INTEGRATED ENVIRONMENTAL
1     DECISION-MAKING IN THE
           21st CENTURY

               EVI EW DRAFI
             May 3,1999
   REPORT FROM THE EPA SCIENCE ADVISORY
     BOARD'S INTEGRATED RISK PROJECT

-------
                      U.S. ENVIRONMENTAL PROTECTION AGENCY
                                SCIENCE ADVISORY BOARD
                                INTEGRATED RISK PROJECT
Steering Committee
  CHAIR
  Or. Genevieve M. Matanoski, The Johns Hopkins University, Baltimore, MD

  MEMBERS
  Dr. Joan M. Oaisey, Lawrence Berkeley Laboratory, Berkeley, CA
  Dr. Paul Deisler (Consultant), Austin, TX
  Dr. Mark A. Harwell, University of Miami,  Miami, PL
  Dr. Wayne Kachel, MELE Associates, Brooks AFB, TX
  Dr. Alan Maki, Exxon Company, USA, Houston, TX
  Dr. Paul R. Portney, Resources for the Future, Washington, DC
  Dr. Milton Russell (Consultant), Joint Institute for Energy and Environment and U. Tenn., Knoxville, TN
  Dr. Ellen Silbergeld, University of Maryland, Baltimore, MD
  Dr. Robert Stavms (Consultant), Harvard  University, Cambridge, MA
  Dr. Paul H. Templet (Consultant), Louisiana State University, Baton Rouge, LA
  Dr. Valerie Thomas (Consultant), Princeton University, Princeton, NJ
  Dr. Bernard Weiss (Consultant), University of Rochester Medical Center, Rochester, NY
  Dr. Marcia Williams (Consultant), Putman, Hayes & Bartlett, Inc., Los Angeles, CA
  Dr. Terry F. Yosie (Consultant), Ruder Finn, Inc., Washington, DC
  Dr. Terry F. Young, Environmental Defense Fund, Oakland, CA

  SCIENCE ADVISORY BOARD STAFF
  Ms. Stephanie Sanzone, Designated Federal Official, US EPA, Science Advisory Board
  Mr. Thomas O. Miller, Designated Federal Official, US EPA, Science Advisory Board
  Ms. Wanda Fields, Management Assistant, US EPA, Science Advisory Board
Ecological Risks Subcommittee
 CHAIR
 Dr. Mark A. Harwell, Rosenstiel School of Marine and Atmospheric Science, University of Miami, Miami, FL

 MEMBERS
 Dr. William Adams, Kennecott Utah Copper Corp, Magna, UT
 Dr. Steven M. Bartell, SENES Oak Ridge, Inc., Oak Ridge, TN
 Dr. Kenneth W. Cummins, South Florida Water Management District, Sanibel, FL
 Dr. Virginia Dale, Environmental Sciences Division, Oak Ridge National Laboratory, Oak Ridge, TN
 Dr. Carol Johnston, Natural Resources Research Institute, University of Minnesota, Duluth, MN
 Dr. Frederick K. Pfaender, Carolina Fed. for Environmental Programs, University of North Carolina, Chapel Hill, NC
 Dr. William H. Smith, School of Forestry and Environmental Studies, Yale University, New Haven, CT
 Dr. Terry F. Young, Environmental Defense Fund, Oakland, California

 SCIENCE ADVISORY BOARD STAFF
 Ms. Stephanie Sanzone, Designated Federal Official, US EPA, Science Advisory Board
 Ms. Wanda R. Fields, Management Assistant, US EPA, Science Advisory Board

-------
Human Exposure and Health Subcommittee
  CO-CHAIRS
  Dr. Joan Daisey, Lawrence Berkeley Laboratory. Berkeley, CA
  Dr. Bernard Weiss, Department of Environmental Medicine, University of Rochester Medical Center, Rochester, NY

  MEMBERS
  Dr. Stephen Ayres, School of Medicine, International Health Programs, Virginia Commonwealth University,
  Richmond, VA
  Dr. Paul Bailey, Mobil Business Resources Corporation, Product Stewardship & Toxicology, Paulsboro, NJ
  Dr. George Daston, Miami Valley Laboratories, The Procter and Gamble Co., Ross, OH
  Dr. Curtis Klaussen, Department of Pharmacology, University of Kansas Medical Center,
  Kansass City, KS
  Dr. Paul Lioy, Environmental and Occupational Health Sciences Institute, Rutgers University, Piscataway, NJ
  Dr. William Pease, Environmental Defense Fund, Oakland, CA
  Dr. Henry Pitot, McArdle Laboratory for Cancer Research, University of Wisconsin, Madison, Wl
  Dr. Jonthan Samet, Department of Epidemiology, Johns Hopkins University, Baltimore, MD
  Dr. Valerie Thomas, Center for Energy and Environmental Studies, Princeton University, Princeton, NJ
  Dr. Lauren Zeise, Office of Environmental Health Hazard Assessment, California Environmental Protection Agency,
  Berkeley, CA

  SCIENCE ADVISORY BOARD STAFF
  Mr. Samuel Rondberg, Designated  Federal Official, U. S. EPA, Science Advisory Board
  Ms. Mary L. Winston, Management Assistant, U. S. EPA, Science Advisory Board
Economic Analysis Subcommittee
 CHAIR
 Dr. Paul R. Portney, Resources for the Future, Washington, D.C.  20036

 MEMBERS
 Dr. Nancy E. Bockstael, Department of Agricultural and Resource Economics, University of Maryland, College Park,
 MD  Dr. Trudy Ann Cameron, Department of Economics, University of California, Los Angeles, CA
 Dr. Maureen L. Cropper, The World Bank, Washington, DC
 Dr. A. Myrick Freeman, Department of Economics, Bowdoin College, Brunswick, ME
 Dr. Charles D. Kolstad, Department of Economics, University of California, Santa Barbara, CA
 Dr. Robert Repetto, Economic Research Program, World Resources Institute, Washington, D. C.
 Dr. Robert N. Stavins, John F. Kennedy School of Government, Harvard University, Cambridge, MA
 Dr. Thomas H. Tietenberg, Dept. of Economics, Colby College, Waterville, ME
 Dr. W. Kip Viscusi, Harvard Law School, Cambridge, MA

 SCIENCE ADVISORY BOARD STAFF
 Mr. Thomas Miller, Designated Federal Official, U.S. EPA, Science Advisory Board
 Ms. Diana L. Pozun, Management Assistant, U.S. EPA, Science Advisory Board

-------
Valuation Subcommittee
  CO-CHAIRS
  Dr. Alan W. Maki, Exxon Company, USA, Houston, TX
  Dr. Milton Russell (Consultant), Joint Institute for Energy & Environment and U. Tennessee, Knoxville, TN

  MEMBERS
  Dr. Stephen M. Ayres, Virginia Commonwealth University, Medical College of Virginia, Richmond, VA
  Dr. Nancy E. Bockstael, Dept. of Agricultural and Resource Economics, University of Maryland, College Park, MD
  Dr. Caron Chess (Consultant), Center for Environmental Communications, Rutgers University, New Brunswick, NJ
  Dr. Virginia Dale, Environmental Sciences Division, Oak Ridge National Laboratory, Oak Ridge, TN
  Dr. William H. Desvousges (Consultant), Triangle Economic Research, Durham, NC
  Dr Thomas Dietz (Consultant), Department of Sociology and Anthropology, George Mason University, Fairfax, VA
  Dr. A. Mynck Freeman, Department of Economics, Bowdom College, Brunswick, ME
  Dr. Mark A. Harwell, Rosenstiel School of Marine and Atmospheric Science, University of Miami, Miami, FL
  Professor Jerry A. Hausman (Consultant), Department of Economics, Massachusetts Institute of Technology,
  Cambridge, MA
  Dr. Douglas E. MacLean (Consultant), Department of Philosophy, University of Maryland, Baltimore, MD
  Dr. John W. Payne (Consultant), Fuqua School of Business, Duke University, Durham, NC
  Dr. Edella Schlager (Consultant), School of Public Administration and Policy, University of Arizone, Tucson, AZ
  Dr. Margaret Shannon (Consultant), Center for Environmental Policy and Administration, Syracuse University,
  Syracuse, NY
  Dr. Paul Templet (Consultant), Institute for Environmental  Studies, Louisiana State University, Baton Rouge, LA
  Dr. Terry F Young, Environmental Defense Fund, Oakland, California 94611
  Dr. James Wilson (Consultant), Department of Resource Economics and Policy, University of Maine, Orono, ME

  Science Advisory Board Staff
  Mr. Thomas Miller, Designated Federal Official, U.S. EPA, Science Advisory Board
  Ms. Diana Pozun, Management Assistant, U.S. EPA, Science Advisory Board
Risk Reduction Options Subcommittee
 CO-CHAIRS
 Dr. Wayne M. Kachel, MELE Associates, Brooks AFB, TX
 Ms. Marcia Williams, (Consultant), Putman, Hayes & Bartlett, Inc., Los Angeles, CA

 MEMBERS
 Ann Bostrom (Consultant), School of Public Policy, Georgia Institute of Technology, Atlanta, GA
 Ms. Dorothy P. Bowers (Consultant), Environmental and Safety Policy, Merck & CO., Inc., Whitehouse Station, NJ
 Mr. Robert Frantz (Consultant), Corporate Environmental Programs, General Electnc Company,
 Dr. Nina Bergan French, SKY+, Oakland, CA
 Ms. Mary A. Gade (Consultant), Illinois Environmental Protection Agency, Springfield, IL
 Mr. Bradford S. Gentry (Consultant), Yale University, The Center for Environmental Law & Policy, New Haven, CT
 Dr. Ricardo R. Gonzalez (Consultant), Department of Radiological Sciences, Univ. of Puerto Rico School of
 Medicine, San Juan, PR
 Dr. Michael Greenberg (Consultant), The State University of New Jersey, Rutgers, Department of Urban Studies &
 Community Health, New Brunswick, NJ
 Dr. Linda E. Greer (Consultant), Natural Resources Defense Council, Washington, DC
 Dr. Hilary I. Inyang, Center for Environmental Engineering and Sciences Technologies (CEEST), University of
 Massachusetts, Lowell, MA
 Dr. Charles D. Kolstad, University of California, Department of Economics, Santa Barbara, CA
 Terrence J. McManus (Consultant), Corporate Environmental Affairs, Intel Corporation, Chandler, AZ
 Dr. Wm. Randall Seeker, Energy & Environmental Research Corp., Irvine, CA 92718

 SCIENCE ADVISORY BOARD STAFF
 Ms. Kathleen W. Conway, Designated Federal Official, U.S. EPA, Science Advisory Board,
 Ms. Dorothy M. Clark, Management Assistant, U.S. EPA, Science Advisory Board

-------
   IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

     INTEGRATED ENVIRONMENTAL DECISION-MAKING  IN THE 21st
                                  CENTURY

                             TABLE OF CONTENTS

PART I — THE FRAMEWORK

1.  Proposed Framework for Integrated Environmental Decision-Making
      1.1 Integrated Environmental Decision-Making	  1-1
      1.2 Scope of the Project	  1-8
      1.3 A Proposed Framework for Integrated Environmental Decision-Making	  1-12
      1.4 Nature of the Framework  	  1 -23
      1.5 Benefits and Challenges of the Framework 	  1-33
      1.6 References Cited 	  1-36

PART II — INPUTS TO ENVIRONMENTAL DECISION-MAKING: RISK COMPARISONS

Preface
2.  Ecological Risks
      2.1 Background	2-1
      2.2 Objectives and Approach  	2-3
      2.3 ERS Ecological Risk Ranking Methodology 	2-4
      2.4 National-Scale Ecological Risk Ranking	  2-23
      2.5 An Effects-Backwards Methodology for Risk Rankings	  2-32
      2.6 References Cited 	  2-36
      Appendix 2A. Ecological Risk Profiles	2-38

3.  Human Health Risks
      3.1 Introduction	  3-1
      3.2. The Environmental Health Risk Rating Methodology  	3-4
      3.3 Analysis and Reporting of Relative Risk Rating Survey Data	  3-16
      3.4 Correspondence Between Ecological and Health Risk Formats	3-18
      3.5. Implications of Ratings	  3-19
      3.6 A Fuzzy Logic Approach	  3-20
      3.7 Extensions and Refinements of the Methodology	  3-25
      3.8 Summary and Conclusions	  3-26
      3.9 References Cited 	  3-27
      Appendix 3A. Health Risk Assessment Introduction  	  3-28
      Appendix 3B. Instructions  	  3-30
      Appendix 3C. Risk Characterization Data Sheets  	  3-33

PART III — INPUTS TO ENVIRONMENTAL DECISION-MAKING: ECONOMICS AND
VALUATION

Preface
4.  Benefit-Cost Analysis for Integrated Risk Decisions
      4.1 Introduction  	4-1
      4.2 Fundamental Questions in the Economic Analysis of Risk	4-3
      4.3. The Benefits of Risk Reduction	4-8
      4.4 Costs of Environmental Protection 	  4-17

-------
   IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

      4.5 Comparing Total Benefits and Total Costs	 4-22
      4.6 Distributional Considerations	 4-27
      4.7 Conclusions	 4-28
      References Cited	 4-31
      Endnotes	 4-33

5. Assessing the Value of Natural Resources
      OVERVIEW	5-1
      5.1 Introduction	5-4
      5.2 Valuation and the Decision Context 	5-6
      5.3 The Nature of Values 	 5-14
      5.4 The Economic Valuation Framework  	 5-20
      5.5 The Importance of Deliberative Processes to Valuation	 5-26
      5.6 Additional Approaches to Valuation of Environmental Systems  	 5-37
      5.7 Summary and Conclusions	 5-47
      5.8 References Cited 	 5-52

PART IV — INPUTS TO ENVIRONMENTAL DECISION-MAKING: RISK REDUCTION
APPROACHES

Preface
6. Risk Reduction Options
      6.1 Introduction and Approach  	 6-1
      6.2 Define the Problem	 6-6
      6.3 Develop Background Information	 6-12
      6.4 Identify the Spectrum of Risk Reduction Options	 6-16
      6.5 Establish Screening and Prioritization Criteria  	 6-33
      6.6 Screen and Prioritize Potential Risk Reduction Options 	 6-40
      6.7 Evaluate the Remaining Risk Reduction Options  	 6-48
      6.8 Optimize the Options 	 6-54
      6.9 Select an Option	 6-59
      6.10 Document the Process	 6-65
      6.11  Quantify Option Effectiveness	 6-65
      6.12 References Cited  	 6-71

PART V — IMPLEMENTATION AND PERFORMANCE EVALUATION

Preface
7. Performance Evaluation — The Design and Use of Environmental Report Cards
      7.1 Introduction  	 7-1
      7.2 Types of Performance Measures 	 7-3
      7.3 Improved Report Cards	 7-8
      7.4 Implications for Existing Monitoring Systems 	 7-17
      7.5 Summary and Recommendations	 7-19
      7.6 References Cited 	 7-22

PART VI — CONCLUSIONS AND RECOMMENDATIONS

8. Conclusions and Recommendations	8-1

APPENDIX: Summary of a Hypothetical Case Example

-------
PARTI THE FRAMEWORK

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote


 1
 2           CHAPTER 1.  PROPOSED FRAMEWORK FOR INTEGRATED
 3                     ENVIRONMENTAL DECISION-MAKING
 4
 5                             TABLE OF CONTENTS
 6
 7
 8      1.1 Integrated Environmental Decision-Making 	1-1
 9           1.1.1  The Call for Integrated Decision-Making 	1-1
10           1.1.2 Signs of Progress	1-3
11
12      1.2 Scope of the Project  	1-8
13
14      1.3 A Proposed Framework for Integrated Environmental Decision-Making	1-12
15           1.3.1  Overview	1-12
16           1.3.2 Phase I: Problem Formulation	1-14
            1.3.3 Phase II: Analysis and Decision-Making	1-17
lu           1.3.4 Phase III: Implementation and Performance Evaluation	1-20
19
20      1.4 Nature of the Framework	1-23
21           1.4.1  Major Characteristics	1-23
22           1.4.2 Types of Integration	1-25
23           1.4.3 Building on Previous Frameworks	1-29
24
25      1.5 Benefits and Challenges of the Framework	1-33
26          1.5.1  Benefits	1-33
27          1.5.2 Challenges	1-34
28
29     1.6 References Cited	1-36
30

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote


 1           CHAPTER 1. PROPOSED FRAMEWORK FOR INTEGRATED
 2                      ENVIRONMENTAL DECISION-MAKING
 3
 4
 5     1.1 Integrated Environmental Decision-Making
 6
 7      1.1.1 The Call for Integrated Decision-Making
 8
 9           Environmental decision-making has progressed over the years with the accretion
10     of experience and knowledge, becoming ever more subtle, inclusive, and powerful.
11     Despite criticisms leveled against them, and despite inherent limitations, previous
12     environmental decisions and policies have spurred significant environmental and health
13     progress. First, the decisions have addressed a multitude of complicated
14     environmental problems. Second, they have offered broad perspectives on how we
15     formulate and support environmental research. Decision-makers currently draw upon
16     an eclectic mixture of tools and information to inform their decisions: ecological and
       human health risk assessment; benefit/cost and cost-effectiveness models; expanded
       risk communication and public participation; and measures for monitoring the results of
19     the decisions themselves. Although these tools provide essential inputs for decision-
20     making, they have typically been applied unevenly and to relatively narrow issues. In
21     the area of human health risks, for example, assessments often have been framed
22     around single stressors or classes of stressors in relatively specific exposure situations.
23     The deficiencies of such highly focused assessments are increasingly apparent; they
24     sidestep the complexities, interrelationships, and subtleties of environmental problems
25     as they actually confront us.  Thus, a number of recent studies have urged the
26     Environmental Protection Agency (EPA) to begin to address environmental issues in a
27     more integrated way (e.g., NAPA, 1995; Presidential/Congressional Commission on
28     Risk Assessment and Risk Management, 1997).
29
30           Much of the fragmentation in EPA's approach to the control of environmental
31     problems has its roots in the statutory framework that guides the work of the Agency.
32     From its formation in  1970, the EPA has been given responsibility for implementing a
33     number of environmental statutes that mandate targeted actions to control specific
34     pollutants in specific media (e.g., Clean Air Act language regarding particulates in air) or
                                           1-1

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1     specific routes of exposure (e.g., Safe Drinking Water Act language regarding priority
 2     pollutants in drinking water).  The focus on assessing and controlling chemical
 3     contaminants pollutant by pollutant in single media has resulted in an evolving
 4     collection of federal laws and regulatory requirements that is neither systematic nor
 5     comprehensive.  Nonetheless, these laws have been largely successful in controlling
 6     many of the targeted pollutants and have provided a strong national underpinning for an
 7     effective environmental protection program  comprised of federal, state, and local
 8     controls.
 9
10           Yet, despite these successes, there is a growing consensus, both within and
11     outside the Agency, that a more integrated  approach to environmental management is
12     needed. Prioritizing and managing risks pollutant by pollutant and medium by medium
13     can be both inefficient at reducing the major burdens of environmental impacts on
14     human health and ecosystems and costly in the face of today's shrinking budgets. Of
15     still greater concern is the possibility that  such a fragmented approach may cause us to
16     overlook significant environmental problems while we busy ourselves with
17     comparatively minor issues that contribute little to the overall protection of human
18     health and ecosystems.  Further, in some instances, current statutes and regulations
19     prevent the Agency from considering all relevant risk, benefit/cost, or other information.
20     A 1995 report from the National Academy of Public Administration (NAPA) pointed out
21     that there are "no established criteria that the Agency might use to set priorities that cut
22     across statutory lines" and called on Congress and the Agency to give serious thought
23     to an "integrated statute that would provide multi-media decision-making authority to the
24     Agency" (NAPA, 1995). The SAB views the issue of statutory integration as a policy
25'    discussion and outside the bounds of the present study.  Even within the current
26     statutory framework, however, there are numerous opportunities for a more holistic
27     assessment of risks and risk management options, and more inclusive decision-making
28     approaches.
29
30           The call for integrated assessment of risks has included the need to consider
31     multiple sources, multiple routes of exposure, and multiple human health endpoints
32     (e.g., cancer, genetic effects, and developmental and reproductive toxicity) (see, for
33     example, NRC, 1994), and aggregate risks posed by multiple agents or stressors  (e.g.,
34     endocrine disruptors or mixtures of polycyclic aromatic hydrocarbons). Although many
35     Agency risk assessments still focus on a single agent (e.g., lead, 1,3-butadiene), there

                                              1-2

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1      clearly has been an evolution in risk assessment methods towards more realistic,
 2     multiple source, multiple pathway, multiple agent assessments.
 3
 4      1.1.2 Signs of Progress
 5
 6           In the 1983 publication entitled Risk Assessment in the Federal Government:
 7     Managing the Process (NRC, 1983), commonly referred to as the "Red Book", an NRC
 8     panel laid out the elements of risk assessment and risk management using terminology
 9     that came to be the standard. These concepts were adopted by EPA Administrator
10     Ruckleshaus in 1984 and have formed the basis for much of the Agency's action to this
11      day.  In summary, the NRC committee described the four steps of risk assessment as
12     hazard identification, dose-response assessment, exposure assessment, and risk
13     characterization, which was defined as "the estimated incidence of the adverse effect in
14     a given population" (NRC, 1983) (Figure 1-1).  In addition, the NRC committee stressed
15     the scientific basis for risk assessment and the need for both quantitative and
16     qualitative expressions of risk. Risk management was viewed as "a decision-making
17     process that entails consideration of political, social, economic, and engineering
       information with risk-related information to develop, analyze, and compare regulatory
19     options and to select the appropriate regulatory response..." (NRC, 1983).
20
21           The Red Book was extremely useful in articulating the risk assessment process
22     and its relationship to risk management. The paradigm was expressed, however, in
23     terms of single agents and single health effects in humans.  Since that time, the Agency
24     has developed risk assessment guidelines to address a number of endpoints (i.e.,
25     cancer, reproductive and  developmental toxicity, and neurotoxicity), as well as exposure
26     assessment, which is a component of the risk assessment model.  In addition, the
27     Agency has taken  steps to consider more integrated exposure scenarios, e.g., multi-
28     route exposures to mixtures of chemical agents associated with Superfund sites (U.S.
29     EPA, 1989) or combustor emissions (EPA, 1990; 1993a), and, in response to the Food
30     Quality Protection  Act, to  consider multiple pesticides and multiple  routes of exposure in
31     assessing children's risk.
32
33
                                            1-3

-------
                                                                                      [FILEAREDBOOK.PDQ]
         RESEARCH
         RISK ASSESSMENT
                         RISK MANAGEMENT
       LaBoratory ind lulu
      ootervations ol advaise
       Main enacts ind
      eipsoiuies 10 oartcHar
          iginii
 Hizara Identrfciuoir
 (Deal IM agent causa
 IM adverse efteef)
        Inloimaien en
     exnaeoiauon tnatnots loi
      (ugn 10 ton dosa and
        inmal 10 luiraan
  Ooia-Raipona
   As!»sjm«nc
(Whal is IM ralannshp
  baiwaan easa and
 ncdanca « Humans1)
      FiaH maasuramants.
      astmaiao aiposuras.
       charactanann ol
         pogdaieia
 Eiposure Aisasunanc
 (Wisai eiposum are
 cunanfly eipnancad at
anueipalad under tittaren
    conditions')
                                                      \
                                 Oatttopmem el
                                regulatory option!
 Rbk Chaiactaroauir
 (What s ma asmaitd
nodanca ol tne adnrsa
  allact in a gwan
   populauoni)
                                Evahaton ol pubic
                               rieitnt. aconomc. socal
                               poSlcal consaquaneaa ol
                                                               \
                                                                                             Agancy oecawns ina
                                                                                                 actions
                                     Elements of risk assessment and risk management.
               \-\
                                                        From:

 National Research Council, Commission  on Life Sciences, Committee on the Institutional Means for
       Assessment of Risks to Public Health. 1983.  Risk Assessment in the Federal Government:
                                                     Managing
the Process. National Academy Press. Washington, D.C.  191pp.
                                                  I-H

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1           The Agency has also made significant progress in adapting the Red Book
 2     paradigm to ecological risk assessment. In 1992, the Agency released its Framework
 3     for Ecological Risk Assessment (U.S. EPA, 1992) which used the term "characterization
 4     of ecological effects" to include both  hazard identification and exposure assessment.
 5     The Framework also added an explicit Problem Formulation phase prior to the analysis
 6     of exposure and effects to emphasize the importance of articulating the problem and a
 7     plan for analyzing and characterizing risk prior to conducting specific risk analyses.  The
 8     resulting framework contained three  phases: Problem Formulation, Analysis, and Risk
 9     Characterization (Figure 1-2). An expanded discussion of ecological risk assessment
10     principles and approaches was subsequently provided by the Agency  in final Guidelines
11     for Ecological Risk Assessment (U.S. EPA, 1998). The guidelines note that "although
12     ecological risk assessments provide  critical information to risk managers, they are only
13     part of the environmental decision-making process" (U.S. EPA, 1998). In addition to
14     assessing the relationship between a particular stressor and a particular effect, the
15     ecorisk guidelines set the stage for considering multiple effects (including cascading
16     effects) associated with a single stressor or source, as well as multiple causes of an
17     observed effect or change in ecological condition.  The Agency has already applied the
       ecological risk assessment paradigm, including the development of a conceptual model
19     relating various stressors and effects, to five watershed cases (for discussion, see SAB,
20     1997).
21
22           Three additional Agency developments merit brief mention here: a) guidance
23     and support for comparative risk analysis;  b) extra-statutory approaches to
24     environmental protection; and c) guidance on planning and scoping for cumulative risk
26     assessment.
26
27           First, Comparative Risk Analysis (CRA) has been defined by the Agency as "both
28     an analytical process and a set of methods used to systematically measure, compare,
29     and rank environmental problems" (U.S. EPA, 1993b).  The Agency, in its Unfinished
                                             1-5

-------
                   Ecological Risk Assessment
   Planning
 (Risk Assessor/
  Risk Manager/
Interested Parties
    Dialogue)

                                             it*?£feiS?S&£::
                                        Charactenzstio n
                                             of
                                          Ecological
                                            Effects
                Characterization
                     of
                   Exposure
                     RJSKLCHARACTERJZATIO
                                       A
                                       T
                              Communicating Results
                               to the Risk Manager
                            ^:~]Risk Management and "/
                            •.Communicating Results to
                            i-..^Interested Parties : :"/:..
Figure]
1992a).
The framework for ecological risk assessment (modified from U.S. EPA,
                                    f    -

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1     Business report (U.S. EPA, 1987), and the SAB, in Reducing Risk (SAB, 1990),
 2     engaged in comparative risk analyses. In its 1990 report, the Board concluded that it
 3     was possible, on a scientific basis, to distinguish between large risks and small risks
 4     using a set of technical criteria.  In the years that followed, the Agency promoted the
 5     wide use of CRA as a process for setting priorities by integrating multiple stressors and
 6     multiple types of risks within whole regions such as cities, states or even the nation
 7     itself.  Comparative risk analysis is intended principally as a policy-development and
 8     broad resource-allocation tool. In contrast to Unfinished Business and Reducing Risk,
 9     however, state and local-level comparative risk analyses have highlighted the role of the
10     public and stakeholder groups, in addition to the scientific/technical community, in
11     defining risk priorities.  Support for broader inclusion of public values in decision-making
12     is a theme that has been echoed by a number of recent reports (e.g., NRC, 1996;
13     Presidential/ Congressional Commission on Risk Assessment and Risk Management,
14     1997).
15
16           Second, during the 1990s, the Agency has experimented with a number of
17     approaches to re-inventing environmental protection, including greater use of
       community-based decision-making, voluntary cross-media emissions reductions (the
19     33/50 project; reference), integrated environmental agreements with states (National
20     Environmental Performance Partnership agreements), and voluntary regulatory reform
21     efforts with an array of stakeholders (e.g., Common Sense Initiative, Project XL).
22
23           Third, the Agency has recently issued guidance directing program offices to
24     "consider a broader scope that integrates multiple sources, effects, pathways,
25     stressors, and populations for cumulative risk analyses in all cases for which relevant
26     data are available" (U.S. EPA, 1997). The cumulative risk guidance also notes that on-
27     going Agency efforts to involve stakeholders "will provide the solid basis for engaging
28     interested and affected parties in risk assessment and risk management issues"
29
30           The SAB has not reviewed the role and adequacy of the science being used in
31     the planning and evaluation of these activities, including CRA. These endeavors are an
32     indication of movement in the Agency toward more integrated and inclusive methods of
33     environmental decision-making.  These experiments, however, do not yet represent the
34     mainstream of EPA's efforts.
35

                                             1-7

-------
        IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1     1.2 Scope of the Project
 2
 3             It is in this atmosphere that the SAB undertook the task of revisiting its 1990
 4     report, Reducing Risk, to update and extend the thinking about how science can best
 5     inform the decision-making process.  The Charge to the SAB from the Agency included
 6     requests to:
 7
 8            a) update the risk rankings in Reducing Risk using explicit scientific criteria and
 9            the judgments of SAB panel members;
10            b) identify risk reduction opportunities and strategies;
11            c) identify uncertainties and data quality issues associated with the risk rankings;
12            d) assess costs and benefits of risk reduction options; and
13            e) propose a new framework for assessing the value of  ecosystems.
14
15            The initial charge also included a request that the SAB explore techniques and
16     criteria for identifying emerging risks. However, the SAB concluded that its recent
17     report, Beyond the Horizon: Using Foresight to Protect the Environmental Future (SAB,
18     1995) provided criteria and suggestions germane to this charge question and so did not
19     elaborate further on the future risks as part of the Integrated Risk Project.
20
21            After careful consideration of the Charge and discussion with Deputy
22     Administrator Fred Hansen, the SAB concluded that it could best assist the Agency by
23     investigating approaches for accomplishing these goals and considering the inter-
24     relationship of these tasks in the broader decision-making context.
26
26            The project, known as the Integrated Risk Project, was  guided by a Steering
27     Committee and five specialized Subcommittees working over several years. The
28     Subcommittees and their respective charges were as follows:
29
30            a)     The Steering Committee (SC), chaired by Dr. Genevieve Matanoski, set
31                  the overall direction for the project by defining scope and timetables.  The
32                  SC met periodically over the course of the project to: (1) assess the
33                  progress and direction of the subcommittees; (2) ensure that the results
34                  could be integrated into a comprehensive decision process for identifying
35                  current and future environmental risks; and (3) review options for reducing

                                             1-8

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1                 risks in a holistic context. The SC's efforts were designed specifically to
 2                illustrate the relationship among the various factors influencing risk
 3                management decisions (e.g., technical assessment of the risks  and risk
 4                reduction options, economic considerations, equity considerations, and so
 5                forth).
 6
 7          b)    The Ecological Risks Subcommittee (ERS), chaired by Dr. Mark
 8                Harwell, was charged to assess and rank risks to ecosystems at the
 9                national scale, as well as to suggest ways in which the risk ranking
10                methodology could be applied at smaller geographical scales (e.g.,
11                 regional, state, or local).  The group was also asked to explore
12                commonalities and differences with the Human Exposure and Health
13                Subcommittee (HEHS) methodology with the aim of integrating  the two
14                ranking schemes.
15
16          c)    The Human Exposure and Health Subcommittee (HEHS), co-chaired
17                by Drs. Joan Daisey and Bernard Weiss, was charged to develop a
                  methodology for assessing and ranking risks to human health, to consider
19                ways in which an integrated risk ranking could be produced that includes
20                both cancer and non-cancer risks, and to test the methodology  for a
21                limited set of environmentally mediated health issues. The Subcommittee
22                was also asked to explore commonalities and differences with the ERS
23                methodology.
24
25          d)    The Risk Reduction Options Subcommittee (RROS),  co-chaired by Dr.
26                Wayne Kachel and Ms. Marcia Williams, was charged with developing a
27                methodology for selecting an optimal set of risk reduction options with due
28                regard for the human health and ecological risks (defined in terms of risks
29                associated with environmental stressors, locations, or exposure/transport
30                media).  Because of the time constraints on the project, the RROS was
31                asked by the SC only to illustrate the methodology for one or more
32                example problems in lieu of addressing the wider range  of risks
33                considered by the ERS and the HEHS.
34
35

                                            1-9

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1           e)    The Economic Analysis Subcommittee (EAS), chaired by Dr. Paul
 2                 Portney, was charged with assessing current methods for estimating costs
 3                 and benefits (either physical or monetary) associated both with the
 4                 implementation of risk reduction strategies and with allowing risks to go
 5                 unaddressed.  The EAS was also asked to consider those aspects of the
 6                 "net benefits" equation that cannot easily be monetized.
 7
 8           f)     The Valuation Subcommittee (VS), co-chaired by Drs. Alan Maki and
 9                 Milton Russell, was charged to consider a new framework for assessing
10                 the value of ecosystems to humans, including ecological services and
11                 environmentally mediated health and quality of life values. The work of
12                 the VS was intended to provide a wider societal view of risk and risk
13                 reduction options than that derived from science-based risk assessments
14                 and current methods of economic analysis.
15
16           Over the course of the project, the IRP SC and subcommittees held over 25
17     public meetings and teleconference calls. Although most of these meetings were held
18     in Washington, D.C., public sessions were also held in Berkeley and San Francisco,
19     CA; Atlanta, GA; New Orleans, LA; and Baltimore, MD.
20
21           The conceptual model that emerged-^the framework for Integrated
22     Environmental Decision-making (IED)—is one that emphasizes the dialogue and
23     interaction between risk assessors, risk reduction options analysts (e.g., engineers,
24     economists, and environmental law/policy experts),  decision-makers, and the public.
26     This report describes the IED framework for making integrated decisions, and
26     recommends needed improvements to the tools and techniques required to both
27     implement and evaluate IED.
28
29           Throughout the study, the Steering Committee used stressors as a common
30     focus to link the components of the framework because stressors are the actual
31     physical, chemical, or biological changes that  affect ecological or human systems, and
32     thus, will be the focus of risk reduction actions. However, the IED framework provides a
33     multi-dimensional approach  to problem definition and options analysis, one that
34     recognizes the linkages between stressors, sources, pathways of exposure, and
35     adverse effects endpoints. It is this multi-dimensional approach that can lead to the

                                            1-10

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1     most effective overall reduction of risk.
 2
 3           The SAB, as a body of scientists, engineers, and economists, is best suited to
 4     describing the scientific and technical analyses that should inform decision-making.
 5     The project focused, therefore, on methods for assessing and comparing multiple risks
 6     to human health or the environment; an approach for designing a risk reduction
 7     program; qualitative and quantitative benefit/cost analyses; and considerations in the
 8     design of performance evaluation systems. With regard to those aspects of the IED
 9     process that rely on the application of public values (e.g., the selection of specific risk
10     goals and tradeoffs, and decision criteria), the SAB cannot represent the public.
11     However, the report does discuss the importance of public values to decision-making
12     and suggests ways in which EPA, and other decision-making bodies, can utilize
13     deliberative processes to engage stakeholders and the public in aspects of
14     environmental decision-making.
15
16           This report represents the SAB's broadest view of environmental decision-
17     making to date.  The proposed framework is designed as a flexible guide that can be
       used to address environmental problems of different size, scope, and location. The
19     effort to develop the IED has required the SAB to venture into risk management areas
20     that do not usually arise in most of its work. However, while  methods for reaching
21     decisions about regulatory matters or policies are part of the IED framework, the SAB
22     offers no recommendations directly impinging on specific decisions or policies.
23
24
                                            1-11

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1     1.3 A Proposed Framework for Integrated Environmental Decision-Making
 2
 3       1.3.1 Overview
 4
 5           In order to encourage a more integrated approach to environmental protection,
 6     the SAB proposes a framework for decision-making based on the recognition that
 7     requires in-depth analysis of projected risk reduction under possible management
 8     scenarios and selection of a preferred scenario, based on criteria such as feasibility,
 9     cost-effectiveness, seriousness of the risks addressed, and equity.  The final phase of
10     IED—Implementation and Performance Evaluation—is one in which the implementation
11     of risk reduction measures occurs and environmental results are monitored and
12     evaluated. Performance Evaluation provides critical feedback so that management
13     approaches can be fine-tuned and the extent and nature of remaining risks, and the
14     means for reducing those risks if necessary, can be re-evaluated.
15
16           The complete IED framework is an integrative scheme for making decisions
17     where many different variables, often interacting across physical, regulatory, and
18     organizational boundaries, can be considered simultaneously rather than in isolation by
19     the many types of participants. It allows for:  a) the consideration of related clusters of
20     risks; b) the development of multiple risk reduction options; c) the definition of markers
21     for evaluating progress toward specific environmental goals; and d) consideration of
22     public preferences and values throughout the process.
23
24           Although the IED process requires the involvement of a broad spectrum of
25'    participants (e.g., scientists, engineers, economists, decision-makers, and the public),
26     the different groups have unique roles to play. In other words, the framework does not
27     imply that "everyone must be involved in everything all the time." For example, just as
28     scientists cannot provide the perspective of the general public, members of the general
29     public cannot do the job of scientists.  Decision-makers, after considering the various
30     sorts of information (data, views, and judgments) generated by the problem formulation
31     and analysis, must make the decision. The important message of the IED  framework is
32     that the various groups involved must maintain effective communication with each other
33     throughout the process so that each can most effectively do its job within the overall
34     context.
35

                                             1-12

-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
       /Information

       Expert
       Judgment

       Values
      /Information

       Expert
       Judgment

       Values

       Legal and
       Institutional
       Milieu
                       Figure 1-3.  Integrated Environmental
                            Decision-Making Framework
       PROBLEM FORMULATION

What are the most important environmental risks?
      What are our environmental goals?
         /•—
          Risk     1} Goal Setting
        vCompansons\
              ^ 	
               Preliminary
             Options Analysis
                                              t
 ANALYSIS AND DECISION-MAKING

What are the best risk reduction opportunities?
How can we achieve our goals and objectives?
                                 IMPLEMENTATION and
                              PERFORMANCE EVALUATION

                                    How are we doing?
                          /frnplementatiorN /Monitoring  \/Information A
                          t           }\and Reporting JI Evaluation J
                                                                          REPC
                                                                          CAR
Is the nature
of the problem
changing?
                                                                     RE
                                                                     CA

                                                                     Are
                                                                     our
                                             »ORT
                                             RD

                                             we meeting
                                             ofyecf/Ves?
     RT
                                        1-13

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1           As illustrated in Figure 1-3, the IED framework is intended to answer a series of
 2     straightforward questions. What are the most important environmental risks? What are
 3     our environmental goals? What are the best risk reduction opportunities? How can we
 4     achieve our goals and objectives? How will we know whether or not we are meeting our
 5     goals? What modifications in our approach are needed to improve environmental
 6     results? Finding answers to these fundamental questions requires application of
 7     scientific and technical assessment and analysis techniques, as well as political, policy,
 8     and values-driven choices.  The following sections describe the IED process, including
 9     the use of the various analytical methods.
10
11       1.3.2  Phase I: Problem Formulation
12
13           The initial phase of the IED framework is Problem Formulation, in which
14     agreement should be reached among all participants—risk assessors, risk managers,
15     and interested and affected parties—about what needs to done by whom and why, if
16     not how. As shown in Figure 1-3, Phase  I includes primarily three related tasks: Risk
17     Comparisons, Goal Setting, and Preliminary Options Analysis.  In this initial Phase, the
18     discussions are at the level of planning, scoping, and screening, rather than the
19     detailed analyses conducted in Phase II.
                                            1-14

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote


     Figure 1-4. Risk Comparisons	
              The term Risk Comparison is used in this report to denote the characterization and ranking of risks
       posed by environmental stressors, where an environmental stressor is any physical, chemical, or biological
       change or agent that could affect ecological or human health systems. In the case of a single stressor, this
       analysis consists essentially of risk assessment approaches that have already been established for human
       health risks (NRC, 1983) and for ecological risks (EPA 1992,1998). The information resulting from this
       component is a characterization of the nature and magnitude of the risks posed by the  stressor and
       characterization of the systems at risk and elements of systems that may be exposed to the stressor.

