REGULATORY IMPACT ANALYSIS FOR PROPOSED
TECHNICAL STANDARDS FOR UNDERGROUND STORAGE TANKS
March 30, 1987
-------
REGULATORY IMPACT ANALYSIS FOR PROPOSED TECHNICAL STANDARDS
FOR UNDERGROUND STORAGE TANKS
Prepared for:
Office of Underground Storage Tanks
U.S. Environmental Protection Agnecy
Washington, D.C. 20460
Prepared by:
Sobotka & Company, Inc.
EPA Project Officer:
Sammy K. Ng
March 30, 1987
-------
DISCLAIMER
This report was prepared under contract to an agency of the United States
Government. Neither the United States Government nor any of its employees,
contractors, subcontractors, or their employees makes any warranty, expressed
or implied, or assumes any legal liability or responsibility for any third
party's use of or the results of such use of any information, apparatus, product,
or process disclosed in this report, or represents that its use by such third
party would not infringe on privately owned rights.
-------
ACKNOWLEDGMENTS
The following firms have contributed to this report by providing technical
)r analytical support, and/or writing chapters or sections of the report.
Pope-Reid Associates developed the UST Simulation Model.
Meridian Research, Inc. with support from Versar, Inc., conducted the economic
impact analysis of the retail petroleum industry and wrote Chapter 6 of this
report. Meridian Research, Inc. also wrote Appendix E: Regulatory Flexibility
Analysis.
ICF, Inc. provided the risk analysis, using outputs from the UST Simulation
Model, and contributed to Chapter 7 of this report.
Research Triangle Institute and Glen Anderson of EPA1s Office of Pol icy Analysis
conducted the Denefits analysis and contributed to Chapter 7 of this report.
EPA, Office of Underground Storage Tanks -- Sammy K. Ng was the EPA project
manager for the study and played a principal role in coordinating the contractor
team and in guiding the design and development of this study. Other EPA staff
also played an active role in guiding this study, from the conceptual design
through the development of the analytical methodologies and quality control
checks on data accuracy.
-------
REGULATORY IMPACT ANALYSIS FOR PROPOSED TECHNICAL STANDARDS
FOR UNDERGROUND STORAGE TANKS
Table of Contents
Page
EXECUTIVE SUMMARY ES-1
Chapter 1: INTRODUCTION
l.A. Overview of the Problem of Leaking Underground Storage Tanks 1-1
l.B. The Purpose of a Regulatory Impact Analysis 1-2
l.C. The Purpose of a Regulatory Flexibility Analysis 1-4
l.D. Questions Addressed in This Report 1-4
I.E. Scope of This Analysis 1-5
l.F. Sources of Information 1-6
1.G. Organization and Chapter Summaries 1-6
Chapter 2: OVERVIEW OF METHODOLOGICAL APPROACH 2-1
2.A. Components of an RIA 2-1
2.A.I. Need for the Proposal 2-1
2.A.l.a. Types of Damages from UST Releases 2-3
2.A.l.b. Type of Analysis 2-4
2.A.2. What Are the Alternative Approaches? 2-5
2.A.2.a. Types of Approaches 2-5
2.A.2.b. Types of Analysis 2-7
2.A.3. What Are the Costs? 2-7
2.A.3.a. Direct Costs 2-7
2,A.3.b. Economic Impacts 2-8
2.A.3.C. Implementation Costs 2-9
2.A.4. What Are the Benefits? 2-10
2.A.4.a. Types of Benefits 2-10
2.A.4.b. Types of Analysis 2-10
2.A.5. How Do the Costs and Benefits Compare? 2-10
2.A.5.a. Types of Information 2-10
2.A.5.b. Types of Analysis 2-10
2.B. Overview of the UST Model 2-11
2.B.I. Summary Description of EPA's UST Model 2-11
2.B.2. Limitations 2-13
2.B.3. Model Outputs 2-14
Chapter 3: DESCRIPTION OF THE PROBLEM: DEFINING THE BASE CASE
3.A. Introduction: The Concept of a Base Case and its Role
in the Analysis 3-1
3.B. Overview of Key Factors 3-1
3.B.I. Why Does a Tank Leak? 3-2
3.B.2. What Happens When a Tank Leaks? 3-2
3.B.3. Factors to Evaluate 3-3
3.C. The Tank Population 3-3
3.D. Location of Tanks in Relation to Population and Drinking Water 3-8
3.E. Importance of Hydrogeological Settings 3-13
3.E.I. Hydrogeological Factors Affecting Failure Rates 3-13
3.E.2. Hydrogeological Factors That Affect Characteristics
of the Floating Plume 3-13
3.E.3. Hydrogeological Factors Affecting Character!sties
of the Aqueous Plume 3-15
-l -
-------
Page
3.E.4. Distribution of Hydrogeological Settings 3-17
3.E.5. Summary 3-19
3.F. Current Practices 3-19
3.F.I. Leak Detection 3-19
3.F.2. Responses to Tank Leaks 3-19
3.F.3. State Regulations 3-21
3.G. The Base Case 3-21
3.G.I. Characteristics of the Rase Case 3-21
3.G.2. Proportion of Tanks Leaking 3-21
3.G.3. Predicted Status of Tanks 3-23
3.G.4. Effect of Hydrogeological Setting
and Resulting Plume Areas 3-29
3.G.5. Base Case Responses to Leaks 3-29
3.G.6. Model Results for Base Case 3-30
Chapter 4: COMPONENTS OF A REGULATORY STRATEGY
4.A. Introduction 4-1
4.B. Alternative Elements of a Strategy for New Tanks 4-1
4.B.I. Tank Construction 4-2
4.B.l.a. Effectiveness in Reducing Failures 4-2
4.B.l.b. Costs of Corrosion-Resistant Systems 4-2
4.B.I.C. Comparisons of the Cost and Effectiveness of
Bare and Protected Tanks 4-2
4.B.l.d. Costs and Effectiveness of Tank Systems Constructed
to Intercept Releases 4-5
4.B.l.e. Marginal Cost Effectiveness of Tank Construction
Alternatives 4-5
4.B.l.f. Influences on the Value of Avoiding Releases 4-5
4.B.2. Alternative Systems for Leak Detection 4-8
4.B.2.a. Operation, Advantages, and Disadvantages of Detection 4-8
Methods
4.B.2.b. Costs and Effectiveness for Leak Detection Methods 4-10
4.C. Alternatives for Existing Tanks 4-10
4.C.I. Mandatory Retirement 4-10
4.C.2. Tank Upgrades 4-13
4.C.3. Leak Detection 4-13
4.D. Alternatives for Corrective Action 4-17
4.D.I. Specification of Alternatives 4-17
4.D.2. Effectiveness of Corrective Action 4-18
4.D.3. Costs of Corrective Action 4-18
Chapter 5: ANALYSIS OF COST AND EFFECTIVENESS OF UST REGULATORY OPTIONS
5.A. Methodology for Calculating Costs and Effectivenesses
of the Options 5-1
5.A.I. The Tank Universe Analyzed 5-1
5.A.2. Measurement of Effectiveness 5-1
5.A.3. Costs as Presented 5-2
5.B. Representations of the Base Case, the Proposal, and the
Alternatives 5-2
5.C. Cost and Effectiveness Results 5-4
5.C.I. Comparisons Without Corrective Action 5-4
-ii -
-------
Page
5.C.l.a. Total Cost-Effectiveness Comparisons 5-4
5.C.l.b. Incremental Cost Comparisons 5-6
5.C.2. Changes in Comparisons When Corrective Action Costs
Are Included 5-9
5.C.2.a. Total Cost-Effectiveness Comparisons 5-9
5.C.2.b. Incremental Cost-Effectiveness Analysis 5-9
5.C.3. Sensitivity of Results to Assumptions 5-12
5.C.3.a. Option Choice 5-12
5.C.3.b. Sensitivity to Corrective Action Costs 5-12
5.C.3.C. Sensitivity to the Effectiveness of Retrofit
Cathodic Protection 5-15
5.C.3.d. Sensitivity to the Effectiveness of Leak Detection 5-17
5.C.3.e. Sensitivity to Changes in Hydrosetting Distribution 5-19
5.C.3.f. Sensitivity to Changes in the Discount Rate 5-19
5.C.3.g. Sensitivity to the Choice of Secondary Containment
to the Prevalence of False Reports of Leaks 5-21
5.C.3.h. Phasing of Option Provisions by Age or Risk 5-24
Chapter 6: ECONOMIC IMPACTS OF UST TECHNICAL STANDARDS AND REGULATIONS
6.A. Introduction and Methodology 5-1
6.B. Impacts on UST-owning Firms in the Retail Motor Fuel Marketing
Sector 6-5
6.B.I. The Retail Motor Fuel Marketing Industry: Current Status
and Future Trends 6-5
6.B.2. Methodology and Assumptions 6-15
6.B.3. Impact Analyses 6-21
6.B.4. Interactions with Financial Responsibility Requirements 6-22
6.C. Impacts on Firms Using USTS for Nonretail Motor Fuel Storage 6-51
6.C.I. Overview of Approach 6-52
6.C.2. Results and Analysis 6-61
6.C.3. Limitations of the Analysis 6-62
6.C.4. Sources of Information 6-63
6.D. Impacts on Firms Using USTs for Chemicals Storage 6-65
6.D.I. Methodology 6-62
6.D.2. Results and Analyses 6-65
6.D.3. Limitations of the Analysis 6-69
6.D.4. Sources of Information 6-69
Chapter 7: BENEFITS OF UST TECHNICAL STANDARDS AND REGULATIONS
7.A. Introduction 7-1
7.B. Methodology 7-1
7.C. Health and Safety Risks 7-3
7.C.I. Approach and Assumptions 7-4
7.C.2. Results 7-5
7.D. Property Damage Benefits 7-7
7.D.I. Property Damages Avoided as a Benefits Measure 7-7
7.D.2. Underlying Assumptions and Requirements in the 7-9
Aggregation Process
7.D.3. Damage Function Estimation 7-18
7.D.4. Property Damages Avoided Benefits Projections 7-21
7.E. Environmental Effects 7-26
7.E.I. Approach and Assumptions 7-26
7.E.2. Results 7-31
-iii-
-------
Page
7.F. Option and Existence Value Losses 7-31
7.F.I. Origins of Option and Existence Value 7-34
7.F.2. Option Value 7-32
7.F.3. Existence Value 7-36
Chapter 8: IMPLEMENTATION CONCERNS AND SUMMARY COMPARISON OF
REGULATORY OPTIONS
8.A. Review of Regulatory Options 8-1
8.B. Implementation Concerns 8-2
8.C. Summary Comparisons of Integrated Options 8-4
8.C.I. Summary of Costs 8-4
8.C.2. Summary of Effectiveness and Benefits 8-6
8.C.3. Summary of Tradeoffs 8-8
8.C.4. Summary of Economic Impacts 8-10
8.C.5. Chapter Summary 8-12
Chapter 9: LIMITATIONS OF THE ANALYSIS
9.A. Uncertainties in Assumptions, Data, and Methodology 9-1
9.A.I. Overview of Uncertainties 9-1
9.A.2. Uncertainties in the UST Model 9-1
9.A.2.a. Uncertain Data Inputs 9-1
9.A.2.b. Phenomena Not Addressed by the Model 9-2
9.A.3. Uncertainties in Benefits Analysis 9-2
9.A.3.a. Unmodeled Aspects of Fate and Transport 9-2
9.A.3.b. Toxicological Issues 9-3
9.A.3.C. Nonhealth Benefits 9-3
9.A.4. Uncertainties in Economic Impact Analysis 9-3
9.A.4.a. Uncertainties in the Regulatory Base Case 9-3
9.A.4.b. Uncertainties in Regulatory Costs 9-3
9.A.4.C. Uncertainties in Economic Impacts 9-4
9.B. Effects of Uncertainties 9-4
APPENDIX A: SUMMARY OF THE UST MODEL
APPENDIX B: ZIP CODE ANALYSIS OF UST LOCATION AND POPULATION DENSITY
APPENDIX C: UST MODEL SPECIFICATIONS
APPENDIX D: ECONOMIC IMPACT ASSESSMENT METHODOLOGY
APPENDIX E: REGULATORY FLEXIBILITY ANALYSIS
APPENDIX F: METHODOLOGY FOR ESTIMATING RISK
-iv-
-------
LIST OF EXHIBITS
Pa^e
2.1 Information and Analysis for UST RIA 2-2
2.2 Simple Flowchart of UST Simulation Model 2-12
3.1 Underground Storage Tank Population By Use 3-4
3.2 Underground Storage Tanks by Industry Sector 3-5
3.3 Distribution of Tank Types 3-6
3.4 Bare Steel Tank Age Distribution 3-7
3.5a Population Densities By State 3-9
3.5b Densities of USTs By State 3-9
3.6 Plot of Service Stations and Population 3-10
3.7 Distribution of Wells, Stations, Population 3-11
3.8a Densities of Ground-water Using Population by State 3-12
3.8b Densities of USTs By State 3-12
3.9 Ground-water Regions of the United States 3-14
3.10 Percentages of Ground Water Used from Confined 3-16
Aquifers for Each Ground-Water Region
3.11 Health Region 3-18
3.12 Current Status of UST Population and 3-24
Size Distribution of Release Incidents
3.13 Current Status of UST Population and 3-25
Size Distribution of Release Incidents
3.14 Current Status of UST Population and 3-26
Size Distribution of Release Incidents
3.15 Base Case Population by Status and Age 3-27
3.16 Number of Floating Plumes by Unsaturated Zone Media 3-28
3.17 Cumulative Floating Plume Acres Over a 30 Year Period 3-32
-v-
-------
LIST OF EXHIBITS
Page
4.1 Failure Frequencies by Tank System Type 4-3
4.2 Total Present Value Costs by Tank System Type 4-3
4.3 Corrosion Resistance Cost-Effectiveness 4-4
4.4 Costs and Release Frequencies by Tank System Type 4-6
4.5 Cost-Effectiveness by System Type 4-7
4.6 Costs of Leak Detection Methods 4-11
4.7 Effectiveness of Detection Methods in Reducing Release 4-11
Volumes and Floating Plume Areas
4.8 Cost-Effectiveness by Detection Type 4-12
4.9 Effect of Replacement at Five Years 4-14
4.10 Summary of Cost and Effectiveness of Actions to Upgrade 4-15
Existing Tanks
4.11 Costs and Effectiveness of Leak Detection Methods 4-15
4.12 Cost-Effectiveness by Detection Type 4-16
4.13 EPA Estimates of Approximate Costs and Probabilities 4-20
of Corrective Action Steps
4.14 UST Model Assumptions for Costs of Corrective Action Steps 4-20
5.1 Cost and Effectiveness of Options 5-5
30-Year Costs, No Corrective Action
5.2 Summary of Simulation Results for Current tJST Population 5-7
5.3 Incremental Cost Effectiveness Comparisons 5-8
5.4 Cost and Effectiveness of Options 5-10
30-Year Costs, with Corrective Action
5.5 Cost and Effectiveness of Options 5-11
30-Year Costs, with C.A., Detail
5.6 Cost and Effectiveness of Options 5-13
Variation Due to Detection Choices
-vi -
-------
LIST OF EXHIBITS
Page
5.7 Cost and Effectiveness of Options 5-14
Variations on Corrective Action Costs
5.8 Cost and Effectiveness of Options 5-16
Ground Water Cleanup at 20% of Sites vs. at 40% of Sites
5.9 Impact of Detection Type, Effectiveness 5-18
Assuming 100% Bare Steel Existing Tanks
5.10 Relative Costs of Options by Setting 5-20
(Bare Steel Existing Tanks Only)
5.11 Effect of Discount Rate on Options 5-22
5.12 Effect of Assumed Costs of False Alarms on Incremental Costs 5-23
6.1 Boundary Conditions for Economic Impact Analysis 6-3
6.2 Distribution of the Number and Percentages of Outlets Owned 6-6
by Firms of Various Sizes Engaged in Retail Motor Fuel
Marketing
6.3 Ownership and Operation of Retail Motor Fuel Outlets 6-8
6.4 Distribution of Total Assets Among Firms Owning Retail 6-9
Motor Fuel Outlets
6.5 Distribution of Net Income to Total Assets Ratios Among 6-10
Firms Owning Retail Motor Fuel Outlets
6.6 Number of Firms and Outlets Potentially Affected by UST 6-12
Regulation, By Size Category and Segment
6.7 Projected Exit of Existing Retail Motor Fuel Outlets 6-16
6.8 Use of the Affordabi1ity Model to Perform Economic Analysis 6-18
of the Impacts of UST Regulatory Options
6.9 Key Assumptions and Their Effects on the Results of the 6-20
Analysis
6.10 Additional Assumptions (Economic and Financial) 6-23
6.11 Additional Assumptions (Technical) 6-24
6.12 Costs and Probabilities of Corrective Action Events and the 6-27
Probability That a Tank is Replaced as a Result of a Corrective
Action Event
vii-
-------
LIST OF EXHIBITS
Page
6.13 Additional Assumptions (Related to Choice of Method 6-28
of Compliance)
6.14 Three Sources of Severe Financial Hardship for the 6-29
Median Open Dealer
6.15 Impact of Regulatory Options 6-33
Percentage of Outlets Owned by Small Firms Surviving
and Exiting by Year 5
6.16 Impact of Regulatory Options, Assuming No Revenue Increase: 6-34
Cumulative Percentage of Outlets Owned by Small Firms
Surviving and Exiting Through Year 5
6.17 Impact of Regulatory Options 6-36
Percentage of Outlets Owned by Small Firms Surviving
and Exiting by Year 10
6.18 Impact of Regulatory Options, Assuming No Revenue Increase: 6-37
Cumulative Percentage of Outlets Owned by Small Firms
Surviving and Exiting Through Year 10
6.19 Impact of Regulatory Options, Percentage of Outlets Owned by 6-38
Small Firms Surviving and Exiting by Year 15
6.20 Impact of Regulatory Options, Assuming No Revenue Increase: 6-39
Cumulative Percentage of Outlets Owned by Small Firms
Surviving and Exiting Through Year 15
6.21 Impacts of Regulatory Expenditures on the Ratios of Net 6-41
Income to Total Assets per Outlet for Large Firms
(Excluding Large Oil Companies), Assuming No Revenue Increase
6.22 Economic Impacts of Regulation Under Option II on Small Firms 6-43
in the Retail Motor Fuel Marketing Industry, Assuming Revenue
Increases of 1, 3, and 5 Percent
6.23 Economic Impact of Regulation Under Option II on the Ratio of 6-45
Net Income to Total Assets per Outlet for Large Firms, Assuming
Revenue Increases of 1, 3, or 5 Percent
6.24 Screening Analysis for Non-Retail Petroleum Tanks 6-54
Regulatory Costs of $500/Tank
6.25 Screening Analysis for Non-Retail Petroleum Tanks 6-56
Regulatory Costs of $5000/Tank
-vni
-------
LIST OF EXHIBITS
Page
6.26 Screening Analysis for Non-Retail Petroleum Tanks 6-58
Regulatory Costs of $50,000/Tank
6.27 Screening Analysis for Underground Chemical Tanks 6-66
Regulatory Costs of $500/Tank
6.28 Screening Analysis for Underground Chemical Tanks 6-67
Regulatory Costs of $5000/Tank
6.29 Screening Analysis for Underground Chemical Tanks 6-68
Regulatory Costs of $50,000/Tank
7.1 Frequency of MEI Risk, Base Case vs. all Options 7-6
7.2 Frequency of MEI Risk (with T/0 Cutoff) 7-8
Base Case vs. all Options
7.3 Coefficients of Plume Size 7-10
7.4 Number of Plumes Per Tank 7-10
7.5 Number of Plumes by Size, Thirty-Year Time Period 7-12
7.6 Total Number of Plumes, by Five-Year Increments 7-12
7.7 Number of Plumes, by Size — Base Case 7-13
7.8 Number of Plumes, by Size — Option I 7-13
7.9 Number of Plumes, by Size -- Option II 7-13
7.10 Number of Plumes, by Size -- Option III 7-14
7.11 Number of Plumes, by Size -- Option IV 7-14
7.12 Number of Plumes, by Size -- Option V 7-14
7.13 Plume Measurements 7-15
7.13(A) Compilation of Well Contamination and Distance 7-17
7.14 Benefits Summary: Benefits of Control Options 7-22
Relative to Base Case (Billions of Dollars)
7.15 Lost Profits of Tank Owners 7-23
7.15(A) Comparison of Lost Profits Relative to Base Case 7-23
-ix-
-------
LIST OF EXHIBITS
PiSi
7.16
Aggregate Damages
7-25
7.17
Comparison of Benefits to Base Case
7-27
7.18
Input Vari ables
7-28
7.19
Summary of Data on U.S. Streams (by Stream Order)
7-30
7.20
Number of USTs with Aquatic Impacts (by Stream Order)
7-32
7.21
Distribution of Streams Impacted (by Stream Order)
7-33
8.1
Summary of Total Costs
8-5
8.2
Summary of Effectiveness and Benefits
8-7
8.3
Tradeoffs
8-9
8.4
Summary of Economic Impacts
8-11
8.5
Summary of Effects
8-13
-x-
-------
EXECUTIVE SUMMARY
INTRODUCTION
Section 9003(a) of the Hazardous and Solid Waste Amendments of 1984 (HSWA)
requires the Environmental Protection Agency (EPA) to "promulgate release
detection, prevention, and correction regulations applicable to all owners and
operators of underground storage tanks, as may be necessary to protect human
health and the environment." In response to this mandate, EPA is proposing
technical standards requirements for owners and operators of underground tanks
(USTs) containing petroleum and other regulated substances, except substances
regulated as hazardous waste under Subtitle C of the Resource Conservation and
Recovery Act (RCRA). The technical standards package includes proposed regula-
tions for new and existing tanks with provisions covering release detection,
general performance standards for tank systems and leak detection equipment,
mandatory upgrading of existing USTs, corrective action, closure, and reporting
and recordkeeping.
Executive Order 12291 requires regulatory agencies to determine whether a
proposed regulation is a major rule, and if so, to conduct a Regulatory Impact
Analysis (RIA) for the proposed rule.*/ The Executive Order defines a major
rule to be one that is likely to result in (1) an annual effect on the economy
of $100 million or more, (2) a major increase in costs or prices for consumers,
individual industries, Federal, State or local government agencies, or geo-
graphic regions, or (3) significant adverse effects on competition, employment,
investment, productivity, innovation, or the ability of United States-based
enterprises to compete in domestic or export markets. EPA1s analysis shows
that the proposed rule for the technical standards for underground storage
tanks will have an annualized cost for prevention and detection greater than
$100 million. This RIA provides an analysis of EPA's requirements for UST tech-
nical standards based on the guidelines contained in the Office of Management
and Budget's "Interim Regulatory Impact Analysis Guidance" £/ and EPA's "Guide-
lines for Performing Regulatory Impact Analyses."^/
Besides the HSWA Amendments to RCRA requiring EPA to establish technical
requirements for underground storage tanks, the 1986 Superfund Amendments and
Reauthorization Act (SARA) also amended Subtitle 1 to require EPA to promulgate
financial responsibility regulations for owners and operators of underground
storage tanks. The financial responsibility regulations and the technical re-
requirements for USTs are being proposed separately. The financial responsi-
bility package includes proposed regulations requiring owners and operators of
USTs to maintain evidence of financial responsibility for corrective action and
compensation of third parties for damages caused by releases from their USTs.
The proposed regulations for technical standards and for financial responsi-
bility are supported by separate regulatory impact analyses and other analyses
(e.g., paperwork burden analyses) required by law. These analytical efforts
have been coordinated to ensure consistency and accuracy, as will be subse-
quently discussed in this Executive Summary.
}_/ Federal Register, Vol. 46, February 19, 1981, p. 13193.
£/ Office of Management and Budget, Interim Regulatory Impact Analyses
Guidance, June, 1981.
3/ U.S. EPA, Guidel ines for Performing Regulatory Impact Analyses,
DecemFer 1983.
-------
ES-2
THE REGULATED COMMUNITY
The proposed requirements for technical standards for underground storage
tanks will affect a regulated community of approximately 450,000 establishments
owning an estimated 1.4 million tanks. }_/ The large majority, 96 percent, of
these underground tanks are used to store gasoline and other petroleum products.
The remaining 4 percent of USTs are used to store oth.er chemicals. More than
half of the petroleum product USTs are used by gas stations and other motor
fuel retailers. Other industry sectors owning petroleum product underground
storage tanks include: agriculture £/, manufacturing, transportation, govern-
ment and wholesale and retail trade.
An overwhelming majority, an estimated 89%, of existing petroleum USTs are
constructed of bare steel. Bare steel tanks are not protected against corrosion
and therefore fail more often and have a shorter average life than cathodically
protected steel tanks and tanks constructed of other corrosion-resistant
materials.
SCOPE OF THIS ANALYSIS
This Regulatory Impact Analysis presents the results of several analyses
of EPA's technical standards for USTs containing gasoline, or petroleum pro-
ducts. The analyses performed for this RIA were done assuming all USTs contain
gasoline. Therefore, except for a screening analysis of economic impacts for
industries operating hazardous substances USTs presented in Chapter 6, the
results presented here do not necessarily apply to USTs used to store any
substance other than gasoline or petroleum products. However, as stated above,
USTs used to store other than petroleum products account for only 4% of the UST
population.
This Regulatory Impact Analysis presents, in detail, the costs, effective-
ness, risks and economic impacts of alternative requirements for preventing
leaks from occuring and/or detecting them more rapidly if they do occur. The
analysis also considers the effect of prevention and detection measures on the
costs of corrective action for plumes that may still result. This RIA analyzes
the cost and economic impacts, but not the effectiveness, of the corrective
action measures that are proposed by EPA — the rationale for the corrective
action policy proposed is presented in the Preamble to the proposed rule.
The impacts resulting from closure requirements are not highlighted in this
RIA, but have been included in the analysis in the sense that when tanks are
replaced, the replaced tanks are assumed to be closed in a manner consistent
with the proposed regulation (i.e., release detection and corrective action as
appropriate). The analysis of economic impacts does address the costs of
closure requirements.
Recordkeeping costs are currently in the process of being developed and
are not analyzed in this RIA.
V Data Resources, Inc., Underground Storage Tanks, Technical/Financial/
Economic Data Collection, October 2, 1985.
2/ Only underground storage tanks with capacities greater than 1,100
gallons will be regulated, thus excluding many tanks found on farms.
-------
ES-3
THE REGULATORY OPTIONS
In developing the proposed rule, EPA considered three approaches for
regulating new tanks and four approaches for regulating existing tanks that
were then combined in different ways to form five regulatory options. For
new tanks, EPA considered the following regulatory approaches:
o protected single-wall tanks with release detection;
o secondary containment with interstitial monitoring; and
o a class approach.
For existing tanks EPA considered the following regulatory approaches:
o mandatory upgrading or replacement of substandard USTs;
o methods of release detection;
o frequency of release detection; and
o phase-in of release detection.
EPA combined these general approaches for new and existing tanks into five regula-
tory options for prevention and detection. This RIA presents the cost and effec-
tiveness of the five options for technical standards, in combination with the
proposed option for corrective action. The analysis of the options was performed
using a single corrective action policy (the proposed policy) for all the options.
Only the proposed corrective action policy was assumed for each option so that
the differences between the technical standards options could be compared on a
consistent basis, and so that incremental differences between options could be
attributed to the prevention and detection requirements and not obscured because
of differences in the corrective action policy. The five technical standards
options and the proposed corrective action policy are described below.
Option I (Baseline level): requires manual inventory control, quarterly leak
detection installed within three years (five years for corrosion-resistant
tanks) for existing and new tanks, and corrosion protection for all new tanks.
Periodic tightness tests may be used as an alternative to quarterly vapor wells.
Option II--The Proposea Rule (Enhanced baseline plus targeted upgrading): i s
similar to Option I with upgrading to new tank standards within ten years,
though leak detection systems must be sampled monthly rather than quarterly and
tightness tests are not to be used in lieu of monthly leak detection after
tanks are replaced or upgraded to new tank standards (corrosion protection).
Option III {Baseline plus secondary containment for new tanks): requires quarterly
leak detection or periodic tightness tests for existing tanks and secondary
containment with interstitial monitoring for new tanks. For existing tanks, this
option is identical to Option I. New tanks, however, must be lined systems (or
double-walled tanks) with interstitial monitoring.
Option IV (Class Option): requires rapid replacement of existing tanks and secon-
dary containment for new tanks within state-designated well-head protection
areas. Tanks in other areas are required to conform to baseline standards (Op-
tion I).
Option V (Emphasis on leak prevention): For existing tanks, this option re-
quires manual inventory control, frequent (continuous) leak detection starting
in three years, and early retirement. Existing tanks must be replaced within
ten years with secondary containment systems.
-------
ES-4
Proposed Corrective Action Policy: Where a release occurs, an investigation and
actions to reduce immediate hazards are followed by limited removal of contamin-
ated soil, and removal of any free product from the ground water. The need for
subsequent cleanup at those sites where ground water is contaminated is to be
determined through site-specific exposure and risk assessment.
APPROACH USED TO ANALYZE THE REGULATORY OPTIONS
Prevention, detection and corrective action are substitutes for one another
in avoiding the same damages resulting from leaking USTs. As a result, it was
necessary to perform an integrated analysis showing how all three requirements
interact together, and to clearly present the integrated results. These aspects
of the analysis and related methodological issues are discussed briefly below.
Integrated Analysis
Once preventive measures are undertaken that prevent plumes from forming,
the value of additional detection measures is reduced. Similarly, detection
measures that allow detection of releases before large plumes are formed can
avoid much of the damages from a release, and therefore reduce the value of
requiring tank systems that are less likely to fail. In addition, different
combinations of tank systems and detection methods can work together to achieve
similar reductions in damages, but at different costs to the tank owner.
Prevention and earlier detection can also reduce the cost of response once
the leak is discovered because they result in fewer and smaller plumes that
can be less expensive to clean up. Conversely, it is not necessary to rely on
prevention and detection alone to avoid damages that result from leaking USTs,
because response measures may also be able to achieve some of these benefits.
However, corrective action will not reduce the damages that occur prior to the
plume's discovery, and there can still be residual damages that remain even
after corrective action is completed. Also, there can be implementation pro-
blems associated with corrective action.
Because prevention, detection and corrective action interact, it is essen-
tial to analyze them together. However, their combined analysis is difficult
because of the complex dynamics that determine the occurence of tank failure,
leak characteristics, effectiveness of detection methods, and variability in
plume characteristics due to hydrogeologic factors. To assist the analysis, EPA
developed the UST Simulation Model which simulates tank failure, leak detection,
and plume development. Using the UST model, we can estimate the combined
effects of prevention and detection options in terms of the number, duration
and magnitude of plumes. The model also estimates the costs of prevention,
detection, and corrective action, as well as other costs associated with product
loss, system repairs and tank replacement.
Using the UST Model results, we can show the cost and effectiveness trade-
offs between different combinations of prevention and detection measures. We
can also show how different combinations of prevention and detection measures
result in lower costs for corrective action.
-------
ES-5
Methodological Issues
Improved prevention and detection measures avoid plume area, and therefore
reduce corrective action costs. If corrective action requirements (and their
resulting costs) were already established, then it would be straightforward
to simply include corrective action costs as regulatory costs, and to present
the savings in corrective action costs that prevention and detection measures
provide. It would also be straightforward to include corrective action require-
ments and their costs in the base case -- the situation in the absence of
additional federal requirements.
However, the corrective action requirements are part of the proposed rule.
Therefore, the presentation of information about corrective action requires
decisions about whether corrective action cost should be shown as a regulatory
compliance cost, how we measure effectiveness of corrective action, and whether
corrective action should be assumed to occur in the base case in the absence of
other federal requirements. These decisions and their implications are discus-
sed below.
In presenting information about corrective action, we generally show the
cost of corrective action as a regulatory compliance cost that is reduced as a
result of prevention and detection measures. For example, the RIA typically
provides information about the cost of avoiding plumes (the cost of prevention
and detection), and the cost associated with the plumes that still result (the
cost of corrective action), to provide a total cost. In this way, it is easy
to see the total costs resulting from the options, as well as the cost trade-
offs between corrective action and prevention/detection.
Alternatively, corrective action costs avoided can be considered to be a
benefit, especially if corrective action costs are viewed as a proxy for the
damages incurred. Corrective action costs avoided might be considered to be a
partial proxy for damages avoided if it is believed that on a site-specific
basis, corrective action would be undertaken to the extent that the benefits
justify the costs. Given the level of information available for effectiveness
measures and benefits at this point, we generally chose not to use this alternate
form of presentation in the RIA. Therefore, we generally discuss "cost savings"
or "cost trade-offs" rather than "net benefits."
In this RIA we measure effectiveness in terms of plume area avoided.
This intermediate measure is fairly straightforward to estimate, easy to visu-
alize, and monotonically (though not linearly) related to most final damage
measures. Ideally, we would prefer to show a more final and comprehensive
measure of effectiveness that monetizes all of the benefits of avoiding plumes.
Although progress has been made in this RIA in developing additional measures
of benefits, plume area avoided is the most direct and reliable measure avail-
able at this point.
When comparing costs with effectiveness as measured by plume area avoided,
it is important to keep the composition of the costs in mind to properly inter-
pret the results and avoid double counting benefits. When considering only
detection and prevention costs, plume area avoided can be considered to be a
proxy for most of the damages associated with leaks, including those addressed
by corrective action. When considering corrective action costs in addition to
-------
ES-6
prevention and detection costs, the plume area avoided should be a proxy only for
those damages that corrective action fails to address (primarily damages prior
to detection, such as property damages and health effects).
In order to facilitate the comparison of alternative prevention and detec-
tion options, we analyzed the costs and effectiveness for all options assuming
only the proposed corrective action policy would be applied for all five options.
We also included the cost of the corrective action policy in the base case,
implicitly assuming that corrective action would be required even if there were
no additional federal requirements for prevention and detection measures.
Failure to include corrective action in the base case would have obscured the
trade-offs between the cost of prevention and detection, and the cost of correc-
tive action. This RIA fully analyzes the cost and effectiveness of the preven-
tion and detection requirements and the cost of the proposed corrective action
policy, but does not analyze the effectiveness of the corrective action policy.
ANNUALIZED COST OF THE PROPOSED RULE
The Cost of the Proposed Rule
The average annualized incremental cost of the requirements for prevention
and detection is estimated to be $0.21 billion (or $150 per UST). This cost
includes incremental costs of installing and maintaining USTs, incremental
detection and monitoring costs, incremental tank removal and replacement costs,
incremental costs for pipe repairs, and adjustments for the value of product
lost. The average annualized incremental costs are calculated from a base case
which includes the statutorily mandated "Interim Prohibition" against bare steel
tanks, and other important assumptions which are detailed in Chapter 3.
The averaye annualized cost of corrective action under the proposed rule
is $3.05 billion. The average annualized incremental costs of corrective action
under the proposed rule are less than in the base case because of the effective-
ness of prevention and detection measures under the proposal. If the same
corrective action requirements are assumed to apply to the base case, annualized
corrective action costs for the base case are estimated to be $4.58 billion. As a
result, the proposal saves $1.53 billion in corrective action costs.
In total, the $0.21 billion incremental annualized cost for prevention and
detection produces an incremental annualized savings of $1.53 billion in correc-
tive action costs, resulting in a net savings of $1.32 billion annually.
If it is assumed that the cost of the corrective action measures does not
exceed the damages that are avoided or mitigated by corrective action, this
result might also be viewed in terms of benefits as follows: to the extent that
corrective action costs are considered to be an appropriate partial proxy for
damages, then the base case is experiencing damages on the order of $4.58
billion annually, which are reduced by the proposal to $3.05 billion at a cost
of $0.21 billion, yielding a net benefit of $1.32 billion annually.
Costs of the Regulatory Options
Exhibit ES.l provides the annualized cost for prevention/detection and
corrective action for the base case and each regulatory option, and also
presents the annualized costs as incremental costs from the base case.
-------
ES-7
Exhibit ES.l
ANNUALIZED
COSTS
OF UST REGULATORY
OPTIONS
Base
Case
Option
I
Option
II
Option
III
Option
IV
Option
V
(a) Total Costs
($ Bi11 ion)
Prevention & Detection V
$1.58
$1.68
$1.79
$2.19
$2.30
$2.88
Corrective Action 2/
4.58
4.13
3.05
3.86
2.56
2.07
Total Annualized Cost £/
$6.16
$5.81
$4.84
$6.05
$4.86
$4.95
(b) Incremental Costs
Prevention & Detection
$0.10
$0.21
$0.61
$0.72
$1.30
Corrective Action (Savings)
(0.45)
(1.53)
(0.72)
(2.02)
(2.51
Total Incr. Cost (Savings)
($0.35) ($1.32) ($0.10) ($1.30) ($1.21)
]_/ Cost includes tank acquisition/installation, detection/monitoring,
tank removal, and an offset for reduction in product lost. Present value
costs includes all capital and operating costs expected to be incurred over
30 years, discounted at 3%.
£/ Estimated costs for corrective action in each scenario, assuming EPA's
corrective action requirements are applied to each scenario (including the
base case).
£/ Total present value cost of prevention and detection + total present
value cost of corrective action.
-------
ES-8
As can be seen from the exhibit, the incremental costs for prevention/
detection and corrective action are very different across the options, but the
differences in total costs are smaller. This reflects the trade-off between
prevention/detection and corrective action. The total costs fall generally into
two yroups: the highest total costs are for the base case and Options I and III;
the lowest total costs are for Options II, IV and V. Between Options II, IV
and V, Option II has the lowest cost of prevention and detection, the highest
cost for corrective action, and the lowest total cost.
EFFECTIVENESS OF THE PROPOSED RULE
Effectiveness can be estimated by the plume area that would be avoided if
the regulation were in place (the difference between the plume area present in
the base case and the plume area that will occur after the regulations take
effect). The plume area avoided by each option is shown in Exhibit ES.2.
Exhibit ES.2
PLUME AREA AVOIDED RELATIVE TO BASE CASE
Base Option Option Option Option Option
Case I II III IV V
Incremental Percent Plume
Area Avoided 54% 67% 55% 68% 83%
Incremental Plume Acres
Avoided 103,500 128,500 105,500 129,000 159,000
All of the options avoid a substantial percentage of the plume area in the
base case -- the options avoid 54% to 83% of the plume area of the base case
over thirty years. Option II, the regulatory option proposed, avoids 128,500
plume acres, or 67% of the plume area associated with the base case.
Options I and III avoid less plume area (54% and 55% respectively).
Option IV avoids slightly more plume area (68%) than Option II, but will tend
to avoid plumes at locations where the damages are likely to be greatest.
However, concerns about implementing the class system proposed in Option IV put
into question the likelihood that this performance can in fact be achieved.
Option V avoids the most plume area (83%), but concerns about the feasibility
of replaciny existing systems with secondary containment systems and potential
economic impacts put into question whether this level of performance can be
achieved.
Although plume acres avoided is a useful proxy for the effectiveness of
options for prevention and detection, it is only an intermediate measure of such
damages as health and safety risks, property damage, damage to aquatic ecosys-
tems, and reduction in option or existence value. Efforts to quantify these
damages are complex and are still ongoing. These efforts are described in
Chapter 7.
-------
ES-9
ECONOMIC IMPACTS
Economic impacts were considered for three classes of facilities which own
or operate USTs: (1) facilities using USTs for storing motor fuels for retail
marketing; (2) facilities using USTs for storing petroleum products for purposes
other than retail marketing; and (3) facilities using USTs for storing regulated
hazardous substances.
Economic impacts are far more likely to be significant for the retail
motor fuel marketing class than for the other two classes mentioned above.
Reasons for this include: (1) there are no substitutes for USTs in retail motor
fuel marketing; (2) there are many outlets owned by small businesses; and (3)
there tend to be at least three USTs per facility, so UST regulatory costs
represent a potentially significant fraction of capital and operating expenses.
Thus, the economic impact analysis focuses on the retail motor- fuel market,
though a screening analysis is performed on the other two classes of facilities
mentioned above.
An economic impact analysis was conducted using a worst-case assumption
that firms will not be able to recover any of the costs of the regulations
through price increases. Exhibit ES.3 summarizes the potential exits of small
firms in the retail motor fuel marketing sector (having annual sales less than
$4.6 million) in the first five years of the regulations, assuming no revenue
increase:
Exhibit ES.3
Percentage Of Outlets Owned By Small Firms Exiting Through Year 5
(Assuming No Revenue Increase)
Reason For Exit
Natural Exit Rate (Current Trend)
Tank Replacement or Upgrade
Corrective Action
Total Exit Through Year 5:
I
II
OPTION
III
IV
V
19%
19%
19%
19%
19%
2%
2%
2%
15%
41%
52%
50%
52%
39%
22%
73%
71%
73%
73%
82%
Under an assumption of no revenue increase per facility and limited ability
of small firms to get loans to cover compliance-related costs, the burden on
small firms could be significant. These potential burdens are largely attribu-
table to corrective action costs, although tank replacement and upgrade costs
are important factors for Option IV, and the most important factor for Option
V. The profitability of larger firms (excluding large oil companies) could
also be significantly affected by corrective action costs, but no estimates of
these exit rates are developed in this RIA.
-------
ES-10
If the assumption of no revenue increase is relaxed, the exit rates are
reduced. For example, a 3% revenue increase might be a reasonable assumption
for small firms since that is the level that would allow larger firms to remain
sufficiently profitable. Under this assumption, for Option II, the exits due
to tank replacement or upgrade do not occur, and the exits due to corrective
action drop from 50% to 37%. Should a revenue increase greater than 3% occur,
survival rates under Option II are projected to approach base case survival
rates over the first ten years.
Because the proposed regulation may have a significant impact on a substan-
tial number of small entities, a Regulatory Flexibility Analysis, as required
by the Regulatory Flexibility Act of 1980 (5 U.S.C. 601-612), has been prepared
and is included as Appendix E. The corrective action policy option chosen by
EPA maximizes flexibility to determine appropriate long-term corrective actions
on a site-specific basis, while at the same time adequately protecting human
health and the environment. This built-in flexibility is an important factor in
evaluating the potential effects of this rulemaking on small business.
IMPLEMENTATION CONCERNS
The regulatory options have been qualitatively assessed relative to one
another in this RIA. The proposed regulation raises relatively fewer implemen-
tation issues relative to its alternatives (except for Option I), though imple-
mentation concerns may still be significant. Options III, IV and V rely on
secondary containment which may be difficult to accomplish because of possible
capacity shortages in tank manufacturing and shortages in installation exper-
tise. The class system in Option IV may create additional implementation
difficulties, while the economic impacts associated with the tank replacement
requirements of Option V may hinder implementation there.
COORDINATION BETWEEN TECHNICAL STANDARDS RIA AND FINANCIAL RESPONSIBILITY RIA
The technical standards RIA and the financial responsibility RIA use simi-
lar assumptions about the size of the regulated community, the financial and
operational characteristics of regulated firms, and the unit costs of complying
with the regulations. However, the two analyses use different probabilities
that corrective action events will occur at regulated tanks. Much of the data
needed to produce precise estimates of event probabilities and costs are just
not available and in the analyses for both RIAs appropriate assumptions for each
analysis had to be made based upon the data and expertise available. Both RIAs
present estimates of event probabilities based upon best available information.
The technical standards RIA develops probabilities and costs of corrective
action for al1 regulated USTs; the financial responsibility RIA, on the other
hand, develops the probabilities and costs of corrective action only for those
firms that will be able to use insurance as a financial assurance mechanism.
The financial responsibility RIA excludes, for example, the probabilities and
costs of corrective action for incidents occurring at facilities with USTs that
will not be able to obtain insurance, incidents occurring before the owner or
operator has insurance, and incidents that are not reported to the insurer.
The technical standards RIA may overstate the probabilites and costs of correc-
tive action events, because the analysis for the technical standards RIA does
-------
ES-11
not account for owner/operator voluntary upgrade and early retirement programs.
UST owners or operators may upgrade their existing tanks to new tank standards
or replace their existing bare steel tanks with corrosion protected tanks in
order to reduce the risk of corrective action requirements or to increase their
ability to obtain liability insurance.
In summary, the two RIAs may present different results for two reasons:
(1) they examined a different population of underground storage tanks; and (2)
in the analysis for both RIAs, the limited availability of data made it necessary
to make assumptions appropriate for each analysis in order to produce estimates
of the effects of different UST regulations. As a result of these differences,
the corrective action event probabilities developed in this RIA and the insurance
claim rates presented in the financial responsibi1ity RIA cannot be directly
compared.
LIMITATIONS
The results presented in this Executive Summary and in the RIA should be
viewed in the context of the limitations of the analysis as presented in
Chapter 9.
-------
Chapter 1
INTRODUCTION
l.A. OVERVIEW OF THE PR08LEM OF LEAKING UNDERGROUND STORAGE TANKS
It is estimated that over one half million establishments in the United
States use approximately 1.4 million underground tanks for the storage of
petroleum products and chemicals.V About 4 percent of all tanks store chem-
icals, and the rest store petroleum either for retail sale (50 percent of
tanks) or for the facility's own use (46 percent of tanks).£/ The tank owning
facilities represent all sectors of private and public enterprise, including
agriculture, mining, construction, manufacturing, transportation, communica-
tion, utilities, wholesale and retail trade, services, and military and non-
military government at all levels.
Most of these underground tanks are made of bare steel, with no protective
measures to prevent corrosion. Most bare steel tanks are quite old, and an
estimated 60 percent have been in use for over 15 years. £/
In recent years, mounting evidence has led to increasing concern that
leakage from underground storage tanks represents a significant hazard to human
health and the environment. Various investigations of the problem have identi-
fied the conditions under which underground storage tanks have a high potential
for failing. While these investigations have pointed to older steel tanks
without corrosion protection as being the tanks most likely to fail, other types
of tanks can also fail, and tanks may fail from reasons other than corrosion.
Leaking underground tanks can result in significant resource losses, both from
resulting contamination of ground water and from the loss of tank contents, in
health risks from contaminated drinking water, and in safety hazards from fires
or explosions.
Possible responses can address tanks already in place ("existing tanks")
to detect leakage and minimize its consequences, or can focus on tanks being
installed in the future ("new tanks") to minimize future leaks and damages.
Existing tank responses include detection measures, tank material and technolog-
ical upgrades, and requirements for early tank retirement. Additional measures
could require corrective action for leaks that form plumes and could impose
financial responsibility requirements. New tank options can include tank
technical specifications and installation protocols, as well as the other
options available for existing tanks. There are many examples of public and
private responses to the problem. The responses have varied, according to the
needs and priorities of the parties involved.
V EPA, Regulation of Underground Storage Tanks, Preamble and Proposed
Regulations, Draft, November 24, 1986, p. 10.
£/ EPA Office of Underground Storage Tanks, Summary of State Reports on
Releases from Underground Storage Tanks, August 1986.
£/ Data Resources, Inc., Underground Storage Tanks, Technical/Financial/
Economic Data Collection, October 2, 1985.
-------
1-2
o In states, counties, and municipalities where leaking underground tanks
are perceived as a particularly important threat, statutes and regula-
tions have been devised to require tank testing, ground-water monitor-
ing, maintaining inventory records, and other responses to prevent or
detect leaks.
o Among gasoline retailers, the industry with the largest number of
underground storage tanks, many owners—particularly the large, major
oil companies and their distributors—have embarked on programs of
systematic replacement of the oldest tanks in their facilities irrespec-
tive of evidence of leaking, and have instituted the practice of regular
tank testing or environmental monitoring.
o Local building codes, local and national fire codes, and national
petroleum trade association codes all specify recommended methods for
installing and maintaining underground storage tanks.
In spite of considerable industry and public sector interest and activity
in detecting and preventing underground tank leaks, information on some aspects
of the problem can still be characterized as anecdotal. To address the need
for systematic information on an extremely complex problem, the U.S. Environ-
mental Protection Agency (EPA) undertook a program of studies starting early in
1984 to evaluate the nature and magnitude of the problem, identify and evaluate
technical options for addressing it, and assess the need for and potential
consequences of regulation.
Under the 1984 Amendments to the Resource Conservation and Recovery Act
(RCRA), enacted on November 8, 1984, EPA was charged with establishing a regula-
tory program for underground tanks storing petroleum products and hazardous
substances other than hazardous waste. Under the statutory requirement of HSWA
69003(a), EPA shall "promulgate release detection, prevention, and correction
regulations ... as may be necessary to protect human health and the environ-
ment ."V
l.B. THE PURPOSE OF A REGULATORY IMPACT ANALYSIS
Executive Order 12291 (supplemented by EPA Guidelines)^/ requires the
Agency to prepare a Regulatory Impact Analysis (RIA) for every major regulation,
defined as one imposing annual costs over 100 million dollars. Because of the
large number of tanks, almost any regulation of underground storage tanks is
major under this definition.
A principal goal of the Executive Order is to ensure that a regulation
confers net benefit—that is, that the benefits of a regulation outweigh its
costs to society. A second goal is that a regulation be cost-effective—that
]_/ Hazardous and Solid Waste Amendments of 1984 (Public Law 98-161),
Subtitle I, 69003(a), November 1984.
_£/ U.S. EPA, Guidelines for Performing Regulatory Impact Analyses, Decem-
ber 1983.
-------
1-3
is, to ensure that a regulation is not selected if, among the group of feasible
regulatory alternatives, there is an approach that is expected to confer bene-
fits at least as high as the proposal but at lower cost, or to confer greater
benefits than the proposal but at no greater cost. The RIA supports these
goals by providing a systematic presentation of information on benefits, costs,
and economic impacts in a way that clarifies the implications of alternative
regulatory approaches. In addition to serving as a tool for selecting from
among alternatives, the RIA process can also prove useful in developing new
regulatory alternatives.
The EPA has provided specific guidance regarding the contents and conduct
of RIAs in the "Guidelines for Performing Regulatory Impact Analyses" and its
appendices. The key elements of an RIA are:
o Stating the need for and consequences of the proposal:
What is the problem being addressed, why is federal action necessary,
and what difference is the regulatory action expected to make?
o Considering alternative approaches:
What are the available regulatory and non-regulatory options, and what
is the base case" that is likely to prevail in the absence of regulation?
o Assessing benefits:
What are the benefits, incremental from the base case, of the proposal,
including reductions in risk to human health and safety, reductions in
economic property damage and resource loss, and reductions in environ-
mental damage?
o Assessing costs:
What are the costs, incremental from the base case, including real
resource costs, government regulatory costs, deadweight welfare losses,
and adjustment costs for displaced resources? What are the impacts
of these costs, and who ultimately bears the costs? In addition, what
are the possible adverse effects on product quality, production, compe-
tition, innovation, and market structure?
o Evaluating costs and benefits:
What are the total and incremental benefits and costs for each alterna-
tive, and which alternative results in the greatest net benefit? If
benefits cannot be easily monetized, what is the cost-effectiveness of
the alternatives?
These components are interdependent. For example, stating the need for the
proposal, considering alternative approaches, and assessing costs all depend
critically on understanding the base case in the absence of regulation. Thus,
each part of an RIA must be developed with the others in mind. Certain data
input and analytical steps occur more than once in addressing the key questions
posed in an RIA.
-------
1-4
l.C THE PURPOSE OF A REGULATORY FLEXIBILITY ANALYSIS
The Regulatory Flexibility Act requires that regulatory agencies carefully
consider the potential effects of regulation on small entities.^./ Small enti-
ties include small businesses (generally firms with less than 500 employees),
small organizations (any not-for-profit enterprise that is independently owned
and operated and not dominant in its field), or small governmental jurisdictions
(any government of a district with a population less than 50,000). If the
proposed regulation will "have a significant economic impact on a substantial
number of small entities," the regulatory agency must prepare a Regulatory
Flexibility Analysis (RFA) addressing these issues.^/ The RFA must contain:
(1) the rationale for the proposed regulatory action;
(2) the objectives of and legal basis for the proposed rule;
(3) a description and count of the small entities to which the
proposed rule will apply;
(4) a description of the projected reporting, recordkeeping and other
compliance requirements of the proposed rule; and
(5) an analysis of alternatives to the proposed rule.
The RFA can be incorporated into the Regulatory Impact Analysis. It is included
here as Appendix E.
1.0. QUESTIONS ADDRESSED IN THIS REPORT
The basic questions addressed in the RIA apply to any regulatory activity,
but are especially difficult to answer for USTs as compared to other pollution
sources for a number of reasons.
o First, there are likely to be many leaking USTs, but it is not known
which ones are leaking. Whether or not tanks leak, when they leak, and
how they leak is largely determined by chance.
o Second, even if we knew which tanks are leaking and how they are
leaking, the potential damages that result are highly variable and
largely affected by site-specific factors. The same leak can be
relatively unimportant in one situation, but can result in significant
damages in another situation.
o Third, even if tanks are not leaking now, they might leak in the future.
Evaluating regulatory options for potential future problems requires
V The Regulatory Flexibility Act (Public Law 96-354) §2(a), September
19, 1980.
2/ Ibid.
-------
1-5
the ability to make predictions about how tanks will leak, the damages
that will result, and the mitigating effects of regulatory options.
o Fourth, the dynamics of failure, release, detection, transport, damages,
and response are interrelated and complex. As a result, the need for
regulations, and the effectiveness and costs of different options are
best understood within a framework that ties these areas together in
a consistent way.
o Fifth, because of the interrelationships and complexities mentioned
above, requirements for new tanks, existing tanks, and corrective
action cannot be analyzed in isolation. Such components of a regulatory
program need to be packaged together and analyzed in an integrated
manner.
o Sixth, because there are so many underground tanks owned by small
businesses, almost any regulatory requirement can pose significant
concerns about total costs, economic impacts, and the feasibility of
implementation.
As a result, it is necessary to perform a number of analyses that must be
integrated together to address the questions posed in an RIA.
The key purpose of this report is to present results to facilitate the
analysis of technical requirements (i.e., requirements for prevention and detec-
tion of releases) for underground storage tanks, taking into account require-
ments for corrective action. Results are presented for costs, effectiveness,
and economic impacts of five regulatory options which have different technical
requirements, but the same corrective action component. This report also
describes the methodologies and assumptions used to obtain these results and
in addition addresses the sensitivity of the findings to changes in the assump-
tions .
I.E. SCOPE OF THIS ANALYSIS
This Regulatory Impact Analysis presents the results of several analyses
which investigated the costs, effectiveness, risks and economic impacts of
EPA's technical standards for IJSTs containing gasoline, or petroleum products.
Except for a screening analysis of economic impacts for industries operating
hazardous substances USTs presented in Chapter 6, the results presented here do
not necessarily apply to USTs used to store any substance other than gasoline
or petroleum products.- However, USTs used to store substances other than
petroleum products account for only 4% of the LIST population.
This Regulatory Impact Analysis presents, in detail, the costs, effective-
ness, risks and economic impacts of alternative requirements for preventing
leaks from occuring and detecting them more rapidly when they do occur. The
analysis also considers the effect of prevention and detection measures on the
costs of corrective action for plumes that still result. This RIA analyzes the
cost and economic impacts, but not the effectiveness, of the corrective action
measures that are proposed by EPA — the rationale for the corrective action
policy chosen is presented in the Preamble to the proposed rule.
-------
1-6
The impacts resulting from closure requirements are not highlighted in the
RIA, but have been included in the analysis in the sense that when tanks are
replaced, the replaced tanks are assumed to be closed in a manner consistent
with the proposed regulation (i.e., release detection and corrective action as
appropriate). The analysis of economic impacts also addresses these costs.
Recordkeeping costs are currently in the process of being developed and
are not analyzed in this RIA.
l.F. SOURCES OF INFORMATION
The information for this report was drawn from several sources:
1. The cost-effectiveness analysis was done by Sobotka A Company, Inc. (SCI)
and Pope-Reid Associates (PRA), using the UST Simulation Model.
2. The Risk analysis was conducted by ICF, Inc., using inputs from the UST
Simulation Model.
3. The economic analysis of retail petroleum USTs was prepared by Meridian
Research, Inc. with support from Versar, Inc. The economic screening analy-
sis of other industries owning USTs was done by SCI.
4. The benefits analysis was conducted by Research Triangle Institute and Glen
Anderson of the Office of Policy Analysis, U.S. Environmental Protection
Agency.
5. Other data sources:
o Analysis of the National Data Base of Underground Storage Tank Release
Incidents, Versar Inc., for the Office of Solid Waste, U.S. Environmental
Protection Agency.
o Underground Motor Fuel Storage Tanks: A National Survey, Office of
Pesticides and Toxic Substances, U.S. Environmental Protection Agency.
o Compliance Cost Calculations for EPA Regulation of Underground Storage
Tanks, Oata Resources, Inc. for Office of Solid Waste, U.S. Environmental
Protection Agency.
o Underground Storage Tanks Technical/Financial/Econornic Data Collection,
Data Resources, Inc. and Ouantum Analytics for the Office of Solid Waste,
U.S. Environmental Protection Agency.
l.G. ORGANIZATION AND CHAPTER SUMMARIES
This RIA discusses the current problem regarding underground storage tanks
and presents the results of several analyses of EPA's regulatory alternatives
and the proposed rule. Chapters one and two introduce the problem associated
with USTs and discuss the methodologies used to analyze the regulatory alterna-
tives. Chapter three discusses the current UST universe and outlines the "no
regulation" base case used in the analysis. Chapters four and five discuss the
-------
1-7
regulatory requirements, the integration of new and existing tank requirements
into the five regulatory options and present the cost and effectiveness analysis
of the five options. Chapter 6 presents the results of the economic impact
analysis and Chapter 7 discusses the benefits of the IJST regulatory options. A
summary comparison of the five regulatory options is presented in Chapter 8 and
the limitations of all the analyses are discussed in detail in Chapter 9.
