin, and R.V.
Tauxe. 1999. Food-related illness and death in the United States. Emerging Infectious Diseases
5(5):607-625.
Means. R.S. 1998. Mechanical Cost Data. Kingston: R.S. Means Company, Inc.
Murray, P.R., E.J. Baron, J.H. Jorgensen, M.L. Landry, and M.A. Pfaller, eds. 2007. Manual of
Clinical Microbiology. Volume 1. Washington, D.C.: ASM Press. 9th Edition. 2256pp.
National Research Council (NRC). 1997. Safe water from every tap, improving water service to
small communities. Washington, DC: National Academy Press.
Office of Management and Budget (OMB). 1996. Economic analysis of Federal regulations
under Executive Order 12866. January 11.
Parashar, U.D., R.C. Holman, M.J. Clarke, J.S. Bresee, and R.I. Glass. 1998. Hospitalizations
associated with rotavirus diarrhea in the United States, 1993 through 1995: surveillance based on
the new ICD-9-CM rotavirus-specific diagnostic code. Journal of Infectious Diseases. 177:13-
17.
Parry, S.M. andR. L. Salmon. 1998. Sporadic STEC 0157 infection: secondary household
transmission in Wales. Emerging Infectious Diseases. 4(4): 1-6.
Perz, J. F., F. K. Ennever, and S. M. Le Blancq. 1998. Cryptosporidium in Tap Water:
Comparison of Predicted Risks with Observed Levels of Disease. American Journal of
Epidemiology. 147(3): 289-301.
Prager, R., S. Annemuller and H. Tschape. 2005. Diversity of virulence patterns among shiga
toxin-producing Escherichia coli from human clinical cases—need for more detailed diagnostics.
International Journal of Medical Microbiology. 295:29-38.
Rangel, J.M., P.H. Sparling, C. Crowe, P.M. Griffin andD. L. Swerdelow. 2005. Epidemiology
of the Escherichia coli 0157:H7 outbreaks, United States, 1982-2002. Emerging Infectious
Diseases. 11(4): 1-11.
Economic Analysis for the Final RTCR
10-5
September 2012
-------
Rolle-Kampczyk, U.E., G.J. Fritz, U. Diez, I. Lehman, M. Richter, and O. Herbarth. 2004. Well
water—one source of Helicobacter pylori colonization. International Journal of Hygiene and
Environmental Health. 207:3 63-3 68.
Rose, J.B. 1997. Environmental ecology of Cryptosporidium and public health implications.
Annual Review of Public Health. 18:135-161.
SAB (Science Advisory Board). 2010. Science Advisory Board Drinking Report on the
Agency's Supporting Analyses for Revisions of the Total Coliform Rule. Final Report. March
19, 2010.
SBAR Panel (Small Business Advocacy Review Panel). 2008. Small Business Advocacy Review
Panel on EPA Planned Revisions to Public Water System Requirements: Revisions to the Total
Coliform Monitoring and Analytical Requirements and Consideration of Distribution System
Issues. Final Report. March 31, 2008.
Skraber, S., J. Schijven, C. Gantzer, and A.M. de Roda Husman. 2005. Pathogenic viruses in
drinking-water biofilms: a public health risk? Biofilms. 2:105-117.
Standridge, J. 2008. E. coli as a public health indicator of drinking water quality. Journal AWWA
100(2):65-75.
Swerdlow, D.L., B.A. Woodruff, R.C. Brady, P.M. Griffin, S. Tippen, H. Donnel, Jr., E.
Geldreich, B.J. Payne, A. Meyer, Jr., J.G. Wells, K.D. Greene, M. Bright, N.H. Bean, and P.A.
Blake. 1992. A waterborne outbreak in Missouri of Escherichia coli 0157:H7 associated with
bloody diarrhea and death. Annals of Internal Medicine. 117(10):812—819.
Tsang, T.H.F., E.K. Denison, H.V. Williams, L.V. Venczel, M.M. Ginsberg and D. J. Vugia.
2000. Acute Hepatitis E infection acquired in California. Clinical Infectious Diseases. 30:618-
619.
Tyler, K.L., E.S. Barton, M.L. Ibach, C. Robinson, J.A. Campbell, S.M. O'Donnell, T. Valy-
Nagy, P. Clarke, J. D. Wetzel, and T. Dermody. 2004. Isolation and molecular characterization
of a novel type 3 reovirus from a child with meningitis. Journal of Infectious Diseases.
189:1664-1675.
United States Census Bureau. 2001a. Households and Families: 2000. Census 2000 Brief.
C2KBR/01-8.
United States Census Bureau. 2001b. Statistical Abstract of the United States: 2001.
Washington, DC. Economics and Statistics Administration.
United States Census Bureau. 2008. Current Population Survey. Annual Social and Economic
Supplement. Population Division, Education & Social Stratification Branch.
United States Environmental Protection Agency (USEPA). 1989. National Primary Drinking
Water Regulations: Total Coliforms {Including Fecal Coliform andE. coli); Final Rule. Federal
Register 54(124):27544. June 29, 1989.
Economic Analysis for the Final RTCR
10-6
September 2012
-------
United States Environmental Protection Agency (USEPA) 1997. The Benefits and Costs of the
Clean Air Act: 1970-1990, Prepared by the Office of Air and Radiation and the Office of Policy,
Planning and Evaluation. EPA/410/R-97/002.
United States Environmental Protection Agency (USEPA). 1998a. Guidance on Implementing
the Capacity Development Provisions of the Safe Drinking Water Act Amendments of1996.
Office of Water. July.
United States Environmental Protection Agency (USEPA). 1998b. National Primary Drinking
Water Regulations: Consumer Confidence Reports: Final Rule. Federal Register 63(160): 44512.