              In the multi-stressor case, it is important to consider a wide range of environmental risks
       simultaneously so that the seriousness of nsks can be characterized relative to one another. This may include
       comparison of  nsks across different stressors affecting human health, ecosystems, and/or quality of life. Risk
       comparisons may be done by comparing quantitative estimates of nsks where that is possible, or by qualitative
       evaluation of nsks against some set of criteria. While quantitative risk comparisons are ideal, qualitative
       compansons may be best suited when comparing non-commensurate risks (e.g., nsks  to humans compared to
       risks to ecosystems). Chapters 2 and 3 descnbe approaches to risk comparisons designed to  utilize the best
       available scientific information and expert judgment to categonze nsks (e.g., high, medium, and low) and make
       transparent the influence of any value judgments or tradeoffs. The two approaches can be used to rank risks
       within, but not across, the categories of human health or ecological nsks. This is because the ranking factors,
       while analogous, differ in specifics for the two types of nsks.

              The I ED framework also allows for the explicit assessment and companson of nsks to  quality of life
       (QOL), often defined as potential non-health impacts on humans from environmental change. Examples of
       nsks often included in this category are aesthetic, economic, and equity impacts, as well as effects on peace of
       mind, cultural or community identity, and recreational opportunities. Although some of  these QOL risks can be
       assessed via analysis of ecological risks or benefit/cost, the SAB did not specifically propose a method for
       ranking QOL risks because the selection of QOL ranking criteria is largely a value-dnven, rather than scientific.
       process and as such is more appropriately conducted by a broad group  of public or stakeholder
       representatives rather than by a technical panel. However, the EPA (1993) has developed a guidebook for
       assessing quality of life risks that provides a starting point for such assessments.  In most of the state
       comparative risk projects, risks to human welfare or quality of life have been considered by a separate, non-
       technical subcommittee that developed criteria and produced a ranked list of QOL risks.

              The SAB also does not propose a formal process for merging the ecological and human health risk
       issues into a single ranked list  because such an activity appropriately would require consideration of many
       non-technical issues, including political acceptability and societal values. However, the development of a
       single pnoritized list of risks of concern, including nsks to ecosystems, human health, and quality of life, can be
       developed with public/values deliberation and is an appropriate activity during Problem Formulation.
1              Risk comparison methods (see Figure 1 -4) are used to identify the sets of risks
2      to ecological, human health, and/or quality of life systems that will be the subject of
3      detailed consideration in Phase II.  Preliminary analysis of risk reduction options is
4      important in Phase I in order to formulate the problem in terms that will be amenable to
5      the greatest overall risk reduction.  Problem Formulation requires considerable dialogue
6      between the participants comparing risks and those involved with preliminary
7      identification of available options. This dialogue is designed to focus the problem and


                                                     1-15

-------
        IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1     to ensure consistency between the stressors and risks being considered and the risk
 2     reduction opportunities that may be available.
 3
 4            Planning, scoping, and screening - including selection of endpoints - also
 5     requires explicit input of societal values and stakeholder participation.  For example,
 6     while some of the ecological endpoints that are selected in the problem formulation
 7     phase may be chosen strictly because of the value attached to their ecological role,
 8     there are also ecological endpoints that will be chosen because of their particular
 9     significance to society. Examples of ecological endpoints in this latter category include
10     both economically important species and endangered species. Similarly, human health
11     endpoints would likely include both risks to the general population and those relevant to
12     specific subsets of the population, such as children or the elderly, because of societal
13     concerns about their protection.
14
15            There is also an important role for decision-makers in this phase; for example,
16     decision-makers will be involved in helping to identify the important environmental
17     problems to be considered, identifying the sets of at-risk systems to address, and
18     identifying the specific ecological or human health endpoints to select.  During Problem
19     Formulation, decision-makers also need to identify clearly the range of potential
20     decisions and management options, examine economic, political, or other constraints
21     on the options to be considered, and to characterize the scope and time frame for IED
22     implementation.
23
24            In summary, in Phase I the participants—scientists, decision-makers, and
25     interested and affected parties—seek agreement through deliberative-analytic dialogue
26     (in the meaning of NRC, 1996) on such issues as:
27
28            a) the goals for the exercise, including environmental goals to be achieved;
29            b) which environmental problems/stressors/systems will be included and which
30            will not;
31            c) effects of concern;
32            d) the spatial, temporal, and organizational dimensions of the problem;
33            e) relevant data, models and analyses;
34            f) possible approaches to data analysis;
35            g) scoping of the uncertainties involved;

                                              1-16

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1           h) research needed to significantly reduce critical uncertainties;
 2           i) a first-cut on the range of options available to reduce risks;
 3           j) the endpoints upon which the health of the ecological or societal systems
 4           ultimately will be judged; and
 5           k) the type of factors that will be considered when reaching a decision.
 6
 7           The intent of Phase I is to have an open, yet structured, exchange of information,
 8     concerns, opinions, and values that will help to address the kinds of issues listed above.
 9     Chapters 2, 3, and 6 describe approaches by which some of the more technical aspects
10     of this information might be developed, and Chapter 5 describes deliberative  processes
11     that might be used to set environmental goals and incorporate values deliberation in
12     Problem Formulation.  The Agency's experience with some of its stakeholder processes
13     provides insights on constructive interaction for the purposes of planning and scoping.
14
15       1.3.3  Phase II: Analysis and Decision-Making
16
17           Phase II is that portion of the decision-making process in which most of the
       traditional "work" is done. Whereas in Phase I the participants formulate the  problem
19     using screening level information gained from risk comparisons, goal setting, and
20     preliminary options analysis, in Phase II the technical specialists employ similar
21     approaches, but with greater specificity and data requirements, to develop the in-depth
22     technical information that helps the risk managers to reach a final decision.
23
24           In practice, the analysts take the information and general directions gained from
25     Problem Formulation and generate more detailed, more fully supported risk
26     assessments and risk reduction options.  In the context of the IED framework, options
27     analysis requires consideration of risk reduction opportunities with regard to their
28     technical feasibility,  overall aggregate risk reduction to be obtained (e.g., reductions in
29     "target"  risks and collateral reduction in all affected risks), full economic consequences
30     of various risk reduction scenarios, and so forth (see Figure 1-5). Decision-makers may
31     also request analysis of potential options with regard to sustainability, equity, and other
32     potential
                                             1-17

-------
        IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
         Figure 1-5.  Options Analysis, Screening, and Selection
                   This component of the I ED framework is focused on identifying the best risk reduction opportunities
           and is applicable to cases of either a single stressor or for multiple stressors. In its simplest form—the form
           that has generally been used historically at the Agency— options analysis involves examination of multiple risk
           reduction options to address a highly ranked risk so as to identify the option(s) that will be most cost-effective in
           reducing that particular nsk.  The  IED framework expands this analysis to emphasize that ancillary reductions of
           other risks should also be assessed and factored into the decision process. For example, control of fly-ash
           emissions to reduce mercury emissions to the environment may not only reduce mercury but may reduce
           emissions of participates and possibly other pollutants as well. In an IED approach, these ancillary benefits
           should be considered explicitly when selecting a management option for a single stressor.

                   In the single stressor case, the approach described in Chapter 6 for designing a risk reduction
           program identifies possible actions that could be taken to reduce either the stressor or its effects on the
           system(s) at risk. An important aspect of this analysis is to examine a broader array of potential options than
           might typically have been done in the past. Criteria are developed in consultation with decision-makers to
           screen potential options, aggregate or disaggregate options, and, through an iterative process, converge on a
           set of options that analysis indicates would optimally reduce the risk. Chapters 4 and 5 provide more detailed
           guidance on assessing the economic and societal consequences of various options, an important aspect of
           options analysis.

                   While this simple form of integrated options analysis can yield a broader view of the benefits after an
           option or set of options has been selected, a more powerful feature of this integrated analysis results from its
           application during Problem Formulation, prior to selection of risk reduction options. In such an application,
           scientists should examine appropriate methods by which one can combine subsets of the ranked risks in order
           to investigate management options that could impact on the combined risks. The basis for such combinations
           of risk might include features such as common sources or pathways through the environment.  The critical link
           between the environmental risks in the subset would then be that they are all affected by a single risk reduction
           option/strategy. While it may not be possible to group all risks of concern on the basis of their technical
           attributes, a scientific analysis of the  risks may well reveal commonalities that indicate which risks will be
           affected by the same risk reduction option. The likely effect of this integrated view is that the option(s) selected
           to reduce a group of risks might differ from that which would be selected to reduce the top ranked risk, if it were
           to be considered in isolation.

                   When applied to the multi-stressor situation, the IED framework calls for an expanded analysis of risk
           reduction options so as to identify those options that may simultaneously reduce, directly or indirectly, risks
           posed by more than one stressor. The goal in this case is to maximize the reduction of the total aggregate nsk
           from multiple stressors, rather than to maximize the reduction of risks posed by any single stressor. This
           approach, requiring as it may the simultaneous consideration of risks from quite different types of stressors,
           has not yet been fully utilized, and will not be trivial to implement Nevertheless, the SAB believes that its
           development and implementation offer tremendous potential for improving environmental health overall.
1

2

3
4

5
 decision criteria.  Methods for assessing the economic and societal consequences of
potential actions are described in Chapters 4 and 5.  A detailed discussion of a process
for designing a risk reduction program is contained in Chapter 6.
                                                         1-18

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1           The Analysis portion of Phase II is generally more "analytic" than "deliberative"
 2     (NRC, 1996) although a continued level of interaction between the participants in the
 3     overall process (scientists, risk managers, and interested and affected parties) is
 4     important. Phase II is also more resource-intensive than Phase I since it can involve
 5     consideration of more options (e.g., various groupings of related stressors, as well as
 6     different risk reduction options) and more detailed analyses (e.g., the effects of different
 7     risk reduction options on multiple stressors or groups of stressors, as well as the
 8     inclusion of cost-effectiveness factors).
 9
10           In the Decision-Making portion of Phase II, the Agency or other decision-makers
11     utilize outputs from the analyses of risk and risk reduction options, consider widely-held
12     public values, the views of stakeholders, and legal and institutional constraints, and,
13     ultimately, make environmental decisions.  Clearly, this is not a totally scientific process.
14     However, the best science should inform and contribute to decision-making. This can
15     be accomplished, for example, by making explicit a) the implications of the chosen
16     management option(s) to the health of ecological or human systems; b) the economic
17     costs and  benefits associated with the option; and c) the societal values that are
       affected by the decision, including both values relating to economic efficiency and
19     values relating to sustainability and equity.
20
21           Another important role for scientific and technical analysis is to make clear the
22     uncertainties associated with estimates of risk, the estimates of risk reduction that may
23     be achieved by different management options, and the economic assessments of
24     various risk management scenarios.  The IED approach does not eliminate the
2&    uncertainties associated with making decisions. However, by encouraging an open and
26     comprehensive examination of environmental problems in an integrated fashion, the
27     IED framework should lead to a more clear identification of the nature and extent of the
28     uncertainties associated with the available information. In any event, environmental
29     decision-making must proceed in the presence of uncertainties, and nothing in the
30     proposed framework should be construed as precluding environmental decisions
31     because uncertainties remain.
32
33
                                             1-19

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1       1.3.4 Phase III:  Implementation and Performance Evaluation
 2
 3           Phase III of the IED framework consists of implementation of the chosen risk
 4     reduction options and evaluation over time of the extent to which the risk reduction
 5     measures are achieving the desired environmental outcomes.  This Phase involves the
 6     articulation and execution of the specific actions that must be taken to implement the
 7     decision, and the establishment of a process to evaluate performance and results of
 8     such action. The specific activities required to implement an environmental decision will
 9     depend on the suite of management options selected for any particular problem or set
10     of problems, and thus we do not address this aspect of the IED in any detail.  The
11     Agency has considerable experience with many of the risk reduction options described
12     in this report (e.g., adopting best available technology and imposing permit limits) and is
13     gaining valuable new experience with others (e.g., regulatory negotiation and National
14     Environmental Performance Partnerships).
15
16           In contrast to implementation, however, the performance evaluation process is
17     fundamentally rooted in science because it is science that can translate the public's
18     overarching goals (e.g., improved health, sustainable ecosystems) into discrete,
19     measurable components. Accordingly, science is essential in deciding what to monitor,
20     i.e., specifying the endpoints of concern for the systems at risk and identifying the
21     specific measures that need to be monitored in order to characterize the status and
22     trends for those selected endpoints with respect to the environmental goals. Further,
23     the scientific issues of spatial and temporal variability,  measurement error, time lags,
24     and so on, must be explicitly addressed in order to demonstrate environmental
25 '   condition and to separate signal from noise. And finally, reference conditions and
26     benchmarks or milestones along the way to the desired system conditions must be
27     defined scientifically so that meaningful and measurable performance criteria for
28     success or failure can be defined.
29
                                             1-20

-------
      IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote


      Figure 1-6.  Performance Evaluation/Report Cards  	
              The Government Performance and Results Act (GPRA) created a requirement that,
       beginning in Fiscal Year 2000, federal agencies report annually on measurable results of their
       various programs and activities. GPRA has resulted in a flurry of activity aimed at measuring
       program performance, including an effort funded by the Office of Science and Technology Policy to
       design a national environmental report card.

              In the IED context, an Environmental Report Card is a tool for communicating to multiple
       audiences the performance of a risk reduction program in measurable terms related to environmental
       outcomes. At the broadest level, the Report Card should inform all of the IED participants about how
       well, in general, the ecological or societal system at risk is responding to the actions taken. At a more
       detailed level, the Report Card should provide scientists with sufficient monitoring data to improve the
       risk assessments and/or the risk reduction options previously selected. The need for monitoring data
       to assess performance in this framework emphasizes again the importance attached to EPA's
       development of monitoring programs which can measure both ecological and human health exposure
       and outcomes.

              The report card should contain specific milestones that can be used to measure progress
       towards achieving the environmental goal(s) agreed upon by the IED participants. Each of the
       selected endpomts defined during Problem Formulation should be a part of the report card, as well as
       the specific measures or indicators that are monitored to characterize those endpoints. The
       frequency of the reporting should be decided upon with all of the participants and should be
       commensurate with the nature of the risk and the time frame for system response. Four types of
       performance measures can be used in evaluating progress: measures of administrative effort, often
       called process measures (e.g. number of permits issued, number of control technologies installed);
       measures of stressor levels; measures of exposure; and measures of environmental outcomes (i.e.,
       measures of adverse effects or condition) that report on changes in the state of the systems at risk
       (e.g. hectares of wetland restored or the number of cancer cases avoided). Decreases in adverse
       effects or improvements in health or environmental condition are the ultimate basis for evaluating risk
       reduction programs, and environmental outcome measures (or early markers of  the final outcome)
       are therefore preferable. It is often necessary, however, to supplement outcome measures with
       shorter term measures (e.g., early markers of the final outcome, as well as process, exposure, and
       stressor measures) for purposes of program accountability and course correction.
1            The IED framework includes the use of Environmental Report Cards to
2      document performance and outcomes of risk reduction activities at several levels and
3      for different audiences (see Figure 1-6). As shown in the IED framework, the
4      performance evaluation process contains several important feedbacks associated with
5      the Report Card. One feedback loop is to the Analysis and Decision-making Phase,
6      reporting on how well the selected risk reduction options and strategies are achieving
7      the environmental goals. This feedback loop allows for adaptive management and
8      changes in implementation activities, including the possible need to identify and analyze
9      additional options to further reduce risks. A second important role for Report Card


                                                1-21

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1     information is to allow re-examination, as needed, of the initial risk rankings or other
 2     aspects of Problem Formulation. As risk reduction options are put into place, for
 3     example, particular risks should be reduced, and a reordering of risk rankings may be
 4     appropriate. Further, there may be a shift in or redefinition of societal values over time,
 5     requiring different sets of environmental goals and, therefore, different environmental
 6     decisions.
 7
 8           In summary, the IED framework emphasizes the need to consider performance
 9     information at several points in the decision-making process and to review
10     environmental decisions in light of new scientific understanding, shifts in societal
11     values, changes in stakeholder preferences and available resources, and/or responses
12     of the environment to previous decisions. The topic of performance evaluation and
13     report cards is discussed more fully in Chapter 7.
14
15
16
                                             1-22

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1     1.4  Nature of the Framework
 2
 3      1.4.1 Major Characteristics
 4
 5           Integrated environmental decisions should exhibit the following characteristics:
 6
 7           a)     Transparency. The IED framework is designed to promote transparency
 8                 so that interested parties will be able to follow the process and be aware
 9                 of the information that was considered in reaching decisions.
10                 Transparency is enhanced by the use of clearly articulated goals,
11                 analytical methods, and criteria; open deliberative processes; and well-
12                 documented decisions.
13
14           b)     Flexibility.  The IED framework can be applied in a flexible manner
15                 depending on the specific circumstance; i.e., where appropriate, to permit
16                 valid short-cuts, to eliminate unnecessary procedures, and so to expedite
117                 the process of decision-making and implementation. Factors such as the
                   extent and nature of a problem, the amount and kind of information
1 g                 available, and the information gaps identified will influence the required
20                 degree of complexity of approach and level of detail of the analyses.
21
22           c)     Dynamic process design. The technical analyses required to implement
23                 the framework should not be conducted in an isolated, stepwise manner.
24                 For example, during the Problem Formulation Phase, problem scoping
25                 and definition and preliminary analysis of options will affect the
26                 development of goals, and vice versa. Some iteration is also required
27                 between Problem Formulation and Analysis, since preliminary analyses
28                 will often point out missing elements in the problem definition or
29                 inconsistencies in goals.
30
31           d)    Explicit feedback, interaction, and cooperation. The IED process
32                 requires cooperation and open, continuing communication between
33                 scientists, managers, members of the public, and others involved in the
34                 different phases of an IED project.  Examples of critical feedback
35                 processes include the evaluation of performance against specified

                                            1-23

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1                 environmental goals, which allows course corrections and may improve
 2                 future decisions, and two-way communication between policy-making
 3                 bodies and the public.
 4
 5           e)    The use of information from many sources. The IED framework
 6                 requires use of concepts and methods originating in many different
 7                 scientific, technical, and scholarly fields (e.g., physical sciences, public
 8                 health, environmental engineering, political science, philosophy, and
 9                 economics), as appropriate for any given case.  In addition to science and
10                 scientific judgments, inclusion of public values is needed to compare
11                 unlike risks, to set acceptable goals, to assess risks to quality of life, to set
12                 priorities, and to reach broadly acceptable decisions.
13
14           f)     A way of thinking about environmental problems. Finally, integrated
15                 environmental decision-making is not just a series of methodologies, but
16                 rather is a way of thinking, in a whole and complete way, about any
17                 environmental decision-making case in order to maximize the efficient
18                 reduction of aggregate risk to populations or ecological systems. The IED
19                 framework should help to focus attention on the multiple aspects of a
20                 problem, a broad range of factors that may influence the decision, and
21                 important evaluation processes that add value after a decision is
22                 implemented. The following section provides a brief synopsis of the
23                 opportunities for integration provided by the framework.
24
                                             1-24

-------
        IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17

19
20
21
22
23
24
25-
26
27
28
29
30
31
32
33
34
35
Figure 1-7.  Types of Integration in the
IED Framework
  1.4.2 Types of Integration

      The IED framework requires that
information and viewpoints be integrated at
multiple points in the decision-making
process. Six critical types of integration are
summarized in Figure 1-7, and discussed
briefly below.

      a) Integrated Risk Assessment

       Focusing on a single agent, a single
medium, and a single outcome to assess
risks is not a realistic representation of the
way environmental exposures impact on
humans and ecosystems. While such an
over-simplified view was arguably
necessary in the early days of risk
assessment, we should now work to
develop tools that can determine risks from
multiple exposures and multiple  outcomes
in order to more accurately represent real
world situations. As already noted, some
early, but significant, progress has been
made in this area and the SAB encourages
the Agency to continue these efforts.  In
response to the Food Quality Protection Act
of 1996, for example, the Agency has
stepped up efforts to consider risks posed
by groups of pesticides exhibiting common
modes of action, rather than setting risk levels based on each pesticide individually. In-
depth exploration of the analytical challenges inherent in integrated or cumulative risk
assessment is beyond the scope of this study, but it is an area that will require
extensive thought and methodologic development.
 Integrated Risk Assessment:
 developing scientific data and analytical methods
 for determining risks from multiple exposures and
 multiple outcomes in order to more accurately
 represent real world situations.

 Risk Comparisons:
 considenng a wide range of environmental risks
 simultaneously so that the seriousness of risks can
 be characterized relative to one another.

 Integrated Analysis of Management
 Options:
 investigation of options to reduce subsets of ranked
 risks, rather than considenng single risks in
 isolation, to achieve greater aggregate nsk
 reduction.

 Integrated Analysis of Economic
 Consequences:
 identifying the full range of benefits and costs, both
 monetized and non-monetized, associated with
 reduction of multiple risks.

 Integration of Performance
 Information:
 using performance evaluation measures to devise
 course-corrections.

 Integrating Multiple Disciplines and
 Points of View
 understanding and utilizing information from all
 concerned parties in the IED process.
                                               1-25

-------
        IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1            b) Risk Comparisons
 2
 3            The second type of integration important for integrated decision-making involves
 4     the consideration of a wide range of environmental risks simultaneously so that the
 5     seriousness of risks can be characterized relative to one another. Chapters 2 and 3
 6     describe approaches to technical risk comparisons based on the best scientific
 7     information and expert judgment. The methods use stressors as a common point of
 8     departure to facilitate an understanding of the interconnections between risks to
 9     humans and risks to natural systems. Environmental concerns may also be expressed
10     in terms of sources, pathways, or endpoints; these entities are interrelated so that no
11     matter which entity is used as a basis of comparison, scientific information on stressors,
12     sources, pathways, and endpoints will be required.
13
14            An important aspect of the IED framework, however, is the explicit notion that
15     technical risk comparisons are not in themselves sufficient to inform an environmental
16     decision or to set environmental priorities per se. Other factors that should be
17     considered in setting risk priorities include the availability of management options,
18     opportunities for overall aggregate risk reduction, economic impacts, and public
19     concerns.
20
21            c) Integrated Analysis of Management Options
22
23            The third type of integration in the IED framework, discussed in Section 1.3.3,
24     occurs during the analysis and selection of risk reduction options. In this process, risk
26     rankings are analyzed with regard to opportunities for risk reduction (i.e., options
26     available to address the risks) in order to determine the best approach to reduce
27     multiple risks simultaneously in the most cost-effective  manner. Clustering of risks is
28     useful both during Problem Formulation (to identify a problem set that will maximize
29     cost-effective risk reduction) and Analysis (to identify the full range of risk reduction
30     benefits associated with any set of control options). A risk reduction program designed
31     to address sets of priority risks will likely differ from a program designed to address a
32     single risk, and should include consideration of the effects (positive or negative) of
33     management options on "non-target" risks.
34
35            The SAB  recognizes that the Agency often faces a mandate (legislative or court-

                                             1-26

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1     ordered) to reduce the risks from certain stressors to no more than some maximum
 2     level.  In these situations, the Agency can use the IED framework to develop risk
 3     reduction options that cost-effectively reduce "collateral11 risks from other stressors while
 4     meeting the mandated risk level for a particular stressor.
 5
 6           The SAB also recognizes that this type of integrated options analysis is a more
 7     difficult undertaking than the traditional approach and that such an exercise would
 8     require many estimations regarding combined effectiveness of management options,
 9     each subject to uncertainties. However, as the Agency gains experience with and
10     develops new tools for such integrated analyses, the quality of the decisions will
11     improve, and the improvement of environmental conditions should become apparent.
12
13           Chapter 6  describes a process for identifying, screening, and selecting risk
14     reduction options for environmental problems defined in terms of single or multiple
15     stressors.
16
17           d) Integrated Analysis of Economic Consequences

19           Approaching environmental problems from the standpoint of integrated risks
20     should improve our ability to identify the full range of benefits and costs, both monetized
21     and non-monetized. In theory, economic analysis is an inherently integrated exercise in
22     which all benefits, values, or gains are  compared to the costs or losses from any
23     environmental intervention. Many benefit/cost analyses  in the past, however, have
24     focused on the control of a single source of a chemical. This narrow focus is largely the
25     result of the way  in which the "problem" has been defined. When the approach to risk
26     is broadened to examine the multiple effects of multiple chemicals stemming from a
27     source and the reduction in multiple risks that will ensue from controlling those
28     chemicals, the scope of the benefit/cost analysis will be broadened to include the
29     multiple benefits from the reductions of multiple risks. Chapter 4 describes the basic
30     principles that underlie benefit/cost analysis and Chapter 5 discusses issues
31     surrounding the assessment of difficult-to-quantify benefits.
32
33           e) Integration of Performance Information
34
35           Many decision-making schemes make provision for later evaluations of the

                                            1-27

-------
        IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1     decisions that have been made. In practice, however, such after-the-fact reviews too
 2     often do not occur. Once a decision has been made, there is a natural tendency to
 3     move on to new problems, rather than to re-visit old ones. And yet, integration of
 4     information on performance and environmental results is a critical feature of any
 5     approach that includes principles of adaptive management.  Today's environmental
 6     problems are often so complex that,  no matter how sophisticated the analysis leading to
 7     an initial judgment, it is from reviewing the effects of and making adjustments to an
 8     earlier decision that a risk manager becomes better equipped to make future decisions.
 9
10            Performance evaluation and feedback is an integral part of the IED framework;
11     this component of the framework is summarized in section 1.3.4 and described in more
12     detail in Chapter 7.
13
14            f) Integrating Multiple Disciplines and Points of View
15
16            Each of the five types of integration described above depends upon, to varying
17     degrees, effective interactions between scientists, decision-makers, and interested and
18     affected parties; i.e., the integration of information and understanding of all concerned
19     parties in the IED process.  As noted above, the IED framework requires not only the
20     use of information and perspectives from a number of scientific disciplines, but also the
21     inclusion of public values throughout the decision-making process.  In practice, this
22     means that, although the in-depth technical tasks are undertaken by experts in discrete,
23     analytic exercises, the experts periodically will be informed and their results reviewed by
24     other IED participants. Although much of the  emphasis in this report is on inclusion of
25*     perspectives from outside the Agency, coordination and communication within the
26     Agency during the course of the various analyses will be critical as well. In this way,
27     through a series of deliberative and analytic processes (NRC, 1996), all of the
28     participants gain an understanding of the value-based concerns of the others.  In fact,
29     the IED calls for recognition, inclusion, and consideration of values throughout the
30     process, it is only through explicit inclusion and integration of the values and
31     perspectives of a diverse set of participants, from within and outside the Agency, that
32     the most acceptable and effective environmental risk reduction will occur. The
33     problems and promises of this important type of integration are discussed in Chapter 5.
34
35

                                             1-28

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1       1.4.3 Building on Previous Frameworks
 2
 3           From the previous section, the reader can see that the goal of IED is to identify
 4     appropriate risk reduction actions to address a mixture of environmental problems (both
 5     human health and ecological), so as to reduce total aggregate risk, rather than focusing
 6     only on single risks in isolation. Deliberative processes are one means of deciding
 7     which mix of  risk reduction goals should be pursued.  The IED framework explicitly
 8     includes a mechanism for evaluation and feedback, so that a strategy of adaptive
 9     management can be easily employed.  In addition, the framework includes substantive
10     interaction with interested and affected parties throughout the process so that public
11     values are reflected appropriately.
12
13           The IED framework proposed by the SAB builds upon several previous
14     frameworks,  in particular the risk assessment/risk management model described by the
15     National Research Council (NRC 1983), the ecological  risk assessment framework
16     (U.S. EPA 1992), and the risk characterization process described in NRC (1996), which
17     focused on the interaction  between analytic and deliberative processes in decision-
       making. The IED framework moves beyond these earlier efforts in three significant
19     areas:
20
21            a) Evaluation of Single Stressors
22
23           Although the SAB emphasizes that the Agency should consider multiple
24     stressors in an integrated approach to risk, it recognizes that the IED framework also
25     should enhance the decision-making process for single stressors, of the type historically
26     considered by the Agency (e.g., a drinking water pollutant).  When a single stressor is
27     considered, the IED framework should expand the previous approaches by:
28
29           (1) Characterizing stress-effects relationships across all systems and
30           populations;
31           (2) Exploring a broader range of risk reduction options, some of which may be
32           qualitatively new;
33           (3) Assessing the benefits and costs of each option, including explicit
34           consideration of non-monetary benefits and costs;
35           (4) Assessing the magnitude and nature of the aggregate risk reduction

                                            1-29

-------
        IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1           associated with each option;
 2           (5) Involving scientists, options analysts, stakeholders, and risk managers
 3           collectively at various points throughout the process; and
 4           (6) Establishing a performance evaluation "report card" to characterize the
 5           efficacy of the implemented risk reduction option and to signal both the need and
 6           opportunity for adapting the original management decision.
 7
 8            b) Evaluation of Multiple Stressors
 9
10           One of the most important extensions in the IED framework is that, unlike
11     previous frameworks, a primary goal of the IED process is to consider multiple stressors
12     when formulating the problem and developing possible solutions.  The IED provides a
13     structured way to begin to explore multiple ecological risks, multiple human health risks,
14     and/or multiple quality of life concerns. In some cases this approach will lead to
15     consideration of combinations or groups of risks; e.g., all organophosphate pesticides,
16     all automobile emissions, or all factors leading to local loss of biodiversity.
17
18           Initial steps in the IED process are designed to produce relative rankings of risks
19     to human health, quality of life, and ecosystems, independently. While the SAB had
20     originally planned to integrate the rankings for the various types of risks, the members
21     concluded that such a merged ranking is based on values,  rather than on science
22     alone. The SAB's review of risks of concern also emphasizes that health and
23     ecological risk assessments are often qualitatively different. For example, the focus of
24     the assessments is different; that is,  the focus of health risk assessment is an  individual
26     within a single species, while the focus of even a narrowly drawn ecological risk
26     assessment is entire populations of any/many species. More generally, ecological risk
27     assessments often address the integrated risks to a prescribed region, such as a
28     watershed. This difference is reflected in the different types of stressors of concern; cf.,
29     carcinogens for humans vs. habitat fragmentation. However, in those instances where
30     there are common stressors of concern (e.g., chlorinated pesticides, climate change) or
31     where effects of a stressor on an ecological system produce effects on human health or
32     quality of life (e.g., habitat alteration that affects the range and activity of disease
33     vectors and infective parasites, or changes in the abundance of commercially important
34     or endangered species), there is an opportunity for some merging of concerns to take
35     place. An understanding of the relationships among different types of risks/stressors is