A short summary of the contents of^each chapter follows:
1. INTRODUCTION: This chapter provides an overview of the leaking underground
storage tank problem by describing the current IJST universe, why USTs leak,
damages from UST releases, the range of possible responses to tank leaks, and
the current level of response. It also describes the purpose of a regulatory
impact analysis and the purpose of a regulatory flexibility analysis. Finally,
it provides a summary of the information sources used for this report.
2. OVERVIEW OF METHODOLOGICAL APPROACH: This chapter presents the key issues
addressed by the RIA and the approaches which are used to analyze these issues.
Key issues include: (1) need for the proposal; (2) alternative regulatory
approaches and their interrelationships; (3) regulatory costs and economic
impacts; (4) regulatory benefits; and (5) comparison of costs and benefits.
This chapter also describes the UST Model, the key tool used to evaluate the
UST regulatory options.
3. DEFINING THE BASE CASE: This chapter introduces the concept of a base case
and its role in the analysis. It describes the key factors that need to be
estimated in order to characterize the base case. These include the distri-
bution of USTs by use, type, age, location in relation to population and drink-
ing water, and hydrogeological setting. For the purpose of using the UST Model
to estimate costs and plumes associated with a given regulatory option, specific
assumptions must be made regarding tank types, a tank age distribution, and a
distribution of hydrogeological settings for USTs. Hydrogeological settings
are important because they affect the performance of leak detection equipment
and affect the characteristics of floating and aqueous plumes.
Current practices for leak detection and responses' to tank leaks also
need to be specified as part of the base case. Leak detection methods in the
base case are sensory detection and manual inventory control. These are simu-
lated in the UST Model through a base detection assumption. This chapter also
provides estimates of the portion of the tank population that is currently
leaking or will leak, and the implications in terms of plume area and costs.
4. COMPONENTS OF A REGULATORY STRATEGY: This chapter presents the alternatives
for tank construction, leak detection, phasing, and corrective action that can
be combined into regulatory options for new tanks and existing tanks. The
chapter is primarily descriptive in nature, though information is presented on
the advantages and disadvantages of each of the various components discussed.
5. COST-EFFECTIVENESS ANALYSIS: This chapter outlines the five integrated regu-
latory options analyzed in this RIA. These options are analyzed using the UST
Model. Effectiveness of a regulatory option is measured as contaminant plume
area that would have appeared under the base case, but is avoided under the
option (i.e., plume area avoided). This effectiveness measure is considered to
be a reasonable proxy for final damage measures. The major categories of costs
-------
1-8
include: (1) initial facility costs (includes the cost of obtaining and instal-
ling new tanks); (2) detection and monitoring costs; (3) value of product lost;
(4) tank removal and replacement, and pipe repairs; and (5) corrective action
costs. Cost element 1 applies only to new tanks. Cost element 2 applies to
all tanks. Cost elements 3, 4, and 5 apply only to leaking tanks. All costs
are based on best available engineering estimates. Only cost elements affected
by the proposed rules are analyzed.
Cost-effectiveness is shown two ways: (1) without corrective action costs
included; and (2) with corrective action costs included as part of the regula-
tory costs. Furthermore, regulatory costs with corrective action are presented
relative to a base case without any corrective action, as well as relative to a
base case with corrective action.
6. ECONOMIC IMPACTS: Economic impacts are analyzed for three UST-using sectors:
(1) facilities using USTs for storing motor fuels for the retail market; (2)
facilities using USTs for storing fuels for non-retail purposes; and (3) facil-
ities using USTs for storing hazardous substances. The analysis concentrated
on the retail motor fuel sector because that is where economic effects are
likely to be most significant. Economic impacts were assessed using a return
on assets approach.
Under assumptions of no revenue increase per facility, limited ability of
small firms to get loans to cover compliance related costs, limited ability of
firms to get insurance for releases, and no corrective action in the base case,
the burden on small firms could be significant, with most of the exit being
attributable to corrective action costs. A sensitivity analysis is used to
explore the effect of relaxing some of these assumptions. For example, under a
3% revenue increase, large firms are not adversely affected, though some small
firms are still projected to exit within the first five years because of the
corrective action requirement. A Regulatory Flexibility Analysis is included
as Appendi x E.
7. BENEFITS ANALYSIS: Forty-four case studies of UST release incident damages
were used to identify the types of damages caused by leaking USTs and place
monetary values on some of the major types of damage. Damages were monetized
by estimating the actual sums spent to repair damages caused by releases (e.g.,
replacing contaminated wells, compensation for damaged structures) plus the
losses to businesses closed or disrupted by the releases and the contamination
they caused.
Riven a distribution of monetary damages associated with serious release
incidents, total damages are calculated using estimates of the number and size
of leaks under the base case, the proposed rule, and various regulatory options.
These estimates are made using the UST Model. Data from the Release Incident
Survey provide estimates of frequency with which releases of various sizes are
likely to have serious consequences (e.g., lost drinking water resources , or
vapor contamination). UST Model predictions of the timing of the damage inci-
dents are used to discount the monetary damages back to the implementation time
of the regulations, providing estimates of the present value of the damages.
The risks avoided by the proposed rule were not considered in the same
framework discussed above, due to the problem of monetizing health risks.
Instead, the UST Model's outputs, in terms of predicted frequencies and
-------
1-9
magnitudes of contamination incidents, were used as inputs into a separate
analysis which examines concentrations of carcinogens at various possible
exposure points, the size of the exposed populations, and the duration of
exposures.
8. IMPLEMENTATION CONCERNS AND COMPARISON OF REGULATORY OPTIONS: This chapter
presents a summary comparison of EPA1s five UST regulatory options. This
comparison draws on the information on costs, effectiveness, economic impacts
and benefits presented in previous chapters. Five exhibits present summary
data from previous chapters.
9. LIMITATIONS OF THE ANALYSIS: This chapter lays out the uncertainties in-
herent in the analyses. With respect to the UST Model, these include uncer-
tainties due to data inputs, as well as uncertainties in the modeling process.
There are also uncertainties in the base case specification, assumed unit
costs, population of affected USTs, economic status of affected parties, and
benefits assessment. These uncertainties combine to imply that a regulatory
option revealed to have a modest advantage in the analysis may not necessarily
the best. However, comparative results between two options generally carry
more confidence than absolute estimates associated with any given option.
APPENDICES:
Appendix A: Summary of the UST Model
Appendix B: Zip Code Analysis of UST Location and Population Density
Appendix C: UST Model Specifications
Appendix D: Economic Impacts Assessment Methodology
Appendix E: Regulatory Flexibility Analysis
Appendix F: Methodology for Estimating Risk
-------
Chapter 2
OVERVIEW OF METHODOLOGICAL APPROACH
This chapter provides an overview of the types of information and analyses
that are encompassed in this RIA. It presents the basic questions to be addres-
sed in the RIA and sets the context for how these questions will be answered
with respect to the UST issue. It then presents a brief description of the LIST
Model, the key tool used to assess the cost and effectiveness of UST regulatory
alternati ves.
2.A. COMPONENTS OF AN RIA
The basic questions addressed by an RIA are:
1. What is the need?
2. What is the proposal and what are its consequences?
3. What are the alternative approaches?
4. What are the costs?
5. What are the benefits?
6. How do the costs and benefits compare for the alternatives?
All of these questions are interrelated and depend on much of the same underly-
ing data and analyses. For example, establishing the need for federal interven-
tion and evaluating the benefits of alternatives is largely dependent on the
same type of information and analysis. The overall purpose of an RIA is to
pull together the various data and analyses in a consistent and comprehensive
way, with the goal of highlighting the trade-offs for the alternatives under
consideration.
The following discussion provides an overview of the types of information
and analyses that are encompassed by this RIA. The discussion is organized
according to the basic questions addressed by an RIA. Generally, we begin by
discussing the types of information needed, and then identifying how different
types of analyses contribute toward providing the needed information. Descrip-
tions of how these analyses are performed are discussed elsewhere in this
report. Exhibit 2.1 provides an overview of the analyses referred to in the
discussion that follows.
2.A.I. Need for the Proposal
The beginning of an RIA is devoted toward establishing the need for federal
intervention. This need is generally established by developing an assessment
of the damages that would occur in the absence of federal intervention.
This needs assessment is often referred to as the "base case," because it
also provides a reference point for determining the consequences of the alter-
-------
Exhi 2.1
INFORMATION AND ANALYSIS FOR UST RIA
-------
2-3
natives being considered. Ideally, the base case includes consideration of the
effects of state and local requirements, and of changes in industry practice
that would prevail in the absence of federal requirements. In practice, this
is difficult to determine. Once the need is established, the bulk of the
remainder of the RIA is oriented toward evaluating the consequences of alter-
natives for federal intervention relative to the base case.
2.A.l.a. Types of Damages from UST Releases
The need for regulatory intervention is best described in terms of the
types of damages that might occur as a result of leaking underground storage
tanks. The discussion here is divided into damages that can occur prior to
detecting leaks, and those damages that may occur after the leak is detected.
This division is useful because it corresponds to different types of regulatory
responses that become effective either before detection (technical standards and
financial responsibility) or after detection (corrective action).
Before Detection
Some of the types of damages that result from leaks, such as the cost
of lost product, are borne directly by the owners of the tanks. In addition,
leaks may cause foundation water-proofing to break down, foul sump and drainage
systems, or damage electrical conduits, buried cables, piping or pumps. Leaked
product may seep into basements or foundations, creating unpleasant or toxic
vapors. These vapors may pose a health risk to workers and a potential fire or
explosion hazard. Product residues in soil may volatilize and pose a potential
health risk to workers and customers.
Damages may also occur beyond the facility, as the released product travels
on or through the ground water, or by other routes such as sewer lines. Some of
the potential damages are similar to those occuring on-site, such as materials
damages, the risks of fire and explosion, and health effects associated with
the volatilization of product. Additional types of damages include potential
health risks to those who drink ground water or surface water that has been
contaminated. The magnitude of these effects will naturally depend on the
proximity of the facility to houses, businesses, and underground utility cables,
and whether the contaminated water is used for drinking or other purposes.
In addition to these off-site effects, the leaked product may contaminate
soil, causing loss of crops and soil productivity, and may also affect plant
and animal life. Ground water that is used for irrigation or industrial purposes
may also be affected; users of such water may find yields decreasing or adverse
effects on the quality of the products they produce. There may also be some
nuisance effects involving taste, odor or visual sensess (such as oil slicks)
as the release becomes detectable.
A final variety of off-site effect that may occur is referred to as option
value damage. Option value is a value placed on a resource because of its
potential to be used in some (possibly unforeseen) way in the future. Even
though a resource (such as ground water downgradient from the UST) is not cur-
rently being used, there may still be economic value to protecting it in an
uncontaminated condition because of the possibility that it will be used in the
future.
-------
2-4
After Detection
After detection, all of these types of damages may continue to occur. In
addition, anxiety may be another effect of the leak. Nearby residents may
fear the contamination of their water supply or fear an explosion; even if the
tangible harm does not occur, there has been some damage to these individuals'
welfare. As a result, nearby property values may fall. ]_/
Once the leak has been detected and stopped, there is a choice of doing
nothing and knowingly accepting further damages, or taking steps to mitigate
future damages. Of course, such mitigating steps have costs associated with
them. Presumably, the only mitigating measures that should be undertaken are
those that confer benefits (avoided future damages) in excess of the costs of
the mitigating measures. In these cases, the cost of the mitigating measures
becomes a measure of damages; care should be taken not to count both the avoided
damages and the cost of avoiding them, as this would be double-counting. The
mitigating measures can be undertaken singly or in various combinations, and
include: degrees of corrective action (removing the contaminated soil at the
tank, removing the floating plume, and removing the dispersed plume), treating
contaminated water prior to use, and using alternative water sources.
Damages are Variable
The incidence and severity of problems in the base case range widely.
Generally, whether or not a tank is now leaking, whether it will leak in the
future, and the leak characteristics depend on several factors including the
characteristics of the tank, the soil setting, and chance. Moreover, the
damages that result from these leaks are highly dependent on site-specific fac-
tors. Even for two leaks with exactly the same leak rate and duration, the
resulting damages can vary greatly and depend on a number of factors including:
the hydrogeologic setting, other avenues of transport, the proximity of people,
structures and surface water, and the current and future uses of the contaminated
water.
2.A.l.b. Type of Analysis
To determine the damages resulting from the base case, we need information
and analysis in several areas. First, we need to evaluate the existing and
future tank population to estimate which tanks will fail, when they will fail,
and how they will fail. Second, we need to estimate the types of damages that
will result due to these expected failures. Third, this evaluation must incor-
porate consideration of the measures that are already being taken or are expec-
ted to he taken as a result of prior requirements (such as the statutory interim
prohibition for new tanks) or by requirements of state and local authorities,
and by industry. The alternatives considered for federal action should not be
burdened by the costs that are borne as a result of measures already taken,
V A change in the market value of nearby properties as a result of
potential ground water contamination from leaking IJSTs is a pecuniary effect,
and not an additional damage itself; instead, it is a market reflection of the
perceived damages that may occur.
-------
2-5
nor have attributed to them the benefits that have already been captured by
these measures.
We evaluated the likelihood that the existing tank population is leaking,
the characteristics of the leaks, and the resulting damages in several ways.
First, we looked at past experience using the State Release Incident Survey.
Approaches that relied on reported information help characterize leaks that
are known and the resulting damages, but do not necessarily represent ongoing
leaks that have not been detected or reported, or that have not yet occurred.
Second, we looked at current experience by examining the National Motor Fuel
Tank Survey. Survey approaches attempt to identify undetected leaks and help
us assess the extent to which tanks are now leaking without our knowledge, and
the severity of these leaks. Third, we examined case studies in depth to help
provide perspective on a variety of aspects of the problem, including the
damages that result and ways in which to evaluate damages. Fourth, we simulated
current and future tank system failures through modeling methods such as the
UST Model (discussed later in this Chapter) that incorporate our best under-
standing of the factors that lead to leaks, determine leak characteristics, and
determine the damages that result. The UST Model simulates potential damages in
terms of plume size and duration, and these results may be interpreted to
determine the total resulting damages. Such a simulation could characterize
undetected and future leaks, which are hard to predict using only data about
known leaks. The UST Model is especially useful for identifying the effect of
regulatory alternatives because it predicts plume size and duration under the
base case as well as under the regulatory alternatives.
2.A.2. What Are the Alternative Approaches?
2.A.2.a. Types of Approaches
Technical Standards
Damages can be avoided by preventing leaks in the first place, or damages
can be reduced by detecting leaks earlier. To a large extent, these choices
can be thought of in terms of requirements for new tanks versus requirements
for existing tanks.
Different types of tank systems fail very differently. Corrosion is a
principal cause of leaks in bare steel tanks. If tanks are corrosion resistant,
they are much less likely to leak. If tanks also have some form of secondary
containment, they are even less likely to have releases into the environment.
To the extent that new tank requirements result in fewer failures, more leaks
will be prevented in the future. Since much of the existing bare steel tank
population is nearing retirement, new tank standards may have a significant
effect on the potential problems remaining from USTs in a decade or so.
Early replacement of tanks might also prevent leaks from occuring, or
might result in detecting leaks that are under way. In determining the
desirability of early replacement, it is important to assess the likelihood
that older tanks are more likely to leak.
-------
2-6
Early detection can reduce the damages that result from leaking tanks,
whether they are leaks from existing tanks, or leaks from new tanks that
meet new tank standards. There are a wide range of continuous and periodic
detection options that can reduce damages by detecting leaks sooner. All of
these methods have different levels of sensitivity and reliability. The
contribution of any detection method is more significant for tanks that are
likely to leak than for tanks that tend not to Teak as much. The relative
performance of the detection method may depend strongly on the character-
istics of the leaks which, in turn, can also depend on the type of tank. Some
detection methods depend directly on the leak rate, others are affected by the
accumulated leak volume, while other detection methods may not detect leaks
well that occur in particular locations in the tank system. In addition, some
methods perform independently of the environmental setting, while the perfor-
mance of other methods depends on the characteristics of the unsaturated or
unsaturated zone.
Interrelationship of Prevention and Detection
Generally, prevention and earlier detection are substitutes for one another
in avoiding the same damages. Therefore, it is important to avoid douhle count-
ing when assessing the benefits associated with prevention and detection. Once
preventive measures are undertaken, the value of additional detection measures
is reduced. Similarly, very effective detection measures can avoid much of the
damages, reducing the value of requiring the use of tank systems that are less
likely to fail. In addition, different combinations of tank systems and detect-
ion methods can achieve similar reductions in damages, but possibly at much
different costs to the tank owner.
Identifying and evaluating the trad-eoffs regarding effectiveness and cost
for different combinations of prevention and detection can be difficult given
the inherent complex interrelationships and variability that exist between tank
failure, hydrogeologic setting, and exposure setting.
Interrelationship of Technical Standards and Corrective Action
Technical standards and corrective action can be substitutes for one
another in avoiding some of the same damages. Technical standards—prevention
and earlier detection--can reduce the cost of response once the leak is
discovered, because technical standards would result in fewer, and smaller
plumes that can be less expensive to clean up; smaller plumes could also reduce
the need for treating water supplies or obtaining alternative water sources.
Conversely, it is not necessary to rely on the technical standards alone to
avoid damages that result from leaking IISTs, because response measures may be
able to achieve some of these benefits also. Ideally, alternatives for preven-
tion, detection and for corrective action should be developed and considered
together in order to avoid double counting benefits, and to make the most
cost-effective use of all three together.
-------
2-7
2.A.2.b. Types of Analysis
Analyzing the regulatory alternatives is far more difficult than developing
and understanding the base case because it is necessary to project how the
alternatives will alter costs and benefits from the base case. This is espe-
cially complex for technical standards for USTs because of the role of chance
in determining tank failure and leak characteristics, the complex dynamics that
determine the effectiveness of alternative detection methods, the variability
in plume characteristics due to hydrogeologic factors, and the variability of
the damages that result due to variability in such factors as the proximity of
people and the uses of the contaminated ground water.
To assist in our analysis, we have developed the UST Simulation Model which
estimates the effects of regulatory alternatives using current information and
consistent assumptions. The outputs of the UST model are then used as inputs
into cost and effectiveness analyses, an economic impact analysis, and the
benefits analysis. These results are then systematically presented in this RIA
in combination with all other relevant studies to clearly present the trade-offs
associated with the alternatives. Because of the central role of the UST Model
in integrating analyses and providing key inputs into follow-on analyses, the
UST Model is briefly described later in this chapter. More complete descriptions
are provided in Appendix A and in the UST Model Documentation. \j
We currently use the UST Model to analyze the effects of the regulatory
options on gasoline-containing USTs. Hazardous substance-containing USTs are
not currently modeled because they represent only 4% of the UST universe.
Modeling USTs that contain hazardous substances is a more complex endeavor and
requires far greater data collection efforts because these USTs could theoret-
ically contain any of hundreds of hazardous substances, each requiring its own
set of modeling parameters. In essence, the 4% of hazardous substance-contain-
ing USTs are modeled as gasoline USTs for purposes of estimating the total cost
and effectiveness for a given option.
2.A.3. What Are the Costs?
There are several different classes of costs to consider in an RIA. These
can be divided into direct costs, economic impacts, and implementation costs.
2.A.3.a. Direct Costs
The direct costs of regulatory alternatives are those costs which can be
attributed to the alternatives. These costs include the costs associated with
installation and operation of detection methods, the costs associated with
meeting new tank requirements, and the costs associated with related items such
as closing a facility for tank testing (which may result in forgone profits).
V Pope-Reid Associates, Inc., Final Report: Underground Storage Tank
Model, December 1986.
-------
2-8
For requirements leading to early retirement, direct costs include the
costs associated with forgoing the use of the existing tank. To assess this, it
is necessary to estimate how long the tank would have been in service, which in
turn may depend on when the tank is expected to fail.
The cost estimates should include both the total costs and the incremental
costs above the base case costs, that would have been incurred in the absence
of the regulatory alternative. However, it is not always clear whether a
certain activity should be included in the base case or be attributed to the
proposed regulation. For example, current state requirements or industry prac-
tice might suggest that a certain activity would take place even in the absence
of regulatory change. Yet the state requirements or industry practice may have
been developed in anticipation of federal requirements. When it is not clear
whether a certain activity should be included in the base case or attributed to
the regulation, it will be attributed to the regulation so as not to understate
regulatory costs.
In addition, the regulatory alternatives may result in some savings over
the base case. For example, more expensive tank systems that fail less fre-
quently might result in less costs for repair and replacement than would occur
in the base case. Other savings include reduced product loss or reductions in
costs of responding to leaks. Care must be taken to avoid double counting
savings, by ensuring that either credit is given as cost savings or as damage
reductions, but not both.
Actual data and estimates from best engineering judgement were used to
establish unit costs for alternative detection and tank requirements. These
were applied to the estimated tank population to develop estimates of total and
incremental costs for the tank population as a whole.
Similarly, the UST Model incorporates these estimates as inputs for estima-
ting direct costs for the groups of tanks being analyzed. The UST Model also
estimates the expected remaining useful lives of existing tanks of different
types and ages, and can be used to estimate the costs associated with early
retirement of existing tanks. Since the IJST Model also tracks the cost of
product loss, repairs and replacement, it can be used to estimate the incremen-
tal savings over the base case for regulatory alternatives.
2.A.3.b. Economic Impacts
The economic impacts of the direct costs depend on their magnitude and
on who bears them: they may be passed forward to consumers in the form of price
increases, absorbed by the owners or operators, or passed backward to production
factors. To the extent that closures occur when costs are absorbed by owners,
unemployment may result. Other types of economic impacts include impacts on
competition, product quality, productivity, and innovation. Concerns about
economic impacts may be high because there are a large number of tank establish-
ments, many of whom are small businesses, that may be sensitive to even rela-
tively small incremental costs due to the regulatory alternatives.
-------
2-9
The economic impact analysis, presented in Chapter 6, uses cost inputs
from engineering analysis and the UST Model as a basis for evaluating potential
impacts of technical requirements. The key economic impact of concern is the
potential effect on the viability of affected entities. A return on assets
approach is used to predict the effect of regulatory alternatives on viability.
Return on assets was chosen because the rate of expected return on assets is
reasonably consistent across industries and size classes and there are docu-
mented benchmarks which correspond to likely failure and severe financial
distress. The economic impact analysis concentrates on the retail motor fuel
sector because this is where economic impacts are likely to be most significant.
This is the case because in this sector there are no substitutes for USTs and
there are many small businesses. For other USTs (i.e., non-retail fuel and
hazardous substances), a less detailed screening analysis is presented.
Because of the concern for the potential effect on small businesses, a
Regulatory Flexibility Analysis, as required by the Regulatory Flexibility Act
of 1980, has been conducted and is included as Appendix E to this report.
2.A.3.C. Implementation Costs
The cost and feasibility of implementation, both to regulators and to the
regulated community, can significantly affect the relative desirability of the
alternatives under consideration. There are numerous regulatory choices that
can significantly affect the cost and feasibility of implementation for the
overall UST program.
For example, the timing and scope of detection requirements can signifi-
cantly affect the timing and magnitude of demands on firms providing corrective
action and closure services, as well as affect the government entities overseeing
corrective action and closure. If, for example, a large percent of existing
tanks are leaking, an immediate requirement for testing would result in signifi-
cant immediate demands for corrective action. Prevention-related requirements
have the opposite effect. These requirements can significantly reduce demands
on the corrective action program in the long run by reducing the number and
size of releases. Detection-related requirements could also reduce demands on
the corrective action program and the private-sector resources it employes,
but would result in more strain on detection equipment manufacturing and
operating capacity, and relatively larger government resource requirements for
ensuring compliance.
Various analyses can help shed light on the cost and feasibility of imple-
mentation. The demands on various aspects of implementation can be assessed
from the same types of analyses used to analyze costs and benefits. Additional
studies, such as an analysis of industry capacity for supplying tanks and
detection services, and an assessment of the capabilities of state programs,
provide additional perspective on feasibility. It is therefore important to
evaluate implementation trade-offs along with all of the other trade-offs asso-
ciated with the different alternatives under consideration. At a minimum, it
is helpful to evaluate regulatory alternatives relative to each other from an
implementation perspective. That is, rather than trying to estimate the abso-
lute magnitude of implementation costs for each regulatory alternative being
considered, it is often sufficient to rank remaining alternatives. Such an
analysis is presented in Chapter 8.
-------
2-10
2.A.4. What Are the Benefits?
2.A.4.a. Types of Benefits
The benefits of the alternatives are the damages avoided which were men
tioned earlier in the discussion of the need for the proposal.
2.A.4.b. Types of Analysis
The UST Model provides a systematic basis for evaluating how different
regulatory options would be expected to reduce the damages that motivate regu-
lation. The IJST Model provides information about the performance of different
combinations of tank types and detection methods in different hydrogeological
settings. The cost outputs include the costs associated with product loss,
tank repair and replacement, and corrective action. Measures of damages include
plume areas and plume durations that result for the leaks that occur during the
period that the leaks remain undetected. These estimates can be used as start-
ing points for additional analyses that estimate the health risks and other
damages that may result. Chapter 7 addresses such valuation issues.
2.A.5. How Do the Costs and Benefits Compare?
2.A.5.a. Types of Information
As previously shown in Exhibit 2.1, a number of studies provide key infor-
mation to help characterize the base case. These studies, in conjunction with
engineering analyses, provide inputs into the UST Model. The UST Model provides
a consistent and systematic basis for analyzing the effects of the regulatory
alternatives in terms of changes in different types of costs, and changes in
plume characteristics. The outputs of the Model are then used in the economic
analysis and benefits analysis. All of these results then must be compared in
ways that facilitate making a decision to select one of the regulatory alterna-
tives, or to develop a new one.
2.A.5.b. Types of Analysis
The cost-effectiveness analysis provides a framework for comparing the
trade-offs associated with the regulatory alternatives. Comparisons are stra-
ight forward when the information is quantitative--for example:
o How different alternatives compare in terms of cost per acre of plume
avoided, or cost per case of statistically expected cancer avoided.
o Whether costs can be substantially reduced, without greatly reducing
benefits, by tailoring requirements to be less stringent where damages
are not expected to be high anyway; or conversely, whether benefits may
greatly increase, with limited increases in cost, by tailoring require-
ments to be more stringent where the damages are expected to be high.
-------
2-11
However, there are many factors that are difficult to incorporate into
this simple cost-effectiveness framework because we cannot place dollar values
on them. These include damages that we know about but cannot fully quantify
and business failures expected to result from the proposed rule. In addition,
other practical concerns, such as implementation, can significantly affect
regulatory choices. Finally, the overall evaluation of whether the benefits of
the proposal exceed' its costs cannot be reduced to a quantitative expression,
but rather is a judgemental balancing decision informed by the RIA and other
sources of information.
2.B. OVERVIEW OF THE UST MODEL
This section provides an overview of the UST Model. A more detailed
description of the inputs and outputs of the model and how the model works is
included in Appendix A, or can be found in Pope-Reid Associate's final report
of the UST Model documentation. }_/
2.R.I. Summary Description of EPA's UST Model
EPA's UST Model is based on a detailed specification of all of the ways in
which different types of tank systems can fail in different environmental set-
tings, the likelihood that these failures will occur over time, and the charac-
teristics of the leaks that result from particular failures. Based on these
specifications, we can simulate the problems that occur month-by-month, allowing
chance to operate during the life of a particular type of tank in a particular
setting. The tank may not leak, or it may leak more than once. For the leaks
that occur, the model simulates the movement of the product through the unsat-
urated zone and the development of the plume until the leak is discovered. The
timing of the discovery of the leak can depend on the point of detection, and
the sensitivity, frequency and reliability of the detection measures that are
assumed to be used. We can also repair tank systems or replace them when the
leaks are discovered, undertake corrective action that is based on the size of
the plumes that result from the leaks, and keep track of all of the costs that
are incurred. Exhibit 2.2 provides a simple flow diagram that illustrates the
sequence of steps undertaken in the model.
By repeating the simulation of individual tanks many times, we can develop
an understanding of what happens to a population of identical tanks in the same
situation. We can do this analysis in environmental settings that have different
soil characteristics, different depths to ground water, and different ground-
water velocities to see how these parameters affect failure and leak character-
istics. We can see how the same spectrum of leaks results in different plume
characteristics in different hydrogeological settings. We can also see how this
all compares for different types of tank systems in combination with different
detection options. We can scale the results to represent the actual tank popu-
lation, based on our understanding of the distribution of tanks across hydro-
geological and exposure settings.
The outputs of the model can then be used as inputs into an exposure
analysis that translates plumes and their durations into risks, an economic
Model-/'Dl££8l3lP1i,9&§sociates' Inc,» Final Report: Underground Storage Tank
-------
2-12
Exhibit 2.2
SIMPLE FLOWCHART OF UST SIMULATION MODEL
-------
2-13
analysis that translates costs and cost savings into economic impacts, and
an implementation analysis that translates information about the performance of
regulatory options into demands on different aspects of implementation.
2.B.2. Limitations
This type of approach requires a great deal of information. To a large
extent, information about underground storage tanks is not available. Some
data, such as the number and type of tanks being used, has been collected in
the past year or two. EPA's Release Incident Study £/ and Motor Fuel Tanks
Survey^/ have provided some information on the types of releases, the extent
of damages and the percentage of tanks that may be leaking. Other information
needed for this analysis, such as data regarding tank performance and the per-
formance of leak detection equipment, is not yet available. Many tank types
and most leak monitoring equipment are relatively new, and therefore the perform-
ance of these tanks and equipment has not yet been documented. Because much
of the information needed to do this analysis is limited or unavailable, assump-
tions based upon the limited data available and best engineering judgement must
be used in place of data. To the extent that there is uncertainty in the
assumptions, there is uncertainty in the analysis and in the results of the
analysis. However, any comprehensive analysis that attempts to accomplish the
same results will face this same limitation.
At present, we use the UST Model to simulate USTs containing motor fuel.
Tanks storing motor fuels represent 96% of the UST universe. The remaining 4%
of the UST universe consists of USTs containing hazardous substances. These
hazardous substances can be any substance designated as hazardous under CERCLA
§101(14), other than hazardous wastes. In order to accurately model this 4% of
the UST universe, it will be necessary to make assumptions about the distribu-
tion of hazardous susbstances over USTs, the fate and transport of hazardous
substances, the age distributions and hydrogeological settings for hazardous
substance-containing USTs, the level of base detection, and other parameters.
Sufficient data is not yet available to take on this extremely complex task.
Therefore, the analysis is currently conducted by assuming all 1.4 million USTs
in the UST universe contain motor fuel.
One advantage of the UST model is that it requires that assumptions and
data be used in an explicit and consistent manner. If there is uncertainty
about particular assumptions used in the analysis, the model can be used to
conduct sensitivity analyses to make the importance of that uncertainty clear.
In addition, given the complex interrelationships that are inherent in the UST
situation, the model results can identify the significance of the assumptions
that otherwise might seem less important when considered in isolation. Finally,
the model can easily be revised or calibrated to reflect data as it becomes
available. As such, the UST Model is a means to fully exploit available data,
other models, and judgement in a systematic, comprehensive and controlled way.
V EPA Office of Underground Storage Tanks, Summary of State Reports on
Releases from Underground Storage Tanks, August 1986.
£/ Office of Pesticides and Toxic Substances, U.S. Environmental Protection
Agency, Underground Motor Fuel Storage Tanks: A National Survey, May 1, 1986.
-------
2-14
2.B.3. Model Outputs
The model provides specific sets of outputs for a given type of tank and
detection method in a given hydrogeological setting. Each time the model is
used, we can tailor or change the assumptions regarding the tank type, detection
method and the hydrogeol ogical setting to reflect information known about
the DST universe. The outputs that result from each of the tailored model runs
become key inputs to other analyses such as cost and effectiveness, benefits,
economic impacts and financial assurance. These outputs include measures of
effectiveness such as the frequency of release incidents, release rates, the
distribution of release volumes for a given number of tanks over a given period
of time, and the time it takes to detect a release given specific hydrogeologic
settings, specific leak monitoring equipment and the distribution of plume
areas. The model also provides total discounted costs and yearly costs for
equipment installation and operation, equipment repairs and replacements,
product lost, and corrective action. Further explanation of the model outputs
as well as examples of how these outputs are incorporated into different analyses
are provided in the remaining chapters of this report.
-------
Chapter 3
DESCRIPTION OF THE PROBLEM: DEFINING THE BASE CASE
3.A. INTRODUCTION: THE CONCEPT OF A RASE CASE AND ITS ROLE IN THE ANALYSIS
When evaluating the benefits anrl costs of potential technical standards,
it is necessary to compare them to the benefits and costs that would have
prevailed in the absence of the new requirements. This "base case" describes
the current universe and practices in the absence of any regulation, and
establishes a reference point for the technical standards. The selection of
the base case can have a significant effect on the performance of options
relative to an alternative of no regulation; however, the base case does not
affect the performance of options relative to one another.
Often, it is not clear what the base case should be. For example, the
base case for new tank requirements may be corrosion protected tanks with min-
imal operational requirements (base inventory control) since this is the current
requirement under the interim prohibition. On the other hand, if state regula-
tions require leak detection for new tanks, then it may be more appropriate
to establish a different base case for evaluating the incremental costs and
benefits of federal requirements. Once the base case is established for a
given set of regulatory requirements, it can be used throughout the analysis
as a representation of the regulated universe in the absence of any regulation
and as a reference point for comparing the results and impacts of the regula-
tory alternatives.
The first sections of this chapter identify those elements of the current
petroleum tank universe that are key to the establishment of the base case and
describe the current knowledge of these factors. The last part of the chapter
describes how the base case was modeled using EPA's UST Simulation Model and
how the base case can be used in evaluating the regulatory alternatives.
3.B. OVERVIEW OF KEY FACTORS
The country has a large number of underground storage tanks, many of which
are old and are not protected against corrosion. They are found near where
people live and work, and to a somewhat lesser extent, they are found near
drinking water sources. If an UST is leaking, the released product can move
through a chain of pathways to affect water sources, air, and structures ad-
versely and the resulting damages to health, property, and the environment can
be costly or impossible to avert or repair. Thus, USTs could potentially pose
serious problems, which could be addressed with a regulatory program.
The following sections provide background information on the key factors
involved in establishing the base case for the underground storage tank regula-
tory analysis. Section 3.G. outlines the established base case and character-
izes the base case in terms of the key factors.
-------
3-2
3.B.I. Why Hoes a Tank Leak?
A large proportion of the current population of underground storage tanks
are constructed of bare, unprotected steel. The average age of bare steel
tanks is between 15 and 20 years .V When steel is exposed to the natural
environment and the elements, it corrodes. If a steel tank is left unprotected
and exposed to natural conditions for many years, corrosion takes its toll, holes
form in the walls of the tank and the tank may leak.
3.8.2. What Happens When a Tank Leaks?
Products released from USTs generally cause damage to property or to health
only after travelling, via a very complex set of physical and chemical processes,
from the leaking system to a receptor population. Transport can occur in the
liquid or vapor phase, as free product or dissolved in ground or surface water,
in environments that may or may not be conducive to volatilization and degrada-
tion. Transport mechanisms can differ radically for different types of product.
For these reasons, it is impossible to describe comprehensively the fate and
transport of leaked products in the limited space available. We can, however,
outline a typical course for the most common type of leak, gasoline, as follows.
When a leak begins, gasoline flows downward, first along the side of the
tank, then straight down through the backfill material and then through the
unsaturated zone. Depending on the volume released, the gasoline may reach the
capillary zone just above the surface of the water table. Because it is less
dense than water, gasoline floats on the ground-water surface, spreading out to
form a "floating plume." Water-soluble compounds from the floating plume, such
as benzene, then begin to dissolve into the ground water to form a dispersed
plume that lengthens in the direction of ground-water movement and widens
slowly. At the same time, some components of the gasoline will volatilize from
the spill wherever it is in contact with air in the pore spaces of the soil or
rock. The volatilized product will diffuse through the air in the soil, and
may or may not reach the ground surface in noticeable amounts. Some components
may also begin to degrade, or become adsorbed to the soil.
The characteristics of the floating and dispersed plumes are dependent on
the hydrogeological setting. In finer grained soils, for example, the floating
plume may be thicker and smaller in area. In regions of rapid ground-water flow,
the dispersed plume will also move more quickly, and the floating plume may be-
come elongated or have a greater surface area. In some types of aquifers,
depending upon the chemical makeup of the aquifer media, the dispersed plume
may be greatly retarded in comparison to the flow of ground water while in
other aquifers a plume containing the same product may be retarded very little.
At some point, the dispersed plume may begin discharging to surface water.
It may also reach a zone of low permeability at the lower limit of the surface
aquifer, and slowly penetrate that layer to reach a deeper, confined aquifer.
Eventually, much of the plume will volatilize: through the soil, from
surface water, from water used for drinking, showering, irrigation, or manu-
facturing, or even during corrective action if air strippers are used without
-------
3-3
incineration. These airborne contaminants could reach large numbers of people,
though at very low concentrations and after long periods of time.
3.B.3. Factors to Evaluate
In order to determine the need to regulate underground storage tanks and
evaluate the regulatory alternatives, a "no regulation" base case that closely
reflects the current state of the regulated universe must first be established.
To accurately evaluate how effective a regulatory alternative may be in prevent-
ing tank leaks or in mitigating damages from tanks that may already be leaking,
we need a firm understanding of several key factors regarding the current tank
universe. We need to know how many tanks may currently be leaking, and we need
to be able to estimate the probability of existing tanks leaking in the near
future. We know that bare steel tanks corrode. The greater the number of bare
steel tanks and the greater the average age of underground tanks, the greater
the magnitude of possible problems from leaking tanks. Therefore, we need to
know the size and the age distribution of the current population of tanks and
the proportion of the current population that is bare steel tanks. The extent
of risk or damages from leaking tanks depends upon the hydrogeological setting
surrounding underground tanks, the proximity of tanks to ground-water wells,
and the current state of leak detection and monitoring practices.
Extensive research has been undertaken by EPA to develop a better under-
standing of the magnitudes of tank populations, leaks, exposures, risks, and
damages. Sections 3.C. through 3.F. describe the current knowledge of several
key factors of the underground storage tank universe. Each of these factors
plays a key role in the establishment of the base case for evaluating regulatory
alternati ves.
3.C. THE TANK POPULATION
Underground storage tanks covered by the regulations are found in large
numbers in many sectors of the economy, and total about 1.4 million. V The
vast majority of these (shown in Exhibit 3.1) are used to store motor fuels and
other petroleum products, with only a few percent storing chemicals. More than
half of the petroleum product tanks are used by gas stations and other motor
fuel retailers; the rest are spread over a spectrum of industries as shown in
Exhibit 3.2.
A major portion of the current population of petroleum underground storage
tanks are made of bare (unprotected) steel. A minority of-tanks are of
corrosion-resistant steel or of noncorroding fiberglass, as shown in Exhibit
3.3. Bare steel tanks corrode, and are commonly considered to have an expected
life of between fifteen and twenty years. By this measure, a substantial
portion of existing tanks are near the end of their useful lives, as shown in
Exhibit 3.4. Some 40,000 petroleum tanks are replaced annually, either because
they are discovered to have failed or through upgrading programs. £/ Still other
}_/ EPA, Regulation of Underground Storage Tanks, Preamble and Proposed
Regulations, Draft, November 24, 1986, p. 10.
2/ SCI estimate based upon The Steel Tank Institute's estimates of total
tank replacements.
-------
UNDERGROUND STORAGE TANK POPULATION BY USE
CHEMICAL STORAGE (4.0%)
RETAIL FUEL SALES
(48.0%)
PETROLEUM PRODUCT
SALES
(48.0%)
x
_J,
O"
c+
co
CO
I
4*
Source: EPA Office of Solid Waste, Summary of State Reports on Releases
from Underground Storage Tanks, August 1986.
-------
3-5
EXHIBIT 3.2
UNDERGROUND STORAGE TANKS BY INDUSTRY SECTOR
NUMBER OF
INDUSTRY SECTOR TANKS
RETAIL MOTOR FUEL SALES 695,000
PETROLEUM PRODUCT STORAGE:
Agriculture 86,000
Mining 14,000
Construction 42,000
Manufacturing 75,000
Transportation 53,000
Communications and Utilities 39,000
Wholesale and Retail Trade 136,000
Services 54,000
Government, Military 49,000
Government, Non-Military 98,000
651,000
SUBTOTAL 1,346,000
CHEMICAL STORAGE: 51,000
TOTAL 1,400,000
Source: DRI, Compliance Cost Calculations for EPA Regulation of Underground
Storage Tanks, December 20, 1985.
-------
DISTRIBUTION OF TANK TYPES
GASOLINE SERVICE STATIONS
CATHODICALLV
PROTECTED
(9.2%)
BARE STEEL
(70.5%)
FIBERGLASS
(20.3%)
Source: DRI, Underground Storage Tanks, Technical/Financial/
Economic Data Collection, October 2, 1985
-------
BARE STEEL TANK AGE DISTRIBUTION
RETAIL MOTOR FUEL ESTABLISHMENTS
0-5 VRS
(4.1%)
6-10 YRS
(7.6%)
>20 VRS
(35.8%)
11-15 YRS
(28.9%)
pi
x
7T
e'-
er
c-
rt
u>
i>
lo
I
16-20 YRS
(23.6%)
Source: DRI, Underground Storage Tanks, Technical/Financial/
Economic Data Collection, October 2, 1985
-------
3-8
tanks are retired without being replaced, as many older service stations have
closed over the last decade.
3.0. LOCATION OF TANKS IN RELATION TO POPULATION AND DRINKING WATER
Knowledge about UST locations is crucial for attempts to estimate the
types and magnitude of damages resulting from UST leaks. The most serious
potential damage from leaking USTs (if no action is taken) is their effect on
the health of users of ground water. In our analysis, more emphasis is placed
upon damages to private well users than users of public ground-water systems
because private wells may be more threatened by leaking USTs than public wells.
Public wells can be tested more efficiently, are more likely to tap less-
vulnerable confined aquifers, are less costly (on a per-gallon basis) to treat
if contaminated, and can be sited in areas less likely to become contaminated.
USTs are found where people are found. Where populations are dense, UST
populations are dense as well. There are common-sense reasons to expect this
to be true, as well as ample data. Exhibits 3.5a and 3.5b show the great
similarity between state-wide UST and population densities: variations in state
populations account for fully 95 percent of the variations in USTs by state.
Even for geographical areas as small as zip codes, service stations are found
to be closely related to populations. ]_/ Exhibit 3.6 shows a plot of service
stations by zip code against population by zip code, and indicates a strong
degree of association between stations and population (many of the outliers,
such as the circled points representing New York City, represent very atypical
situations).
USTs, then, are not isolated from people in the way that, say, hazardous
waste disposal facilities are isolated. Many leaks could be close to residences,
other structures, or buried conduits or sewers, and therefore cause threat of
fire or explosion if the leaked product is ignited. On the other hand, the
immediate threat to private ground-water wells could be lower than might be
expected, since, in the urban areas that contain most of the population and
therefore most USTs, very few private wells are found. Exhibit 3.7 shows that
while population and service stations are concentrated in areas of high popula-
tion density (urban areas are shown at the left of the graph) private wells are
more typically found in areas of low or moderately high population density.
Damage to public ground-water systems may show a different pattern. Though
the use of ground water (as opposed to surface water) for public water systems
varies from region to region, the-density of public well users is still corre-
lated with population density across the country—and this means that it is
correlated with the density of USTs, at least on a state-wide basis. (Even in
relatively densely populated areas, however, public water systems may be sited
to avoid having IISTs or other sources of contamination nearby.) Exhibits 3.8a
and 3.8b show that states with high UST densities coincide to a large degree
with states where ground-water users are densely packed.
\j A discussion of the analysis undertaken to establish the correlation
between UST densities and location and population densities and location is
included in Appendix B.
-------
3-9
Exhibit 3.5a
POPULATION DENSITIES BY STATE
* \\\ iv-v.w- •• •• f i iH
-v ¦¦¦ •• \\ ¦>", •. \ •¦ v -.\'t " -
«\\\\ ,-. .,
IP4 •
HP
BLANK:
HORIZONTAL SHADING:
DIAGONAL SHADING:
Persons per
165 > Persons per
40 > Persons per
square mile 2 165
square mile 2. 40
square mile
Exhibit 3.5b
DENSITIES OF USTs BY STATE
¦\W \V\ W - J .TT. -¦••¦V-.-.v. v •> •. ¦ <:h >—
^\yv\\y;iV-^ /
*. -\w^ i\v. v.%.\in •:•/•> w.\\-r-:v.-. v,v,-\ ¦> ¦ ¦>
v {••.••••¦••:•.••.•••• •• ¦••¦£ i
•-«; \'\Y< •. ' I n ¦*-
____r V. Wvi-\\\\V-\'\.
-1\•. \ v \\v.; \\\v-\
—-.j
&=
BLANK:
HORIZONTAL SHADING: 0.34 >
DIAGONAL SHADING: 0.08 >
USTs per square mile J> 0.34
USTs per square mile > 0.08
USTs per square mile
Source: SCI, based on 1980 census data.
-------
45
40
35
30
25
20
15
10 ¦
5 ¦
0 -
'LOT OF SERVICE STATIONS AND POPULATION
REGRESSION LINE
1~ 1 T
40 60
(Thousands)
POPULATION OF ZIP CODES
NYC POINTS CIRCLED
100
Source: SCI estimate from 1980 census data.
-------
DISTRIBUTION OF WELLS, STATIONS, POP.
ACROSS AREAS OF DM-T LIVING POP DENSITY
14.64 4.82
WELLS
1 .88
0 75
0.36
O 21
0.1 3
0.08
O 06
0.00
POP U LAI 101 >l/ACR E
SERVICE STATIONS
POPULATION
Source: SCI and PRA Estimates Using UST Model
-------
3-12
Exhibit 3.8a
DENSITIES OF GROUND-WATER USING POPULATION BY STATE
BLANK: Persons per square mile > 150
VERTICAL SHADING: 150 > Persons per square mile > 50
DIAGONAL SHADING: 50 > Persons per square mile
Exhibit 3.3b
DENSITIES OF USTs 3Y STATE
I . 1
,N >N\V^ % \*\* *.* « •»». .V.
> ; V.s » v.» \ ••A •. *. \ \ \ -.1 ¦ ?_
a/"
BLANK:
HORIZONTAL SHADING:
DIAGONAL SHADING:
USTs per square mile > 0.34
0.34 > USTs per square mile > 0.08
0.08 > USTs per square mile
Source:
SCI, based on 1980 census data.
-------
3-13
3.E. Importance of Hydrogeological Settings
Information on the physical characteristics of the sites where USTs are
found is also important to allow estimates of the number of tanks leaking, the
formation of plumes and the movement of contaminants. Currently, USTs are
distributed across eleven distinct, USGS-identified, hydrogeologic regions
nationwide (see Exhibit 3.9), and, within these regions, in distinct hydro-
geological settings. Leaks in some of these settings are much more threatening
(in terms of likelihood of damages) than in others, and cross tabulation of the
USTs by setting with proximity of water use will allow a truer picture of
site-by-site risks to emerge.
3.E.I. Hydrogeological Factors Affecting Failure Rates
There are two hydrogeological factors that have a major effect on leak
rates: aggressiveness of the soil and water table depth. The aggressiveness
of the soil is a function of several physico-chemical properties including soil
pH, sulfide concentration, and oxidation/reduction potential. The UST Simula-
tion Model and several independent analyses have characterized tank corrosion as
one of the chief failure mechanisms for unprotected tank systems in terms of
both frequency and release volume. For unprotected steel tanks, the rate of
corrosion is accelerated in aggressive soils. Natural soil materials vary
widely in terms of the factors that affect aggressiveness, and like other hydro-
geologic parameters, there can be significant variations on a local level.
Water table depth (also known as unsaturated zone thickness or vadose zone
thickness) also plays an important role in determining leak rates. When the
water table is permanently or occasionally higher than the bottom of the tank
it has the following effects:
o Periodic or permanent inundation affects the rate of corrosion, depend-
ing on the quality of the ground water;
o If the tank is not anchored by means of a ballast and strap system, its
buoyancy can cause it to shift, resulting in ruptures in the piping
system;
o If the water table is generally or always above the gasoline level in
the tank, a leak in the tank will result in water flowing in rather
than product flowing out (this is probably not true for the pipe system,
which is under pressure). In this situation, tank leaks are unlikely
to result in any health or environmental damage, and are likely to be
discovered and corrected without government intervention.
3.E.2. Hydrogeological Factors That Affect Characteristies of the Floating Plume
There are three primary factors that control the size of the floating
plume: permeability of the vadose, or unsaturated, zone material, pore size
distribution of the vadose zone material, and water table depth. In permeable
material, the gasoline can move more rapidly, and the plume can spread out so
that it is relatively wide and shallow. If large pore sizes are relatively abun-
dant, the capillary forces that restrict the downward flow of gasoline are not
as strong; otherwise, the fine-grained material can essentially 'lock up' the
gasoline, preventing it from reaching the water table. And if the water table
-------
Exhibit 3.9
Gi*ci«i«KJ
Ctitim
I
600KU.OM£I£NS
Groundwater region* ol »>« Unlltd Stain (AHm Healh, 1904).
-------
3-15
is deep, a floating plume may not form, provided that the vadose zone comprises
relatively finegrained material.
3.E.3. Hydrogeological Factors Affecting Characteristics of the Aqueous Plume
Several hydrogeologic factors control the introduction and movement of
gasoline contaminants in aqueous plumes and the vulnerability of aquifers to
contamination:
o Whether the aquifer is confined or unconfined;
o Ground water velocity;
o Aquifer thickness; and
o Amount of organic carbon present in the aquifer material.
The most important of these is probably whether the aquifer is confined or
unconfined. When the uppermost surface of an aquifer coincides with the water
table (i.e., the boundary between the vadose zone and the saturated zone), the
aquifer is unconfined. Unconfined aquifers are recharged by infiltrating
rain water, and are relatively vulnerable to contamination from sources like
USTs that release pollutants directly above them.
Confined aquifers are bounded above by an aquitard, i.e., a geologic unit
with low permeability that limits water flow. Confined aquifers are generally
recharged in areas where the confining layer (overlying aquitard) is absent.
These aquifers are less vulnerable to UST contamination because:
o The floating plume will not come in direct contact with the aquifer.
Benzene and other constituents will slowly diffuse and advect through
the aquitard, but the mass flux (pollutant loading rate) will be lower
than for an unconfined aquifer.
o Contaminants are generally less mobile in aquitards. Organic carbon
content of clays and silts (common aquitard materials) is relatively
high, resulting in strong adsorption and retardation of organic pollu-
tants.
o The aquifer may be pressurized, i.e., hydraulic head is generally
higher than the base of the aquitard, so that the water in the aquifer
is 'pushing' upward.
Many of the aquifers that are used for current drinking water supplies are
confined. Confined aquifers are sometimes preferred over unconfined because
they are more dependable during droughts and can provide high well yields.
This is particularly important for municipal well systems. Private drinking
water wells do not require high well yields, and thus can often tap unconfined
aquifers (saving on well drilling costs, which are increased with well depth).
Exhibit 3.10 provides preliminary estimates of the percentage of public
and private wells drawing from confined aquifers in each of ten hydrogeologic
regions identified by USGS. In some regions, such as the High Plains and Western
-------
3-16
Exhibit 3. 10
Percentages of Ground Water Used from Confined
Aquifers for Each Ground-Water Region 1/
Ground-Water
Region
Percentage of Ground-Water Used from
Confined Aquifers
Municipal Supply
Domestic Wells
Western Mountain Ranges
Arid Basins
Columbia Plateau
Colorado Plateau
High Plains
Central
Glaciated Central
Unglaciated Appalachian
Glaciated Appalachian
Coastal Plain
0
45
40
90
0
5
50
0
10
90
5%
50%
45%
95%
5%
10%
55%
5%
15%
95%
0
0
10
20
0
5
20
0
0
65
5%
5%
15%
25%
51
10%
25%
5%
5%
70%
V Source: ICF, Inc. estimates
-------
3-17
Mountain Ranges, almost all wells tap unconfined aquifers. In other regions,
like the Atlantic and Gulf Coastal Plain, most wells tap confined aquifers.
Ground water velocity also determines aqueous plume characteristics.
Velocity is a function of hydraulic gradient and permeability. It controls time
of travel (TOT) and thus the potential for degradation, and also bears on the
amount of water available for dilution. In very low-velocity systems, leaks
are likely to be detected before contaminants reach wells. In very high-veloc-
ity systems, more wells may be affected, but the contaminants are likely to be
in more dilute concentrations. Aquifer thickness also affects dilution rates;
thicker aquifers are expected to exhibit lower contaminant concentrations.
Velocities range over about five orders of magnitude, although most aquifers
have velocities on the order of 0.01 m/day to 5 m/day; aquifer thickness is
generally in the range of a few meters to 100 meters.
The fraction of organic carbon (foe) in the aquifer material also affects
aqueous plume characteristics. Benzene and other organic contaminants move at
velocities that are somewhat slower than that of ground water; the higher the
foe, the more strongly they are adsorbed, and the slower they move. This
effect is similar to that of ground water velocity on plume characteristics.
The foe's for aquifer materials vary over about two orders of magnitude.
3.E.4. Distribution of Hydrogeological Settings
Because damages caused by leaking IJSTs are so highly dependent on hydro-
geological settings, we attempted to develop a first cut at the distribution of
some important parameters across the nation. One mapping of hydrogeological
settings that is a useful starting point is the USGS (Heath) ground-water
regions (shown previously in Exhibit 3.9). USGS regions distinguish hydro-
geological settings across the nation according to large-scale variations in
geologic and ground water characteristics. However, there is still significant
variation in hydrogeological settings within each USGS Region.
As we have explained, the plumes resulting from IJSTs vary widely with
ground water velocity, vadose zone medium, and depth to ground water. Therefore,
we developed a distribution of hydrogeological settings in categories which re-
present combinations of these three variables. Exhibit 3.11 presents the number
of underground storage tanks located in each of 27 hydrogeological settings
(based on all possible combinations of the three variables, each subdivided
into three ranges). The information in Exhibit 3.11 is useful for analyzing
cost and effectiveness of regulatory alternatives in areas of different hydro-
geological vulnerability. As outlined above, areas which have USTs in vulner-
able hydrogeological settings, and are therefore potentially at higher risk,
are those areas of the country which have unconfined aquifers with high water
tables (not including those settings where the water table is above the bottom
of the tank), high ground water velocities and more permeable soil types.
However, the distribution of hydrogeological settings alone does not ade-
quately characterize potential risks from leaking tanks. A vulnerable hydrogeo-
logical setting may contain few underground storage tanks or be situated where
ground water is not suitable for use, and therefore few, if any, benefits will
accrue from regulations for that setting. Ground water may actually be more
threatened in an area with a high density of IJSTs even if the hydrogeological
setting there is intrinsically less vulnerable.