August 19, 1998.
United States Environmental Protection Agency (USEPA). 1998c. National Primary Drinking
Water Regulations: Disinfectants and Disinfection Byproducts; Final Rule. Office of Ground
Water and Drinking Water. Federal Register 63(241):69389. December 16, 1998.
United States Environmental Protection Agency (USEPA). 2000. Guidelines for Preparing
Economic Analyses. U.S. EPA Office of the Administrator. EPA/240-R-00-003. September
2000.
United States Environmental Protection Agency (USEPA). 2001. National Primary Drinking
Water Regulations; Arsenic and Clarifications to Compliance and New Source Contaminants
Monitoring; Final Rule. Federal Register 66(14):6976. January 22, 2001.
United States Environmental Protection Agency (USEPA). 2002a. National Primary Drinking
Water Regulations: Long Term 1 Enhanced Surface Water Treatment Rule : Final Rule. Federal
Register 67(9):1812. January 14, 2002.
United States Environmental Protection Agency (USEPA). 2002b. National Primary Drinking
Water Regulations; Announcement of the Results of EPA 's Review of Existing Drinking Water
Standards and Request for Public Comment. Federal Register 67(74): 19030. April 17, 2002.
United States Environmental Protection Agency (USEPA). 2003a. Labor Costs for National
Drinking Water Rules.
United States Environmental Protection Agency (USEPA). 2003b. National Primary Drinking
Water Regulations; Announcement of Completion of EPA 's Review of Existing Drinking Water
Standards: Notice. Federal Register 68(138):42907. July 18 2003.
United States Environmental Protection Agency (USEPA). 2005. The Safe Drinking Water
Information System—Federal Version (SDWIS/FED) Violation data. Available online at:
http ://epa. gov/enviro/html/sdwi s/ [14 Apr 2009],
United States Environmental Protection Agency (USEPA). 2006a. Ground Water Rule Economic
Analysis. Office of Water. EPA 815-R-06-014. October 2006. United States Environmental
Protection Agency (USEPA). 2007a. Establishment of the Total Coliform Rule Distribution
System Advisory Committee: Federal Register 72(125): 35869. June 29, 2007.
Economic Analysis for the Final RTCR
10-7
September 2012
-------
United States Environmental Protection Agency (USEPA). 2006b. Information Collection
Request for Contaminant Occurrence Data in Support of EPA's Second Six-Year Review of
National Primary Drinking Water Regulations. EPA ICR Number 2231.01., OMB Control No.
2040-0275.
United States Environmental Protection Agency (USEPA). 2007. The Safe Drinking Water
Information System—Federal Version (SDWIS/FED) data. Available online at:
http ://epa. gov/enviro/html/sdwi s/ [29 Apr 2009],
United States Environmental Protection Agency (USEPA). 2008a. 2006 Drinking Water Data
Reliability Analysis and Action Plan, EPA 816-R-07-010. March 2008. Available online at:
www, epa. gov/ safewater.
United States Environmental Protection Agency (USEPA). 2008b. Information Collection
Request for the Microbial Rules. OMB Control Number: 2040-0205. EPA Tracking Number:
1895.04.
United States Environmental Protection Agency (USEPA). 2008c. Information Collection
Request for the Public Water System Supervision Program. OMB Control Number: 2040-0090.
EPA Tracking Number: 0270.43.
United States Environmental Protection Agency (USEPA). 2008d. Assessment of the
Vulnerability of Non-Community Water Systems to SDWA Cost Increases.
United States Environmental Protection Agency (USEPA). 2009. Total Coliform Rule /
Distribution Systems Advisory Committee Agreement in Principle: Notice of agreement in
principle. Federal Register 74(8): 1683. January 13, 2009.
United States Environmental Protection Agency (USEPA). 2010a. Information Collection
Request for the Proposed Revised Total Coliform Rule.
United States Environmental Protection Agency (USEPA). 2010b. National Primary Drinking
Water Regulations: Revisions to the Total Coliform Rule; Proposed Rule. Federal Register
75(143): 40926. July 14, 2010.
United States Environmental Protection Agency (USEPA). 2010c. Preamble for the Final
Revised Total Coliform Rule. June 2010.
United States Environmental Protection Agency (USEPA). 2010d. Technology and Cost
Document for the Final Revised Total Coliform Rule. June 2010.
United States Environmental Protection Agency (USEPA). 2010e. Total Coliform Rule
Compliance Monitoring Data Quality and Completion Report. June 2010.
Ventura, S. J., Mosher, W.D., Curtin, S.C., Abma J.C., and S. Henshaw. 2000. Trends in
pregnancies and pregnancy rates by outcome: estimates for the United States, 1976-96. Vital
and health statistics. Series 21, Data on natality, marriage, and divorce 2000;(56):l-47.
Economic Analysis for the Final RTCR
10-8
September 2012
-------
Vesikari T. 2006. Safety and Efficacy of a Pentavalent Human-Bovine (WC3) Reassortant
Rotavirus Vaccine. The New England Journal of Medicine. 354:23-33.
Vreugdenhil G.R., N.C. Schloot, A. Hoorens, C. Rongen, D.G. Pipeleers, W.J.G. Melchers, B.O
Roep, and J.M.D. Galama. 2000. Acute onset of type 1 diabetes mellitus after severe echovirus 9
infection: putative pathogenic pathways. Clinical Infectious Diseases 31: 1025-1031.
WHO. 2004. Guidelines for Drinking-water Quality. Third Edition, Volume 1,
Recommendations. World Health Organization. Geneva, Switzerland.
Economic Analysis for the Final RTCR
10-9
September 2012
------- |