                                             1-30

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1     critical to Problem Formulation, both in terms of setting environmental priorities and
 2     goals, and for designing subsequent risk and risk reduction analyses.
 3
 4           In addition to ranking of risks, which allows scientists to compare the expected
 5     impacts of each risk, the IED framework suggests combining risks into logical
 6     groupings, e.g., those with a common source or pathway, in order to identify risk
 7     reduction opportunities across stressors.  In order to be successful, this analysis
 8     requires open, publicly-accessible, and frequent dialogue among those who assess and
 9     compare risks, those who determine methods for reducing risks, and those who make
10     the final decisions.
11
12           In summary, as the IED framework is implemented to address multiple stressors,
13     it should:
14
15           (1)    Lead to a more realistic ranking of risks to humans and to ecosystems ,
16                 where some of those risks may be  posed by combinations of related
17                 stressors;
             (2)    Lead the Agency to consider in a systematic fashion all of the appropriate
19                 factors related to risks in a given circumstance, including aggregate risk,
20                 economic factors, and societal values; and
21           (3)    Lead to action that will increase the reduction in aggregate risk posed by a
22                 combination of stressors in  a given  circumstance.
23
24           As the Agency gains experience and develops more sophisticated analytical
25     tools, the benefits of IED should become  more apparent. At first, it may be easiest
26     simply to examine the total reductions in exposures. More complete evaluations of
27     outcomes are complicated by the need to estimate reductions in several types of risk
28     (e.g., health, ecological, and quality of life risks) resulting from the reductions of multiple
29     stressors affecting multiple receptors. Nonetheless, appropriate methods should be
30     developed to measure the total impact and benefits of risk reduction programs in the
31     future.
32
33            c) Considering Environmental Values
34
35           The third extension that the IED framework provides is a more explicit

                                            1-31

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1     consideration of the analytic/deliberative process described by the MAS Panel on Risk
 2     Characterization (NRC, 1996). It is through such a process that societal values
 3     intersect with the scientific risk characterization and risk reduction analyses. The
 4     proposed I ED framework emphasizes the role and timing of stakeholder and decision-
 5     maker inputs to the analytic processes.  It explores more deeply the valuation of
 6     environmental outcomes and risks and the need to include not only the concepts of
 7     economic efficiency and wiHingness-to-pay, but also issues of environmental
 8     sustainability and equity. The IED framework recognizes that societal values constitute
 9     the milieu in which integrated environmental decision-making occurs forming the basis
10     for societal goals for improved social welfare, improved ecological/health conditions,
11     and long-term sustainability and equity.  It is in the realm of social values that the
12     success or failure of  environmental decisions will be judged.
13
14
                                             1-32

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1     1.5  Benefits and Challenges of the Framework
 2
 3      1.5.1 Benefits
 4
 5           A fully implemented IED framework should result in a new level of environmental
 6     protection. The IED framework refines and moves beyond the risk ranking exercises
 7     contained in the SAB's Reducing Risk report.  The techniques described in this report
 8     for ranking health and ecological risks offer a clear and systematic approach for using
 9     the best scientific information and expert judgment and are substantially more
10     transparent than were those in the earlier effort. In addition, the IED framework goes
11     beyond the comparison of risks and risk reduction options to spotlight the as-yet
12     unrealized benefits to be gained from addressing environmental problems in a truly
13     integrated fashion - including integration of technical information on risk (including
14     information on multiple stressors, sources, and effects), risk reduction options, and
15     economic consequences; integration of values and goals into the decision-making
16     process; and integration of performance information so that needed course corrections
17     can be identified. While the Agency has made some significant first steps in some of
       these areas, the full benefits are yet to be achieved.
19
20           The expected benefits of the IED framework relate directly to the characteristics
21     of the framework described in this chapter. Transparency of the decision-making
22     process should increase acceptance of the outcome by affected parties, flexibility
23     allows the process to be adapted to suit the needs of a particular situation, open and
24     inclusive dialogue should lead to broader consideration of perspectives and options,
25*    and so forth. Broadly speaking, then, implementation of the IED framework should
26     result in:
27
28           a)    Enhanced ability to improve human health and quality of life and to protect
29                 the  integrity of ecological systems;
30
31           b)    Improved targeting of resources for risk reduction; and
32
33           c)    Greater accountability for results;
34
35

                                            1-33

-------
        IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

  1       1.5.2 Challenges
  2
  3            At the same time, the IED framework presents the Agency with significant
  4     challenges.  The techniques recommended for ranking risks in the IED will need further
  5     refinement during their development and application. Likewise, the concept of
  6     combining risks into related subsets for risk management purposes offers considerable
  7     promise but introduces new uncertainties that must be acknowledged and addressed.
  8     The integration of economic concepts, both for setting environmental goals and for
  9     evaluating how best to meet those goals, is important but may also be controversial
10     and, therefore, must be introduced clearly and objectively to avoid unwarranted
11     criticism. The consideration of management options from the total aggregate risk vs.
12     single risk perspective, which is a major feature of the framework, could also be
13     controversial, even in the face of its technical soundness. Finally, the concept of
14     evaluating the impact of decisions and reflecting that information in subsequent
15     management actions (i.e., adaptive management) is not new; however, the challenge of
16     using this approach in a consistent and continuing way remains.
17
18            In addition to the technical and methodological issues, it is important to
19     recognize that successful implementation of the  IED framework will require some
20     adjustments in the manner and degree to which the many participants interact over the
21     course of the decision-making process. Integrated environmental decision-making
22     requires the sharing of information, ideas,  approaches, and management deliberations
23     to a degree now seldom  practiced among individuals of very different backgrounds.
24     Although this sharing is a positive aspect of the IED framework, it may require
25'    significant adaptation on the part of individual policy-makers and institutions. For
26     example, decision-makers will need to interact more extensively with scientific and
27     technical analysts and the public in the course of developing integrated approaches to
28     environmental risks. Likewise, scientific and technical experts will need to recognize
29     the role (and limitations)  of science in decision-making, and also to recognize the
30     legitimate role of values in establishing environmental goals and selecting management
31     approaches.  This culture change will be assisted by familiarity and experience with the
32     IED process. Experience, combined with discipline, will also be needed to apply the
33     IED framework with discretion to the depth  of detail necessary for a particular problem,
34     and no more; i.e., the IED framework should not become a barrier to decision-making.
35

                                            1-34

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1           Additional challenges likely to be encountered when implementing the IED
 2     framework include:
 3
 4           a)     problems of understanding arising from differences in terminology and
 5                 outlook imbedded in the different disciplines and backgrounds of the
 6                 participants in the process;
 7
 8           b)     difficulties of using both qualitative and quantitative measures
 9                 concurrently in the decision process;
10
11           c)     the need to compare different types of risks (e.g., health, ecological and
12                 quality of life risks) within a common decision framework and to discern
13                 and define the inter-relationships among risks so as to define common
14                 goals across the different risk types; and
15
16           d)     time-lags between implementation of risk reduction plans and the
17                 detection of results and effects, which make the selection of appropriate
                   performance measures particularly important.
19
20           Facing and surmounting the challenges, learning how to use the IED framework
21     flexibly with discretion and care, and improving the underlying methodologies over time,
22     will lead to more effective environmental decision-making and to measurable
23     environmental progress.  The SAB expects that the proposed framework and the
24     recommended improvements to the component analyses will assist the Agency to
25     develop a more comprehensive, integrated, and transparent process and culture for
26     setting environmental  priorities and  making and implementing decisions.
                                            1-35

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1     1.6 References Cited
 2
 3     National Academy of Public Administration. 1995. Setting Priorities, Getting Results:
 4           A New Direction for EPA.
 5
 6     National Research Council.  1983.  Risk Assessment in the Federal Government:
 7           Managing the Process. National Academy Press, Washington, DC.
 8
 9     National Research Council.  1994.  Science and Judgment in Risk Assessment.
10           National Academy Press, Washington, DC.
11
12     National Research Council.  1996. Stem, P.C. and H.V. Fineberg (ed.s).
13           Understanding Risk: Informing Decisions in a Democratic Society.  National
14           Academy Press, Washington, DC.
15
16     Presidential/Congressional Commission on Risk Assessment and Risk Management.
17           1997. Risk Assessment and Risk Management in Regulatory Decision-Making.
18
19     Science Advisory Board. 1990. Reducing Risk: Setting Priorities and Strategies for
20           Environmental Protection (EPA-SAB-EC-90-021).
21
22     Science Advisory Board. 1995. Beyond the Horizon: Using Foresight to Protect the
23           Environmental Future (EPA-SAB-EC-95-007).
24
25     Science Advisory Board. 1997. Advisory on the Problem Formulation Phase of EPA's
26           Watershed Ecological Risk Assessment Case Studies (EPA-SAB-EPEC-ADV-
27           97-001).
28
29     U.S. Environmental Protection Agency. 1987.  Unfinished Business: A Comparative
30           Assessment of Environmental Problems. Washington, DC.
31
32     U.S. Environmental Protection Agency. 1982.  Framework for Ecological Risk
33           Assessment (EPA/630/R-92/001). Office of Research and Development,
34           Washington, DC.
35

                                           1-36

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1     U.S. Environmental Protection Agency.  1989.  Risk Assessment Guidance for
 2           Superfund (EPA/
 3
 4     U.S. Environmental Protection Agency.  1990.  Methodology for Assessing Health Risks
 5           Associated with Indirect Exposure to Combustor Emissions-Interim Final.
 6
 7     U.S. Environmental Protection Agency.  1993a. Draft Addendum to the Indirect
 8           Exposure Document.
 9
10     U.S. Environmental Protection Agency.  1993b. A Guidebook to Comparing Risks and
11           Setting Environmental Priorities (EPA 230-B-93-003). Office of Policy, Planning,
12           and Evaluation, Washington, DC.
13
14     U.S. Environmental Protection Agency.  1997.  Guidance on Cumulative Risk
15           Assessment, Part 1: Planning and Scoping. July 3, 1997.
16
17     U.S. Environmental Protection Agency.  1998.  Guidelines for Ecological Risk
             Assessment (EPA/630/R-95/002Fa).
19
20
                                           1-37

-------
PART II  INPUTS TO ENVIRONMENTAL DECISION-MAKING:




              RISK COMPARISONS

-------
      IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote


        PART II — INPUTS TO ENVIRONMENTAL DECISION-MAKING: RISK
                                    COMPARISONS

                                         Preface

 1
 2           The process of comparing environmental health and ecological risks is a key
 3     component of the Problem Formulation Phase of the IED framework.  In the early
 4     stages of problem formulation, a draft list of stressors of concern will be identified using
 5     existing knowledge and expertise of a group of scientists (including human exposure
 6     and health experts and/or ecological experts) and/or on the basis of a review of
 7     available scientific literature. A comprehensive list of possible stressors is developed to
 8     provide a complete set of possible comparisons. The draft list developed by risk
 9     scientists would then be discussed with others in the IED process, including the options
10     analysts, decision-makers, and stakeholders.  As a result of these discussions,
      additional stressors might then be added or removed, some groups of related stressors
      might be aggregated into a single composite stressor (e.g., "waterbome infectious
13     microbes" vs. "Cryptosporidiurrf), and some stressors might be disaggregated into
u     more specific stressors (e.g., "mercury" vs. "heavy metals") for subsequent analysis.
15     Quantitative and qualitative information on the stressors would be assembled into a
16     format, such as a stressor risk data sheet, to facilitate comparisons. Important
17     information would include: stressor intensity; the adverse effects likely to be associated
18     with exposure to the stressor; observed effects in the population or ecosystem; and
19     some assessment of the causal relationships between the environmental stressor and
20     the observed effects. This latter could include assessments of the probable
21     contribution of the stressor to the observed effects of concern, the so-called attributable
22     risk. Major gaps in knowledge should also be identified as part of Problem Formulation
23     and Analysis in order to inform research planning and help set research priorities.
24
25            During the design of the IED framework, the SAB participants acknowledged that
26     technical risk rankings,  in isolation, offered insufficient guidance for policy decisions.
27     Given the multitude of problems and issues to be addressed, a more  comprehensive
28     and systematic framework for analyzing and reducing environmental health, ecological,
29     and quality of life risks appeared necessary.  Analysis and comparison of various

                                            n-i

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1     environmental health and ecosystem risks, however, remain the fundamental basis of
 2     such a framework, and the foundation for defining risk reduction priorities.
 3
 4           As symbolized by the interlocking circles in the IED framework, the assessment
 5     and comparison of risks and the definition of environmental goals are inter-related. In
 6     other words, scientific information on the nature and extent of various risks influences
 7     the relative priority that society places on those risks. In addition, however, non-
 8     scientific issues such as  dread, previous experiences, and degree to which exposure is
 9     voluntary, for example, also influence the relative priority assigned by the public to
10     environmental risks.  For this reason, the ultimate risk priorities that emerge during
11     Problem Formulation will be a product of the interaction between the risk comparison
12     and goal setting processes, and (as described in Part IV) the preliminary assessment of
13     risk reduction potential.
14
15           As the IED process proceeds from the problem formulation to the analysis
16     phase, the risk assessors develop more detailed risk analyses for the priority stressors,
17     which can be used to help focus the development of options.  The stressor
18     comparisons will  be used to help organize thinking about potential options for reducing
19     risks and to examine opportunities for reducing several risks through a single risk
20     reduction option. The more detailed  information on exposures and risks also provides
21     the basis for estimating the risk reduction achievable through various options, and for
22     determining the specific environmental outcomes that will  be used to evaluate
23     performance against the risk  reduction goals.
24
25'           During Problem Formulation, it is important to identify linkages between health,
26     ecological, and quality of life risks and to seek opportunities to address all three types
27     of risks through a risk reduction program.  For this reason, it is helpful to define the risk
28     problems using a common dimension; the SAB Subcommittees chose to define risks in
29     terms of stressors.
30
31           Stressors  of concern may be chemical (e.g., ozone, heavy metals), physical
32     (e.g., radiation, changes in land  use), or biological (e.g., infectious agents in drinking
33     water, introduced exotic species). Further, the stressors may cause adverse health and
34     ecological effects directly (through exposures via environmental media) or may act
35     indirectly.  For example, freons released into the atmosphere do not directly influence

                                             n-ii

-------
      IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1     health, but, instead, reduce the thickness of the protective ozone layer of the
 2     atmosphere, consequently increasing the risks of skin cancer. A methodology for
 3     comparing environmental risks to quality of life was not developed by the SAB because
 4     the selection of QOL ranking criteria is more appropriately done by a broad group of
 5     public or stakeholder representatives rather than by a technical panel. However, the
 6     SAB recognizes the need to develop and incorporate such a process into integrated
 7     environmental decision-making.
 8
 g           Two Subcommittees of the IRP developed risk comparison approaches  whose
10     results could be used in the IED process. One approach, which was applied to the
11     comparison of ecological risks, utilizes an expert group to develop and weight risk
12     ranking factors in order to produce the group's consensus judgment of the relative risks
13     associated with various stressors.  The second approach, which is discussed using
u     human  health  risk issues, involves polling experts individually for their professional
15     judgment of relative degree of risk associated with various stressors, while soliciting
16     information on which factors most influenced this judgment and the degree of
      confidence each expert assigns to his/her assessments. The latter method provides
      information not only on the "average" or "median" risk ranking, but also on the range of
19     expert opinions. The consensus approach or the individual polling approach could be
20     applied to either human health or ecological risk comparisons using stressors and rating
21     factors  specific to each group.  The methodologies also could be used to elicit
22     stakeholder or public views on risk priorities by expanding the composition of the
23     surveyed or empaneled group.
24
25           The Ecological Risk Subcommittee (ERS), as a group of experts, developed a
26     set of decision rules for ranking ecological risks. The Subcommittee identified a set of
27     stressors of ecological importance and established a systematic template for
28     considering each stressor with respect to issues of ecological importance (e.g., spatial
29    extent and duration of the stress, intensity of the ecological effect, recoverability once
so    the stressor is  removed, and other issues of special ecological significance). The group
31     then  developed a process to characterize the co-occurrence of stressor intensity and
32    ecological consequences for an at-risk ecosystem. For each identified ecological factor
33    of importance, a multiplicative value was assigned that would adjust the risk level up or
34    down for each individual stressor.  Once all the factors were considered, the initial risk
35    assignment was adjusted by each factor, leading to a summary risk value for each

                                            n-iii

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1     stressor.  Stressors were then placed in narrative risk categories on the basis of the
 2     numerical risk scores, which derived directly from the application of the ecologically
 3     based criteria.
 4
 5           An important advantage of the ERS method is the transparency of the factors
 6     and their assigned relative weights used for deriving the risk rankings.  The dialogue
 7     among the convened experts helps to ensure that there is a common understanding of
 8     terminology and of the specific criteria for assigning ranking factors, a process that is
 9     very difficult to accomplish when polling individuals.  A potential disadvantage of this
10     approach is that the quantitative scores calculated for the risk rankings could be
11     misused by attributing greater precision to the numbers than the methodology warrants.
12
13           The Human Exposure and Health Subcommittee (HEHS) adopted a process
14     based on the use of the World Wide Web to solicit risk comparisons from a large
15     number of experts in fields related to human health risks.  One reason for the choice of
16     such a methodology was the extreme breadth of disciplines pertaining to environmental
17     health, e.g., epidemiology, exposure analysis, and toxicology. The respondents are
18     provided with a list of stressors, baseline information regarding these stressors and
19     their associated outcomes, and a consistent template to provide their comparisons of
20     risks to human health attributable to exposures to the stressors. As a part of the
21     survey, a set of about ten factors that might relate to the risk comparison is provided,
22     and the respondent is asked to indicate those influencing her/his rating for each
23     stressor.
24
25'          An important advantage of the HEHS methodology is that it provides information
26     on the variability in expert opinion about the relative  degree of risk associated with
27     various stressors, as well as on the confidence of the experts in their risk comparisons.
28     A disadvantage  of the methodology is that the opportunity for interaction and
29     information-sharing among the expert participants is lost, and the effort required to
30     prepare a common set of information for all risk areas to be compared is substantial.
31
32
33
                                             n-iv

-------
      IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
 1
 2

 3

 4

 5

 6


 7
 8

 9

10

11

12
13
 •4
 J
16
17
18
19
20
21
2?
23
24
25
26
27
28
29
30
31
32
Table PII-1.  Correspondence of Human Health and Ecological Risk Comparison
            Factors
Human Health Risk Factors
Size of population affected
Particular subpopulations at risk
Severity and persistence of health effects
Persistence in the environment and/or
human body
Percent of attributable incidence
Potential future risk


Ecological Risk Factors
Proportion of resource at risk
Distribution of "hot spots"
Recovery potential,
Species depletion
Duration of stress-effects co-occurrence


Secondary stress induction
Special ecological significance
       Despite the apparent differences in the ERS and HEHS approaches to risk
comparisons, the factors influencing the relative ranking of risks share many
commonalities (Table PII-1). The primary difference in the two approaches lies in the
manner in which the decision rules are derived.  The relationship between the
methodologies is illustrated in Figure PII-1.  Following from left to right in this figure, the
initial comprehensive list of stressors is identified and information concerning stressor-
effect relationships is collected.  This information is provided to a group of experts,
either constituted as an expert panel (as in the ERS approach, depicted by the
feedback arrow from the Risk Comparison box to the Decision Rules box) or as
individual experts polled separately (as in the HEHS approach). In the HEHS
methodology, the risk comparisons are derived directly from the experts. The experts
are also asked to indicate the factors that influenced their risk ratings for each stressor.
A multivariate  regression analysis could be used to relate the risk ratings to the various
factors that the experts indicate  influenced those ratings. The factor weights from the
regression analysis then define the Decision Rules.  In the ERS approach, the experts
develop the Decision Rules directly, then apply them to the stressor database to derive
the Risk Comparisons.
                                             n-v

-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
                                  Consistent Framework for Risk Comparisons
 1           One clear advantage of illustrating the two approaches in the context of a unified
 2     framework is the provision for verification of the ratings generated using either
 3     approach. In the case of the
 4     ecological risk comparisons, a
 5     larger group of individual experts
 6     could be polled explicitly with
 7     respect to the ecological risk
 8     adjustment factors, and a
 9     regression analysis on that
10     database could be compared with
11     the Decision Rules developed by
12     the expert panel. Consistency of
13     the weighting factors would
14     provide an increased confidence
15     in the overall risk comparisons.
16     Correspondingly, following the
17     HEHS process, a subsequent
18     stage might be to convene a panel
19     of human health experts with        Rgure Pll-1.
20     sufficient diversity of perspectives
21     and assign them the task of weighing the information generated by the surveyed
22     experts then formulating and applying explicit decision rules to arrive at a list of health
23     risk ratings.
24
25           The two subcommittees also considered the possibility of merging human and
26     ecological risks into a single science-based ranking, but realized that the comparison of
27     health risks to ecological risks was not possible on an objective scientific basis because
28     it is fundamentally based upon  considerations of societal values.  Nevertheless, the
29     use of analogous methods for developing technical risk comparisons for both human
30     health risks and ecological risks, as described in the following chapters, should facilitate
31     the subsequent setting of societal values-based risk priorities during Problem
32     Formulation.
33
34           An additional contribution of the risk comparison methods proposed by the two
35     subcommittees lies in their transparent use of decision rules for comparing risks. The
                                      n-vi

-------
      IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

1     explicit statement of decision rules, and the relative weights or importance assigned by
2     the participants to various considerations, makes clear what assumptions, scientific
3     judgments, and values went into the final results.  In this way, users of the approaches
4     can see how the results would change if the judgments about decision rules were to
5     change.
6
7           Chapters 2 and 3 describe in detail the approaches developed by the ERS and
8     HEHS, respectively, for comparing multiple environmental risks.
9
                                           H-vii

-------
IRP Integrated Draft Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
                   CHAPTER 2:  ECOLOGICAL RISKS

                          TABLE OF CONTENTS


2.1  Background	2-1

2.2  Objectives and Approach 	2-3

2.3  EPS Ecological Risk Ranking Methodology 	2-4
      2.3.1 Overview and Rationale	2-4
      2.3.2  Ecological Risk Characterization	2-5
            2.3.2.1  Selection and Aggregation of Stressors and Ecosystems  .... 2-5
            2.3.2.2 Development of Ecological Risk Profiles	2-8
      2.3.3 Development of Relative Ranking of Ecological Risks	2-14
            2.3.3.1 Multiplicative Factors Used to Assign Stressor Risk Values .. 2-14
            2.3.3.2 Sample Calculation for the Stressor Pesticides  	2-21
      2.3.4 Sources of Uncertainty	2-21

2.4  National-Scale Ecological Risk Ranking 	2-23
      2.4.1 Results of the ERS National-Scale Ecological Risk Ranking	2-23
      2.4.2 Synthesis and Conclusions	2-28

2.5  An Effects-Backwards Methodology for Risk Rankings  	2-32

2.6  References Cited	2-36

Appendix 2A: Ecological Risk Profiles  	2-38

-------
       IRP Integrated Draft Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
 1                         CHAPTER 2. ECOLOGICAL RISKS
 2
 3
 4     2.1 Background
 5
 6          During the 1980s, the U.S. Environmental Protection Agency (EPA) initiated
 7     extramural research to develop new approaches to ecotoxicology and to conducting
 8     ecological risk assessments (Bamthouse and Suter, 1986; Levin et al., 1989; Harwell
 9     and Harwell, 1989; Kelly and Harwell, 1990).  The initial development of a methodology
10     to compare and rank ecological risks began in 1986 with the cross-agency Unfinished
1 1     Business Project (EPA 1987a, b). As part of that project, EPA convened an ecological
12     risk panel to develop and apply a methodology to rank the risks to  ecological systems of
13     31 listed "environmental problem areas." The methodology and results of the
14     ecological panel were reported in Harwell and Kelly (1986) and in EPA (1987b).  The
15     ecological panel decided that the list of environmental problem areas that had been
       developed by EPA was influenced more by programmatic considerations than by what
       an ecological system might experience. Consequently, the panel developed a
18     comprehensive list of environmental stressors, defined as those physical, chemical, or
1 9     biological changes that might affect an ecological system. Many of the stressors were
20     related directly to the list of environmental problems, but other stressors were identified
21     that were not captured by the list of environmental problems.  The ecological risk panel
22     also developed a list of 16 ecological system  types, defined so that differences in
23     exposures and/or responses to the stressors  could be identified.
24
25           In 1 989, as part of the EPA Science Advisory Board's Reducing Risk Project, the
26     SAB established a subcommittee on ecological and welfare risks,  the latter addressing
27     environmentally mediated factors affecting the quality of human life. That SAB
28     ecological subcommittee adopted and further refined the methodology of the previous
29     Unfinished Business panel and developed an updated set of relative ecological risk
30     rankings for the same set of environmental stressors (SAB, 1 990a). The summary of
31     the SAB Reducing Risk project and reports of the other subcommittees are contained in
32     the report entitled, Reducing Risk: Setting Priorities and Strategies for Environmental
33     Protection (SAB, 1990b).
                                            2-1

-------
       IRP Integrated Draft Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1           In Reducing Risk, the SAB recommended that EPA enhance the consideration of
 2     ecological risks to a level comparable to human health risks and develop guidance for
 3     implementing ecological risk assessments.  In response, the Agency established a
 4     program in the Risk Assessment Forum to develop a new framework for conducting
 5     ecological risk assessments.  Two major advances were necessary: a) while chemical
 6     ecological stressors are analogous to the chemicals considered in human health risk
 7     assessments, the domain of ecological risk assessment was broadened also to include
 8     biological and physical insults that may cause significant ecological effects; and b) while
 9     human health risk assessments address only one species and often only a single
10     endpoint (human cancers), assessment of ecological risks almost always requires
11     considering multiple endpoints because of pervasive differences across species,
12     ecological systems, stressors, and  organizational hierarchy (i.e., from species to
13     landscape and larger scales of ecosystems). Consequently, based on two
14     EPA-sponsored workshops (Harwell and Gentile, 1992; Fava et al., 1992), EPA issued
15     a new ecological risk assessment framework (EPA, 1992) which expanded on the NRC
16     (1983) Red Book risk assessment paradigm to include the full suite of natural and
17     anthropogenic stressors (i.e., not just chemicals) that affect the environment at
18     population, community, ecosystem, and landscape levels. The EPA ecological  risk
19     assessment framework calls for identification of the at-risk components of ecological
20     systems, selection of ecological endpoints, development of a conceptual model that
21     describes the ecological system  and its stress-response relationships, mutual
22     characterization of the stress regime and ecological effects in terms of the selected
23     endpoints, and integration into an overall assessment of ecological risk with explicit
24     accounting of uncertainties. Since issuing the framework, the Agency has issued
25 '   ecological risk assessment guidelines (EPA/630/R-95/002B) that expand on and modify
26     the framework document, and it has issued guidance on cumulative risk assessment
27     (U.S. EPA, 1997) that further builds on the  ecological risk framework.  Neither the
28     ecological risk assessment framework nor EPA's proposed ecological and
29     cumulative risk assessment guidelines,  however, explicitly describe the methodology
30     needed to rank the relative importance of various ecological risks. Developing an
31     improved methodology for the relative ranking of ecological risks is the focus of this
32     chapter.
33
34
                                             2-2

-------
       IRP Integrated Draft Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1     2.2 Objectives and Approach
 2
 3           As part of the SAB Integrated Risk Project (IRP), the Ecological Risks
 4     Subcommittee (ERS) was formed to address relative risks to ecological systems.  The
 5     objectives of the ERS were to:
 6
 7           a)    examine and refine the methodology for assessing relative ecological
 8                 risks;
 9
10           b)    develop criteria to rank ecological risks at the national level;
11
12           c)    apply those criteria to perform a national-level relative ranking of
13                 ecological risks;
14
15           d)    develop a methodology for implementing relative ecological risk
16                 assessments at regional or local scales; and

             e)    assist in integrating ecological risk assessments into a broader conceptual
19                 framework that includes ecological, human health, and welfare risks.
20
21           To accomplish these objectives, the ERS carefully considered the relative risk
22     methodologies developed in Unfinished Business and  Reducing Risk as well as the
23     more recent ecological risk assessment framework. The Subcommittee concluded that
24     a new methodological framework for evaluating relative ecological risks was needed
25     that expands upon the previous methodologies but adds transparency and clarity and is
26     adaptable to multiple scales of analysis. As detailed in the following sections, the new
27     relative ecological risk methodology focuses primarily on the stressor-effects model to
28     assess ecological risks, but it can also be adapted to apportion the relative contributions
29     of various stressors to an observed effect in the environment.  As a part of the
30     methodology development process, the ERS updated the list of environmental
31     stressors and ecological systems of concern, and used a matrix approach to capture
32     information about stressor-effect relationships so that both quantitative and qualitative
33     assignment of risk rankings could be accomplished.  The ERS ecological risk ranking
34     methodology was applied at the national scale, and the necessary modifications were
""3     identified to apply the methodology to regional, state, or local levels. Close dialogue

                                            2-3

-------
       IRP Integrated Draft Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1     with the IRP Human Exposure and Health Subcommittee (HEHS), charged with
 2     assessing relative human health risks, has resulted in a consistent framework for both
 3     types of environmental risks and allows explicit consideration of the linkages
 4     between ecological risks and human health risks.
 5
 6     2.3 ERS Ecological Risk Ranking Methodology
 7
 8       2.3.1  Overview and Rationale
 9
10           The basic approach proposed by the ERS for ranking the relative significance of
11     ecological risks follows the ecological risk assessment framework (U.S. EPA, 1992),
12     particularly with regard to addressing all types of ecological stressors (physical,
13     chemical,  and biological) affecting the important attributes (endpoints) of ecological
14     systems.  This approach requires consideration of the two fundamental components of
15     ecological risks for each stressor: a stress or exposure regime, and a response or
16     ecological effects regime.  The stressors, exposure patterns, ecological endpoints, and
17     ecological effects are aggregated as appropriate for the scale of the analysis,
18     recognizing that the goal of comparative ecological risk assessments is to characterize
19     the dominant relationships between environmental stressors and ecological effects.
20     For example, in the  national-scale comparative risk assessment presented here,
21     ecological endpoints and effects generally were characterized only at the ecosystem
22     and landscape levels, even though ecological effects from a stressor may be observed
23     at the population and community levels as well.  Similarly, ecosystem-specific
24     differences in response were generally noted, but the Subcommittee chose to focus on
25'    those types of ecological systems considered to be most at risk.
26
27           The first step in the comparative ecological risk assessment is to determine the
28     potential ecological  importance of each stressor at the ecosystem level or landscape
29     level for ecological systems at risk.  This ecological risk characterization is based on a
30     stress-effects profile, i.e., a graphical representation of a stressor's exposure regime
31     with its ecological effects for each at-risk ecosystem type. The stressor exposure is
32     categorized into high, medium, and low levels, normalized to actual exposures in the
33     real world, and effects distributions associated with each stressor levels are estimated.
34     The ecological effects are assigned high, medium, and low intensity categories, based
35     on criteria of the ecological significance of the effect (e.g., high-level effects reflect

                                             2-4

-------
       IRP Integrated Draft Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1     major disruptions to the fundamental structure or processes of an ecosystem). The
 2     next step in the methodology is to transform these generic ecosystem- or
 3     landscape-level stress-effect relationships into a relative ranking of risk at a specific
 4     scale (e.g., regional, national, or global) by applying a series of numerical modification
 5     factors that relate to the ecological significance of the effect. The resulting modified
 6     numerical value constitutes the risk score for a stressor at the specified scale of
 7     interest. The relative risk posed by the stressors can then be determined by comparing
 8     scores across the stressors  and by grouping these scores into qualitative risk
 9     categories (e.g., very high, high, medium, and low relative risks).  This categorization
10     reflects the semi-quantitative nature of the procedure and identifies clusters of stressors
11     with similar levels of risks. Thus, the primary import of the aggregated rankings is their
12     relative value, i.e., what stressors have relatively the same level of ecological risk, what
13     set of stressors is much higher in risk than others, and so forth. Details of the risk
14     ranking methodology and its application by the EPS to the national scale are described
15     in the remainder of this chapter.
16
17       2.3.2 Ecological Risk Characterization

19            2.3.2.1 Selection and Aggregation of Stressors and Ecosystems
20
21            The ecological risk ranking process begins by identifying the stressors that
22     potentially pose environmental risks and the set of ecosystem types for consideration of
23     the risks.  For this project, lists of stressors identified by the EPS and by the HEHS
24     were merged. Table 2-1 notes those stressors, or groups of stressors, that may affect
25     ecological and/or human health endpoints. Identifying a common set of stressors
26     facilitates direct comparison of risks to humans and ecosystems.  For the ecological risk
27     ranking process, the stressors were characterized at the level of detail sufficient to
28     distinguish between different types of stressor-effect relationships, but not so detailed
29     that an unmanageable number of stressors had to be evaluated.  For example,
30     pesticides were evaluated as a class of stressors, rather than on a
31     chemical-by-chemical basis. Some general categories of stressors were disaggregated
32     where appropriate; for example, the stressor class of habitat alteration was divided into
33     specific types of habitat alterations (e.g., habitat fragmentation, habitat conversion).
34
35

                                              2-5

-------
Table 2-1: Environmental Stressors of Concern
  Stressor                              Ecological Risk                Human Health Risk

  I. CHEMICAL STRESSORS
  Criteria Air Pollutants
          CO                                    NA                           /
          NO,                                   /                            /
          O,                                    /                            /
          SO2                                   /                            /
          Airborne Particulates                     •                             /
  Toxic Inorganics
          Hg                                    /                            /
          Other Heavy Metals                      /                            (/)
          Other Toxic Inorganics (As, Se, B)         /                            (/)
  Asbestos                                      NA                           /
  Persistent Toxic Organics                        /                            (/)
          Endocnne Disrupters                     /                            /
          Pesticides, Herbicides                    /                            /
  Environmental Tobacco Smoke                    NA                           /
  Volatile Organics                                NA                           (/)
  Toxic Anions
          Nitrate, nitrite                           NA                           /
  Acid Deposition                                 /                            NA
  Nutnents                                      s                            NA
  Radionuclides
          Radon                                 NA                           /
          Others (e.g., Cs, Sr)                      /                            /
  Oil/Fuel Leaks, Spills, and Use                    /                            /•
  Dissolved Oxygen/BOD                          /                            NA
  Acid Mine Drainage                             /                            NA
  Contaminated Ground Water                      /                            NA

  II. BIOLOGICAL STRESSORS
  Bioaerosols/Allergens                            NA                           (/)
  Human Disease Agents
          Airborne Viruses, Infectious Agents         NA                           /
          Waterbome Infectious Microbes           NA                           /
  Non-human Disease/Pest Outbreaks               /                            NA
  Introduced Species                              /                            /
  Genetically Engineered Organisms                 /                            /

  III. PHYSICAL STRESSORS
  Climate Change (global, regional, and micro-climate)  /                            /
  Noise                                          /                            /
  Light Pollution                                  /                            /
  Habitat Alteration                                                             ?
          Hydrologic Alteration                     /
          Habitat Fragmentation                    /
          Habitat Conversion                      /
          Physical Habitat Disruption                /
          Turbidity and Sedimentation               /
          Altered Fire Regime                      /
          Altered Salinity Regime                   /
  Thermal Pollution                               /                            NA
  Harvesting Living Resources                                                   ?
          Coastal                                /
          Freshwater                             s
  UV-B                                          /                            /
  EMF                                           NA                           /
                                                2-6

-------
       IRP Integrated Draft Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1           Several factors have to be considered in selecting the appropriate level of
 2     aggregation or disaggregation of stressors, including the exposure and effects data that
 3     are available, differential impacts on the ecological system types of interest, potential
 4     management options, and so forth. There are no simple rules for this classification
 5     process. For example, the Subcommittee examined alternative ways to aggregate
 6     chemicals, ranging from classifications based upon mode-of-action of toxicity to
 7     structural characteristics,  uses of the chemical, or nature of the ecological effects. No
 8     single scheme was considered ideal for the purposes of this risk ranking activity. The
 9     classifications that the Subcommittee selected for ecological stressors were based
10     primarily on the source of the chemical (e.g., criteria air pollutants), the physical and
11     chemical nature of the chemical (e.g., nutrients, radionuclides), or, in the case of
12     persistent toxic organics, the usage of the chemical (e.g., pesticides).  A more detailed
13     ecological risk ranking across  a more disaggregated set of chemical stressors could be
14     done following the same methodology as proposed here but using a different
15     classification scheme; that is, the list the Subcommittee developed (Table 2-1) is not
16     presented as the only or even the best way to classify chemical stressors.  In fact, the
17     appropriate level of stressor aggregation for relative human  health risk assessment
       often differed, as noted in Table 2-1.
19
20           The set of ecological system types that were selected by ERS (Table 2-2) were
21     modified from the set of ecological system types used in the Unfinished Business and
22     Reducing Risk projects.  The intent of this
23     classification was to capture the range of      Table 2-2. Types of Ecological Systems
24     ecological systems that occur in the U.S.,
25     but without having an unnecessarily large set
26     of ecosystem types to consider. The specific
27     set of ecological systems to use for risk
28     rankings at less than national  scales may be
29     considerably different; for example, a risk
30     ranking done on the Southeastern U.S.
31     would not need to have the category of
32     semi-arid and arid ecosystems, but it might
33     have several different categories that better
34     refine forested ecosystems (e.g., coastal hardwood forests and swamps, pinelands,
35     oak-hickory upland forests, etc.).