-------
HEATH REGION
HYDRO-GEO?/REGION REGION REGION REGION REGION REGION REGION
SETTING
1
2
3
4
5
6
7
1.1.1
0
0
0
0
0
10959
0
1.1.2
0
0
0
0
0
347
0
1.1.3
0
0
0
0
0
0
0
1.2,1
0
0
0
0
0
0
0
1.2.2
0
0
0
0
0
0
4344
1.2,3
0
0
0
0
0
2106
4681
1.3.1
0
0
0
0
0
0
118
1.3.2
115
0
11
0
0
0
10635
1.3.3
1203
47
0
48
68
1287
1283
2.1.1
0
0
0
731
0
1288
0
2.1.2
79
0
324
0
0
541
0
2.1.3
0
0
0
0
0
1668
0
2.2,1
0
0
0
0
0
0
1213
2.2,2
0
0
0
0
0
0
70 75
2.2,3
0
0
0
0
0
0
741
2.3,1
0
0
0
0
0
0
0
2,3,2
11
15444
0
0
238
0
42
2,3,3
0
0
0
0
109
0
0
3.1.1
0
0
0
0
0
478
0
3.1.2
286
0
0
0
0
0
0
3,1.3
101
0
0
0
0
0
0
3.2.1
0
0
0
0
0
0
0
3.2,2
0
0
0
I)
0
0
0
3.2,3
0
0
0
0
0
0
0
3.3,1
0
0
0
0
0
0
0
3.3,2
0
33
0
0
0
0
0
3,3,3
0
0
221
0
972
0
(J
Other
1232
335
113
0
0
0
0
TOTALS
3027
15859
669
779
1387
18647
30132
% TOTAL
2.66
13.92
0.59
0.68
1.22
16.37
26.45
\f See Heath Region Map, Page 3-14
£/ Depth to Ground Mater, Vadose Zone Medium,
1 = 0-10 Meters 1 = S. Stone, Lime Stone, Shale
2 = 10 - 20 Meters 2 = Silt/Clay
3 = >20 Meters 3 » Sand f. Gravel
REGION REGION REGION REGION REGION REGION
8 9 10 11 12 13
%
TOTAL
0
0
0
0
0
0
9.36
0
0
0
0
0
0
0.30
0
0
0
966
0
0
0.85
0
5858
0
0
0
0
5.14
4689
514
3250
0
0
0
11.2
0
967
0
0
0
0
6.8
0
23
0
0
0
0
0.1
3
0
333
3852
0
0
13.12
2075
735
9883
1635
251
75
16.32
0
0
0
0
0
0
1.77
0
0
0
0
0
0
0.80
253
0
0
0
0
0
1.69
0
0
0
0
0
0
1.06
0
0
0
0
0
0
6.21
0
4088
0
0
0
0
4.24
0
0
0
0
0
0
0
0
0
0
0
0
0
13.81
0
0
0
0
0
0
0.01
0
0
0
0
0
0
0.42
0
0
0
0
0
0
0.25
0
0
0
0
0
0
0.09
0
0
0
0
0
0
0
0
0
3173
0
0
0
2.79
0
0
0
0
0
n
0
0
685
0
0
0
0
0.60
0
0
0
0
0
0
0.03
0
0
0
0
0
0
1.05
197
109
0
0
115
0
1.84
7217
12979
16639
6453
366
75
6.33
11.39
14.60
5.66
0.3?
0.07
Ground Water Velocity
1 = < 0.3? Meters/Oay
2 = 0.32 - 2.23 Meters/Oay
3 = > 2.23 Meters/Oay
-------
3-19
Cross-referencing the hydrogeological settings and tank densities by USGS
Region can provide an idea of which areas of the country have relatively vulner-
able hydrogeological settings and high concentrations of underground storage
tanks. For example, a relatively high proportion of USTs are found in USGS
Region 7, but about one-half of the underground storage tanks in USGS Region 7
are located in clay soils. Clay soils impede the flow of leaking petroleum to
the ground water table, and therefore, stringent controls for tanks in clay soils
may not be as cost-effective as stringent controls for tanks in other soils.
3.E.5. Summary
Hydrogeologic characteristics have a strong influence on leak rates from
USTs, characteristics of floating plumes, and characteristics of aqueous
(dispersed) plumes. These characteristics vary widely on a regional and local
basis. Some aquifers are much more vulnerable to contamination than others,
and we can identify some settings where leaks are not likely to cause significant
problems (e.g., water table permanently higher than the top of the tank).
Risks from leaking underground tanks cannot, however, be predicted on the
basis of the hydrogeological setting alone. As discussed earlier, tank density,
population density, and ground water usage also play vital roles in establishing
the level of potential risks from leaking tanks.
3.F. CURRENT PRACTICES
3.F.I. Leak Detection
There is no evidence that any form of leak detection or leak monitoring,
other than manual inventory control, is widely practiced by current tank owners.
According to EPA's Release Incident Survey, visual, detection and detection by
odor or vapors are the most commonly reported methods of detecting leaks from
underground tanks. In fact, more than 65 percent of reported releases were
detected either visually or due to the presence of odor or vapor. }_/ This
evidence suggests that leak detection equipment is not widely used by the
current population of tank owners. Inventory control may or may not be widely
practiced by tank owners, but evidence reported by both the Release Incident
Survey and the Retail Motor Fuel Tank Survey suggests that if inventory control
is practiced, it is not done often or well, and has not figured prominently in
the detection of tank leaks.
3.F.2. Responses to Tank Leaks
As mentioned in Chapter 2, after a leak has been detected a choice can be
made to accept further damages, or to take steps to mitigate these future
damages. The following discussion covers three potential responses: do nothing,
clean-up, and treatment or use of alternative water sources.
Do Nothing: If the source is cut off, the leaked product could continue
to mi grate. Tn many cases it would form a plume in, or on top of, the surficial
aquifer if it had not already done so before detection, and the plume would
\j EPA Office of Underground Storage Tanks, Summary of State Reports on
Releases from Underground Storage Tanks, August 1986.
-------
3-20
proceed to migrate, slowly dispersing and possibly affecting wells before
discharging to surface water. As time went on, the concentrations of contami-
nants in additional affected wells would drop. The chance that a public system
would be encountered would rise over time, though the odds that this will
happen before the concentrations in the plume are greatly diluted are small in
the typical case.
Of course, if the plume were to intercept a dense private well field, a
public well field, or cause other damages, then there may be interest in explor-
ing ways to avoid these damages. These potential responses are discussed
below.
Undertake Clean-up: Corrective action can take several forms. The con-
taminated soil in the unsaturated zone might be excavated and treated. Or the
floating plume could be removed by sinking wells just into the saturated zone,
and using an oil/water separator to recover the free product. This step may or
may not be accompanied by some form of treatment of the water (which will have
some gasoline dissolved in it) pumped up with the free product. This step would
largely eliminate the source for the dispersed plume, which would then fall more
rapidly in concentration as it grew. Costs for this action are quite variable.
The removal of a large floating plume could cost several hundreds of thousands of
dollars. Costs of removing the floating plume from a smaller release might be
under $100,000. EPA is in the process of refining its estimates of clean-up
costs.
Removal of the dispersed plume could eliminate much of the remaining risk,
though this could add substantially to costs. Some large spills have been known
to cost in the millions of dollars to clean up, where the goals for the purity of
the ground water after the clean-up have been strict.
Treat Drinking Water or Use Alternatives: Upon discovering that their
drinking water is contaminated, residents of a home with a tainted well have
various alternatives. They could continue to drink the contaminated water and
bear the health risks, or they could switch to bottled water for drinking and
cooking. Bottled water would avoid about half of the health risk, as much of
the risks from inhalation would remain.V Alternatively, the residents could
purchase carbon adsorption units to treat their water, eliminating nearly
all the health risks as long as they were maintained properly. Finally, the
residents might be able to drill a new or deeper well to clean water or tie
into a municipal water system.
If a public water supply well were to be affected, risks could again be
accepted or eliminated via treatment, at an annualized cost in the tens or
hundreds of thousands of dollars, depending on the size of the system. In a
typical case, the discounted cost of a treatment system for a public system
would be low, given that decades could be expected to pass before the plume
reached public wells and treatment would be necessary.
V EPA Office of Drinking Water, Economic Impact Analysis of Proposed
Regulations to Control Volatile Synthetic Organic Chemicals in Drinking Water,
October 1985.
-------
3-21
3.F.3. State Regulations
Forty-two states have at least some form of LIST legislation in place.
Most of these state programs reflect state fire code restrictions for petroleum
storage and petroleum industry codes and recommended practices. Approximately
twelve states have comprehensive programs for regulating underground storage
tanks. States with comprehensive programs are, for the most part, states with
relatively high ground water usage.
Since the majority of states do not have comprehensive UST programs
regulating tank types or leak detection and monitoring practices, state
regulatory requirements are not incorporated into the established base case.
EPA did, however, evaluate state regulatory approaches and sought state input
when establishing its regulatory framework and in structuring the regulatory
alternatives for underground tanks.
3.G. THE BASE CASE
The base case depicts, or illustrates, the alternative of no federal
regulation, or what will prevail in the absence of any regulation. The base
case is very important in the analysis of the regulatory alternatives because
it provides a base level from which we can evaluate the incremental costs and
benefits of each regulatory alternative. By comparing each alternative with
the established base case we obtain the incremental costs and benefits of the
alternative over the alternative of no regulation.
3.G.I. Characteristics of the Base Case
All the available information that is outlined in the previous sections
of this chapter was evaluated to establish the "no regulation" base case used
in analyzing the costs and benefits of EPA's regulatory alternatives. The base
case consists of 1.4 million tanks: 89% are constructed of bare (unprotected)
steel and 11% are constructed of fiberglass reinforced plastic (FRP).V The DRI
tank age distribution (presented in Exhibit 3.4) is the assumed age distribu-
tion of tanks in the base case. As will be explained below, a certain level
of base leak detection (manual or sensory) is incorporated into the base case,
but no supplemental leak detection or monitoring equipment is included in the
base case. All replacement tanks, or newly installed tanks, are single-wall,
coated and cathodically protected steel tanks. This tank type is the minimum
tank construction requirements allowed under the current interim prohibition
for new tanks.
3.G.2. Proportion of Tanks Leaking
EPA's UST Simulation Model provides a method for assessing what proportion
of the UST universe is currently leaking. Using the model, we can simulate the
lives of tanks of different types and different ages. By repeating the simula-
tion of individual tanks many times, we can develop an understanding of what
V EPA Office of Pesticides and Toxic Substances, Underground Motor Fuel
Storage Tanks: A National Survey, Vol. 1 Technical Report, May H 1986, p.
2-10. The distribution of tank types in Exhibit 3.3 was taken from DRI's
report of October 3, 1985. We chose to use EPA's survey results in our analysis.
-------
3-22
happens to a population of identical tanks in the same situation. We can also
compare model runs of different tank types to draw conclusions about the effects
of different tank technologies, and the correlation between tank characteristics
and release profiles. We use model runs simulating different tank types, tank
ages and environmental situations to scale the model results to the current
tank population.
The modeling approach to assessing the current situation for underground
storage tanks uses information that is currently available, but also requires
that explicit assumptions be made when information is not available. The LIST
Model uses currently available data regarding tank characteristics and failure
probabilities, and incorporates physical properties of tank construction mater-
ials and the tank's environmental surroundings. Best engineering judgement is
used to provide needed assumptions when current data is unavailable. The model
does require that assumptions and data be used in an explicit and consistent
manner and, if there is uncertainty about particular assumptions, the model can
be used to conduct sensitivity analyses over the range of uncertainty in the
assumptions.
One assumption that is very important in the IJST simulation is the assump-
tion made in regard to the level of base detection. Base detection is the
assumed level of leak detection that is currently in place at tank facilities.
Some leaks will be detected in the absence of regulation, without leak detection
equipment and without stringent inventory control. Some leaks are detected by
product inventory. Other leaks are detected because water leaks into a tank,
petroleum appears in the basement of a nearby property, or a nearby ground-water
well is contaminated. The base detection assumption in the UST model is an
attempt to simulate these current practices or the current level of leak detec-
tion.
We represent base detection in the UST Model by adjusting the effectiveness
of monthly inventory control. The level assumed for base detection signifi-
cantly affects the conclusion that can be drawn about the number of underground
tanks that are currently leaking undetected. The better the level of base
detection is assumed to be, the fewer the number of tanks that can be concluded
to be leaking undetected. If we assume that base detection is virtually non-
existent, we can conclude that a large percentage of the current tank population
is leaking and will not be detected without requiring more stringent levels of
leak detection.
The level of base detection is entered into the UST Model as a percentage
of tank capacity for daily and weekly detection, and is entered as a percentage
of monthly throughput for monthly detection. For example, a base detection level
of 10% - 10% - 3% means that a leaking tank will be detected if 10% of the tank
capacity is lost in a day, if 10% of the tank capacity is lost in a week, or if
3% of the monthly throughput is lost in a month. Exhibits 3.12-3.14 illustrate
the effect of different assumptions for base detection upon the percentage of
tanks currently leaking and the distribution of release volumes for bare steel
tanks. The exhibits illustrate three levels of base detection: 4% of monthly
throughput, 3% of monthly throughput and 2% of monthly throughput. The level
of performance attributed to base detection has an important implication for
evaluating regulatory alternatives. The better base detection is, the less
-------
3-23
value is accrued from technical standards imposed by government regulation. If
we compare the modeling results to survey data from the EPA1s National Survey
of Underground Motor Fuel Storage Tanks and the National Release Incident
Survey, a base detection level that assumes that a loss of 3% of monthly tank
throughput can be detected in a month calibrates the model to closely reflect
the current UST situation depicted by these surveys. This level of base
detection (10% - 10% - 3%) is the assumption used throughout our analysis.
As illustrated in Exhibits 3.12-3.14, a more stringent base detection
assumption results in a smaller estimated proportion of facilities with
undetected leaks. Relaxing the base detection assumption results in a higher
prediction of undetected leaks, or a decrease in the proportion of tight facili-
ties, and larger release volume estimates. By assuming a more stringent level
for base detection (Exhibit 3.12) we are assuming that fewer gallons of product
are released before a leak is detected. Therefore, leaking tanks are detected
more quickly, less tanks may have undetected leaks at any given time, and the
distribution of release volumes will show a greater number of small release
volumes relative to large release volumes. In contrast, if we assume a less
stringent level for base detection (Exhibit 3.14), more gallons of product will
be released before a leak is detected. The result is a greater number of
undetected leaks (or fewer tight facilities) and larger release volumes.
3.G.3. Predicted Status of Tanks
The IJST Simulation Model allows us to use the information available re-
garding the current tank universe to predict the status or failure state of the
tank universe now, and over a given period of years. Exhibit 3.15 illustrates
the model's predicted status of bare steel tanks for the base case. This exhibit
shows the proportion of the current population of tanks that are either leaking
because of pipe failures or tank failures, or are tight.
Exhibit 3.15 depicts the model's predicted status of the current popu-
lation of bare steel tanks, given a particular set of inputs and assumptions.
These inputs and assumptions reflect EPA's best available information regarding
the current characteristics of the UST population. Not every aspect of the UST
universe can be accounted for, and therefore, the predicted status of the bare
steel tank population shown in Exhibit 3.15 is only a best estimate. The
actual numbers of leaking and tight UST facilities may vary due to the lack of
full or precise information. For example, EPA is aware that some of the major
oil companies, and in fact some independent dealers, have undertaken voluntary
tank upgradings or tank replacement policies. Because the Interim Prohibition
requires all new tanks, or replacement tanks to be corrosion resistant, the
effect of these voluntary upgrading policies is to reduce the number of existing
bare steel tanks and therefore reduce the likelihood of facility failures.
Because little information is known about the extent of, or the details
of, voluntary upgrading or tank replacement policies, these policies are not
included in the base case assumptions, and the base case and options were
modeled assuming no voluntary replacement policies. The result is that the
number of tanks predicted to be leaking could be overstated to the extent that
a number of bare steel tanks are being replaced before they leak through these
voluntary upgrading and tank replacement policies.
-------
3-24
Exhibit 3.12
CURRENT STATUS OF UST POPULATION
BASE DETECTION: 10%- 10«/4-2«/4
LEAKING
FACILITIES
(25 6%
TIGHT
FACILITIES
(74.HV®)
SIZE DISTRIBUTION OF RELEASE INCIDENTS
BASE DETECTION: 10®/i-10°/i-2Vo
10000
SIZE RELEASE (GALLONS)
Source: SCI Estimates using UST Model results.
-------
3-25
Exhibit 3.13
CURRENT STATUS OF UST POPULATION
BASE DETECTION lOVi-IO'i-3%
SIZE DISTRIBUTION OF RELEASE INCIDENTS
BASE DETECTION. 10%- 10®/o-3"A
SIZE RELEASE (GALLONS)
Source: SCI Estimates using UST Model results..
-------
3-?fi
Exhibit 3.14
CURRENT STATUS OF UST POPULATION
BASE DETECTION: 10V«-
leaking
FACILITIES
(32 6%)
TIGHT
FACILITIES
(67 4%)
SIZE DISTRIBUTION OF RELEASE INCIDENTS
z
Ul
o
G
z
w
2 ~
< c
ui o
-j —
Ui -
u.
O
Ic
iu
CD
5
z
1 <4
1 3
1 2
1.1
1
0 3
0.8
0.7
0 6
0 S
0 >4
0 3
0.2
0.1
0
BASE DETECTION: 10°/o- 10%-4%
<500
500-3000 3001-10000
SIZE RELEASE (GALLONS)
>10000
Source: SCI Estimates using UST Model results
-------
BASE CASE POPULATION BY STATUS AND AGE
110
100
90
80
70
SO
50
40
30
20
10
15,541
28,451
108,573
88.477
134,292
0-5 Years
6-10
11-15
Tank Age (years)
18-20
Over 20
V/A TiSh* Fac. Pipe Leak E\>^l Tank Leak
Source: DRI Age Distribution; SCI and PRA Status Estimates Using UST Model
-------
NUMBER OF FLOATING PLUMES BY UNSATURATED ZONE MEDIA
800
700
600
500
w
TJ
C
D
w 400
3
O
¦C
1mm*
300
200
100
0
Y//A Plumes < 25 sq m. fv\\N Plumes > 25 sq. m.
Base Case Population
Unsaturated Zone Media
Source: SCI and PRA Estimates Using UST Model
-------
3-29
3.G.4. Effect of Hydrogeological Setting and Resulting Plume Areas
The UST Simulation Model also allows us to analyze the effect that differ-
ent parameters or external variables may have on the costs and benefits of
regulations. Exhibit 3.16 illustrates the effect that different hydrogeolog-
ical settings have on the number of floating plumes that occur as a result of
leaking USTs. This exhibit shows the estimated number of floating plumes attrib-
uted to the base case population of tanks distributed across five unsaturated
zone mediums. The distribution of floating plumes was segregated into large
plumes (greater than 25 square meters) and small plumes (less than 25 square
meters) to greater illustrate the affect of soil types upon plume character-
istics. Sand and gravel soils, which are generally more porous and more per-
meable than clay soils, do not restrict the downward flow of gasoline as much
as clay soils and allow the product to move more rapidly and to spread out.
Sand and gravel soils therefore result in more plumes and a greater proportion
of large plumes. Clay is the least permeable and the least porous of the soil
types shown and therefore results in the greatest proportion of small plumes.
Water table depth also affects the number and size of plumes. For a given soil
type, a more shallow water table will generally result in a greater number of
plumes. If the water table is relatively deep, small releases and releases
that are detected relatively quickly may never reach ground water; product is
released only to the unsaturated zone and there are relatively fewer plumes than
with shallower water depths. In Exhibit 3.16, a distribution of ground water
depths was estimated based upon analysis done by Pope-Reid Associates regarding
the distribution of USTs across Heath Regions (see Exhibit 3.11).
3.G.5. Rase Case Responses to Leaks
The base case is established to provide a snapshot of the present situation
in order to obtain estimates of the incremental costs and the incremental
effectiveness of proposed regulations. In most cases, the characteristics of
the base case, or the present situation, can be estimated rather easily. For
some characteristics, the estimate of present practices is much more difficult,
as is the case in estimating the current level of corrective action that is
undertaken when USTs are determined to be leaking.
We believe some level of corrective action currently takes place. Of the
reported incidents covered in EPA's Release Incident Survey, approximately fifty
percent of the incident responses reported taking some type of remedial or
corrective action following leak detection. A majority of these responses
reported tank removal and/or tank replacement only. Eighteen percent of these
incidents reported undertaking some soil excavation. Ground-water treatment
techniques, such as steam stripping, were reported by less than 12% of the
responses. }_/
]_/ EPA Office of Underground Storage Tanks, Summary of State Reports on
Releases from Underground Storage Tanks, August 1986, Pp. 8-6 - 8-9.
-------
3-30
Given the fact that the Release Incident Survey reports evidence of very
little corrective action taking place, and given EPA1s uncertainty regarding
the levels of corrective action currently occuring at leaking UST sites, we
chose not to include corrective action measures, beyond pipe repairs and tank
replacements, in the base case for our analysis.
When we include consideration of the cost of EPA1s corrective action re-
quirements, we incorporate the cost of corrective action into the cost of the
regulatory options. We also include the cost of corrective action due to EPA1s
requirements that would result for the base case if current practices continued.
Whenever corrective action costs are incorporated into the base case, it is
appropriately labeled to avoid any confusion. In all other instances, the base
case includes no corrective action measures beyond pipe repairs and tank
replacements.
3.G.6. Model Results for Base Case
Table 3.1 lists the UST Model outputs for the base case population of
tanks. The release volume, number of plumes and the floating plume area are
the totals for all tanks in the base case (1.4 million tanks) over a thirty-
year period. Total cost includes the cost of tank repairs, replacements and
upgrades, and the cost of product lost due to tank system failures. Costs are
shown for a period of thirty years and are discounted using a discount rate of
three percent. \j The base case estimates provide a picture of what could
result under the alternative of no regulation. Ry comparing similar results
of model runs simulating the imposition of each of the regulatory alternatives
with the base case estimates, we can obtain estimates of incremental costs and
benefits of each alternative. Using the model in a similar fashion, we can also
compare the relative costs and benefits between the alternative regulatory
approaches.
Exhibit 3.17 provides an estimate of the cumulative contamination from
floating plumes for the base case population of tanks over a thirty-year period.
After thirty years, the cumulative plume area begins to level off. Contamina-
tion is measured in acres of floating plume and does not include dispersed plume
contamination to the aquifer. Again, these same results can be simulated for a
tank population under each regulatory alternative and compared to the base case
results to obtain estimates of incremental benefits of regulating underground
storage tanks.
The following chapter, Chapter 4, introduces a range of actions that
must be addressed by federal regulations to mitigate the problems of leaking
USTs. Chapter 5 provides the results of cost and effectiveness analyses of
UST regulatory alternatives. Chapter 5 compares the cost and effectiveness of
EPA's proposed regulatory alternative to the base case and to other selected
regulatory alternatives.
V A thirty year period is used in order to provide estimates of tank
failure probabilities and costs over an extended tank life. The average age of a
bare steel tank is approximately 17 years, but many tanks survive for 20 or more
years. A 3% discount rate is used to reflect the current real rate of interest.
-------
3-31
TABLE 3.1
MODEL RESULTS FOR BASE CASE TANK POPULATION
(1.4 Million Tanks Over a Period of 30 Years)
Number of Release Incidents 2,667,660
Number of Floating Plumes 1,975,380
Number of Floating Plumes
less than 25 square meters 537,520
Total Plume Area (Acres) 190,387
Total Cost V $ 31.0 Billion
\f Cost for 30 years discounted at 3%. Total Cost includes the cost of
tank repairs, replacements and upgrades, and the cost of product lost due to tank
system failures. Total cost does not include any cost for corrective action
(we assume no corrective action in the base case).
-------
CUMULATIVE FLOATING PLUME ACRES OVER A 30 YEAR PERIOD
BASE CASE POPULATION
=r
co
i
CO
ro
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 T"
S 10 11 12 13 14 15 16 17 18 IS 20 21 22 23 24 25 28 27 28 29 30
YEAR
Source: SCI and PRA Estimate Using UST Model
-------
Chapter 4
COMPONENTS OF A REGULATORY STRATEGY
This chapter introduces the broad range of individual actions that could
be required by the regulation to address the problem of leaking USTs. We pre-
sent data on the costs of each, as used in the calculation of cost-effectiveness,
and present some illustrative UST Model results on the effectiveness of each in
avoiding failures, release incidents and volumes, and contaminant plumes.
4.A. INTRODUCTION
Numerous strategies and combinations of strategies could be used to miti-
gate the damages likely to occur in the base case. Generally, these strategies
can be divided into actions that address (through corrective action) releases
that occur in order to reduce the damages they may cause; actions that prevent
or minimize potential releases from new tanks; actions that prevent or minimize
potential releases from existing tanks; and regulations that ensure that those
responsible for releases will be financially able to deal with them. The
following sections take up each of these divisions, and introduce a number of
component actions which could be used to make up a regulatory strategy within
each division.
Some UST Model results are presented to indicate the relative costs and
levels of performance of the components. A meaningful comparison of the
overall cost and effectiveness of isolated, individual components of a regu-
latory strategy is difficult to make, however, because of the many potential
interactions among the components. For instance, the relative advantages of
one type of leak detection for existing tanks over another type will vary
depending on the other leak detection methods used in conjunction with that
type of leak detection, and on the provisions for new tanks and for corrective
action. For this reason, an overall comparison of cost and effectiveness is
postponed until the following chapter. That chapter describes the group of
integrated regulatory options, each composed of a combination of these com-
ponents, that EPA constructed as potential regulatory packages.
4.B. ALTERNATIVE ELEMENTS OF A STRATEGY FOR NEW TANKS
The principle choices to be made for regulations of new USTs involve the
construction of the tank (which determines the frequency of failure and/or
release) and the method and frequency of leak detection (which affects the
eventual size and longevity of any release that occurs). The main alternatives
for each of these areas are described in the following sections. A number of
other choices, for instance, those involving regulation of installation prac-
tices, although important, are basically secondary and are not discussed expli-
citly here.
-------
4-2
4.B.I. Tank Construction
A basic choice in the regulation of new tanks is the specification of the
construction of the tanks. Protection against corrosion (the main cause of
releases from existing systems) and provision for containment of releases from
the system are the most important goals for tank construction standards.
Choices for materials for tank construction include bare steel, steel that
is coated to help protect against corrosive conditions, cathodically protected
steel (protected either by sacrificial anodes or impressed current, and gener-
ally given an insulating coating as well), and non-corroding fiberglass. The
Interim Prohibition (HSWA §9003(g)) allows bare steel tanks only in selected
settings where corrosion is considered to be a less serious problem. Thus, the
question of requiring corrosion protection or not in the new tank regulations
is largely moot because of this statutory mandate.
4.B.l.a Effectiveness in Reducing Failures
Exhibit 4.1 displays the frequency of failures over a thirty-year life as
predicted by the UST Model for three systems: bare steel tank and piping;
coated and cathodically protected steel tank and piping; and fiberglass tank
with cathodically protected steel pipes.
4.B.l.b. Costs of Corrosion-Resistant Systems
The capital cost of corrosion protection is not trivial, but it is none-
theless relatively small compared to the total costs of purchasing and instal-
ling a new tank. In addition, once the life-cycle savings in reduced repair
and replacement costs are factored in, corrosion-resistance can pay for itself.
Exhibit 4.2 shows the approximate purchase and installation costs, repair and
replacement costs, and costs associated with testing and maintaining the ca-
thodic protection systems. Costs other than the purchase and installation
costs are discounted at 3 percent per annum. Not included are costs of lost
product or corrective action costs; inclusion of these costs would work in
favor of the protected tanks compared to the bare steel tanks.
4.B.1.C Comparisons of the Cost and Effectiveness of Bare Steel and Protected Tanks
The data from the previous exhibits is presented in the form of a two-way
plot in Exhibit 4.3. The costs of the systems (broken into first costs and
other costs) are measured on the vertical axis, while the effectiveness of the
systems in terms of preventing the releases allowed by a bare-steel system is
measured on the horizontal axis. Presented in this way, the data allow us to
compare'the alternatives over two dimensions simultaneously. Preferred systems
are those falling further toward the lower right—those higher in effectiveness
and at the same time lower in cost. If any system is below and to the right of
all other systems, it is said to dominate the others. If two systems are found
at the same distance along the effectiveness axis, then the one with the lower
cost is said to be more cost-effective than the other. If a comparison of two
systems shows one to be above and to the right of the other--more costly and
more effective—then the incremental cost per unit of effectiveness may be
measured by comparing the vertical difference between the two to the horizontal
difference. This framework is used repeatedly in the analysis that follows to
al low elements of the options, and the options themselves, to be compared.
-------
4-3
Exhibit 4.1
FAILURE FREQUENCIES BY TANK SYSTEM TYPE
SYSTEM TYPE FAILURES % REDUCTION VS.
PER SYSTEM RARE STEEL
PER 30 YEARS
BARE STEEL TANK AND PIPING 3.81 NA
COATED & CATHODICALLY PROTECTED TANK AND PIPING 0.65 83%
FIBERGLASS TANK, PROTECTED PIPING 0.54 86%
Exhibit 4.2
TOTAL PRESENT VALUE COSTS BY TANK SYSTEM TYPE
(4,000 gal Ion tanks)
CATHODIC
PURCHASE AND REPAIR AND PROTECTION TOTAL
SYSTEM TYPE INSTALLATION REPLACEMENT MAINTAINENCE COST
BARE STEEL TANK AND PIPING S 19,995 $ 8,076 NA $28,071
COATED A CATHODICALLY
PROTECTED TANK AND PIPING $ 21,100 $ 544 $392 $22,036
FIBERGLASS TANK,
PROTECTED PIPING $ 26,825 $ 701 $392 $27,918
Note: Present values calculated assuming a 30-year life and 3% discount rate.
Estimated costs are per tank, assuming a 3-tank facility. They do not
include costs of product lost to releases, or the estimated salvage value
of the systems at the end of the 30-year period.
-------
in
en
<
L±J
>
o
N
ro
@ ^J)
T>
> §
CL tf)
D
5 °
UJ (-
|—
CO
y~
10
cc
UJ
CL
(/)
O
U
CORROSION RESISTANLq COST-EFFECTIVENESS
$45
W/ & W/OUT REPAIR & REPLACEMENT COSTS
$40
$35 -
$30 -
$25
$20
$15-
$10
$5 -
BARE STEEL
BARE STEEL
NOTE: Costs do not include cost of lost product.
T
T
T
0.00
PURCHASE,INST. ONLY
1.00 2.00
FAILURES AVOIDED VS. BARE STEEL
+ W/ REPAIRS, ETC.
+ FIBERGLASS
~ FIBERGLASS
+ C&CP
~ C&CP
3.00
4.00
O 100% ITFFXTIVENES
Exhibit 4.3
-------
4-5
The exhibit shows that, not counting repair and replacement, the bare-
steel system is the cheapest and least effective. The fiberglass system is the
most costly and, by a slight margin, the most effective. Counting repairs and
replacements, discounted moderately, the corrosion-resistant tanks both dominate
the bare-steel system. The cathodically protected steel tank is somewhat less
costly than the fiberglass tank, but appears slightly less effective.
4.B.l.d. Costs and Effectiveness of Tank Systems Constructed to Intercept Releases
Even corrosion-resistant tank systems are subject to releases. Containment
of these releases before they can enter the environment can be accomplished
either through the use of impermeable liners under the tank and piping or with
a second wall for the tanks and/or pipes. Due to the added cost of these
systems, they would almost certainly be specified as corrosion-resistant, if
only to protect the additional investment. It is natural to combine secondary
containment systems with a leak detection system able to identify a breach in
one wall before the second wall (or liner) fails or is overwhelmed by the
release. Total costs for these systems are presented alongside the costs for
other tank systems and the frequency of releases to the environment in Exhibit
4.4. The costs shown in this exhibit include costs of lost product, but not
the salvage value of the systems at the end of the period of the analysis.
These data are displayed graphically in Exhibit 4.5.
4.B.l.e. Marginal Cost Effectiveness of Tank Construction Alternatives
The data presented in Exhibits 4.4 and 4.5 can be used to examine the
marginal cost effectiveness of release prevention through tank system design.
Single-wall, cathodically protected steel tanks dominate both unprotected steel
tanks and fiberglass tanks, showing lower total costs and fewer release inci-
dents. For this reason, it makes little sense to discuss the marginal cost of
reducing releases through the use of a single-wall cathodically protected
system. We can calculate, though, that moving from a single-wall protected
system to a protected system with a liner will cost about $7,500 more per tank
and will prevent about 0.4 releases over thirty years. Thus, the incremental
cost per release avoided would be about $18,500. (The incremental cost of
secondary containment is higher for the larger tanks which are typically used
as replacement tanks. Releases avoided by secondary containment would be
smaller if a shorter average time horizon were used for the analysis, as is
appropriate for replacement tanks which will not be installed immediately but
only as existing tanks fail. These differences between the analysis presented
here, for new tanks, and the analysis in Chapter 5, for replacement tanks,
explains why liners appear much more cost-effective in this section than in
Chapter 5.) Whether this cost is worth incurring will depend on the seriousness
of the avoided releases, whether they would have to be cleaned up and to what
degree, what the cleanups would cost, and for how many years and at what rate
the costs would be discounted. For instance, if releases typically required
$100,000 in cleanup expenditures (or, in the absence of cleanup, would do
$100,000 or more in damage), then the cost of the lined system would be well
worth paying.
4.B.l.f. Influences on the Value of Avoiding Releases
The damage done by a typical release depends in large part on the volume
of the release and the duration over which it is allowed to continue before it is
detected. If a device or system is used with a single-wall tank system that is
-------
4-6
Exhibit 4.4
COSTS AND RELEASE FREQUENCIES BY TANK SYSTEM TYPE
SYSTEM TYPE
TOTAL COST PER TANK,
3 TANK FACILITY
NUMBER OF RELEASES
TO THE ENVIRONMENT
PER TANK, OVER 30 YEARS
BARE STEEL TANK AND PIPING
COATED & CATHODICALLY
PROTECTED TANK AND PIPING
FIBERGLASS TANK,
PROTECTED PIPING
LINER WITH INTERSTITIAL
MONITOR
$ 42,602
$ 25,130
$ 29,825
$ 32,515
3.36
0.60
0.50
0.20
DOUBLE WALL TANK AND PIPES
S 34,152
0.25
Note: Costs are the present value of all capital and operating costs for
30 years, discounted at 3 percent per year. They include costs of lost product,
but do not include a credit for the salvage value of systems remaining at the
end of the 30-year period.
-------
COST-EFFECTIVENlSS by system type
(/)
q:
<
Ld
o
fO
N
ro
@ \~n
x>
> §
Q_ m
D
2 °
LlI H
I— s—'
en
y~
if)
a.
LJ
Q_
I—
if)
O
u
IN AVOIDED RELEASES TO THE ENVIRONMENT
$45
$40 H
[] BARE STEEL
$35 -
$30 -
$25
$20
$15
$10
~ 2-WA
~ LINE
LL
R
~ FRP
~ C&CP
-P*
I
NOTE: Costs do include the cost of lost product.
0.00 1.00 2.00
RELEASES AVOIDED VS. BARE STEEL
3.00
o
100% EFFECTIVENESS
Exhibit 4.5
-------
4-8
capable of detecting releases promptly, many releases can be stopped before they
have contaminated the ground water. This alone greatly reduces the damage that
can be done by the release, as well as the costs of cleaning it up. Even if
a release cannot be detected before it forms a plume on the ground water, more
rapid detection can reduce health risks dramatically by keeping concentrations
of contaminants in the ground water low and by reducing the time of exposure.
The next section turns to the alternatives available for the detection of leaks.
4.B.2. Alternative Systems for Leak Detection
Numerous systems have been developed to allow detection of leaks or releases
from underground tank systems, and still more have been proposed. In this sec-
tion, a partial list of systems that might be used is presented and described
briefly in terms of method of operation, cost, and effectiveness in minimizing
ground-water contamination. Given the choice of a method of detection, there
will often be further choices of the sensitivity of the sensors used and of the
frequency with which the method is used. Systems addressed include:
o Tank and pipe tightness tests;
o Line leak detectors;
o Vapor welIs;
o Floating liquid sensors or observation wells;
o Manually-operated inventory monitoring programs;
o Automatic inventory control; and
o Interstitial space monitor (for secondary containment systems).
Each is discussed below.
4.B.2.a. Operation, Advantages, and Disadvantages of Detection Methods
Tank and Pipe Tightness Tests
Various methods are used to determine whether a tank system is tight or
not. Commonly, a tank will be taken out of service for a number of hours,
filled to capacity, allowed to reach equilibrium in temperature and shape, and
then monitored accurately to detect any drop in the level of liquid. (Variations
include applying a partial vacuum and listening for bubbles, or filling with an
inert gas and using vapor sensors outside the tank.) Advantages: Can detect
slow (0.15 gallon per hour) leaks with a relatively high degree of certainty.
Disadvantages: Too costly (at about $500 per test) and disruptive to perform
except at infrequent intervals (perhaps no more than once a year); some small
but still significant leaks can escape detection; inconsistent if testers have
not had adequate training; might indicate a leak when the only gap in the
system is a loose fitting at the top of the tank which would not release
product under actual operating conditions.
Line Leak Detectors
Line leak detectors are designed to signal piping failures by identifying
an abnormal pressure drop in pressurized delivery lines. Advantages: Operates
continuously; some are low in price; some are able to detect relatively slow
leaks; automatic in operation and therefore less dependent on skillful operators
to be effective. Disadvantages: Low priced devices are relatively insensitive
and may be subject to wear, and sensitive devices are relatively expensive.
They can miss some significant leaks, and are applicable only to pressure
-------
4-9
Vapor Wells
This detection system consists of an array of wells around the tanks
equipped with continuously operating sensors to detect the vapors which are
generated by released product and which diffuse through the vadose zone (unsat-
urated material above the water table). Alternatively, samples can be drawn
periodically from the wells and tested for traces of product vapors. Advan-
tages : Can trigger an alarm soon after product is released; relatively inexpen-
sive to install and operate, given the level of protection it can provide; may
be able to detect very slow leaks. Disadvantages: May be unreliable due to
unpredictable patterns of vapor movement; may not indicate which tank is leaking;
might be subject to false alarms, especially if releases occured at the site in
past years; solid data on performance in real-world applications is not yet
available; may not work in some settings (fractured rock, high water table).
Floating Liquid Sensors or Observation Wells
These are similar in some ways to vapor monitors, but rely on finding ac-
tual product in wells (using automatic sensors or periodic sampling) instead of
vapors. Advantages: Able to detect some very slow leaks; allow identification
of which tank is leaking; little danger of false alarm. Pisadvantages: Might
miss some leaks all together; might detect a release only after a long time
lag; floating sensors by definition are unable to detect a release until it has
already contaminated some ground water, although an observation well with
sensors above the ground water table could detect some releases before they
reached the ground water.
Manually-Operated Inventory Monitoring Programs
Over time, records of product levels in tanks, deliveries, and sales
generate data which can be analyzed statistically to detect a pattern of disap-
pearing product. Advantages: Low in cost, given that the data should be
collected for accounting purposes and to detect theft in any case; potentially
able to detect fairly slow leaks; designed to be used frequently enough to
detect many leaks before much damage is done. Disadvantages: Requires training
and some skill to use effectively; will miss some small but significant leaks,
and will detect others only after a time lag.
Automatic Inventory Monitoring and Control
These systems automate the inventory tracking procedures described above,
employing a set of sensors in the tanks connected to a central control box.
Advantage: Eliminates the need for training, skill, and diligence in measuring
and recording product levels. Pi sadvantages: Somewhat expensive (though they
reduce the burden on the persons who would otherwise have had to make, record
and analyze stick readings) and are untested for detecting slow leaks.
Interstitial Monitors
These devices use a variety of means (most notably pressure change or
conductivity sensors) to detect the presence of product in the interstitial
space (between the inner and outer containment barriers) in secondary contain-
ment systems. Advantages: Very rapid and reliable detection. Their superior
-------
4-10
performance is due not to their own special characteristics, but to the con-
trolled environment between the inner and outer walls. The space between
the walls of a double-wall tank can be made air-tight, so that any hole in
either wall can be quickly detected by a change in pressure or a change in the
level of a liquid filler. Alternatively, the outer wall (or the liner in a
lined system) will channel any product released from the inner wall towards a
sensor at the lowest point in the system, while simultaneously simplifying
detection by keeping ground water out of the system. Disadvantage: Can be
used only in conjunction with secondary containment, which is expensive.
4.B.2.b. Costs and Effectiveness for Leak Detection Methods
Costs for various systems of leak detection are presented in Exhibit 4.6.
The costs are built up from the initial costs, and the detection method's
per-use costs (i.e.: one tightness test is $520) multiplied by frequency of use
to yield annual costs. These are combined to give a present discounted cost of
leak detection for a thirty-year period. Automatic inventory control systems
have not been included because of the difficulty of predicting their level of
sensitivity.
The effectiveness of the different systems is difficult to predict, as it
is influenced both by the pattern of leaks expected and chance factors (such as
the behavior of vapors in the complex physico-chemical environment which exists
underground) on which little information is available. We have used the UST
Model, and a set of assumptions about the reliability of the methods, to illus-
trate the relative abilities of the systems to reduce release volumes from one
particular type of tank in one type of aquifer over a thirty-year period. The
results of this limited analysis are presented in Exhibit 4.7.
The information from the previous exhibits can be combined to show the
cost-effectiveness of the detection methods, in terms of release volumes and
plume areas avoided. This comparison is depicted in Exhibit 4.8. The exhibit
shows that the costs of monthly vapor monitoring (VW-M), yearly tightness testing
(TT-l/YR), and especially manual inventory control (MAN. INV.) are out of
proportion to their effectiveness. These are the three systems with high
continuing costs. By contrast, vapor wells with a continuous sensor (VW-C)
which have high initial costs but have little or no annual costs, are shown to
dominate the other options, as they are the most effective and the least costly
systems.
4.C. ALTERNATIVES FOR EXISTING TANKS
In some ways, the alternative actions for reducing releases from existing
tanks are more limited than for new tanks, since no choice of tank type can be
made and because some detection methods are impractical to retrofit. The choice
of actions is, however, still quite broad. Many alternatives exist for systems
to allow early detection of releases; it appears to be possible to add effective
corrosion protection to existing bare steel tanks; and the option of requiring
early retirement of tanks and replacement with new tanks is available.
4.C.I. Mandatory Retirement
Forcing early retirement of existing tanks, thereby making way for the
installation of new tank systems less prone to corrosion and undetected releases,
is obviously one effective response to the problem of existing tanks. Its
-------
4-11
Exhibit 4.6
COSTS OF LEAK DETECTION METHODS
(Per Tank, for New 3-Tank Facilities)
PRESENT
INITIAL
COST PER
VALUE OF
SYSTEM
COST
YEAR
COST, 30 YRS
MANUAL INVENTORY
$ o
$730
$14,308
TIGHTNESS TEST (ANNUAL)
0
520
10,192
TIGHTNESS TEST (3 YEAR)
0
173
3,397
FLOATING SENSOR (SAMPLED MONTHLY)
180
33
833
VAPOR SENSOR (SAMPLED QUARTERLY)
180
100 t
2,140
VAPOR SENSOR (SAMPLED MONTHLY)
180
300 t
6,060
VAPOR SENSOR (CONTINUOUS)
1753
0
1,753
LINE LEAK DETECTOR (LOW COST)
350
0
350
LINE LEAK DETECTOR (IMPROVED)
1585
0
1,585
t The annual operating cost for vapor wells represents a cost of a quarterly or
monthly visit to the site by an outside contractor (lab) to check the well sensor
at a cost of $25 a visit. The annual operating cost may be lower if the station
operator elects to check the well sensor himself; however, the level of effec-
tiveness may than deteriorate.
Exhibit 4.7
EFFECTIVENESS OF DETECTION METHODS IN REDUCING RELEASE VOLUMES AND FLOATING
PLUME AREAS OVER 30 YEARS
SYSTEM
% REDUCTION OF
RELEASE VS. BASE
DETECTION
% REDUCTION OF
PLUME AREA VS.
BASE DETECTION
BASE DETECTION ONLY
MANUAL INVENTORY
TIGHTNESS TEST (ANNUAL)
TIGHTNESS TEST (3 YEAR)
FLOATING SENSOR (MONTHLY)
VAPOR SENSOR (QUARTERLY)
VAPOR SENSOR (MONTHLY)
VAPOR SENSOR (CONTINUOUS)
LINE LEAK DETECTOR (LOW COST)
LINE LEAK DETECTOR (IMPROVED)
NA
8.9 %
46.4
31.3
41.3
63.2
65.0
65.2
31.3 (0.0 vs.
39.0 (7.7 vs.
3 yr TT)
3 yr TT)
NA
8.9 %
59.6
44.0
56.5
76.5
76.7
76.8
44.0 (0 vs. TT
45.6 (1.6 vs.
* Line leak detectors were run along with three year tightness tests;
the release and plume area reductions shown are the additional reductions in
the percentage of base detection releases and plumes attributable to the line
leak detectors.
-------
20
19
18
17
16
15
14
13
12
11
10
9
8
7
6
5
4
3
2
1
0
COST-EFFECTIVENESS BY DETECTION TYPE
NEW TANKS; AVOIDED RELEASE VOLUMES
~ MAN. INV.
BASE BASE DETECT I ON
MAN. INV. MANUAL INVENTORY ~ TT-1/VR
TT-l/YR TIGHTNESS TEST — EVER* YEAR '' " V ^
TT-3/YR TIGHTNESS TEST — EVERY 3 YEARS
FS-M FLOATING SENSOR — MONTHLY
VW-Q VAPOR WELL — QUARTERLY
VW-M VAPOR WELL — MONTHLY
VM-C VAPOR WELL — CONTINUOUS SENSOR
TT-3 YR U/LL LINE LEAK DETECTOR (LOW COST — WITH 3 YR TT)
TT-3 YR W/IMP LL LINE LEAK DETECTOR (IMPROVED — WITH 3 YR TT)
~ VW-f
~ TT-3 YR
VW-Q
D TT-3 VR W/IMP LL D ~
~ FS-M VW-C
, , , ° TT~3 VR Y,LL , 1
6 20% 40% 60%
RELEASES AVOIDED VS. BASE DETECTION
Exhibit 4.8
-------
4-13
advantages are clear: because new, protected tanks are likely to have dramati-
cally lower releases than the old and inferior tanks they replace, wholesale
replacement of tanks could almost eliminate future damages from USTs. This is
illustrated in Exhibit 4.9, which shows cumulative plume areas over time with
and without a policy of mandatory retirement five years after the promulgation
of the UST regulations.
The disadvantages of. this alternative are that it would be very costly,
and might be hard to implement without causing capacity problems in tank manu-
facturing and installation.
To judge the ultimate cost and effectiveness of many potential alterna-
tives, including programs of mandatory retirement, the costs and performance of
new tanks must be considered along with the cost and performance of the choices
made for the existing tanks. This issue is dealt with in Chapter 5, which
takes up the question of comparisons among comprehensive, integrated regulatory
options.
4.C.2. Tank Upgrades
Upgrading of existing bare steel tanks to offer some protection against
corrosion or releases is an accepted practice. These actions for existing
tanks are analogous to the choice of a method of construction for new tanks,
since they are aimed at release prevention. We have some information about the
costs of these steps, but no data on effectiveness.
o Retrofit of Cathodic Protection
o Interior Coating
o Retrofit of Lining or Partial Lining of Tanks or Pipes
Exhibit 4.10 provides a summary of what is known about cost and effectiveness
of these upgrade options.
4.C.3 Leak Detection
Leak detection options for existing tanks are essentially the same as for
new tanks, though it is generally more expensive to add leak detection devices
to existing tanks than to incorporate them into the design of a new tank system.
Exhibit 4.11 presents the costs and effectiveness of several leak detection
systems for existing bare steel tanks, as estimated using the UST Model. The
data are presented graphically in Exhibit 4.12. Again, the vapor monitoring
systems are the most attractive. Although for new tanks the continuous monitor
is more advantageous than the intermittent monitors, the continuous vapor
monitor is less advantageous compared to the intermittent monitors for existing
tanks, largely because with an existing tank there will be fewer years of use
for the system compared to a new facility.
An important side issue is the phasing in of leak detection. It will not
be practical to equip all UST facilities with leak detection immediately, due
to capacity problems in the industry and uncertainties about the best designs
for leak detection systems. Given, then, that some facilities will be equipped
with leak detection sooner than others, the question of which sites should be
given priority arises. Two considerations that should influence the strategy
for the phasing in of leak detection (and upgrading or early retirement as
well) are the expectation of release at a given facility, and the expectation of
-------
EFFECT OF REPLACEMENT AT FIVE YEARS
YEARS
Exhibit 4.9
-------
4-15
Exhibit 4.10
SUMMARY OF COST AND EFFECTIVENESS OF ACTIONS TO UPGRADE EXISTING TANKS
(Per Tank, for Existing 3-Tank Facilities)
APPROXIMATE LIKELY DEGREE OF
OPTION COST EFFECTIVENESS
RETROFIT OF CATHODIC PROTECTION S3,050 Very high
INTERIOR COATING $1,420 - $4,100 High only for
corrosive products
RETROFIT OF LINER AND SUMP MONITOR* $12,350 - $18,450 Very high
RETROFIT OF PARTIAL LINER (BOTTOM
OF EXCAVATION ZONE) AND SUMP MONITOR*... $11,100 - $17,200 Possibly very high
* Includes $900 for sump monitor.
Exhibit 4.11
COSTS AND EFFECTIVENESS OF LEAK DETECTION METHODS
(Per Tank, for Existing 3-Tank Facilities)
PRESENT
% REDUCTION OF
INITIAL
COST
VALUE
RELEASE VOLUMES
SYSTEM
COST
PER YEAR
COST **
VS. BASE DETECTION
MANUAL INVENTORY
$ 0
$730
$7,570
19.9 %
TIGHTNESS TEST (ANNUAL)
0
520
5,052
56.8
TIGHTNESS TEST (3 YEAR)
0
173
1,627
34.1
FLOATING SENSOR (MONTHLY)
500
33
1,711
24.5
VAPOR SENSOR (QUARTERLY)
500
100
1,668
57.7
VAPOR SENSOR (MONTHLY)
500
300
3,630
60.4
VAPOR SENSOR (CONTINUOUS)
2073
0
2,847
63.2
** Costs shown here are the average present value costs of detection for the
remaining life of an existing tank. The period of time over which the per-tank
costs are discounted varies depending upon when the tank fails, or is replaced
with a new tank system. In the simulation, an existing tank, on average, lasted
12 years before it failed and/or was replaced.
-------
COST-EFFECTIVENESS BY DETECTION TYPE
EXISTING TANKS; AVOIDED RELEASE VOLUMES
$7 -
$6 -
$5 -
$3 -
;i -
~ MAN. INV.
BASE
MAN. INV.
TT - 1/YR
TT - 3/YR
FS-M
VU-Q
VU-M
VW-C
$0 -£] DASE
RASE DETECTION
MANUAL INVENTORY
TIGHTNESS TEST -- EVERY YEAR
TIGHTNESS TEST -- EVERY 3 YEARS
FLOATING SENSOR — MONTHLY
VAPOR HELL -- QUARTERLY
VAPOR WELL — MONTHLY
VAPOR WELL -- CONTINUOUS SENSOR
~ TT— 1 /YR
~ VW-M
I
~ VW-C
~ FS-M DTT-3 YR
~ VW-Q
0%
20%
40%
60%
RELEASES AVOIDED VS. BASE DETECTION
Exhibit 4.12
-------
4-17
damages and risks if a release occurred at the facility. Older, unprotected
USTs and/or USTs near drinking water sources should presumably be required to
install leak detection first. A quantitative examination of the contribution
that a proper phase-in strategy can make is included in Section 5.C.3.h. which
discusses the sensitivity of the cost effectiveness results.
4.D. ALTERNATIVES FOR CORRECTIVE ACTION
4.D.I. Specification of Alternatives
As discussed earlier, damage can be avoided by preventing releases or
by halting and cleaning them up once they have occured. Options for cleaning
up the releases, referred to as corrective action, can be broken down into
immediate actions to reduce the hazards associated with releases and longer-term
actions to remove the portion of the release dissolved in the ground water.
Based on a consensus in the state programs on immediate actions to be
taken upon discovery of a release (referred to as stage-one actions), few options
are open. It is considered necessary to take certain steps, as described below.
First, once a leak is indicated or suspected, the operator must notify the
state agency and then either confirm that the system is leaking or assume that
it is leaking. The leak must be stopped immediately and steps must be taken to
mitigate any fire or safety hazard. The operator must excavate deep enough to
isolate the release, and must remove the released product within the excavation
zone and the immediately surrounding area. After an investigation of the leak
to determine whether, and to what extent, product has reached ground water or
soil outside the excavation zone, any free product mounded on the ground-water
surface (the floating plume) must be removed.
Three potential approaches for determining the required extent of cleanup
activities beyond the first stage are among those used in State UST programs.
They are: 1) fixed cleanup standards with a variance provision; 2) site-specific
standards selected with reference to the risk posed by particular releases; and
3) a predetermined class approach.
Under the first approach, owner/operators would generally be required to
clean up soil and ground water contaminated by an UST release until specified
standards are met. The standards would be national health-based standards,
including either a few indicator chemicals or all constituents for which EPA
has previously developed health-based standards. In the absence of health-based
standards, background concentrations would be used to establish cleanup targets.
A variance procedure could be used in some circumstances, however, to exempt a
particular release from the need to meet the health-based standards. Variances
could be based on showing that the health-based standards either could not be
met, and/or the release did not present a substantial hazard.
This approach has the potential advantage of providing a degree of certainty
and uniformity in cleanups and savings of the costs of setting standards on a
site-by-site basis, while allowing for variances to be used where it can be
shown that meeting the nationwide standard would be too difficult and/or unnec-
essary to protect human health. Among its disadvantages are that its implemen-
tation would require a considerable delay to allow nationwide standards to be
-------
4-18
selected; it would not be able to ensure nationwide uniformity in standards
because states could set more stringent levels and because the variance procedure
would probably have to be used frequently; and the likelihood that even with
the variance process, many sites would be subjected to overly stringent cleanups.
A second approach, and the one proposed by the EPA, is to base all long-
term (stage-two) cleanup decisions on case-specific exposure and risk assessment
techniques. Factors considered under this approach include the quantity, toxi-
city, mobility, and persistence of the product released, the exposure pathways,
the extent of contamination, levels of background contamination, and established
standards. The standards established by the site-specific risk analysis could
be adjusted under some circumstances on the basis of evidence that they are not
practically achievable. These are the same factors that would have to be
considered in order to obtain a variance under the nationwide, health-based
standards approach.
The advantages of this approach are that it would allow for all of the
flexibility of the variance option in the first approach, would save the time
required for determining nationwide standards, and would meet the need to
protect human health and the environment. One potential disadvantage is that
it could be time consuming to implement at each site. However, it is believed
that the risk analysis procedures could be streamlined so as to minimize this
drawback.
A third approach would combine the national health-based standard and the
site-specific standard approaches. It would provide for different cleanup
standards for distinct, predetermined situations or classes. This could have
advantages similar to the nationwide standards approach, in that it would
promote consistency and could avoid delays for setting standards at individual
sites, once the class standards had been set and once the class to which each
site belonged had been determined. The disadvantage of this approach is that
it would involve more time and controversy to set more than one nationwide
standard, it would be difficult and time-consuming to determine which sites
fell into which classes, and it still might not avoid the need to allow
variances in some circumstances.
4.D.2. Effectiveness of Corrective Action
It is very difficult at this time to assess the effectiveness of different
approaches to long-term corrective actions. The difficulties stem from uncer-
tainties about the efficiency of free product removal techniques, the cost and
effectiveness of dispersed plume cleanups, and the distribution of situations
that might fall into one risk and damage class or another. Therefore, rather
than try to estimate the effectiveness of different corrective action assump-
tions, across regulatory options, the analysis for the RIA holds the corrective
action option constant, assuming a case-by-case approach to long-term action in
which the dispersed plume is removed in a minority of cases. This allows for a
truer comparison of options for prevention and detection.
4.D.3 Costs of Corrective Action
The costs of corrective action are difficult to predict because they are
subject to so many influences. Costs will be much higher if more corrective
action steps are needed (for instance, if floating or dispersed plumes must be
-------
4-19
removed, rather than contaminated soil only); if release volumes are large and
the extent of the contamination is great; and if the standards for cleanup of
the soil and ground water and/or the disposal of removed soil and water are
stringent. A variety of cost estimates might be obtained even if all of the
characteristics of a release and the required cleanup were specified, simply
because of differences in technologies and markups used by cleanup contractors.
Nonetheless, EPA has made estimates of the costs of corrective action that
vary according to the seriousness of the release and the effect (number of
cleanup steps) needed to address it. Exhibit 4.13 shows the typical costs,
along with preliminary assumptions about the frequency with which a release
would fall into each category. Note that stage two cleanup actions (dispersed
plume cleanups) are assumed to be needed for only a fraction of plumes that
reach ground water (about 40 percent), and will impose a wide range of costs.
Information about the cost per action and the relative frequencies of each
type of action allow us to estimate the average corrective action expenditure.
The average expenditure, weighted by its frequency, is about $69,000. (These
figures do not include the costs of tank removal and disposal, which can be
considered part of the costs of replacing the tank rather than a corrective
action cost).
The approach used in the UST Model for determining corrective action costs
was very similar, and indeed was based on many of the same underlying assumptions
about corrective action technologies and their costs. It differs in that it
explicitly recognizes that plumes are less costly to clean up if they have not
moved far from their sources. The cost schedule used in the model is shown in
Exhibit 4.14. Based on the frequencies of releases of various sizes estimated
by the UST Model, the average cost of corrective action in response to a release
is about $54,000. This is moderately lower than EPA's weighted average estimate
of the costs of responding to a release, though the costing algorithms were
designed to approximate EPA's estimate closely. The difference is due to new
information on the size of typical floating plumes, showing them to be smaller
than previously thought in a given aquifer setting, and new estimates of the
distribution of USTs by aquifer type, showing more settings in which plumes
tend to be quite small and presumably lower in corrective action costs. The
effect of this small difference in corrective action cost assumptions on options
selection can be seen in Chapter 5, in the section on the sensitivity of the
results to changes in corrective action costs.