                                              2-7
Forests
Lakes
Rivers and Streams
Wetlands
Grasslands
Agroecosystems (Rangeland and Cropland)
Deserts and And Systems
Tundra
Estuarine and Near-Coastal Ecosystems
Marine Ecosystems

-------
        IRP Integrated Draft Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
      2.3.2.2 Development of Ecological Risk Profiles

      The foundation of the comparative risk analysis
is a set of individual ecological risk profiles (see
Appendix A) that summarize for each listed ecological
stressor the nature of the stressor and its potential
ecological effects; define what is and what is not
included within the stressor category; discuss how
reliable and certain the information base is; and
provide information on any other factors that should be
considered when the relative rankings are assigned.
Because these risk profiles are modular, they can be
revised as new information  is developed, thereby
facilitating a revaluation of the risk score for that
stressor. In short, the risk profile provides not only the
information used to develop the quantitative risk score
(discussed below), but also a summary of the level of
understanding about a stressor,  references for that
understanding, and a description of the uncertainties
associated with it.

      An important component of each ecological risk
profile is the stress-effects profile, a visual tool for
illustrating the relationship between the stressor and
effect regimes for each stressor at the ecosystem or
landscape level. In essence, the stress-effects profile
illustrates the significance of the adverse  effects on
at-risk ecosystems that are expected to occur as a
result of current exposures to the stressor.

       The first element of the stress-effects profile is
the stressor profile, i.e., a frequency distribution that
depicts the observed stressor intensity affecting the
ecological systems at risk in the  region of interest.
The at-risk ecosystems may differ for each stressor;
  In
  I
  I
        Habitat Fragmentation
         Stressor Profile
         (Profile C)
  a

      Low  Medium High
     Observed Stressor Intensity
Rgure 2-1 a: Habitat Fragmentation
Stressor Profile
        Introduced Species
         Stressor Profile
         (Profile B)
Freque
      Low  Medium High
     Observed Stressor Intensity
Figure 2-1 b: Introduced Species
Stressor Profile
    in
        Acid Deposition
         Stressor Profile
         (Profile A)
  £
 u.
 (0
       Low Medium High
     Observed Stressor Intensity
                                                2-8
                                                      Figure 2-1 c: Acid Deposition
                                                      Stressor Profile

-------
       IRP Integrated Draft Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
"7

19
20
21
22
23
24
ys
26
e.g., all aquatic and terrestrial systems in the U.S. are the at-risk ecosystems for the
stressors of habitat fragmentation (Fig. 2-1 a) and introduced species (Fig. 2-1 b), but
only certain aquatic and forest systems are at risk from the stressor of acid deposition
(Fig. 2-1 c). The x-axis of the stressor frequency distribution is normalized to a relative
scale of low, medium, and high intensity of occurrence of the stressor, based on an
estimated actual distribution of the stressor in the environment. Thus, a high level of a
stressor would relate to the highest levels found in the environment, not necessarily at
the highest levels that have been tested  in the laboratory. This procedure allows the
ecological risk rankings to be based on relevant levels of the stressor, and it allows
identification of the risk reduction that would be achieved if the stressor distribution in
the environment were reduced. The y-axis of the stressor profile reflects the relative
frequency of low, medium, and high occurrences of the stressor as experienced by the
at-risk ecological systems, normalized on a scale of 0 to 1.  Thus, if the ecosystem at
risk is, for example, estuaries, this curve would illustrate the relative frequency or
percentage of all estuaries that are exposed to high stressor intensities, the percentage
that are exposed to medium stressor levels, and the percentage that are exposed to  low
stressor levels.

      In  practice, the ERS concluded that there are only a limited number of types of
environmental stressor frequency distributions, such as log-normal, normal, and
skewed.  Thus,  in  order to simplify the ecological risk  ranking process, a template of
four types of stressor-profiles was developed (Fig. 2-2), and each stressor was
assigned to one of these stressor-profile types to represent its frequency distribution.
For example, stressor Profile A depicts the most common situation, in which there is the
highest frequency of low levels of exposures, and lowest frequency for high levels of
exposures; this  would be typical, for example, for exposures of a chemical in the
         I
                                   Standard Stressor Profiles
                                IB            ic
                 L M  H
                            L  M   H           L  M   H
                             Observed Stressor Intensity
L  M  H
27
       Figure 2-2: Standard Stressor Profiles
                                             2-9

-------
       IRP Integrated Draft Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1      environment.  Profile C represents the opposite situation, in which there is highest
 2     frequency of high-level of exposure; a case in point would be habitat fragmentation, in
 3     which most landscapes have become highly fragmented and only a relatively few have
 4     low fragmentation. Note that for a more detailed relative risk assessment, the basic
 5     template of profiles developed by the ERS could be replaced by the best available
 6     information on actual stressor distribution profiles. Also, note that the frequency
 7     distributions for a particular stressor are location- and scale-specific; for example, the
 8     frequency distribution of acid precipitation is different in the Northeastern U.S. than in
 9     the Southeast.
                                                 H
                   LMH          LMH        LMH         LMH
               Observed Stressor Intensity  Observed Stressor Intensity  Observed Stressor Intensity Observed Stressor Intensity
                              H

                              M
MH
                19
                                              H

                                              M

                                              L
H
                                                             MH
                 Observed Stressor Intensity Observed Stressor Intensity  Observed Stressor Intensity
       Figure 2-3: Standard Effects Curves

10

11            The other half of the stress-effects profile is the effects profile, i.e., the frequency
12     distribution of ecological effects associated with each of the stressor levels.  That is, the
13     effects profile shows the intensity of adverse ecological effects that would result from
14     exposure to high, medium, or low levels of a particular stressor. Again, a standard set
15     of effects profiles was developed that the ERS believes represents most environmental
16     situations (Fig. 2-3), and one profile was assigned by ERS to each stressor for the
17     national-scale ranking. For example, Profile 4 illustrates a threshold situation, in which
18     a medium-level stressor causes high-level effects, compared with Profile 10, in  which
                                               2-10

-------
       IRP Integrated Draft Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
 1     only exposure to a high-level stressor causes a
 2     high-level ecological effect.
 3
 4           Effects are evaluated for the ecological systems
 5     considered to be most susceptible to the stressor. The
 6     intent is to focus on the most important ecological
 7     endpoints for those ecological systems most at risk.  For
 8     the purposes of the ERS national ranking, the ecological
 9     endpoints were constrained to include ecosystem- and
10     landscape-level effects on the structure, function, and/or
11     composition of the system.  Examples include an
12     ecological process endpoint such as decomposition
13     rates, a community composition endpoint such as
14     species richness, and a trophic structure endpoint, such
15     as the health of a critical or habitat-creating species.
16     Because of the focus on national risks, ERS did not
1 ~*     address other potential ecological endpoints, such as the
       population levels of individual species. However, for
19     ecological risk ranking conducted on local or regional
20     scales,  population-level or other ecological endpoints
21     might supplement the ecosystem- and landscape-level
22     endpoints.
23
24           Example effects profiles are shown in Fig. 2-4.
25     Fig. 2-4a illustrates the case of habitat fragmentation, for
26     which Profile 2 was selected to reflect that even
27     high-levels of intensity of habitat fragmentation only
28     cause medium-levels of ecological effects.  This is in
29     contrast to the greater potential ecological effect that
30     could result from high levels of introduced species (Fig.
31     2-4b, Profile 1). For the acid deposition example (Fig.
32     2-4c), two profiles are shown, one for the more sensitive
33     unbuffered lakes category (Profile 7), the other for
34     forests  (Profile 15), although both effects profiles relate
35     to the same  stressor profile. Again, in more detailed
        Habitat Fragmentation
        Effect Profile
        (Profile 2)
                    i-H
     Obseived Stressor Intensity
Figure 2-4a. Habitat
Fragmentation Profile
        Introduced Species
        Effect Profile
        (Profile 1)
                     -H
      Obseived Stressor Levels
Figure 2-4b. Introduced Species
Effects Profile
        Acid Deposition
        •Effect Profiles
         (Profiles 7 and 15)

              / Forests,
             '
                    J-M  «
                          UJ
      Observed Stressor Levels
Figure 2-4c. Acid Deposition
Effects Profiles
                                               2-11

-------
        IRP Integrated Draft Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
  1      relative risk ranking assessments, a suite of
  2      stressor-effects profiles could be developed for each
  3      and every ecosystem type of importance in the area
  4      of concern, or even for each and every specific
  5      ecological endpoint affected by the stressor.  While
  6      that is beyond the scope of the present application to
  7      national relative risk ranking by the EPS, it does
  8      illustrate the flexibility of the ERS methodology for
  9      specific risk ranking applications.
10
11            The combined stressor-effects regime, then,
12      is depicted by overlaying the effects profile (i.e., the
13      distribution of the intensity of effects caused by
14      varying levels of the stressor)  onto the stressor
15      profile (i.e., the actual frequency distribution of
16      stressor in the environment of concern) (Fig.  2-5). In
17      this combined stressor-effects profile, the effects
18      profile y-axis is scaled to reflect  the intensity of the
19      ecological effect (on a scale of high, medium, and
20      low) associated with the intensities of the stressor
21      that occur in the environment. The stressor profile
22      y-axis is scaled to the frequency of occurrence (0-1)
23      of the stressor in the specific environment of
24      concern. The x-axis, however, is identical for the two
25"     profiles, i.e., reflecting low, medium, and high levels
26      of stressor intensity. Thus, even a high-intensity
27      stressor exposure (defined in  terms of the levels
28      actually experienced in the environment) might result
29      in only a low-level ecological effect (e.g., acid
30      deposition effects on forest ecosystems shown in Fig.
31      2-5c).
32
33            In general, the high ecological effects category
34      was reserved for those ecological changes that are
35      very significant, involving major changes to the
                                               2-12
  cr
  S
  u.
 CO Q.
        Habitat Fragmentation
        Stress-Effect Profile
        (Profile C2)
                     TH
                     I
 i
 I
 I
_LL
       Low  Medium High
     Observed Stressor Intensity
Figure 2-5a. Habitat
Fragmentation Stress-Effect
Profile
Mressor Frequency
P ~
Intn
St
(P

)duced Speci
ress-Effect Pr
rofileBI)
/
/


BS
ofile
-H j
1
LM "
M ,
!
L
*
\
\
\
\
E3
Low Medium High
Observed Stressor Intensity
Figure 2-5b. Introduced Species
Stress-Effect Profile
        Acid Deposition
       Stress-Effect Profile
       (Profiles A7.A15)
  f
  89
  £
  to o
                  .--TH
                    I

        Low  Medium High
      Observed Stressor Intensity
                                                            Figure 2-5c. Acid Deposition
                                                            Stress-Effect Profile

-------
       IRP Integrated Draft Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1     structure, composition, and/or function of the system; an example is a major physical
 2     disturbance that removes the major structural components of an ecological system.
 3     Medium-level ecological effects were considered to be significant changes in the
 4     ecosystem's structure, composition, and/or function; for example, the replacement of
 5     one or more native species in the community by introduced exotic species would
 6     constitute a medium-level effect.  Low-level ecological changes were those that, while
 7     detectable (i.e., non-zero), do not entail significant changes to either the structure,
 8     composition, or function of the system; an  example would be the shift in the salinity
 9     isopleth near the freshwater outflow from a canal.  The frequency of the no (or de
10     minimus) level of ecological effects was not assessed by ERS; this category applies to
11     any ecological changes that would be non-detectable.  For example, in the
12     stress-effects profile for acid deposition (Fig. 2-5c), the largest fraction of affected lakes
13     are judged to have low-level ecological consequences (i.e., detectable but not highly
14     significant changes), but there may be a larger set of lakes that do not experience any
15     detectable effects at all and therefore effects would be judged to be insignificant for the
16     ecosystem.
17
             As with the stressor regime, the effects profile can be modified if and when
19     additional data are acquired. Also, the use of multiple effects profiles, such as  in Fig.
20     2-4c, allows the ready identification of those ecological effects that  might differ  across
21     different ecosystem types, thereby highlighting differential ecological sensitivities and
22     risks. In principle, such a distinction could also be made within a particular ecosystem
23     type, showing differential effects for various ecological endpoints for that ecosystem.
24     This approach would be particularly useful for localized comparative risk assessments.
25
26           It should be noted that there are insufficient data on many types of ecological
27     responses to stressors, and much of the available information excludes major
28     constituents of ecological systems (e.g., microbes and amphibians). The
29     Subcommittee recommends using the best available information for characterizing
30     ecological effects from stressors and, if there are quite different responses among
31     different components of ecological systems, to focus on  the more sensitive or more
32     ecologically important components.  Further, as new information is acquired on
33     stress-response relationships, the database upon which the ecological risk profiles and
34     resulting relative risk rankings are based should be updated and new rankings  done
35     periodically.  Note that different stress-effects profiles might be appropriate for

                                              2-13

-------
        IRP Integrated Draft Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

  1      conducting this risk ranking exercise at regional or local scales.  That is, at a regional
  2      scale the frequency distribution of the stressor (stress profile) may well differ from the
  3      distribution at the national scale.  On the other hand, since the effects profile is based
  4      on the responses of a particular ecosystem type to the stressor, effects profiles are not
  5      different for risk rankings at different scales. This is an important distinction, as it
  6      means that an effects profile database can continually be updated as more information
  7      becomes available, and that same database could be used for different risk ranking
  8      exercises, but the stressor distribution database must be developed for each specific
  9      area of concern, e.g., a particular region or the entire nation.
10
11       2.3.3 Development of Relative Ranking of Ecological Risks
12
13            Although the stress-effects profiles described in the previous section provide an
14      initial, visual characterization of the ecological risks for vulnerable ecological systems,
15      the ERS concluded that in order to characterize relative ecological risks, it is necessary
16      to convert each  profile into a quantitative score. The Subcommittee developed such a
17      methodology that goes well beyond the relative risk ranking approaches of Unfinished
18      Business and Reducing Risk.  The ERS methodology reflects the proportion of the
19      ecological resource at-risk to a particular stressor and adjusts the risk level assigned to
20      each stressor in the context of the specific scale of concern (e.g., regional, or national).
21      In order to produce a national ranking of relative ecological risks, the ERS developed
22      several specific  risk modification factors that should be considered in making this risk
23      adjustment. These multiplicative  factors, and the rationale for selecting their specific
24      numerical values, are discussed below and summarized in Table 2-2.  For ecological
25      risk rankings done at other than the national scale, different factors, or perhaps different
26      numerical values for the same factors, might be needed.
27
28            2.3.3.1 Multiplicative Factors Used to Assign Stressor Risk Values
29
30            The multiplicative factors proposed for the national-scale ecological risk ranking
31      were developed by expert judgment, expressed through consensus of the
32      Subcommittee.  The intent was to establish the weights to apply to various
33      scale-specific considerations that would lead to a higher or lower assignment of a risk
34      score and would more accurately, and more transparently, represent the ecological risk
35      of each stressor vis a vis the other stressors. In essence,  each multiplicative factor

                                             2-14

-------
       IRP Integrated Draft Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1     addresses the question of what is there about this particular stressor that makes it more
 2     or less an ecological risk compared to another stressor.
 3
 4          At the outset, the ERS assigned a numerical value to each of the three levels of
 5     ecological effects as defined above: High = 4, Medium = 2, and Low = 1. The
 6     calculated risk score for a stressor  is quantified by applying the appropriate
 7     multiplicative factors to each of three  possible effects intensities (H, M, L), then
 8     summing the results, as discussed  in more detail below. Since the initial effects
 9     intensity scores were based on a factor-of-two basis (i.e., high = 2x medium = 4x low),
10     the multiplicative factors were scaled  so that a factor that warranted a jump in the major
11     risk category (e.g., from medium to high) would be assigned a multiplicative value of 2.
12     For example, the proportion of the  resource experiencing a high level of stress from
13     harvesting of coastal fisheries populations was considered so extensive that a jump of
14     the risk value one full category was assigned (Table 2-3), whereas the proportion of
15     ecological systems experiencing a  medium level of effects from UV-B exposure was
16     considered so low that a factor of 0.5 was assigned, resulting in one full drop in  risk
17     category.

19           Less significant factors were scaled relative to this two-fold standard.  For
20     example, a multiplicative factor of 1.1 means that the risk value would be increased by
21     a small increment, and a string of about 7 such increments would be needed among the
22     various ranking factors to equal one full category jump (i.e., 1.17). The Subcommittee
23     sought consistency across the factors, so that two factors assigned the same
24     multiplicative value were considered to be relatively equivalent in terms of the risk
25     ranking. Thus, a recovery potential requiring a centuries time scale for recovery once
26     the stressor was removed (multiplicative factor 1.25) was considered equivalent to a
27     medium level of species depletion  in  terms of ecological significance to the ranking
28     (Table 2-3).
29
30           While these multiplicative factors represent the best judgment of the
31     Subcommittee, other considerations  might result in somewhat different assigned
32     values.  The opportunity to implement the integrated ecological-human health risk
33     ranking framework through an expert opinion survey (discussed in Chapter 3) could
34     provide a verification of the weights assigned by the Subcommittee for the
35     national-scale  ecological risk ranking. Again, the transparency of the ERS methodology

                                             2-15

-------
        IRP Integrated Draft Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

  1      readily allows examination of the assumptions and the values assigned by the expert
  2      judgment process.  If it were determined that the weighting factors needed adjustments,
  3      it would be a simple process to revise the multiplicative values and recalculate the risk
  4      rankings without having to redesign the relative risk ranking methodology.
  5
  6            Once the risk ranking weighting factors are established, their application to the
  7      stressor-effects profiles is straightforward: First, each level of ecological effect is
  8      considered separately to calculate partial risk  scores. The first partial risk score for a
  9      stressor focuses on the high level of ecological effects (initial value of 4).  The
10      proportion of resources affected at the high intensity of effects (if any) is assessed,
11      based on the stress-effects profiles (Fig. 2-5). This is followed by each of the other
12      weighting factors described below, all multiplied together to develop a partial risk score
13      reflecting the contribution to the total risk ranking score from the high intensity of the
14      stressor. If the stress-effects profile indicates that no ecological systems experience
15      the high level of ecological effects from that stressor, even when the stressor is at the
16      high level of intensity (e.g., habitat fragmentation profile [C2] shown in Fig. 2-5a and the
17      acid deposition-forest profile [A15] shown in Fig. 2-5c), then the multiplier of zero will
18      apply for the factor "proportion of resources at risk."  Consequently, the assigned partial
19      risk contribution from the high intensity effect level is zero. This illustrates an important
20      reason why the weighting factors are multiplicative rather than additive; i.e., if any factor
21      is assigned a value of 0, then that partial risk score is 0.
22
23            Similarly, the second partial risk score is calculated for the medium level of
24      ecological effects caused by the stressor (which begins with an initial assigned value of
25'     2).  The third partial risk score is based on the low level of ecological effects caused by
26      the stressor (which begins with  an initial assigned value of 1).  In these cases, the
27      partial risk scores are calculated by applying the appropriate multiplicative factors in the
28      same manner as described for the high-level partial risk score.
29
30            Finally, the three partial risk scores (for the H, M, and L levels of effects) are
31      summed to result in the final total risk ranking score for the specific stressor, calculated
32      for the specific scale of concern (in our case, the national scale).  The details of the
33      partial and total risk ranking scores developed by ERS are shown in a matrix form in
34      Table 2-4.
35

                                               2-16

-------
            Table 2-3. Multiplicative Factors for Deriving National Risk Rankings
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17

19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
Factor (Assessed at H/M/L Effect Level)
1. Proportion of Resource At Risk
>60 percent
30-60 percent
10-30 percent
1-10 percent
<1 percent
effect level (H/M/L) does not occur
2. Existence of Hot Spots
Nationally distributed
Regional
Local or None
3. Recovery Potential
Irreversible or >centunes
Centuries
Decades
< Decades
4. Duration of Stress-Effects
>centunes
Decades

-------
        IRP Integrated Draft Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1            The following sections describe each multiplicative factor and its assigned
 2     weighting values for national-scale rankings.
 3
 4            a) Proportion of Resource at Risk - The first entry into the risk scoring process
 5     characterizes the proportion of the ecological resource that experiences effects at the
 6     specified exposure intensity.  This information is the same as shown graphically in the
 7     stress-effects profile for each stressor. The question asked is what proportion of the
 8     ecological systems at risk in the region of concern for the ranking (e.g., the proportion of
 9     all lakes in the nation) is exposed to the stressor such that the high level of ecological
10     effects is realized.  If the assigned effects distribution does not show severe ecological
11     effects at high intensities of the stressor, then this proportionality factor is assigned the
12     value of zero. If high-level ecological effects do occur, then the proportion of the
13     ecological systems of concern  experiencing high ecological effects is estimated as one
14     of the following categories: <1 %, 1-10%, 10-30%, 30-60%, or >60%. Then the
15     multiplicative factor (Table 2-3) is applied based on that proportion category. For
16     example, if more than 60% of the at-risk ecological resources experience the high
17     effect, then the factor would be 2x.  On the other hand, if less than 1 % experienced that
18     level of effect, then the multiplicative factor assigned would be 0.25x.  Similarly, the
19     same question is asked for the proportion  of the ecological resource that experiences
20     medium-level effects from the stressor, and finally the proportion at the low-level of
21     effects.  For  example, in the case of habitat fragmentation, shown in Fig. 2-5a, no
22     ecological systems experience high-level ecological effects, even at high levels of
23     fragmentation, so the partial risk score for the high-level effects is 0. The medium-level
24     effect results from high levels of habitat fragmentation, which the figure illustrates
25 .   occurs at more than 60% of the systems, resulting in the initial value for medium effects
26     (2) being multiplied by 2x (for the >60% proportionality factor). Similarly, only low-level
27     effects occurs at medium or at low levels of habitat fragmentation,  and those conditions
28     occur 10-30% of the systems,  resulting in  a multiplicative factor of  1 x for the low effects
29     partial risk score.
30
31            b) Distribution of "hot  spots" - This multiplicative factor adjusts the risk score
32     if the stressor is distributed in localized hot spots, with disproportionate ecological
33      effects from  what would be predicted from the stressor frequency distribution profile.
34     This approach acknowledges that adverse ecological effects may be significant in
35      highly localized areas and, therefore, warrant national concern even though they do not

                                              2-18

-------
       IRP Integrated Draft Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1      affect a large proportion of the resource and thus would not show up in the profile (?).
 2     This factor is assigned at the highest level of effect for which ecological effects occur
 3     (i.e., at the high level if there are high effects, at the medium level if there are medium
 4     but no high effects, and at the low level if there are no medium effects).  If no or very
 5     few hot spots occur, the multiplication factor is 1.0; if hot spots do occur within a few
 6     regions of the nation, the multiplication factor is 1.25; and if they occur in most regions
 7     of the nation, the factor is 1.5.  For example, a 1.5 factor for hot spots was assigned to
 8     the stressor heavy metals, which occur at high concentrations in localized areas across
 9     the nation.
10
11           c) Recovery potential - This multiplicative factor relates to ecosystem
12     resilience, i.e., the hypothetical, relative potential of the affected ecosystem to recover if
13     the stressor were to be removed, even if such removal does not appear realistic given
14     physical, societal, or economic constraints.  The factor reflects the fact that effects that
15     are irreversible or very  long-lasting (e.g., over geological time) are more significant
16     ecologically than those effects that can be reversed quickly. For example, the recovery
17     factor for nutrient effects on estuaries was assigned a value of 0.75 since the ERS
       estimated that once the nutrient inputs were removed, the system would recover within
19     a decade.
20
21            d) Duration of the stress-effect - This multiplicative factor takes into account
22     the time into the future that the  stress could be expected to occur, so that an adverse
23     effect that is expected to last a  long time would be given a higher risk assignment than
24     one that is of short duration. For example, hydrologic modification was assigned the
25     duration factor of 1.25, in part because major changes in streambeds, such as from
26     dams, are expected to persist for centuries.
27
28            e) Secondary stress induction - This multiplicative factor relates to the number
29     and strength of interactions between the selected stressor and other stressors on the
30     list, i.e., if the stressor  induces or predisposes the occurrence of secondary stressors.
31     The purpose of this multiplicative factor is to note those stressors that create cascading
32     effects through the induction of other stressors.  If two or more significant interactions
33     exist, the score is multiplied by a factor of 1.1; otherwise, the factor is 1.0. For
                                              2-19

-------
       IRP Integrated Draft Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1     example, hydrologic alteration would be expected to cause increases in the nutrient,
 2     sedimentation, salinity, and other stressors, so its secondary stress induction factor is
 3     1.1.
 4
 5           f) Species depletion - This multiplicative factor relates to the potential for the
 6     stressor to result in depletion of species, from local loss of a population (extirpation) up
 7     to global extinction of a species. The species depletion factor is included because of
 8     the importance of the loss of species and the reduction in biodiversity that has occurred
 9     during the past century, significantly affecting the structure,  composition, and function of
10     many ecological systems. It also has particular importance to society in that some
11     species are of special concern (e.g., threatened or endangered species). The
12     Subcommittee defined high, medium, and low-level species depletion effects as follows:
13
14           High-level effects: extinction or extirpation of many species, resulting in loss of
15           species diversity or richness at one or more ecological systems; an example of
15           high-level species depletion occurs for high intensities of habitat conversion,
17           such as when  grasslands are converted to agricultural systems.
18
19            Medium-level effects: extinction or extirpation of a number of ecologically and/or
20            societally important species, but not necessarily resulting in overall decreased
21            species diversity; an  example of medium-level species depletion is when
22            pesticides occurring at high intensities cause the loss of many species of insects
23            or other invertebrates in an ecological system adjacent to an agricultural field.
24
25            Low-level effects: extinction or extirpation of one to a few species; an example is
26            the loss of a single species because of a pest or disease outbreak.
27
28            g) Special ecological significance - This multiplicative factor was developed to
29      capture any special significance of the ecological effects from the stressor on the
30      affected ecosystems and/or ecological endpoints. One such modification captures the
31      exceptional importance of certain ecological attributes, i.e., ecosystems or effects that
32      are disproportionate  in importance relative to their spatial extent or frequency of
33      occurrence.  For example, effects that are particularly consequential to coastal wetlands
34      as a group, or a stressor that could eliminate an entire class of ecological system, such
 35     as vernal ponds, or that has broad-based effects on a group of organisms (e.g., all

                                               2-20

-------
       IRP Integrated Draft Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1     reptiles) would trigger the special ecological significance multiplier. This multiplier was
 2     reserved for ecological risk considerations not captured by the other factors and was
 3     only infrequently invoked.
 4
 5           2.3.3.2 Sample Calculation for the Stressor Pesticides
 6
 7           As discussed above, the quantitative step in the ecological risk ranking
 8     methodology is to apply the multiplicative adjustment factors to the initial scores for
 9     each of the high, medium, and low levels of effects occurrence. To demonstrate this
10     process, consider the calculation of the total risk ranking score for the stressor
11     pesticides (see Tables 2-3 and 2-4). For the high-level effects partial risk score, the
12     Subcommittee concluded that the high intensity of ecological effects (initial value 4)
13     occur only in 1-10% of ecosystems (thus, the multiplicative factor is 0.5); hot spots are
14     distributed nationally (multiplicative factor is 1.5); recovery would be expected to occur
15     in decades (multiplicative factor is  1.0); the duration of stressor-effects would likely be
16     decadal (multiplicative factor is  1.1); there are no secondary induced stressors
'7     (multiplicative factor is 1.0); species depletion is highly significant (multiplicative factor is
 j     1.25); and there are disproportionate effects on higher trophic-levels and critical species
19     (multiplicative factor is1.1). Thus, the application of multiplicative factors results in a
20     modified partial risk score of 4.54 for the high-intensity effect.  Similarly, calculations
21     were done for the medium-intensity partial risk score (product = 1.10) and for the
22     low-intensity partial risk score (product =  1.10). These three partial risk calculations
23     were then summed to produce a total risk ranking score of 6.74. To complete the risk
24     ranking process, this total risk score for pesticides is compared to the total risk scores
25"    calculated for all the other stressors of concern.
26
27       2.3.4 Sources of Uncertainty
28
29            There are many sources of  uncertainty inherent in relative ranking of risks.
30     Some causes of uncertainty are reducible through further research and better scientific
31     information, such as having an  improved understanding of stress-effect relationships for
32     a specific stressor affecting a specific ecological endpoint.  Other sources of uncertainty
33     are essentially irreducible, such as those caused by the intrinsic complexities and
34     variability of natural ecosystems.  Uncertainty is only important here, however, if it
35     significantly changes  the relative assignment of risk estimates across the stressors.

                                              2-21

-------
       IRP Integrated Draft Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1           Examples of the sources of uncertainty in the ERS national ecological risk
 2     ranking process include:
 3
 4           a)     aggregation of stressors for the analysis into a class of stressor, i.e.,
 5                  grouping of different specific stressors with different exposure or effects
 6                  characteristics - an example is the category "persistent toxic organics";
 7
 8           b)     aggregation across specific ecological system types, i.e., inclusion of
 9                  different specific ecosystems with different effects characteristics;
10
11           c)     aggregation across specific ecological endpoints;
12
13           d)     lack of information on the exposure regime;
14
15           e)     lack of information on ecological effects;
16
17           f)     interactions among multiple stressors; and
18
19           g)     composition and expertise of the members of the ranking panel.
20
21           Clearly, the ERS process of developing a national-scale ranking was limited by
22     the availability of information and time for the exercise, factors that could be mitigated
23     by a concerted effort within EPA or another institution to acquire and analyze the full set
24     of available information on each stressor. A major advantage of a comparative risk
25     ranking process is that many of the associated uncertainties do not make a difference
26     in the ranking, whereas a process of doing an assignment of absolute risk values would
27     be much more subject to uncertainties and much less reproducible. Moreover, when
28     the risk scores are converted to qualitative characterizations of risk, as opposed to
29     numerical risk values, the importance of uncertainties vis a vis the overall ranking is
30     even further reduced.  In practice, the Subcommittee found that the quantitative
31     ecological risk scores for each stressor tended to clump into distinct ranges, allowing
32     separation of the stressors into qualitative categories of risk. In most cases,
33     uncertainties in specific multiplication factors applied to a stressor did not cause a
34     reassignment of the stressor from one major, qualitative category to another. This
35     significantly enhances the confidence that can  be placed on the relative rankings. As

                                             2-22

-------
       IRP Integrated Draft Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1     noted in the next section, however, in some cases uncertainties may be so great that
 2     risk associated with a particular stressor (e.g., endocrine disrupting chemicals and
 3     genetically engineered organisms) cannot be ranked.
 4
 5     2.4 National-Scale Ecological Risk Ranking
 6
 7       2.4.1 Results of the ERS National-Scale Ecological Risk Ranking
 8
 9           In order to apply the ecological risk ranking methodology to the national scale of
10     relevance, the ERS developed ecological stressor-effect profiles (Appendix 2A) for 32
11     of the stressors listed in Table 2-1 and scored each stressor using the partial and total
12     ecological risk scores procedures described above. The results of the ERS
13     national-scale risk ranking process are summarized in Table 2-4. The table shows for
14     each stressor the primary ecosystem type considered by ERS to be at risk.  The
15     stressor-effect profile curves are  next listed to indicate the stressor and effects
16     distributions that the Subcommittee considered was appropriate for that stressor at the
       national scale.  Then, for each effect intensity level, the partial risk scores are
       calculated by taking the intensity value and multiplying by the number assigned at each
19     cell along the row, resulting in the partial scores at the right of the table. These three
20     partial scores are added and shown in the "sum score" column. The next entry in the
21     table is the ERS's judgment as to whether the trend for risk from the stressor is
22     increasing, decreasing, or remaining the same nationally; for example, the risks from
23     metals were judged to be declining, but the risks from climate change were judged to be
24     increasing. Additionally, ERS identified the potential for surprises in its assignment of
25     risk levels; for example, ERS judged there to be little likelihood of surprises for nutrient
26     additions to the environment, but a high potential for surprises from introduced exotic
27     species.  Note that the latter two  columns (trends and surprises) were added to
28     characterize further the Subcommittee's judgment of the stressor, but were not
29     incorporated into the relative risk rankings per se.  Finally, it should be repeated that: a)
30     ERS assigned the various values based on the assumption that current environmental
31     laws and regulations continue to  be enforced; b) these relative rankings apply only at
32     the national scale and might be quite different for regional or local scales; c) only
33     ecosystem- and landscape-level  endpoints were considered; d) these rankings were
34     based on the understanding of the  ERS members and did not entail analysis of
?K     extensive data or models; and e) limitations in the available data may in some cases

                                           2-23

-------
Table 2-4: Summary of Ecological Risk Ranking Scores
stressor
NOx
ecosystem type stress- effects
at risk effects Intensity
curve
terrestrial A15


1
2
4
proportion
of resource
0.25
0
0
hot
spots
1


recovery duration secondary species eco
stress depletion slgnlf
Induction
0.75 1 1


1 1


Partial
SCORE
019
000
000
SUM trends
SCORE
0.19 1
potential for
surprises
L

tropospheric
ozone
forests, shrubs A10
and grasslands

1
2
4
OS
0.25
0.25
1
1
1
075 11 1
075 1.1 1
0 75 11 1
1 1
1 1
1 1
041
0.41
083
1.65 =
L

tropospheric
ozone
agric A7


1
2
4
1
0.5
025
1
1
1
0.75 1 1
0.75 1.1 1
0.75 1.1 1
1 1
1 1
1 1
0.75
0.83
0.83
2.40 =
L

SO2
terrestrial A12


1
2
4
0.25
0
0
1


1 1.1 1


1 1


028
000
000
0.28 1
L

Hg
freshwater and esturarie: A1 2


1
2
4
0.25
0
0
1.25


1 11 1


1 1


034
000
000
0.34 ?
UH?

other
heavy metals
freshwater and esturartes A13


1
2
4
025
025
0.25
1
1
1.5
1 1.
1 1
1 1.

toxic
Inorganics
(As. Se. B)
freshwater, eslurarles A10
and agriculture

1
2
4
0.25
025
0.25
1
1
125

.


persistent toxic
organlcs
aquatic and terrestrial A10


1
2
4
05
0.25
0.25
1
1
1.5
f
.
.

pesticides
aquatic and terrestrial A7


1
2
4
1
05
0.5
1
1
1.5
1
1



1
1
1
028
055
165
2.48 1
H



1
1
1
1 1
028
0.55
1.51
2.34 =
H

1
1
1
1
1
1 1
0.55
0.55
165
2.75 1
H

1
1
1 .1 1
1 1
1 1
1.25 1.1
1 10
1.10
4.54
6.74 A
H

-------
                Table 2-4 (continued)
w
stressor
acid deposition
ecosystem type stress- effects
at risk effects Intensity
curve
lakes A7


1
2
4
proportion
of resource
1
0.5
0.25
hot
spots
1
1
1
recovery duration secondary
stress
Induction
1 1 1
1 1.1 1
1 1.1 11
species eco
depletion slgnlf
1
1
1
1
1
1
Partial SUM trends potential for
SCORE SCORE surprises
1.00 3.31 |
1.10 V
1.21
L

acid deposition
forest A15


1
2
4
0.25
0
0
1


125 1.1 11


1


1


0 38 0.38 1
0.00 T
000
L
,
nutrients
freshwater and esturaries B19


1
2
4
1.5
1
0.25
1
1
1 5
075 1.1 1
075 11 1
075 1.1 11
1
1
1
1
1
1
1.24 4.25
1.65
1.36
L

radlonuclldes
aquallc and terrestrial A3


1
2
4
0.25
0
0
1


1 1 1


1


1


0 25 0.25 -
000
0.00
'

oil spills
freshwater, eslu rarl as A 1 1
and tundra

1
2
4
0.25
0.25
0
1
1

1 1 1
1 1 1

1
1

1
125

0 25 0.88
0.63
0.00
L
DO/BOD



acid mine
drainage

freshwater and esturarle: B1 9



freshwater A7


1
2
4

1
2
4
0.5
05
0.25

0.25
0.25
0.25
1
1
1.5

1
1
125
075
075
075

075
0.75
075
1 1
1 1
1.1 1.1

1 1 1
1 1 1
11 1
1 0.38
1 0.75
1 136

1 0.21
1 041
1 103
2.49 I
V


1.65 I
T

L



L


contaminated
groundwater
freshwater and eslurarles A12


1
2
4
025
0
0
1


075


1.25 1


1


1


0 23 0.23 ?
0.00
000
L

disease/pest
outbreaks
aquallc and terrestrial A7


1
2
4
0.5
0.5
0.5
1
1
1.5
1
1
1
1 1
1 1
1 1 1
1
1
1
1
1
1
0 50 4.80 =
100
3.30
H

-------
Table 2-4 (continued)















1
N
^


















stressor
Introduced
exotic
species

climate change



noise







hydrologlc
alteration


habitat
fragmentation


habitat
conversion


physical habitat
disruption


turbidity
sedimentation

ecosystem type stress-
at risk effects
curve
aquatic and terrestrial B1



aqualic and terrestrial C7



aquatic and terrestrial A1 2
(fauna only)


aquatic and terrestrial A1 2
(fauna only)


streams, wetlands. D7
and estuaries


aquatic and terrestrial C2



aquatic and terrestrial C7



aquatic and terrrestrial A1



aquatic and wetland B1


effects
Intensity
1
2
4

1
2
4

1
2
4

1
2
4

1
2
4

1
2
4

1
2
4

1
2
4

1
2
4
proportion
of resource
1
15
1

1
1
1.5

025
0
0

025
0
0

1 5
0.5
1.5

1
2
0

0
1
1.5

1
05
025

1
1.5
1
hot
spots








1



1



1
1
1

1
1



1
1

1
1
15

1
1
1
recovery
1
1
1

1
1
1

1



1



075
1
1.25

1
1










0.75
0.75
1
duration secondary
stress
Induction
1 1
1 25 1
1.25 1.1

1.25 1
1 25 1.1
1.25 1.1

1 1



1 1



1.25 1
1.25 1.1
1.25 1.1

1.25 1
1 25 1.1



1.25 1.1
1.25 1.1

1
1.1
1.1

1
1 1
1.1
F species eco
depletion slgnlf
1
1.25
1.5

1
1
1.25

1 1



1 1



1 1
125 1
1.5 1.1

1 1
1.25 1



1.25
15

1
1
1

1
1
1
Partial
SCORE
1.00
4.69
825

1.25
2.75
1031

0.25
000
000

025
000
000

141
1.72
17.02

125
688
000

000
344
12.38

1 10
1.21
182

083
272
484
SUM trends |
SCORE
13.94 A
T


14.31 A
T


0.25 A
T


0.25 A
T


20.14 =



8.13 A
T


15.81 A
T


4.13 A
T


8.39


totentlal for
surprises
H



H



L



H



L



L



L



L



L



-------
Table 2-4 (continued)
stressor
altered
fire regime
ecosystem type
at risk
terrestrial and wetlands
stress-
effects
curve
A13
effects
Intensity
1
2
4
proportion
of resource
0.5
0.5
0.25
hot
spots
1
1
1
recovery duration secondary species eco
stress depletion slgnlf
Induction
1 1
1 1
1 1
.1 1
1 1.1
1 1.1

altered
salinity regime
freshwater, esturaries
and wetlands
A7
1
2
4
1
025
0.25
1
1
1
1
t
1


freshwater
populations at-risk
68


1
2
4
1
1
0
1
1

0.75
0.75

1
1
1.1 1
1
1
1
Partial
SCORE
055
1.21
121
SUM trends
SCORE
2.97 II
potential tor
surprises
L

1
1
1.1
1.10
055
133
2.98 A
L

1
1 1

t
1

083
1.65
000
2.48 A
L


aquatic and terrestrial
Including ocean




2
4

0.5
0
1
1

0.75 1 25 1
0.75 125 1 1


1
1

0.94
0.94
0.00
1.88 A
H

-------
IRP Integrated Draft Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
                                            Table 2-5: Relative Ranking of Ecological
                                            Risks at the National Scale
 1     influence the risk rankings; for example,
 2     some chemicals may rank low for ecological
 3     risks because there are few data on
 4     ecological effects other than for the aquatic
 5     and pesticide endpoints required by law.
 6     Nevertheless, the Subcommittee is
 7     reasonably confident that the overall
 8     ecological risk rankings accurately reflect the
 9     relative risks to ecological systems in the
10     U.S. from anthropogenic stressors.
11
12       2.4.2 Synthesis and Conclusions
13
14            The results from the relative ranking
15     of national ecological risks derived by the
16     ERS using the stressor-based methodology
17     are shown in Table 2-5 and Figure 2-7. The
18     process for assigning the stressors into
19     qualitative risk categories was to rank order
20     the stressor total risk scores and identify,
21     where possible, distinct breaks between
22     groups of stressors. Inspection of Figure 2-7
23     shows the clear discontinuities that were
24     used to group the stressors into the four
25     qualitative ecological risk categories (highest
26     risks, high risks, medium risks, and low
27     risks).  Note that the division between
28     medium and low risks was less distinct and,
29     therefore, somewhat more arbitrary than the
30     distinction between the highest and high
31     categories, and the distinction between the
32     high and medium categories. Other
33     aggregations or dividing  lines could be
34     assigned to these numerical rankings, but
35     the important consideration is that there is a clear distinction across the stressors, so
                                              HIGHEST ECOLOGICAL RISKS
                                                    hydrologic alterations
                                                    harvesting marine living resources
                                                    habitat conversion
                                                    climate change
                                                    introduction of exotic species
                                              HIGH ECOLOGICAL RISKS
                                                    turbidity/sedimentation
                                                    habitat fragmentation
                                                    pesticides
                                              MEDIUM ECOLOGICAL RISKS
                                                    disease/pest outbreaks
                                                    nutrient additions
                                                    physical habitat disruption
                                                    acid deposition (lakes)
                                                    altered fire regime
                                                    altered salinity regime
                                                    persistent toxic organics
                                                    heavy metals other than Hg
                                                    DO/BOD
                                                    harvesting freshwater living
                                                           resources
                                                    troposphenc ozone
                                                    toxic inorganics
                                                    UV-B
                                                    acid mine drainage
                                              Low ECOLOGICAL RISKS
                                                    oil spills
                                                    acid deposition (forests)
                                                    Hg
                                                    S02
                                                    radionuclides
                                                    noise pollution
                                                    light pollution
                                                    groundwater contamination
                                                    thermal pollution
                                                    NO,
                                              UNKNOWN BUT POTENTIALLY
                                               IMPORTANT RISKS
                                                    endocrine disrupters
                                                    genetically engineered organisms
                                        2-28

-------
N
i
M
                       Figure 2-7. Ecological Risk Ranking Scores
     25
        Highest Risks
     20
     15
            tl
     E
10
      5 -
            High
            Risks
               t
     Medium Risks
L
           Low Risks
                                                       •sum score
                                                  U.tt.u
       I
                                      I  i f i i I !
             i
1 !  I i
                               r I
                                      I
  «  =
  I
« 1

                                        •8  s
                                          E
                                          E
                                   Stressor

-------
        IRP Integrated Draft Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

  1      that, for example, the assessment of the ecological risks from hydrologic alteration was
  2      clearly quite different from the assessment of ecological risks from oil spills. The
  3      Subcommittee believes that its assignments of relative risks into the qualitative
  4      categories in Figure 2-7 is a reasonable interpretation of the data.
  5
  6            Further, the Subcommittee believes this categorization is relatively robust, as
  7      demonstrated by an exercise EPS conducted of substitution of alternate plausible
  8      values for weighting factors assigned to specific stressors, i.e.,  a type of sensitivity
  9      analysis. In all cases the Subcommittee examined, to shift significantly the position of a
10      particular stressor relative to the other stressors would require changes in the
11      multiplicative factors that  do not seem warranted based on the  information available.
12      Obviously, as new or more complete data become available, or as the exercise is
13      applied to specific regions or locales, then there may be shifts in the relative risk
14      rankings.
15
16            Consistent and at least semi-quantitative relationships are maintained in these
17      final, total risk categorizations:  The initial entry value of 4.0 that was assigned to
18      high-level ecosystem effects has now been modified so that all total risk scores above
19      5.0 are assigned to the high risk category, and all total risk scores above 10 are in the
20      highest ecological risk category.  Similarly, from the initial value of 2.0 for medium-level
21      effects, the total risk scores in the range of 1.0 to 5.0 now fall into the medium
22      ecological risk category. And anything less than a total ecological risk score of 1.0 was
23      considered to be in the low ecological risk category.
24
25            At the national scale, the highest level of ecological risk was assigned by ERS to
26      two classes of habitat alteration (hydrologic modification and habitat conversion), as
27      well as to climate  change, introduced exotic species, and over-exploitation of living
28      marine resources through harvesting. Grouped at the high level of ecological risk are
29      turbidity/sedimentation, habitat fragmentation, and pesticides. The medium ecological
30      risk category includes disease and pest outbreaks, nutrients, physical habitat disruption,
31      acid deposition (lakes), and altered fire or salinity regimes. Note that two stressor
32      categories (endocrine disrupters and genetically engineered organisms) did not have
33      calculated risk scores and were not assigned to a narrative risk category because the
34      Subcommittee believed the potential for ecological risks may be very high but is
35      currently largely unknown. The Subcommittee, however, felt that it was important to

                                              2-30

-------
       IRP Integrated Draft Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1     distinguish between unknown, but potentially important, risks and low risks. Further
 2     research and experience with these stressors may lead at some point in the future to a
 3     confident assignment of these to a risk category.
 4
 5           The results shown in Table 2-5 indicate that the greatest ecological  risks in the
 6     U.S. relate predominantly to physical and biological stressors, not to chemical
 7     stressors.  That conclusion is consistent with the relative ecological risk ranking of the
 8     Unfinished Business and Reducing Risk projects (EPA 1990a, 1990b; Harwell et al.,
 9     1992).
10
11           It should be recalled that the risk ranking process done by ERS focused on the
12     national scale of concern and explicitly considers that current environmental laws and
13     regulations will continue to be enforced.  At less than national scales, the ecological risk
14     rankings might be different; for example, at a local level the most important risk might
15     be the acid drainage from a mine entering a valuable freshwater stream. The national
16     scale of ranking is meant to look more broadly, and such local-scale risks only become
•|~7     nationally important if captured by the criteria for scaling-up; that is,  acid mine drainage
       would have had a higher national-scale ranking if it occurred at many locations all
19     across the nation.
20
21           Another noteworthy conclusion is that many environmental problems that have a
22     high public perception of risks and/or have a high allocation of resources for
23     environmental protection and management may, in fact, constitute relatively lower
24     ecological risk as nationally ranked. Included in this are ecological risks from
25     radionuclides, oil spills, heavy metals, toxic organic and inorganic chemicals, and SO2
26     and NOX air pollution. Again, this conclusion is consistent with Reducing Risk, and
27     comparisons of its ecological risk rankings with a poll of public perceptions of ecological
28     risks (Harwell et al., 1992).
29
30           There are at least two possible reasons for this discrepancy:  The first plausible
31     reason is the success that the environmental regulatory system has had over the past
32     several decades  The ERS risk ranking was predicated on the assumption that existing
33     environmental laws and regulations would remain in effect and continue to be enforced;
34     the same risk ranking done in the 1960s might have given quite different results. In that
35     case,  risks that have been reduced may not yet be perceived as having been  reduced,

                                            2-31

-------
        IRP Integrated Draft Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1      illustrating a time-lag in the perception of environmental successes. Second, there is
 2      an anthropocentric focus of concern about environmental risk agents, i.e., historically
 3      there has been primary attention to human health effects rather than to ecological
 4      effects from environmental stressors, so the public has much better understanding of
 5      those risks. Both of these discrepancies can be mitigated by increased education of
 6      the public on the scientific bases of risk ranking.
 7
 8            The results of the SAB ERS relative ecological risk rankings have considerable
 9      implications for the ability of EPA and other regulatory and resource management
10      agencies to reduce risks to ecological systems.  If society desires to restore, protect,
11      and sustain the environment with high ecological quality, then the remaining highest
12      priority risks to ecological resources must be a focus of attention. At present, the
13      highest ranked risks are not well regulated by EPA. In addition, examination of the risk
14      rankings in conjunction with expected trends in the associated stressors (Fig. 2-7)
15      highlights the fact that, in the Subcommittee's assessment, many of the top-ranked
16      ecological risks are the very ones that are predicted to experience an upward trend.
17
18      2.5 An Effects-Backwards Methodology for Risk Rankings
19
20            In the ecological risk ranking procedure described above, the major ecological
21      effects caused by a stressor or group of stressors were analyzed. The logical
22      progression began with the stressor, then proceeded to rank the resulting ecological
23      effects.  It is often useful, however, to apportion the major, multiple causes of a
24      particular adverse ecological effect, i.e., to work backwards from an observed effect to
25'     estimate the relative contribution to that ecological effect from  each of several
26      stressors.  In this section, we briefly present the outline of a methodology for beginning
27      with an effect and determining the relative contribution of various stressors to the effect
28      at a national, regional, local, or other scale of interest.
29
30            The first step in an effects-backwards relative risk assessment is to identify and
31      characterize the ecological effect, including the spatial and temporal scale of interest
32      and the resources or habitat types affected. For example, the causes of species
33      depletion, defined as loss of native species within the exposed ecosystems for all types
34      of ecosystems nationwide,  might be analyzed. A smaller-scale example could be a
35      reduction in forest ecosystem productivity for a particular watershed, which might be

                                              2-32

-------
       IRP Integrated Draft Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1     caused by more than one stressor.
 2
 3           The second step in the procedure would be to identify all of the major stressors
 4     for which a cause-and-effect linkage could be established or reasonably hypothesized.
 5     For this purpose, the stressor list presented in Table 2-1 could be used as a starting
 6     point. Depending upon the particular ecological effect being analyzed, groups of
 7     stressors could be aggregated to simplify the analysis. Logical aggregations might
 8     consist of stressors likely to occur in the same locations or groups of similar classes
 9     (such as chemicals) whose independent effects may be small but whose cumulative
10     effects may be significant.
11
12           Once the stressor list is developed, a stress-effect profile, similar to those
13     described earlier, could be used to establish the relative strength of each stressor-effect
14     relationship. If desired, a screening-level analysis could be performed by considering
15     only those stressors that exhibit a strong relationship to the effect. Note that in most
16     cases, the stress-effect profiles generated in this context will not be the same as those
•1 "*     developed for the stressor-based analysis because in the latter case the effects profile
       represents an aggregation of multiple ecosystem (or landscape-level)  effects, rather
19     than the occurrence of a particular effect.  This is because the adverse effects on
20     structure, composition, and function of one or more ecosystem types or landscapes
21     were lumped to streamline the national analysis.
22
23           In order to complete the effects-backwards stress-effect profile, it is important
24     explicitly to define high, medium, and low-level effects. In many cases, this step will not
25     be trivial because it entails quantifying or otherwise describing the total effect in order to
26     normalize the y-axis and thereby derive the high, medium, and low effect ranges.  In the
27     species depletion example, therefore, the number of species reported extirpated or at
28     significant risk of extirpation within the next decade in all types of terrestrial and aquatic
29     systems could be estimated to provide the 100% line in the  normalization.  High,
30     medium, and low levels  of effect could then be defined accordingly.  Alternatively, a
31     rough measure of species at-risk or already affected might be developed for a
32     screening-level analysis; for example:
33
34           high-level effects:        the extinction or extirpation of a number of different
35                                    species in different types of ecological systems

                                             2-33

-------
        IRP Integrated Draft Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1                                     located in many areas;
 2
 3            medium-level effects:      the extinction or extirpation of a few species, species
 4                                     in only a few ecological systems, or species only in a
 5                                     few localized areas; and
 6
 7            low-level effects:          extinction or extirpation of only a few species in a few
 8                                     ecological systems in selected locations.
 9
10     In the forest ecosystem productivity example, data may be available to assess the
11     decrease in forest productivity, defined here as the decrease in the amount of energy
12     stored by the forest ecosystem, using measurements of tree diameter, height, volume,
13     basal area, or mean annual increments of growth.  High, medium, and low levels of
14     effect would then be normalized to this total.
15
16            Using the information in this stress-effect profile, the next step in the
17     effects-backwards relative risk methodology is to assign an ecosystem- or
18     landscape-level partial score and modify that score with appropriate multiplicative
19     factors as in the stressor-based methodology.  Once a numerical score for each
20     significant stressor is developed, the stressor-effect scores can be summed.  The
21     proportion of each stressor's partial score relative to the total score provides a measure
22     of the relative  importance of the stressor in producing the effect of concern. Note that
23     the particular multiplicative factors may well be different from those presented in the
24     national risk ranking methodology above, because of the characteristics of the particular
25     effect or scale of interest. For example, for species depletion at the national scale, the
26     spatial extent, hot spot, and ecological significance modification factors might remain
27     the same. Other factors might need revision: for example, secondary stress induction
28     might be redefined to refer only to an effect on a keystone species, and duration and
29     species depletion potential would not be relevant.  Additional factors, such as
30     secondary effects induction, could be added.
31
32            While the Subcommittee did not further develop these considerations into a
33     more detailed methodology nor apply an effects-backwards methodology for any
34     environmental problem, we suggest the Agency consider instituting research to
35     complete the development and application  of an effects-backward methodology to

                                             2-34

-------
      IRP Integrated Draft Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

1     apportion causes to effects that are manifested in the environment.
2
                                          2-35

-------
        IRP Integrated Draft Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1     2.6 References Cited
 2
 3     Bamthouse, L.W. and G.W. Suter, II (Ed.s.).  1986. User's manual for ecological risk
 4           assessment. ORNL-6251. Oak Ridge National Laboratory, Oak Ridge, TN.
 5
 6     Fava, J. A., L.W. Bamthouse, J. Falco, M.A. Harwell, and K. Reckhow.  1992.
 7           Chairpersons' Summary Report on the EPA Draft Document: Framework for
 8           Ecological Risk Assessment. EPA/625/R-91/022. U.S. Environmental Protection
 9           Agency, Risk Assessment Forum, Washington, DC.
10
11     Harwell, M. A. and J. Gentile.  1992. Report of the EPA Ecological Risk Assessment
12           Guidelines Strategic Planning Workshop, Miami, FL, May 1991.
13           EPA/630/R-92/002. U.S. Environmental Protection Agency, Risk Assessment
14           Forum, Washington, DC.
15
16     Harwell, M. A. and J.R. Kelly.  1986. Ecosystems Research Center Workshop on
17           Ecological Effects from Environmental Stresses. ERC-140, Ecosystems
18           Research Center.
19
20     Harwell, M.A., and C.C. Harwell. 1989 Environmental decision-making in the presence
21           of uncertainty.  Chapter 18, pp 517-540. ]n: Levin, S. A., M. A. Harwell, J. R.
22           Kelly, and K. Kimball (Ed.s.).  1989.  Ecotoxicology: Problems and Approaches.
23           Advanced Texts in the Ecological Sciences Series. Springer-Verlag, New York.
24           547 pp.
25
26     Harwell, M. A., W.  Cooper, and R. Flaak. 1992. Prioritizing ecological and human
27           welfare risks from environmental stresses. Environmental Management
28           16(4) :451-464.
29
30     Kelly, J. R. and M. A. Harwell. 1990. Indicators of ecosystem recovery. Environmental
31           Management 15(5) :527-545.
32
33     Levin, S. A., M.A. Harwell, J.R. Kelly, and K. Kimball.  (Ed.s.).  1989. Ecotoxicology:
34           Problems and Approaches. Advanced Texts in the Ecological Sciences Series.
35           Springer-Verlag, New York. 547 pp.

                                           2-36

-------
       IRP Integrated Draft Report-Peer Review Draft, May 3,1999—Do Not Cite or Quote

 1     National Research Council. 1983. Risk Assessment in the Federal Government:
 2           Managing the Process. National Research Council, National Academy Press,
 3           Washington, DC.
 4
 5     Science Advisory Board. 1990a. The Report of the Ecology and Welfare
 6           Subcommittee, Relative Risk Reduction Project. Reducing Risk Appendix A.
 7           EPA SAB-EC-90-021A. U.S. Environmental Protection Agency, Science
 8           Advisory Board. Washington, DC.
 9
10     Science Advisory Board. 1990b. Reducing Risk: Setting Priorities and Strategies for
11           Environmental Protection. SAB-EC-90-021. U.S. Environmental Protection
12           Agency, Science Advisory Board. Washington, DC.
13
14     U.S. Environmental Protection Agency. 1987a. Unfinished Business: A Comparative
15           Assessment of Environmental Problems.  Appendix III. Ecological Risk Work
16           Group, Office of Policy Analysis, U.S. Environmental Protection Agency,
••"•           Washington, DC.

19     U.S. Environmental Protection Agency. 1987b. Unfinished Business: A Comparative
20           Assessment of Environmental Problems.  Overview Report. Office of Policy
21           Analysis, U.S. Environmental Protection Agency, Washington, DC.
22
23     U.S. Environmental Protection Agency. 1992. Framework for Ecological Risk
24           Assessment. EPA/630/R-92/001.  Risk Assessment Forum, U.S.  Environmental
25           Protection Agency, Washington, DC.
26
27
28
29
30
31
32
33
34
                                           2-37

-------
      IRP Integrated Draft Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote





1     Appendix 2A: Ecological Risk Profiles
                                        2-38

-------
IRP Integrated Draft Report-Peer Review Draft, May 3,1999—Do Not Cite or Quote


         CHAPTER 3. HEALTH RISK RATING METHODOLOGY

                         TABLE OF CONTENTS

3.1  Introduction	3-1
      3.1.1 Background	3-1
      3.1.2 Approach 	3-1
      3.1.3 Role and Significance of Expert Opinion in Risk Analysis  	3-2

3.2. The Environmental Health Risk Rating Methodology 	3-4
      3.2.1 Nature of the Methodology	3-4
      3.2.2 Scope of Stressor List	3-8
      3.2.3 Stressor Risk Characterization Data Sheets	3-10
      3.2.4 Rating of Confidence in Relative Risk Rating of a Stressor	3-14
      3.2.5 Factors Influencing Relative Risk Rating	3-14
      3.2.6 Survey Design  	3-15

3.3  Analysis and Reporting of Relative Risk Rating Survey Data 	3-16

3.4  Implications of Ratings  	3-18

3.5  A Fuzzy Logic Approach  	3-19

3.6  Extensions and Refinements of the Methodology	3-25

3.7  Summary and Conclusions	3-26

3.8  References Cited	3-27

Appendix 3A. Health Risk Assessment Introduction	3-28

Appendix 3B. Instructions 	3-30

Appendix 3C. Risk Characterization Data Sheets  	3-33

-------
       IRP Integrated Draft Report-Peer Review Draft, May 3,1999—Do Not Cite or Quote


 1             CHAPTER 3. HEALTH RISK RATING METHODOLOGY
 2
 3     3.1 Introduction
 4
 5      3.1.1 Background
 6
 7          Adverse health effects in the U.S. population arise from a wide variety of
 8     individual causes and their combinations.  EPA is broadly mandated by Congress to
 9     address those causes that are environmentally-mediated, i.e., to reduce risks from
10     environmental stressors or conditions that impair health.
11
12          As noted in Chapter 1, EPA initiated its first effort to assign relative risk ratings to
13     environmental problems in 1986 to improve allocation of Agency resources. This effort
14     was reported in Unfinished Business (EPA, 1987). EPA's Science Advisory Board
15     (SAB) expanded upon this work in the report Reducing Risk: Setting Priorities and
16     Strategies for Environmental Protection (SAB, 1990).  For that report, the
 "*     Environmental Health Effects Subcommittee conceptualized health  risks as a
-4     multidimensional matrix whose axes comprised variables such as exposure sources,
19     exposure routes, health endpoints, and stressors.  It noted that the practical translation
20     of this conceptual structure could be achieved by a relational database design allowing
21     independent ratings of the different variables across various dimensions. At the end of
22     the rating process, each of its elements could then be evaluated independently to
23     achieve a global view of where the major problems resided and where resources could
24     most profitably be directed. This perspective, in essence, has been  adopted as a guide
25     to the selection of risk reduction options, a topic discussed at length in Chapter 6.
26
27      3.1.2 Approach
28
29          To meet the requirements of the IRP, the Human Exposure and Health
30     Subcommittee (HEHS)  took as its objective the development of a methodology for
31     relative rating of the human health risks from exposures to environmental stressors.
32     The HEHS also viewed  the matrix structure as a useful conceptual plan, but, as in
33     1990, was compelled to consider the pragmatic question of its translation into  practice.
34     It concluded that any method developed by the Subcommittee should offer, in so far as

                                            3-1

-------
       IRP Integrated Draft Report-Peer Review Draft, May 3,1999—Do Not Cite or Quote

 1      possible, certain key attributes. Specifically, the method should be:
 2
 3            a)    as simple as possible to understand, implement, and use;
 4            b)    capable of gathering judgments from many respondents quickly and
 5                  inexpensively;
 6            c)    responsive to the technical knowledge of individual experts and able to
 7                  capture some of the reasoning of the experts;
 8            d)    responsive to the concerns of other interested groups; and
 9            e)    suitable for analysis of correlations among various respondents.
10
11      The method also should explicitly utilize the best existing scientific information on
12      environmental health risks  and should be a scientifically based and transparent
13      methodology.
14
15            Because evaluation of sources and source categories was to be a focus of the
16      Risk Reduction Options Subcommittee and because coordination of ecological and
17      human health risks was seen as one of the crucial aims of the current exercise, HEHS
18      adopted the Ecological Risks Subcommittee's strategy of structuring its methodology
19      around environmental stressors, i.e., chemical, biological, and physical agents released
20     into the environment by or modified through human activities. Although many
21      comparative risk projects have been structured around the ranking of problems, such as
22      drinking water contamination or urban air pollution, stressors provide a more
23     homogeneous basis for rating adverse health effects.  Furthermore, reducing human
24     environmental health risks eventually requires reductions of exposure to environmental
25,    agents or stressors, largely by controlling sources.  The analysis of stressors can
26     provide a starting point for the development of policy options.
27
28       3.1.3 Role and Significance of Expert Opinion in  Risk Analysis
29
30           Expert opinion plays a special role in the architecture of risk assessment,
31     especially when data are incomplete, contradictory, or multidimensional (Morgan and
32     Henrion, 1990).  One method for incorporating expert  opinion relies upon a small group
33     of respondents whose implicit or direct aim is to provide a consensus judgment. The
34     Delphi method, described in many publications (Adler and Ziglio,  1996; Linstone and
35     Turoff, 1975^, exemplifies such an approach.  In striving for consensus, such a group

                                             3-2

-------
       IRP Integrated Draft Report-Peer Review Draft, May 3,1999—Do Not Cite or Quote

 1      exercise may deliberately suppress significant disagreement among members of the
 2      expert panel. The extent of disagreement, however, is a crucial piece of information.
 3      Presenting only a consensus position is equivalent, in some respects, to presenting
 4      experimental data only as arithmetic means unaccompanied by any measures of
 5      variability.  As noted by Morgan and Keith (1995), "...when uncertainty is high because
 6      of fundamentally different views about underlying...processes, a consensus summary
 7      may not best serve policy analysis needs.  An alternative approach, widely used in
 8      applied Bayesian decision analysis, formalizes and quantifies the judgment of individual
 9      experts through expert solicitation."
10
11            A high degree of variability in risk judgments among experts chosen for their
12      specialized knowledge may indicate deficiencies in the data upon which judgments are
13      based, or disagreements in how experts interpret the data.  Moreover, an aspect of
14      interpretation often ignored is the degree of confidence the judges place in their
15      assigned ratings, a judgment vector, so to speak, that largely reflects the breadth and
16      depth of the available data. It may also reflect the complexity of such data; a rich data
n      set may, paradoxically, point in several directions simultaneously.

19            Any approach or methodology aimed at health risk assessment must deal
20     simultaneously with both current questions and future projections.  For most currently
21      perceived problems, the primary impediments to establishing a risk ranking or rating
22      reside in the nature of the available data. Although a sparse data base is an obvious
23      source of uncertainty, it is by no means the sole one. For some agents or conditions,
24     even extensive data does not guarantee a shrinkage of  uncertainty. Different observers
25      may interpret the same information from different points of view. A process that takes
26     account of both variations in data availability and interpretation is required to conform to
27     current standards of risk characterization such as those described  in Understanding
28     fl/s/f(NRC, 1996).
29
30           Another feature that should be incorporated into  a health risk rating methodology
31      is geographic flexibility.  Although the IRP has focused primarily on national concerns,
32     particularly in its choice of stressors, the risk rating process should be able to
33     accommodate state and local concerns as well. Without some uniformity in how ratings
34     are achieved, consistency between levels of geographic integration will be difficult to
35     achieve.

                                             3-3

-------
       IRP Integrated Draft Report-Peer Review Draft, May 3,1999—Do Not Cite or Quote

 1            A third aspect of desirable methodological flexibility is the capacity to adapt to
 2      changes in currently perceived or mandated stressors, exposure standards, or public
 3      concerns.  Barriers to flexibility include factors ranging from the inability to convene
 4      expert rating panels to the formidable difficulties of publishing and distributing new
 5      printed documents addressing a particular risk.
 6
 7            The rating process itself, beyond logistics, should also be able to define its limits.
 8      An assignment of low risk, for example, must be presented  in context.  For example,
 9      although experts might assign a low overall rating to a specific stressor, such a rating
10      might have to be modified for a specific subpopulation, or when the available
11      information is tentative.
12
13      3.2. The Environmental Health Risk Rating Methodology
14
15        3.2.1 Nature of the Methodology
16
17            The HEHS developed a highly simplified system for polling and characterizing
18     expert judgements of the relative risks of a set of stressors. In the method developed
19     by HEHS, each expert, or other respondent, is asked to rate the health risks of
20     environmental stressors, from "very low" to "very high." Information is provided on the
21      effects of, and population exposures to, each stressor, although the respondent is
22     instructed to draw on his or her own knowledge and scientific judgment to interpret the
23     information provided.  The respondent is also asked to identify the determinants of his
24     or her rating. Such determinants might include severity of effect, the size of the
25     affected population, risks to subpopulations, or other factors.
26
27           The final results of the rating exercise include not only the average rating of each
28     stressor, but also the distributions of these ratings, and information on the chief
29     considerations upon which  different respondents based their ratings.  From these
30     results, one  can see not only the average rating for each stressor, but one can also
31      identify those stressors for which there is substantial disagreement and variation in
32     ratings.
33
34           To implement this approach, HEHS constructed a World Wide Web site
35     designed to  collect two types of information: a) ratings of health risks,  from a list of

                                              3-4

-------
      IRP Integrated Draft Report-Peer Review Draft, May 3,1999—Do Not Cite or Quote

 1     environmental stressors, provided by a sample of respondents not constrained to attain
 2     consensus, and b) ratings of the confidence these respondents place in their ratings.
 3     The Web site design aimed for the five fundamental properties listed in the previous
 4     section.
 5
 6           Another desirable property was the ability to identify the backgrounds of the
 7     respondents and to correlate them with their responses.  Although this method was
 8     designed for the elicitation of expert judgements, an additional advantage of a
 9     web-based process is that it can also be used by stakeholders, members of a specific
10     community or interest group, or members of the general public.
11
12           The Entry Page for the Web Survey (Appendix 3A) explains the purpose of the
13     survey and presents the instructions. The respondent then begins the process  by
14     registering (Figure 3-1). The respondent enters his or her name, affiliation, and  e-mail
15     address. To help ensure the integrity of the process,  passwords would be issued to
16     restrict access. In addition, however, each respondent, in transmitting the ratings,  also
17     automatically provides the Internet address from which they are sent. The respondent
      then proceeds to the rating page, shown in Figure 3-2. The respondent chooses a
19     stressor for rating by a mouse click on the stressor name. Table 3-1 gives the current
20     stressor list.
21
22           Once a stressor is selected,  the respondent enters his or her ratings of Risk
23     Rating and Confidence, which range from Very High to Very Low, on the appropriate
24    buttons. The rater can also, by selecting the Information button, access the appropriate
25     Stressor Risk Characterization Data Sheet; risk data sheets on the web site (Appendix
26    3C) summarize key available scientific information on each of the stressors, e.g.,
27     information on exposure route, populations exposed, identified health effects, and  so
28     forth. Categories of information that should be presented in the risk data sheets are
29    described further in Section 3.2.3.  In rating the stressors, respondents can access the
30    risk data sheet for each stressor and use all, some, or none of this information plus their
31     own expert knowledge to arrive at a risk rating.
                                             3-5

-------
        IRP Integrated Draft Report-Peer Review Draft, May 3,1999—Do Not Cite or Quote
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
              Stressor Risk Characterization User Name registration.
        Please Register your User Name and Password:
                  Register yoa name ltd pauvtrd HERE. 3« compfete and nrefall
                                 Motto's
           Em*
          Address:
                                       iOSNWSc
       Figure 3-1.  Registration form for the risk
       survey. Only valid registrants are permitted to
       submit ratings.