-------
4-20
Exhibit 4.13
EPA ESTIMATES OF APPROXIMATE COSTS AND PROBABILITIES OF CORRECTIVE ACTION STEPS
STEP
IMMEDIATE ASSESSMENT
HAZARD MITIGATION
SITE INVESTIGATION
REMOVE A DISPOSE SATURATED SOIL
REMOVE FLOATING PLUME
REMOVE DISPERSED PLUME:
LOW
MODERATE
HIGH
EXPECTED
UNIT COST
700
2,000
9,500
8,000
33,000
75,000
150,000
225,000
PROBABILITY
PER RELEASE
1.00
1.00
0.90
0.90
0.50
0.05
0.05
0.10
EXPECTED COST
PER RELEASE
S
700
2,000
8,550
7,200
16,500
3,750
7,500
22,500
TOTAL EXPECTED COST:
$ 68,700
Source: Based on data provided by John Heffelfinger, OUST, EPA
Exhibit 4.14
UST MODEL ASSUMPTIONS FOR COSTS OF CORRECTIVE ACTION STEPS
FLOATING PLUME SIZE IN SQUARE METERS:
UP TO:
(NONE) 25
750 1500 3500 7500 12500 >12500
(costs are in thousands of dollars)
INITIAL
CLEANUP AND
INVESTIGATION,
PLUS SOIL REMOVAL $15.0 $15.0 $15.0 $15.0 $15.0 $15.0 $15.0 $15.0
FLOATING PLUME
REMOVAL
DISPERSED PLUME
CLEANUP
0 11.0 24.4 27.3 33.0 39.7 50.8 70.4
0 25.2 100.8 120.8 144.2 170.8 218.7 303.0
WEIGHTED AVERAGE COST PER RELEASE ACROSS ALL HYDROSETTINGS FOR BASE CASE: $54,000
Source: UST Model Documentation, assuming half of releases are from tanks and
half are from pipes, and with reduced costs ($11,000 for floating plume
removal and $25,200 for dispersed plume) for very small plumes (less
than the excavation area of the tank, or about 25 square meters).
-------
Chapter 5
ANALYSIS OF COST AND EFFECTIVENESS OF UST REGULATORY OPTIONS
This chapter presents and discusses the results of our analysis of the
cost-effectiveness of the UST regulatory options. The framework of the analysis
is similar to that of Chapter 4 in that the costs of various regulatory options
are presented along with effectiveness estimates. The regulatory options ana-
lyzed are packaged options, in that they include rules for new tanks, existing
tanks, and corrective action. This shift makes it more difficult to illuminate
which components of a given package make the greatest contribution to cost and
effectiveness, but allows for the interactions of the various components. In
addition, the data are presented for the tank universe as a whole, rather than
on a per-tank basis.
5.A. METHODOLOGY FOR CALCULATING COSTS AND EFFECTIVENESS OF THE OPTIONS
This section briefly restates the methodology for generating the costs and
effectiveness, which was discussed in detail in Chapters 2 and 3.
5.A.1 The Tank Universe Analyzed
The cost-effectiveness results presented in this chapter are intended to
represent results over a 30-year period from all petroleum underground storage
tanks covered by EPA's proposed rule. Included explicitly in the analysis are
existing tanks in various environmental and exposure settings, of various ages
and types, as well as the tanks that will replace them over time. Although the
extremely large and varied tank population cannot be modeled in all its detail,
efforts were made to allow for the effects of as many important sources of
variability as possible. For example, we modeled the age distributions for
bare steel and fiberglass tanks as described in Chapter 3 and we modeled bare
steel tanks in three hydrosettings (sand, sandstone, and clay). Recause fiber-
glass tanks make up only 11% of the UST population and have fewer releases than
bare steel tanks, fiberglass tanks were modeled in only the most prominent
hydrosetting (sand). In simulating the UST population we used a scenario-based
approach where each scenario represents a subset of the total UST population and
is defined by a number of variables, or characteristics, such as hydrosetting,
type of monitoring, and tank type. The results were then weighted to reflect
what is known about the prevalence of the individual scenarios in the actual
population.^/
5.A.2 Measurement of Effectiveness
As described earlier, the effectiveness of the options is estimated by
simulating the life cycles of a large number of individual UST facilities with
the UST Model to determine the nature of the damages that will occur given a
set of regulations. Leak frequencies, timing, and sizes are generated by the
failure component of the UST Model. The interaction of these releases with
detection and the environment determines how long they continue, how far they
spread, and in what concentrations.
V The variables considered in building up the scenarios, the distributions
of frequencies used, and the number of tank simulations run for each scenario
are provided in Appendix C.
-------
5-2
In presenting the results of the simulations, a variety of effectiveness
measures may be used. Each has advantages and disadvantages, and none is
unambiguously superior. Generally, measures like failure frequencies and total
release volumes, which can be defined precisely and calculated in straightfor-
ward ways, are difficult to relate to the ultimate impacts, like health risks
or cleanup costs imposed. Measures more immediately related to these ultimate
impacts, though, can be computed and described only with difficulty and uncer-
tainty. This dilemma is fundamental, arising from the complex chain of events
that must occur before a failure in an UST system affects society in real ways.
The measure of effectiveness used in this analysis is an output part way
between tank failures and the ultimate damages caused by leaks: the areas of the
floating contaminant plumes. Effectiveness of the proposed rule is defined as
the part of the floating plume area that would have appeared under the base
case, but is avoided under the proposal. This intermediate measure of the
impact of UST releases has a number of advantages, including that it is fairly
straightforward to estimate, easy to visualize, and is directly (though not
linearly) related to most final damage measures. Its disadvantages are that it
can be translated into ultimate damage measures only with difficulty, it does not
include a time dimension, and it does not relate well to damage reduction contri-
buted by corrective action.
5.A.3 Costs as Presented
The same UST Model simulations used to estimate avoided plume areas also
calculate the total costs associated with the regulatory options. Costs are
limited to those phenomena affected by the proposed rule and include new tank
initial facility purchase and installation costs, detection and monitoring
costs, value of product lost, tank removal and replacement costs, and pipe
repair costs. These costs are summed, discounted to time of promulgation, and
summed across the population. A separate category of costs is the cost of
corrective action, which will differ according to the stringency of the cleanup.
More detailed lists, definitions, and derivations of cost categories are pro-
vided in the UST Model documentation.^/
All of these costs, when summed, are the costs associated with the regula-
tory options, and can be compared with the effectiveness of the option in redu-
cing the damages that would occur under the base case.
5.B. REPRESENTATIONS OF THE BASE CASE, THE PROPOSAL, AND THE ALTERNATIVES
The UST Model runs used for the cost-effectiveness analysis are intended
to characterize the proposal and other options closely in terms of major provi-
sions, but do not include every detail. The options, as described by EPA, are
simulated using the UST Model and the main differences are discussed below.
"No Further Regulation" (Base case): assumes no regulation beyond the interim
prohibition requirements; existing tanks are mostly bare steel tanks without
supplementary detection or inventory control. Tanks found to be leaking are
replaced with corrosion-resistant tanks, which are assumed to be coated and
cathodically protected.
\l Pope-Reid Associates, Inc., Final Report: Underground Storage Tank
Model, December 1986
-------
5-3
Option I (Baseline level): requires manual inventory control, periodic leak
detection within three years for existing and new tanks (five years for
corrosion-resistant tanks) , and corrosion-resistance for all replacement tanks.
Given that the regulated community is allowed to choose from a variety of detec-
tion measures, the option is modeled as though half of operators choose quarterly
vapor well monitoring, and the other half choose the less effective tank tight-
ness tests every three years (five years for corrosion-resistant tanks). Tanks
are assumed to be replaced with coated and cathodically protected tanks with
line leak detectors, and either quarterly vapor well monitoring or tightness
tests every five years.
Option II—The Proposed Rule (Enhanced baseline plus targeted upgrading): is
similar to Option I with upgrading to new tank standards within ten years,
though leak detection systems must be sampled monthly rather than quarterly and
tightness tests are not permitted after tanks are replaced. For modeling
purposes, operators are assumed to retrofit with cathodic protection and monthly
vapor wells to meet new tank standards after eight years. Replacement tanks
are assumed to be coated and cathodically protected, to have line leak detec-
tors, and to have monthly vapor well monitoring.
Option III (Baseline plus secondary containment for new tanks): requires per-
iodic leak detection for existing tanks and secondary containment with inter-
stitial monitoring for new tanks. For existing tanks, this option was assumed
to be identical to Option I; replacement tanks are assumed to be lined systems
with interstitial monitoring.
Option IV (Class Option): requires rapid replacement of existing tanks and sec-
ondary containment for replacement tanks at state-designated well-head protec-
tion areas. Tanks in other areas are required to conform to baseline standards
(Option I). It is assumed that 40 percent of the tank population is located
within a well-head protection area. Tanks located in these state-designated
areas are assumed to be fitted with continuous vapor wells after one year, and
then replaced before the fifth year with protected tanks with liners. The other
60 percent of tanks (those outside well-head protection areas) are modeled the
same as Option I. As in other options, ground water is cleaned up at 40 percent
of sites where the release has reached ground water; all of these cleanups are
assumed to be performed in the well-head protection areas.
Option V (Emphasis on prevention): For existing tanks, this option requires
manual inventory control, frequent leak detection starting in three years, and
early retirement. Replacement tanks must have secondary containment. Half of
all existing tanks are modeled with continuous vapor well monitoring, and half
are modeled with monthly vapor well monitoring and three-year tightness tests.
Existing tanks are replaced with lined systems either when they fail, or after
eight years.
Steps assumed to be taken in response to a release are the same for all options.
Where a release has occurred, an investigation and actions to reduce immediate
hazards is followed by limited removal of contaminated soil, removal of any
free product from the ground water, and ground-water cleanup at 40 percent of
sites where ground water has been contaminated.
-------
5-4
For all options (except for the "no
tory control is modeled as effective in
monthly throughput. Inventory control is
of all establishments.
further regulation" base case) inven-
detecting a loss of 0.5 percent of
assumed to be practiced at 25 percent
5.C. COST AND EFFECTIVENESS RESULTS
The cost and effectiveness results are presented in three parts. First,
the plume acres avoided by each option are presented along with costs not
counting corrective action. The dominated alternatives are identified, and the
marginal cost of avoiding damages as more stringent options are used is computed.
Next, the difference in cost rankings once corrective action costs are included
is shown.
The scales measuring effectiveness in the sections comparing costs
and effectiveness with and without corrective action are identical: plume area
avoided. Ideally, we would prefer to show some measure of the effectiveness of
the options once the benefits of corrective action were considered. Unfortun-
ately, there is little reliable information at this point on the effectiveness
of corrective action at alleviating problems once a product has escaped and
formed a plume. Until this information is available, it is reasonable to
assume that corrective action will not eliminate all of the problems associated
with contamination of soil and ground water, and that it will always be more
beneficial to avoid a given area of contamination than to clean it up. For
this reason, plume area avoided is a useful proxy for the benefits of the
standards, even if corrective action is undertaken. It also is a measure of
the health risks that are avoided during the time that the plume would have
been growing undetected, as well as other "irreversible" damages. In both
sections, the incremental cost-effectiveness of one option compared to others
is discussed where appropriate.
Finally, the sensitivity of the cost-effectiveness rankings to changes in
assumptions is discussed. Among the most important assumptions varied in this
section are choices by UST operators among techniques allowed by individual
options; the effectiveness of leak detection methods, costs of corrective
action, and phasing of the provisions of the options.
The analysis shows that all options could provide a great degree of pro-
tection relative to a "no further regulation" base case. The results could
also be used to identify the most effective and least costly option. We must
caution, though, that this use of the limited analysis presented could be
misleading because the ordering of the options in terms of cost and effective-
ness is quite sensitive to changes in assumptions (e.g., assumptions about the
effectiveness of detection methods). In addition, the identity of the apparent
"low cost option" for technical standards is significantly affected by assump-
tions regarding the appropriate level of corrective action.
5.C.I. Comparisons Without Corrective Action
5.C.l.a. Total Cost-Effectiveness Comparisons
Exhibit 5.1 shows the cost and effectiveness of the technical standards of
each of the five regulatory options compared to the base case. It does not
-------
$58
$56
-
$54
-
$52
-
$50
-
$48
-
$46
-
$44
-
$42
-
$40
-
$38
-
$36
-
$34
-
$32
$30
II
COST AND EFFECTIVENESS OF OPTIONS
30-VEAR COSTS, NO CORRECTIVE ACTION
0% 20% 40 60Vo
PERCENT OF BASE-CASE PLUME AREA AVOIDED
Source: SCI and PRA Estimate Using UST Model
-------
5-6
include the costs of corrective action regulations. Exhibit 5.1 shows that all
the options can provide substantial improvements over the base case: at least
55 percent and as much as 83 percent of the floating plume area that would
occur under the base case is avoided by each option. The more protective
options are generally shown to be more costly (taking into account any savings
in costs of repair, replacement, and lost product, but excluding consideration
of any savings of corrective action costs), since they require earlier and
more expensive upgrading of existing tanks. The data on which Exhibit 5.1 was
based are presented, for reference, in Exhibit 5.2.
The main reasons for this improvement over the base case are the inventory
control requirement, the requirement of periodic to frequent monitoring for
existing tanks in each option, and tank upgrading or replacement. Any program
requiring at least inventory control in combination with periodic leak monitor-
ing can alleviate a large portion of the potential damage from leaking under-
ground storage tanks, and early retirement or upgrade will reduce damages even
more.
5.C.l.b Incremental Cost Comparisons
As discussed earlier, it is often useful to calculate the incremental costs
per unit of increased effectiveness as we move from a given option to one that
is both more costly and more effective. When examining Exhibit 5.1, for in-
stance, we are interested in knowing how costly Option I is per unit of in-
creased effectiveness compared to the base case. By dividing the increased
cost by the reduction in the number of plumes that occur with Option I, we
obtain an incremental cost per plume of $31,654. This may be compared to EPA's
estimate of the cost of cleaning up a typical base-case plume of at least
$33,000 (for floating plume removal alone; see Chapter 4) suggesting that the
cost of Option I is worth paying when Option I is compared to the base case.
Similar comparisons can be made between Option I and Option II and between
Option II and Option V. The cost per plume avoided by choosing Option II over
Option I is only $3,862, suggesting that Option II is likely to be a cost-
effective choice compared to Option I. The next comparison, between Option II
and Option V, shows a significantly higher cost per avoided plume of $39,325.
Depending on the costs of cleaning up plumes and the value of avoiding the
residual or irreversible damages that remain even after the plumes have been
cleaned up, Option V may be a better choice than Option II. This question will
be further illuminated in the following section, which examines the options
with corrective action costs included.
Finally, it is interesting to compare Option III to Option I, even though
Option III is dominated by Option II, because this comparison shows the incre-
mental cost per avoided plume of requiring secondary containment for replacement
tanks. The incremental cost (which, like all of the costs presented here, is a
discounted present value cost) comes to $61,723 per plume. This is a high cost
compared to the costs per plume avoided for the other comparisons. However, it
still might be worth paying for the costs of secondary containment under some
circumstances.
The incremental cost comparisons, including comparisons after corrective
action costs are included, are presented in Exhibit 5.3.
-------
5-7
Exhibit 5.2
SUMMARY OF SIMULATION RESULTS FOR CURRENT UST POPULATION
(1.4 MILLION TANKS)
TOTAL
RELEASE
VOLUME NUMBER OF
(Millions INCIDENTS
of Gallons) (Thousands)
NUMBER OF
NUMBER OF PLUMES
PLUMES > 25 m2
(Thousands) (Thousands)
BASE CASE 8,647 2,668 1,975 1,438
OPTION I 4,055 2,921 1,913 999
OPTION II 2,912 2,158 1,390 590
OPTION III 3,889 2,645 1,751 944
OPTION IV 2,866 2,114 1,396 671
OPTION V 1,530 1,271 844 289
PLUME
AREA COST COST
(Thousands % AREA NO CA W/CA
of Acres) AVOIDED ($Bil1ions) ($Bi11ions)
BASE CASE 190 0.0% $31.00 $120.81
OPTION I 87 54.4% 32.96 113.90
OPTION II 62 67.5% 34.98 94.75
OPTION III 85 55.4% 42.96 118.56
OPTION IV 61 67.8% 44.97 95.04
OPTION V 31 83.5% 56.47 97.08
-------
5-8
Exhibit 5.3
INCREMENTAL COST EFFECTIVENESS COMPARISONS
INCREMENTAL DISCOUNTED COST PER:
WITHOUT CORRECTIVE ACTION:
FROM BASE CASE TO OPTION I
FROM OPTION I TO OPTION II
FROM OPTION II TO OPTION V
FROM OPTION I TO OPTION III
PLUME
AVOIDED
$31,654
3,862
39,325
61,723
PLUME
> 25 M2
AVOIDED
$
4,476
4,491
71,424
182,063
PLUME
ACRE
AVOIDED
$ 18,980
80,900
706,228
5,068,266
GALLONS OF
RELEASE
AVOIDED
0.43
1.77
15.55
60.35
WITH CORRECTIVE ACTION:
FROM OPTION II TO OPTION V $4,264 $ 7,745 $ 76,576 $ 1.69
FROM OPTION I TO OPTION III 28,785 84,908 2,363,669 $ 28.15
-------
5-9
5.C.2. Changes in Comparisons When Corrective Action Costs Are Included
5.C.2.a. Total Cost-Effectiveness Comparisons
Exhibit 5.4 shows the dramatic change in both absolute and relative costs
of the options when the costs of EPA's corrective action requirements are
included. If corrective action costs are included in the base case, it jumps
from the lowest in cost to the highest in cost, this is because it is so
lacking in protection that the cost of cleaning up the plumes it allows to
occur is extremely high. At the opposite extreme, the most costly and protec-
tive technical options are revealed to save more in avoided corrective action
than their incremental cost for prevention and detection.
Exhibit 5.5, by focusing only on the options, makes it easy to compare the
relative cost and effectiveness between the options. Options II, IV, and V are
shown to be the most cost-effective options. This illustrates the effectiveness
of both early retirement, or upgrade, of bare steel tanks and frequent leak de-
tection. The relative cost effectiveness of Option V also illustrates that the
most effective regulatory alternatives can, in theory, be close to the least
costly in the long run. This is because effective leak detection and early
replacement with corrosion-protected, lined tanks eliminates a large portion of
potential corrective action costs. This result, of course, depends on the unit
cost and intensity of corrective action assumed to be appropriate and does not
reflect implementation concerns associated with the options.
Option III is the most costly alternative because tanks are replaced with
secondary containment only after they leak. Corrective action costs are higher
than the corrective action costs for the options requiring mandatory upgrade or
retirement before the existing tank fails. Option III also has high facility
costs due to the secondary containment requirement for new tanks. Though
Option III is effective, eliminating 56 percent of the potential plume area, it
is more costly than other options which have the same or better effectiveness.
Option IV, which is a combination of the less-stringent Option I at some
sites and a very stringent set of rules at other sites, is shown to be virtually
the same in cost and effectiveness as the moderately-stringent Option II. That
these options are so close together is largely coincidental. Small changes in
some of the assumptions of the analysis or in the definition of Option IV might
demonstrate that a class option could be the least-cost option, before accounting
for implementation costs.
One important note regarding Exhibit 5.5: it tends to make the cost differ-
ences and the differentials in effectiveness between options appear dramatic if
one does not keep in mind the small scale. In Exhibit 5.3 it is more obvious
that, compared to the base case, any of the options can make a large difference
in the damages that can be attributed to leaking underground storage tanks in
the near future; the differences across the options are somewhat smaller by
comparison. In addition, the differences highlighted by 5.5 may not be reliable,
because of the sensitivity of the exact results to changes in assumptions.
This issue is discussed in Section 5.C.3.
5,C.2.b Incremental Cost-Effectiveness Analysis
Referring to the last two lines of Exhibit 5.3, we see the extent to which
the added protection provided by secondary containment for replacement tanks
-------
COST AND EFFECTIVENESS OF OPTIONS
30-VEAR COSTS, WITH CORRECTIVE ACTION
cm
2tm eovo
PERCENT OF BASE-CASE PLUME AREA AVOIDED
80 %
Source: SCI and PRA Estimate Using UST Model
-------
$120 -r
$118 -
$116 -
$114 "I
$112 -
$110 -
$108 -
$106 -
$104 -
$102 -
$100 -
$98 -
$96 -
$94 -
$92 -
$90
III
COST AND EFFECTIVENESS OF OPTIONS
30-YEAR COSTS, WITH C.A., DETAIL
54%
ll .'v
58%
62% 66% 70% 74%
PERCENT OF BASE-CASE PLUME AREA AVOIDED
78%
82%
Source: SCI and PRA Estimate Using UST Model
-------
5-12
and/or early retirement fall short of paying for themselves in reduced correc-
tive action costs. We can interpret the results as follows: those plume-
related damages that cannot be reversed by corrective action would have to be
valued at more than $4,264 per plume to make Option V attractive relative to
Option II, and $28,785 per plume to make Option III preferable to Option I.
5.C.3 Sensitivity of Results to Assumptions
The cost and effectiveness of the options is strongly affected by several
of the assumptions used in the analysis of the options: operators' choices for
meeting the requirements of an option, (e.g.: replacement rather than upgrad-
ing), corrective action costs, detection effectiveness, retrofit effectiveness,
hydrosetting distributions, discount rate, and phasing deadlines for implementa-
tion. The sensitivity of the results to these issues is discussed in the sec-
tions that follow.
5.C.3.a. Option Choice
One important source of uncertainty in predicting the effects of the op-
tions is that many options allow choices. Option II, for instance, allows the
operator to choose between monthly vapor wells and tightness tests every three
years. It is assumed that half of tank owners choose each method, but there is
no way to be sure how the owners would actually choose. These choices are
important because the two methods result in a difference in cost and effective-
ness for Option II, as illustrated in Exhibit 5.6. Exhibit 5.6 shows the
options as modeled, and in addition it shows where Option II would fall if all
facility owners chose tightness tests (Il-tt) and if all owners chose vapor
wells (II-vw). The difference is very small in terms of cost, but larger in
terms of effectiveness. Still, it affects the ranking of Option II only in
comparison with Option IV.
5.C.3.b. Sensitivity to Corrective Action Costs
The level of corrective action required and the estimated cost of correc-
tive action will affect both the absolute cost of the options and the relative
cost between options. As corrective action costs increase, the overall (abso-
lute) cost of the options increases. Also, increasing the estimated costs for
corrective action will cause more stringent options, those requiring frequent
leak detection or early upgrading/retirement for existing tanks and/or secondary
containment for new tanks, to appear to be (relatively) more cost effective.
Frequent leak detection, early retirement and secondary containment decrease
the likelihood of having to do extensive corrective action, and therefore the
total cost advantage of these options compared to less-protective options will be
enhanced as per-incident corrective action costs rise. This effect is illus-
trated in Exhibit 5.7, which shows the effectiveness of the options compared to
their costs calculated three ways: using the UST Model estimates of per-incident
corrective action costs (which, as discussed in Chapter 4, are slightly lower
than EPA's current estimates of costs); using per-incident costs half as great
as the UST Model costs; and using costs twice as high. Under the low corrective
action cost scenario, Option II has the lowest total cost, and two of the
options (III and V) are actually higher in total cost than the base case (even
if corrective action costs are attributed to it). Under the high cost scenario,
the relative positions of the options change dramatically: IV and V drop below II
in total cost, and III drops below I. This shows that more protective steps,
-------
COST AND EFFECTIVENESS OF OPTIONS
VARIATION DUE TO DETECTION CHOICES
$120
$118
$116
$111
$112 -
$110 -
$108 -
$106 -
$101 -
$102 -
$100 -
$98 -
$96 "
$94 -
$92 -
$90
III
54%
ll-tt ~
II.
¦I
A ll-vw
58%
EPA OPTIONS
—I 1 1 1 1 1 I I I I
62% 66% 70% 74% 78%
PERCENT OF BASE-CASE PLUME AREA AVOIDED
~ ll-T.T. ONLY 4
Source: SCI and PRA Estimate Using UST Model
82%
ll-WELLS ONLV
-------
COST AND EFFECTIVENESS OF OPTIONS
VARIATIONS ON CORRECTIVE ACTION COSTS
$220
$210
$200
$190
$180
$170
$160
$150
$140
$130
$120
$110
$100
$90
$80
$70
$60
I
BASE, 2 X EPA C.A.
I, III
-A ~ BASE W/EPA CA
~ III
~ I
BASE, 1/2 EPA C.A.
All
AIV
~II ,IV
IIV
II
—I 1 1 1 1
20
-------
5-15
including secondary containment for replacement tanks, could be cost-effective
if unit corrective action costs turn out to be higher than current estimates. In
addition, it means that if hydrogeological settings resulting in unusually high
corrective action costs could be identified a priori, then more stringent
options could be cost-effective in those settings even if they are too expensive
to require everywhere.
Total corrective action costs could vary from those presented here even
if the unit costs of given corrective action steps (excavating contaminated
soil, floating plume removal, ground-water cleanup) have been estimated correc-
tly. Current EPA estimates are that ground-water cleanups will be required for
40 percent of contamination, but it is possible that fewer cleanups (or more
cleanups) will ultimately be performed. Exhibit 5.8 shows the changes in the
rankings of the options assuming ground water is cleaned up at 20 percent of all
sites instead of the current EPA assumption of 40 percent. Again, the less-
protective options look relatively better compared to the more stringent
options, with Option II definitely lowest in cost.
5.C.3.C. Sensitivity to the Effectiveness of Retrofitting Cathodic Protection
One of the most important features of the proposed option (Option II) is
the requirement that existing tank systems be upgraded to new tank standards
within ten years. As described earlier in this chapter, the requirement that
tank systems be protected from corrosion was assumed to be met by retrofitting
cathodic protection after eight years, a step that was assumed to be completely
effective in stopping localized corrosion. (Many tank operators would actually
choose to replace their tanks instead of retrofitting them. This choice was
not modeled because it goes beyond the requirements of the proposal.) Because
there is no extensive body of data on the effectiveness of retrofitted cathodic
protection (RCP), it is at least possible that it will not work as well as it is
modeled. To show the sensitivity of the results to assumptions about the
effectiveness of RCP, a few additional model runs were made that allow us to
make rough comparisons between the proposal with 100% effective RCP (Option
11-100% RCP) and the same proposal with RCP operating at an arbitrarily selected
lower effectiveness level of 50% (Option 11-50% RCP). In other words, at half
of all tanks where CP is added to existing tanks, it imposes costs but has no
effect at all on corrosion (a fact which is assumed not to be detectable in the
inspection of the CP system). V
The results show that Option 11-50% RCP would fall roughly halfway between
the Option 11-100% RCP, and Option I. Just over 2% less base-case plume area
would be avoided compared to Option 11-100% RCP. The costs of lost product and
added corrective action (mostly the latter) would be about $7 billion higher,
and added tank replacement due to increased existing tank failures would add
about $3 billion. The current measured difference between Option 11-100% RCP
and Option I is about S20 billion, so Option II would still be superior to
Option I even if RCP were effective for only half of the tank population. On
the other hand, Options IV and V would clearly dominate Option II in terms of
cost and reduced plume area if RCP worked only half the time.
}_/ The added model runs were not comprehensive. Rather, they examined only cases
in which the existing tanks were bare steel. In addition, the runs for the
replacement tanks were not redone; instead, the results for the replacement
tanks were approximated closely using very similar replacement runs for Option I.
-------
COST AND EFFECTIVENESS OF OPTIONS
Ground Water Cleanup at 20% of Sites vs. at 40% of Sites
$120
$118 -
$116 -
$114 -I
$112 -
$110 -
$108 -
$106 -
$104 -
$102 -<
$100 -
$98 -
$96 -
$94 -
$92 -
$90 -
$88 -
~ Hi
$86
\l .'V
~ iv
54%
"i 1 1 1 1 1 1 r
58% 62% 66% 70% 74%
PERCENT OF BASE-CASE PLUME AREA AVOIDED
78%
82%
G.W. Cleanup at 40% of Sites ~ G.W. Cleanup at 20% of Sites
Source: SCI and PRA Estimate Using UST Model
-------
5-17
5.C.3.d. Sensitivity to the Effectiveness of Leak Detection
Much of the effectiveness of the proposed option is attributable to the
use of leak detection to allow leaks to be stopped either before they contamin-
ate ground water or while the extent of the contamination is still small.
The rapid detection of small leaks from USTs is, however, a difficult task
under field conditions. It can be taken for granted that any given detection
method will in some cases fail to detect leaks of the magnitude that they were
designed to handle. For this reason, tightness tests and vapor wells were not
modeled to be perfect.
Unfortunately, while test data could be used to establish that per-test
reliability of tightness tests is at least 90% (which is the level used in the
modeling), no comparable information could be located for vapor wells. EPA's
decision, based on engineering judgment, was to assume that vapor wells could
be expected to work as designed at between 50 and 90 percent of the sites at
which they were installed. The model runs assumed that vapor wells worked at
70 percent of sites.
A number of additional runs were performed assuming 50 percent effective-
ness or 90 percent effectiveness to test the sensitivity of the results to this
important parameter. The results show that the effectiveness of Option II
varies significantly with the assumed reliability of vapor wells, while the costs
of the option, including corrective action, are relatively unaffected. Exhibit
5.9 shows a cost-effectiveness display similar to that of Exhibit 5.6 (which
concerned sensitivity to the choice of detection methods) except that it assumes
that all existing tanks are bare steel. It also omits Option IV for clarity,
and includes two added data points: II-vw@90% and Option II-vw@50%. These two
points indicate the cost and effectiveness of Option II if all operators chose
vapor wells, and if vapor wells worked at 90 percent and 50 percent of sites,
respectively. £/ The modeled differences in vapor well effectiveness are more
important than the differences between vapor wells and tightness tests. The
differences between Option II-vw@50% and Option II-vwP90% are of about the same
magnitude as the differences among Options I, II, III, IV, and V.
The differences in costs are smaller, apparently because the primary
effect of properly functioning monthly vapor wells is to reduce the size of
plumes rather than to prevent them. Costs of corrective action do not show a
strong enough dependence on plume size to allow monthly vapor wells to have a
large effect on the costs.
V As with the sensitivity runs for the effectiveness of RCP, these runs were
performed only for bare steel and existing tanks. Thus, the comparisons approxi-
mate closely, but not exactly, the results that would be obtained if all of the
runs had been redone.
£/ Note that for the other options, the reliability of vapor wells is held con-
stant at 70 percent, and that of tightness tests is assumed to be 90 percent for
al 1 of the runs.
-------
$1 22
$1 20
$118
$11 6
$114
$112
$110
$108
$106
$104
$102
$100
$98
$96
MPACT OF DETECTION TYPE, EFFECTIVENESS
ASSUMING 100% BARE STEEL EXISTING TANKS
I
~
-vw@50%
X
II—tt
t
I—vw@70%
-vw@9 0%
A
T
—I—
78%
V
~
—I—
82%
T
T
T
T
T
T
%
58%
62% 66% 70% 74%
PERCENT OF BASE CASE PLUME AREA AVOIDED
Exhibit 5.9
-------
5-19
5.C.3.e. Sensitivity of Option Rankings to Estimates of Hydrosetting Distribu-
tions
The basic analysis of the cost effectiveness of the options assumes that
USTs are distributed across hydrogeological settings with three representee
soil (or subsoil) types: sandstone/limestone/shale; sand; and clay/silt. These
media types have moderate, high, and low degrees of permeability, respectively.
The modeling assumed that about 21 percent of USTs are found in the moderately
permeable settings, 40 percent in the highly permeable settings, and the remain-
ing 39 percent in the low permeability settings. These distributions were
based on careful assessments of available county-level data on USTs and hydro-
geological characteristics, but it is still possible that they differ from
the (unknown) actual distributions.
The permeability of the subsurface media surrounding an UST can have
significant effects on the relative effectiveness of options for prevention and
detection of releases because it affects the time it takes for a release to hit
ground water and effects the ultimate size of any floating plumes. Thus, an
option that is valuable if it detects leaks before they contaminate ground
water or before the contamination spreads might be much more effective if most
USTs are in low permeability settings rather than in high permeability settings,
depending on the time needed for the detection method to catch a leak. For this
reason, the sensitivity of the ranking of the options to differences in the
distribution of settings was assessed by re-weighting and combining existing
model runs in the three settings. Exhibit 5.10 shows the relative costs of the
options assuming three different setting distributions: first, the distribution
developed by PRA Inc. for the analysis; next, a distribution with many more
highly permeable settings and fewer low permeability settings (half of the
clay/silt settings were changed to sand settings); and finally one with many
more low permeability settings (half of the sand settings were moved into the
clay/silt category). The exhibit shows that the cost rankings change only for
groups of options that are very close together in costs: Option II is the
least expensive option with relatively more low permeability settings, but it
would be somewhat more costly than the more stringent options if relatively
more settings were in the high permeability category. Relative differences in
effectiveness are similarly small.
In conclusion, the basic findings of the cost-effectiveness analysis are
robust even to substantial changes in soil/subsoil types. Tailoring of options
with respect to setting may still be worthwhile if more stringent requirements
could be imposed at sites with more permeable hydrogeological settings.
5.C.3.f. Sensitivity to Changes in the Discount Rate
Costs in this analysis were discounted to the first year of the regulations
at three percent per year, the rate chosen by EPA to represent the difference
between the value to society of a dollar today compared to a dollar next year.
The discount rate is important because many of the regulatory choices involve
the decision to spend money now in order to avert costs that would appear many
years in the future.
-------
RELATIVE COSTS OF OPTIONS BY SETTING
(BARE STEEL EXISTING TANKS ONLY)
1 10%
AS WEIGHTED
ZZI '
MORE CLAY
^ IV
MORE SAND
\KKI v
Exhibit 5.10
-------
5-21
Because the use of different discount rates could change the relative
attractiveness of the options, we recalculated the present value of the total
costs of the options at two additional rates: ten percent per year, the rate
recommended by 0MB, and zero percent. Total costs are much lower with a ten
percent discount rate than with the lower rates, because costs that will be
incurred years in the future (e.g., costs of tank replacement and corrective
action) are dramatically reduced at high discount rates. Perhaps more impor-
tant, from the standpoint of option choice, the individual options are affected
differently by the discount rate. Exhibit 5.11 shows that the more stringent
options, IV and V, are much more attractive compared to Option II at low dis-
count rates. Careful attention should be paid, therefore, to the question of
whether the appropriate discount rate is high or low.
5.C.3.g. Sensitivity of the Choice of Secondary Containment to the Prevalence
of False Reports of Leaks
We have already discussed the fact that some leak detection methods are
less likely than others to miss an actual leak. There is also concern that
some methods will be more likely to indicate that a release is taking place
when it is not. This can be called a "false positive" or a "false alarm." Two
of the most likely cases in which this could happen are if a tightness test is
used on a tank with a loose fitting at a connection above the highest levels at
which that product is stored, and if a vapor well is used at a site where very
small surface spills take place, or where there is residual contamination from
an earlier release. A false alarm does no damage to the environment, but could
impose significant costs on operators who must either expend resources on start-
ing a corrective action procedure or on trying to confirm the existence of a
release.
The questions raised by the added costs imposed by false alarms are most
important when we compare two options that we expect to differ greatly in
vulnerability to sounding false alarms. For instance, because the interstitial
monitor of a system with secondary containment can be easily isolated from
outside influences, it might be much less likely to sound false alarms than a
vapor well, which operates outside a single-wall tank. Though secondary con-
tainment is expensive, it may be less costly than single wall systems with
vapor wells, even excluding the costs of corrective action, if it prevents
expensive false alarms.
Unfortunately, there is no reliable information available on either the
costs or frequencies of false alarms for various detection systems. We can,
however, calculate the effect of different possible values for the costs and
frequencies of false alarms on the choice of options as a first step toward ad-
dressing this issue.
Exhibit 5.12 shows the impact of a set of possible false alarm frequencies
for given present value costs per false alarm on the relative merits of two
options: Option I and Option III. These options differ only in that Option III
requires replacement systems to employ secondary containment (assumed to be a
liner), while Option I requires single-wall, protected tanks and leak detection.
Option III is more costly, without accounting for the corrective action costs
it would save, but would prevent most of the plumes that would occur under
Option I.
-------
EFFECT OF DISCOUNT RATE ON OPTIONS
#170
[] BASE
(/)
Z
Q
£D
h-
W
O
O
LI
D
_J
H
Z
LlJ
(fi
Id
a:
0.
a:
<
UJ
>-
Q
n
$160 -
$150 -
$1 40 -
$130 -
$120 -; BASE
$110 -
$100 -
$90 -
$80 -
$70 -
$60 -
$50 -
$40 -
$30 ->BASE
$20 -
$10 -
$0
0%
~
~
~ I
+
+1
¦aiv
-HI.IV
*Vn
^ II. IV
20%
0% DISCOUNT RATE
40%
PLUME AREA AVOIDED
60%
3%
~ V
+ V
rv>
ro
O V
80%
10%
Exhibit 5.11
-------
5-23
Exhibit 5.12
EFFECT OF ASSUMED COSTS OF FALSE ALARMS ON INCREMENTAL COSTS
Incremental Cost Per Plume (of any size) Avoided between Options I and III
ASSUMED NUMBER ASSUMED P.V. OF
OF FALSE ALARMS COST PER FALSE ALARM:
PER SYSTEM
OVER 30 YEARS $1,000 $1,500 $3,000
0 $ 61,723 $ 61,723 $ 61,723
2 44,439 35,798 9,873
4 27,156 9,873 (41,978)
6 9,873 (16,052) (93,828)
-------
5-24
The table shows the discounted incremental cost of choosing Option III over
Option I per plume avoided. Unless the discounted costs of corrective action
are substantial, or false alarms are a significant problem for Option I, Option
III will not be cost-effective compared to Option I.
If, however, we expect two more false alarms under Option I per thirty
years, no false alarms under Option III, and a cost per false alarm of SI,500,
the incremental cost per plume avoided is cut almost in half. This is because
the net cost of choosing Option III over Option I is reduced by several thousand
dollars. If four more false alarms are expected (or if each one imposes costs
twice as high), the two options would be nearly equal in costs excluding correc-
tive action: the incremental cost per plume avoided for Option III would drop
below $10,000. With the assumption of six more false alarms per thirty years,
each at a cost of $1,500, Option III dominates Option I, saving money and
reducing the number of plumes. The fact that the false alarm issue shows the
potential to swing the cost rankings of single-wall construction vs. secondary
containment argues that research into the expected frequency of false alarms
and their costs should be part of any investigation of leak detection methods.
5.C.3.h. Phasing of Option Provisions by Age or Risk
It seems clear that the sooner the detection and upgrading provisions of
the proposed regulation (as well as similar provisions of the other options)
are implemented, the more damage from leaking USTs can be avoided. The proposal
allows leak detection to be implemented over the next three to five years, and
upgrading over the next ten years, largely because it is recognized that capac-
ity problems would prevent the industry from implementing detection and up-
grading al 1 at once.
The fact that the industry will end up complying with the regulations
slowly, over a number of years, suggests that it might be important to control
which tanks meet the regulations sooner, and which ones later. That is, the
regulations could be phased in over a number of years, with provisions to
ensure that the tanks posing the most danger, if left alone, are brought into
compliance first.
Two attractive systems for the phasing of leak detection and upgrading are
by tank age and by potential risk imposed by a release. Under the first system,
older tanks could be required to be in compliance sooner than younger tanks.
Under the second, tanks within a relatively short distance of drinking water
sources (the same tanks at which, it is assumed, ground-water cleanups would be
needed) would comply first.
To test the potential of these phasing plans, some additional runs of the
UST Model were made. Phasing by age was tested only for Option II, only for
bare steel existing tanks, and in only one hydrosetting (sandstone/limestone/
shale, which has moderate permeability). For the oldest 36% of tanks in the
run, those 25 years or older, detection was assumed to be implemented in two
years instead of the three years assumed in most of the runs for Option II.
Similarly, tank upgrading was assumed to take place after five years for the
older tanks, instead of eight years. For the younger tanks, detection and
upgrading were delayed to the fourth and tenth years respectively.
-------
5-25
Phasing of Option II by risk was examined in a manner very similar to that
used for phasing by age. The differences being that, for risk, phasing all three
hydrosetting types were modeled. Age was ignored in determining which tanks
would phase in detection and upgrading at years 2 and 5 instead of 4 and 10.
It was also assumed that dispersed plume cleanups are required only for those
tanks where detection and upgrading are phased in early.
The results for phasing by age are ambiguous. Compared to an unphased
Option II, age-based phasing reduces cumulative releases and plume areas moder-
ately, with plume area avoided in comparison to the base case rising from 67
percent to almost 72 percent. Costs after corrective action, however, turned
out to be marginally higher under age-based phasing, largely because the total
number of plumes grew slightly. The results on comparative costs are probably
sensitive to the degree to which corrective action costs rise with the magnitude
of the release.
The results for phasing by risk are quite different from the age-based
results. Plume areas are barely reduced at all, since no attempt was made in
this phasing scheme to implement the regulations sooner for tanks that are more
likely to fail. Total costs, however, are lower by a substantial S4.65 billion,
or are lower by over 7 percent of the total costs of the regulations for existing
tanks. The cost savings arise from the fact that it is assumed the regulations
would be phased in first at those sites where ground-water cleanups would be
required in the event of a plume.
-------
Chapter 6
ECONOMIC IMPACTS OF LIST TECHNICAL STANDARDS AND REGULATIONS
6.A INTRODUCTION AND METHODOLOGY
The regulated community affected by the proposed regulations can be
divided into three sectors, each of whose impacts can be evaluated separa-
tely. They are, in order of decreasing number of tanks owned:
o Facilities using USTs for storing motor fuels for the retail market
(see Section 6.B);
o Facilities using USTs for storing motor fuels for nonretail purposes
(see Section 6.C); and
o Facilities using USTs for storing regulated chemicals (see Section 6.D).
Of these, the retail motor fuel sector presents the greatest potential for
economic dislocation because:
o There are no substitutes for USTs;
o There are at least three USTs per facility (one for each grade of gaso-
line) ;
o IJST costs represent a significant fraction of capital and operating
costs for the establishment; and
o There are many small establishments (though economic impacts are clouded
by questions of ownership and operation for the retail establishment,
the land occupied, and the tanks themselves).
Thus, the economic analysis has concentrated on the retail motor fuel sector.
The other two sectors are addressed in less detail.
Methodology
The EPA Guidelines for performing regulatory impact analysisV recommend
an analytical approach based on combined financial and market analysis, con-
sisting of the following steps:
o Segment the regulated industry into groups by relevant characteristics;
o Perform baseline financial analysis on segments or on model plants;
o Separate resource costs from transfers, such as taxes, that distribute
the resource costs to parties other than the directly regulated commu-
nity;
V EPA, Office of Policy Analysis, "Guidelines for Performing Regulatory
Impact Analysis," EPA-230-01-84-003, December 1983.
-------
6-2
o Estimate effects of cost pass-throughs, including the assumption
(and estimated likelihood) of no pass-through.
o Where data allow, perform discounted cash flow analysis or return
-on-investment analysis.
The results will estimate the impacts of regulatory costs on firm or facility
revenue and profit for the segments of the regulated industry. This is the
basic approach here.
According to the EPA guidelines, the general framework to be used is
based on the static partial equilibrium model of supply and demand relationships
in the affected markets. In this model, if regulations on suppliers increase
the costs of providing a particular good or service, the market price of the
affected good or service will increase and the amount provided to the market
will decrease at any given price. The magnitude of the effects will depend
on the relative sensitivity of supply and demand to changes in price (that is,
the price elasticities of supply and demand). Economic effects (i.e., changes
in profitability, plant closures, employment, inflation, capital availability,
etc.) all flow from the changes in prices and quantities predicted by appro-
priate shifts in supply/demand schedules within the partial equilibrium frame-
work. At a minimum it is helpful to examine the boundary conditions of perfectly
elastic demand (full absorption) and perfectly inelastic demand (full pass-
through), as shown in Exhibit 6.1.
For the three sectors addressed here, the boundary condition of full cost
absorption is likely to be of most concern. This is the case because UST
regulatory requirements will not affect any of these sectors uniformly. For
the retail motor fuel segment, the following observations are pertinent:
o UST regulatory costs are independent of quantity of gasoline pumped.
Therefore, regulatory costs per gallon are likely to be much less for
high volume stations, thus limiting the potential for pass-through.
o UST regulatory costs per station depend significantly on whether or not
a release occurs and the costs of responding to it. These costs will
not be spread evenly over stations, thus making it more likely that
those who incur these costs will not be able to pass them through.^/
o Some stations are much further along in release prevention/detection
programs than others, thus potentially spreading the regulatory burden
unevenly and therefore limiting the likelihood of pass-through.
o Although market demand for retail gasoline is relatively inelastic,
demand at individual stations can be highly sensitive to price (i.e.,
there are few substitutes for gasoline, but there are many substitutes
for the gasoline from an individual station).
V This assumes that insurance is not available to cover corrective action
costs. If such insurance were available, corrective action costs could be
smoothed over output and therefore be treated, as variable costs.
-------
6-3
Exhibit 6.1
BOUNDARY CONDITIONS FOR ECONOMIC IMPACT ANALYSIS
a. Perfectly Elastic Demand (Full Absorption)
-------
6-4
These reasons suggest that, particularly in the short run, the economic analysis
for the retail motor fuels sector needs to focus on the scenarios of cost-absorp-
tion and its implications for potential exit from the industry. Similarly, for
the other sectors, USTs are not spread uniformly through the affected industries,
suggesting that full absorption is also the primary scenario of concern.
The importance of estimating potential exit attributable to regulatory
alternatives suggests a need to predict expected exit in the absence of regula-
tory change. For the retail motor fuels segment, this is an important part of
the baseline. It is necessary to distinguish these independently caused trends
—notably, a continuous decrease in number of retail outlets, resulting in higher
sales per station—from the added impacts of UST regulations. It should also
be noted that two other environmental regulations, affecting used oil and vapor
emissions controls, are also expected to be implemented during the early years
of UST regulation. The costs of UST regulations, added to the costs of these
two new regulations can be expected to affect the viability of facilities and
firms in the motor fuel retailing industry, but not all business failures
and/or closings during the early years of UST regulation can be attributed to
UST regulations.
An additional factor which complicates the economic analysis for retail
motor fuel is the complex pattern of ownership and operation for stations,
tanks, and the land occupied. This makes it much more difficult to say pre-
cisely who bears the initial incidence of the regulatory burden and therefore
muddies the issue of who bears the ultimate burden. This is explained in more
detail in Section 6.B.
For the nonretail petroleum and chemical UST-using sectors, the key
analytical tool used is the screening analysis. This screening analysis
examines levels of regulatory costs relative to profit levels for UST-using
industries (defined as four-digit SIC codes). It therefore serves as a rough
indicator of the potential for economic dislocation. However, it is safe to
conclude that these sectors will not undergo as much economic dislocation as
retail motor fuel because:
o Only a fraction of establishments maintain USTs (thus suggesting the
existence of substitutes for USTs);
o The average number of USTs per establishment is generally less than
three; and
o USTs represent a smaller fraction of total costs for these industries
than for retail motor fuel.
However, corrective action burdens may affect the viability of some firms (both
large and small) because a firm cannot escape corrective action for existing
releases by substituting away from USTs. Nonretail petroleum tanks are
addressed in Section 6.C while chemical tanks are addressed in Section 6.D.
-------
6-5
6.B IMPACTS ON UST-OWNING FIRMS IN THE RETAIL MOTOR FUEL MARKETING SECTOR
6.B.1 The Retail Motor Fuel Marketing Industry: Current Status
and Future Trends
The retail motor fuel marketing industry is composed of a large number of
firms that are widely diverse in their financial and operational characteris-
tics. In 1984, the base year for this study, the industry was composed of
almost 90,000 firms owning 193,000 retail motor fuel outlets and 43,UU0 firms
operating retail motor fuel outlets that they leased from their owners. 1/
The firms owning retail motor fuel outlets range in size from some of
the largest corporations in the world to very small businesses with
no reported payroll.
The firms owning retail motor fuel outlets may own as few as one
outlet or as many as several thousand outlets. Exhibit b.Z illus-
trates the distribution of the number of outlets owned by firms en-
gaged in the retail motor fuel marketing sector.
Most firms owning more than one outlet engage in lines of business
other than retail motor fuel marketing. Some engage in other types
of retail operations, such as grocery sales or car washes. Many are
involved in the wholesale marketing of gasoline or other petroleum
products. The largest firms in this industry may be involved in all
aspects of petroleum production, refining, distribution, and market-
ing. Such large firms are also often in other diverse lines of busi-
ness (e.g., production and sale of chemicals, steel, etc.).
Retail motor fuel outlets vary substantially from one another. Some out-
lets (i.e., pumpers) specialize in low-cost, high-volume fuel sales and pro-
vide few if any ancillary services. Others (e.g., convenience stores) provide
retail motor fuel primarily as an item to be purchased at the same time that
customers purchase food, newspapers, or cigarettes. Other outlets specialize
in providing a wide range of automobile-related services (e.g., repairs and
routine maintenance).
Because of this diversity, there is no one data source that provides fin-
ancial and operational information for firms or facilities engaged in retail
motor fuel marketing as we have defined it. The Department of Commerce, for
example, compiles data for firms deriving SO percent or more of their revenues
from the sale of motor fuels, but EPA could not use this definition because it
does not include a substantial proportion of the UST-owning firms that sell
retail motor fuels. Instead, EPA used information from government data
sources, individual firms, and industry associations to construct the owner-
ship and operational profile for retail motor fuel outlets shown in Exhibit
6.3. J/ The types of owners of retail motor fuel outlets are defined below.
1/ Lundberg Letter, Vol. 13, No. 1, Nov. 1, 1985.
-------
6-6
Exhibit 6.2
DISTRIBUTION OF THE NUMBER AND PERCENTAGES OF
OUTLETS OWNED BY FIRMS OF VARIOUS SIZES ENGAGED
IN RETAIL MOTOR FUEL MARKETING
Total Number of
Number of Outlets Number of Firms Outlets Owned by
per Firm in This Group Firms in This Group
1
80,304 (89.49) 1/
80,304
(41.61)
2-9
8,081
(9.01)
28,991
(15.02)
10-24
1,190
(1.33)
20,239
(10.49)
25-49
58
(0.06)
2,004
(1.04)
50-99
48
(0.05)
3,483
(1.80)
100-499
35
(0.04)
8,619
(4.47)
500-999
5
(0.01)
3,102
(1.61)
1,000+
18
(0.02)
46,255
(23.97)
TOTAL
89,738 2/
193,000
Source: Meridian Research, Inc. and Versar Inc., Financial Responsibility for
Underground Storage Tank Releases; Financial Profile of the Retail
Motor Fuel Marketing Industry Sector, Draft Report, March 1987.
1/ Numbers in parentheses are percentages of all firms or all outlets in
this sector that are represented by this group.
Zj Columns may not total because of rounding; percentages are calculated
for the rounded total.
-------
b-7
Exhibit fa.3
OWNERSHIP AND OPERATION OF RETAIL MOTOR FUEL OUTLETS
Number
Number of Retail
Number
of Retail
Total Number
of
Outlets Owned
Outlets Owned
of Retail
Segment
Firms
and Operated
and
Leased
Outlets Owned
Refiners
27
y,9b4
36,817
4b,781
Jobbers
8,766
25,333
20,713
4b,04b
Convenience Stores 1/
516
14,732
0
14,732
Independent Chains Lj
125
4,010
1,127
5,137
Open Dealers
80,304
80,304
0
80,304
TOTAL
8y,738
134,343
58,b57
193,000
Source: Meridian Research, Inc. and Versar Inc. Financial Responsibility for
Underground Storage Tank Releases: Financial Profile of the Retail Motor
Fuel Marketing Industry Sector, Draft Report, March 1987.
1/ Convenience store owners are defined to exclude jobbers.
Id Independent chains are defined to exclude jobbers and convenience store
owners.
-------
6-8
• Refiners are large, vertically integrated oil companies owning
refineries that produce petroleum products distributed through thou-
sands of their wholesale and retail "branded" outlets. They include
such companies as Amoco, Exxon, Chevron, Mobil, etc. The "semi-
majors" are large, integrated oil companies that may own fewer refin-
eries or supply fewer wholesaler retail outlets than the majors.
• Jobbers are primarily wholesalers of petroleum products who may
also own retail service stations or convenience store outlets.
• Convenience stores are chains of retail stores that for our pur-
poses exclude jobbers and include only those retail outlets that sell
motor fuels.
• Independent chain marketers are owners of chains of retail motor
fuel marketing outlets; they often sell "unbranded" or private brand
petroleum products. For the purpose of this analysis, independent
chain marketers are defined to exclude jobbers and convenience store
owners.
• Open dealers both own and operate their gasoline marketing opera-
tions, usually at single-site locations. In many cases, open dealers
are former lessee dealers who have bought their locations from the
major oil companies or jobbers.
In addition to these owners, lessee dealers (also called independent deal-
ers) operate outlets under lease arrangements, generally with refiners,
jobbers, or independent chains.
Exhibit 6.4 shows the wide range of asset sizes among all of the types of
firms owning retail motor fuel outlets. Although the vast majority of firms
(93.6 percent) have less than $600,001 in assets, those firms with assets of
$600,001 or more own the majority of retail outlets. The median firm has
assets between $200,000 and $400,000, while the median outlet is owned by a
firm with assets in the $600,000 to $1 million range. With the exception of
the largest convenience store chain, all the firms with more than $1 billion
in assets are refiners; all the firms with less than $200,000 in assets are
open dealers. There is a relative absence of firms with assets in the $100
million to $1 billion range; this size range is too small for most refiners
but too large for large jobbers or independent chains.
Exhibit 6.5 illustrates the distribution of net income to total assets
ratios (commonly called the rate of return on assets) for all of the firms
owning retail motor fuel outlets. This is the ratio used in our analysis (see
Section 6.B.2) to characterize firms' financial health or profitability; the
1/ Meridian Research, Inc. and Versar Inc., Financial Responsibility for
Underground Storage Tank Releases: Financial Profile of the Retail Motor Fuel
Marketing Industry Sector, Draft Report, March 1987.
-------
6-9
Exhibit 6.4
DISTRIBUTION OF TOTAL ASSETS AMONG FIRMS
OWNING RETAIL MOTOR FUEL OUTLETS
Number of Firms Number of Outlets Owned
Total Assets in This Group by Firms in This Group
0-$200,000
30,114
(33.56)
30,114
(15.60)
$200,001-$400,000
33,410 (37.23)
36,705
(19.02)
$400,001-$600,000
20,478
(22.82)
21,684
(11.24)
$600,001-51,000,000
3,567
(3.97)
14,268
(7.39)
$1,000,001-$10,000,000
2,063
(2.30)
28,722
(14.88)
$10,000,001-$100,000,000
76
(0.08)
9,572
(4.96)
$100,000,001-$!,000,000,000
4
(0)
2,562
(1.33)
$1,000,000,000+
27
(0.03)
49,371
(25.58)
TOTAL
89,738
2/
193,000
Source: Meridian Research, Inc. and Versar Inc., Financial Responsibility for
Underground Storage Tank Releases: Financial Profile of the Retail
Motor Fuel" Marketing Industry Sector, Draft Report, March 1987.
Numbers in parentheses are percentages of all outlets or all firms in
this sector represented by the firms in this asset group.
Zl Columns may not total because of rounding; percentages are calculated
for the rounded total.
-------
6-10
Exhibit 6.5
DISTRIBUTION OF NET INCOME TO TOTAL ASSETS
RATIOS AMONG FIRMS OWNING RETAIL MOTOR
FUEL OUTLETS
Ratio of Net Income Number of Firms Number of Outlets Owned
to Total Assets in This Group by Firms in This Group
Less than 0
1
(0)
185
(0.10)
0-0.02
30,573
(34.07)
48,801
(25.29)
0.02-0.04
1,540
(1.72)
25,891
(13.42)
0.04-0.06
6,941
(7.73)
45,840
(23.75)
0.06-0.08
30,590
(34.09)
42,054
(21.79)
0.08+
20,094
(22.39)
30,225.
(15.66)
TOTAL
89,738
2/
193,000
Source: Meridian Research, Inc. and Versar Inc., Financial Responsibility for
Underground Storage Tank Releases: Financial Profile of the Retail
Motor Fuel Marketing Industry Sector, Draft Report, March 1987.
1/ Numbers in parentheses are percentages of the total population of
outlets or firms in this net income to total assets group.
Zj Columns may not total because of rounding; percentages are calculated
for the rounded total.
-------
6-11
lower a firm's return on assets, the greater its likelihood of failing or of
being in severe financial distress. The median net income to total assets
ratio for these firms is between 0.06 and 0.08 (i.e., between 6 percent and
8 percent), a fairly typical return on assets for U.S. firms that are not en-
gaged in banking or financial services. Such a value shows that firms in the
retail motor fuel marketing sector are, on average, neither more nor less
profitable than firms engaged in most other lines of business.
Most of the net income to total assets ratio categories include both large
and small firms: although a large convenience store chain is the only firm
represented in the negative (i.e., less than 0) return on assets category, the
second lowest category of return on assets (0 to 0.02) includes both single-
outlet open dealers and the Texaco Corporation. Firms in the highest rate of
return category include both Exxon and single-outlet open dealers.