Stressor Risk Characterization and Health Risk Ranking
                            Data Sheet
                                 For
                            Dr. Bernard Weiss
                            Univ. of Rochester
                        weiss@envmedrocrtester.edu
 If the above information is NOT correct for YOUretumto ttie login page and enter YOUR name and PASSWORD.
           Vlew/Edit/Moditv Your Current Stressor Ranking Data
               Select a Stressor to Rank: i
    Get INFORMATION ON:
                               Relative
                              Risk Rank
            Confidence in   Factors innuencing
              Ranking         rankmg
Please rank Stressor and indicate r
   your confidence to your ranking:
     Review of General Instructions
  VH
<• H
r M
i- L
c VL
- 7
                                                        ChKklll that Apply
f VH
r H
<• M
r L
<• VL
r •>
I r StZ(ofpopidrfonalT«atd
i cfBtiaitarsubpopiilxliora
! r S«vei9y of health effects
  P«t«* of
  incidence
  Pmi^cnc
  irtvirontTwrit
                     Figure 3-2. Risk rating form. The respondent selects a Stressor,
                     then accesses the data sheet (Appendix 2) by selecting the
                     Information Window.
                                                      3-6

-------
        IRP Integrated Draft Report-Peer Review Draft, May 3,1999—Do Not Cite or Quote
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
.17
  B
 19
20
21
22
23
24
25,
26
27
28
29
30
31
32
33
34
35
A Review of the General Instructions can also be selected with a click.

      Lastly, the respondent checks the appropriate boxes to indicate the factors
influencing his or her rating.  One or more factors may be selected. A description of
how these factors may be interpreted is also given in the Instructions  (Appendix 3B). A
respondent may also indicate other reasons for the rating and enter a specific
comment.  Provision for such entries appears at the bottom of the form and is shown in
Figure 3-3.

      The respondent then proceeds systematically through the list of stressors. The
rater can review his or her ratings and accept or modify them, as shown in Figure 3-4.
Finally, once all of the stressors have been rated and accepted by the respondent, the
results are sent via the Internet for analysis with the ratings submitted by other
respondents.
      Other reasons for ranking;
         General Comments:
                                                               1
                                      V.fes»vS^^S!*^^H*<»^'^fe-Saft'4S
                                                               1
                                        >';>:W^''<-"^AvSS—j*:f^.^i'i!j:^;3
-------
       IRP Integrated Draft Report-Peer Review Draft, May 3,1999—Do Not Cite or Quote
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
IS
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
                                         	
                            gW&iSS^g^tS;:
                          Edit/Update Risk Ranking Data
                                       For
                                   Dr. Bernard Weiss
                                   Univ. of Rochester
              If the above information is MOT correct for YOU return to the Login oacs and enter YOUR name and PASSWORD.
             IStressor
          Risk  Confidence
          Rank    Rank
General Comments  Other Reasons

                                   5
                                Incidence-of poisonings jn
                                dtidran
             .tiabQenated
             Hydrocarbons
tl-Dec-1997
;09:54:16AM
           NewdataonreOTductionj
                            E-Mail a copy of my current Entries to:
            Figure 3-4. Selection summary and revision form. This
            feature allows the respondent to review and modify previous
            selections.
  3.2.2 Scope of Stressor List

       A multitude of physical, chemical, and biological stressors are associated with
environmental health risks.  For the organic compounds alone one can easily
enumerate thousands of compounds distributed among air, water, soil, and food.
Clearly, the Risk Rating Methodology could be overwhelmed by undertaking to evaluate
too large a list of stressors, which would leave the process paralyzed at the stage of
acquisition and development of information data sheets. Such a broad evaluation
would come at the expense of a more thorough examination of those stressors offering
an established reason for potential or actual concern.  The proposed list of stressors to
be rated might include the Criteria Pollutants (air, water), Hazardous Waste Priority
                                               3-8

-------
       IRP Integrated Draft Report-Peer Review Draft, May 3,1999—Do Not Cite or Quote
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
20
21
22
23
24
25
26
27
28
29
30
31
32
Chemicals, Persistent Organic Pollutants,
contaminants regulated by emissions or effluent
standards, radionuclides regulated by EPA and the
Department of Energy, common indoor pollutants,
pesticides, metals, indirect stressors (e.g., ozone
depleting chemicals), and emerging issues (e.g.,
hormone-disrupting chemicals).

      From a large list of stressors, EPA needs to
dedicate the resources to reduce the  stressors to
be rated to a manageable number, probably no
more than 100 at most. This might be done by
aggregating some of the stressors on the basis of
criteria such as commonality of toxic targets,
endpoints, and sources. Since EPA resources will
be limited, especially at the beginning, the Agency
might wish to begin with a deliberately limited list of
stressors, but one whose members intersect a
number of groupings.  For example, HEHS
developed a short list  of stressors with which to test
and illustrate the methodology (Table 3-1).

      This list, with categories that are not always
mutually exclusive, is  presented as an example and
should be modified as needed based on review and
analysis of broader lists of pollutants  and other
stressors  noted above. The approach should be
iterative, with stressors added and deleted over
time based upon new information about exposures,
health risks, and stakeholder concerns.
Table 3-1. List of Current Stressors
Selected for Risk Ratings
        Criteria Air Pollutants
        Carbon Monoxide
        Ozone
        PM-10

        Heavy Metals
        Arsenic
        Lead
        Mercury

        Persistent Organic Chemicals
        Dioxins, furans, PCBs
        Hormone disrupting chemicals
        Pesticides, Herbicides

        Environmental Tobacco Smoke

        Polycyclic Aromatic
        Hydrocarbons

        VOCs (Air Toxics)
        Benzene
        Tetrachloroethylene
        Formaldehyde

        Bioaerosols
        House dust mites allergen and
        fungi

        Physical Agents
        Radon
        EMF

        Indirect Stressors
        CO2 and other greenhouse gases
        Freons and other ozone layer
        depleting chemicals
                                              3-9

-------
       IRP Integrated Draft Report-Peer Review Draft, May 3,1999—Do Not Cite or Quote

 1       3.2.3 Stressor Risk Characterization Data Sheets
 2
 3           The Stressor Risk Characterization Data Sheets provide key information useful
 4     in rating risks.  This summary information should be based primarily on studies reported
 5     in the peer reviewed scientific literature and presented succinctly.  Peer-reviewed
 6     comprehensive risk assessments developed by EPA, such as the those for the criteria
 7     air pollutants, or those developed for environmental tobacco smoke (EPA, 1992), radon
 8     (EPA, 1992), dioxins (EPA, 1995), can provide key conclusions and information for
 9     inclusion in the risk data sheets. For stressors not yet subjected to a comprehensive
10     risk assessment, including a review and analysis of the literature, the information
11     included in the data sheet should be based primarily on peer-reviewed papers reported
12     in the scientific literature that vary in experimental designs and protocols. Risk data
13     sheets to be used in the rating process should themselves be subjected to peer review
14     before use and should include references to the sources of information. An example of
15     a Stressor Risk Characterization Data Sheet is shown in Table 3-2, and the full
16     collection of  risk data sheets prepared by HEHS is presented in Appendix 3C. The
17     following kinds of information should be included for each risk data sheet, if available:
18
19           a)    Exposure routes and pathways  pertaining to the Stressor. Route of
20                 exposure might be inhalation, ingestion, dermal or some combination,
21                 depending upon the particular Stressor. Pathway of exposure should
22                 provide some information on  how  the stressor travels from the source to
23                 the human receptor. For example, radon is advectiveiy transported from
24                 soil gases into homes and its decay progeny are inhaled. The dioxins are
25,                generally emitted from combustion sources, transported through the
26                 environment and accumulated in food.
27
28           b)    Population exposed to the stressor, e.g., total U.S. population or a
29                 special subpopulation(s). For many stressors, e.g., criteria pollutants,
30                 virtually all of the U.S. population is exposed to some degree. An example
31                 of an exposed sub-population of concern would be subsistence fishermen
32                 and their families, who may have a very high exposure to PCBs, dioxins or
33                 methylmercury.
34
35           c)    Average dose to (or exposure concentration) for the exposed population.

                                            3-10

-------
      IRP Integrated Draft Report-Peer Review Draft, May 3,1999—Do Not Cite or Quote

 1                 Ideally, the distribution of exposures for the population would be
 2                 presented here. However, this information is currently available only for a
 3                 limited number of stressors, e.g., radon. Information on high end
 4                 exposures in the exposed population should also be described, if
 5                 possible.
 6
 7           d)    Relevant animal toxicological data, e.g., NOAELs or LOAELs if
 8                 available. It is generally accepted that an agent that produces an adverse
 9                 effect in experimental animal studies will potentially pose a hazard to
10                 humans following sufficient exposure.  In many instances, animal
11                 toxicology data can provide supporting evidence of adverse health effects
12                 in exposed humans.  In the absence of adequate data on the toxicity of an
13                 agent in humans, we assume that effects in animal species are probably
14                 indicative of effects in humans. This is a default assumption, and is
15                 abandoned only when all relevant information on aspects of differential
16                 absorption, distribution, metabolism, and toxicokinetics between species
17                 have been considered and indicate that effects observed in animals will
                  not occur in humans. Minor adverse effects will not be discounted unless
19                there is evidence that they are not relevant to humans.
20
21            e)    Health effects from environmental exposures. Information on the
22                observed or potential adverse health effects for human populations at
23                current environmental exposure levels should be summarized. Well-
24                designed and conducted epidemiological studies of various
25'               sub-populations, such as highly exposed workers or the general
26                population (e.g., PM-10, ozone), which have been peer reviewed, provide
27                good evidence that exposed populations are at risk of adverse health
28                 effects.
29
30           f)     Biological half life in humans, if known.
31
32           g)     Sizes of populations experiencing specific adverse health  effects due
33                 to current exposures to the stressor. This information will not be available
34                 for every stressor.
35

                                            3-11

-------
       IRP Integrated Draft Report-Peer Review Draft, May 3,1999—Do Not Cite or Quote

 1            h)    Severity and persistence of adverse health effects. For example,
 2                  mortality, morbidity, and reversibility of endpoint.
 3
 4            i)     Occupational exposure limit(s). (Optional) Although Threshold Limit
 5                  Values and Permissible Exposure Limits developed for the protection of
 6                  healthy workers in industry are much higher than would be appropriate for
 7                  the general population, which includes children and susceptible
 8                  sub-populations such as asthmatics, these limits are often very close to
 9                  exposures at which adverse health effects have been observed in human
10                  populations. They provide a useful reference point for comparison to
11                  average and high-end environmental exposures.
12
13            j)     Environmental half-life. Environmentally persistent stressors, such as
14                  dioxins and freons, can accumulate in various environmental
15                  compartments and may be of greater concern than those which degrade
16                  rapidly under environmental conditions.
17
18            k)    Comments. Any additional information which might be of use in rating
19                  the.relative risk of a particular stressor should be included here.
20
21            Although it will not be possible to provide these kinds of data for every stressor,
22     the data sheets should provide a space for each. Gaps in the data sheets will enable
23     the experts (and the Agency) to easily identify critical gaps in information. In addition, it
24     might be desirable to include an estimate of the fraction of disease incidence
1%     attributable to exposures to the environmental stressor. A stressor for which a
26     substantial fraction of the total annual disease incidence is attributable to exposures to
27     the stressor might be rated somewhat higher than one which contributed little to  overall
28     incidence. There are, however,  some methodological pitfalls in making such estimates,
29     e.g., double counting across two stressors.
30
                                             3-12

-------
IRP Integrated Draft Report-Peer Review Draft, May 3,1999—Do Not Cite or Quote
 Table 3-2. Stressor Risk Characterization Data Sheet for PM-10
STRESSOR
Exposure Routes and
Pathways Considered
Population Exposed
Average (Potential)* Dose for
Exposed Populations
Animal Toxicological Data
Health Effects from
Environmental Exposures
Size of Populations
(estimated) experiencing
health effects:
Severity and Persistence of
Adverse Health Effects:
Estimated Fraction of
Disease Incidence
Attributable to
Environmental Exposures (if
Known):
Other Comments
PM-10
(Airborne Paniculate Matter, <10 ^m MMAD )
Inhalation of outdoor air (PM in outdoor air infiltrates into indoor
environments)
All of U.S. population
• - 1 1 m3/day X (9 - 34 ng/m3) = 99 ^g/day to 374 ^g/day
inhaled, depending upon where one lives
• Doses to patients with chronic obstructive pulmonary
disease (COPD) may be 3-times greater than for healthy
adults (U.S. EPA, 1996)

• Increased death rates in elderly from cardiopulmonary
disease
• Increased hospital admissions for COPD (emphysema,
bronchitis) patients
• Increased risk of acute respiratory disease
• Decreased lung function in children and asthmatics
• -60,000 premature deaths in the elderly annually from
cardiopulmonary disease (American Journal of Respiratory
and Critical Care Medicine, 1995)
• Relative risk values range from about 1 .02 to 1 .08.
Although RR is not very high, exposed population is so
large than incidences of health effects large.
• Premature death
• Hospitalizations due to exacerbation of COPD
• Decreased lung function in children and asthmatic is
irreversible effect
PM is estimated to account for about 60,000 premature
cardiovascular deaths annually; this is about 8% of the annual
cardiovascular deaths in the U.S.

REFERENCES:
      U.S. EPA. 1996. Air Quality Criteria for Particulate Matter.  EPA/600/P-95/001cF, April 1996.
      U.S. EPA. 1992. Technical Support Document for the 1992 Citizen's Guide to Radon. U.S.
      Environmental Protection Agency.
                                      3-13

-------
       IRP Integrated Draft Report-Peer Review Draft, May 3,1999—Do Not Cite or Quote

 1       3.2.4  Rating of Confidence in Relative Risk Rating of a Stressor
 2
 3           Respondents are asked to indicate their level of confidence in the risk rating
 4     given for each stressor. The confidence ratings are Very High, High, Medium, Low and
 5     Very Low.  This rating can include both the respondent's judgment of the state of
 6     existing knowledge  as well as the weight accorded his or her individual expertise about
 7     a particular stressor. The Committee considered separate ratings for these items but,
 8     after much discussion, decided to include only a single confidence rating in the interest
 9     of maintaining a simple methodology.  If a respondent believes the state of scientific
10     knowledge to be inadequate for a rating, this information can be supplied in the
11     comments section.
12
13       3.2.5  Factors Influencing Relative Risk Rating
14
15           For each stressor rated, the respondent is also asked to indicate the major
16     factors which influenced his or her judgment of relative risk by checking the appropriate
17     boxes.  These factors are:
18
19           a)    Size of the population affected. A high risk rating might be based on
20                 evidence that many members of the population are experiencing adverse
21                 health effects from environmental exposures to the stressor, while a low
22                 risk rating based on size of population would indicate that only a smalt
23                 fraction of the population is being exposed at present.
24
25           b)    Particular sub-populations at risk. A check for this factor in
26                 combination with a high risk rating for the stressor might arise if some
27                 small  sub-population(s) are subjected to very high exposures and risks.
28
29           c)    Severity and persistence of health effects.  A high risk rating might be
30                 given  for a stressor based on health effects that are life-threatening, life
31                 shortening, and/or irreversible. A low risk rating might be assigned to a
32                 stressor because the most significant adverse health effect is transient
33                 and apparently reversible.
34
35

                                             3-14

-------
       IRP Integrated Draft Report-Peer Review Draft, May 3,1999—Do Not Cite or Quote

 1           d)     Percentage of attributable incidence.  A respondent might assign a high
 2                 risk rating to a stressor because exposure to the environmental stressor
 3                 accounts for a significant percentage of the total annual incidence of a
 4                 given disease or health endpoint. For example, based on the estimated
 5                 annual number of lung cancers from radon exposures (EPA report on
 6                 radon, 1992), it can be estimated that about 1 -2 % of the annual incidence
 7                 of lung cancer in the U.S. is attributable to radon exposures.
 8
 9           e)     Persistence in the environment.  A respondent might rate a stressor as
10                 very high or high because it is not rapidly degraded in the environment but
11                 tends to accumulate and, because of this property, poses an
12                 environmental health risk.
13
14           f)     Potential future risk. A respondent might rate a particular stressor as
15                 very high or high because the stressor is very likely to pose a significant
16                 environmental health risk in the future if no actions are taken to reduce
17                 anticipated exposures arising from identified sources. The respondent
                   might also check this factor as influencing his/her low risk rating because
19                 he/she judges that no increased health risk is likely to arise in the future.
20
21           g)     Other. The respondent is asked to check this if some major factor, other
22                 than those provided, influenced his or her rating. The respondent is also
23                 asked to identify this other factor(s).
24
25'          h)     Comments. Experts may have important knowledge that has not yet
26                 become widely known. The "Comments" box, as well as the "Other"
27                 category, provide means to alert EPA to this knowledge.
28
29      3.2.6 Survey Design
30
31           Implementation of an expert survey requires a specific description of the
32     population to be surveyed and how the sample is drawn.  Even the Delphi process, or
33     any other process that relies on a select committee to make judgments or issue  ratings
34     is, sometimes implicitly, a hostage to sample selection procedures and possible  bias.
35     One virtue of the Web design is the enlarged possibilities it presents  for explicit

                                            3-15

-------
       IRP Integrated Draft Report-Peer Review Draft, May 3,1999—Do Not Cite or Quote
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
samples. For example, a risk rating exercise might draw from the total SAB
membership; or, for some purposes, it might sample from the membership of the
Inhalation Specialty Section of the Society of Toxicology or the International Society for
Exposure Analysis; or, its sampling population might come from a specific community.
Because expert groups may be viewed as inherently biased, the ability of EPA to use
an unbiased survey technique should prove useful.

3.3 Analysis and Reporting of Relative Risk Rating Survey Data

      A leading purpose of the Web survey instrument is to evaluate the diversity of
rater estimates. The importance of rater diversity is illustrated in Table 3-3, which
shows some of the patterns of rater judgments that would yield equivalent average
(mean) risk ratings.

 Table 3-3. Distributions Of Risk Ratings For a Stressor That Would Yield an Average
Rating of "Medium"
Very High
oo
0

ooo
oooo
High
oo
o

oo
0
Medium
00
oooooo
oooooooooo


Low
oo
0

00
o
Very Low
oo
o

ooo
oooo
      Patterns of rater responses convey important messages. A stressor whose
ratings are concentrated at the extremes reflects a high degree of polarization among
the population sampled. Similarly, the degree of consensus among raters on the
dimension of confidence also reflects the characteristics of the sample population and
its interpretation of current information. Figure 3-5 depicts four synthesized examples
of the kinds of patterns such plots might produce (based on 12 fictional respondents).
As discussed by the Subcommittee, they convey data potentially helpful to the
formulation of  risk options. For example, plot (C) might characterize a phenomenon
such as global warming whose possible health impacts might be widespread and
                                            3-16

-------
      IRP Integrated Draft Report-Peer Review Draft, May 3,1999—Do Not Cite or Quote
1     severe but that currently lacks enough supporting data to elicit much confidence in that
2     evaluation.

A
f
cc.
•5
S


C
?
1
•5
£


VI


•


VI
• •*«





I
«

«
•
• •
C.
i

« •



C
H

•
•
•

atfdtrc
U





orviocn
H


»
•

•
H





;«
VM



•

VM







B
H ?
a «
L £
VI

m D
N ?
B «
L S
VI

VL





VL






I





0
I





C
il



• * •

inndcni
H





oflfldtfi
N



— —

•
H





ec
VH




• ••«
VH







nt
N
•
L
VL

nt
<
I
•
A

        Figure 3-5. Synthetic plots of risk versus confidence (12 respondents): A) high
        variability; B) low risk, mid to high confidence; C) high risk, low confidence; D) high
        risk, high confidence
 3          The Subcommittee views distributions of respondents' risk ratings and
 4    confidence in those ratings as the key pieces of information provided by the rating
 5    exercise.  In many instances, these raw data may prove sufficient for conveying the
 6,    relative degree of risk and confidence associated with each rated stressor. However,
 7    the Agency may apply data reduction techniques to these data; for example, many
 8    readers of EPA publications will be comfortable with means and standard deviations,
 9    which can be expressed by converting rater responses into numerical scores.  When
10    the distributions are clearly asymmetrical, Box Plots showing medians and 10th, 25th,
11    75th and 90th percentiles will prove more informative. Other kinds of transformations
12    are also available, but are constrained by the limitation that the ratings are ordinal, not
13    ratio scales. The Agency could also use non-numerical, visual presentations of the
14    results  of the survey that would provide a different way of looking at the results, and
15    might entail fewer underlying assumptions than do numerical data reduction techniques.
16
                                            3-17

-------
       IRP Integrated Draft Report-Peer Review Draft, May 3,1999—Do Not Cite or Quote

 1            In essence, the Agency is encouraged to develop a set of rules governing how
 2      the various combinations of risk/confidence ratings (25 in the current scheme) are
 3      ordered. These are subtle judgments, and might be assigned to an expert panel for
 4      resolution once the survey data have been gathered.
 5
 6            A second analysis of the Web data will be based on rater responses to the
 7      prompt requesting the factors upon which the ratings were based (Figure 3-2). These
 8      data provide considerable amounts of information, but are not readily transformed into a
 9      simple score.  However, they could be tallied as distributions, and when tallied, would
10      illuminate the risk ratings and provide the Agency with potential initiatives for
11      remediation or regulation, and specific goals for health research.
12
13            One major advantage of a Web survey is the ease of modification.  The original
14      design sought to include sources, pathways (exposure route), and endpoints as well as
15      the  basis for selection. In this first version, respondents were asked to rank primary,
16      secondary, and tertiary relationships for all variables.  For example, several different
17      sources might be responsible for environmental exposure to a particular stressor (e.g.,
18      both transportation and point sources for air pollutants), or a stressor might traverse
19      different exposure routes  (e.g., ingestion and inhalation for arsenic), or might affect
20     more than one endpoint (e.g., brain development and blood pressure for lead).
21      Although HEHS deemed this ranking form too complex for its current purposes, it
22      illustrates the flexibility afforded by electronic survey techniques. Some of the
23      modifications that might be incorporated in future versions, at the cost of increased
24     complexity, include provision for rater choices of endpoint, default hazard
25     identifications, default q1* and RfD values, graphical  data, and the submission of the
26     rater's own quantitative risk figures. Building in such  modifications, however, would
27     constitute a major software design project.
28
29     3.4 Implications of Ratings
30
31           The Relative Risk Rating Survey will provide information from a variety of experts
32     and stakeholders that can yield ratings for a stressor(s) that reflects the category of risk,
33     and assignment of the confidence that the responders to the survey made about the
34     rating. Currently, assignment of  a stressor to a priority category has typically been
35     formulated by a single number or classification.  Separating risk and reliability estimates

                                             3-18

-------
       IRP Integrated Draft Report-Peer Review Draft, May 3,1999—Do Not Cite or Quote

  1     provides a regulator, policy maker, scientist, or other interested party, with the types of
 2     information necessary to begin establishing a more meaningful priority for the particular
 3     stressor. Based upon the n'sk and confidence ratings, actions or potential responses
 4     could range from no additional actions at the time, to new research, or to
 5     implementation of new exposure and risk reduction strategies.  These should be
 6     considered, at the outset, as the initial priorities, and be revisited periodically to retain or
 7     to modify the initial risk assignment. At the same time, new stressors can be added to
 8     the list based upon new environmental release or formation, exposure, or health data.
 9
10     3.5 A Fuzzy Logic Approach
11
12           An important element in the process of soliciting probabilistic judgments about
13     risks from experts is obtaining their views of the uncertainties bounding their estimates.
14     Morgan and Keith (1995) describe how they tried to obtain both mean and standard
15     deviation judgments about climate change from a small group of authorities and
16     Morgan and Henrion (1990) describe several ways by which such information might be
17     presented graphically. The survey technique described in this chapter deals with
       uncertainty by eliciting the judge's evaluation of the confidence he or she places in the
19     health risk rating.  A low degree of confidence corresponds, in this instance, to wide
20     uncertainty and a high degree to narrow uncertainty.
21
22           Different combinations of risk and confidence ratings, as depicted in Figure 3-5,
23     can help illuminate which stressors deserve closer examination or which might be
24     assigned high or low priorities. Modifications of the survey protocol could be designed,
25     however, to procure indices of uncertainty more directly.  Fuzzy logic approaches could
26     prove useful for doing so.
27
28           Fuzzy logic is the term invoked by Lotfi Zadeh, the originator of the discipline, to
29     designate a tactic, in essence, for dealing with uncertainty.  As noted by Morgan and
30     Henrion (1990), even experts find it more comfortable to deal with uncertain
31     probabilities by applying verbal labels such as "likely" and "unlikely" rather than by
32     issuing numerical estimates.  Fuzzy logic, rather than representing uncertainty in the
33     conventional sense, as in soliciting estimates of variance, deals directly with ambiguities
34     such as those inherent in the terms above. Conventional logic treats every propositon
35     as either true or false. Conventional mathematical definitions of set membership also

                                            3-19

-------
       IRP Integrated Draft Report-Peer Review Draft, May 3,1999—Do Not Cite or Quote
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
rely on binary definitions.  Fuzzy systems view truth definitions and fuzzy set
membership as possessing any value in the range [0.0,1.0].

      Consider a speaker's description of the weather as "hot."  It conveys
considerable information although the adjective is imprecise and subjective and is a
function of the speaker's personal experience and  environment.  Figure 3-6 is a Fuzzy
depiction of "hot." Almost everyone would consider an ambient temperature of 40°C to
be hot, so that an ambient temperature that high might be assigned a truth value of,
say, 0.90. In Zadeh's terminology, that value corresponds to its membership in the set,
"hot." Below that temperature, your definition might depend largely on where you live.
Some inhabitants of Rochester, New York, would classify a reading of 25°C as "hot,"
but residents of Miami might consider such a reading as somewhat cool. At 25°C,
membership in the set, "hot," might elicit a truth value of 0.15. In traditional, Boolean,
set theory, "hot" would be assigned a crisp definition such as greater than 30°C and
symbolize a clear dividing line certain to evoke debate between Rochesterians and
Floridians.
                   20
  30         40
Temperature (°C)
           Figure 3-6. Membership in the class, "Hot," is not fixed,
           but depends upon individual definitions. The shading is
           not necessary, but is designed to show that, in this
           instance, the proportion of judges applying the adjective
           is less dense at the lower temperatures.
                                            3-20

-------
       IRP Integrated Draft Report-Peer Review Draft, May 3,1999—Do Not Cite or Quote
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17

19
20
21
22
23
24
2§r
26
27
28
29
30
31
32
33
34
35
                                Fuzzy Rating Patterns

                                 (Best Guess and Bounds)
                                                  -StressorA
                                                  —StressorS
                                                  —StressorC
                      VL
H
VH
                                Risk Rating
                 Figure 3-7. The vertical line, situated there by a
                 hypothetical rater, can be said to represent divided
                 membership in the risk classes, "Medium" and
                 "High." In place of a numerical value, the
                 respondent's rating can be interpreted as an estimate
                 of a truth or membership value in the fuzzy set,
                 "Medium" of about 0.7 and a membership value in
                 "High" of 0.3.
      One way to display the elements of fuzzy logic classification is shown in Figure
3-7.  Here, the usual numerical scale has been translated into verbal labels from Very
Low to Very High, in conformity with the categories used earlier. The figure depicts the
set (or rating) of "Medium," for example, as extending from Low to High, with maximum
membership (or truth) value at Medium and minimum membership values at Low and
High. Suppose a judge were asked to locate the point on the abscissa corresponding
to his or her best estimate of risk and marked it at the position designated by the
vertical line. The position of the line indicates majority membership in "medium," and
what might be called minority membership in "high."  Roughly translated into
percentages, the line coincides with about 70% membership in "medium" and 30% in
"high."
                                             3-21

-------
      IRP Integrated Draft Report-Peer Review Draft, May 3,1999—Do Not Cite or Quote
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
      The estimate defined by the vertical line in Figure 3-7 provides no indication of
the rater's uncertainty about her judgment. As we noted earlier, a rater's uncertainty
can be translated into what statisticians term "confidence limits." Morgan and Keith
(1995) secured analogous information from their judges by asking for standard
deviation estimates. With a fuzzy logic approach, a similar kind of estimate can be
obtained more simply.

      Figure 3-8 depicts fuzzy risk ratings of three hypothetical stressors.  The triangle
labeled "Stressor A" reflects the rater's judgment that this agent presents a Low risk
that is unlikely to be much lower or higher. The fuzzy rating for "Stressor B" is centered
at Medium, and is asymmetric to reflect the judgment that it could be as small as Low or
as great as Very High. The third triangle, centered between High and Very High, shows
the rater inclined towards the latter, but also locating the lower bound between Low and
Medium.
                                     M    H    VH
                                 Risk Rating
             Figure 3-8. Hypothetical ratings for three different
             stressors. The peak represents the rater's best guess
             of where along the abscissa the risk belongs; that is,
             where membership is believed to be 1.0. The lower
             comers of the triangle are set to reflect the rater's
             judgment of the lower and upper bounds of risk.  To
             summarize the judgments of many raters about a
             particular Stressor, the abscissa locations are defined
             by a scale, say, from 0 to 100 (Figure 9) and the heights
             summed at each integer position from 0 to 100.
                                             3-22

-------
       IRP Integrated Draft Report-Peer Review Draft, May 3,1999—Do Not Cite or Quote
 1
 2
 3
 4
 5
 6
 7
 S
 9
10
11
12
13
14
15
16
IT.

19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
      An example of the kind of visual display that can be used to elicit such
judgments from raters is shown in Figure 3-9. The rater views three horizontal bars,
each equipped with a slider that can be manipulated by a computer mouse. As the
slider is moved, its corresponding value on the abscissa, which ranges from 0 to 100,
appears above the bar. The uppermost bar is used to set the location of the rater's
best guess about the risk posed by the particular stressor. The left bar allows the rater
to set her estimated lower bound, defined as, "Unlikely to be any lower than this." The
right bar sets the location of the estimated upper bound, defined as, "Unlikely to be any
higher than this."  Respondents tested so far find it relatively easy to manipulate the
sliders. The triangles plotted as a result may be asymmetric, as shown in Figure 3-7.
For summarizing the responses of a specified sample of raters, the abscissa is divided
into, say, 100 units, and the summed ordinate values then calculated.
            Stressor Risk Characterization Ranking Form
                                Best guess
                                     64
                       Lower bound   Upper bound
            Figure 3-9. Depiction of a visual display presented to a
            rater asked to indicate where she would position her best
            guess of the health risk posed by a specific stressor and
            where she would place the lower and upper bounds of her
            estimate.  With the computer mouse, the respondent
            locates the slider at the appropriate numerical index. For
            this display, the scale runs from 0 to 100, but a second
            parallel scale, could show the corresponding verbal labels
            from, say, Very Low to Very High. These numerical
            positions can be converted into the kind of visual display
            shown in Figure 3-8.
                                           3-23

-------
       IRP Integrated Draft Report-Peer Review Draft, May 3,1999—Do Not Cite or Quote

 1           Fuzzy logic can also be applied to the criteria upon which the raters have based
 2     their risk estimates. For any particular stressor, they could be asked to also submit
 3     their degree of reliance on these criteria. That is, they could be asked to supply the
 4     weight accorded them, from Very Low to Very High, in making their risk judgment.  For
 5     example, the rating for a particular stressor might have been based largely upon the
 6     size of the population affected while another stressor rating might have been based
 7     largely on outcome severity, and its impact on a particular subpopulation such as
 8     children. The weights given these criteria and the associated uncertainty about their
 9     relevance can be translated into the kind of responses depicted in  Figure 3-8.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
                                             3-24

-------
       IRP Integrated Draft Report-Peer Review Draft, May 3,1999—Do Not Cite or Quote

 1     3.6  Extensions and Refinements of the Methodology
 2
 3           The Health Risk Rating Methodology was developed by the Human Exposure
 4     and Health Subcommittee to provide a unique tool for assigning relative risk to selected
 5     stressors. Its design was governed by three determinants that are seen as lacking in
 6     many risk evaluation exercises. First, it explicitly recognizes uncertainties by requesting
 7     raters to provide estimates of their confidence in the available data and in their own
 8     grasp of this information. Second, it offers the Agency a mechanism with which to
 9     secure risk ratings from designated populations not limited by the constraints of  time
10     and geography imposed by the usual committee deliberations.  Third, it permits
11     improved estimates of rater variability, a key policy index that often is not  available or is
12     explicitly concealed by focusing on consensus.
13
14           For this prototype exercise, the selection of stressors and the characterizations
15     of each stressor in the data sheets were based upon the experiences and professional
16     judgment of the members of the Subcommittee. Because of the limited resources
17     available to it, the Subcommittee was unable to extensively test and refine the
       methodology. In practice, the EPA would likely convene a working group and identify
19     resources for conducting the described activities.
20
21           These resources will provide the basis for developing an implementation  plan for
22     the proposed rating methodology. This must include: a) selection of the stressors to
23     submit for risk rating at the present time, b) development of a sampling program for
24     expert selection and completion of data sheets, c) statistical methods for  combining the
25     risks and the confidence to obtain a weighed priority, d) methods that link the data
26     sheets to basic resources or sources of information or exposure, effects,  and other
27     variables, e) approaches for peer review of data sheets prior to implementation of a
28     expert survey, f) activities to promote the availability and use of the Stressor Risk
29     Characterization  Data Sheets, ratings and prioritizations of individual stressors by
30     stakeholders, e.g., on the Internet, and g) identification of a timetable for
31     implementation of the project, and completion of respondents on the initial data  sets
32     and newly identified stressors.
33
34
                                            3-25

-------
       IRP Integrated Draft Report-Peer Review Draft, May 3,1999—Do Not Cite or Quote

 1      3.7  Summary and Conclusions
 2
 3            a)     Risk ratings can inform environmental decision-making, but in and of
 4                  themselves cannot be the sole basis of these decisions.
 5
 6            b)     Risk ratings are not purely scientific since value judgments concerning the
 7                  relative importance of a number of factors must be used to arrive at a
 8                  rating for each stressor.  Furthermore, when information on a particular
 9                  stressor is limited, judgments must be made about the uncertainties and
10                  the degree of protectiveness that is appropriate.
11
12            c)     Human health risk ratings, however, can be scientifically based, justifiable,
13                  well-defined, and transparent.
14
15            d)     Risk ratings are traditionally the product of a committee instructed to
16                  reach a consensus.  Elicitation of the opinions of many individual experts,
17                  through a formal process, can provide an alternative approach  to rating
18                  environmental health risks, especially when data are incomplete,
19                 contradictory, or multidimensional. This approach also provides the
20                 means to formally address uncertainties in the ratings,  and to elicit
21                  information on the factors that influenced the ratings.
22
23           e)    Development of a single, merged ecological and human health risk list of
24                 ratings requires value judgments.  Ecological and health risk ratings can
25                 be developed within  a consistent conceptual framework, with many
26                 commonalities in the factors used to rate risks.
27
                                             3-26

-------
      IRP Integrated Draft Report-Peer Review Draft, May 3,1999—Do Not Cite or Quote
 1
 2     3.8 References Cited
 3
 4     Adler M. and E. Ziglio. 1996. Gazing into the Oracle: The Delphi Method and its
 5           Application to Social Policy and Public Health. [London: Jessica Kingsley
 6           Publishers]
 7
 8     tinstone H. A., and Turoff, M. 1975. The Delphi Method: Techniques and Applications.
 9           [London: Addison-Wesley Publishing Co.]
10
11     Morgan M. G. and M. Henrion. 1990. Uncertainty: A Guide to Dealing with Uncertainty
12           in Quantitative Risk and Policy Analysis. Cambridge: Cambridge University
13           Press.
14
15     Morgan M.G. and D.W. Keith. 1995. Subjective judgments by climate experts. Environ.
16           Sci. Technol. 29:468A-476A.
17
      National Research Council. 1996. Stem, P.C. and H.V. Fineberg (Eds).
19           Understanding Risk: Informing Decisions in a Democratic Society.  National
20           Academy Press, Washington, DC.
21
22     Science Advisory Board. 1990.  Reducing Risk: Setting Priorities and Strategies for
23           Environmental Protection (EPA-SAB-EC-90-021). U.S. Environmental Protection
24           Agency, Science Advisory Board. Washington,  DC.
25
26    U.S. Environmental Protection Agency. 1987. Unfinished Business: A Comparative
27           Assessment of Environmental Problems. Office of Policy Analysis,  U.S.
28           Environmental Protection Agency, Washington, DC.
29
30    U.S. Environmental Protection Agency. 1992. RADON REPORT??
31
                                           3-27

-------
       IRP Integrated Draft Report-Peer Review Draft, May 3,1999—Do Not Cite or Quote


 i                 Appendix 3A. Health Risk Assessment Introduction
 2
 3     Thank you for agreeing to participate in this survey. Its purpose is to inquire about how you
 4     rate the health risks of a variety of environmental stressors. It also asks about the degree of
 5     confidence you place in your ratings or in the available scientific data. We are also interested in
 6     the reasons for your risk ratings. This information will be incorporated into a report for EPA's
 7     Integrated Risk Project, an undertaking of the Science Advisory Board.
 8     Here is how it works. At the bottom of this page, you select Go to the Register Form. An
 9     entry form will appear. To register using the Registration Form, enter the information
10     requested and choose a password, as indicated. You can also practice using the system by
11     entering test and fesf for the username and password, respectively. When you select Login,
12     the Stressor Risk Characterization and Health Risk Rating Data Sheet will appear. On the
13     same page, you can select Review of General Instructions, which describes your task. It also
14     contains descriptions of some of the criteria (Factors) you may have used to assign your
15     ratings. After reviewing the instructions, you Exit the page (or click on Go to Stressor Risk
16     Characterization Data Sheet) and proceed to rate the stressors. When you choose Select a
17     Stressor to Rate, a list of stressors will appear. Once you have made a selection, you are
18     asked to make a rating of Health Risk from Very High to Very Low. Your will also mark your
19     confidence in your rating. "Confidence" is a way of dealing with the largely subjective nature of
20     such ratings, and permits us to join combinations of risk and confidence ratings for the
21     assignments of priorities or ratings. For each of the stressors, we have tried to assemble the
22     most relevant information in a highly compressed format. You access this information by
23     selecting Open Information Window. After you have reviewed this information, you return to
24     the rating form either by selecting Close Information Window (if you have the rating form
25     open concurrently) or by exiting the information page.
26     Once you have rated the first Stressor, you then go on to the next Stressor in the list. If you do
27     not wish to rate a particular sstressor either because of lack of information or because of the
28     nature of the Stressor, there is no reason for you to do so; that is important  information in itself.
29     We have provided opportunities for you to offer additional reasons for your  ratings and for
30     general comments.
31     We have confined this demonstration to a limited number of stressors. The forms can be
32     expanded. Try it so we can see how it works and how we might want to proceed.
33     When you have finished making your selections, check your answers and then push the "Send
34     Results" button. That will mail the answers to me (Bernie Weiss). For technical details or
35     problems e-mail Geoff Inglis (inglis@envmed.rochester.edu).
36     You may send additional comments on these pages to Bernie Weiss. Your comments will be
37     posted for the group to read on the "comments page"

                                                3-28

-------
    IRP Integrated Draft Report-Peer Review Draft, May 3, 1999— Do Not Cite or Quote

                 Go to the Register Form
                       If \/mi arp alrpaHv rociistpfffri'        ___ —
3                 Go to the Login Form
4
5
6
7
                               3-29

-------
       IRP Integrated Draft Report-Peer Review Draft, May 3,1999—Do Not Cite or Quote


 i                            Appendix 3B.  Instructions
 2
 3     For each environmental stressor, please indicate:
 4     Your expert judgment of the relative environmental health risk that it poses to the
 5     U.S. population, either current or potential future risk. Assume that all current
 6     controls and regulations remain in place for the rating. Please rate as:
 7     Very High (VH), High (H), Medium (M), Low (L) and Very Low (VL) or (?).
 8
 9     We have provided some information on each stressor in a "Risk Data Sheet." For
10     the relative risk rating, please use your expert judgment, and all, some, or none of
11     the information provided on the risk data sheet.
12     If you feel that you cannot assign a relative risk rating to any given stressor, for
13     any reason, please indicate this with a "  ?" in the entry for rank.
14     Your level of confidence in your rating.
15     Very High (VH), High (H), Medium (M), Low (L) and Very Low (VL) or (?}.
16     Your confidence in the relative health risk  rank that you assign to a given
17     stressor can include both your judgment of the state of the existing scientific
18     knowledge about the health risks of the  stressor plus a judgment of your own
19     expertise regarding any particular stressor.
20      Please indicate also (by checking) the major factor or factors that influenced
21     your relative risk rating. For example, one  particular factor might have been
22     decisive. Alternatively, a collection of several factors might have determined your
23     rating.
24     Check all factors that apply.
25     Size of population affected: A high risk rating might be based on evidence that
26     many members of the population are experiencing adverse health effects due to
27     environmental exposures to the stressor. Conversely, a low risk rating might be
28     based on evidence that few, if any, members of the population are experiencing
29     adverse health effects as a result of exposure to the stressor.
30     Particular subpopulations : Although most of the population may not be exposed
31     to levels of the stressor great enough to cause adverse health effects, a rating of
32     high risk might arise if some smaller sub-populations are subjected to very high
33     exposures and risks; or, the stressor is  might be viewed as acting on an
34     especially susceptible subpopulation. A rating of low risk might arise if little

                                          3-30

-------
      IRP Integrated Draft Report-Peer Review Draft, May 3,1999—Do Not Cite or Quote

 1     evidence points to high exposures or risks for any subpopulations or if extreme
 2     sensitivity has not been identified in any subpopulations.
 3     Severity of health effects: A high risk rating might be based on the presence of
 4     pronounced, irreversible, or life-threatening adverse health effects irrespective of
 5     the size of population affected. In parallel, a low risk rating might be based on
 6     mild, transient health consequences arising from exposure to the stressor
 7     especially if they occur in a limited number of individuals.
 8     Percent of attributable incidence: A high risk rating might arise if a significant,
 9     though small, proportion of the total U.S. incidence of a health effect is seen to
10     arise from exposures to the environmental stressor; e.g., it is estimated that
11     about 1-2% of the annual lung cancer incidence in  the U.S. is due to radon. In
12     contrast, a low risk rating might be based on lack of evidence that a significant
13     proportion of the incidence or prevalence of an identified health effect can
14     reasonably be attributed to the stressor.
15     Persistence in the environment: A high risk rating might follow if the stressor is
16     believed not to degrade rapidly in the environment but, instead, tends to
      accumulate over time in one or more environmental compartments. A low risk
      rating might follow if the stressor is believed to degrade relatively rapidly in the
19     environment.
20     Potential future risk: A high risk rating might be based on predictions of a
21     significant potential  for environmental health risks in the future if no actions are
22     taken to reduce anticipated exposures arising from identified sources. A low risk
23     rating might be based on the assumption that no rise in or relatively minor
24     increases in exposures to the stressor are likely to be seen in the future.
25     Other. Please describe briefly any other reason for your ratings.	
26     Comments: Provide, if you wish, comments further amplifying or describing the
27     reasons for your choice of ratings.
28     To Summarize: Please be sure that you have responded  to all three questions:
29     3Your relative risk rating
30     3Your confidence in your rating
31     SThe factors determining your responses
32
33
34

                                          3-31

-------
     IRP Integrated Draft Report-Peer Review Draft, May 3,1999—Do Not Cite or Quote

1                        Go to Stressor Risk Introduction page.
2
3                  Go to Stressor Risk Characterization Ranking Sheet
4
5
6
7
                                         3-32

-------
     IRP Integrated Draft Report-Peer Review Draft, May 3,1999—Do Not Cite or Quote

I
2              Appendix 3C. Risk Characterization Data Sheets
                                     3-33

-------
PART III  INPUTS TO ENVIRONMENTAL DECISION-MAKING:




            ECONOMICS AND VALUATION

-------
       IRP Draft integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote


 1          PART III—INPUTS TO ENVIRONMENTAL DECISION-MAKING:
 2                          ECONOMICS AND VALUATION
 3
 4                                       Preface
 5
 6
 7           Integrated environmental decision-making at its heart is about how people,
 8     acting as individuals or through their government, can make themselves best off
 9     through actions that affect the environment directly and indirectly. Judging which
10     among alternative actions will produce the most "well being" requires information not
11     only on risks, discussed in the previous chapters, but also on the definitions of "well
12     being", i.e., goals, and the relationships among goals and among alternative choices or
13     actions. Just as the IED framework requires that we look at the full range of risks, it is
14     also necessary to consider the full economic consequences of decisions of whether and
15     how to address certain environmental risks or sets of risks. An understanding of the
16     tradeoffs implied by these choices is crucial both during Problem Formulation and
       Analysis and Decision-Making.
18
19           The Economic Analysis Subcommittee (EAS) was given the task of describing
20     the ways in which benefit/cost analysis can be used to frame choices about different
21     uses of societal (public and private) resources. In addition, at the request of the Deputy
22     Administrator and in recognition of the fact that some environmental changes are
23     difficult to monetize and value, a complementary subcommittee, the Valuation
24     Subcommittee, was formed to focus explicitly on improved methods for assessing
25     ecosystem values and incorporating that information into decision-making.  The VS was
26     a multi-disciplinary group that included ecologists, economists, and other social
27     scientists. The work of these subcommittees is contained in the following two chapters,
28     which taken together provide a framework within which choices about the environment
29     and its protection can be understood and implemented to best effect.
30
31           As discussed in previous chapters, environmental decision- making requires an
32     understanding of the physical and biological relationships between stressors-
33     substances and conditions that can cause harm-and organisms or ecological systems
34     of concern, including the relationship between alternative levels of these stressors (or of

                                            Ill-i

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1     substitutes for them) and their effects.  This is the province of science, and with other
 2     factors included, is known by the term "risk assessment."
 3
 4           By itself, however, such information is inadequate to guide decisions. The
 5     reason is that there are a near-infinite number of substances and conditions that can
 6     and do impose harm, and it is impossible and unwise to act to eliminate or even to
 7     reduce them all. The ability to lessen the risks they pose is limited by the resources of
 8     labor, capital, knowledge (technology), and physical endowment that is devoted to the
 9     task.  And these same resources are those on which people depend to provide the
10     other things desired and indeed necessary for life, including food, shelter, fiber,
11     manufactured goods, amenities, and investment to provide for future populations and
12     for increases in ability to satisfy  wants in the future. A critical task during Problem
13     Formulation, therefore, is to choose which of the risks (potentials for harm) or
14     combinations of risks are to be acted upon, and to what degree; i.e., to set
15     environmental or risk reduction goals.  Intimately related, of course, is the choice of the
16     most effective and efficient mechanisms and instruments to be used in actually
17     accomplishing the task; these aspects of options selection are discussed in Part IV.
18
19           The premise our democratic society has selected as the fundamental building
20     block for such choices is that they should be guided by the individual goals and desires
21     of people as expressed directly  and in and through their governmental and other
22     institutions.  Logic then suggests that the task of the decision-maker is to maximize the
23     attainment of these goals and desires with respect to both environment and all the other
24     things that people want. These include, of course, things material and non-material,
25     and also people's goals for future generations and in fulfilling their stewardship
26     responsibilities for natural systems. Put directly, then, it follows that the task of
27     environmental protection activities is to provide-for now and for the future-the
28     healthiest, safest, most ecologically secure set of conditions that the American people
29     are willing to pay for (in terms of other things they must give up to get them).
30
31            Chapters 4 and  5 provide a framework to guide decision-makers to deliver on
32     this task.
33
34
                                               l-n

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1           Chapter 4 adopts the commonsense framework for decision-making of
 2     comparing the advantages (benefits) and disadvantages (costs) of the range of
 3     plausible actions presented by a situation. It first addresses the difficult question of how
 4     to sort out the true benefits and costs of actions designed to reduce risks and improve
 5     the environment, and in the process identifies the many ways in which unsophisticated
 6     efforts to assess economic consequences both underestimate and  overestimate some
 7     elements of benefits and costs, and totally ignore others. It then discusses techniques
 8     for measuring and eliciting the benefits and costs of environmental  actions, and
 9     describes the requirements for and difficulties and uncertainties of such techniques.
10
11           Chapter 4 demonstrates that the development of benefit/cost measures in terms
12     useful to decision makers is not a simple task. It requires disciplined treatment of a
13     number of factors. For one thing, the treatment of a stream of benefits and costs as
14     they are realized over time is crucially important. The chapter concludes that
15     discounting of those streams in a consistent way to show current decision-makers the
16     full implications (present and future) of alternative choices is essential to provide
17     meaningful  information. It also recognizes that this discounting process can be
       controversial, and that it cannot be undertaken blindly.
19
20            Another matter of concern in presenting benefit/cost information is how to treat
21     the fact that while analysis can demonstrate which risk management scenario  provides
22     the greatest net benefits, it cannot take account of the fact that the distribution of the
23     costs and benefits is not uniform. Sometimes the costs are borne by some people and
24     the benefits gained by others. More frequently, some may lose a little and others a lot,
25     while at the same time others are benefiting disproportionately or not at all.
26     Complicating all of this is the fact that individuals affected by the environmental action
27     are unequally endowed in wealth and income. This influences both the benefits and
28     costs as they enter the evaluation framework, and the sense of fairness that pervade
29     the results. Chapter 4 recognizes that these distributional issues cannot be handled
30     within the framework of benefit/cost analysis, but it asserts that this makes them no less
31     important in the choice of actions to be taken. Consequently, distributional issues are
32     seen as matters on which the analysis can shed light, but which must be taken into
33     account outside the analysis at the final decision stage.
34
35           Chapter 4 conveys a mixture of confidence and humility. The framework of

                                             Ill-iii

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1     benefits and costs is seen as a consistent, coherent, transparent, robust tool that
 2     provides decision-makers with a firm basis on which to proceed. When well done, the
 3     products of specific benefit/cost analyses are seen as important inputs for decisions,
 4     sufficient to the task to be relied upon with confidence most of the time, always limited
 5     by data and sometimes by methodology, and sometimes sufficiently uncertain as to
 6     provide only indicative information—which nonetheless can be useful.
 7
 8           As with risk assessment and comparison, however, even the best and most
 9     complete benefit/cost analyses are also shown to be inadequate in themselves to yield
10     "answers." Other factors always are present and must be taken into account.  In
11     essence, integrated environmental decision-making requires consideration of the broad
12     implications of actions taken to improve the environment across stressors and across
13     options to relieve them.  The framework outlined in Chapter 4 provides a
14     methodological approach to such decision-making.
15
16           Chapter 5 is a vital complement to Chapter 4. It addresses the important issue
17     of how the benefits (and sometimes costs) of changes in ecological outcomes can be
18     properly determined and incorporated into the decision process. By their nature,
19     changes in ecological conditions are often not easily observed in quantifiable terms that
20     can be rendered in the monetary units that most often are used to compare possible
21     outcomes in a benefit/cost analysis.  For this reason, it is often thought that the benefits
22     received from ecological protection are under-estimated, and consequently, that
23     ecological systems are under-protected as compared to other goals and desires of
24     people.
25,
26           Chapter 5 examines the issue of ecological valuation in detail. In doing so, it
27     starts with an examination of ecological values and concludes that in principle they are
28     not different from other values, and that they need not enter into the decision process in
29     any unique way. At the  same time, however, measuring and incorporating the values
30     ascribed to anthropogenic changes in ecological conditions do present serious
31     difficulties that require that special care be taken.
32
33           For one thing, citizens doing the valuing often have insufficient knowledge  of how
34     changes in ecological factors affect the things they care about.  They need expert
35     scientific assistance in making the important connections. For another, many of the

                                             Ill-iv

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1      benefits from ecological systems are subtle and do not enter into conscious
 2     consideration in the normal course of events, as do market goods and services.  For
 3     these, special techniques for eliciting preferences are required.  In addition, some of the
 4     values ascribed to ecological outcomes arise in a social context and may not surface
 5     from commonly used individual preference measures.  Equity, sustainability, and
 6     stewardship are often cited as examples of such values.
 7
 8           Most importantly, because ecological services often do not enter into markets or
 9     enter incompletely, it is difficult if not impossible to provide monetized measures of the
10     benefits they deliver. Often it is not possible to even determine quantitative measures
11     for differences in outcomes.  Qualitative measures of benefits and costs, therefore,
12     must be arrived at and then incorporated into decision processes. In short, the
13     valuation of ecological costs and benefits  is prone to error because elements valued by
14     people may be omitted or incorrectly specified and because measurement is inherently
15     more difficult than with goods and services for which market and market-like measures
16     are available.
17
             Chapter 5 offers suggestions to practitioners  of benefit/cost analysis and
19     decision-makers that will improve the incorporation of ecological matters into decisions.
20.    It goes further, however, in making two recommendations. The first is that there be an
21     expanded use of deliberative processes to assure that all relevant elements are
22     included in decisions, and that they are valued properly. While deliberative processes
23     are seen to have an important role, Chapter 5 also notes that such processes should  be
24     used discriminately and be tailored to the situation.  When this  is done, deliberative
25,    processes will  not only generally lead to outcomes that are more satisfactory and
26     robust, but also will not cause inordinate delay and indeed may even speed the delivery
27     of ecological protection because they lessen post-decision controversy.
28
29           The other key recommendation of  Chapter 5 is for further research on and
30     experimentation with additional approaches to valuing benefits of environmental
31     systems.  Existing approaches are seen to be inadequate in their treatment of such
32     values as fairness and sustainability. They also have difficulty  in incorporating the
33     systemic benefits of such matters as biodiversity, and are incomplete in their treatment
34     of dynamic responses to change. More holistic approaches that take account of the
35     web of interactions involved in decisions surrounding ecological systems and their

                                             Ill-v

-------
      IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

1     interactions with other production and consumption activities would be helpful.
2
3           The overall message from Chapters 4 and 5 is that integrated environmental
4     decision-making requires a framework within which the decision-maker can meld the
5     results of science and the goals and values of the people served and formulate
6     acceptable decisions. The further message of these chapters is that such a framework
7     exists, but that its use requires skill and artistry, and indeed that in some respects the
8     framework itself remains a work in progress.
9
                                            Ill-vi

-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote


   CHAPTER 4.  BENEFIT/COST ANALYSIS FOR INTEGRATED RISK
                              DECISIONS

                        TABLE OF CONTENTS


4.1 Introduction	4-1

4.2 Fundamental Questions in the Economic Analysis of Risk	4-3

4.3 The Benefits of Risk Reduction	4-8
      4.3.1  Revealed Preference	4-13
      4.3.2 Stated Preference	4-15

4.4 Costs of Environmental Protection 	4-17

4.5 Comparing Total Benefits and Total Costs	4-22
      4.5.1 Calculating Net Benefits	4-22
      4.5.2 Present Values and Discounting	4-22
      4.5.3 The Net Benefit Criterion and the Scope of Public Projects	4-24
      4.5.4 Uncertainty in the Measurement of Benefits and Costs	4-26

4.6  Distributional Considerations 	4-27

4.7 Conclusions 	4-28

4.8 References Cited	4-31

Endnotes  	4-33

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote


 1        CHAPTER 4.  BENEFIT/COST ANALYSIS FOR INTEGRATED RISK
 2                                     DECISIONS
 3
 4
 5     4.1 Introduction
 6
 7           No society, no matter how wealthy, has available the resources to "solve" every
 8     problem that confronts it.  Schools will never be as good, nor food, clothing, shelter and
 9     access to medical care as abundant as we would all like. Similarly, no society-modem
10     or otherwise - will ever be free of all environmentally mediated risks to human health
11     and ecosystems. Thus, it is of great importance to be able to identify the most serious
12     threats to health and the environment, and also to target certain of them for possible
13     mitigation.
14
15           Identifying what we might call the total risks associated with air and water
16     pollution, solid and hazardous waste disposal, drinking water contamination and other
       activities is largely the province of risk assessment.  Determining these risks requires us
       to ascertain which pollutants are extant in the ambient environment, the route(s)
19     through which humans or ecosystems are exposed to these pollutants and the duration
20     of these exposures, and the resulting consequences. At a minimum, this requires the
21     skills of engineers and chemists; atmospheric, terrestrial, and aquatic scientists;
22     toxicologists, epidemiologists and biostatisticians, and clinical health specialists; and
23     ecologists (including botanists, forest and fisheries scientists and others). Some risks
24     may be described quantitatively, occasionally with some statistical precision. Often, we
25     can do no more than suggest that a particular pollutant may have adverse effects on
26     human populations, or on ecological systems. As described in Part II, assessment and
27     comparison of risks associated with different stressors or activities is an important part
28     of both formulating "the problem" and analyzing possible risk reduction options.
29
30           Economics typically concerns itself with the way in which people allocate scarce
31     resources among many competing needs so as to make themselves-broadly
32     speaking-as well off as possible. As such, the possible contributions of economics to
33     Integrated Environmental Decision-making lie largely-though not exclusively- in
34     deciding which risks it makes sense for society to make incremental changes in

                                            4-1

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote


 1     (sometimes called "marginal" changes) once they have been characterized, in deciding
 2     to what degree those risks ought to be reduced, and in identifying appropriate policy
 3     instruments that can be employed to achieve such risk reductions.
 4
 5            Economics, along with other behavioral sciences, can also help in understanding
 6     risks by characterizing the extent to which changes in behavior can effect exposure to
 7     stressors. For example, although many people spend time outdoors on a clear day, it
 8     would be incorrect to count them as the population at risk on a very polluted day - for
 9     the simple reason that some of those people would alter their behavior on account of
10     the pollution. Similarly, contamination of an aquifer used for drinking water may well
11     increase some people's risk of acute or chronic illness, but may not affect all users if
12     everyone  were informed about the risk and most chose instead to drink bottled water.
13     In both of these cases, economics would suggest that there were "damages" associated
14     with the pollution, but these damages can take forms other than just increased risk of
15     illness.
16
17            It is very important to note the distinction drawn above between risk posed by a
18     certain exposure level and the possible "changes" in risk that would result from control
19     actions that reduce either stressor or exposure levels, or both. Economic analysis is
20     especially useful in determining which changes in risk level make sense to pursue via
21     policy measures and which  do not.
22
23            The focus of this chapter is on the role of economic analysis, particularly
24     benefit-cost analysis (or BCA), in helping society to decide which environmental risks to
25 *    address first and which to leave for a later time. The IED framework recognizes that
26     such analysis should be an important component of decision-making, but not the sole
27     consideration. In fact, there are instances where benefit-cost considerations do not
28     play a direct role in decision-making.  For instance, although it would be most unusual,
29     decisions about which risks to address and how to address them could be determined
30     in a national plebiscite in which voters directly established risk management priorities;
31     voters in  California often do make decisions in statewide referenda (including-decisions
32     on important environmental matters) that are elsewhere made by legislatures, or even
33     by administrative agencies to which legislatures have delegated power. As suggested
34     above, such decisions are more frequently made by elected officials, who  sometimes
35     not only specify which environmental  risks are to be controlled, but also to what level of

                                              4-2

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote


 1     residual risk and using which control technologies or techniques.
 2
 3           Most commonly, however, risk management decisions are made by appointed
 4     officials and the civil servants who work for them, acting under statutes that delegate to
 5     federal, state or local regulatory agencies the power to establish priorities and
 6     determine who must take actions to reduce particular risks.  Their decisions can be,
 7     and generally are, informed by a variety of types of analysis, of which economic
 8     analysis is one. The IED framework is intended to help ensure that the broadest range
 9     of analyses are applied in an integrated fashion.
10
11     4.2  Fundamental Questions in the Economic Analysis of Risk
12
13           Once risks have been characterized and compared, at least two important
14     questions remain. How much, if at all, should each of the risks be reduced, and in what
15     order? How should these reductions be accomplished? Economics offers insights on
16     the answers to both questions. Let us address the second question first, because it is
17     perhaps the easier of the two. Because resources are scarce relative to human wants
       even in the richest of societies, economists answer the second question - how shall
19     risk reductions be accomplished? -- quite directly.  Generally, their answer is, "In  the
20     least expensive way possible." For, by reducing as cheaply as possible the risks
21     associated with indoor radon, heavy metals in aquatic ecosystems, or radioactive
22     contamination in landfills, to pick but three examples, more of society's resources are
23     available for education, national defense, health care or the private goods and services
24     that people desire. So long as we accomplish what we set out to do (in risk reduction),
25     why spend more than we have to?
26
27           In fact, there is a "cousin" of BCA - called cost-effectiveness analysis (CEA)~
28     that deals explicitly with this question. If, and this is a big "if," the only output of a
29     regulatory program is to reduce cancer incidence,  say, CEA would rank cancer
30     prevention programs on the basis of "cost per cancer case avoided."  Then the least
31     expensive option would be pursued first, followed by the next most attractive one, and
32     so on until a decision had been made to no longer pursue reductions in cancer risk.
33     The point at which this process stops  is the subject of a subsequent section.  This
34     approach would maximize the reduction in cancer incidence for the amount of money
35     spent. Other patterns of spending the same amount would prevent fewer cancers.

                                            4-3

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote


 1           There are several problems with CEA that bear mention here. First, it only works
 2     for "apples vs. apples" comparisons. If one program reduces cancer cases and also
 3     improves the visibility of the air, while another reduces cancer and protects aquatic
 4     ecosystems, too, they cannot be compared on a simple cost per cancer case avoided
 5     basis.  Most environmental regulatory programs have multiple outputs (that is, they
 6     provide more than one kind of beneficial effect). Thus, CEA is often not of much
 7     practical assistance in risk ranking and  management. Another problem with CEA was
 8     alluded to above.  While it is easy to see the common sense in starting with the
 9     cheapest control opportunities first, and working our way up the list, CEA provides no
10     obvious "stopping rule" unless Congress, for instance, has limited the size or cost of the
11     cleanup effort. That is, it does not tell us when we have reached a cost per cancer
12     case avoided that is too much from society's standpoint.
13
14            Still another problem with CEA is that it treats cases of cancer - or other
15     endpoints - as being homogenous, when in fact they may not be. For example,
16     suppose the cheapest way to prevent cancer cases is to pay the cost of smokers'
17     participation in smoking cessation programs that they would not otherwise attend.
18     Some  members of society, perhaps many, might object to such a program on the
19     grounds that these risks are bome voluntarily - at least more voluntarily, say, than
20     those that arise from  airborne exposures to other carcinogens. Even though the latter
21     might be more expensive to reduce, society might prefer to address them first.  In other
22     words, we care about dimensions of risk other than sheer statistical magnitude. In
23     addition to the voluntary or involuntary  nature of the risk, these other dimensions may
24     include such things as the degree of "dread" associated with the risk; for instance, for
25,    whatever  reason,  people seem to fear  radiation-induced cancers more than those
26     associated with other causes. Despite this and its other limitations, CEA has been a
27     very useful tool in assisting regulatory officials interested in rationalizing risk reduction
28     programs.
29
30            One final word about doing things for least cost.  The attractiveness of this idea
31     is the reason why economists and others are often enthusiastic about what have come
32     to be called market-based or economic-incentive approaches to  environmental policy.
33     These include such things as taxes on pollution, tradable  discharge permits,  and
34     deposit-refund schemes, among others. The attraction of these approaches is that they
35     are able not only to provide incentives  to meet an overall environmental goal (mercury

                                              4-4

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote


 1     emissions are reduced to x tons per year, for instance), but also that the costs of
 2     complying with this limit are minimized across society. Furthermore, the burden of
 3     identifying the least cost options falls on the private sector, which is especially good at
 4     reducing costs. This is not the place to go into the reasons why this is so, though it is
 5     worth pointing out that such approaches have been used quite successfully in the
 6     United States (for the phase out of leaded gasoline in the 1980's, and for the control of
 7     sulfur dioxide in the 1990 amendments to the Clean Air Act) and in a number of
 8     countries in Europe and elsewhere around the world.
 9
10           This brings us to the question of how much reduction we should seek for the
11     various risks that society faces. There are several ways this question can be answered.
12     One approach would be to direct the regulatory authorities to set standards for the risky
13     pollutants so as to provide a margin of safety against adverse health effects - an
14     approach premised on thresholds. This is the directive Congress has given to the EPA
15     in important parts of the Clean Air Act, the Clean Water Act, the Resource Conservation
16     and Recovery Act, and in  other environmental statutes as well.
17
             This seemingly attractive approach often founders when the underlying science
19     suggests - as it often does - that there is no "safe" level; that is, a threshold
20     concentration cannot be found that guarantees the health of affected individuals or
21     ecosystems. This issue has arisen recently in the context of the revision of the National
22     Ambient Air Quality Standard for ozone, and the establishment of a new standard for
23     particulate matter of 2.5 microns and less. In both cases, the Clean Air Science
24     Advisory Committee has indicated that no clear "bright line" can be found such that
25-     concentrations below that level can be regarded as being safe.
26
27           There is another problem with the threshold approach.  It might sometimes be
28     the case that even if a safe level could be found, society might deem it too expensive to
29     provide, especially if a somewhat less protective level costs substantially less.  A
30     safe-standard approach denies the legitimacy of such tradeoffs. Once again, this issue
31     has arisen in the context of air quality standard setting, with the EPA having publicly
32     acknowledged that setting the ozone standard at too low a level could have significant
33     adverse economic repercussions.
34
35           Another way to decide the question of how much protection to provide is to tie

                                             4-5

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote


 1     the answer to affordability. That is, one could decide that any risk reductions that
 2     regulated entities (whether corporations, lower levels of government, or individuals) can
 3     "afford" are worth undertaking.  This is the approach often taken when the EPA writes
 4     discharge standards for new sources under the Clean Air or Clean Water Acts.  Indeed,
 5     they have been directed to do so by Congress, which has drafted statutes in which
 6     "economic achievability" is one of the relevant criteria. While this approach may seem
 7     to deal satisfactorily with all economic concerns, this is not so.  First, controls on some
 8     sources might be desirable even if the source could not afford them - in some
 9     instances, completely shutting down a plant may turn out to be in the best interests of
10     society if there is no other alternative measure, and if that plant discharges substances
11     that poses serious threats to human health and/or the environment.
12
13           Second, some controls may be affordable but nevertheless not worth pursuing.
14     This would be the case for relatively trivial risks that would be expensive to ameliorate.
15     Even  if the sources of these risks were quite profitable, many people would be
16     uncomfortable spending a lot of money on not-so-serious problems.  The problem with
17     requiring successful firms to do more than those that are just breaking even  is that it
18     creates exactly the wrong incentives for firms: do well and you will be more  heavily
19     regulated, do poorly and we'll take it easy on you. That is neither the right
20     environmental nor the right economic signal to be sending.
21
22           There is, of course, a third way to make decisions about which risks to control
23     and by how much. This approach involves an effort to strike a balance between, on the
24     one hand, the good that will be done for human health and the environment when one
25  ,   or more risks are reduced and, on the other, the costs or other adverse consequences
26     associated with taking such actions. The criticisms above of both the threshold and the
27     affordability approaches go right to the heart of this matter.
28
29            Benefit-cost analysis is the technique most commonly used to assist in balancing
30     the favorable effects of risk reductions (the benefits) with the adverse consequences
31     (the costs).  Underlying BCA is the notion that it may be possible to measure both the
32     good  and bad effects of a policy change (a measure to reduce one or more  risks in this
33     case) using a common  denominator - money.   If so, one can then ascertain whether
34     the gains to the gainers outweigh the losses to the losers, and thus determine, on net,
35     whether society as a whole is made better or worse off as a result of the change. A

                                             4-6

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote


 1     whole host of possible risk reduction policies could then be compared on the basis of
 2     which ones do the greatest net good and which do little or no net good.
 3
 4          We offer several observations about BCA before moving on to a discussion
 5     about how one might value benefits and costs, make comparisons among them, and
 6     take into account effects that are felt at different points in time.
 7
 8          First, and perhaps most importantly, most economists think of BCA as no more
 9     than a tool to assist in decision making. This is important because most
10     non-economists (and perhaps some economists) believe that BCA must be used as a
11     decision rule.  That is, they believe that the use of BCA means that only those policy
12     changes that pass a benefit-cost test should be put in place. To repeat, most
13     economists do not hold this position, although all economists would argue that the
14     information in a well-done BCA can be of great value in helping to make decisions
15     about risk reduction policies (Arrow et  al., 1996).
16
17          Second, when trying to translate benefits and costs into dollar terms, it is the
       preference of individuals that count. That is, we look to each individual to determine
19     how to value a given benefit or cost.
20
21          This is often a hard pill for non-economists to swallow. Some people reason that
22     if the public knew what they themselves know about a particular problem, surely
23     everyone would ascribe greater  benefits to solving the problem than they do currently.
24     And sometimes it is the case that willingness to pay to reduce a particular risk is small
25     solely because affected parties do not understand the consequences. Often, however,
26     they understand the problem quite well, but simply do not care as much about the
27     effects as someone else. It is unclear how one would attempt to value the
28     consequences of a policy change to someone without attempting to determine that
29     person's considered valuation of the change.
30
31           Another difficulty with BCA has to do with its reliance on  individuals' valuations,
32     however they might be revealed. When willingness to pay is used as the monetary
33     value of a change, as it generally is, the values that individuals  reveal are conditioned
34     by their incomes. Thus, an individual  will be willing to pay no more each year for a risk
35     reduction than the total amount  of money he/she has available  after taxes.  Accordingly,

                                            4-7

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote


 1     the social benefit (as counted in a BCA) of reducing the risk of premature mortality for a
 2     wealthy person will generally be greater than doing so for a poor person (because the
 3     former will be able to pay more for the same statistical reduction in risk than the latter).
 4     This is a problematic aspect of BCA for almost all economists.  From time to time,
 5     people have attempted to derive weights that might be used to inflate the preferences
 6     of the poor so that they count for more in BCA.  The problem with this approach is
 7     obvious to all - who shall be delegated the responsibility of assigning the weights,  and
 8     on what basis should these weights be derived? The final section of this chapter takes
 9     up such questions related to the distribution of benefits and costs.
10
11           We turn now to a discussion of the benefits of risk reduction policies, and how
12     they can be estimated and expressed in dollar terms.
13
14     4.3  The Benefits of Risk Reduction
15
16           Economic valuation of  environmental resources has been characterized by some
17     other disciplines as being anthropocentric and utilitarian, and inattentive to the intrinsic
18     value of these resources. To some degree this is true. If an environmental change
19     never matters in any way to any human - today or in the future - then it will not, even in
20     principle, show up in any economic valuation or assessment. As Freeman (1997)
21     points out, however, it is important to distinguish between ecosystem functions (e.g.,
22     photosynthesis, absorption, and dispersal) and the environmental services produced by
23     ecosystems that are valued by humans.  The range of these services  is great. They
24     include obvious environmental products such as food or fiber and services such as
25     flood protection, but also include the quality of recreational experiences, the aesthetics
26     of the landscape, and such desires (for whatever reasons) as the protection of marine
27     mammals.  Humans may "passively"  value environmental services. For example,
28     "existence value" reflects human recognition of the so-called "intrinsic" value of an
29     ecosystem.  If an environmental measure produces a change that matters to humans,
30     for whatever reasons, then it  is an item to be counted in the assessment, at least in
31     principle.  In practice, there are, however, numerous problems.  Table 4-1 presents
32     examples of environmental and resource service flows for which value measures might
33     be desired (Freeman, 1993).
34
35            Policies for environmental protection often affect the functioning of ecosystems.