Small Businesses in the Retail Motor Fuel Marketing Industry
Exhibit 6.fa shows the numbers of small businesses, by industry segment,
that will be affected by this regulation. For the purpose of tnis analysis,
small businesses are defined using the Small Business Administration's defini-
tion for this industry sector: firms with less than $4.6 million in annual
sales. 1/ In 1984, small businesses meeting this definition either owned or
operated more than 75 percent of the 193,000 retail motor fuel outlets in the
United States.
Of this number, open dealers were estimated to own approximately 80,000,
or 42 percent, of all retail motor fuel outlets. Open dealer firms vary
widely in size and age; some open dealers have new outlets and over $500,000
in assets, while others have 30-year-old outlets and only $42,000 in assets.
EPA estimates that the typical (the statistical median) open dealer has
$90,000 in net worth, $210,000 in assets, and $14,000 in annual after-tax
profits. Such a typical open dealer firm is thus a business earning a reason-
able return on investment and having a reasonable expectation of continuing in
business.
In addition to open dealers, small business owners in the retail motor
fuel marketing sector include owners of small chains of retail outlets. It is
common for owners of small chains to own two or three retail outlets and also
to act as wholesale suppliers for several open dealers. (This business pat-
tern is particularly common in rural areas.) It is also common for firms in
this sector to own a chain of several convenience stores, some of which do not
sell gasoline. For such small convenience store chains, gasoline sales are
not generally the primary line of business. EPA estimates that there were
3,700 such chains owning 8,200 retail motor fuel outlets in 1984.
The Agency estimates that 59,000 retail motor fuel outlets are operated by
lessee dealers, whose outlets thus represent 30 percent of all retail motor
fuel outlets. The majority of these lessee-dealer-operated outlets are owned
1/ Federal Register, Vol. 49, No. 28, p. 5032, February 9, 1984.
-------
6-12
Exhibit 6.6
NUMBER OF FIRMS AND OUTLETS POTENTIALLY AFFECTED
BY UST REGULATION, BY SIZE CATEGORY AND SEGMENT
Segment
Number of
Firms
Number of Outlets
Owned or Leased
Small Business Owners
Small Jobbers
Small Convenience Stores
Open Dealers
Small Business Operators
Lessees
Total Small Businesses
(Owners and Operators)
Small Businesses 1/
3,296 6,591
402 1,608
80,304 80,304
43,131 58,657
127,133 147,159
Large Businesses
Large Business Owners
Refiners
27
46,781
Large Jobbers
5,470
39,455
Large Convenience Stores
114
13,124
Independent Chains
125
5,137
Total Large Businesses
5,736
104,497
Source: Meridian Research, Inc. and Versar Inc., Financial Responsibility for
Underground Storage Tank Releases: Financial Profile of the Retail Motor
Fuel Marketing Industry Sector, Draft Report, March 1987.
1/ Defined as firms with less than $4.6 million in annual sales.
-------
6-13
by large, vertically integrated firms that engage in petroleum production,
refining, and marketing, but many are owned by independent marketers who own
chains consisting of between 2 and 100 retail motor fuel outlets. The outlets
operated by lessee dealers range in characteristics from some of the most
modern and efficient outlets in the country to some of the most financially
marginal operations in the retail motor fuel marketing sector. EPA estimates
that the typical (statistical median) single-station lessee dealer is a firm
with 582,000 in assets, £62,000 in net worth, and £6,000 a year in after-tax
profits; the typical lessee dealer is thus a very small firm, but one which
nevertheless has reasonable profits for the size of the business and a reason-
able expectation of continuing in business.
The impacts of UST regulation will differ for different segments of the
small-business portion of this industry. For example, open dealers, who both
own and operate an outlet, will have to meet all of the costs of UST regula-
tions. However, for lessee dealers the situation is more complex because,
although the terms and conditions of leases vary widely, the most common
arrangement makes the lessee dealer responsible for "sounding the alarm,"
i.e., for operating whatever leak detection and inventory control systems have
been agreed to by both the owner and the lessee, but not for maintaining or
replacing tanks or paying for corrective action. 1/ However, EPA is aware
that some lessors are attempting to alter these arrangements to increase the
responsibilities of lessees by, for example, holding the lessee responsible
for releases or requiring the lessee to buy the tanks. In this Regulatory
Impact Analysis, EPA has assumed that the terms of traditional lease arrange-
ments will prevail, and the analysis therefore does not assess the iiipacts of
any changes in lease terms on lessee dealers. However, the analysis does
consider cases in which severe economic impacts could cause trie owner of a
leased outlet to close the outlet and thus force the lessee dealer out of
business.
Although the typical open dealer and lessee dealer are sound businesses,
there are marginal firms in both categories. A marginal firm is defined as
one that is making very low profits or that has an aging outlet and cannot
afford to invest any substantial amount of money in this outlet. Marginal
firms are likely to fail or close their outlets in the event of significant
regulatory expenditures; however, many of these firms would also leave the
industry even in the absence of regulatory costs.
Current Trends
Three sets of long-term annual data provide information on current trends
in the number of retail motor fuel outlets. These three data series, and the
trends they illustrate, are described below.
• U.S. Census Bureau Service Station Population: this time series esti-
mates the number of service stations, defined as facilities receiving
more than bO percent of their revenues from sales of gasoline and
related products. This definition does not include a substantial
number of retail motor fuel outlets that receive sales from other
sources. Census data show that, from 1974 to 15*84, the number of
service stations declined from 196,000 to 132,000, a decline of
1/ Notes of meeting with the Service Station Dealers of America, 1966.
-------
fa-14
33 percent over 11 years, for an average annual rate of decline of
3.7 percent per year. 1/
• American Petroleum Institute Reports: These cover deactivations and
openi ngs of retail motor fuel outlets reported by 30 member
companies; this time series does not, however, provide data on the
total number of facilities owned by API member companies. This
source defines deactivation as the closing of an outlet (including
removal of the outlet's equipment) for which no reopening is
contemplated. For the period 1974 to 1984, this data series shows
that refiners deactivated 48,852 outlets and opened 2,973 outlets,
for a net decline in refiner-owned outlets of 4b,879 outlets. This
net decline in refiner-owned outlets is equal to 72 percent of the
total decline in service stations reported by the dureau of Census
source, which suggests that the decline in outlets may be explained
in large part by refiner closings of marginal lessee operations, i-/
o The Annual Pollars-Per-Day-Survey of Convenience Stores: This survey
provides estimates of the number of convenience stores selling gaso-
line. From 1974 to 1984, the Dol lars-Per-Day Survey reports that the
number of convenience stores selling gasoline has risen from 3,520 to
22,475, an average annual growth rate of 20 percent per year. '*/
These time series enable us to conclude that the total number of retail
motor fuel outlets is declining, that convenience stores are increasing both
in numbers and share of all retail motor fuel outlets, and that the closure of
outlets is as important a phenomenon for refiners as for smaller businesses.
Future Trends
For our purposes, the data series described above do not provide a basis
for projecting future trends. For example, except for refiners, the number of
outlets that actually exit, as against the net decline or growth in outlets,
is not reported. Further, the data are not segmented in a way that can be
applied directly to the categorization and representative firm methodology for
this analysis of the industry.
The annual exit rates for outlets owned by small and large firms were
produced in the following manner:
1/ U.S. Department of Coiimierce, Franchising in the Economy 1983-1985,
January 1985; cited in National Petroleum News, 198b Factbook.
1/ American Petroleum Institute, New Construction Report, May 1984;
cited in National Petroleum News, 1986 Factbook.
2/ John F. Roscoe, 14th Annual Dollars-Per-Day Survey, Marcn 1985; cited
in National Petroleum News, 198b Factbook.
-------
6-15
® Based on the time series data described above, we estimated that, in
1984, 4,200 retail motor fuel outlets owned by firms other than re-
finers exited the industry. (Refiner-owned outlets were excluded
from the economic impact analysis because we assumed that the costs
of these regulations would play only a minor role in the decisions of
refiners to close their outlets.)
t We attributed outlet closures to representative firms on the basis of
the number of firms they represent and their rate of return on
assets, because firms with higher rates of return will probably close
fewer outlets than firms with lower rates of return.
• We devised a system for allocating the estimated 1 y84 closures among
representative firms by assessing the effects on rates of return that
the costs of tank replacement would have on these firms. We assumed
that 4.7 percent of all tanks would have to be replaced, 1/ at a
cost of $20,000 per tank. Although there are many types of costs
that could cause a firm to close its outlets, it is critical for this
analysis to allocate some closures to the fact that the firm was
required to replace tanks at such outlets; unless this methodology is
used, it is difficult to assess the impacts associated with the in-
crease in tank replacement caused by regulatory requirements.
Exhibit 6.7 shows the annual exit rate of outlets owned by small busi-
nesses and large businesses other than refiners in 1984, and the projected
percentages of existing retail motor fuel outlets in each category that will
exit 5, 10, or 15 years in the future. It should be noted that these exit
trends do not attribute any exit to corrective action expenditures. All exit
due to corrective action is attributed to EPA corrective action requirements.
6.B.2. Methodology and Assumptions
Impacts
The economic impact analysis was performed using an affordability
model, £/ which provided estimates for the following economic impacts:
• The current industry baseline ratio of net (i.e., after-tax) income
to total assets (i.e., the rate of return on assets) for each repre-
sentative firm, and the changes that would occur in this ratio as a
result of the inposition of various levels and types of regulatory
costs.
1/ Based on SCI estimates using UST Model results; analysis of the "base
case."
Zj Meridian Research, Inc. and Versar Inc., Documentation for the
Affordability Model, Draft Report, March 1987; see Appendix C: Methodology
Used to Estimate Economic Impacts, for a summary of the structure of the
affordabi1ity model.
-------
6-16
Exhibit 6.7
PROJECTED EXIT OF EXISTING RETAIL MOTOR FUEL OUTLETS
Projected Exit
Projected Exit
Projected Exit
as a Percentage
as a Percentage
as a Percentage
Annual
of Current
of Current
of Current
Exit Rate
Outlets Through
Outlets Through
Outlets Through
Ownership
in 1984
Year 5
Year 10
Year 15
Small Business 1/
4.06%
19%
28%
36%
Large Business
1.00%
5%
10%
14%
Other Than
Refiners
Source: Meridian
Research, Inc.
and Versar Inc.,
Documentation for the Affordability
Model, Draft Report, March 1987.
1/ Defined as firms with less than $4.6 million in sales.
-------
6-17
o The percentage of outlets voluntarily exiting the industry. "Volun-
tary exit" is defined as an owner or operator's choice to close one
or more outlets. It includes both baseline exit (defined as exit
unrelated to the imposition of new regulatory costs) and the exit of
outlets owned by firms that decide that paying expected regulatory
compliance costs would decrease profits to levels lower than accept-
able. Owners of outlets with tanks that have releases and thus re-
quire corrective action do not have the option of voluntarily closing
their outlets. These outlets either fail, if the firms owning them
fail, or remain in business.
• The percentage of failing firms and the percentage of failed outlets
(the percentage of all outlets owned by tne failed firms). It is
assumed that all firms that incur regulatory costs will attempt to
meet them, and that no firm incurring a corrective action may volun-
tarily close any of its outlets. The latter assumption is not de-
scriptive of the situation for large firms; thus, separate analyses
are required for small and large firms.
• The percentage of firms in severe financial distress (defined as
firms that do not file for bankruptcy but that exhibit signs of
severe financial distress, such as restructuring their debt or miss-
ing loan payments).
In the small firm analysis, the overall percentage of outlets surviving at
a given point in time (i.e., at the end of Year 5, Year 10, or Year 15 after
the promulgation of tank technical requirements) was calculated based on both
the percentage of outlets voluntarily exiting and tne percentage of outlets
closing because the firm had failed.
Affordability Model
The affordability model contains financial (e.g., net income, total
assets, etc.) and operating (e.g., number of outlets owned, number of outlets
operated, etc.) data for 69 firms taken to represent the approximately 90,00u
firms that own and the 43,000 firms that operate and lease 193,000 retail
motor fuel outlets. 2/
Exhibit 6.8 shows how the affordability model can be used to perforin the
economic impact analysis of UST regulatory options and scenarios. First, the
initial net income and total assets of owners of representative firms are
reduced to reflect the impact of various levels and types of regulatory
costs. Regulatory costs are divided into capital and other costs; capital
costs are treated differently for tax reasons (e.g., the investment tax credit
is applied). The model has a scalar to adjust net income to take account of
revenue increases. The model also incorporates the probability that USTs
owned by firms will experience release incidents requiring corrective action
and thus that firms will incur the costs associated with sucn incidents (i.e.,
for assessment, cleanup, tank upgrading, tank closure, tank replacement) and
1/ Meridian Research, Inc. and Versar Inc., Financial Responsibility for
Underground Storage Tank Releases: Financial Profile of the Retail Motor Fuel
Marketing Industry Sector, Draft Report, March 1987.
-------
USE OF THE AFFORDABILITY MODEL TC PERFORM ECONOMIC
ANALYSIS OF THE IMPACTS OF ' REGULATORY OPTIONS
SOURCE: MERIDIAN RESEARCH. INC ANO VERSAR INC.
AFFOROABILITY MODEL DOCUMENTATION. 19B7.
-------
6-19
accounts for payments from insurers to cover these costs. The model assumes
that those firms that are able to pass financial tests (tests that require a
firm to have a certain ratio of net income to total assets and loan size to
total assets) can borrow to pay the costs of corrective action, tank replace-
ment, and upgrading. The magnitude of the UST financial responsibility that
will be left unfunded is determined by the percentage of failing firms, the
number of outlets owned by these firms, and the amount of corrective action
costs these firms are predicted to incur.
The percentage of firms failing was estimated based on the results of a
study that used enpirical data that established a relationship between ratios
of net income to total assets for firms in various asset classes and the prob-
ability that firms in these classes would fail. 1/ These data show that
firms whose net income to total assets ratio is below -.30 almost always
fail. The model assigns appropriate failure probabilities to firms with
higher ratios of net income to total assets. Severe financial distress rates
are based on the same enpirical data. These data show that firms whose net
income to total assets ratios fall in the range of -.04 to -.30 are likely to
experience financial distress even if they do not fail.
Assumptions
The assumptions used in this analysis are classified into two groups:
o Key assumptions
o Additional assumptions.
Key assumptions are those most critical in defining the general limits of
the analysis and thus of the study's results. The six key assumptions are
that:
• No steel tanks are cathodically protected. Eighty-nine percent of
tanks are bare steel and 11 percent are fiberglass.
« No insurance or other form of private risk pooling will be available
for UST releases.
• Leaks will be discovered with the probabilities developed by SCI
using the UST Model. (Option II, for example, predicts that ZS per-
cent of tanks will have high-cost releases within 5 years and that
33 percent of tanks will have low-cost releases within b years).
• The National Leaking UST Trust Fund will not alleviate the economic
impact of corrective action costs.
• No State UST Funds will be used to pay corrective action costs.
1/ Meridian Research, Inc. and Versar Inc., Documentation for the
Affordability Model, Draft Report, March 1987.
-------
fa-20
• With the beginning of each new year, each surviving firm's annual net
income and net worth will revert to their original values.
Exhibit 6.9 describes the effect that each key assumption has on the re-
sults of the analysis and evaluates the strength of each effect.
The additional assumptions underlying the economic inpact analysis are
divided into three categories:
• Economic and financial
• Technical
• Those related to choice of compliance method.
The additional assumptions are presented, by category, in Exhibits 6.1U
through 6.13. Unlike the key assumptions, these additional assumptions do not
define the limits of the analysis; instead, they present data, describe data
sources, and illustrate how parts of the analysis work.
6.B.3. Impact Analyses
This section describes the results of the economic impact analyses per-
formed for the regulatory options being considered. First, it analyzes the
impacts of the costs being considered on the typical open dealer firm.
Second, the impacts of regulatory costs on small firms are assessed, assuming
both that there will not be a revenue increase and that there will be a reve-
nue increase. Finally, large-firm impacts are evaluated assuming both that
there will and will not be revenue increases. Small-firm impacts are ex-
pressed in terms of the percentage of outlets owned by small firms that will
survive, while large-firm inpacts are presented in terms of the impacts of
regulatory expenditures on the ratios of net income to total assets per out-
let. A separate inpact analysis is performed for lessee dealer firms. Im-
pacts on these firms are dependent on the economic impacts on large firms.
Analysis of Inpacts on the Median Open Dealer Firm
This economic impact assessment analyzes the costs likely to be imposed on
UST-owning firms under each of the regulatory options under consideration.
The model uses 69 representative firms to represent the 133,000 firms in the
industry. (See Appendix D for further discussion of represented firms). The
most important results of the analysis can be seen best by examining the im-
pacts of selected regulatory scenarios on one representative firm in a single
year. The one firm chosen for this purpose is the median open dealer, defined
as a firm that owns and operates one retail outlet and that has the median
level of net income, total assets, and net worth for open dealer firms. This
firm is assumed to represent 10,000 of the "80,000 firms in the open dealer
segment of the industry, and can be considered typical of the firms in the
open dealer segment because it has median values for financial characteristics.
-------
6-21
Exhibit 6.9
KEY ASSUMPTIONS AND THEIR EFFECTS ON THE RESULTS OF THE ANALYSIS
Assumption
No steel tanks are cathodi'
cally protected. Eighty-
nine percent of tanks are
bare steel and 11 percent
are fiberglass.
No insurance or other
form of private risk
pooling will be available
for UST releases.
Releases will occur with
the probabilities developed
by SCI using the UST Model
(Option II, for example,
predicts that 26 percent of
tanks will have high cost
releases within 5 years and
that 33 percent tanks will
have low cost releases
within 5 years.)
The National Leaking UST
Trust Fund will not
alleviate the economic
impact of corrective action
costs.
State UST funds will not be
used to pay corrective
action costs.
Effect of Assumption on Results of the Analysis
Slightly overestimates the impact of corrective
action and other regulatory compliance costs.
In fact, 13 percent of all steel tanks are
cathodically protected.
Overestimates the impact of corrective action
costs because insurance or other forms of
private risk pooling will cover some corrective
action costs; however, this overestimate is not
large because (1) tank upgrading and replacement
costs will not be covered, and (2) many small
fimis cannot obtain insurance.
Section 4 discusses the assumptions used as a
basis for determining the probabilities of
corrective action. If the estimated
probabilities are, for example, too low, the
numbers of corrective action incidents in this
analysis are understated, and thus their
economic impacts are also understated.
Slightly overestimates the impact of corrective
action costs on firms. With the passage of SARA,
a national UST Trust Fund in the amount of $500
million has been authorized, but the Fund cannot,
in most cases, be used to cover the corrective
action costs of solvent firms that lack
financial assurance at the required levels, and
the Fund is not large enough to significantly
change the economic impacts of corrective action
costs.
The degree to which the impacts are overesti-
mated depends on: how actual funds are
designed, their adequacy in covering the
corrective actions that occur, and the number of
States implementing them. Impacts would be
seriously overestimated if many States implement
State funds that provide adequate coverage for
releases, and include credit assistance for tank
closure and replacement.
-------
6-22
Assumption
With the beginning of
each new year, each sur-
viving firm's annual net
income and net worth are
assumed to revert to their
original values.
Exhibit 6.9 (Continued)
Effect of Assumption on Results of the Analysis
Underestimates economic impacts. The percentage
of surviving outlets is slightly overestimated
for small firms; impacts on the profitability of
large firms are seriously underestimated.
-------
fa-i;3
Exhibit b.10
ADDITIONAL ASSUMPTIONS (ECONOMIC AND FINANCIAL)
o Loan availabi1ity
- In the impact analyses, it is assumed that loans are
available to firms that meet certain financial criteria. 1/
- In the median open dealer analysis, it is assumed that the
median open dealer is not able to obtain a loan.
• A firm's net income and the value of its total assets determine
its financial condition.
The underlying assumptions are as follows:
- A firm's net income for a year is its after-tax profits.
- The ratio of net income to total assets determines the
financial condition of a firm. If a firm's ratio is less
than -0.3, it is assumed to fail. If its ratio is -0.3 or
greater, a probability of failure is assigned to the firm.
The higher the ratio, the lower the probability of failure.
• Costs are financed either internally, first through net income
and then through asset sales, or externally through a loan. A
firm may sell assets only to cover those costs that cannot be
met by its net income.
• Value of a firm's total assets
- The value of a firm's total assets will decrease by the value
of the assets sold when assets are sold to cover costs.
- A firm's value of total assets will remain constant unless
assets are sold; the value of firms' total assets are assumed
to revert to their original value at the beginning of each
year.
Source: Meridian Research, Inc. and Versar Inc., Documentation for the
Affordabi1ity Model, Draft Report, March 1987.
1/ Tne financial criteria are based on the notes of a meeting With
First Virginia Bank, 1986.
-------
6-24
Exhibit 6.11
ADDITIONAL ASSUMPTIONS (TECHNICAL)
• All firms have 3.5 tanks at each outlet.
• Releases are detected only as a result of a tank test or another
method of leak detection/monitoring.
• The corrective action, 1/ tank upgrade, 1/ and leak
detection costs 1/ used in the analysis are based on the UST
Model.
Average site-by-site corrective action costs, including leak
verification costs of $4,000 but excluding costs for tank
replacement, are:
• $23,100 for a non-plume release
• $37,200 for corrective action to clean up the floating
plume only for plumes of less than 25 square meters, and
$63,200 for corrective action to clean up both the
floating and dispersed plumes for plumes of less than 25
square meters.
• $127,700 for a plume release with an area of greater than
25 square meters.
— Costs for the upgrading of one tank are $3,050
— Costs for the mandatory replacement of one tank 1/ are:
• Closure: $12,500
• Tank replacement: either $20,000 for a cathodically
protected tank, or $23,000 for a protected tank with a
liner.
• Testing of the tank one year later: cost as listed below.
• Leak detection/monitoring costs are based on those provided in
the Cost Effectiveness Analysis and are as follows:
— Tank test: $500/tank.
1/ See: Section 4.D.3.
lj See: Section 4.C.2.
1/ See: Section 4.B.2.b.
£/ The costs for mandatory tank replacement are based on
engineering estimates.
-------
6-25
Exhibit 6.11 (Continued)
Monthly vapor well monitoring: initial capital cost of $ 1,bUO
per facility, plus $900 in annual costs thereafter.
Inventory control costs are assumed to be $0 because
inventory control is current industry practice.
• The probabilities that corrective actions will occur and their
severity were taken from the Cost Effectiveness Analysis.
Specific assumptions are as follows:
Yearly probabilities of plume releases of less than 25 square
meters, plume releases of greater than 25 square meters, and
non-plume releases were computed for each of the options.
• The ^probabilities provided by SCI using the UST
model 1/ for eacn combination of option and release type
were arranged into 3 groups: (1) the probabilities for
years 1-5; (2) the probabilities for years fa-10; and (3)
the probabilities for years 11-15.
• Each group's 5 annual probabilities were totaled. An
average annual probability was computed by dividing this
total by 5. This average was used as the probability of
corrective action for each of the years defined by the
group.
Probabilities for years 1-5, 6-10, and 11-15 were determined
for releases with corrective action costs of over $50,000,
(called high cost release events) and under $50,000 (called,
low cost release events), using the following assumptions; L/
• All plume releases of greater than 25 square meters cost
more than $50,000.
• Sixty percent of plume releases of less than 25 square
meters involve cleanup of both the floating and dispersed
plume, and have costs of greater than $50,000.
• Forty percent of plume releases of less than 25 square
meters involve cleanup of tne floating plume only, and
have costs of less tnan $50,000.
• All non-plume releases have costs of less than $50,000.
1/ Based on SCI estimates, using UST Model.
i/ Meridian Research, Inc. and Versar Inc., based on SCI estimates
using the UST Model.
-------
6-26
Exhibit 6.11 (Continued)
• Corrective action costs for high-cost and low-cost release
events were determined for each option by the weighted average
of costs for plume release of greater than Zb square meters,
plume release of less than Zb square meters, and non-plume
releases. The resulting weighted costs for each option are
presented in Exhibit 6.12.
• Tanks are assumed to be replaced based on the UST Model
estimates for each option. These tank replacement rates are
presented in Exhibit 6.12.
• Mandatory tank retirement or upgrading is assumed to take
place at a rate of 6.4 percent of tanks per year for Options
II and V, and at a rate of 3 percent of tanks per year for
Option IV.
-------
Exhi' 6.12
COSTS AND PROBABILITIES OF CORRECTIVE A EVENTS AND THE PROBABILITY THAT A TANK
IS REPLACED AS A RESULT u * CORRECTIVE ACTION EVENT
High-Cost Release Event
Low-Cost Release
Event
Probabi1lty That
a Tank is Replaced
5-Year
Interval
Corrective Action Probability Corrective Action
Cost of Event Cost
Probabi1ity
of Event
As a Result of
a Corrective
Action Event
Option 1
Each of
Each of
Each of
Years
Years
Years
I-5
6-10
II-15
$108,149
107,356
108,375
0.0590
0.0308
0.0286
Option II
$30,650
28,512
28,22b
0.0600
0.04b2
0.0424
O.Obb
0.023
0.030
Each of
Each of
Each of
Years
Years
Years
I-5
6-10
II-15
$102,431
101,836
104,000
0.0520
0.0176
0.0100
Option III
$30,650
28,104
28,133
0.0660
0.0344
0.0180
0.0b4
O.OId
0.011
Each of
Each of
Each of
Years
Years
Years
I-5
6-10
II-15
$108,767
108,533
108,872
0.0582
0.0300
0.0268
Option IV
$30,b24
28,493
28,318
0.0578
0.0420
0.0382
O.Obb
0.024
0.030
Each of
Each of
Each of
Years
Years
Years
I-5
6-10
II-15
$103,583
104,299
103,676
0.0522
0.0182
0.0168
Option V
$30,626
28,7b3
28,7b3
0.0b38
0.0288
0.0272
0.0b3
0.01b
0.020
Each of
Each of
Each of
Years
Years
Years
I-5
6-10
II-15
$98,845
98,98y
63,200
0.0496
0.0152
0.0004
$30,802
28,228
28,763
0.0b94
0.0318
0.0016
O.Obb
0.01b
0.002
Source:
Meridian Research, Inc. and
Versar Inc., based on SCI
estimates using
the UST Model.
-------
6-28
Exhibit b.13
ADDITIONAL ASSUMPTIONS (RELATED TO CHOICE OF
METHOD OF COMPLIANCE)
• If an option allows for a choice between replacing and upgrading
a tank, it is assumed that firms will choose to upgrade.
• If an option allows for a choice in the method of leak detection/
monitoring used, the choices assumed in Section 4.
-------
b-29
The analysis shows that two events largely determine the magnitude of the
economic impacts associated with Options I-V: corrective action and replace-
ment of a tank. Exhibit 6.14 compares the economic impact on the meaian open
dealer of two different types of corrective action events, a plume release and
a non-plume release, and of replacing a tank. The costs of a plurne release
are the average costs of responding to a release having a plume of more than
25 square meters. (Although the scenario depicted in Exhibit fa.14 involves a
corrective action that requires replacing a tank, the economic impact analysis
assumes that only 60 percent of corrective actions will require tank
replacement.) Exhibit fa.14 shows that:
o Having to perform corrective action for a non-plume release forces
the median open dealer firm into a "severe financial distress" condi-
tion. Firms in this condition have an increased probability of fail-
ure, have difficulty in meeting existing loan or credit obligations,
and consider leaving the business.
o Having to perform a corrective action for a plume release causes the
median open dealer firm to fail even before the leakiny tank is
closed and replaced.
• Having to replace one tank also forces the median firm into a condi-
tion of severe financial distress. (Having to replace 3 tanks at
once would bankrupt the firm.)
The costs of these events are significant compared with the net worth and
profits of the median open dealer firm. The corrective action expenditures
required to clean up a plume release exceed the net worth of such firms arid
are equal to 10 years' worth of their profits. Replacement of one tank has a
cost equal to more than one-third of the net worth of the median open dealer,
or more than 2 years of such a dealer's profits. (The impact of mandatory
tank upgrading as permitted under Option II has a relatively small impact on
the financial condition of the median open dealer. The £3,050 cost 0f cathod-
ically protecting a tank would leave such a firm in good financial condition,
with a return on assets of 5.4 percent in the year in which this cost is in-
curred. )
Costs of the magnitude of those for tank closure and replacement present
two problems for the median open dealer. First, because of limited access to
credit, the median open dealer may be unable to raise the funds to pay for
such expenditures. Second, in a business that typically has a planning hori-
zon of less than 5 years, an expense large enough to absorb 2-10 years of
profits could easily persuade the owner to leave the industry.
Although it may initially seem odd that a firm confronted with corrective
action or tank replacement costs will not be able to raise the necessary capi-
tal, especially since less capital is required for these expenses than was
required to start the business in the first place, there are important differ-
ences between regulatory expenditures (e.g., replacing all tanks in one year)
and investing in a business. First, a new business has substantial resaleable
assets that can be used as collateral for a loan, while a corrective action
cleanup or replacement of a tank system cannot be used to provide security for
a loan. Second, although the business can be expected to provide a regular
-------
Ncn-olume ^e lease
Exhib. -.14
THREE SOURCES OF SEVERE FINANCIAL HARDSHIP FOR THE MEDIAN OPEN
DEALER (1)
Impacts on firm. i4)
Net Income
Net Income/Assets
Condition
2 Sara
Steel
ranio
Teat 3
LeaK
varific a 11on
34. 000
Non-a1ume
Re\ease
Reolaco : TanK
to Mew TanK
31. 200 y
'¦^Lean in *.
T an*
Suaoecteo
Laax
Con-
firmad
Clean-up (21
Coats
319 100
Standards '3)
320. 000
314. 000
312. 720
39. 320
-38. 140
-328. 400
5 57*
6.0BX
4 . 44*
-4 03*
-13 64*
jOOO
Gooo
Fa ir
Severe Financial
Severe Financial
Source #2: A Plume Release
3 Bare
Steal
TanKa
X Teat 3 X
LaaK
Variflcatlon
34. 000
Plume Relaeaa
Clean-uo
Coats (SJ
3123. 700
Replace l TanK
to New Tank
$1. 300 >
in l
Tank
Suaoactao
LaaK
Con-
fIrmad
Standards 13)
320. 000
I mo acts on Firm: (4)
Net Income
314. 000
312, 720
39. 320
-3112. 740
Fai lure
Net Incoma/Assets
6.67*
6.06*
4 44*
-113.91*
-inanciai Condition
3ood
Good
Fair
¦ a 11 *j r e
I
OJ
O
Source: Meridian Research. Inc. ana Verssr Inc
-iBing tna afforaeoi11 ty moael. 1987.
(1) median ooen dealer is tne owner of a single outlet Mho possesses 314.000 in net income. 3210.000 in total assets, and 390.000 in net wor-n These
iteaian financial cnarscteriscici -ere determined from survey oata cdmoiled ay tne Service Station Dealers of America (SSOaj ana from tne FInstat aata
oaae provided By tne Small Business Adminiatr at ion (SBAJ .
21 ~-\e ^on-oluree release cleen-«..o coat is a weignteo average of tne corrective action :osts soecified Dy EPa :a:n corrective action cost is «eisnted ay
tne crooaaility of its occurrence. as aoecifiea cy £PA. The soecif'.e calculation is as fallals: (0.2 x S3. 700) «. ,o 3 x 123. 200J - $19. 100
'3) -rotectea »ingla--e1 lea tsnn system «i»n irnoroveo line Usk detectors, caoita
maenaivt tvce of reolacement tan* system and is allowed unoar Motions I and
:oat of 37. OOO installation cost of 313. 000 This Is tne least
ij| -aauming that tnere are no once increases: tnat net income calculations include tne effect of taxes:
ma net *ortn return to tneir original levels *itn tne aeginning of eecn new year
and tnat tne firm's net income, total assets.
SI The olurne release clean-»-jo cost :s a weighted average of tne corrective action costs specified oy £Pa Eacn corrective action cost is -eigniea ay tne
crooaDillty or its occurrence, as soecifiea oy £Pa . Tne specific calculation is as follows 0 5 * 356. 200J *¦ .0.1 x 3131. 2Q0J • «'Q 1 x 3206 200) *
0 2 x 3261.200) - 3123. TOO.
-------
Exhibit 6.14 (continued)
THREE SOURCES OF SEVERE FINANCIAL HARDSHIP FOR THE MEDIAN OPEN
DEALER (1)
Source #3: The Replacement of One Tank
Impacts on Firm: (3)
Net Income $14. OOO 33. 370 -$16. 300
Net Income/Assets 6.67% 1.60X -0.42X
Financial Conoltlon Good Fair Severe Financial
Olatreaa
Source: Merlaian Researcn. Inc. ana Versar Inc.. using the affordap 1111y model. 1907.
(1) The Median dealer 13 the owner af a single cutlet who possesses $14. 000 In net Income, 3210.000 in total
assets, ana 590.000 in net wortn. These median financial cnaracter i at l cs were aeterminea from survey data cr>
comoileo oy the Service Station Dealers of America (5S0A) and from the rINSTAT data Pase provided Dy tha Small 1
Business Administration (SBAJ . ^
(2) Protected 3 lng 1 i-«s11ed tank system with Improved line leak detectors, capital cost of 37.000; installation
cost af 313.000. This is the least expensive type of replacement tank system, and is allowed under Options I
and II.
(3) Assuming that there are no price increases: tnat net income calculations include the effect of taxes; ana that
the firm's .net income, total assets, and net worth return to their original levels with the Peglnnlng of each
new year .
-------
6-32
income, corrective action and tank replacement expenditures provide no new
income. A roughly parallel comparison would be the difference between buying
a new car, which represents a valuable asset with an expected lifetime of more
than 10 years, and paying to rebuild the engine of an 3-year-old car, which
might easily cost more than the potential sales value of the car. Addition-
ally, although low-cost financing might be available for a new car, because
the car itself can be used as collateral for the loan, such financing would
certainly not be available to pay for the old car's engine job.
Analysis of the Inpacts of Regulatory Options on Outlets Owned by Small Firms,
Assuming No Revenue Increase
This economic iirpact analysis assesses the percentage of outlets owned by
small firms that will exit the retail motor fuel marketing sector and identi-
fies the factors responsible for these exits. The following section presents
the cumulative percentage of smal1-firm-owned outlets surviving at the close
of years 5, 10, and 15 after iiiplementation of any of the UST regulatory
options under consideration.
Some retail outlets in the motor fuel marketing sector will exit the
industry even without the imposition of regulatory costs on their owners;
exits occurring in the absence of regulatory costs are termed "natural"
exits. Natural exits were estimated by dividing the number of exits predicted
by past industry exit trends, which were assumed to continue for the next
15 years, by the number of retail outlets owned by small firms (defined as
firms with less than $4.6 million in assets) now in the industry.
To determine the percentage of total exits attributable to a given regula-
tory option, the percentage of natural exits was subtracted from the percent-
age of exits predicted to occur under that option. The difference between
these percentages was then attributed to the inpact of regulatory costs.
Exits caused by regulatory costs are divided into two categories:
• Exits attributed to tank replacement--this category of exit includes
all exits (above the percentage of natural exits) attributable to the
impact of the costs of leak detection, mandatory tank upgrading,
replacement of tanks that cannot be repaired after a release, or
mandatory tank closure and replacement.
• Exits attributed to corrective action—this category of exit includes
those where the firm owning the outlet could meet all UST regulatory
costs other than those of corrective action but would fail if forced
to incur corrective action costs.
When a firm fails as a result of a large release it cannot afford to clean
up, the exit of its affected outlets could be attributed either to the costs
of tank replacement or the costs of corrective action, because either cost by
itself may be sufficient to cause the firm to fail. In this analysis, such
exits are attributed to the costs of tank replacement.
Exhibits 6.15 and 6.16 show the percentage of sma11-firm-owned outlets
surviving and exiting by the end of year 5 for each regulatory option. These
-------
IMPACT OF REGULATORY OPTIONS
Percentage of Outlets Owned by Small
Firms Surviving and Exiting by Year 5 *
Survival and Exit Rate Percentages
100% uvyyyj uvwj—r
75%
50% -
25% -
'//A Exit-Natural
Exit-Tank Replace
Exit-Correc. Action
Surviving
Options
cr>
i
oo
CO
This analysis assumes no revenue
increases.
-------
Ex 6.1b
IMPACT OF REGULATORY OPTIC ASSUMING NO REVENUE INCREASE:
CUMULATIVE PERCENTAGE OF OUTLETS OWNED BY SMALL FIRMS 1/
SURVIVING AND EXITING THROUGH YEAR 5
Option
1 II 111 IV V
2/
Natural Exit Through Year 5 -
3/
Exit - Tank Replacement or Upgrade Through Year S —
4/
Exit - Corrective Action Through Year 5 -
TOTAL EXIT Through Year 5
Surviving Through Year b
19
19
19
19
19
2
2
2
lb
41
52
bU
b2
39
22
73
71
73
73
lb
27
2y
27
27
lb
Source: Meridian Research, Inc. and Versar Inc., using affordabi1ity model results.
1/ Small firms are defined as firms with annual sales of less than $4.b million.
1/ Natural exit is based on an extrapolation of past exit trends for outlets owned by small firms
and the replacement of failed tanks that these firms would have replaced even in the absence of further UST
regulation.
1/ Exit caused by leak detection, replacement of leaking tanks, and mandatory tank retirement or
upgrading costs.
i/ Exit caused by corrective action costs.
-------
6-35
exhibits show that natural exit would account for the exit of ly percent of
such outlets by the end of year 5.
At the close of year 5, all of the Options are very similar in their
effects on the percentage of outlets surviving: between lb and 2y percent of
smal1-firm-owned outlets are predicted to be in business. Option II has the
highest percentage of outlets surviving (29 percent), and Option V has the
lowest (18 percent). Tne reason that Option V has the lowest percentage of
surviving outlets is that, under this option, firms must incur the high costs
of replacing their tanks with new tanks having secondary containment.
In Options I, II, and III, the impacts of corrective action costs cause
almost all of the exits for outlets owned by small firms (between 50 and
52 percent); the impacts of replacing or upgrading tanks are very minor by
comparison, causing only 2 percent of all exits under these three options.
Exits attributed to corrective action are high because almost all small firms
that experience a high-cost release are forced into bankruptcy, and all im-
pacts of corrective action are attributed to corrective action requirements.
Under Option IV, the impact of tank replacement or upgrading causes
15 percent of all exits above natural exits. This is equal to 38 percent of
the exits caused by corrective action (39 percent of all exits above natural
exits). Under Option V, 41 percent of all exits above natural exit are caused
by tank replacement or upgrading and 22 percent of exits are caused by the
impact of corrective action expenditures. Option V, which requires mandatory
tank retirement and the most stringent leak detection measures, thus prevents
many more exits due to corrective action than the other options but still
results in a large percentage of exits because of the high costs of replacing
tanks.
Exhibits 6.17 and 6.18 show the percentage of smal1-firm-owned outlets
that are predicted to survive or exit at the end of year 10 for each regula-
tory option; Exhibits 6.19 and 6.20 show the same data for the end of year
15. As would be expected, the percentage of outlets surviving is lower in
subsequent years. Under all options, natural exits account for an increasing
percentage of total exits for outlets owned by small firms at the close of
years 10 and 15. During these later years, Options I, II, and III remain
similar in their impacts, i.e., the impacts of corrective action costs con-
tinue to be the primary source of exit above natural exit. It is interesting
that by years 6-10 under Option II (see Exhibit 6.18) tank replacement or
upgrading does not cause any more exits than in the baseline and that by years
11-15 (see Exhibit 6.20), exit caused by tank replacement or upgrading under
Option II would actually be lower than natural exit.
Under Option IV, the proportion of exits caused by tank replacement or
upgrade compared to corrective action shifts in later years. In years fa-10,
tank replacement or upgrade causes 35 percent of all exits above natural exit
and by years 11-15, 42 percent. Under Option V, the iiiportance of tank
replacement or upgrade as a source of exit increases in later years to the
point where it accounts for yb percent of exit above natural exit in years
11-15.
-------
IMPACT OF REGULATORY OPTIONS
Percentage of Outlets Owned by Small
Firms Surviving and Exiting by Year 10*
100%
Survival and Exit Rate Percentages
75%
50%
25%
IV
Options
HI
IP IP IP IP
W/,
— • • •
1
(I ji
§i§
1 1
1MB •¦••• ¦
:
v
222 Exit-Natural
J Exit-Tank Replace
Exit-Correc. Action
Surviving
i
OJ
This analysis assumes no revenue
increases.
-------
Er t 6.18
IMPACT OF REGULATORY OPTI ASSUMING NO REVENUE INCREASE:
CUMULATIVE PERCENTAGE OF uUTLETS OWNED BY SMALL FIRMS 1/
SURVIVING AND EXITING THROUGH YEAR 10
Option
I II III IV V
Natural Exit Through Year 10 -
3/
Exit - Tank Replacement or Upgrade Through Year 10 -
4/
Exit - Corrective Action Through Year 10 -
TOTAL EXIT Through Year 1
Surviving Through Year 10 14 21 14 18
28
28
28
28
28
2
0
2
19
b3
56
51
bb
3b
8
86
79
86
82
89
cn
i
CO
-vj
Source: Meridian Research, Inc. and Versar Inc., using affordability model results.
1/ Small firms are defined as firms with annual sales of less than $4.6 million.
1/ Natural exit is based on an extrapolation of past exit trends for outlets owned by small firms and
the replacement of failed tanks that these firms would have replaced even in the absence of further UST
regulation.
1/ Exit caused by leak detection, replacement of leaking tanks, and mandatory tank retirement or
upgrading costs.
1/ Exit caused by corrective action costs.
-------
IMPACT OF REGULATORY OPTIONS
Percentage of Outlets Owned by Small
Firms Surviving and Exiting by Year 15*
Survival and Exit Rate Percentages
100% [777773 777773 7
75% -
50% -
25% -
II III IV
Options
Y/A Exit-Natural
3 Exit-Tank Replace
Exit-Correc. Action
Surviving
cr>
i
OJ
00
* This analysis assumes no revenue
Increases.
-------
Exh b.20
IMPACT OF REGULATORY OPTIONS, ASSUMING NO REVENUE INCREASE:
CUMULATIVE PERCENTAGE OF OUTLETS OWNED BY SMALL FIRMS 1/
SURVIVING AND EXITING THROUGH YEAR lb
Option
I II III IV V
Natural Exit Through Year lb -
3/
Exit - Tank Replacement or Upgrade Through Year 15 -
4/
Exit - Corrective Action Through Year 15 -
TOTAL EXIT Through Year 15
Surviving Through Year 15 8 15 8 12
36
36
3b
3b
3b
1
-2
1
22
54
55
51
55
30
3
92
85
92
88
93
Source: Meridian Research, Inc. and Versar Inc., using affordabi1ity model results.
1/ Small firms are defined as firms with annual sales of less than $4.6 million.
hj Natural exit is based on an extrapolation of past exit trends for outlets owned by small firms and
the replacement of failed tanks that these firms would have replaced even in the absence of further UST
regulation.
2/ Exit caused by leak detection, replacement, of leaking tanks, and mandatory tank retirement or
upgrading costs.
1/ Exit caused by corrective action costs.
-------
fa-40
These results show that in terms of the percentage of surviving outlets
owned by small firms, Option II is at least as good as the other options con-
sidered. In terms of the economic inpacts of tank replacement or upgrade
alone, Option II is superior to the other options considered.
Analysis of the Impact of Regulatory Options on the Profitability of Large
Firms, Assuming No Revenue Increase
The inpacts of the regulatory expenditures potentially iiiposed by Op-
tions I through V on the ratios of net income to total assets of large firms
(excluding large oil companies) were analyzed, assuming no revenue increases.
This analysis compared the initial (i.e., 1983-84) ratio of net income to
total assets with the value of this ratio after these firms had incurred one
year of regulatory costs. Because the analysis did not consider the impact of
any decline in assets that might have been caused by the need of these firms
to finance previous UST regulatory expenditures or the costs of servicing any
debt incurred to finance such previous expenditures, the ratios of net income
to total assets reported here are accurate only for the first year iri which
regulatory costs are assumed to be incurred; in subsequent years, the cumula-
tive impact of regulatory costs on these ratios is seriously underestimated in
this analysis. As a consequence, the large-firm net income to total asset
ratios reported for years 6-10 and years 11-15 are useful primarily for making
gross comparisons of the levels and types of costs that would be required in
these two subsequent b-year intervals with those imposed in the first s-year
period.
The impacts on large firms of the regulatory costs associated with the
options under consideration are reported in two stages. First, the impacts of
tank replacement costs (i.e., leak detection, upgrading of tanks, mandatory
tank retirement) are reported. Next, the impacts of all regulatory costs,
including corrective action clean-up costs, are reported.
Exhibit 6.21 shows the initial average value of the ratio of net income to
total assets per outlet for large firms and the average value of this ratio
after regulatory costs have been incurred during years 1-b, 6-10, and 11-lb
under the five regulatory options. The initial average ratio of net income to
total assets per large-firm-owned outlet in 1983-84 is .042 (or a 4.2 percent
return on assets), well below the 7 percent long-term average for this ratio
for most businesses. Though 1986 return on assets data are not yet available
for large firms in this sector, profits for these large firms (other than
refining companies) are expected to be higher in 1986 than in recent years,
although they may return to their former levels if oil prices stabilize or
increase.
As Exhibit 6.21 shows, the impact of the total of average annual
regulatory costs incurred during years 1 through 5 causes the net income per
outlet for large firms to become negative. This iiiplies that, in the absence
of revenue increases, large firms can expect b years of losses under any of
the options considered. By years 6-10, regulatory costs would cause a smaller
impact on profits; average net income per outlet continues to be negative
under all options except Option II, where it is zero. By years 11-lb, the
regulatory costs associated with options II, IV, and V would allow average
-------
6-41
Exhibit 6.21
IMPACTS OF REGULATORY EXPENDITURES ON THE RATIOS OF NET
INCOME TO TOTAL ASSETS PER OUTLET FOR LARGE FIRMS 1/
(EXCLUDING LARGE OIL COMPANIES), ASSUMING NO
REVENUE INCREASE
Option
II
III
IV
Initial Average Per Outlet Ratio of
Net Income to Total Assets (for 1983-1984)
.042
.042
.042
.042
.042
Revised Ratio of Net Income to Total Assets
Caused by Tank Upgrade or Replacement
Costsb Incurred Annually During Years 1-5
Revised Ratio of Net Income to Total Assets
Caused by All Regulatory Costs0
Incurred Annually During Years 1-5
.040
.038
-.099 -.086
,040
100
. 03b
¦.101
. 009
-. 02/
'ised Ratio of Net Income to Total Assets
Jaused by Tank Upgrade or Replacement
Costs Incurred Annually During Years 6-10 Id .048 .048
Revised Ratio of Net Income to Total Assets
Caused by All Regulatory Costs Incurred
Annually During Years 6-10C -.026 .000
.048
.042
-.02b -.010
.021
•.027
Revised Ratio of Net Income to Total Assets
Caused by Tank Upgrade or Replacement
Costs Incurred Annually During Years 11-15 iL/ .047 .Obi .049 .054 .051
Revised Ratio of Net Income to Total Assets
Caused by All Regulatory Costs Incurred
Annually During Years 11-15 b -.023 .020 -.020 .00b .038
Source: Meridian Research, Inc. and Versar Inc., using affordability model results.
1/ Large f irms are defined as firms with annual sales greater than $4.6 million that
do not engage in petroleum production or refining.
Zj Tank replacement or upgrade costs include costs of leak detection/monitoring,
mandatory tank upgrading, replacement of tanks that cannot be repaired after a release,
mandatory tank closure and replacement.
1/ Total annual regulatory costs include tank replacement and corrective action
costs.
-------
fa-42
profits per large-firm-owned outlet to be positive. Profits per large-firm-
owned outlet would continue to be negative under Options I and III.
The relative impacts of the options vary by time period. For years 1-b,
option II has the least inpact on profitability; options 1, II and IV are very
similar; and option V has the most severe impacts on profitability. By years
11-15, those options involving mandatory tank retirement or upgrading (options
II, IV and V) allow greater profitability for large firms than options I and
II, which permit older tanks to remain in place (as long as they do not have
to be replaced because of releases).
The effect of paying tank upgrade or replacement costs alone causes pro-
fits in years 1-5 to rise under all options by years 11-15. Under Options I,
II, III, and IV, pre-regulation profit levels are reached by years b-10 if
only the costs of tank upgrading and replacement are considered. Under Option
V, if only the costs of tank upgrading and replacement are considered, large
firms have higher rates of return by years 11-15 than their pre-regulation
rates of return.
Analysis of the Impacts of Regulatory Options on Outlets Owned by Small Finns,
Assuming That Revenue Increases Are Possible
The analysis was performed assuming that the revenues of motor fuel mar-
keting firms will not increase to cover their increased regulatory costs.
However, the revenues of these firms could increase, for two different reasons:
• The price of retail motor fuels could rise and provide a higher mar-
gin that could be used to pay regulatory costs.
• As outlets exit the industry, the volume of motor fuel sales at the
remaining outlets could increase. This increase in volume would
provide the firm with higher profits, because, on the margin, the
relative fixed costs of operation would decline. This would enable
owners to pay their regulatory costs from this higher profit margin.
In the retail motor fuel marketing sector, a 1 percent revenue increase with
no cost increase can be achieved either by a 1 percent price increase (with no
change in the volume of motor fuels sold) or a 12-15 percent increase in the
volume of motor fuels sold (with no change in price).
Exhibit 6.22 compares the percentage of outlets owned by small firms, sur-
viving at the close of years 5, 10, and 15 under the following scenarios: if
no additional UST regulations are inposed; if Option II regulations are im-
posed and there is no revenue increase; and if Option II regulations are im-
posed and revenue increases of 1, 3, or 5 percent occur, respectively. The
"base case," i.e., the percentage of outlets surviving if no additional UST
regulations are iiiposed, is the benchmark for measuring the incremental im-
pacts associated with the other scenarios. As Exhibit 6.22 shows, if existing
outlets continue to exit the industry at this base rate, 81 percent of small-
firm-owned outlets existing today will continue to survive at the end of b
years. Even a 5 percent revenue increase is not quite sufficient to maintain
the base case survival rate at the end of year 5 if the owners incur all Op-
tion II regulatory costs (including corrective action costs). However, by
-------
b-43
Exhibit 6.22
ECONOMIC IMPACTS OF REGULATION UNDER OPTION II ON
SMALL FIRMS IN THE RETAIL MOTOR FUEL MARKETING INDUSTRY,
ASSUMING REVENUE INCREASES OF 1, 3, OR 5 PERCENT
Percentage of Outlets Surviving
Scenario
Year S
Year 10
Year lb
Base Case - No Further UST Regulation
Survival Based on the Continuation
of Past Exit Trends
81
72
b4
Regulation Under Option II
No Revenue Increase
Tank Upgrade Only
All Regulatory Costs
1 percent Revenue Increase
Tank Upgrade Only
All Regulatory Costs
3 percent Revenue Increase
Tank Upgrade Only
All Regulatory Costs
5 percent Revenue Increase
Tank Upgrade Only
All Regulatory Costs
79
29
97
41
100
b2
100
74
72
21
96
32
99
b4
100
b8
bb
15
94
2b
99
47
99
b3
Source: Meridian Research, Inc. and Versar Inc., using affordability model
results.
-------
6-44
year 10, a 5 percent revenue increase will almost maintain base case survival
rates. Thus it is possible that a relatively modest price increase, or a
relatively large increase in the volume of retail motor fuel sold (or some
combination of these two) could help to mitigate the adverse impacts of Option
II UST regulations on the survival of sma11-firm-owned outlets in this segment.
Exhibit 6.23 compares the average per outlet ratio of net income to total
assets for large firms for the base case and the same revenue increase scen-
arios as were shown in Exhibit 6.22 for small firms. For large firms, a
3 percent revenue increase would be adequate to achieve a higher ratio of net
income to total assets per outlet than exists in the base case. By years
6-10, a 1 percent revenue increase would be adequate to achieve slightly
greater than base case rates of return on assets.
If all of the firms in an industry incur the same cost increase, economic
theory suggests that, in the long run, prices will rise to cover these costs
(although the industry's total output may decline). However, if all of the
firms in the industry do not incur the same cost increase, the ability of
those firms incurring higher costs to increase prices to cover these costs is
constrained.
In the UST regulatory context, all firms will not incur the same cost
increases. The cost increases that are the source of most of the impacts for
Option II are those for corrective action. These costs will rarely be in-
curred by new outlets or those with upgraded tanks. Even among firms owning
older bare steel tanks, the costs of corrective action will vary randomly in
any given year. Some small firms will, through good luck alone, have no re-
leases. Others may only have releases costing less than $60,000, and some
small firms will have releases costing more than $100,000. Large firms, with
greater numbers of outlets, will have less variation in their annual costs.
For example, a firm with 100 outlets can expect in any given year to incur
costs for all outlets approximately equal to the averaye cost per outlet times
the number of outlets (approximately $27,000 per outlet for years 1-b of Op-
tion II). A large firm will have better access to credit to cover an unusu-
ally costly year than a small firm will. As a result of these factors, large
firms will typically have higher costs per outlet than small firms that are
fortunate enough to have no releases, but much lower costs than small firms
with a large plume release. In the absence of insurance or other risk pooling
measures (risk retention associations or State funds), a price increase ade-
quate to ensure profitability for large firms would still not be adequate to
forestall closure for those small firms incurring the costs of an expensive
corrective action.
The results in Exhibit 6.23 show that a 3 percent revenue increase is
adequate to ensure large firms a considerably better rate of return in any
post-regulatory interval than they have today. However, with this same reve-
nue increase, 19 percent of the outlets owned by small firms will be forced to
exit as a result of regulatory costs by year 5.
However, an increase of 3 percent may not be possible even for older out-
lets owned by large firms. Attempts by firms with older outlets to maintain
prices at a level significantly above the price at which gasoline can be sold
at newly constructed outlets will not succeed over the long run, particularly
-------
Ext 6.23
ECONOMIC IMPACT OF REGULATION UNDt.x OPTION II OH THE KATIO OF NET INCOME
TO TOTAL ASSETS PER OUTLET FOR LARGE FIRMS, !/ ASSUMING REVENUE
INCREASES OF 1, 3, OR b PERCENT
Scenario
Value of the Ratio of Net Income to
Total Assets
per Outlet During:
Years 1 - S Years
6-10
Years 11 - lb
Base Case - No Further UST Regulation
No Revenue Increase
.042
.042
.042
Regulation Under Option II
No Revenue Increase
Tank Upgrade Only
.038
.048
.Obi
All Regulatory Costs
-.086
.000
.020
1 Percent Revenue Increase
Tank Upgrade Only
.077
.06b
.0b8
All Regulatory Costs
-.019
. 04b
.061
3 Percent Revenue Increase
Tank Upgrade Only
.147
. lbb
.lb/
All Regulatory Costs
.071
.11!#
. 132
5 Percent Revenue Increase
Tank Upgrade Only
.216
.224
.22b
All Regulatory Costs
.142
.188
.201
Source: Meridian Research, Inc. and Versar, Inc., using affordability model results.
1/ Large firms are defined as firms with annual sales greater- than $4.6 million that ao not engage in
petroleum production or refining.
-------
6-46
in an industry such as retail gasoline marketing, where the cost of entry is
relatively low. For example, a new convenience store with both gasoline and
grocery sales can be constructed and stocked for less than $2b0,0U0; this
figure is only three times the cost of replacing the tanks at an existing
station and is less than the cost of a large corrective action.
The primary limitation on the ability of existing outlets to increase
their sales when other outlets exit the industry is that new entrants witn
protected tanks may be able to price their motor fuels at a level lower than
that of existing outlets, which may have to recoup corrective action or tank
retrofit costs they have incurred.
Impact on Outlets Operated by Lessee Dealers
Outlets operated by lessee dealers represent a peculiar combination of
large and small firms. As shown in Exhibit 6.3, 36,817 leased outlets
(62.8 percent) are owned by refiners, 20,713 leased outlets (35.3 percent) are
owned by jobbers, and 1,127 leased outlets (1.9 percent) are owned by indepen-
dent chains. All of the refiners and independent chains, and virtually all of
the jobbers, are large businesses. All lessee dealers themselves, however,
are small businesses.
The median financial statistics for single-station lessee dealers (who
comprise about three-quarters of all lessee dealers) are: $82,000 in assets,
$62,000 in net worth, $6,000 in annual after-tax profits, and a rate of return
on assets of 7.2 percent. This compares with $210,000 in assets, $90,000 in
net worth, $14,000 in annual after-tax profits, and a rate of return on assets
of 6.67 percent for the median open dealer; thus the average lessee dealer has
less than half the assets of the average open dealer but has a similar rate of
return.
The inpact on lessee dealers depends critically on who bears the initial
impact and whether it can be passed on. Under the most common current lease
arrangements, the owner of the outlet also owns the tank, and thus the owner
bears the principal impact. In the long run, however, there is considerable
question about whether the owner will actually bear the burden. One possibil-
ity, which some lessors are already attempting, is to change the terms of the
lease to make the lessee bear greater direct responsibility for the tanks
themselves. Another possibility, at least for leases that expire within five
or ten years, is to raise the rent to cover at least the routine regulatory
costs (if not the costs associated with an actual release and tank replace-
ment) .
To the extent that the owner bears the impacts of costs, this lessee
dealer analysis is similar to that for outlets owned by large firms. As shown
in Exhibit 6.21, the impact on large firms of tank replacement costs only is
relatively small, and it is unlikely that these costs would cause an owner to
close a lessee dealership or substantially change the terms of the lease.
Corrective action costs under all options, however, have substantial impacts
on the profitability of large firms, which may cause them to close lessee-
operated outlets. If large firms can increase revenues by 3 percent, there
would be little if any inpact on their lessee dealers under Option II.
-------
6-47
If rates of return are reduced to less than 3 percent, tnere is a strong
incentive to pass costs through to the lessee. Thus it is instructive to look
at the potential inpacts on lessees if they are required to bear these costs.
This analysis is similar to that for open dealers (presented above). An
analysis of the median lessee dealer snows that:
• The median lessee would be forced into severe financial distress as a
result of having to clean up a non-plume release and would fail as a
result of having to replace one tank.
• The median lessee—like the median open dealer—would fail in the
event of a plume release even before the leaking tank is closed and
replaced.
• The median lessee would be pushed at least to the brink of failure if
required to replace one tank.
For the median lessee-operated outlet, the costs of a non-plume release
are equal to two-thirds of net worth; the costs of a plume release are over
twice such a dealer's net worth, and the costs of closing and replacing one
tank exceed half of the lessee's net worth. In contrast to these impacts, the
impacts on the typical lessee dealer of upgrading a tank are relatively
modest. The $3,050 cost of cathodically protecting a tank would leave the
median lessee in good financial condition, with a return on assets of 4.1
percent in the year in which this cost is incurred.
The impact of the regulatory options on lessees is difficult to judge, for
several reasons. The owner of the outlet is clearly somewhat better able to
sustain the impact of regulatory costs than the lessee, since in the long run
the lessor firm's profits remain positive and large owners are better able to
sustain the short-run losses related to a release than is a lessee of a single
outlet. The exact terms of a lease, however, will play a critical role in who
bears the ultimate burden. The strategic goals of the owner may also play a
role. If the outlet is marginal, for example, the owner may not want to
assume the risk and may aggressively try to shift the risk and burden to the
lessee, which would tend to increase the chances of the lessee's failure.
There is also a possibility that an owner would close a site to minimize out-
lays for corrective action (e.g., replacing a tank) or would take the oppor-
tunity to change the use of the site (e.g., from full service to "pumper") if
the site location were especially valuable. In this type of situation a
lessee could be forced out of business even_ if he did not directly bear the
costs of regulation or of corrective action. Lease arrangements that clearly
hold the owner responsible for major costs of corrective action and tank re-
placement provide a lessee with .considerable protection. Yet if these costs
can be passed through, their impact on lessees will generally be greater than
the impacts on open dealers.
6.B.4 Interactions with Financial Responsibility Requirements
This economic impact analysis assumes that there will be no insurance or
State corrective action and compensation funds available to pay the costs of
corrective action or other expenditures required by the UST technical stand-
ards. EPA is also issuing proposed financial responsibility requirements for
-------
6-48
owners and operators of USTs. These requirements are covered in a separate
regulatory impact analysis (RIA).
The proposed financial responsibi1ity requirements will require that all
UST owners or operators provide financial responsibility in the amount of
$1,000,000 per occurrence and an annual aggregate amount in accordance with
the following schedule:
Aggregate Level
$1 million
$2 mi 11 ion
$3 mill ion
$4 million
$5 million
$6 million
Number of Tanks Covered
1-12
13-6U
61-140
141-250
251-340
341 or more
The proposed regulation allows an owner or operator to satisfy these financial
responsibility requirements by means of any one, or a combination, of the
following financial mechanisms: insurance, guarantee, indemnity agreement,
surety bond, letter of credit, qualification as a self-insurer, or State or
local corrective action and compensation fund. Subtitle I allows the Adminis-
trator to suspend the Agency's enforcement of these financial responsibility
requirements if a class of UST facilities is unable to obtain mechanisms for
financial responsibility and additionally meets certain conditions with regard
to taking steps to form a risk retention group or if the State takes certain
steps toward forming a State fund.
Of the mechanisms that can be used to satisfy an owner or operator's fin-
ancial responsibility requirements, only insurance and State corrective action
and compensation funds have the potential to mitigate significantly the
economic impacts of the technical standards. Insurance is currently available
only to multi-facility chains other than refiners. In the small business
segment of the retail motor fuel marketing sector, insurance is available only
to small jobbers and small C-store chains and is not available to open
dealers. There is no active market for UST insurance in the many UST-using
industry sectors other than retail motor fuel marketing.
Both the future availability and future costs of insurance to cover cor-
rective action and compensation of third parties resulting from UST releases
are uncertain. In the Financial Responsibility RIA, EPA estimated the costs
of insurance based on current UST insurance premiums and on projections of
insurance premium costs for 1987 provided by the leading current insurer of
USTs. The Financial Responsibility RIA examines the availability of UST in-
surance using various alternative scenarios that estimate how many USTs that
are not currently insured will be able to obtain insurance in the future.
The availability of insurance is currently limited both by capacity con-
straints within the insurance industry and by the number of USTs that fit the
-------
fa-49
profile that insurers are willing to insure. Currently, insurers limit UST
insurance to firtns:
• Having USTs with an average age of no greater than 17 years, or, in
the case of some insurers, with an age of less than 20 years;
• Belonging to an industry association; and
• Having records documenting a minimum of 6 months of daily inventory
control before applying for insurance.
In the future, insurers may drop the requirement that a firm belong to an
industry association, but a more stringent requirement than inventory control
for the prior identification of releases, such as tank tightness testing, may
become a precondition to obtaining UST insurance. Prior identification of
releases is important to insurers because policies do not cover claims for
damages from UST releases identified before the effective date of the policy.
As a result of these restrictions and the nature of the insurance claims
process, the number of UST releases covered by insurance will be significantly
less than the total number of UST releases. Specific reasons for these dif-
ferences are the following:
• To the extent possible, insurance excludes coverage for prior re-
leases because it requires owners to furnish inventory control
records or to perform other release identification measures before
a policy is issued; however, prior releases are expected to
constitute a substantial portion of all UST releases identified
within the next 5 years.
• The tank age restrictions inposed by insurers exclude that segment
of the total UST population most likely to have the highest
probability of releases.
• Firms covered by insurance do not report all releases to insurers.
In particular, releases having costs that are less than the
policy's deductible may not be reported to insurers.
As a result of the factors listed above, the insurance claims rates and
premium costs used in the Financial Responsibility RIA are considerably lower
than implied by the probabilities of corrective action estimated in this RIA.
A further factor differentiating the two analyses is that insurance claims
rates are based on existing tank replacement practices while the technical
standards RIA assumes that tanks are not replaced unless they have a release;
which is an assumption necessary to the development of the maximum impact
scenario for that RIA. As a result of these differences, the insurance claims
rates cannot be compared to the probabilities and costs of corrective action
events developed elsewhere in this RIA.
Insurance will thus be available only to a portion of small UST-using
businesses, and then only to cover some of the potential costs of corrective
action. Initially, insurance will oe available only to a limited number of
-------
6-50
small businesses, and these will be those businesses least likely to experi-
ence a release requiring corrective action. Over the long run, a larger por-
tion of the affected small businesses may become eligible for insurance as
existing releases are corrected and as old tanks are replaced or upgraded.
For those small businesses able to obtain insurance, the Financial
Responsibility RIA found the mitigative effect of insurance on the economic
impacts of the technical standards to be minimal. Leak detection and tank
upgrading/replacement costs are not covered by UST insurance, and thus insur-
ance has no effect on the inpacts of these costs. Insurance will pay for
corrective action costs above the level of the deductible in the policy
(normally $5,000 for small businesses).
For those small businesses able to obtain insurance, the Financial
Responsibility RIA found that initially an estimated 0.7 percent of small
business-owned outlets would close as a result of the costs of insurance pre-
miums. However, over the long run (10 years), EPA estimates that fewer firms
would exit the industry if they had UST insurance than if they did not. The
costs of insurance premiums thus force some low-profit, marginal open dealer
firms to close. However, among larger, more profitable open dealer firms and
small business chains, fewer outlets would close because of paying insurance
premiums than would close as a result of meeting the costs of their UST-
related corrective action and third-party liability awards from their own
funds.
Currently, there are no State compensation and liability funds that fully
meet Subtitle I's requirements for use as a financial responsibility mechan-
ism. The extent to which such programs mitigate economic impacts depends on
how they are set up. At one extreme, a fund that paid for all corrective
actions and provided low-interest loans to small businesses for tank replace-
ment and that was based on a gasoline tax would virtually eliminate the econ-
omic inpacts on small business of the technical standards. At the other ex-
treme a fund based on tank fees and that paid only corrective action costs for
financially insolvent UST owners or operators would do nothing to mitigate the
economic impacts of these standards on small businesses.
In summary, unless the availability of insurance and types of firms able
to obtain it alter greatly, insurance will not significantly mitigate small
business economic impacts. State compensation and liability funds may miti-
gate the economic impacts of the technical standards, but it is uncertain
whether such funds will come into being and whether they will be designed in a
way that permits them to mitigate economic impacts.
-------
6-51
6.C. IMPACTS ON FIRMS USING USTS FOR NONRETAIL MOTOR FUEL STORAGE
The UST regulations will impose requirements on facilities in many other
industries besides those that sell motor fuels at retail, including those
storing petroleum products for their own use or for the wholesale market.
This section analyzes potential economic impacts in industries (other than
retail gasoline sellers) that have been identified as including significant
numbers of petroleum tank owners.
This section identifies those industries that are likely to be affected
by UST regulations because a significant fraction of the facilities in the
industry own or operate underground tanks which contain petroleum products.
The industries selected are based on a detailed analysis of underground tanks
in California. These nonretail petroleum product tanks are primarily used to
service in-house vehicles (noncoirenercial tanks of less than 1100 gallons are
exempt). Depending on the nature of competition in the specific industry and
the relative position of the firm, a variety of responses to the imposition of
regulatory costs are possible.
o In many industries, the costs of tank ownership are small relative
to the overall cost and activity structure of the average facility.
Expected levels of regulatory costs could be absorbed without sig-
nificant impacts on profitability.
o In industries in which a large number of firms must meet the regulatory
requirements (i.e., where a large percent of the firms in the industry
have underground tanks), the fact that a firm's competitors face similar
increases in costs may permit cost pass-throughs.
o In industries in which there is little competition among the firms—for
example, because of geographic market segmentation or a high degree of
product differentiation—a firm facing increased regulatory costs may
also pass all or part of these costs on to consumers, as it is not
likely to lose the customers to competitors. Tank-owning firms that
are defense contractors to the federal government typically meet this
description.
o The study identifies many industries in which only a small percent of
firms have tanks, the tank-owning firms have only one or two petroleum
tanks, and the firm is likely to have close competitors that do not
operate underground tanks. If the regulatory costs are relatively
high, they can be expected to take tanks out of service and purchase
the needed gasoline or other petroleum products from retailers. In
this case the firm could save the costs of owning and operating the
tank, capture some salvage value from the used tanks, and avoid the
costs of regulatory compliance. These would be partially offset by
the increase in retail costs over wholesale costs for supplying its
petroleum needs.
o Part of the cost of any regulatory requirement will be shifted to the
public, including both consumers and nonconsumers, through the lowering
-------
6-52
of corporate taxable income. This arises from deductions for higher
operating costs, tax credits for investment, and--if price rises result
in decreased sales—lower net income from sales. The public is also
likely to pay when the tank owner is a government entity rather than a
private corporation.
6.C.I. Overview of Approach
The first step in the analysis is identifying the affected industries.
The analysis was performed for those industries that own nonretail petroleum
underground storage tanks. Available data were used to estimate the fraction of
each industry that owns USTs and the average number of tanks per facility for
that fraction of the industry. The potential burden for tank-owning firms in
each industry was determined by a screening analysis. This analysis essentially
compares potential regulatory costs to profits for firms of different size
classes within the affected 4-digit SIC code industry groups.
Identify the Industries with Relatively Large Numbers of USTs
A complete registry of underground storage tanks exists for the State of
California. (Similar data are not yet available from other states or on a
national basis.) This database has been processed to identify the industry
category of each UST-owning firm, the number of tanks, their size, and their
contents. This serves as the data base for the assumed distribution of nonretail
petroleum tanks across SICs and within SICs. Based on available data, nonretail
petroleum tanks are used in all major sectors of the economy: mining and construc-
tion, manufacturing, wholesale trade, retail trade, and services. (USTs are also
prevalent in agriculture, utilities, and transportation sectors, but are not in-
cluded [in this draft] because of difficulties in developing financial profiles.)
There are also tanks owned hy public-sector entities and nonprofit organizations.
However, data on their incidence are not available, so they have not been included
in this analysis.
Estimate the Fraction of Each Industry Group that Owns Tanks and
Estimate the Average Number of Tanks per Tank-Owning Facility
An EPA contractor report identifies the number of UST-owning facilities in
California for each SIC code in the database, and identifies the number of tanks
in each SIC category. Using the "County Business Patterns" publication of the
1982 Census, we-identified the total number of establishments in the state for
each relevant SIC code, and thus we were able to calculate the percentage of
establishments in each SIC code that uses USTs for nonretail petroleum products.
This percentage is used as an estimate of the nationwide percentage of firms
owning tanks for each industry. For petroleum tanks other than retail gasoline
sellers, the value ranges from 100 percent (paper mills) to less than one percent
(liquor stores, engineering and architectural services). (Since the number of
facilities with tanks and the number of facilities in the state come from different
sources several years apart, the percent of each industry owning tanks is only a
rough approximation.) In the screening results presented below, only those SICs
-------
6-53
for which greater than five percent of establishments have nonretail petroleum
tanks are presented. It is assumed that, if only a small percentage of estab-
lishments within the industry maintain underground tanks, there are alternatives
which would tend to limit the potential for economic dislocation.
By dividing the number of tanks by the number of UST-owning facilities, we
were able to calculate the average number of tanks per tank-owning facility.
For nonretail petroleum tanks this ranges from 5.6 tanks per firm (electronic
computing equipment) to 1.1 (furniture stores). However, within each industry
group, the number of tanks per firm ranges from zero to several times the
industry's average. Presumably, bigger firms would have a larger number of
tanks per firm (if only because there would be a larger number of establishments
per firm). However, no data are available from which to estimate a relationship
betwen firm size and number of tanks per firm in the various industries.
Compare Regulatory Compliance Costs to Significant Descriptors of
FinancTal Performance of the Tndustry
We examined a number of sources of data on financial statistics of firms
by industry. The one selected for use, on the basis of timeliness and complete-
ness, is "'85 Annual Statement Studies" published by Robert Morris Associates,
Philadelphia, Pa. Robert Morris Associates (RMA) is a national association of
bank loan and credit officers, so the sample that forms the basis of these
financial statements are presumably firms that have applied for credit in
recent months.
RMA statement studies include data for many four-digit SIC code in several
size categories: assets of less that one million dollars, one to ten million
dollars, 10 to 50 million dollars, and 50 to 100 million dollars. (Data are
provided for an asset size group only if the sample in that size range consists
of at least 10 observations.) Two larger asset size groups are included, but
for most industries there are fewer than 10 observations for these size groups,
so detail is not presented. Furthermore, firms with large assets are likely to
have multiple facilities, which would increase the regulatory costs per firm.
Regarding regulatory costs, rather than using specific costs for regulatory
alternatives for this exercise, screening was performed in a more generic way.
Three screening exhibits are presented:
o Exhibit 6.25 assumes regulatory costs of $500 per tank (detection).
o Exhibit 6.26 assumes regulatory costs of $5,000 per tank (replacement).
o Exhibit 6.27 assumes regulatory costs of $50,000 per tank (corrective
action).
The $500-per-tank detection cost may be thought of as the annual incremen-
tal costs of performing an annual tank test and would be borne by all affected
tanks. Other potential regulatory costs include the incremental costs of tank
replacement (i.e., those costs above and beyond the costs of normal tank replace-
ment) and incremental costs of corrective action (i.e., those costs above and
beyond the corrective action costs that would be undertaken in the absence of
regulatory change). In a crude sense, incremental tank replacement costs might
-------
1311
1442
1611
1761
1794
2011
2016
2033
2048
2051
2084
2086
24H
2431
2448
2621
2711
2831
2851
2861
2911
3273
3312
3321
3443
3444
3523
3662
3671
3728
3731
4225
4463
4811
4953
ibit 6.25
SCREENBC ANALlsIS PCR MW-RETAIL PETOdELM TANKS
Regulatory Costs of $500/Tank
sic industry group
Nutter of
establishments
(all sizes)
Percent
of Industry
with tanks
Pretax profit per firm
by firm size
UST costs
per
establishment
Snail Medium
($1000)-
Large
($1000)
Pretax return on assets
Snail
Before After
Mediun
Before After
Large
Before After
CRUDE PETRCLELN & NATURAL GAS
8,712
54.3
109
241
418
0.9
21.IX
20.9X
4.9X
4.9X
ooNsnsjcncN sand and gravel
2,270
54.8
32
207
~
1.5
6.3X
6.OX
7.2X
7.2X
HIGHWAY AND STREET OCNSTRlXjnCH
9,651
7.0
29
89
462
1.4
6.9X
6.6X
4.IX
4.IX
ROOFING AND SHEET METAL MXK
17,487
5.1
24
72
—
0.8
9.OX
8.7X
5.9X
5.8X
EXCAVATION AND FOUNDATION WORK
11,050
9.0
21
167
719
1.1
6.IX
5.8X
10.7X
10.6X
1.7X
4.8%
9.0Z
1.7X
4.8X
9.OX
MEAT PAOCTtC PLANTS
1,689
42.0
25
529
1,755
1.1
4.OX
3.9X
12.3X
12.3X
7.7X
7.7X
POULTRY CRESS I1C PLANTS
364
35.7
—
664
4,365
1.5
—
—
15.IX
15.IX
15.9X
15.9X
CANNES FRUITS AND VEGETABLES
676
35.0
—
8
3,011
1.2
—
—
0.2X
0.2X
12.OX
12. OX
PREPARED FEIDS, NEC
1,767
39.6
(42)
360
1,950
1.2
-6.7X
-6.9X
10.7X
10.6X
8.5X
8.5X
BREAD, CAKE AND RELATED PRODUCTS
2,112
21.9
33
426
1,703
1.3
8.7X
8.4X
9.6X
9.6X
8.6X
8.6X
WINES, BRANDY AND SPIRITS
348
28.4
—
(16)
309
1.0
—
—
-0.4X
-0.4X
1.3X
1.3X
BOTTLED AND CANNED SOFT QUNKS
1,547
93.3
25
463
2,400
1.0
5.IX
4.9X
10.4X
10.4X
9.5X
9.5X
SAMOLLS i PLANING MILLS
5,531
34.3
39
299
177
1.4
7.2X
7.OX
9.OX
8.9X
0.7X
0.7X
MILLWSK
2,121
6.9
52
430
3,610
1.1
10. IX
9.9X
13.4X
13.4X
19.6X
19.5X
WOOD PALLETS & SKIDS
1,469
17.5
54
—
—
1.1
12.3X
12.OX
-
—
—
—
I
tn
-Ck
PAPER MILLS, EXCEPT BUILDING PAPER
NEWSPAPERS
PHARMACEUTICAL ITCPARATICNS
PAINTS & ALLIED PRCDUCTS
INDUSTRIAL INORGANIC CHEMICALS, NEC
350
8,223
666
1,379
582
100.0
6.3
13.1
19.9
30.6
90
353
3,086
1.7
16.4X
16. IX
10. IX
10.OX
12.2X
12. IX
42
277
2,599
1.3
8.3X
8.OX
7.3X
7.2X
12.IX
12. IX
75
351
3,163
1.4
14.5X
14.2X
11.2X
11.2X
12.8X
12.8X
49
310
3,013
0.8
8.7X
8.6X
8.5X
8.4X
12.6X
12.6X
38
326
1,618
0.8
7.OX
6.9X
9.IX
9. IX
6.IX
6.IX
PEIHXELM REFINING
460
53.1
—
1,202
3,741
1.4
—
—
32.IX
32.7X
17. IX
17. IX
READY-MIXED OCNCRETE
5,088
67.5
54
241
1,741
1.1
9.IX
8.9X
7.9X
7.9 X
7.8X
7.8X
BLAST FURNACES & Sim. MILLS
521
29.2
61
126
1,291
1.9
11.7X
11.4X
2.8X
2.8X
5.6X
5.6X
GRAY IRON POUNEEUES
867
44.4
63
148
623
1.0
14. OX
13.8X
4.OX
4.OX
2.9X
2.9X
FABRICATED PLAHMCRK
1,742
9.0
20
114
1,267
1.1
3.4X
3.2X
3.6X
3.6X
6.4X
6.4X
SHEET METAL WORK
3,456
7.0
35
275
2,246
0.7
7.6X
7.5X
9.3X
9.3X
10.5X
10. sx
FARM MACHINERY & EQUHMEMT
1,748
8.5
(9)
121
785
1.0
-1.7X
-1.8X
3.4X
3.4X
3.6X
3.6X
RADIO i TV CCtMJNICAXIGN EQUHWENT
2,059
5.5
1
288
2,632
2.0
0.2X
-0.2X
7.6X
7.5X
11.9X
11.9X
SEMICONDUCTORS & REIATED DEVICES
89
9.3
73
381
3,135
0.7
13.6X
13.5X
10.7X
10.7X
13.3X
13.2X
AIRCRAFT EQUIPMENT, NEC
849
9.7
43
301
—
2.2
7.2X
6.8X
8.7X
8.7X
—
—
SHIP BUILDING & REPAIRING
GENERAL WARHOLS DC AND SICRtCE
MARINE CARGO HANDLBC
TELEEHUE OOMJNI CATION
REFUSE SYSTE16
616
2,284
796
13,124
4,251
23.9
26.2
21.9
42.1
35.8
(35)
45
104
821
189
547
458
328
228
(420)
3,265
3,309
1.5
0.8
1.0
1.0
1.4
-7.3X
9.4X
20.9*
188.3X
-7.7X
9.3X
20.7X
187.9X
5.OX
14.9X
9.7X
10. OX
9.5X
4.9X
14.9X
9.7X
9.9X
9.5X
-1.9X
13.21
13.9X
-2. OX
13.2X
13.9X
-------
5012
5031
5039
5051
5074
5083
5084
5093
5141
5143
5148
5153
5161
5171
5172
5181
5211
5261
5311
5451
5531
7211
7213
7261
7342
7391
7394
7512
7692
7997
SIC INDUSTRY GROUP
AUTCMBILES & OTHER MJICR VEHICLES
UMBER, PLYWOOD & MTLUOPK
CCNSISUCTIGN MATERIALS, NBC
METALS SERVICE OUTERS & OFFICES
PliWBDC AND HYERCNIC HEATING SUPPLES
FARM MACHINERY, WO&ESALE
INDUSTRIAL MACHINHIY, WBXESALE
SCRAP AND HASTE MAHRI/LS
GROQRIES, GENERAL
DAIRY PRODUCTS
FRESH FRUITS & VEGETABLES, VHXSALE
GRAIN
CHEMICALS, UBOLESALERS
PETWXELM SULK STATICHS L TEMGNALS
PETRCLELH PRCDICTS, WELESALE
BE2R & ALE
LUMBiR, BUHDIHS MATERIALS DEALERS
DETAIL NURSERIES & GARDEN STCRES
DEPARTMENT STCRES
DAIRY PRODUCT STCRES
AUTO & HOME SUPPLY STCRES
POWER LAIKDERIES
LINEN SUPPLY
FUNERAL SERVICES i CREMATCRIES
disinfectm: and exierminatiig
RESEARCH & DEVELOPMENT LABCGATCRIES
EQUIPMENT REMTAL & LEASING
PASSENGER CAR RENTAL i. LEASItC
WELDHC REPAIR
MEMBERSHIP SPCRTS & RECREATION CLUBS
Exhib ,25 (Continued)
SCREEinfC ANALYSIS FOR NON-RETAIL PEIKXELM TANKS
Regulatory Costs of $500/Tank
Nunber of Percent UST coats
establishments of Industry Pretax profit per flan per Pretax return an assets
(all sizes) with tanks by firm size establishment
Small Medium Large
Snail Medina Large Before After Before After Before After
($1000) ($1000)
5,829
5.2
49
268
3,342
0.9
8.4X
8.3*
8.4*
8.4*
16.3*
16.3*
6,889
9.6
58
303
1,893
0.9
10.6Z
10.4*
9.4*
9.4*
9.8*
9.8*
9,068
18.7
38
261
1,776
0.9
7.4*
7.2*
8.6*
8.6*
8.7*
8.7*
8,754
5.2
35
211
1,138
0.9
6.7X
6.5*
5.5*
5.5*
5.4*
5.4*
7,955
9.3
35
194
1,054
0.7
7.0*
6.8*
6.8*
6.8*
7.0*
7.0*
13,617
16.6
4
69
1,411
1.2
0.7X
0.5*
2.4*
2.4*
7.2*
7.2*
23,413
6.7
39
185
689
0.9
7.5*
7.3*
6.0*
6.0*
3.3*
3.3*
8,454
8.9
53
161
1,824
0.9
10.5*
10.3*
5.1*
5.0*
9.1*
9.1*
3,594
6.2
40
183
1,401
1.4
7.2*
6.9*
5.1*
5.0*
6.8*
6.8*
3,645
8.4
(37)
268
3,140
0.8
-6.2*
-6.3*
8.1*
8.1*
12.0*
12. OX
5,176
5.6
62
198
1,122
1.0
11.7*
11.5*
7.1*
7.1*
6.5*
6.5*
8,377
14.8
88
152
2,512
1.0
14.5*
14.4*
4.9*
4.9*
11.4*
11.4*
9,380
5.2
52
200
1,218
0.8
10.1*
10.0*
6.2*
6.2*
5.9*
5.9*
13,155
43.9
44
238
194
1.9
7.5*
7.1*
7.6*
7.5*
1.1*
1.1*
4,990
44.1
64
237
6,029
1.7
11.7*
11.4*
8.1*
8.1*
30.1*
30.1*
4,483
25.7
42
268
2,265
1.3
7.5*
7.3*
8.0*
7.9*
12.0*
12.0*
24,268
12.3
42
198
1,405
0.8
8.0*
7.8*
6.6*
6.6*
7.1*
7.1*
7,162
19.4
25
127
—
0.9
5.7*
5.5*
5.0*
5.0*
—
—
9,767
12.3
41
125
1,348
0.9
7.6*
7.4*
3.7*
3.6*
5.7*
5.7*
5,375
10.0
32
535
—
1.4
9.1*
8.7*
12.3*
12.3*
—
—
39,071
5.0
26
139
2,296
0.8
6.5*
6.3*
5.5*
5.4*
11.0*
11.0*
2,122
8.6
36
228
—
1.1
9.5*
9.2*
9.2*
9.2*
—
—
1,294
29.3
62
111
—
1.3
12.9*
12.6*
3.1*
3.1*
—
—
14,951
5.0
30
66
—
0.7
8.6*
8.4*
2.3*
2.3*
—
—
6,575
8.6
27
340
—
0.7
9.8*
9.6*
9.5*
9.5*
—
—
2,717
5.9
35
439
1,773
1.4
7.3*
7.0*
13.0*
12.9*
6.8*
6.8*
15,594
10.4
48
338
722
1.2
11.0*
10.7*
10.5*
10.4*
3.6*
3.6*
4,568
26.7
55
158
1,335
1.1
10.7*
10.5*
4.4*
4.3*
5.8*
5.8*
4,411
8.1
42
408
—
0.6
11.1*
10.9*
12.2*
12.2*
—
—
11,231
13.8
17
45
—
0.8
3.1*
3.0*
1.7*
1.7*
—
~
-------
1311
1442
1611
1761
1794
2011
2016
2033
2048
2051
2084
2086
2421
2431
2448
2621
2711
2831
2851
2861
2911
3273
3312
3321
3443
3444
3523
3662
3671
3728
3731
4225
4463
4811
4953
libit 6.26
sommic ANALYSIS FOR NCH-BEIAIL PETROLEUM TANKS
Regulatory Costs of $5000/Tank
SIC INDUSTRY GROUP
Nmfcer of
flafflhl I «|impnfiy
(all sizes)
Percent
of Industry
with tanks
Pretax profit per firm
by fins size
Snail Medlun Large
($1000)
UST costs
per
establishment
($1000)
Pretax return on assets
Smll
Before After
Medlun
Before After
Large
Before After
CRUDE PETRGLELM £ NATURAL GAS
8,712
54.3
109
241
418
9.0
21.1*
19.3*
4.9*
4.8*
1.7*
1.7*
CONSIHCnOH SAND AND GRAVEL
2,270
54.8
32
207
—
14.8
6.3*
3.4*
7.2*
6.7*
—
—
HIGHWAY AND STOEET OQKSmCTICN
9,651
7.0
29
89
462
14.2
6.9*
3.5*
4.1*
3.5*
4.8*
4.7*
ROOFING AND SHEET METAL UOK
17,487
5.1
24
72
—
7.8
9.0*
6.1*
5.9*
5.3*
—
—
EfCAVATICW AND FOUNDATION UOHC
11,050
9.0
21
167
719
11.1
6.1*
2.8*
10.7*
10.0*
9.0*
8.9*
MEAT PACKING PLANTS
1,689
42.0
25
529
1,755
11.2
4.0*
2.3*
12.3*
12.OX
7.7*
7.6*
POULTRY CRESS DC HANTS
364
35.7
—
664
4,365
15.0
—
—
15.1*
14.8*
15.9*
15.9*
CANNED FRUITS AND VEGETABLES
676
35.0
—
8
3,011
11.7
—
—
0.2*
-0.1*
12.0*
12.0*
PREPARED FEEDS, NEC
1,767
39.6
(42)
360
1,950
12.2
-6.7*
-8.7*
10.7*
10.3*
8.5*
8.4*
BREAD, CAKE AND RELATES PRODUCTS
2,112
21.9
33
426
1,703
13.0
8.7*
5.3*
9.6*
9.3*
8.6*
8.6*
WINES, BRANDY AND SPIRITS
348
28.4
(16)
309
10.3
-0.4*
-0.7*
1.3*
1.3*
BOTTLED AND CANNED SOFT DRINKS
1,547
93.3
25
463
2,400
10.2
5.1*
3.0*
10.4*
10.2*
9.5*
9.5*
SAWGLLS & FLAKDC MILLS
5,531
34.3
39
299
177
13.6
7.2*
4.7*
9.0*
8.6*
0.7*
0.7*
MULUORK
2,121
6.9
52
430
3,610
11.0
10.1*
7.9*
13.4*
13.1*
19.6*
19.5*
VOCD PALLETS i SKIDS
1,469
17.5
54
—
—
11.1
12.3*
9.7*
—
—
—
—
PAPER MILLS, BXXPT BUIIMIC PAPER
350
100.0
90
353
3,086
17.1
16.4*
13.3*
10.1*
9.6*
12.2*
12.1*
NEWSPAPERS
8,223
6.3
42
277
2,599
13.1
8.3*
5.7*
7.3*
6.9*
12.1*
12.1*
PHARMACEUTICAL PREPARATIONS
666
13.1
75
351
3,163
14.1
14.5*
11.7*
11.2*
10.8*
12.8*
12.8*
PAINTS & ALLIED PRODUCTS
1,379
19.9
49
310
3,013
8.1
8.7*
7.3*
8.5*
8.2*
12.6*
12.6*
INDUSTRIAL DERGANIC CHEMICALS, NEK
582
30.6
38
326
1,618
8.2
7.0*
5.5*
9.1*
8.9*
6.1*
6.1*
PKTROTfJM REETKDG
460
53.1
1,202
3,741
13.8
32.7*
32.4*
17.1*
17.0*
READY-MIXED OCKSETE
5,088
67.5
54
241
1,741
11.1
9.1*
7.2X
7.9*
7.6*
7.8*
7.8*
BLAST FURNACES & SHU. MILLS
521
29.2
61
126
1,291
18.6
11.7*
8.1*
2.8*
2.4*
5.6*
5.6*
GRAY IRON POUNTRIES
867
44.4
63
148
623
10.0
14.0*
11.8*
4.0*
3.8*
2.9*
2.9*
FABRICATED PLAUHCRK
1,742
9.0
20
114
1,267
10.9
3.4*
1.5*
3.6*
3.3*
6.4*
6.4*
SHEET MEIAL UCRK
3,456
7.0
35
275
2,246
6.6
7.6*
6.2*
9.3*
9.1*
10.5*
10.5*
FAm MACHINERY & EQUIPMEHT
1,748
8.5
(9)
121
785
10.3
-1.7*
-3.5*
3.4*
3.1*
3.6*
3.5X
RADIO & TV OCtMUNICATICN ECJUIIWENr
2,059
5.5
1
288
2,632
19.6
0.2*
-3.7*
7.6*
7.1*
11.9*
11.8*
SEMIGCNDUCTCRS t RELATES DEVICES
89
9.3
73
381
3,135
7.2
13.6*
12.3*
10.7*
10.5*
13.3*
13.2*
AIRCRAFT EXJUimajT, NBC
849
9.7
43
301
—
22.3
7.2*
3.4*
8.7*
8.1*
—
—
CTi
I
in
CTv
SHIP BUHDDC & REPAIRHC
GJHERAL WAREH0USIN3 AND STCRiCE
MARINE CARGO HANDLING
TELEPHOKE CCmjNICATICN
RECUSE SYSTEMS
616
2,284
796
13,124
4,251
23.9
26.2
21.9
42.1
35.8
(35)
45
104
821
189
547
458
328
228
(420)
3,265
3,309
15.0
7.7
10.0
9.6
14.4
-7.3*
9.4*
20.92
188.3*
-10.5*
7.8*
19.0*
185.0*
5.0*
14.9*
9.7*
10.0*
9.5*
4.6*
14.7*
9.5*
9.7*
8.9*
-1.9*
13.2*
13.9*
-2.0*
13.2*
13.9*
-------
5012
5031
5039
5051
5074
5083
5084
5093
5141
5143
5148
5153
5161
5171
5172
5181
5211
5261
5311
5451
5531
7211
7213
7261
7342
7391
7394
7512
7692
7997
Exhib .26 (Continued)
SCREENHC ANALYSIS FOR KH-BETA1L PETROLEUM TANKS
Regulatory Costs of $5000/Tank
Ntidier of Percent UST costs
establishments of Industry Pretax profit per film per Pretax return on assets
SIC DIDUSlIQf GROUP (all sizes) with by firm size establlstnent — ¦
Stall Hedlua Large
Snail Mediim Large Before After Before After Before After
($1000) ($1000)
AUTCMBILES i OTHER KJIER VEHICLES
5,829
5.2
49
268
3,342
8.6
8.4Z
7.0*
8.4*
8.2*
16.3*
16.3*
LUMBER, PLYWOOD & MULWCRK
6,889
9.6
58
303
1,893
9.2
10.6*
8.9*
9.4*
9.1*
9.8*
9.8*
ocnstructicn materials, nee
9,068
18.7
38
261
1,776
9.2
7.4*
5.6*
8.6*
8.3*
8.7*
8.6*
HEIALS SERVICX OUTERS & OFFICES
8,754
5.2
35
211
1,138
8.7
6.7*
5.0*
5.5*
5.3*
5.4*
5.3*
PUMBHC AND HYERCHIC HEAIDC SUPPLES
7,955
9.3
35
194
1,054
6.6
7.0*
5.7*
6.8*
6.6*
7.0*
7.0*
FA8M MACHJHHOf, WBCLESALE
13,617
16.6
4
69
1,411
11.7
0.7*
-1.5*
2.4*
2.0*
7.2*
7.2*
INDUSTRIAL MACSINBOF, W9CLESALE
23,413
6.7
39
185
689
9.1
7.5*
5.7*
6.0*
5.7*
3.3*
3.3*
SCRAP AND WASTE MATERIALS
8,454
8.9
53
161
1,824
9.1
10.5*
8.7*
5.1*
4.8*
9.1*
9.1*
GROCERIES, GENERAL
3,594
6.2
40
183
1,401
13.6
7.2*
4.8*
5.1*
4.7*
6.8*
6.7*
DAIRY PRODUCTS
3,645
8.4
(37)
268
3,140
8.2
-6.2*
-7.6*
8.1*
7.9*
12.0*
12.0*
fresh nairrs s. vegetables, whxsale
5,176
5.6
62
198
1,122
10.3
11.7*
9.7*
7.1*
6.8*
6.5*
6.4*
GRAIN
8,377
14.8
88
152
2,512
9.8
14.5*
12.9*
4.9*
4.6*
11.4*
11.3*
CHEMICALS, UBXESALStS
9,380
5.2
52
200
1,218
8.3
10.1*
8.5*
6.2*
6.0*
5.9*
5.9Z
PETRCLELM BULK STATIONS & TEHGNALS
13,155
43.9
44
238
194
18.7
7.5*
4.3*
7.6*
7.0*
1.1*
1.0*
PEIRXELM PRODUCTS, VHXESALE
4,990
44.1
64
237
6,029
17.3
11.7*
8.5*
8.1*
7.6*
30.1*
30.0*
BEER & ALE
4,483
25.7
42
268
2,265
12.8
7.5*
5.3*
8.0*
7.6X
12.0*
11.9*
UMBER, BUHDING MATERIALS DEALERS
24,268
12.3
42
198
1,405
8.4
8.0*
6.4*
6.6*
6.3*
7.1*
7.1*
RETAIL NURSERIES & GARDEN STCRES
7,162
19.4
25
127
—
8.8
5.7*
3.7*
5.0*
4.7*
—
—
DEPARMNT STCRES
9,767
12.3
41
125
1,348
9.1
7.6*
5.9*
3.7*
3.4*
5.7*
5.7*
DAIRY PRODUCT STCRES
5,375
10.0
32
535
—
14.5
9.1*
4.9*
12.3*
12.0*
—
—
AUTO i SME SUPPLY STCRES
39,071
5.0
26
139
2,296
7.7
6.5*
4.6X
5.5*
5.2*
11.0*
11.0*
POWER LAJJNDERIES
2,122
8.6
36
228
—
10.7
9.5*
6.7*
9.2*
8.8*
—
—
LINEN SUPPLY
1,294
29.3
62
111
—
12.6
12.9*
10.2*
3.1*
2.8*
—
—
FUNERAL SERVICES & OttMAICRIES
14,951
5.0
30
66
—
6.9
8.6*
6.6*
2.3*
2.1*
—
—
DISDffECTDG AND EXEEEMDWIIG
6,575
8.6
27
340
—
7.0
9.8*
7.3*
9.5*
9.3*
—
—
RESEARCH & DEVEUSMWT LABCRAKREES
2,717
5.9
35
439
1,773
14.4
7.3*
4.3*
13.0*
12.5*
6.8*
6.8*
ErjarPMENr rental & leasing
15,594
10.4
48
338
722
12.2
11.0*
8.1*
10.5*
10.1*
3.6*
3.5*
PASSEN2R CAR RENTAL & LEAS DC
4,568
26.7
55
158
1,335
10.7
10.7*
8.6*
4.4*
4.1*
5.8*
5.7*
WELDDG REPAIR
4,411
8.1
42
408
—
5.8
11.1*
9.6*
12.2*
12.1*
—
MEMBERSHIP SPCRTS & RECREATION mms
11,231
13.8
17
45
—
7.9
3.1*
1.7*
1.7*
1.4*
—
—
-------
1311
1442
1611
1761
1794
2011
2016
2033
2048
2051
2084
2086
2421
2431
2448
2621
2711
2831
2851
2861
2911
3273
3312
3321
3443
3444
3523
3662
3671
3728
3731
4225
4463
4811
4953
Regulatory Costa of $50,0C0/Tank
Nurber of Percent
establishments of Industry
SIC DfflUSira GROUP (all sices) with tanks
CKUDE PEQKXBJN & NATURAL GAS 8,712 54.3
ooNsntucnoN sand and gravel 2,270 54.8
HIGHWAY AND STREET CONSTRUCTICN 9,651 7.0
ROOFING AND SHEET METAL WCHC 17,487 5.1
EXCAVAXICM AND KXJNDATICN WCKK 11,050 9.0
MEAT PACKING PLANTS 1,689 42.0
POULTRY ERESSHC PLANTS 364 35.7
CANNES FRUITS AND VEGETABLES 676 35.0
PREPARED FEEDS, NEC 1,767 39.6
BREAD, CAKE AND RELATED PRCDUCTS 2,112 21.9
WINES, BRAND? AND SPIRITS 348 28.4
BOTTLED AND CANNES SCfT HUNKS 1,547 93.3
SAMGLLS & PLANING MILLS 5,531 34.3
taLLUCRK 2,121 6.9
WOOD PALLETS & SODS 1,469 17.5
PAPER MILLS, EXCEPT BUILDING PAPER 350 100.0
NEWSPAPERS 8,223 6.3
HttntAOVnCAL PREPARATIONS 666 13.1
PAINTS & ALLIED PRODUCTS 1,379 19.9
INDUSTRIAL INORGANIC CHEMICALS, NEC 582 30.6
PETROLEUM RETINHC 460 53.1
READY-MIXED OONCXETE 5,088 67.5
BLAST FURNACES & SHE. MILLS 521 29.2
GRAY IRON FOUNDRIES 867 44.4
FABRICATED PLATOON 1,742 9.0
SHEET METAL MKK 3,456 7.0
FARM MACHINERY & EXJUHWEHT 1,748 8.5
RADIO & TV aMUNICATICN EQUHMBfT 2,059 5.5
SEMICONDUCTORS & RELATED DEVICES 89 9.3
AIRCRAFT EXjUHHENT, NEC 849 9.7
SHIP BUILDIIC & REPAIRING 616 23.9
GENERAL WAREHOUSING AND STCR/CE 2,284 26.2
MARINE CARGO HANDLING 796 21.9
TELEPHONE CtMJNICATICN 13,124 42.1
REFUSE SYSTEMS 4,251 35.8
Pretax profit per firm
by fins size
UST costs
per
establishment
Smll
Modlun
($1000)-
Large
($1000)
Pretax return an assets
Small.
Before After
Medlun
Before After
Large
Before After
109
241
418
89.6
21.1*
3.8*
4.9*
3.1*
1.7*
1.4*
32
207
—
148.2
6.3*
-22.6*
7.2*
2.0*
—
—
29
89
462
142.4
6.9*
-26.9*
4.1Z
-2.5*
4.8*
3.3*
24
72
—
77.8
9.0*
-19.7* .
5.9*
-0.5*
—
—
21
167
719
110.9
6.1*
-26.4*
10.7*
3.6*
9.0*
7.6*
25
529
1,755
111.8
4.0*
-13.7*
12.3*
9.7*
7.7*
7.2*
—
664
4,365
150.0
—
—
15.1*
11.7*
15.9*
15.4*
—
8
3,011
117.4
—
—
0.22
-3.1*
12.0*
11.5*
(42)
360
1,950
121.6
-6.7*
-26.3*
10.7*
7.1*
8.5*
7.9*
33
426
1,703
130.0
8.7*
-25.2*
9.6*
6.7*
8.6*
8.OX
—
(16)
309
103.2
—
~
-0.4*
-3.1*
1.3*
0.9*
25
463
2,400
102.4
5.1*
-15.4*
10.4*
8.1*
9.5*
9.1*
39
299
177
136.5
7.2*
-18.3*
9.0*
4.9*
0.7*
0.2*
52
430
3,610
109.5
10.1*
-11.3*
13.4*
10.0*
19.6*
19.0*
54
—
—
110.7
12.3*
-13.1*
—
—
—
—
90
353
3,086
170.8
16.4*
-14.9*
10.1*
5.2*
12.2*
11.5*
42
277
2,599
131.3
8.3*
-17.9*
7.3*
3.8*
12.1*
11.5*
75
351
3,163
140.9
14.5*
-12.7*
11.2*
6.7*
12.8*
12.2*
49
310
3,013
81.3
8.7*
-5.6*
8.5*
6.3*
12.6*
12.3*
38
326
1,618
81.6
7.0*
-8.1*
9.1*
6.9*
6.1*
5.8*
—
1,202
3,741
138.2
32.7*
29.0*
17.1*
16.5*
54
241
1,741
110.5
9.1*
-9.5*
7.9*
4.3*
7.8*
7.3*
61
126
1,291
185.7
11.7*
-24.2*
2.8*
-1.3*
5.6*
4.8*
63
148
623
100.0
14.0*
-8.3*
4.0*
1.3*
2.9*
2.5*
20
114
1,267
109.4
3.4*
-14.9*
3.6*
0.1*
6.4*
5.9*
35
275
2,246
66.2
7.6*
-6.7*
9.3*
7.0*
10.5*
10.2*
(9)
121
785
103.1
-1.7*
-20.5*
3.4*
0.5*
3.6*
3.1*
1
288
2,632
196.4
0.2*
-38.9*
7.6*
2.4*
11.9*
11.0*
73
381
3,135
72.0
13.6*
0.2*
10.7*
8.7*
13.3*
12.9*
43
301
—
223.2
7.2*
-30.4*
8.7*
2.2*
—
—
(35)
45
104
821
189
547
458
328
228
(420)
3,265
3,309
150.0
76.6
100.0
95.9
143.8
-7.3*
9.4*
20.9X
188.32
-39.2*
-6.5*
1.6*
155.3*
5.0*
14.9*
9.7*
10.0*
9.5*
1.0*
12.8*
7.6*
7.1*
3.5*
-1.9*
13.2*
13.9*
-2.6*
12.8*
13.5*
-------
3012
5031
5039
5051
5074
5083
5084
5093
5141
5143
5148
5153
5161
5171
5172
5181
5211
5261
5311
5451
5531
7211
7213
7261
7342
7391
7394
7512
7692
7997
Exh 6.27 (Continued)
SCREENING ANALYSIS FCR NON-RETAIL PEIWXEUM TANKS
Regulatory Costs of $50,000/Tank
SIC INDUSTRY GROUP
Uraber of
(all sizes)
Percent
of industry
with tanks
Pretax profit per firm
by flan size
UST costs
per
establishment
Stall
Mediun
($1000)-
Large
($1000)
Pretax return on assets
&salL
Before After
Madiun
Before After
Large
Before After
AUTCKBILES & OTHER MOTOR VEHICLES
5,829
5.2
49
268
3,342
86.0
8.4Z
-6.3*
8.4*
5.7*
16.3*
15.9*
LUMBER, PLYVJOCD & MH1KSK
6,889
9.6
58
303
1,893
91.7
10.61
-6.2*
9.4*
6.6*
9.8*
9.4*
CCNSTRDCTICN MATERIALS, NEC
9,068
18.7
38
261
1,776
92.2
7.4*
-10.5*
8.6*
5.6*
8.7*
8.2*
METALS SERVICE OUTERS i CFFICES
8,754
5.2
35
211
1,138
87.0
6.7*
-9.8*
5.5*
3.3*
5.4*
4.9*
PLUMBING AND HYDRCNIC HEATING SUPPLES
7,955
9.3
35
194
1,054
66.2
7.0*
-6.1*
6.8*
4.5*
7.0*
6.6*
FARM MACHINERY, WHXESALE
IMJUSIRIAL MACHINERY, tHIESALE
SCRAP AND WASTE MATERIALS
GROCERIES, GENERAL
DAISY PRODUCTS
FRESH FRUITS & VEGETABLES, WBXSALE
GRAIN
CHEMICALS, WELESALERS
PETRCLEUM BULK STATIONS & TE5MTNALS
PETRCLELM PRODUCTS, WHOLESALE
BEER & ALE
LUMBER, BUHDIN3 MATERIALS EEALQtS
RETAIL NURSERIES i GARDEN STORES
DEPARTMENT STCRES
DAIRY PRODUCT STCRES
AIJTO & HHE SUPPLY STCRES
POWER LAUNDERIES
LINEN SUPPLY
FUNERAL SERVICES i CREMATORIES
DISINFECTING AND EXTEMflNATDC
RESEARCH & DEVELCSMENT LABCRA1ERIES
EQUHMENT RENTAL I LEAS DC
PASSEH3R CAR RENTAL & LEASING
WELDIN3 REPAIR
MEMBERSHIP SPCRTS & RECREATION CLUBS
13,617
23,413
8,454
3,594
3,645
5,176
8,377
9,380
13,155
4,990
4,483
24,268
7,162
9,767
5,375
39,071
2,122
1,294
14,951
6,575
2,717
15,594
4,568
4,411
11,231
16.6
6.7
8.9
6.2
8.4
5.6
14.8
5.2
43.9
44.1
25.7
12.3
19.4
12.3
10.0
5.0
8.6
29.3
5.0
8.6
5.9
10.4
26.7
8.1
13.8
4
69
1,411
117.1
39
185
689
90.5
53
161
1,824
90.8
40
183
1,401
135.9
(37)
268
3,140
81.7
62
198
1,122
102.9
88
152
2,512
97.6
52
200
1,218
83.0
44
238
194
187.0
64
237
6,029
173.2
42
268
2,265
127.5
42
198
1,405
83.6
25
127
—
87.6
41
125
1,348
90.6
32
535
—
144.5
26
139
2,296
76.6
36
228
—
107.1
62
111
—
126.1
30
66
—
68.6
27
340
—
70.3
35
439
1,773
143.9
48
338
722
122.2
55
158
1,335
106.7
42
408
—
58.1
17
45
—
79.1
0.7*
-21.1*
2.4*
-1.7*
7.2*
6.6*
7.5*
-10.0*
6.0*
3.1*
3.3*
2.9*
10.5*
-7.4*
5.1*
2.2*
9.1*
8.6*
7.2*
-17.2*
5.1*
1.3*
6.8*
6.1*
-6.2*
-20.1*
8.1*
5.7*
12.0*
11.7*
11.7*
-7.7*
7.1*
3.4*
6.5*
5.9*
14.5*
-1.6*
4.9*
1.8*
11.4*
10.9*
10.1*
-«.l*
6.2*
3.7*
5.9*
5.5*
7.5*
-24.3*
7.6*
1.6Z
1.1*
0.0*
11.7*
-20.2*
8.1*
2.2*
30.1*
29.2*
7.5*
-15.3*
8.0*
4.2*
12.0*
11.3*
8.0*
-8.0*
6.6*
3.8*
7.1*
6.7*
5.7*
-14.2*
5.0*
1.6*
—
—
7.6*
-9.3*
3.7*
1.0*
5.7*
5.4*
9.1*
-32.2*
12.3*
9.0*
—
—
6.5*
-12.4*
5.5*
2.5*
11.0*
10.6*
9.5*
-18.6*
9.2*
4.9*
—
—
12.9*
-13.5*
3.1*
-0.4*
—
—
8.6*
-11.1*
2.3*
-0.1*
—
—
9.8*
-15.6*
9.5*
7.5*
—
~
7.3*
-22.3*
13.0*
8.7*
6.8*
6.3*
11.0*
-17.2*
10.5*
6.7*
3.6*
3.0*
10.7*
-10.2*
4.4*
1.4*
5.8*
5.3*
11.1*
-4.2*
12.2*
10.5*
—
—
3.1*
-11.4*
1.7*
-1.3*
—
—
CT>
I
in
VO
-------
6-60
be roughly $5,000 per tank, assuming a baseline tank replacement program already
exists. For purposes of this screening analysis, incremental corrective action
costs are assigned to be an order of magnitude higher than tank replacement, or
$50,000 per tank. However, the probability of having to undergo tank replace-
ment or corrective action varies over regulatory options. In addition, the
expected cost of tank replacement and corrective action varies over regulatory
options.
For each SIC code in each exhibit, the following information is presented:
o Number of establishments in SIC code (includes all establishments
nationwide, not just those with USTs);
o Percentage of establishments with nonretail petroleum USTs;
o Pre-tax profits for three size classes of firm:
— small: less than $1 million in assets
— medium: $1-10 million in assets
— large: $10-50 million in assets;
o Costs per establishment: the number of tanks per establishment times
the assumed regulatory cost per tank;
o Return on assets (before regulatory costs): presented for three size
classes, based on data in Robert Morris statement studies; and
o Return on assets (after regulatory costs): presented for three size
classes after adjusting for regulatory costs.
Determine which Industries Show Potential for Significant Economic Impacts
As explained in Section 6.B, return on assets (R0A) was selected as the
measure to assess viability because expected R0A is reasonably consistent across
industries and size classes. An R0A of less than -30 percent is interpreted as
certain failure, while an R0A between -4 percent and -30 percent is characterized
as severe financial distress.
Note that this screening analysis only examines indicators of potential
impacts. It does not attempt to determine whether the costs of meeting the
regulations are passed on to consumers of the facility, or passed "backwards"
to suppliers of the facility. Either or both of these can occur under some
competitive conditions, resulting in insignificant impacts on profits even if
regulatory costs appear high relative to profits. This is one of the limita-
tions of a screening analysis.
In addition, as stated earlier, many industries would presumably have
alternatives to using nonretail petroleum USTs. This would include purchase
of motor fuels at retail. The existence of this alternative tends to mitigate
the potential for adverse economic effects, although corrective action respon-
sibility for existing releases cannot be avoided by switching from USTs.
-------
6-61
6.C.2. Results and Analysis
Exhibits 6.25, 6.26, and 6.27 provide data on the effect of regulatory costs
of $500, $5,000, and $50,000 per tank on return on assets for small, medium,
and large firms in 4-digit SIC codes where more than 5% of the establishments
maintain USTs for non-retail petroleum storage. Exhibit 6.25 shows that regula-
tory costs of $500 per tank do not affect return on assets to any significant
extent for small, medium, or large firms. Therefore, it seems likely that reg-
ulatory costs of this magnitude will not create the potential for significant
economic dislocation among firms using non-retail petroleum USTs. Similarly,
regulatory costs of $5,000 per tank do not seem likely to create the potential
for significant economic dislocation for the same reasons as stated above-
returns on assets are not significantly affected by costs of this magnitude.
In contrast, Exhibit 6.8 shows that regulatory costs of $50,000 per tank
will move almost all small businesses in the SIC codes examined into the severe
financial distress or certain closure category. However this finding must be
tempered by the following observations:
o Non-retail petroleum USTs are maintained in only a fraction of the
establishments within the industries affected. In addition, it is
likely that larger establishments would be more likely to maintain
USTs than smaller establishments, so the extent to which small
businesses are actually affected is not clear.
o Not all establishments with non-retail petroleum USTs will need to
undertake corrective actions. It can probably be assumed that only
those establishments with existing leaks will incur corrective action
costs. Furthermore, not all of these costs are attributable to to the
regulatory alternatives under consideration, as some corrective action
for existing releases will be performed even in the absence of the
promulgation of federal requirements.
Thus, although regulatory costs of $50,000 per tank are clearly significant, the
likelihood that small firms will have to incur such costs as a result of rules
promulgated by EPA is unclear.
With respect to medium firms, it seems likely that regulatory costs of
$50,000 per tank will potentially weaken a significant number of firms, although
their viability does not seem to be threatened in the short-run. Return on assets
remains positive for most medium firms in Exhibit 6.8, although seven of the 65
SIC codes listed go slightly negative. None of these reach a level which can
be termed severe financial distress, however. As with small firms, questions
of UST incidence among medium firms and probabilities of having to undertake
corrective action are pertinent. Also pertinent is the possibility that cor-
rective action could cost considerably more than $50,000.
Return on assets for large firms seems relatively unaffected by regulatory
costs of $50,000 per tank. However, USTs per establishment were assumed not to
vary over size category, even though it is possible that large firms have more
tanks per establishment than smaller firms. It is also possible that large
firms maintain multiple establishments, thus making them more susceptible to
corrective actions from existing releases. Given that corrective action costs
can significantly exceed $50,000, corrective action could produce significant
economic dislocation even among larger firms.
-------
6-62
6.C.3. Limitations of the Analysis
o The data on compliance costs are based on costs incurred by facilities,
but the financial data describes firms.
The California data are on the number of tanks and number of facilities.
But the financial data by industry are on a corporate basis, i.e., tax filers
or loan applicants. It is highly likely that even the smallest size category
(asset size up to $1 million) includes some firms with more than one facility
or more than one tank site. The larger categories for many industries almost
certainly represent multiple site facilities. Therefore, regulatory costs for
larger firms may be understated.
o SIC codes may not categorize the industries in an optimal manner for
tFTs analysis.
The SIC code of a firm is that representing the largest component of its
net sales, but many firms practice a mix of industrial activities. This can
include a mix in type of products involved, and/or a mix between manufacturing,
wholesaling, retailing, and providing services. As a result, the SIC code that
describes the principal activity of a firm may not be the one describing the
activity for which tanks are used. Thus, regulatory costs are assigned to SIC
codes based on tank use patterns by establishments while financial statistics
are assigned to SIC codes by the principal activity of fi rms. Such a limitation
is not thought to create a consistent bias, however.
o The number of tanks per firm may vary with firm size.
In calculating the multiple tank costs, we have assumed that for all asset
size ranges, each firm owns the industry average number of tanks. For this
reason, the regulatory costs relative to profits are largest for the smallest
firms. If in fact the smaller firms have fewer tanks, which is likely to be
the case in some industries, the regulatory burden on those firms may be rela-
tively less. Thus, the screening study may overstate impacts on relatively
small businesses in each industry category.
o Industrial patterns in California may not be typical of the United
States.
California data are used to determine which industries own tanks, the
percent of firms in each industry that own tanks, and the number of tanks per
firm. Because of age of firms; locations relative to suppliers, pipelines, or
transportation; the mix of industries, and other factors, these data may not be
typical of other regions of the country. For example some industries that are
(are not) prevalent in California or that operate with (without) underground
tanks may be different in other states. No other data were available to test
the degree to which California's use of underground tanks is typical. Conse-
quently, the screening analysis is a rough indicator or potentially affected
industri es.
-------
6-63
o Existing releases may vary by firm size.
The likelihood of being liabile for corrective action costs is a function
of whether or not USTs are currently leaking. Because baseline detection and
replacement policies may differ by firm size, it is not clear whether the like-
lihood of an existing release is independent of firm size.
o Because of the different sources used, results should be viewed as
approximate.
The percent of firms in the industry with tanks is a ratio of numbers
derived from a recent California tank registry and from the 1982 County Business
Patterns. Financial statistics are from a source compiled from 1985 data. It
is possible that different economic conditions in the different periods produce
results different than synchronous observations; and that firms and facilities
are classified into SIC codes using different criteria among the sources.
Nevertheless, the ranking and general magnitude of results should not be
significantly affected.
o A^ screening measure does not estimate impacts.
This screening measure is designed to identify those industries for which
the compliance costs are large enough relative to corporate profits that the
economic impacts could be significant if they were completely absorbed out of
profits. Some impacted firms may be able to pass the costs on—in part or in
total—to their customers or to their suppliers. In this case, the few indus-
tries identified in this analysis as showing potential for significant economic
impacts may in fact experience insignificant impacts. The likelihood for this
to occur depends on the degree of competition among firms in the affected
industries. The extent to which firms will incur the larger costs of tank
replacement and corrective action vary with each regulatory alternative.
Modelling these alternative-specific impacts for the large number of SIC codes
potentially affected was beyond the scope of this analysis.
6.C.4. Sources of Information
The following sources of information were used in writing this section
of the RIA:
o SIC Industries with large numbers of USTs:
Data Resources, Inc. and Quantum Analytics, Inc., for U.S.-Environmental
Protection Agency; Draft Report: Underground Storage Tanks Technical Data
Collection; October 24, 1985. October 24, 1985, Appendix D.
California data includes SIC code, number of facilities with tanks, and
number of tanks. SCI calculated the number of tanks per facility with tanks
as the ratio of the latter two.
o Number of faci1ities in California within each SIC industry group:
U.S. Department of Commerce, Bureau of the Census; County Business Pat-
terns, 1982: California; CBP-82-6. pp. 1-16.
-------
6-64
SCI calculated the percentage of facilities in each SIC code with tanks by
dividing the number of facilities with tanks by the number of facilities in the
state.
o Financial Data by SIC Code:
Robert Morris Associates; '85 Annual Statement Studies; Philadelphia, PA;
September 1985.
For samples of varying sizes, provides (among other data) total assets,
net sales, and the ratios of various expense and profit categories to net sales.
SCI calculated absolute values per firm in the sample from the sample aggregates
and from the ratios.
-------
6-65
6.D. IMPACTS ON FIRMS USING USTs FOR CHEMICALS STORAGE
6.0.1. Methodology
The methodology used for the chemical UST screening analysis is identical to
that used for the nonretail petroleum UST screening study. The ten four-digit
SIC industries identified as owning the greatest number of chemical tanks are:
2819 Industrial inorganic chemicals, NEC
2821 Plastics materials and resins
2851 Paints and allied products
2869 Industrial organic chemicals, NEC
2899 Chemical preparations, NEC
3471 Plating and polishing
5161 Chemical Wholesalers
5171 Petroleum Bulk Stations and Terminals
5172 Petroleum Product Whole Salers
7216 Drycleaning Plants
The estimation of the fraction of each industry owning tanks reveals that
even for those industries in which chemical tanks are most prevalent, they are
far from universal. The precent of the industry owning USTs for chemicals
storage ranges from 40 percent (paint and allied products) to one percent
(drycleaning). Another difference in ownership patterns between chemical and
petroleum tanks is that in some industries the average number of tanks (in
facilities that actually have tanks) is much higher. For chemical tanks, the
value ranges from thirteen tanks per facility (large paint facilities) to one
(petroleum product wholesaler). Substitutes for underground chemical tanks
include above-ground tanks, drums (especially for small businesses) and process
changes. As one example of a process change, a switch from oil-based paint to
water-based paint eliminates the need for solvents, which are the chemicals
generally stored in underground tanks at paint producing facilities. As another
example, underground chemical tanks at petroleum bulk stations generally include
additives for gasoline. These could be blended in at an earlier point in the
distribution chain. The existence of such substitutes tends to mitigate the
potential economic effects of regulations for underground chemical tanks.
However, corrective action responsibilities for existing tanks cannot be avoided
by switching to substitutes, so corrective action requirements pose some poten-
tial for economic dislocation.
6.D.2. Results and Analysis
The analysis for chemical tanks was performed in the same manner as the
analysis for petroleum tanks, but using the ten industries identified in the
data base as owning the largest number of underground chemical tanks.
Three exhibits are provided:
o Exhibit 6.28 at $500/tank (approximates detection);
o Exhibit 6.29 at $5,000/tank (approximates replacement); and
o Exhibit 6.30 at $50,000/tank (approximates corrective action).
-------
Exhibit 6.28
SCREENING ANALYSIS FOR UNDERGROUND CHEMICAL TANKS
Regulatory Costs of $500/Tank
Hunter of
Percent of
Pretax profit per film:
UST costs
Pretax return on assets:
establishments
firms with
per
Small
Medlun
Large
SIC INDUSTRY CROUP
(all sizes)
olwnl/*a1 farifa
Small
Medlun
Large
establishment
/fri rmm™
Before
After
Before
After
Before After
(a)
(b)
(b)
iCUUUUJ™
2819
INDUSTRIAL INORGANIC CHEMICALS, NEE
622
16
38
326
1,618
1.3
7.OX
6.8X
9. IX
9.IX
6.IX
6.IX
2821
PLASTIC MATERIALS & SYNTHETIC RESINS
518
31
131
312
2,160
1.6
22.3X
22.OX
9.IX
9.IX
11.8X
11.8X
2851
PAINTS & ALLIED PRODUCTS
1,379
40
49
310
3,103
2.8
8.7X
8.2X
8.5X
8.4X
13. OX
13. OX
2869
INDUSTRIAL ORGANIC CHEMICALS, NBC
582
20
38
326
1,618
1.8
7.OX
6.7X
9. IX
9.IX
6.IX
6. IX
2899
CHEMICAL PREPARATIONS, NBC
1,309
15
38
326
1,618
1.9
7.OX
6.7X
9. IX
9.IX
6.IX
6.IX
3471
PAH1TDC & POLISHING
3,156
7
53
381
--
2.1
11.2X
10.7X
13.5X
13.5X
—
—
5161
CHEMICALS, UB3LESALERS
9,380
3
52
200
1,218
5.1
10.2X
9.2X
6.2X
6.IX
5.9X
5.9X
5171
PETROLEUM BULK STATIONS & HHGNALS
13,155
9
44
238
1,937
1.9
7.5X
7.IX
7.6X
7.5X
8.IX
8.IX
5172
PETROLEUM PRODUCTS, WHOLESALE
4,990
8
64
237
6,029
0.7
11.8Z
11.7X
8.2X
8.IX
30.IX
30. IX
7216
ERWXEANING, EXCEPT RUGS
18,293
1
41
228
—
1.0
10.8X
10.5X
9.2X
9.2X
—
(a) Estimate of all establishments In each SIC code (not Just those with chemical tanks).
SOURCE: County Business Patterns in the US, 1982
(b) SOURCE: Screening Study for Regulatory Inpacts of UST Regulations, Appendix, March 31, 1986.
-------
Exhibit 6.29
SCREENING ANALYSIS FCR UNDERGROUND CHEMICAL TANKS
Regulatory Costa of $5000/Tank
Nuriber of
Percent of
UST costs
establishments
flints with
Pretax profit per firm
per
Pretax return an assets
SIC INDUSTRY GROUP
(all sizes)
chemical tanks
by firm size (b)
establishment
Small
NddluD
ULTgfi
(a)
(b)
Snail
Med Lira
Large
Before
After
Before
After
Before After
./fil Aftn\__
. a ^ /WWW
" ^^lUUU)
($1000)
2819
INDUSTRIAL INORGANIC CHEMICALS, NEC
622
16
38
326
1,618
13.0
7.OX
4.6X
9.IX
8.8X
6.IX
6.IX
2821
PLASTIC MATERIALS & SYNTHETIC RESINS
518
31
131
312
2,160
16.0
22.3Z
19.6X
9. IX
8.7X *
11.8X
11.7*
2851
PAINTS & A" Tin PRODUCTS
1,379
40
49
310
3,103
27.5
8.7X
3.8X
8.5X
7.7X
13. OX
12.9X
2869
INDUSTRIAL ORGANIC CHEMICALS, NEC
582
20
38
326
1,618
17.5
7.OX
3.8X
9. IX
8.7X
6.IX
6.IX
2899
CHEMICAL BOTARATICNS, NBC
1,309
15
38
326'
1,618
18.5
7.OX
3.6X
9.IX
8.6X
6.IX
6.IX
3471
painthc & paLiamc
3,156
7
53
381
--
20.5
11.2X
6.9X
13.5X
12.8X
—
-
5161
CHEMICALS, UBCLESALSiS
9,380
3
52
200
1,218
50.5
10.2X
0.3X
6.2X
4.7X
5.9X
5.7X
5171
PETROLEUM BULK STATICHS & TERMINALS
13,155
9
44
238
1,937
18.7
7.5X
4.3X
7.6X
7.OX
8.IX
8.OX
5172
PEISOiUM raODOCTS, WELESALE
4,990
8
64
237
6,029
7.0
11.8X
10.5X
8.2X
7.9X
30.IX
30. IX
7216
CRYOEANI1C, EXCHTKJGS
18,293
1
41
228
10.0
10.8X
8.2X
9.2X
8.8X
—
—
(a) Estimate of all establishments In each SIC code (not Just those with chemical tanks).
SOURCE: County Business Patterns in the US, 1982
(b) SCXIRCE: Screening Study for Regulatory Inpacts of UST Regulations, Appendix, March 31, 1986.
-------
2819
2821
2851
2869
2899
3471
5161
5171
5172
7216
Exhibit 6.30
SfREEUItr, ANALYSIS ECU UNDERGROUND CHEMICAL TANKS
Regulatory Costs of $50,000/Tank
SIC INDUSTRY GROUP
Nunber of
establishments
(all sizes)
(a)
Percent of
fizna with
chfmlral tanks
(b)
Pretax profit per fLxm
by flnn size (b)
UST costs
per
establishment
Small Mediiin Large
($1000)
($1000)
Pretax return on assets
&nll
Before After
Medlizn
Before After
Large
Before After
INDUSTRIAL HCRGANIC CHEMICALS, NEC
622
16
38
326
1,618
130.0
7.0*
-17.0*
9.1*
5.5*
6.1*
5.6*
PLASTIC MATERIALS 6, SYNTHETIC RESINS
518
31
131
312
2,160
160.0
22.3*
-4.9*
9.1*
4.4*
11.8*
10.9*
FAINTS & ALLIED FRCDUCTS
1,379
40
49
310
3,103
275.0
8.7*
-39.9*
8.5*
1.0*
13.0*
11.8*
INDUSTRIAL ORGANIC CHEMICALS, NEC
582
20
38
326
1,618
175.0
7.0*
-25.3*
9.1*
4.2*
6.1*
5.5*
CHEMICAL PREPARATIONS, NEC
1,309
15
38
326
1,618
185.0
7.0*
-27.2*
9.1*
4.0*
6.1*
5.4*
PAINTDG & PCLISHHC
3,156
7
53
381
—
205.0
11.2*
-32.1*
13.5*
6.3*
-
—
CHEMICALS, WHOLESALERS
9,380
3
52
200
1,218
505.0
10.2*
-88.9*
6.2*
-9.5*
5.9*
3.5*
PETROLEUM BULK STATIONS & THWINALS
13,155
9
44
238
1,937
187.0
7.5*
-24.3*
7.6*
1.6*
8.1*
7.3*
PETRCLEIM PRODUCTS, UBOLESALE
4,990
8
64
237
6,029
70.0
11.8*
-1.1*
8.2*
5.8*
30.1*
29.7*
ervcleanik:, ehht bugs
18,293
1
41
228
100.0
10.8*
-15.5*
9.2*
5.2*
—
—
-------
6-69
Our principal conclusion from this analysis is similar to the previously
presented screening analysis of non-retail petroleum USTs. Regulatory costs of
$500 per tank do not significantly affect return on assets for any size category
in any of the ten SIC codes analyzed. Regulatory costs of $5,000 per tank do
not appear to bring any firms to the point of severe financial distress or
closure, but do weaken smaller firms somewhat. Regulatory costs of $50,000 per
tank will cause many small firms to close, will weaken medium firms, and will
only marginally affect larger firms. There are questions still to be addressed
regarding the likelihood of firms of each size class incurring costs of this
magnitude.
6.D.3. Limitations of the Analysis
The limitations of the screening analysis for chemical tanks are the same
as those for petroleum product tanks, described above.
6.D.4. Sources of Information
The sources of information for the screening analysis for chemical tanks
are the same as those for petroleum product tanks, described above.
-------
Chapter 7
BENEFITS OF UST TECHNICAL STANDARDS AND REGULATIONS
7.A. INTRODUCTION
In Chapter 5, the costs of requirements for prevention and detection and
costs for corrective action under each UST regulatory option were examined.
Chapter 5 also looked at the effectiveness of each option in terms of the
trade-offs between compliance costs and aggregate plume acreage of releases.
In addition to cost-effectiveness, Executive Order 12291 directs agencies to
quantify and monetize benefits to the extent possible and examine the trade-offs
between benefits and costs for the various options considered in the RIA. The
objective of this chapter is to develop estimates of aggregate benefits resulting
from UST requirements under each option when compared to the base-case scenario.
To measure benefits, the EPA Guidelines for Performing Regulatory Impact
Analyses recommends that one examine the following chain of events: (1) the
release of pollutants (in this case from underground storage tanks); (2) the
impact of these releases on ambient environmental quality; and (3) exposure of
people, plants, animals, and materials through various media (air, water,
etc.). The analyst should strive to consider the entire spectrum of impacts
and attempt to quantify those impacts that can be assigned a dollar value as
well as those which can only be described qualitatively. The EPA Guidelines
also recommends that in addition to most likely estimates, upper and lower
confidence limits be presented.
The next section discusses the methodology used to estimate benefits.
Considerable attention in the methodology section is focused on the tricky
problem of dealing with corrective action costs in benefit/cost analysis for
this rule. A rationale is provided for casting the basic trade-off costs of
prevention and detection versus corrective action costs avoided and residual
damages. Subsequent sections of this chapter examine health and safety benefits
(7.C), avoided property damages and foregone profits (7.D), environmental
effects defined broadly (7.E) and option and existence values (7.F).
7.B. METHODOLOGY
Although the chapter focuses on the costs of UST releases, the analysis is
termed "benefits analysis" for the following reason: each regulatory option, as
discussed in Chapter 5, results in significant reductions in aggregate plume
acreage in comparison to the base-case scenario. Provided social costs (i.e.,
damages) positively correlate with plume acreage, each option avoids costs (in
comparison to the base case) and thus conveys benefits. These incremental
benefits are calculated as follows:
PVBi = PVDb - PVDi (1)
where PVBi = present value of benefits for option "i"
PVDb = present value of social costs for the
base case
PVDi = present value of social costs for option "i"
-------
7-2
Of particular concern in the benefits analysis is the treatment of
corrective action costs. There are two ways in which corrective action costs
can be represented in benefit/cost analysis. First, corrective action costs
can be viewed as a program compliance cost and entered on the cost side of the
ledger. Alternatively, corrective action costs can be viewed as a component
of the social costs associated with LIST releases. When corrective action costs
are added to damages, they are considered in the benefits analysis as social
costs avoided.
Is one approach preferred over the other? Assuming that corrective action
costs and environmental damages are calculated so that there is no double
counting (discussed below), both approaches yield the same relative rankings
of the alternatives. However, the magnitude of incremental benefits will be
lower if corrective action costs are included as a cost term.
There is another argument that can be made for incorporating corrective
action costs into the benefits analysis. The magnitude of corrective action
costs is both a function of other actions taken under the UST rule (new tank
standards, monitoring and detection requirements) and a policy variable. As
more stringent prevention and detection measures are required, the number and
size of releases decreases and, other things being equal, aggregate corrective
action costs will also decrease. However, for a given set of preventive mea-
sures, corrective action costs will vary as the level of cleanup required by
the regulatory agency varies; more stringent cleanup standards imply greater
corrective action costs.
Thus, as pointed out in Chapter 2, there is an interrelationship between
prevention and detection policies on the one hand, and corrective action policies
on the other hand. Improved prevention and detection reduces contamination
incidents and therefore reduces corrective action costs. Therefore, in order
to analyze prevention and detection policies against each other, it is necessary
to hold corrective action policy constant. Similarly, in order to analyze
alternative corrective action policies, it is necessary to hold prevention and
detection policy constant. In this analysis, corrective action policy is held
constant across all regulatory options, thus facilitating the comparison of
prevention and detection options. In order to choose the most cost-effective
policy for prevention and detection, it is necessary that the base case also
include the same corrective action policy as the regulatory options. To do
otherwise would distort the capability of the analysis to present meaningful
data for evaluating trade-offs. The rationale for using the corrective action
policy chosen is presented in the Preamble to the proposed rule. Thus, one way
to view this issue is that the costs of increased prevention and detection are
traded off against the reduced costs of corrective action. Also included on
the benefits side of the ledger are the residual damages which are not remedied
by corrective action (i.e., damages which occur before the corrective action is
initiated).
As mentioned earlier, there is some concern about double counting correc-
tive action costs and damages. In benefit/cost analysis, the measure of benefits
should reflect social values and not private values, unless they are the same.
Once a release occurs, the social objective should be to select that level of
-------
7-3
corrective action which theoretically minimizes the sum of damages and correc-
tive action costs. Thus, corrective action is taken if the benefits (future
health, property and environmental damages avoided) exceed the costs of the
corrective action. Unfortunately, this theoretical trade-off is quite diffi-
cult to make in the real world.
In practice, corrective action decisions do not always reflect consider-
ation of these trade-offs. Corrective actions which sometimes appear excessive
may be explained by several factors, such as local or state requirements that
releases be cleaned up to background levels, policies that polluters should
clean up releases regardless of damages, uncertainty about the effectiveness
of corrective action, or imperfect information on the potential for future
damages. In the latter case, if there is a small probability of significant
damages, the local or state agency is likely to err on the side of more rather
than less corrective action.
In some cases, however, too little corrective action may be taken because
owners lack financial assets to pay for corrective action or local or state
agencies lack the resources to effectively coordinate corrective actions at all
sites. On net, it will be assumed that the sum of corrective action costs and
monetized damages does not involve significant double counting.
Corrective action costs have already been discussed and estimated in
Chapter 5. The remainder of this chapter is devoted to the description of four
broad categories of damages associated with UST releases:
o Section 7.C discusses health and safety risks;
o Section 7.D discusses property damages and foregone profits;
o Section 7.E discusses environmental effects; and
o Section 7.F discusses option and existence value.
For only one category, property damages and reduced firm profits, are aggregate
damages expressed in monetary terms. For health and safety effects, aggregate
incidents are estimated but not monetized. The analysis in Section 7.E suggests
that there is significant potential for environmental effects but these effects
are very difficult to quantify. The final category, reductions in option and
existence values, can be described but quantification would involve substan-
tially more analysis. It is assumed that society's willingness to clean up
releases, even when there is no obvious justification in terms of quantifiable
damages avoided, is partly a reflection of our desire to protect intrinsic
values, such as option and existence value.
7.C. HEALTH AND SAFETY RISKS
Protecting human health is one of the primary goals of RCRA in general and
of the UST regulations in particular. Releases from USTs threaten human health
in two ways: (1) they represent a safety hazard due to the potential for fire
and explosion, and (2) exposure to contaminants in water supplies or the atmos-
phere poses risks of chronic health effects. For the benefits analysis, we
-------
7-4
have focused on chronic health effects rather than the safety hazard. Although
the safety issue is not a trivial one (the release incident survey documented
141 fire and explosion incidents out of a total of 10,000 release incidents^/),
it appears that there is a serious threat only in certain circumstances, (e.g.,
where floating plumes seep into basements excavated below the water table).
Subsequent analysis may need to explore safety risks in more detail.
7.C.I. Approach and Assumptions
We analyzed human health risks from leaking USTs using many of the same
outputs from the UST model that support other parts of this RIA. Our approach
is explained in detail in Appendix F; a very brief summary follows.
Our analysis estimates the incremental cancer risks resulting from exposure
to the benzene component of gasoline released from a population of USTs. Our
methodology is intended to provide a logical basis for determining the extent
of risk in the base case and under the regulatory options. Although we believe
the methodology provides a solid analytical framework, we caution that our
results should be regarded as preliminary, and invite comments on how to improve
the approach and underlying data base.
We limited the scope of our analysis to risks from ingesting ground water
contaminated by gasoline released by USTs. We used the leak rates, floating
plume sizes, and leak durations output by the LIST model, and the same assump-
tions used elsewhere in this RIA regarding the proportion of existing tanks
that are bare steel and fiberglass. We simulated risks under two sets of
assumptions: in the first, exposure stops as soon as either the leak is
detected or the pollutant concentrations in ground water exceed the taste
threshold; in the second, leak detection stops exposure, but concentrations
above the taste threshold do not. Based on a screening analysis, we assumed
that most of the health risk posed by gasoline-contaminated water is associated
with the benzene fraction of gasoline, and thus limited our analysis to risks
from benzene exposures. We modeled concentrations of benzene in ground water
over a thirty-year period, estimated lifetime exposures, and used EPA's estimate
of the carcinogenic potency of benzene to predict the upper-bound cancer risk for
different exposure scenarios.
We developed 9,720 different exposure scenarios for each regulatory option.
These comprise combinations of three tank designs (existing bare steel, existing
fiberglass, and replacement tank design), four vadose (unsaturated) zone types,
nine floating plume sizes, three ground-water velocities, and 30 exposure wells.
Each of these factors has an effect on estimated risk:
o Tank design affects the frequency with which leaks develop, and the
leak rate once the tank has failed.
o Vadose zone type influences the dimensions of the floating gasoline
plume for a given leak.
W EPA, Office of Solid Waste, Analysis of the National Data Base of UST
Release Incidents, January 31, 1986
-------
7-5
o Floating plume size affects the transfer of benzene from the floating
plume to the aqueous (ground-water) phase and the degree to which the
contaminant is concentrated along the center line of the dispersed
piume.
o Ground-water velocity affects the transit time for benzene to reach
the wells and its concentration in the water reaching the wells. All
of the aquifers modeled are water table (unconfined) aquifers; confined
aquifers would have lower concentrations of benzene.
o Well locations relate to the intensity and duration of exposure to
benzene in the drinking water.
The replacement tank design and leak detection methods vary among each of
the regulatory scenarios, in turn- affecting the frequency of failure and leak
rates. We combined all of the information elements above for the base case and
each of the regulatory scenarios to predict cancer risk for individuals drinking
from wells. In addition, we estimated the number of people potentially exposed
at each well, and were thus able to estimate population risk (i.e., the total
number of statistically expected cases of cancer across the entire potentially
exposed population).
Finally, we estimated or simulated frequency distributions for each of
the factors that, when combined, define the scenarios. Tank design and popu-
lation at exposure wells are independent distributions. We estimated a joint
frequency distribution for tanks within each of the twelve vadose zone and
ground-water velocity combinations, and simulated the joint distribution for
tank design/vadose zone/floating plume size. We calculated an estimated fre-
quency for each exposure scenario, and used these to generate a weighted
distribution of individual risk for each regulatory option.
7.C.2. Results
Exhibit 7.1 shows the frequency of tanks in different risk intervals for
the base case and the regulatory options under the assumption that exposure
does not stop when pollutant concentrations exceed the taste threshold. Our
key findings are:
o If exposure stops immediately after detection of a leak, but is not
limited by the taste threshold, base case risks range as high as one
in 1,000. About 20% of the tanks have risks greater than 10~6. EPA
often uses a lifetime cancer risk of one in a million (10"®) as a
threshold for regulatory concern.
o If exposure ends immediately upon detection of a leak or when concen-
trations exceed the taste threshold for benzene, upper-bound lifetime
cancer risks from drinking ground water downgradient of leaking USTs
range from about one in 100,000 down to zero in the base case. About
seven (7) percent of tanks have risks to the most exposed individual
(MEI) greater than 10~®.
-------
>,
u
c
V
D
CX
(J
u
L.
1/ /I b«)«;c
- 7
opt 1
Exh " : 7.1
Frequency of ME I Risk
Base Case vs. all Options
- 6
i
Upper Bound of Risk Inlnvul
>l>l 7
- 3
opt b
-------
7-7
o Of the environmental factors we evaluated (vadose zone type, ground-
water velocity, well distance, and time of travel [TOT] from the tank
to the well), TOT has the greatest influence on risk. All of the
scenarios where predicted risks exceed 10"® have TOTs of three years
or less.
o A11 of the regulatory options reduce risks. When we assume that taste
does limit exposure, about 7% of tanks have risks exceeding 10"® in
the base case, and the percentage is reduced to 1.6% by Option I,
2.9% by Option II, 1.7% by Option III, zero by Option IV, and 1.9%
by Option V. When taste does not limit exposure, in the base case
20% of tanks have risks greater than 10~6. The percentage is reduced
to 5% in Option I, 8% in Option II, 6% in Option III, 0.3% in Option IV,
and 3% in Option V.
o Option IV is the most effective, and virtually eliminates risks above
10~6. The reason this option performs so well is that in our simulation
we applied the most stringent regulatory requirements to exposure
scenarios with the shortest TOT from tank to well. All of the high
risk scenarios have short TOTs.
o There is little difference in performance between Options I, II, III,
and V (see Exhibits 7.1 and 7.2).
7.D. Property Damage Benefits
7.D.I. Property Damages Avoided as a Benefits Measure
Damages to property constitute a major detrimental effect of UST leaks.
To the extent that regulatory options reduce the number and severity of UST
leaks, property damages are avoided. These desirable outcomes are consequently
benefits of the regulatory options.
For convenience, three types of "property" are considered in the context
of UST leaks: on-site, off-site business, and residential. On-site damages
are those that would theoretically be reflected in the value of the tank owner's
property. Off-site business damages would theoretically affect the value of
an affected business property. Residential damages are also off-site, and
affect residential property values. It is worth noting that if a leak damages
residential rental property, it is best thought of as an off-site business
damage, at least in the long-run. In the short-run, the renter may incur certain
costs. In the long-run though, competitive pressures should shift the burden
of these costs to the property owner in the form of reduced property values.
There may be other artifacts of the classification scheme selected for this
analysis, but it should serve its purpose well.
Property damages caused by an UST leak typically occur when a contaminated
groundwater plume contacts a well or structural foundation (usually a basement
or sump system). Contact with a well effectively renders the well water unsuit-
able for most uses. Contact with a foundation may result in unpleasant or
dangerous vapors in a home or other building.
-------
Exh 7.2
u
c
«
u
rr
u
u
Ll
Frequency of MEI Risk (with T/0 Cutoff)
Base Case vs. nil Options
2^
X
&
v
A
8
/
/
/
/
z-±
i
00
- 7
4
- 3
1//1 buw
S3 opt 1
Upper Bound of Risk Interval
V//A •>(>< ?
KXI " 4
i>|i( !S
-------
7-9
One alternative for measuring the benefits of avoiding damages of this
type is to use the replacement cost method. The replacement cost method deter-
mines the value of benefits by examining the avoided costs of having to replace
the services once provided by a damaged resource. In the present context, a
well constitutes a private resource owned by a business or household. The well
provides water for drinking, bathing, washing, cooking, etc. A building pro-
vides shelter for households or businesses.
Well and building services can be quickly and severely interrupted by a
contaminated ground-water plume. Fortunately, these services can usually be
replaced. However, replacement can be costly. By estimating the costs of
replacing the services of affected wells and buildings, we can estimate the
benefits of avoiding well and vapor damages of the type caused by UST leaks.
The replacement cost method is an imperfect technique for valuing benefits
of avoiding well and vapor contamination. It may be unable to estimate certain
of the costs imposed by damages. Consider the case of a contaminated well on
the property of a homeowner. "Replacing" the well services may involve the
purchase of bottled water for a short period of time, followed by connection
of the home to a municipal water supply. Estimating the costs of these two
items may seem straightforward. However, it is difficult to estimate the
inconvenience cost of having to identify the solution and implement it. In
the short-run, the household may incur the cost associated with having to use
less water. Once connected to a municipal line, the quality of that water may
be higher or lower than the quality of the water from the well before it was
connected. Consequently, replacement cost estimates should be interpreted
with caution. We cannot even say with certainty that benefits so estimated
will be systematically under- or over-stated.
7.D.2. Underlying Assumptions and Requirements in the Aggregation Process
This benefits analysis involves combining replacement cost estimates
with a damage function that relates UST leaks to well and vapor damages. The
point of departure in the damage function approach is an estimate of the number
and size of contamination plumes that would occur in the base case and under
each of five regulatory options. An output of the EPA UST model, these esti-
mates are provided for five-year increments during a thirty year period. There
are nine plume size categories. Plume sizes range from 1 square meter to
10,000 square meters.
Exhibit 7.3 reports the set_of plume size coefficients used in this anal-
ysis. The estimates are on a per-tank basis for the entire thirty-year
period. For example, Exhibit 7.3 indicates that, in the base case, there would
be 0.18 1000-square-meter plumes per tank over the entire thirty-year period.
We do not know how these 1000-square-meter plumes would be distributed over
time. We do have estimates, however, of how all plumes would be distributed
over time. These coefficients are reported in Exhibit 7.4. Exhibit 7.4 shows,
for example, that in the base case there would be 0.45 plumes per tank during
the first five years, 0.25 plumes per tank during the second five years, etc.
-------
7-10
Exhibit 7.3
Coefficients of Plume Size
Plume size
1
10
25
100
500
1000
2000
5000
10000
Base case
0.1
0.17
0.12
0.29
0.39
0.18
0.14
0.02
0.01
Qoticn 1
0.23
0.28
0.24
0.24
0.24
0.09
0.03
0.01
0
Option 2
0.21
0.21
0.16
0.18
0.14
0.06
0.03
0.01
0
Cotion 3
0.19
0.26
0.24
0.21
0.23
0.09
0.03
0.01
0
Option 4
0.18
0.23
0.16
0.16
0.16
0.06
0.02
0
0
Option 5
0.01
0.19
0.09
0.08
0.08
0.03
0.01
0
0
Exhibit 7.4
Number of Plumes Per Tank
33B333S3S332
II
II
II
II
II
II
II
H
SS33S3SSS
II
II
II
II
II
II
II
II
===_====_
1-5
6-10
11-15
16-20
21-25
26-30
years
years
years
years
years
years
Base case
0.45
0.25
0.20
0.19
0.17
0.16
Option 1
0.40
0.15
O.IO
0.30
0.21
0.20
Option 2
0.45
0.25
0.20
0.05
0.03
0.02
Option 3
0.45
0.20
0.20
0.17
0.13
O.ll
Option 4
0.40
0.15
0.15
0.12
0.10
0.05
Option 5
0.40
0.09
0.00
0.00
0.00
0.00
-------
7-11
Examination of Exhibit 7.3 reveals that the different options have differ-
ent effects on the size of plumes over time.-- Under Option II, the percentage
of leaks that are of the three smallest sizes, 1, 10, and 25 square meters,
increases. At the same time, the percentage of larger leaks falls substantially.
Since the coefficients in Exhibits 7.3 and 7.4 are on a per-tank basis,
the absolute number of plumes can be estimated by multiplying the coefficients
by the number of tanks that are expected to be in service during the thirty-year
period: 1.4 million. Exhibits 7.5 and 7.6 report these results. Exhibit 7.5
indicates, for example, that under Option II there would be a total of 224,000
25-square-meter plumes over thirty years. Exhibit 7.6 reports that under Op-
tion II, there would be 280,000 plumes of all sizes during years 11 through 15.
Estimates of the numbers of plumes of different sizes that would occur
during each five-year increment can be estimated by assuming that the distribu-
tion of all plumes over time would hold for individual plume sizes. Exhibits
7.7 through 7.12 report these results. Consider the results in Exhibit 7.7,
which show projected plumes of different sizes over time in the base case.
This table shows that there would be 79,859 plumes sized 1,000 square meters
during years 1-5, 44,366 plumes of this size during years 6-10, etc. In
total, there would be 252,000 such plumes over the thirty-year period in the base
case. Note that this is the number of 1,000-square-meter plumes projected in
the base case in Exhibit 7.5.
The assumption that the distribution over time would be constant for all
plume sizes is questionable. It seems likely that, at least under some options,
there would be a tendency for large plumes that do occur to occur sooner, due
to detection requirements. Consequently, the benefits estimates that are derive
from this assumption may be biased.
Plume shape is another component of the damage function. Due to ground-
water flow, plumes tend to be irregular in shape — perhaps elliptical or
triangular — rather than round. While plumes are probably not typically
normal ellipses or triangles, we have, for convenience, assumed that they are.
This makes it relatively simple to compute the distance from the tank to the
leading edge of the plume.
Exhibit 7.13 shows the nine plume size categories measured three ways:
square meters (the way the data are provided), distance in feet from tank to
leading edge, and acres. It is worth noting that as plume area increases,
distance from tank to the leading edge of the plume increases far less than
proportionally. For example, a 2,000-square-meter plume is twice as large as
a 1,000-square-meter plume. However, the distance from the tank to the leading
edge of the plume is only 1.4 times greater for the larger plume.
The choice of a triangular versus elliptical proxy for plume shape is
important in that the two different shapes yield very different areas per
plume. For instance, take the example of well contamination. Exhibit 7.13A
shows the farthest distance given from tank to a well for the well cases in
our case studies. As is clear from the table, the elliptical proxy yields an
area more than three times as large as the triangular one does. If the real-
world situation is that most plumes are triangular, then the elliptical model
is an inaccurate one, and vice versa.
-------
7-12
Exhibit 7.5
Number of Plumes by Size, Thirty-Year Time Period
?lume size
1
10
25
100
500
1000
2000
5000
10000
Total
Base case
140,000
238,000
166,000
406,000
546,000
252,000
196,000
28,000
14,000
1,988,000
Option 1
322,000
392,000
336,000
336,000
336,000
126,000
42,000
14,000
0
1,904,000
Option 2
294,000
294,000
224,000
252,000
196,000
84,000
42,000
14,000
0
1,400,000
Option 3
266,000
364,000
336,000
294,000
322,000
126,000
42,000
14,000
0
1,764,000
Option 4
252,000
322,000
224,000
224,000
224,000
84,000
28,000
0
0
1,358,000
Option 5
14,000
266,000
126,000
112,000
112,000
42,000
14,000
0
0
686,000
Exhibit 7.6
Total
Number of
Plumes,
by Five-
Year Increments
sszaaesaa:
:aaaaaaca
SSCBSSSSSSSSBSSSS8
==========
===========
1-5
6-10
11-15
16-20
21-25
26-30
years
years
years
years
years
years
Total
Base case
630,000
350,OOO
280,000
266,000
238,000
224,000
1,988,000
Option
1
560,000
210,000
140,000
420,000
294.000
280,000
1,904.000
Option
2
630,000
350,OOO
280.000
70,000
42,000
28,000
1,400,000
Option
3
630.000
280,OOO
280,OOO
238.000
182.000
154,000
1,764,000
Option
4
560,000
210,000
210,000
168,000
140,000
70,000
1,358,000
Option
5
560,000
126,000
0
0
0
0
686,000
-------
7-13
Exhibit 7.7
Number of
PIumes
, by Size —
Base Case
iosss = s = sa
asssssesaa
„n„s„ =
esn=3BB3
Plume
1-5
6-10
11-15
16-20
21-25
26-30
size
years
years
years
years
years
years
1
44,366
24.648
19.718
18.732
16,761
15.775
10
75.423
41.901
33.521
31.845
28.493
26.817
25
53,239
29,577
23.662
22.479
20.113
18.930
100
128.662
71,479
57,183
54.324
48.606
45.746
500
173,028
96.127
76.901
73.056
65.366
61,521
1 ,000
79,859
44.366
35.493
33.718
30.169
28,394
2.000
62. 113
34.507
27.606
26.225
23.465
22.085
5,000
8,873
4,930
3,944
3,746
3.352
3,155
10,000
4 .437
2.465
1 ,972
1,873
1 .676
1. 577
TOTAL 630.000 350,000 230.000 266.000 238.000 224.000
Exhibit 7.8
Number of Plumes, by Size — Option 1
Plume 1-5 6-10 11-15 16-20 21-25 26-30
size years years years years years years
1 94.706 35,515 23.676 71.029 49,721 47.353
10 115.294 43.235 28.824 86.471 60.529 57.647
25 98.824 37.059 24.706 74.118 51,882 49.412
100 98,824 37,059 24,706 74.118 51,882 49,412
500 98,824 37.059 24.706 74,118 51.882 49.412
1,000 37.059 13.897 9,265 27,794 19.456 18,529
2.000 12.353 4.632 3.088 9.265 6.485 6.176
5.000 4.118 1.544 1.029 3.088 2.162 2.059
10.000 0 0 0 0 0 0
TOTAL 560.000 210,000 140,000 420.000 294.000 280.000
Exhibit 7.9
Number of Plumes, by Size -- Option II
Plume
1-5
0
1
11-15
16-20
21-25
26-30
s ize
years
years
years
years
years
years
1
132.300
73.500
58.800
14.700
8.820
5.880
10
132.300
73.500
58.800
14.700
8.820
5,880
25
100.800
56.000
44.800
11.200
6.720
1 .480
100
113.400
63.000
50.400
12.600
7 .560
5.040
500
88.200
49.000
39.200
9,800
5.880
3.920
1 .000
37,800
21,000
16.800
4.200
2,520
1 .680
2.000
18,900
10.500
8,400
2. 100
1 ,260
840
5.000
6,300
3.500
2.800
700
420
280
10.000
0
0
0
0
0
0
TOTAL
630,000
350.000
280.000
70.000
42,000
28,000
5SSXS3C3S3
33S3aa:::
= = = = S3~*3S
-------
7-14
Exhibit 7.10
Number of
PIumes,
by Size
-- Option III
PlUBC
1-5
6-10
11-15
16-20
21-25
26-30
size
years
years
years
years
years
years
I
9S.000
42,222
42.222
35,889
27.444
23.222
10
130.000
57.778
57.778
49.111
37.556
31,778
25
120.000
53.333
53,333
45,333
34.667
29.333
100
105,000
46,667
46.667
39,667
30,333
25.667
500
115.000
51, 111
51.Ill
43,444
33.222
28,111
1.000
45.000
20.000
20.000
17.000
13.000
11.000
2.000
15,000
6,667
6.667
5.667
4.333
3,667
5.000
5,000
2.222
2.222
1.889
1 .444
1 .222
10.000
0
0
0
0
0
0
TOTAL
630,000
280.000
280,000 238,000
182.000
154.000
aaaaaaaaa
aaaaaaaaa
a a aa aa a ss:
saeaaaaaaaa
i a a s a 3 = a
aasaamssas
a as OB¦a J
Exhibit 7.11
Number of Plumes, by Size — Option IV
PI use
1-5
6-10
11-15
16-20
21-25
26-30
size
years
years
years
years
years
years
l
103.918
38.969
38.969
31.175
25.979
12.990
10
132.784
49.794
49.794
39.835
33.196
16.598
25
92.371
34,639
34.639
27,711
23.093
11.546
100
92.371
34,639
34.639
27.711
23.093
11.546
500
92.371
34.639
34.639
27,711
23.093
11.546
1 .000
34.639
12.990
12.990
10.392
8.660
4.330
2,000
11.546
4.330
4 .330
3.464
2.887
1.443
5.000
0
0
0
0
0
0
10.000
0
0
0
0
0'
0
TOTAL
560.000
210.000
210.000
168.000
140.000
70.000
Exhibit 7.12
Number of Plumes, by Size — Option V
aaaaaaaaaaaa
Plume
size
1-5
years
aaaaaaaaa
6-10
years
11-15
years
16-20
years
21-25
years
26-30
years
1
11.429
2.571
0
0
0
0
10
217.143
48.857
0
0
0
0
25
102.857
23.143
0
0
0
0
100
91.429
20.571
0
0
0
0
500
91,429
20.571
0
0
0
0
1.000
34.286
7.714
0
0
0
0
2.000
11.429
2,571
0
0
0
0
5.000
0
0
0
0
0
0
10.000
0
0
0
0
0
0
TOTAL
560.000
126,000
0
0
0
0
-------
7-15
Exhibit 7.13
Plume Measurements
Pluae Distance Pluae
size fro» source size
(sq. asters) (feet) (acres)
1 3.88 0.0002
10 12.28 0.0025
25 19.40 0.0062
100 38.82 0.0247
500 86.80 0.1235
1000 122.76 0.2471
2000 173.60 0.4942
5000 274.48 1.2355
10000 388.18 2.4710
-------
7-16
Fortunately, the shape of the plume is not the only way to determine a
maximum number of contaminated wells and structures for each size plume. For
well cases, the number of wells contaminated per size plume was derived in the
following manner. Two approaches were used to develop the maxima. One approach
was based on a geometric analysis of the possible numbers of wells in an area
over a given size plume. Using a standard of one-quarter-acre plots of land
per structure, it was determined that each 1000 square meters would contain
at a maximum one structure. This holds for both triangular and elliptical
shaped plumes. In the case of a 10,000-square-meter plume, a maximum of
10 structures could be accommodated. However, this would not allow any space
for on-site land, roads, and other public areas, so it is unlikely that 10
units would exist in a 10,000-square-meter area. Some number less than ten
would be more accurate.
The other method used to determine the number of affected wells was an
analysis of empirical data available to us. These data indicate that the
number of wells affected drops gradually as the plume size declines. The
maximum number of wells found in a 10,000-square-meter site averaged about 5.
Given this fact and the belief that there would have to be less than 10 wells
in such an area, the number 5 was selected for the maximum number of wells in
the 10,000-square-meter case.
For the smaller size cases, the number of wells affected generally dropped
as the plume size decreased. We therefore chose to have the maximum drop
gradually as wel1,.declining from 5 wells to 3 wells for a 5,000-square-meter
plume, to 2 for a 2,000-square-meter plume, to 2 for a 1,000-square-meter plume,
and to 1 for all other plume sizes. Such a decrease also squares with the
geometric analysis, which clearly indicates that the number of prospective well
sites will drop as land area decreases.
In vapor cases, the geometric approach was found to be unrealistic, based on
the real-world observations that were available. The data indicate that it is
extremely rare for any given leak to contaminate more than one structure.
Therefore we have set a limit of one structure per leak for vapor cases.
Another important assumption in the damage function process involves the
proximity of plumes to one another. Exhibit 7.5 showed that, in the base case,
there are projected to be 1,988,000 plumes over a thirty-year period. It is
entirely possible that some of these plumes would intersect one another.
Indeed, observation of actual UST leak incidents suggests this is very likely.
For instance, in a Florida case where several wells were contaminated, there
were three possible sources of contamination, that is from three leaking tanks,
in the general vicinity. How many plumes would intersect and how plume sizes
and shapes would be affected is unknown. In this analysis, it is assumed that
each plume would be separate; there would not be any intersection of plumes.
It is likely that this assumption introduces upward bias into the benefits
estimates since it projects more damaged wells and buildings in the base case
than might actually occur. This follows if it is assumed that a well damaged
by "two" plumes is no more costly than a well damaged by a single plume.
Turning to the Florida case once more, it did not matter to the affected resi-
dents whether one or ten leaking tanks were the source of their well contamina-
tion; all that mattered was the subsequent loss of water use and the steps that
had to be taken to obtain a new source. This did not depend on the number of
piumes.
-------
7-17
Exhibit 7.13(A)
Compilation of Well Contamination and Hi stance
Number of Distance Elliptical Area Triangular Area
Wells (m) (m**2) (m**2)
1
3
6
1
1
6
26
6
1
11
86
20
1
12
103
23
1
15
160
36
1
15
160
36
1
15
160
36
1
15
160
36
1
18
231
53
1
21
314
72
23
377
86
1
24
411
93
1
30
642
146
1
30
642
146
1
30
642
146
2
30
642
146
2
30
642
146
2
30
642
146
2
42
1,258
286
1
45
1 ,444
328
1
45
1 .444
328
2
45
1 ,444
328
3
46
1 ,509
343
3
61
2,654
603
1
76
4, 119
937
2
76
4,119
937
3
92
6,036
1 .373
2
114
9,268
2,107
1
152
16,476
3,747
1
152
16,476
3,747
1
182
23,622
5,371
1
183
23,882
5,431
1
200
28,526
6,486
1
213
32,355
7,357
2
274
53,540
12,174
7
274
53,540
12,174
6
396
111,832
25,430
4
403
115,821
26,337
(a) The ellipse is assumed to have an area A=pi(0.227)x~2, where x is the
distance between the furthest ends of the ellipse.
(b) The triangle is assumed to have an area A=(6/37)x~2, where x is the
distance between the base and the height of the triangle.
-------
7-18
Finally, many of the replacement costs would be incurred in the future.
Thus, the benefits of avoiding such costs should be expressed in present value
terms to allow comparison with regulatory costs. Two discount rates are employed
in this analysis: 3 and 10 percent.
The only explicitly on-site benefits considered in this section are avoided
profit losses that tank owners incur following leaks. We assume that each
plume is from a different site (primarily gasoline retailer). Losses are
measured as foregone profits — the difference between gross revenues and
operating costs. The estimated loss per establishment with a plume varies
across the three scenarios as the assumed net-revenue loss per week and duration
of loss varies. These estimates are derived from a small sub-set of the 141
case studies of known LIST plumes which was used in this analysis.
7.D.3. Damage Function Estimation
Due to the high level of uncertainty involved in estimating the national
benefits of the regulatory options, three different scenarios are considered.
Under Scenario A, which might be thought of as a conservative estimate of
benefits, variable values are selected that minimize damage estimates, and
consequently benefits. Scenarios B and C are medium and high estimates, respec-
tively.
The national benefits of avoiding vapor damages is expressed as a simple
multiplicative function of four variables. These variables are:
o the expected number of plumes;
o the probability, per plume, that some sort of vapor damage
will occur;
o the expected number of affected structures per vapor incident;
o the estimated replacement/restoration cost per structure.
The number of plumes of various sizes expected each year under each
regulatory option have already been presented in Exhibits 7.7-7.12. These
estimates do not vary across Scenarios A, B, and C.
The probability per plume of vapor damage occurring has been estimated
using actual field data on UST leaks. Data from 141 known UST plumes have been
analyzed. Probabilities of a damaging leak occurring were calculated for each
plume size and for each kind of incident. To calculate the probabilities, the
10,000-square-meter plume was used to set a limit on the maximum probability of
any size leak causing contamination. For example, let us examine the vapor
case category. Of the 72 plumes for which distance data were available, 19
were at least 10,000 square meters in size. Of the 19, 5 were known to have
caused vapor damages. Thus, approximately 26 percent (5/19) of these size cases
had vapor impacts.
-------
7-19
Given that 26 percent of the 10,000-square-meter plumes cause vapor
damages, it is now possible to determine the probability for other size vapor
cases. This was done using the following method. It is assumed that as a
plume's size decreases, the probability that any damages occurring from that
plume will decrease. Consequently, the probability for a given size plume is
scaled down according to its size in relation to the 10,000-square-meter plume.
For example, take the case of a 500-square-meter plume. This plume has a
minimum distance (based on the elliptical model of plume shape) of 40 feet from
the tank, while the 10,000-square-meter plume has a minimum distance of 275
feet. Dividing 40 into 275 yields that the distance for the 500-square-meter
plume is 14.5 percent that of the 10,000-square-meter plume. Multiplying this
14.5 percent by the 26 percent probability of the large plume causing vapor
damages indicates that there is about a 4 percent chance of the 500-square-meter
plume causing any vapor damages. This method was applied to all the different
size plumes for vapor and private well cases.
Another important consideration is that the 141 UST cases were from "known"
plumes; i.e., plumes that have been detected by state authorities. It is
universally agreed that many plumes go undetected altogether. Assume that if a
plume does vapor damage, it gets reported and detected. This implies that none
of the undetected plumes do vapor damage. Several state environmental authori-
ties were asked to estimate the proportion of total UST leaks that go undetected.
While many declined to estimate, several others did offer opinions. There was
agreement that 50 to 70 percent of leaks go undetected; i.e., that 30 to 50
percent of leaks are detected. This implies, for instance, that far fewer than
26 percent of all 10,000-square-meter plumes do vapor damages.
To compensate for this fact, the probabilities of damages occurring from a
given leak size were modified using an adjustment factor accounting for the
probability of detection. This factor increases as plume size increases, since
it is assumed that the rate of detection will increase as plume size increases.
The 500-square-meter plume was chosen as a mid-point, where 50 percent of all
leaks will be detected. The adjustment factor rises or decreases by 10 percent
per plume size, to the point where, for example, any 10,000-square-meter plume
will have an 80 percent probability of detection. The adjustment factor may
then be multiplied by the probabilities derived previously to yield a much
better proxy for the probability of a given leak's causing known damages. It
should be remembered that these probabilities are based on a very small and
biased dataset.
The expected number of affected structures per vapor incident has also
been estimated from the field data. Of the 33 vapor cases, only two are known
to have involved more than a single structure. They affected two structures
in one case, and four structures in the other. As previously noted, we have
assumed that a single structure is affected under all scenarios.
The final variable is the estimated damage cost per structure. The 141
field cases indicate a wide range of costs -- from $3,000 to $150,000. Some of
the affected structures were business establishments, others were homes. The
mean cost per structure was about $40,000, while the median cost was closer to
$60,000. These estimates represent fewer than a dozen observations. Due to
the uncertainty of these figures, the estimated cost per structure varies by
scenario.
-------
7-20
The benefits of avoiding private (residential or business) well contamina-
tion are considered separately from the benefits of avoiding public well
contamination.
The national benefits of avoiding private well damages is expressed as a
simple multiplicative function of four variables. These variables are:
o the expected number of plumes;
o the probability, per plume, that some sort of private well damage
will occur;
o the expected number of affected private wells per incident;
o the estimated replacement/restoration cost per private well.
The number of plumes of various sizes expected each year under each
regulatory option has already been presented in Exhibit 7.5. These estimates
do not vary across scenarios.
The method used to derive the probability of a given size leak was the same
as that used for vapor cases. The 10,000-square-meter plume had a 47 percent
chance of causing damages, and this was used as a base, as 26 percent was used
in the vapor cases. After the other probabilities were obtained, they were
multiplied by the same detection adjustment factor used in the vapor cases,
which was 50 percent of the 500-square-meter plumes, 1 percent of the 1-square-
meter cases, and 80 percent of the 10,000-square-meter plumes being discovered.
The expected number of affected private wells per well incident has also
been estimated from the field data. Of the 51 well cases, 20 are known to have
involved more than a single well. Six cases involved five or more wells. The
mean number of affected wells per well incident is three. As stated in Section
2, the maximum number of wells affected per well incident ranges from 5 wells
for the 10,000-square-meter case to 1 for all cases 500 square meters or less.
The final variable is the estimated damage cost per private well contam-
inated. Based on the few case studies for which restoration/replacement costs
were known, three estimates have been derived: $1,000, $3,000, and $10,000
under scenarios A, B, and C, respectively.
The probability of public well contamination is apparently lower than the
probability of private well contamination, but the costs per incident are
potentially much higher. Of the 141 UST leaks examined, 3 resulted in public
well contamination. The data indicate a probability of about 4 percent that
any 10,000-square-meter plume could contaminate a public well. However, this
may be a high estimate, since the data that was collected was heavily biased in
favor of leaks that cause substantial damage. Therefore, the probability for
the 10,000-square-meter plume has been reduced to 1 percent to reduce the bias.
The probability drops for the lower plume sizes by 0.25 percent for each conse-
cutive size above the 500-square-meter plume. At this point, the probability
is zero percent and remains at this level for the smaller sizes. It is assumed
under all scenarios that a single public well is contaminated when such an
incident occurs. The cost per public well affected does vary by scenario --
from $10,000 under Scenario A to $100,000 under Scenario C.
-------
7-21
The only explicitly on-site benefits considered in this section are avoided
profit losses that tank owners incur following leaks. We assume that each
plume is from a different site (primarily gasoline retailer). Losses are
measured as foregone profits — the difference between gross revenues and
operating costs. The estimated loss per establishment with a plume varies
across the three scenarios as the assumed net-revenue loss per week and duration
of loss varies. These estimates are derived from a small sub-set of the 141
case studies.
7.D.4. Property-Damages-Avoided Benefits Projections
Before turning to the model input data and intermediate results, the
summary findings are reported. The data are presented in several forms to
show the differences between the options. In addition, each table contains
aggregate data by scenario for better comparison.
Exhibit 7.14 summarizes the benefits of the five control options relative
to the base case at two discount rates — 3 and 10 percent. The numbers in
Exhibit 7.14 are the sum of the lost profits and benefits and are taken from
Exhibits 7.15A and 7.17, respectively. Therefore, each entry in Exhibit 7.14
represents benefits from avoided vapor damages, private well damages, public
well damages, and lost profits. For Option II under Scenario B (present value
at 3 percent), 69 percent of the benefits are avoided lost profits, while
avoided vapor damages, private well damages, and public well damages represent
24 percent, 6 percent, and 1 percent, respectively.
Exhibit 7.15 presents the lost profits of tank owners discounted at 3 and
10 percent. The lost profits result from the time a station experiencing a
leak closes down. It is important to note that currently no damage probabilities
are included (i.e., it is assumed that a leak always causes a closure of some
length of time).
At a discount rate of 3 percent, damages are lower under Option II than
under Option I. However, at a 10 percent discount rate, damages are slightly
higher under Option II than under Option I. This can be explained with the
aid of Exhibit 7.6. Note in Exhibit 7.6 that the aggregate number of detected
plumes drops from 1.9 million under Option I to 1.4 million under Option II.
This reduction comes solely during years 16-30. In fact, there are more
plumes detected in years 1-15 under Option II. Consequently, present value
damages under Option II are lower at low discount rates but higher at high
discount rates.
The contamination from UST plumes must be examined by local authorities
first-hand. In most cases, this requires the visual inspection of the tanks
and surrounding ground areas. The digging necessary to remove the tanks causes
many commercial facilities to shut down for a period of time. Examination of
the available files has yielded the following data about the impacts on busi-
nesses that sell gasoline. The closing time for a station can vary widely.
In some instances the station may close for only a week while the tanks are
dug up; in other cases, the station may close for a period of months, years, or
permanently. It appears from available data that most stations will shut down
for a period of between 1 and 10 weeks.
-------
7-22
Exhibit 7.14
Benefits Summary: Benefits of Control Options
Relative to Base Case (Billions of Hollars)
Present Value at 3 Percent
Option Scenario A Scenario B Scenario C
I
0.3
2.0
5.7
II
0.6
3.4
9.0
III
0.3
2.0
5.7
IV
0.7
4.3
11.1
V
1.3
7.7
18.8
Present Value at
10 Percent
Scenario A
Scenario B
Scenario C
I
0.3
1.5
4.2
II
0.2
1.1
3.2
III
0.1
0.9
2.8
IV
0.3
2.1
5.5
V
0.6
3.3
8.3
Note: The figures in this table are the sum of the lost profits and
benefits from Exhibits 7.15(A) and 7.17 respectively.
-------
7-23
Net costs for the stations were available from the case files. The net
cost for lost business during closure varied from $1,400 to $1,736 per week,
with a mean of $1,568.
Given the number of plumes under each regulatory option for thirty years
after implementation, lost profits of tank owners can be calculated. These
plume numbers were derived earlier in this section. Three possible scenarios
were examined: Scenarios A, B, and C. Under Scenario A, it was assumed that
cost of closure to tank owners was $1,400 per week, and that the station was
closed for one week. For Scenario B, it was assumed that cost of closure was
$1,568 per week, and that the facility was shut down for a period of five
weeks. Finally, under Scenario C, the assumption was that the tank owner had a
cost of $1,736 per week, and that the facility closed for a period of ten
weeks.
All of the cost figures were derived by multiplying the number of plumes
in a given time period by the cost and period of closure for the specific
scenario. Costs were discounted to the present using discount factors of 3
and 10 percent. The resulting estimated lost profits are presented in Exhibit
7.15.
Exhibit 7.15A reports the benefits of avoiding lost profits under Options
I through V relative to the base case. It is interesting to note that the
benefits are greater under Option II than under Option I when the benefits are
discounted at a rate of 3 percent. The opposite is true when the discount rate
is 10 percent. This has already been explained in the discussion of Exhibit
7.15.
Exhibit 7.16 reports the present value of vapor and well damages under
each option and each scenario, discounted at 3 and 10 percent. The vapor and
well damages are also presented at a 0 percent discount rate, which is just
the total undiscounted damages. Damage estimates increase substantially under
Scenarios A, B, and C for several reasons: probabilities of damages increase
and costs per damaged well or structure increase. Vapor and public well damage
costs become very important (proportionally) under Scenarios B and C.
Damages are lower under Option II than under Option I at discount rates
of 0 and 3 percent, but are higher discounting at 10 percent. This occurs for
the same reason that was explained for lost profits during the discussion of
Exhibit 7.15 — Option II avoids plumes beyond year 15.
Exhibit 7.17 presents projected benefits of Options I through V relative
to the base case. Exhibit 7.18 presents the input assumptions which distinguish
Scenarios A, B, and C from each other.
-------
7-24
Exhibit 7.15
Lost Profits of Tank Owners
Millions of Dollars Discounted
at
3 Percent
Option
Scenario A
Scenario B
Scenario C
Baseline
1,987
11,129
24,643
I
1,810
10,135
22,442
II
1,567
8,775
19,430
III
1,804
10,104
22,373
IV
1,431
8,016
17,749
V
858
4,802
10,633
Millions of Dollars Discounted
at
10 Percent
Option
Scenario A
Scenario B
Scenario C
Baseline
1,141
6,389
14,146
I
970
5,434
12,032
II
1,041
5,829
12,908
III
1,072
6,002
13,291
IV
890
4,985
11,038
V
677
3,793
8,399
Exhibit 7.15(A)
Comparison of Lost
Profits Relative
to Rase Case
Millions of Dollars Discounted
at
3 Percent
Option
Scenario A
Scenario B
Scenario C
I
177
994
2,201
II
420
2,354
5,213
III
183
1,025
2,270
IV
556
3,113
6,893
V
1,130
6,327
14,010
Millions of Dollars Discounted
at
10 Percent
Option
Scenario A
Scenario,B
Scenario C
I
170
955
2,114
II
100
559
1,239
III
69
386
855
IV
251
1,404
3,108
V
463
2,595
5,747
-------
7-25
Exhibit 7.16
Aggregate Damages
PRESENT VALUE OF DAMAGES, R ='3* ($10*6)
Option
Scenario A
Scenario
B
Scenario C
Base case
211.67
1.591.42
5,710.60
I
76.15
611.51
2.211.21
II
68.46
539.58
1,945.78
III
80.31
642.36
2,321.40
IV
48.88
402.85
1.460.18
V
29.20
240.30
871.00
PRESENT VALUE
OF DAMAGES.
R =
¦ 10%
($10*6)
Opt ion
Scenario A
Snenar io
B
Scenario C
Base case
121.51
913.56
3 . 278 .19
I
40.83
327.87
1.185.58
II
45.48
358.45
1.292.62
III
47.71
381.60
1.379 Ofi
IV
30.40
250.53
908.07
V
23.07
189.82
688.02
PRESENT VALUE
OF DAMAGES.
R =
= 0%
($10*6)
Option
Scenario A
Scenario
B
Scenario C
Base case
296.44
2,228.77
7.997.64
I
112.15
900.68
3.256.82
II
85.64
674.94
2.433.90
III
109.93
879.24
3,177.44
IV
64.92
535.07
1,939.42
V
32.70
269.14
975.52
-------
7-26
Exhibit 7.17
Comparison of Benefits to Base Case
PRESENT VALUE OF BENEFITS RELATIVE TO BASE CASE
R = 3* ($10*6)
Option
Scenario A Scenario B
Scenario C
I
135.52 979.91
1 3,499.38
II
143.20 1,051.84
3,764.82
III
131.35 949.06
3,389.20
IV
162.79 1,188.57
4,250.41
V
182.47 1,351.12
4,839.60
PRESENT
VALUE OF BENEFITS RELATIVE TO BASE
R = 10% ($10*6)
CASE
Option
Scenario A Scenario B
Scenario C
I
80.68 585.69
2,092.61
II
76.03 555.11
1,985.5"
III
73.80 531.96
1.899. 13
IV
91.11 663.03
2,370.1?
V
98.44 723.74
2,590.17
-------
7-27
Exhibit 7.18
Input Variables
Scenario A -
Input Variables
No. of
No. of private
Expected value
Plume size
P(vapor
Vapor damages
structures P(private
Damages per wells per
P(public
Oarages per
of damages
(sq. asters)
damages)
per structure
per plume well damages) private well plune well damages)
public well
($/plune)
1
0.000
$3,000
1 0.000
$1,000 1
0.0000
$10,000
$0.00
10
0.000
3,000
1 0.001
1,000 1
0.0000
10,000
1.00
25
0.002
3,000
1 0.004
1,000 1
0.0000
10,000
10.00
100
0.006
3,000
1 0.009
1,000 1
0.0000
10,000
27.00
500
0.016
3,000
1 0.028
1,000 1
0.0000
10,000
76.00
1000
0.040
3,000
1 0.075
1,000 2
0.0025
10,000
295.00
2000
0.072
3,000
1 0.126
1,000 2
0.0050
10,000
518.00
5000
0.119
3,000
1 0.210
1,000 3
0.0075
10,000
1,062.00
10000
0.208
3,000
1 0.376
1,000 5
0.0100
10,000
2,604.00
Scenario B -
Input Variables
No. of
No. of private
Expected value
Plune size
P(vapor
Vapor damages
structures P(private
Parages per wells per
P(public
Oarages per
of damages
(sq. seters)
damages)
per structure per plume well damages) private well plune well damages)
public well
($/plume)
1
0.000
$40,000
1 0.000
$3,000 1
0.0000
$50,000
$0.00
10
0.000
40,000
1 0.001
3,000 1
0.0000
50,000
3.00
25
0.002
40,000
1 0.004
3,000 1
0.0000
50,000
92.00
100
0.006
40,000
1 0.009
3,000 1
0.0000
50,000
267.00
500
0.016
40,000
1 0.028
3,000 1
0.0000
50,000
724.00
1000
0.040
40,000
1 0.075
3,000 2
0.0025
50,000
2,175.00
2000
0.072
40,000
1 0.126
3,000 2
0.0050
50,000
3,886.00
5000
0.119
40,000
1 0.210
3,000 3
0.0075
50,000
7,025.00
10000
0.208
40,000
1 0.376
3,000 5
0.0100
50,000
14,460.00
Scenario C - Input Variables
No. of No. of private Expected value
Plume size P(vapcr Vapor damages structures P(private Damages per wells per P(public Oarages per of damages
(sq. meters) damages) per structure per plume xell damages) private well plune well damages) public well ($/plume)
1
0.000
$150,000 1
0.000
$10,000
1
0.0000
$100,000
$0.00
10
0.000
150,000 1
0.001
10,000
1
0.0000
100,000
10.00
25
0.002
150,000 1
0.004
10,000
1
0.0000
100.000
340.00
100
0.006
150,000 1
0.009
10,000
1
0.0000
100,000
990.00
500
0.016
150,000 1
0.028
10,000
1
0.0000
100,000
2,680.00
1000
0.040
150,000 1
0.075
10,000
2
0.0025
100,000
7,750.00
2000
0.072
150,000 1
0.126
10,000
2
0.0050
100,000
13,820.00
5000
0.119
150,000 1
0.210
10,000
3
0.0075
100,000
24,900.00
10000
0.208
150,000 1
0.376
10,000
5
0.0100
100,000
51.000.00
-------
7-28
7.E. ENVIRONMENTAL EFFECTS
(JST leaks have environmental effects, and thus reducing leaks has a benefit
in terms of environmental effects avoided. The primary environmental effect is
on aquatic ecosystems, and occurs when the release is transported on or in
ground water and discharges to surface water. We performed a screening-level
analysis on the aquatic impacts of gasoline releases from USTs in the base case
to provide an indication of how often UST releases have the potential to pollute
streams. As with the risk assessment discussed in Section 7.C, this analysis
represents a first cut at estimating potential impacts from leaking USTs; we
intend to develop a more complete methodology for subsequent analyses and
invite comments on how to improve the underlying data and modeling approach.
As described below, we used some of the same UST model outputs as in other
parts of the benefits analysis, i.e., the predicted rate of occurrence of
leaks and the leak rates. However, there are several fundamentally different
(and more conservative) assumptions employed in this analysis:
o All gasoline present in the floating plume travels to the nearest
stream and discharges into the stream. This mass of gasoline is
slightly less than the total leak volume as some gasoline is retained
in the unsaturated zone. However, in most cases it is much more than
the mass available for human exposure via ground water, (i.e., the mass
of gasoline that disperses into the ground water).
o The floating plume reaches the nearest stream regardless of whether or
not the leak is detected.
Other key assumptions, and our basic analytic framework, are summarized
below.
7.E.I. Approach and Assumptions
Our approach has three components. First, we devised a set of exposure
scenarios. Each scenario comprises a different combination of stream size
(with a corresponding flow) and gasoline discharge rate, yielding a predicted
gasoline concentration in the water column. Second, we attempted to determine
the concentration of gasoline that is toxic to aquatic organisms. Third, we
estimated the frequency with which each of the exposure scenarios occurs.
We used the hydrologic concept of stream order to provide the basis for
the environmental settings for our exposure scenarios. Streams can be classi-
fied into orders, whereby a first order stream has no tributary channels, a
second order stream is formed when two first order streams merge, a third order
stream is formed when two second order streams merge, and so on. The highest
stream order in the U.S. is tenth order, and is represented by the Mississippi
River. Each stream order has characteristic properties, including mean flow
and length]/. Exhibit 7.19 lists the stream characteristics used to develop
our exposure scenarios.
Keup, L.E., "Flowing Water Resources", Water Resources Bulletin,
Vol. 21, No. 2, April 1985, Pp. 291 - 296.
-------
7-29
Exhibit 7.19
Summary of Data on U.S. Streams by Stream Order
Stream Number of Total Calculated
Order Streams Length (miles) Discharge(ft3/sec)
1
1,570,000
1,570,000
0.6
2
350,000
810,000
3.7
3
80,000
420,000
15.6
4
18,000
220,000
73
5
4,200
116,000
380
6
950
61,000
1,800
7
200
30,000
8,500
8
41
14,000
38,000
9
8
6,200
211,000
10
1
1,800
900,000
Source: Lowell E. Keup, "Flowing Water Resources", Water
Resources Bulletin, Vol. 21, No.2, April 1985.
-------
7-30
We used a simple dilution model to predict concentrations. The mass
loading rate was derived using the UST model's outputs on floating plume size
and duration. We assumed that all gasoline in the floating plume discharges
to the nearest stream. The mass loading rate divided by flow yields the concen-
tration in the stream. As mentioned earlier, our concentration estimates are
conservatively high because we assume that the entire plume enters the stream.
We did not account for dilution, degradation, and other fate and transport
processes that may affect concentration. Another limitation of our approach
is that we do not consider the effect of time on plume travel, or the prob-
ability of plumes in lower order streams merging with plumes in higher order
streams.
Unce concentrations were derived, the next step was to determine whether
they exceeded a chronic toxicity threshold. The EPA has not established an
Ambient Water Quality Criterion for gasoline or any of the specific hydrocarbons
constituting gasoline. Therefore, we had to derive a concentration threshold
that represents the level above which toxic effects may occur.
We were able to obtain only two sets of aquatic toxicity test results for
gasoline, one using rainbow trout and the other using shad. Both tests were
acute tests; we wanted to derive an estimate for the concentration that would
cause chronic effects. We used three different extrapolation methods to esti-
mate a maximum acceptable toxicant concentration (MATC) for chronic exposure,
based on the two acute tests. In addition, we 'back-calculated' the chronic
MATC for gasoline based on available toxicity data for benzene, toluene, xylene,
and naphthalene. The MATCs derived by the four methods were surprisingly
similar, ranging from 0.7 mg/1 to 2.7 mg/1. The figure we used for our analysis
was 0.8 mg/1; the overall results would not change appreciably throughout the
range of MATCs.
Finally, to estimate the distribution of USTs with potential toxic dis-
charges, we calculated a joint frequency distribution for gasoline discharge
rate and stream flow. The distribution of floating plume sizes was estimated
as before for the risk assessment, based on frequencies of different tank
designs and vadose zone types. We assumed that USTs were randomly distributed
among stream orders in proportion to the number of stream miles nationwide
for each stream order (for example, of the total of about 3.24 million stream
miles nationwide, about 1.57 million comprise first order streams; thus 48% of
USTs are assumed to potentially affect first order streams).
-------
7-31
7.E.2. Results
Exhibit 7.20 presents the results of our analysis. Over the thirty-year
simulation period, a very large number of streams of first and second order
have the potential to be contaminated by USTs. We estimate that out of a total
of 1.4 million USTs, up to 560,000 (39%) will have potentially toxic discharges
into first order streams, and about 220,000 are potentially toxic to second
order streams over the thirty-year period. Even streams of orders 3 and 4
could be affected, although to a much lesser degree than first and second order
streams. Up to 5,000 tanks could damage third order streams and about 130
leaking tanks could affect fourth order streams. Streams higher than fourth
order are not affected by leaking UST discharges. Exhibit 7.21 shows the
percentage of all streams in each order that could be contaminated, assuming no
more than one UST per stream. Although leaking USTs are unlikely to cause
direct toxicity to aquatic life in larger streams (which tend to be most impor-
tant in terms of fisheries and other recreational and commercial values), they
have the potential to cause serious damage to a large proportion of the nation's
small streams.
Although we made a number of conservative assumptions in our methodology
and ignored some important physical phenomena, these preliminary results indi-
cate a real potential problem from leaking underground storage tanks. Moreover,
they indicate that reducing leaks from USTs could have a significant benefit in
terms of avoiding aquatic ecosystem impacts. More work is needed to define
better the extent and severity of the current problem, and the effectiveness of
the regulatory options in reducing this problem.
7.F. OPTION AND EXISTENCE VALUE LOSSES
Section 7.D provides dollar estimates of some of the damages borne by
individuals and firms whose use of privately or publicly owned facilities,
operating at the time of an UST leak, would be impaired by ground water-borne
contaminants. While these effects certainly do constitute a portion of the
damages due to UST contamination of ground water, there are other categories of
damages that also should be considered when evaluating the selection of an UST
standard. Because the ground water has been degraded and made unfit for certain
uses, there are numerous individuals and firms that don't currently use that
water but who are made worse off as a result of the contamination. These
additional damages due to an UST leak arise from losses in option value and
existence value, values individuals and firms attach to ground water in its
original condition.
This section of the benefits analysis provides a general description of
option and existence benefits as they relate to ground water and UST regulation.
Some of the principal factors affecting the magnitude of such benefits are
discussed but no quantitative estimates of these benefits occuring from regulat-
ing USTs are made.
-------
Exhibi 20
Number of USTs with Aquatic Impacts
by stream order
Stream Order
-------
"U
u
~
c
E
o
V
c
o
u
10
E
a
u
L
+>
V)
c
u
o
L
V
Q.
Distribution of Streams Impacted
by stream order
Y / / / / /1
~r
5
Stream Order
-------
7-34
7.F.I. Origins of Option and Existence Value
We believe it is helpful to think of nonuser damages as originating in the
lack of perfected ownership of ground-water resources. If ground water could
be owned in a conventional sense, individuals and firms could purchase and sell
ground water, and put it to any use they please, including holding it for
future use or withdrawing it from use altogether. In such a situation, the
differential in the market price for ground water in various places, of various
qualities and for different times of use would provide a measure of the damage
done, both in the present and in the future, by an UST leak. Of course, ground
water is not a conventional good. Either under riparian or prior appropriation
law, ownership of ground-water resources is imperfect: the amount, timing, and
quality of the ground-water resource claimed are not well established. We
believe that some option or existence value is associated with ground water
resources but that the lack of perfected ownership prevents complete expression
of such values in the marketplace.
7.F.2. Option Value
Option value is derived from the willingness of an uncertain future user
of a ground-water resource to pay for an option that would guarantee that the
resource, at its current quality and quantity, would be available for future
use should the need arise.J/ Most of the essential notions connected with
option value are conveyed by the following illustration. Consider a farmer who
wants to retain the option of using ground water for irrigation during periods
of low rainfall or runoff, even though he doesn't currently use ground water
for that purpose. Indeed, if the farmer values that prospective option, he
would be willing to pay some amount today to retain that option in the future,
an option that he might never in fact exercise. Other potential users such as
households, manufacturers, and municipalities might also be willing to pay to
protect their option to use ground water to meet their needs in the future if
there were a market in which they could purchase such options. Contamination
of ground water from an UST leak deprives these individuals and organizations
of the option value they attached to the resource prior to contamination.
How large might option value be? The answer obviously depends on a large
number of factors, many of which relate to local hydrologic and economic
conditions. First, since the value is associated with possible future use of
ground water, current valuation involves discounting the value of such uses to
the present. At most discount rates used for benefits analysis, this effectively
means that option value on prospective uses with a time horizon of fifty years
or greater is virtually zero. The prospective uses must be relatively near at
hand in order to contibute much to the benefits of UST regulation. Second, the
2/ Option value is often defined as the value over and above expected
utility in an uncertain setting. In this analysis we include in option value
the additional expected utility derived by risk-neutral individuals from the
option to participate in a contingent claims market for ground-water resources
This characterization of option value seems appropriate in the case of ground-
water contamination in as much as ownership conditions that are a prerequisite
for the operation of contingent claim markets for ground water have not generally
been satisfied.
-------
7-35
option value is directly related to the cost of the alternative to the prospec-
tive service provided by ground water. For the farmer in the example described
above, an inexpensive alternate surface supply would diminish the option value
of the ground-water resource. The same reasoning, however, supports a high
option value when the ground-water resource is a relatively inexpensive source
to produce and essentially non-renewab]e. Such is the case, for example, for
the municipal water supply of El Paso, Texas and the domestic, municipal, and
agricultural water supplies of Tucson, Arizona and Roswel1 Basin of New Mexico.
Option value is also directly related to two probabilities: the probabil-
ity that the ground water will be contaminated and the probability that the
ground water will be used in the future. If the probability of contamination
is high initially and will be substantially reduced by an UST regulation, then
benefits related to option value will be relatively high. Similarly, if the
likelihood that the potential user will desire to use ground water within the
next fifty years is high, then the benefits of UST regulation will be commen-
surately higher.
There are other influences whose effect on option value is more subtle.
Generally speaking, the more elastic the derived demand for water resources,
the less the option value. This implies that uses whose demand is conventionally
regarded as inelastic, such as indoor domestic use, will, all else equal, have
higher option values than demand related to those uses which are more elastic
e.g., industrial use. Option values for ground water under uncertainty also
increase with the risk aversion of the relevant individuals or organizations.
Those who are averse to risk will be willing to pay a premium above the expected
cost of contamination damage for the chance that the ground water will not be
contaminated. Furthermore, an additional, quasi-option value related to a
willingness to pay for a delay in putting the ground water at risk may be found
for unique ground-water resources where contamination would be irreversible.
There have been few studies valuing ground water or estimating the bene-
fits of preventing ground water contamination e.g., EPA (1985)V and Raucher
(1986)£/. None that we are aware of specifically estimate option value or
aggregate to the national level. Most involve case studies whose value to our
analysis is limited by the special conditions of the cases selected and the
types of cost and benefit categories employed.
\j Policy Planning & Evaluation, "Value of Ground Water in Regional
Cases," Draft Report to U.S. Environmental Protection Agency,
Office of Drinking Water, McLean Va., September, 1986.
£/ Raucher, Robert L., "The Benefits and Costs of Policies Related to
Ground-Water Contamination," Land Economics, Vol. 62, No. 1,
pp. 33-45, February, 1986.
-------
7-36
7.F.3. Existence Value
Existence value has been defined by Krutilia and Fisher as, "the value
some individuals place on the knowledge of the mere existence of gifts of nature,
even when they feel certain they will never have or choose an opportunity to
experience them in situ"2/ (1975, p. 124). The motives underlying existence
value, of satisfaction, methods for estimating it, and the relationships between
existence value and other measures of value have received much attention in
recent resource economics literature. While perhaps a majority of resource
economists recognize existence value to be a legitimate component of benefits
analysis c.f., Madariaga and McConnel1 (1985)£/ and Smith and Desvouges
(1986)V, there are those, such as Brookshire, Eubanks, and Sorg (1986)4/ who
question the validity of existence value concepts when applied in efficTency-
based economic analysis.
By definition, existence value is positive. Within the context of ground
water contamination by USTs, existence value is the amount some decision makers
are willing to pay to assure the unimpaired existence of a ground water resource.
If this is the case, existence values preserved by an effective UST regulation
should be added to the other benefits categories associated with the regulation.
Omitting such existence values underestimates the benefits of the regulation.
Brookshire, Eubanks and Sorg (1986, pp. 51-2)fy summarize eleven studies
through 1983 that have reported estimates that appear to relate to existence
value. These studies all used a contingent valuation framework to produce
estimates but employed a wide variety of ways to illicit existence value from
respondents. These and most of the more recent studies attempt to estimate
existence value for wildlife, special natural environments such as the Grand
Canyon, or surface water quality. None that we could find directly addresses
the individuals willingness to pay for the existence of a ground-water acquifer
uncontaminated by an UST leak. As a consequence, the literature offers no
specific guidance as to what the existence benefits of an UST regulation pro-
tecting ground water might be.
V Krutilla, John V., and Anthony C. Fisher, The Economics of Natural
Environments: Studies in the Valuation of Commodity and Amenity
Resources, Baltimore, Johns Hopkins Press for Resources for the
Future, 1975.
£/ Madariaga, Bruce and K.E. McConnel1, Exploring Existence Value,
draft prepared for the AERE Workshop on Recreation Demand Modeling,
Boulder, Colorado, May 1985.
2/ Smith, V. Kerry and William H. Desvouges, Measuring Water Quality
Benefits, Boston: K1uwer-Nijhoff Publishing, 1986.
V Brookshire, David S., Larry S. Eubanks, and Cindy F. Sorg, "Existence
Value and Normative Economics: Implications for Valuing Water
Resources," Water Resources Research, Vol. 22, No. 11, pp 1509-1518,
October, 1986.
5/ Brookshire, David S., Larry S. Eubanks, and Cindy F. Sorg, "Existence
Value and Normative Economics: Implications for Valuing Water
Resources," Water Resources Research, Vol. 22, No. 11, pp 1509-1518,
October, 1986.
-------
Chapter 8
IMPLEMENTATION CONCERNS AND SUMMARY COMPARISON OF REGULATORY OPTIONS
This chapter addresses implementation concerns and presents a summary compar-
ison of the five regulatory options presented previously in Chapter 5. This
summary comparison draws on the information on costs, effectiveness, economic
impacts and benefits presented previously in Chapters 5, 6, and 7. Section 8.A
reviews the regulatory options, while Section 8.B discusses implementation con-
cerns. Section 8.C then provides summary data and discussion.
8.A. REVIEW OF REGULATORY OPTIONS
For the reader's convenience, the assumed corrective action policy and
the five regulatory options are restated here:
Assumed Corrective Action Policy: Where a release has occured, an investigation of
and actions to reduce immediate hazards are followed by limited removal of con-
taminated soil and removal of any free product from the ground water. The need
for more extensive ground water clean up is determined through site-specific
exposure and risk assessment. This corrective action policy is assumed to apply
to the base case and the regulatory options.
"No Further Regulation" Base Case: assumes no regulation beyond the interim pro-
hibition requirements; existing tanks are mostly bare steel tanks without supple-
mentary detection or inventory control. Tanks found to be leaking are replaced
with coated and cathodically protected tanks.
Option I (Baseline level): requires manual inventory control, periodic leak detec-
tection within three years for existing tanks (five years for corrosion-resistant
tanks), and corrosion-resistance for all replacement (new) tanks. Given that the
regulated community is allowed to choose from a variety of detection measures,
this option is modeled as though half of all operators choose quarterly vapor
well monitoring, and the other half choose to do tank tightness tests every three
years (five years for corrosion-resistant tanks). Tanks are assumed to be replac-
ed with coated and cathodically-protected tanks with line leak detectors, and
either quarterly vapor well monitoring or a tightness tests every five years.
Option II—The Proposed Rule (Enhanced baseline plus targeted upgrading): is simi
lar to Option I, but requires upgrading to new tank standards within ten years,
requires leak detection systems to be sampled monthly rather than quarterly and
tightness tests are not required after tanks are replaced. For modeling purposes,
operators are assumed to retrofit with cathodic protection and monthly vapor wells
to meet new tank standards within eight years of the date that the regulations
take affect. New tanks are assumed to be coated and cathodically protected, to
have line leak detectors, and to be equipped with vapor wells that are sampled
monthly.
Option III (Baseline plus secondary containment for new tanks): requires periodic
leak detection for existing tanks and secondary containment with interstitial
monitoring for new tanks. For existing tanks, this option is modeled identical
to Option I. Replacement tanks are modeled as lined systems with interstitial
moni tori ng.
-------
8-2
Option IV (Class Option): requires rapid replacement of existing tanks and
secondary containment for replacement tanks at state-designated well-head
protection areas.V Tanks in other areas are required to conform to baseline
standards (Option I). It is assumed that 40 percent of the tank population is
located within a well-head protection area. Tanks located in these state-
designated areas are assumed to be fitted with continuous vapor wells after one
year, and then replaced before the fifth year with protected tanks in liners.
The other 60 percent of tanks (those outside well-head protection areas) are
modeled the same as Option I. As in other options, ground water is cleaned up
at 40 percent of sites where the release has reached ground water; all of these
clean ups are assumed to be performed in the well-head protection areas.
Option V (Emphasis on prevention): For existing tanks, this option requires
manual inventory control, frequent leak detection starting three years after
the regulations take effect, and early retirement. Replacement tanks must have
secondary containment. Half of all existing tanks are modeled with continuous
vapor well monitoring, and half are modeled with monthly vapor well monitoring
and three year tightness tests. Existing tanks are replaced with lined systems
either when they fail, or within eight years of the effective date of the regu-
lations.
8.B. IMPLEMENTATION CONCERNS
The cost and feasibility of implementation, both to regulators and to the
regulated community, can significantly affect the relative desirability of the
options under consideration. Factors to be considered include:
(1) The degree to which a proposal is self-implementing (e.g., the interim
prohibition induces tank manufacturers to produce only corrosion-
resistant tanks);
(2) The likelihood of compliance;
(3) The costs of ensuring adequate compliance (ease of enforcement);
(4) Capacity constraints inhibiting compliance; and
(5) The extent to which the regulated community can afford compliance
costs.
For UST regulatory options, there are interrelationships among prevention,
detection, and corrective action that need to be addressed from an implementa-
tion perspective. For example, the timing and scope of detection requirements
2/ Under the requirements set forth in the 1986 Amendments to the Safe
Drinking Water Act, states must designate well-head protection areas as part of
a complete ground water protection plan in order to receive EPA grants to
cover the costs of state ground-water programs. The bill defines a well-head
protection area as the "surface and subsurface area surrounding a water well or
wellfield supplying a public water system, through which contaminants are
reasonably likely to move toward and reach such water well or wellfield".
-------
8-3
can significantly affect the timing and magnitude of demands on firms providing
leak detection equipment and corrective action and closure services, as well as
affect the government entities overseeing corrective action and closure. If,
for example, a large number of existing tanks are leaking, immediate require-
ments for testing would result in significant immediate demands for leak detec-
tion equipment and for corrective action services. Prevention-related require-
ments could have the opposite effect: these requirements could significantly
reduce demands on the corrective action program in the long run by reducing the
number and size of releases. However, if compliance deadlines for prevention-
related requirements are not phased over an adequate period of time, they may
impose capacity constraint problems and high initial compliance costs, which in
turn may lead to economic dislocation.
Phasing compliance deadlines is one way that regulators can address some
implementation concerns. Usually, the longer the phase-in period, the fewer
the implementation concerns. This is true because there is more time to resolve
capacity constraints, or because compliance can be more easily phased into
regular capital replacement and maintenance schedules.
The following discussion of the regulatory options from an implementation
perspective is more qualitative than quantitative. Furthermore, the options
are only evaluated relative to each other, rather than in an absolute sense.
Thus, if an option is shown to have low implementation concerns relative to
other options, it does not necessarily follow that the option is easy to imple-
ment in an absolute sense. This evaluation is presented below.
Option I (Baseline): This option raises very few significant implementation
questions. The requirement that replacement tanks be cathodically protected
does not go beyond the statutorily mandated interim prohibition. The interim
prohibition probably ensures that very few bare steel tanks will even be manu-
factured in the future, thus reducing the chances for non-compliance. The
detection requirement can be met by a tank tightness test every three to five
years, so the rate at which releases are detected is not likely to lead to a
short-run capacity problem with respect to corrective action.
Option II (Baseline plus Target Upgrading): This option also raises very few
significant implementation questions. The upgrading requirement is phased in
over ten years, thus reducing the chance of transitional implementation prob-
lems. New tanks need only be corrosion protected, a provision which should be
easily self-implemented. This option also allows for retrofitting of cathodic
protection and specifies operation and maintenance requirements for cathodic
protection. The relative affordability of these provisions should facilitate
compliance. Deadlines for meeting leak detection standards are sufficiently
phased in so that no serious implementation problems should result.
Option III (Baseline plus Secondary Containment for New Tanks): The secondary
containment requirement for new tanks may be difficult to implement because of
possible capacity shortages in the tank manufacturing market and shortages in
installation expertise even though the need for replacements arises only as
existing tanks are found to be leaking. Also, the additional cost of secondary
containment may cause potential economic dislocation, and thus hinder implemen-
tation. In short, implementation concerns for Option III are estimated to be
greater than the implementation concerns for Option II.
-------
8-4
Option IV (Class Option): Implementation concerns play a key role in consi-
dering this option. Although it seems reasonable for the stringency of a
regulation to match the potential magnitude of the problem, implementation of
this regulatory approach may be impeded by difficulty and confusion associated
with assigning individual facilities to a particular class. Such difficulty
and confusion may exist for both the enforcers (federal and state) and the
regulated community. Given that there are already regulatory distinctions
based on UST contents (petroleum/hazardous substances) and UST type (protected/
unprotected), additional distinctions based on tank location may further
complicate the regulatory framework. Furthermore, it may take a number of
years for well-head protection zones to be sufficiently defined to be used
as the basis for UST regulation. These implementation concerns detract from
the theoretical attractiveness of this option.
Option V (Emphasis on Prevention): Mandatory tank replacement within ten years
potentially exacerbates any capacity problems associated with production and
installation of tanks with secondary containment. In addition, the requirement
for mandatory replacement and secondary containment for new tanks raises ques-
tions of affordability (as discussed in Chapter 6), and therefore could hinder
implementation. Also, the provisions for frequent and rapid leak detection
raise questions of resource adequacy to meet increased demands for corrective
action and tank replacement. In short, it is likely that this option is more
difficult to implement than Option II.
8.C. SUMMARY COMPARISONS OF INTEGRATED OPTIONS
This section is presented in five sub-sections as follows:
o Section 8.C.1 summarizes costs;
o Section 8.C.2 summarizes effectiveness and benefits;
o Section 8.C.3 summarizes trade-offs;
o Section B.C.4 summarizes economic impacts; and
o Section 8.C.5 summarizes all key findings
8.C.I. Summary of Costs
Exhibit 8.1 provides sunmary data on the costs expected to be incurred
under the base case and under each of the regulatory options. Costs are
provided for prevention and detection (which is defined to include tank acqui-
sition /ins tal lTHorrr^itin7Fr?7monTtori ng, tank removal, and an offset for
reduction in product lost), corrective action (assuming EPA's corrective action
requirements are applied to each scenario, including the base case), and total
costs (sum of prevention/detection and corrective action). Exhibit 8.1(a)
provides cost data on a total present value basis (calculated by discounting a
future thirty-year cost stream back to the present using a 3 percent discount
rate), while Exhibit 8.1(b) presents the same cost data on an annualized.basis
(smoothing the present value costs over each of the thirty-years using the same
3 percent discount rate).
-------
8-5
Exhibit 8.1
SUMMARY OF TOTAL COSTS
(a) Total Present Value Costs
Total PV Cost of Prevention and Detection (1)
Total PV Cost of Corrective Action (2)
Total PV Cost (3)
Base Case Option I Option XI Option III Option IV Option V
($ Billions)
$31.0 $33.0 $35.0 $43.0 $45.0 $56.5
B9.8 80.9 59.B 75.6 50.1 40.6
$120.8 $113.9 $94.B $118.6 $95.1 $97.1
(1) Cost includes tank acquisition/installation, detection/monitoring, tank removal, and an offset for reduction
in product lost. Present Value includes all capital and operating costs expected to be incurred over 30 years,
discounted at 3X.
'2) Estimated costs for corrective action in each scenario, assuming EPA's corrective action requirements
are applied to each scenario (including the base case).
(3) Total PV Cost of Prevention and Detection + Total PV Cost of Corrective Action.
(b) Annualized Costs *
Annualized Cost of Prevention and Detection
Annualized Cost of Corrective Action
Total Annualized Cost
Base Case Option I Option II Option III Option IV Option V
($ Billions)
$1.58 $1.68 $1.79 $2.19 $2.30 $2.88
4.58 4.13 3.05 3.86 2.56 2.07
$6.16 $5.81 $4.84 $6.05 $4.86 $4.95
* Annualized costs derived by smoothing PV costs over 30 years at a 3X discount rate (Annualization factor = 0.051).
-------
8-6
Exhibit 8.1 accounts for both capital costs and annual operating and mainten-
ance costs. Because capital expenditures are staggered throughout the thirty-
year period, a total present value approach is the preferred approach to account
for timing differences. Exhibit 8.1 represents the cost data derived from the
runs of the UST Model, and therefore provides the basis for incremental costing
of regulatory options, as will be discussed in Section B.C.3.
Exhibit 8.1 suggests that Options II, IV, and V have the lowest total
present value costs, when costs of prevention and detection and corrective
action are both considered. However, these three options differ in terms of
how total present value costs are distributed between prevention and detection
costs and corrective action costs. This is addressed in Section 8.C.3.
B.C.2 Summary of Effectiveness and Renefits
Findings on effectiveness and benefits are summarized in Exhibit 8.2.
As discussed in Chapter 5, plume area avoided is a reasonable effectiveness
measure on which to compare regulatory options. Although corrective action
also mitigates damages, there are damages which occur before cleanup begins,
there may be some possibility of residual damages remaining once cleanup has
been completed, and there are some feasibility and implementation issues asso-
ciated with corrective action. Exhibit 8.2 provides data on the plume area
that would occur in the base case and under each of the options, and provides
the percentage of plume area avoided relative to the base case for each option.
The options avoid between 54% and 83% of the plume area associated with the
base case.
Benefits data are provided in Exhibit 8.2 with respect to property damage
and profits lost, cancer risks, explosion risks, and damage to fish. Property
damage and profits lost are provided for the base case and each of the regula-
tory options. Low, middle, and high estimates are provided for these damages,
based on differing assumptions about vapor damages per structure, damages per
private well, and damages per public well, as explained in Chapter 7. Data is
also provided for cancer risks associated with consuming ground water contami-
nated with benzene due to a petroleum UST leak. In addition, Exhibit 8.2 shows
there are base-case risks due to explosion hazards, as well as base-case risks
associated with potential damage to fish. These base-case risks have not yet
been quantified, nor have they been analyzed with respect to how the various
regulatory options might mitigate them.
In addition to the benefits presented in Exhibit 8.2, there are also the
benefits associated with option value and existence value. Although not subject
to quantification, some corrective actions are currently undertaken even though
damages apparently avoided do not seem to justify the corrective action expendi-
tures. This suggests that option value and existence value can be significant.
-------
8-7
Exhibit 8.2
SUMMARY OF EFFECTIVENESS AND BENEFITS
EFFECTIVENESS
Base Case Option I Option II Option III Option IV Option V
Floating Flume Acres (1)
Percent of Plume Area Avoided (2)
190,387 86,859 61,886 84,886 61,371 31,465
54X 67X 55X 68X 83X
BENEFITS
PV of Property Damage and Profits Lost
($ billions): (3)
Lou Estimate
Middle Estimate
High Estimate
$2.2
12.7
30. 4
$1.9
10.7
24.7
$1.6
9.3
21.4
$1.9
10.7
24.7
$1.5
8.4
19.2
$0.9
5.0
11.5
Cancer Risks:
Percentage of USTs with MEI Risks > 10 -6 (4)
(limited by taste & odor threshold)
Percentage of USTs with MEI Risks > 10 -6 (5)
(not limited by taste & odor threshold)
7%
20*
2X
5X
3X
8%
22
6X
OX
OX
2X
3X
Explosion Risks
Damage to Fish
(6)
(7)
(1) Floating plume acres is a measure of the quantity of contaminated ground water that would occur under any scenario,
(including the base case), given the level of prevention and detection for that scenario.
(2) For each option, calculated as ((Plume Acres in Base Case) - (Plume Acres under Option)) / (Plume Acres in Base Case).
(3) Present value of property damage and profits lost due to plumes occuring in that scenario (Including base case). Low,
medium and high estimates derived by adjusting estimates of vapor damages per structure, damages per private well,
and damages per public well (see Chapter 7).
(4) Percentage of USTs where the most exposed individual has a risk greater than 10 -6 of contracting cancer due to
lifetime consumption of ground water contaminated with benzene due to a petroleum UST leak (assuming that consumption
stops when the taste or odor threshold is reached).
(5) Percentage of USTs where the roost exposed individual has a risk greater than 10 -6 of contracting cancer due to
lifetime consumption of ground water contaminated with benzene due to a petroleum UST leak (assuming that consumption
is not limited by the taste or odor threshold).
(6) The Release Incident Survey documented 141 fire and explosion incidents out of a total of 10,000 reported release
incidents, though the risk of death or injury is not known.
'7) Concentrations of gasoline as low as 0.005 mg/1 can cause organoleptic effects (bad tastes or smells) in fish,
impairing commercial and recreational fishing.
-------
8-8
8.C.3. Summary of Trade-offs
Exhibit 8.3 provides a summary of relevant trade-offs. Exhibit 8.3(a)
presents the trade-off between costs of prevention and detection with corrective
action costs avoided. As stated previously, corrective action costs avoided in
some ways represent the savings of improved prevention and detection. After
reviewing the cost data for the base case and each regulatory option already
presented in Exhibit 8.1, Exhibit 8.3(a) presents incremental prevention and
detection costs and incremental corrective action costs avoided relative to the
base case.
Exhibit 8.3(a) shows that Option I entails $2.0 billion in incremental
prevention and detection costs, which result in an incremental $8.9 billion in
corrective action costs avoided, relative to the base case, or a net savings of
$6.9 billion. Similarly, Option II entails $4.0 billion in incremental preven-
tion and detection costs, which result in $30.0 billion in corrective action
costs avoided, relative to the hase case, or a net savings of $26.0 billion.
Thus, both Options I and II provide net savings when viewing the basic trade-
off in this manner. Relative to Option I, Option II includes an incremental
$2.0 billion in prevention and detection costs, but yields an incremental
savings of $21.1 billion in corrective action costs.
However when the trade-off is viewed in this manner, Options III, IV, and
V do not provide net savings relative to Option II. Option III entails greater
incremental costs for prevention and detection than Option II, yet does not
produce the same level of incremental corrective action costs avoided. Option
II therefore clearly dominates Option III. It is less clear whether Option II
dominates Options IV and V, however, because there is less of a difference
between them in net savings from the base case. Therefore, additional dollars
spent on prevention and detection in Options IV and V just about pay for them-
selves in terms of corrective action costs avoided.
Exhibit 8.3(b) displays the trade-off between prevention and detection
costs and plume acres avoided, not taking corrective action costs into account.
When using plume acres avoided as the effectiveness criterion, Option III is
still dominated by Option II because Option III avoids fewer plume acres than
Option II for a greater incremental cost. In addition, it seems reasonable to
assume that Option IV is dominated by Option II when using plume acres avoided
as the effectiveness criterion, because Option IV produces roughly the same
level of incremental plume area avoided as Option II, but at a greater incre-
mental cost of prevention and detection.^/ Option V has the highest incremental
cost of prevention and detection and avoids the most plume area. However, the
cost per additional plume acre avoided associated with Option V is much higher
than the cost per additional plume acre avoided associated with Option II.
Comparisons between these options also depend on findings from the analysis of
implementation concerns and economic impacts.
V Because Option IV requires greater levels of prevention and detection
in state-designated well-head protection areas, it produces fewer plumes acres
in these areas, while perhaps allowing more plume acres in other areas. Thus,
it is reasonable that Option IV avoids corrective action costs (because the most
extensive corrective actions would occur in well-head protection areas), but
not necessarily plume acres, relative to Option II.
-------
8-9
Exhibit 8.3
TRADE-OFFS
(a) Cost of Prevention & Detection vs. Cost of Corrective Action Avoided
Base Case Option I Option II Option III Option IV Option V
PRESENT VALUE COSTS
($ Billions)
1.
PV Costs of Prevention and Detection
$31.0
$33.0
$35.0
$43.0
$45.0
$56.5
2.
PV Costs of Corrective Action
89.8
80.9
59.8
75.6
50.1
40.6
3.
Total PV Cost (Row 1 + Row 2)
$120.8
$113.9
$94.8
$118.6
$95.1
$97.1
INCREMENTAL COST FROM BASE CASE
4.
Incremental Costs of Prevention and Detection
$2.0
$4.0
$12.0
$14.0
$25.5
5.
Incremental Corrective Action Costs Avoided
8.9
30.0
14.2
39.7
49.2
6.
Net Savings (Row 5 - Row 4)
$6.9
$26.0
$2.2
$25.7
$23.7
(b) Cost of Prevention and Detection vs. Plume Acres Avoided
Base Case Option I Option II Option III Option IV Option V
Incremental PV Costs of Prevention and Detection
(from base case) ($ billions) $2.0 34.0 $12.0 $14.0 $25.5
Incremental Percent Flume Area Avoided
(from base case) 54X 611 55% 68% 83*
Incremental Plume Area Avoided
(from base case) (acres)
103,500 128,500 105,500 129,000 159,000
-------
8-10
8.C.4. Summary of Economic Impacts
Economic impacts associated with the regulatory alternatives have been
addressed in Chapter 6 and are summarized in Exhibit 8.4. The economic impact
analysis focused on the retail motor fuel industry because: (1) there are no
substitutes for USTs in retail motor fuel marketing, (2) UST costs represent a
significant fraction of capital and operating costs for these facilities, and
(3) there are many small establishments. Exhibit 8.4 presents projected exits
within the first five years of regulation for retail motor fuel outlets owned
by small firms under each of the regulatory options. Exit projections under an
assumption of no revenue increase are presented in Exhibit 8.4(a), and Exhibit
8.4(b) provides a projection for Option II under a 3 percent revenue increase.^/
The data in Exhibit 8.4(a) first indicate that under a no revenue increase
assumption, 19 percent of outlets owned by small firms would naturally exit in
the first five years. Exhibit 8.4(a) further indicates little difference
between Options I, II, and III in expected economic impacts on small firms
owning retail motor fuel outlets. Tank replacement or upgrade costs eliminate
2 percent of these firms under each of these options. Although the secondary
containment requirements for new tanks in Option III might be expected to
produce relatively more exits, the fact that, under this option, replacement
need not occur until leaks are detected suggests that corrective action costs
will produce the exit. In fact, under all three of these options, corrective
action burdens are projected to cause an additional exit of approximately 56
percent of these small firms in the first five years under the (worst case) no
revenue increase assumption.^/
In the case of options IV and V, a higher proportion of the total exits
are due to the tank replacement or upgrade requirements. In the case of Option
IV, exit due to tank replacement and upgrade is 15 percent while exit due to
corrective action is 39 percent. Under Option V, exit due to tank replacement
or upgrade is 41 percent while corrective action-induced exit is 22 percent.
Exhibit 8.4(b) provides some indication of the effects of relaxing the no
revenue increase assumption. A 3 percent revenue increase assumption is used
because, as Chapter 6 suggests, such a level would be adequate for larger firms
to remain sufficiently profitable. The exhibit suggests that potential economic
impacts are somewhat mitigated by relaxing the no revenue increase assumption.
If the 19 percent natural exit assuming no revenue increase is subtracted from
the 37 percent exit due to corrective action assuming a 3 percent revenue
increase, incremental exit due to the proposed rule is 18 percent. Should a
revenue increase greater than 3 percent occur, survival rates under Option II
are projected to approach base case survival rates in ten years, as discussed
in Chapter 6. Only Option II has been run under a revenue increase assumption.
V Revenue increases can occur due to increases in price or quantity. Argu-
ments have been presented in Chapter 6 that the fixed cost nature of compliance
costs mitigates the likelihood of compliance costs being passed forward in the
form of increased prices. However, as outlets close, it is not unreasonable to
expect quantity demanded to be redistributed among remaining outlets.
£/ Chapter 6 suggests that profitability in larger firms (excluding large oil
companies) cogld also be significantly affected by corrective action burdens
under assumptions of no revenue increase.
-------
8-11
Exhibit 8.4
SUMMARY OF ECONOMIC IMPACTS
(a) Cumulative Percentage of Outlets Owned by Small Firms Exiting Through Year Five **
(Assuming No Revenue Increase)
Option I Option II Option III Option IV Option V
Natural Exit Through Year 5 (1) 19X 19X 19% 19X 19X
Exit - Tank Replacement or Upgrade Through Year 5 (2) 2X 2% 2% 15X 41X
Exit - Corrective Action Through Year 5 (3) 52X 50X 52X 393£ 22X
Total Exit Through Year 5 73X 71X 73X 73X 82X
(b) Cumulative Percentage of Outlets Owned by Small Firms Exiting Through Year 5
Assuming a 3X Revenue Increase (Option II only)
"xit - Tank Replacement or Upgrade Through Year 5 OX
^xit - Corrective Action Through Year 5 37X
** Small firms in the retail motor fuels marketing sector are defined as having annual sales less than $4.6 million.
(1) Based on extrapolation of past exit rate trends for outlets owned by small firms.
(2) Exit caused by leak detection, replacement of leaking tanks, and mandatory tank retirement or upgrading costs.
(3) Exit caused by corrective action costs.
-------
8-12
8.C.5 Chapter Summary
This chapter has drawn from all previous chapters in this analysis to
present a summary of the key findings. These key findings are condensed
into Exhibit 8.5. The key findings can be summarized as follows:
o All of the options are substantially effective in mitigating the problems
associated with the base case. The options avoid between 54 percent
and 83 percent of the plume area associated with the base case. Plume
area avoided is a reasonable measure of effectiveness because the alter-
native, corrective action costs avoided, does not eliminate all damages
associated with plumes.
o When viewing the basic trade-off as costs of prevention and detection
versus corrective action costs avoided, the proposed rule (Option II)
dominates Option I and III. Relative to the base case, Option II pro-
duces more corrective action savings than Option III at a lower cost of
prevention and detection. Relative to Option I, Option II produces
approximately $21 billion in incremental corrective action costs avoided
for an incremental investment of $2 billion in prevention and detection.
o Option II does not necessarily dominate Options IV or V when viewing the
basic trade-off as above. The extra dollars spent on prevention and
detection in Options IV and V just about pay for themselves in terms of
corrective action cost savings. This is the case because Options II,
IV, and V all produce about the same level of net savings from the base
case.
o However, Option IV has the most serious implementation concerns associ-
ated with it because of it requires classification of UST locations
classes. Option V has the most significant potential economic impacts
for prevention and detection, because of high costs associated with man-
datory replacement of existing tanks with tank systems utilizing secon-
dary containment. Option V also raises questions of capacity with
respect to producing and installing tank systems utilizing secondary
containment.
o In general, the potential for significant economic impacts associated
with corrective action costs suggests that the undertaking of corrective
action measures may sometimes be impeded by affordability considerations.
Firms that go out of business will not be able to cover the full cost of
the corrective action measures needed.
o Any and all analytical results must be viewed in the context of the
limitations of this analysis. These limitations are addressed in Chap-
ter 9.
-------
8-13
Exhibit 8.5
SUMMARY OF EFFECTS
Base Case
Option I
Option II
Option III Option IV
Option V
INCREMENTAL COSTS FROM BASE CASE
1.
Prevention and Detection Costs (S billions)
$2.0
$4.0
$12.0
$14.0
$25.5
2.
Corrective Action Cost Savings ($ billions)
a.9
30.0
14.2
39.7
49.2
3.
Net Savings ($ billions) (Line 2 - Line 1)
EFFECTIVENESS RELEATIVE TO BASE CASE
$6.9
$26.0
$2.2
$25.7
$23.7
4.
Percent of Plume Acres Avoided
54*
67*
55*
68*
83*
5.
Total Plume Acres Avoided
CLOSURES ASSUMING NO REVENUE INCREASE
103,500
128,500
105,500
129,000
159,000
6.
Percent of Retail Petroleum Outlets Owned by Small
Firms Closed In First 5 Years Due to Tank Replacement
2*
2*
22
15*
41*
7.
Percent of Retail Petroleum Outlets Owned by Small
Closed In First 5 Years Due to Corrective Action Cost
CLOSURES ASSUMING A 3* REVENUE INCREASE
57*
55*
56*
43*
17*
8.
Percent of Retail Petroleum Outlets Owned by Small
Firms Closed In First 5 Years Due to Tank Replacements
OX
9.
Percent of Retail Petroleum Outlets Owned by Small
Closed In First 5 Years Due to Corrective Action Costs
IMPLEMENTATION CONCERNS
41*
(1)
10.
Relative Comparisons
lov
low
medium
high
medium
(1) If the expected 19X natural exit assuming no revenue increase is subtracted from the 412 exit due to corrective action,
assuming a 32 revenue increase, incremental exit due to the proposed rule is 222. Under a 32 revenue increase, there
would be no natural exit.
-------
Chapter 9
LIMITATIONS OF THE ANALYSIS
9.A. UNCERTAINTIES IN ASSUMPTIONS, DATA, AND METHODOLOGY
9.A.I. Overview of Uncertainties
The problems of discerning the damages caused by leaking USTs, of designing
a regulatory response to these damages, and of measuring the expected effects
of the response are technically very complex. A complete understanding of tank
systems and the risks they pose requires a tremendous amount of knowledge
in a large number of fields. In many instances the regulatory analysis must
proceed on the basis of partial information and estimates or assumptions
derived from the data that is available. The conclusions of the analysis, as a
result, will be subject to limitations caused by uncertainties with available
data and the realm of error implicit in the assumptions.
In this chapter, we describe the most important areas of uncertainty in the
analysis and identify the major limitations of our analysis. Many uncertainties
are the result of data limitations due to our inability to model phenomena that
cannot be observed in the real world. Other uncertainties can be attributed to
limitations implicit to the modeling approach. To the extent that there is
uncertainty with available data and in the assumptions, there is uncertainty in
the analysis and in the results of the analysis. However, any comprehensive
analysis that attempts to accomplish the same results will face most of these
same limitations.
9.A.2. Uncertainties in the UST Model
9.A.2.a. Uncertain Data Inputs
The ability of the UST model to estimate the extent of the current problem
and to predict the effects of proposed regulations depends, as discussed in
Chapter 2, on the quality of the data used to estimate the failure probabilities
and other model parameters and inputs. The data are imperfect, and so the
model cannot track reality with precision.
In order to simulate the failure probabilities of underground storage tanks
and the resultant effects, or consequences, of leaking tanks, a great deal of
information regarding the current state of the existing tank population is
needed. Due to the size and diversity of the existing UST universe and our
inability to observe the release and transport of product from a leaking tank,
much of the data needed to accurately model tank failures is unavailable. We
do have estimates of the numbers and types of tanks and estimates of tank
locations and densities. By combining this information with engineering esti-
mates of the physical properties of tank materials, the effectiveness of leak
detection, and the behavior of stored product once it is released into the
environment, EPA has developed a model that reasonably simulates actual events.
Because the UST Simulation Model was developed and calibrated using limited
available data and best estimates of probable events, uncertainty is injected
into all areas of the analysis; and therefore, the results of the analysis can
be viewed only as estimates of present and future events and not as exact
reflections, or precise predictions, of these events.
-------
9-2
9.A.2.b. Phenomena Not Addressed by the Model
The modeling process allows us to consider only a limited number of factors.
The simulated tank system is a simplified version of a real system. We
model what we consider to be the most important phenomena; but by excluding
even the most minor of events or the least likely situations, we bias the
probabilities of simulated occurances and increase the likelihood of the model
estimates deviating from the true probabilities of actual events. By using
the best data available and by drawing assumptions from expert knowlege and
best engineering judgement, we can limit the amount of error introduced into the
analysis, but we cannot eliminate it. One advantage of the UST Model is that
it requires that assumptions and data be used in an explicit and consistent
manner. When there is uncertainty about particular assumptions used in the
analysis, we use the model to conduct sensitivity analyses to make the importance
of the uncertainty clear or to provide a range of possible values rather than
a point estimate of a parameter or event.
At present, we use the UST Model to simulate USTs containing gasoline.
We estimate failure probabilities and probable damages assuming all underground
storage tanks contain gasoline. The lack of accurate, or detailed, information
currently prevents us from modeling other petroleum products and hazardous
substances. Other phenomena not covered by the model include: variations in
the location of the leak within the tank (we assume all leaks occur at the
bottom of a tank); early tank retirement not due to tank system problems when
unprofitable stations close; fractured flow aquifers; multi-layer aquifers; and
false positive readings from detection methods other than tightness tests.
Other events which are not modeled explicitly, but are accounted for through the
calibration of model inputs or assumptions include: off-site leak detection due
to sheens on surface water; leak detection by noticing a decrease in a facility's
profitability over months or years, and leak detection by happenstance or by
(human) sensory detection. Each of these small gaps in the model adds to our
uncertainty over the model results and limits the accuracy of our analysis.
9.A.3. Uncertainties in Benefits Analysis
9.A.3.a. Unmodeled Aspects of Fate and Transport
The benefits analysis of the UST regulation is driven by estimates of the
concentrations of contaminants released from leaking tanks over time and at
different distances from the tanks. These estimates, in turn, are based in
part on results from the UST Model, and on modeling techniques and assumptions
outside the scope of the UST Simulation Model. Some of the important phenomena
that affect concentration estimates that are not part of the UST model include:
degradation; interception by surface water; transport of released product
through the atmosphere; exposure assumptions (i.e., the number of wells and water
users found at given distances from USTs); and the prevalence of the use of
less-vulnerable confined aquifers for drinking water. Significant uncertainty
remains over the best way to model these phenomena, meaning that estimates of
exposure concentrations are subject to significant uncertainty.
-------
9-3
9.A.3.b. Toxicological Issues
Another important element of the benefits analysis is the prediction of
health impacts given exposure levels. Though we believe we identified the most
dangerous components of gasoline with respect to carcinogenic potential, there
may be components whose ability to cause cancer is unknown. There may also be
noncarcinogenic effects, as yet unidentified, from gasoline at the doses likely
to be encountered near leaking USTs. To the extent that noncarcinogenic effects
exist, the damages from leaking USTs are underestimated and the benefits of regu-
latory alternatives are, therefore, underestimated in this analysis.
9.A.3.C. Nonhealth Benefits
Even more uncertainty exists over the magnitude and frequency of non-
health damages. These include both the value of preventing tangible property
damage, as well as the value of preventing less tangible non-market damages—
damage to option value, for example. Uncertainties in estimating nonhealth
benefits can be attributed to difficulties in observing nonhealth damages and
in measuring the value, or cost, of such damages. It is difficult to identify
the incremental change in the market value of property due to damages from
leaking USTs. It is also difficult to estimate the value that currently unused
resources may have in the future.
9.A.4. Uncertainties in Economic Impact Analysis
9.A.4.a. Uncertainties in the Regulatory Base Case
Measuring the costs of regulatory alternatives relative to what would
prevail in the absence of federal intervention requires estimating what would
prevail in the base case. The base case is difficult to characterize because
there are thousands of facilities which employe different leak detection and
inventory control practices. This diversity in operating practices introduces
uncertainty with respect to the resources that may be expended in avoiding and
detecting leaks, and in responding to detected leaks in the absence of regulatory
intervention. The degree of uncertainty inherent in estimating market behavior
and future trends in market prices limits our ability to accurately estimate
the costs of the regulatory alternatives.
9.A.4.b. Uncertainties in Regulatory Costs
Cost estimates for regulatory proposals are based on current technologies
and market conditions. However, it is possible that an increased market for leak
prevention, detection, and cleanup technologies and services could induce
changes in market prices. Price changes may occur as a result of: short term
scarcity due to industry capacity constraints in supplying needed products and
services; innovation and diffusion of new technologies; and economies of scale
as demand growth allows fuller utilization of production capacity. Another
uncertainty is the path of future gasoline prices and the prices of other
substances stored in USTs. The current average price of gasoline is used to
measure the value of product loss resulting from leaks.
------- |