                                             4-8

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote


 1     The social benefits of this protection depend on the link between these functions and
 2     some associated service flow that is valued by humans.  In most cases an
 3     environmental protection measure will enhance environmental services marginally, and
 4     it is this change in services that economists seek to value.  Before the benefits of
 5     environmental protection can be calculated, therefore, it is necessary to predict the
 6     actual consequences on natural systems from failing to engage in protective activities.
 7     Good science is essential to competent environmental benefit-cost analysis.
 8
 9           In some cases the physical consequences of pollution are clear-cut and easy to
10     quantify. For example, a single "point source" of water pollution upstream renders
11     cooling water for a downstream factory sufficiently impure that special pre-treatment
12     equipment must be installed before the water can be used.  In other cases,  however,
13     the connection is far less clear.  Deteriorating air quality due to pollutants from a wide
14     range of sources (vehicles, factories, and perhaps even natural vegetation)  may reduce
15     local average life expectancy by several months for susceptible individuals.  About the
16     most that can be claimed is that the poor air quality changes the probability  that any
17     given individual will live to be, say, at least 70 years of age.  But so many  other
1      behavioral and genetic factors influence the individual's actual life-span that it is
19     impossible to say for sure what would be any individual's actual benefits from  reducing
20     air pollution.  Often, it is not even scientifically clear what effect reducing pollution from
21     one source (say vehicles) will actually have on overall ambient pollutant levels.
22
23           In still other cases, such as the dramatic alteration of ecosystems through
24     destruction of wetlands or through global warming, scientists are uncertain even as to
25     the nature of  the outcome. In these cases, economic valuation of potential  gains from
26     environmental protection measures falls short.  This is not because benefit-cost
27     methods ignore environmental services that do not produce market goods, but because
28     the pathways by which humans are ultimately affected cannot be well articulated.
29
30           It is obviously very much easier to value the benefits of environmental protection
31     when the effect is straightforward, as in the first example, but increasingly difficult as
32     the scientific uncertainty about the physical effects increases. Even in the intermediate
33     case, where the environmental service effects of the policy in question can be
34     articulated no more clearly than as a change in average life expectancies or a change
35     in the probability of respiratory disease, the task of ascertaining individuals'  values of

                                              4-9

-------
      IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote


1     these changes is somewhat complicated.
2
3           From an economic perspective, the environment can be viewed as a form of
4     natural asset that provides service flows used by people in the production of things like
5     recreation, agricultural output, the assimilation of pollutants, or even an amorphous
6     'good" such as "quality of life."  This is analogous to the manner in which real physical
7     capital assets (e.g., factories and equipment) provide service flows used in
8     manufacturing production. Like real physical capital, if the natural environment (as a
                                            4-10

-------
        IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
          TABLE 4-1.  Examples of Environmental and Resource Service Flows
        Resource or environmental media
              (Air, Water, Fishery, Forest)
        Effects
              Direct impacts on humans
                     Human health (morbidity/mortality associated with pollution)
                     Odor, visibility, visual aesthetic
              Ecosystem impacts (biological mechanisms)
                     Impacts on the economic productivity of ecological systems
                           Agricultural productivity
                           Forestry
                           Commercial fisheries
                     Other ecosystem impacts
                           Recreational uses of ecosystems-fishing, hunting
                           Ecological diversity, stability
              Impacts through nonliving systems
                     Materials damage, soiling, production costs
                     Weather, climate
        Economic channels
              Market values
                     Changes in income to producers
                     Changes in the availability/price for marketed goods/
              services to consumers.
              Non-Market Values, i.e.,  changes in availability of:
                     Health
                     Environmental amenities
                           Visibility
                           Opportunities for recreation
       Adapted from Freeman, 1993, pp. 13-14

 productive asset) is allowed to deteriorate, this lessens the flow of services it is capable
of providing.

       A decision to devote society's scarce resources to environmental protection,
however, means that these resources will not be available to be used for other
purposes. These other purposes might be more food produced today or greater current
manufacturing output, or it might be more investment in productive capital equipment or
research to enhance output in the future, all of which are also valuable to society.  It is
                                                4-11

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote


 1     not surprising that poor countries spend little on environmental protection. Banning
 2     DDT, or other harmful but cheap and effective pesticides, has little appeal for people
 3     whose current battle to protect themselves from immediate illness and to feed
 4     themselves today depends on these chemicals. The concept of valuation of
 5     environmental goods and services is couched in terms of society's willingness to make
 6     trade-offs between competing uses of limited resources, and by aggregating over
 7     individuals' willingness to make these trade-offs.
 8
 9           Economists' tools of valuation were originally developed in a more limited
10     context, one in which policy changes mostly caused changes in an  individual's income
11     and/or prices that he faced in the market. Over the last twenty years, however, these
12     ideas have been extended to accommodate changes in the qualities of goods, to public
13     goods that are "shared" by individuals, and to other nonmarket services such as
14     environmental quality and human health.  Many environmental goods and services are
15     considered by economists to be public goods because one person's enjoyment of them
16     does not diminish the ability of others to enjoy them  also.
17
18           We can think of people as having preferences among alternative "bundles" that
19     include both market and nonmarket goods.  Some of these nonmarket goods include
20     environmental services.  Typically people's preferences allow substitution - if the
21     quantity of one good in a bundle is decreased, the quantity of some other good would
22     have to be increased in order to leave the individual just as well off as before the
23     change.  This substitutability between goods establishes the trade-off between pairs of
24     goods that both matter to people. If one of the goods has a monetary value, then this
25     willingness to make a trade-off reveals the monetary value of the other good.  It is not
26     necessary that money be the metric for comparison, it is merely convenient.
27
28           When the alternative 'bundle" includes something dramatic, such as loss of life
29     or loss of a loved one, the individual will be understandably unwilling to  make trade-offs
30     with other goods. However, environmental actions rarely, if ever, imply such dramatic
31     outcomes with certainty.  An environmental protection measure is more likely to alter
32     slightly the probability of illness or loss of life, much like wearing a seat belt affects the
33     probability of death, and people are observably willing to trade off money and  time to
34     alter this probability.
35

                                             4-12

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote


 1           Rarely is it necessary to determine the value of the environment as a whole. The
 2     typical challenge is to ascertain the social value derived from maintaining some type of
 3     ecosystem function at its current level through pollution prevention or increasing its
 4     functions somewhat through more stringent regulations. Failing to implement the policy
 5     would compromise the ecosystem's functioning to some extent, thereby lessening the
 6     quantity of services the environment is capable of providing for human enjoyment.
 7
 8           When it is possible to observe freely functioning markets for goods whose
 9     availability is affected by policy, economists are fairly comfortable with the
10     appropriateness of their standard tools for estimating the benefits or loss of benefits
11     resulting from that policy. If a policy results in higher incomes or lower prices for them,
12     people are generally made  better off.  Higher incomes or lower prices mean an
13     increased difference between what the individual is willing to pay for a good and the
14     amount that he is required to pay. This difference is called a "surplus," and modem
15     welfare economics recognizes that welfare gains (or losses) as a result of some policy
16     are approximated well by increases (or decreases) in this surplus that result from the
17     policy's implementation.

19           The same logic still holds despite the absence of standard markets for many
20     environmental services. Standard markets make the task easier because consumers'
21     decisions as to how much of a good to purchase at different market prices helps reveal
22     something about the surplus they gain at any given price.  With non-market
23     environmental goods, it is necessary to infer this willingness to trade off dollars for
24     additional quantities of environmental services using other techniques.  Environmental
25     economists have developed a repertoire of techniques over the past thirty years that fall
26     roughly into two categories: a) indirect measurement (revealed preference), and b)
27     direct questioning (stated preference). The following is a brief survey.
28
29           4.3.1 Revealed Preference
30
31           Economists have always preferred to measure trade-offs by the observed
32     decisions of consumers in real markets whenever possible. These are called "revealed
33     preference" methods, because the consumers' actions reveal something about their
34     willingness to trade-off one  good for another.  In other words, we get to see them put
35     their money where their preferences are. The task is made easier if the researcher can

                                            4-13

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote


 1     exploit relationships that might exist between the non-marketed (environmental) good
 2     and a good that has a market price. For example, suppose a change in the
 3     environmental good only matters to a particular person if he is also a purchaser of
 4     some good that has a price (e.g., water quality is important to those who buy fishing
 5     rods).  Then the environmental good and this market good can be seen to be
 6     complements of sorts. The individuals' behavior with respect to the marketed good,
 7     when the environmental quality changes, will reveal something about how he values
 8     that environmental change. Likewise, if the individual can mitigate the negative
 9     consequences of the decline in some environmental service by purchasing some
10     market good (buying bottled water when aquifers are contaminated, for instance), then
11     the environmental good and the market good are substitutes of some sort. We can
12     leam something about the lost benefits of the environmental service from these
13     individuals' actions with regard to this substitute market good.
14
15            First, the Travel Cost Method (TCM) is perhaps the oldest of these revealed
16     preferences methods for non-market valuation and relies on the complementary
17     relationship. An individual will decide whether to take a trip (generally a recreational
18     trip) to consume an environmental good if the perceived benefits of the visit exceed the
19     costs of the visit.  In this case, environmental service might be viewed as a quality
20     characteristic of the trip that matters oniy if the individual visits the relevant site.  By
21     observing how people trade off such things as distance traveled to access sites of
22     better environmental quality, researchers can leam something about the value people
23     place on differences in this quality.
24
25            Second, economic methods that model individuals' decisions to avert or mitigate
26     the consequences of environmental deterioration may shed light on how people value
27     other types of changes in environmental quality.  These techniques are most applicable
28     where the failure to institute a given policy would increase risk of illness or loss of life,
29     although they have other applications as well. For example, if a household installs filter
30     systems or purchases bottled water to avoid contaminants in the groundwater that they
31     drink, then they are attempting to avert or mitigate the consequences of water quality
32     deterioration. By analyzing this type of behavior, economists  can at the least establish
33     bounds on the willingness of individuals to pay for improved water quality.  In this case,
34     bottled water is a substitute for clean groundwater in altering the risk of illness from
35     drinking water.

                                             4-14

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote


 1           Third, in some cases, individuals reveal their preferences for environmental
 2     goods in the housing market  Realtors know that the price paid for a property varies
 3     with the characteristics of that property. Additional bathrooms or bedrooms or a better
 4     view will show up in the transaction price of a dwelling, but local environmental
 5     characteristics will show up as well. For example, researchers have found that housing
 6     prices vary with air quality in many cities.  Home-buyers' willingness to pay for this
 7     additional environmental quality reveal something about their willingness to trade-off
 8     money for this type of environmental improvement.  This approach is called the hedonic
 9     property value method of valuation.
10
11           Fourth, wages for similar jobs in different locations may vary with the local level
12     of environmental quality. Higher wages may have to be paid to induce workers to take
13     jobs in areas with more pollution, greater health risks, or fewer environmental
14     amenities.  In fact, the values used by the EPA and other regulatory agencies for
15     reductions in mortality risk come almost exclusively from studies linking compensation
16     to the riskiness of the job. Through the hedonic wage method, it is sometimes possible
1"7     to attribute  differences in wages to differences in the quality of the environment, and
       thereby indirectly to value variations in environmental quality.
19
20           The  above methods are well established for measuring the conceptual trade-offs
21     that economists consider as the basis of environmental valuation. However they are
22     applicable only in special cases. For the first method - the travel cost approach - to be
23     complete, individuals must care about changes in the environmental good only if they
24     visit a particular site.  For the second to be complete, (that dealing with complementary
25     purchases), individuals' values for changes in environmental goods must depend solely
26     on their ability to produce some ultimate  service which can be produced as well using
27     market goods, but at some increased cost. In all of these methods, individuals must be
28     able to discern the environmental consequences and the researcher must be able to
29     find a marketed good with a special relationship to the environmental good that can be
30     exploited.  These conditions hold only in a subset of the cases in which environmental
31     protection measures need to be evaluated.
32
33           4.3.2 Stated Preference
34
35            Economists generally  mistrust trade-off information that has not been observed

                                             4-15

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote


 1     (even indirectly) in real markets, but for some changes in environmental goods and
 2     services, we simply cannot observe behavior that reveals their values to people. This is
 3     particularly true when the value is a passive one. For example, an individual may value
 4     a change in an environmental good because he wants to preserve the option of
 5     consuming it in the future (option demand) or because he desires to preserve the good
 6     for his heirs (bequest demand).  Still others envision no current or future personal use
 7     by themselves or their heirs, but still wish to protect the good because they believe it
 8     should be protected or because they derive satisfaction from knowing it exists in a
 9     protected state (existence demand). With no standard market trade-offs to observe,
10     economists often resort to surveys  in which they construct hypothetical markets. Value
11     information elicited this way is described as "stated preference" (as opposed to
12     revealed preference) information.
13
14            In this approach (often called contingent or hypothetical valuation), survey
15     respondents are presented with a well-defined scenario that requires them,
16     hypothetically, to trade-off something (generally money) for a change in the
17     environmental good or service in question.  While rarely asked in such a direct way,
18     researchers seek answers to questions such as: "How much would you be willing to
19     pay to prevent an additional  5% of the population of sea otters from being killed due to
20     a  hazardous waste spill?" (Or, alternatively, what compensation would you require in
21     order to willingly accept a reduction in this environmental amenity?) Early attempts to
22     use this method were fraught with problems. As might be imagined, the values that
23     result are susceptible to manipulation through the manner in which the valuation
24     questions are posed.  Protocols have been continuously refined, however, and the
25     reliability of these methods is improving. (For further discussion of the problems
26     associated with elicitation of values, see Chapter 5).
27
28            An alternative but related method is to infer values from individuals' hypothetical
29     behavior (as opposed to hypothetical payments). This approach can be particularly
30     useful when paying to prevent environmental degradation seems too implausible or
31     subject to bias.  Respondents might be asked  "If visibility were x miles and travel costs
32     were $y, how many trips would you take to site z?" In yet another approach (contingent
33     choice models), respondents are given descriptions of several scenarios and asked
34     which one they would prefer. One characteristic of each scenario is a level of  prices or
35     net income and another is environmental quality, allowing the researcher to infer

                                             4-16

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote


 1     something about respondents' willingness to trade dollars for variations in the level of
 2     environmental amenities.
 3
 4           The types of methods described briefly here can sometimes be used in
 5     conjunction with one another, either to measure different types of values that are
 6     generated by a change in an environmental resource or to better estimate one type of
 7     value through added information.  Whatever methods are used, they are least
 8     successful when the protection measure to be valued has little-understood, amorphous,
 9     or very long-term consequences.  In these cases it is difficult to articulate to individuals
10     the way in which the policy will affect society.
11
12     4.4  Costs of Environmental Protection
13
14           The task of estimating the costs of specific environmental protection efforts may
15     seem straightforward, compared with the conceptual problems and empirical difficulties
16     associated with estimating environmental protection benefits. To some degree, this is
17     true. In most cases, it is easier to develop cost estimates than benefit estimates to a
       commensurate degree of precision and reliability. But as we move towards developing
19     more precise and  reliable cost estimates, significant conceptual and empirical issues
20     arise.
21
22           The economist's notion of cost, or more precisely, "opportunity cost," is linked
23     with- but distinct from-everyday usage of the word, "cost."  In this part of the chapter,
24     we explore what this concept means in the context of environmental protection efforts,
25     and we briefly examine the ways in which it can be made operational through
26     quantitative, empirical analysis.
27
28           Conceptually, there are four steps required to appraise the cost of an
29     environmental-protection measure. First, we need to identify the specific  policy
30     instrument that is associated with the measure. This is because the same target, such
31     as a given reduction in ambient pollutant concentration, may be achieved at very
32     different total costs with different policy instruments. For example, it is well known that
33     under a variety of  circumstances,  a market-based policy instrument -  such as an
34     emission tax or a tradeable permit system - can enable a regulated sector to achieve
35     an ambient target  at relatively low aggregate cost, compared with a conventional

                                            4-17

-------
       1RP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote


 1     standard that has the effect of requiring all sources to adopt the same abatement
 2     technology (Baumol and Gates 1988). This example leads to the second conceptual
 3     step:  identifying the specific actions that sources will take to comply with the statute or
 4     regulation as implemented with the given policy instrument. Some of these actions may
 5     involve the adoption of a new piece of equipment, but others may involve a change in
 6     process. Third, it is necessary to identify the true cost of each action, which - as we
 7     emphasize below - requires much more than assessing required monetary outlays in
 8     an accounting sense. Fourth, it is necessary to aggregate these costs across society
 9     and over the relevant time frame.
10
11           Economists take cost to be an indication of what must be sacrificed in order to
12     obtain something (through purchase, exchange, or production). The concept of
13     opportunity cost provides a measure of the value of all of the things that must be
14     sacrificed if one or more risks are to be reduced. Opportunity costs typically do not
15     coincide with an accountant's measure of costs, namely monetary outlays.  This may
16     simply be because out-of-pocket costs do not capture all of the explicit and implicit
17     costs that have been incurred, such as the cost of time associated with waiting in line
18     for a vehicle inspection.  Or it may be  because the monetary prices of the resources
19     required to produce a good or service - in our case, environmental quality - may
20     themselves  provide inaccurate indications of the opportunity costs of those resources,
21     because of failures by markets that result in transaction prices not fully reflecting social
22     values. Likewise, it is important to distinguish between the private costs of some good
23     or service to producers or consumers of that good or service, and the social costs
24     imposed on society as a whole. The cost concept that is relevant for environmental
25     policy analysis refers to overall social  opportunity costs. Thus, the costs of
26     environmental protection are essentially the forgone social benefits  due to  employing
27     scarce resources for environmental protection purposes, instead of  putting these
28     resources to their next best use.
29
30            Thus, costs and benefits are two sides of the same coin. Environmental benefits
31     are created  by taking some environmental policy action, while other (presumably largely
32     non-environmental) benefits are thereby foregone. Hence, in keeping with the definition
33     of benefits established above, the cost of an environmental protection measure may be
34     defined as the (gross) decrease in consumer and producer surpluses1 associated with
35     the measure and with any price and/or income changes that may result (Cropper and

                                             4-18

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote


 1     Dates, 1992).
 2
 3           Environmental protection measures are intended to have desirable
 4     consequences, labeled "benefits," and are likely also to have undesirable
 5     consequences, labeled "costs," but there are many situations in which there may be
 6     some ambiguity regarding whether a specific regulatory consequence should be
 7     counted as an increased cost or a decreased benefit. Take, for example, an
 8     environmental regulation that provides direct protection of human health (benefits) and
 9     requires expenditures of capital and labor (costs) to achieve this purpose.  But the
10     positive health impacts of the regulation may have the effect of increasing labor
11     productivity, thereby decreasing the costs of producing other goods and services.  Is
12     this a decrease in the costs of the regulation or an increase in its benefits? To avoid
13     this sort of confusion, economists argue for the use of net benefits (benefits minus
14     costs), as opposed to benefit-cost ratios as an evaluative criteria (see below).
15
16           With these definitions in mind, we provide in Table 4-2 a view of the costs of
17     environmental protection measures, beginning with the most obvious and moving
       towards the least direct.2 First, many policy makers and much of the general public
19     would identify the on-budget costs to government of administering (monitoring and
20     enforcing) environmental laws and regulations as the cost of environmental regulation.
21     Most economic analysts, on the other hand, would identify the capital and operating
22     expenditures associated with regulatory compliance as a substantial portion of the
23     overall costs of regulation, although a considerable share of compliance costs for some
24     regulations fall on governments rather than private firms - a good example being the
25'    regulation of contaminants in drinking water, a Federal regulation, the cost of which is
26     borne primarily by municipal governments. Additional direct costs include legal and
27     other transaction costs, the effects of refocused management attention, and the
28     possibility of disrupted production.
29
30           Next, there are what have sometimes been called "negative costs" (in our
31     conceptual framework, non-environmental benefits) of environmental regulation,
32     including the beneficial productivity impacts of a cleaner environment and the potential
33     innovation-stimulating effects of regulation.3 "General equilibrium" or multi-market
34     effects associated with product substitution, discouraged investment4, and retarded
35     innovation constitute another important layer of costs5, as do the transition costs of

                                             4-19

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
 3
 4
 5

 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29.
30
31
32
33
34
35
36
37
38
39
                  TABLE 4-2. Costs of Environmental Regulation
   Government Administration of Environmental Statutes and Regulations
         Monitoring
          Enforcement
   Private Sector Compliance Expenditures
          Capital
          Operating
   Other Direct Costs
          Legal and Other Transactional
          Shifted Management Focus
          Disrupted Production
   General Equilibrium Effects
          Product Substitution
          Discouraged Investment
          Retarded Innovation
   Transition Costs
          Temporary Unemployment
          Obsolete Capital
   Social Impacts
          Loss and Change of Jobs for Some Workers
          Economic Security Impacts
Source: Jaffe et al., 1995

 real-world economies responding over time to regulatory changes. Finally, there are
other potential social impacts that are given substantial weight in political forums,
including those involving jobs and economic security.

      Within the category of direct compliance costs, business expenditures for
pollution abatement in the United States represent, on average, about 61  percent of
total direct abatement costs (Rutledge and Leonard, 1992); personal consumption
abatement, 11 percent; government abatement, 23 percent; government regulation and
monitoring, 2 percent; and research and development, 3 percent. These averages
must be taken with the appropriate grain of salt, since - as suggested earlier -
                                              4-20

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote


 1     measuring even direct costs is by no means a trivial undertaking. For example, there is
 2     significant variation between cost estimates produced by the U.S. Environmental
 3     Protection Agency (1990) and the U.S. Department of Commerce (1993) for these
 4     activities.
 5
 6           There are a number of potential problems of interpretation associated with any
 7     cost data.  For example, the questionnaire used by the U.S. Department of Commerce
 8     (1993) to collect data for its Pollution Abatement Costs and Expenditures (PACE)
 9     survey6 asked corporate and government officials how capital expenditures compare to
10     what they would have been in the absence of environmental regulations. This creates
11     two problems.7 The first involves the determination of an appropriate baseline. Absent
12     any regulation, firms might still engage in some - perhaps a great deal of - pollution
13     control to limit tort liability, stay on good terms with communities in which they are
14     located, maintain a good environmental image, etc. Should such expenditures be
15     included or excluded in the no-regulation baseline?
16
17           Second, when additional capital expenditures are made for end-of-the-pipe
       abatement equipment, respondents have relatively little difficulty in calculating these
19     expenditures. But when new capital equipment is installed, which has the effect of both
20     reducing emissions and improving the final product or enhancing the efficiency with
21     which it is produced, it is far more difficult to calculate how much of the expenditures
22     are attributable to environmental standards.8
23
24           In the give-and-take of environmental policy debates, it has frequently been the
25     case that abatement costs of proposed regulations have been over-estimated
26     (Hammitt,  1997). This may be due partly to the adversarial nature of the environmental
27     policy process, but it is also a natural consequence of employing short-term cost
28     analyses that do not take  into account potential, future cost savings due to
29     technological change, some of which may be endogenous to the regulatory regime.
30
31           As we said at the outset, the task of estimating the costs of environmental
32     protection efforts is relatively straightforward, at least compared with that of estimating
33     environmental protection benefits, but producing high-quality cost estimates still
34     requires careful analysis.
35

                                            4-21

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote


 1     4.5 Comparing Total Benefits and Total Costs
 2
 3           After the streams of benefits and costs associated with a pollution control
 4     program have been monetized, benefit-cost analysis involves evaluation according to
 5     the discounted sum of the net benefits it provides. This section discusses how net
 6     benefits are computed, and then aggregated over time. It also discusses in greater
 7     detail what it means for a program to "pass a benefit-cost test."
 8
 9           4.5.1 Calculating Net Benefits
10
11           The net benefits of a program in a given year (year T) are the difference
12     between the total benefits of the program in that year and its total costs in that year.  In
13     the case in which most of the costs of the program are incurred in the early years,
14     possibly in the form of capital investment, while the benefits of the program are spread
15     over time, net benefits may initially be negative, but eventually become positive. In a
16     case where the annual costs and benefits of the program are constant over time, net
17     benefits will also be constant.
18
19           Two caveats are in order. When net benefits are computed in each year, it is
20     customary to express them in constant dollars (in real terms). That is, any increases in
21     costs and benefits due to inflation are subtracted out, so that costs and benefits in each
22     year are expressed in, for example, 1997 dollars. A second  caveat concerns
23     comparing constant dollars at different points in time. When adding the dollar value of
24     net benefits over several years, it is important to realize that  people are not indifferent
25     between receiving a dollar in 1997 and a dollar in the year 2007.  The dollar received in
26     the year 1997, if invested at 5%, will grow to $(1.05)10 = $1.63, by the year 2007.
27     Equivalently, a dollar received in the year 2007 is worth only $1/(1.05)'° = 610 today,
28     since this is all one need invest today to obtain $1 in the year 2007.  For this reason,
29     economists express all future net benefits in terms of the present year's dollars; that is,
30     they compute the present value of net benefits in each year before aggregating net
31     benefits.
32
33           4.5.2 Present Values and Discounting
34
35            The present value of a dollar of net benefits occurring 10 years from today is the

                                             4-22

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote


 1     amount that must be invested today to yield one dollar in ten years. Generally, the
 2     present discounted value of $1 received in t years is $1/(1+r)', where V is the annual
 3     social rate of discount.
 4
 5           How should the annual social rate of discount (V) be determined?  Many
 6     economists would argue that T should represent the real rate of return on a riskless
 7     capital investment. By real return, we mean the return actually paid minus the rate of
 8     inflation. To illustrate, in 1980 the rate paid on a Treasury-Bill was about 11.6%, but the
 9     rate of inflation (as measured by the GDP deflator) averaged 9.4%. The real return on
10     government bonds was therefore 2.2%. Since we assume that benefits and costs will
11     be measured in real terms (adjusted for the rate of inflation), discount rates  should be
12     measured in real terms also.  The reason for using a riskless rate of return (such as the
13     rate of return on Treasury-Bills) is that the rate of return applied to public projects
14     should not include a risk premium. This is because public investments are financed by
15     all members of society, thus spreading the risk of the investment.
16
17           There is, however, another way to estimate the social discount rate.  This is to
       say that the costs and benefits in year T should be discounted to the present by a
19     consumption discount factor - the rate at which people are willing to trade consumption
20     in year T for consumption today.  In a world of perfect markets, the consumption
21     discount factor should equal the discount factor 1/(1+r)' where V is the return on capital
22     investment.  There are, however, cases where the two may differ, and this leads to
23     complications.
24
25           For simplicity, suppose that we measure the annual social rate of discount ("r")
26     by the real rate of return on riskless investments. In the U.S., this has ranged
27     historically from 0% to 4%. This suggests that the appropriate social discount rate lies
28     in this interval. Whether a rate of 0% or 4% is used, however, will often make a great
29     deal of difference to the magnitude of the present value of net benefits. A higher
30     discount rate, other things equal, will reduce the net present value of a program.  A
31     discount rate of 2% for example, implies that one dollar received in 20 years is worth
32     670 today, whereas it is worth only 460 at a discount rate of 4%.  Furthermore, the
33     discounted sum of the net benefits of a program will be more sensitive to changes in
34     the discount rate the longer the horizon over which benefits and costs extend and the
35     greater the disparity in the time profiles of benefits and costs.

                                            4-23

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote


 1           While the concept of discounting has a sound rationale, it can lead to
 2     conclusions that many people, including economists, find unpalatable. Even at a
 3     discount rate of 2% per annum, a dollar of benefits received 200 years from now counts
 4     for less than 20 today. This leads to the conclusion that the present value of net
 5     benefits will sometimes be negative for projects whose costs are concentrated in the
 6     present but whose benefits stretch far into the future.  To avoid rejecting such
 7     future-oriented programs, some people have called for not discounting future costs and
 8     benefits at all, especially for projects with long horizons. At first, this might seem to be
 9     a course of action that would favor future generations.  In an important sense, however,
10     it does not. If, by using a zero discount rate,  we adopt programs that do not pay off
11     until the distant future, we are implicitly passing up opportunities to invest in other
12     projects that yield higher rates of return. Since these projects may increase the capital
13     available to future generations, it is not clear that we have made them better off by
14     using a zero discount rate.
15
16            4.5.3 The Net Benefit Criterion and the Scope of Public Projects
17
18            Suppose that we have calculated the present value of the net benefits of an
19     environmental project in each year of the project's life. Adding these together will
20     produce the discounted sum of net benefits of the project.9  What can be  done with this
21     information? As stated at the beginning of this section, positive net benefits imply that,
22     at its current level, the project increases social welfare, in the sense that the gainers
23     could, in principle, reimburse the losers and still have  something left over for
24     themselves (see Section 4.5.I). This does not, however, mean that the size  of the
25     project is optimal. Figure 4-1 shows how the present value of the marginal benefits and
26     marginal costs of an environmental project vary with the scope of the project. In the
27     illustration, the scope of the project is the reduction in pollutant emissions from a given
28     baseline. The optimal reduction in emissions, according to the diagram, is "A" ~ the
29     point where the increase in benefits (or marginal benefits) from increasing the percent
30     reduction, just equal the additional cost of the reduction. Beyond this point, the
31     additional cost of further reductions exceeds the marginal benefits of the  reductions.
32
33            The total benefits of a given pollutant reduction, however, exceed the total costs
34     of the reduction at program levels that are much larger than optimal. The total benefits
35     of the reduction - the area under the marginal benefit curve - equal or exceed the total

                                              4-24

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
 8
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
                                                          Marginal
                                                          Cost
                                                          Marginal
                                                          Benefit
                                                        B
                 Reduction in Pollutant Emissions
       Figure 4-1. Marginal costs and marginal benefits associated with an
         environmental project.
costs of the reduction - the area under the marginal cost curve - for all reductions from
0 to "B". Thus, even programs that propose reductions greater than optimal still pass a
benefit-cost test.

      The lesson of Rgure 4-1 is that a program that passes a benefit-cost test at its
current level of implementation may still be suboptimal, in the sense that a similar
program at a different level of implementation could yield larger net benefits.  A
corollary to this result is that a program that consists of many components, for example,
a regulation that covers many sources, may, overall, pass the benefit-cost test, even
though individual components of the program (regulations on individual sources) may
                                            4-25

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote


 1     not pass the test.  In the context of IED, one objective of integrated options analysis is
 2     to improve net benefits by careful selection of risks to be addressed and risk reduction
 3     options to be pursued.
 4
 5           4.5.4 Uncertainty in the Measurement of Benefits and Costs
 6
 7           The state of the art of measurement of the benefit and costs of reducing risks is
 8     not sufficiently developed to produce exact measures of economic value.  This leads to
 9     the question, must policy makers wait for further research to produce exact measures
10     before they can use benefit and cost information as an aid  in decision making? If not,
11     how should they interpret the ranges of values that current  research has produced?
12
13           To counsel waiting for exact measures is equivalent to saying that in  many cases
14     value measures should never be used. The state of the art cannot be expected to
15     advance to the point of producing exact values for all kinds of environmental change.
16     This is because of the inherent uncertainty and imprecision in measurement
17     techniques, and because of uncertainties about which models are appropriate in
18     specific circumstances. So how are policy makers to proceed in the face of continued
19     and inherent uncertainty about economic values?
20
21            A simple approach with a basic intuitive appeal is to perform sets of calculations
22     with the upper and lower bounds of the range, and perhaps with the midpoint of the
23     range as well. This is the essence of the approach for valuing reductions in mortality
24     risks outlined in the EPA guidelines for performing regulatory impact analyses  (U.S.
25     EPA, 1983).
26
27            Clearly, if the benefits of a policy calculated with the upper end of the ranges are
28     less than the lower end of the range of estimated costs, the policy is unlikely to be
29     justifiable on economic grounds; and if the benefits calculated with the lower end of the
30     range exceed the upper end of the range of costs, the economic case for the policy is
31     quite strong. This simple-minded approach is a step in the right direction; but it does
32     not make use of all of the relevant information contained in the set of available
33     estimates.   This range reflects only the information contained in the two estimates
34     yielding the highest and lowest values; it ignores the information on the quality of these
35     two estimates; and it ignores the information contained in the other estimates that yield

                                             4-26

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote


 1     intermediate values.
 2
 3           It is possible to make use of ail available estimates and to incorporate judgments
 4     about the quality of each of these estimates. This can be done by viewing probabilities
 5     as statements about the degree of confidence held about the occurrence of some
 6     possible event.  The approach involves assigning probabilities to all of the values
 7     produced by the available estimates, where a higher probability reflects a greater
 8     degree of confidence in that estimate.  For example, the assignment of a probability of
 9     unity to a particular estimate means we are certain that this study has produced the
10     correct value. Once the probabilities have been assigned, various useful summary
11     statistics can be generated.  For example, the expected value of the parameter in
12     question (the probability distribution) can be calculated and used for benefit-cost
13     calculations. The variance of the distribution can be used to determine confidence
14     intervals on the  value to be used, thus preserving for policy makers  information on the
15     uncertainty about economic values.
16
17     4.6  Distributional Considerations
  i
19           This discussion of benefits and costs, as well as the way the  two are compared,
20     glosses over an important point - one that has been raised increasingly in discussions
21     about which risks the government ought to be addressing.  Specifically, BCA is silent
22     about the distributional implications of risk reduction measures. This has given rise to
23     concerns about the absence of environmental equity or environmental justice in the
24     decision making processes of regulatory agencies.
25-
26           To illustrate this point, suppose hypothetical^ that emissions from a series of
27     small auto paint shops in Los Angeles were resulting in slight increases in  ambient
28     ozone levels in Beverly Hills.  This in turn was slightly increasing the risks of eye
29     irritation among wealthy joggers there. Because the latter have very high incomes, the
30     calculated benefit from reducing this risk is very great - the joggers would  be willing to
31     pay substantial  amounts to be free of this annoyance.  Suppose further that closing the
32     paint shop, and in the process rendering unemployed  some of the unskilled people who
33     work there, is the only way to reduce the risk - control equipment could not be afforded.
34     In this case, the calculated benefits of the risk reduction could be well  in excess of the
35     costs, but because all the benefits go to the wealthy while the costs are borne by the

                                            4-27

-------
        IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote


 1      poor, one might be quite uncomfortable with enacting the controls.
 2
 3            In other situations, the situation is reversed. That is, the physical benefits of a
 4      risk reduction program would go to a poor population - say, those who live around a
 5      Superfund site - while the costs of the cleanup would be shared much more broadly
 6      and, in general, by those with higher incomes.  But because the willingness to pay of
 7      those affected by the site is quite low, the benefits as calculated in a BCA would be
 8      small relative to the costs. Thus, such a cleanup might fail a benefit-cost test even
 9      though it would have what many would regard as favorable distributional
10      consequences. Simply put, BCA counts a dollar's worth of benefit or cost to a wealthy
11      person the same as a dollar's worth of benefit to a pauper. This can result in some
12      efficient programs (programs for which the net benefits are positive) having unattractive
13      features. Economists have long acknowledged that efficiency and equity are not
14      always compatible objectives.
15
16            In the introductory section, we mentioned this concern with BCA, and suggested
17      that much thought has been given over the years to the incorporation of distributional
18      weights for BCA that would attempt to incorporate such considerations into
19      determinations of efficiency. We believe this is a poor idea, not because we are
20      unconcerned with the distribution of income, but rather because we see no way to
21      derive any kind of consensus on what the weights should be.  It seems better to
22      estimate the benefits and costs as best we can, and also provide as much information
23      as possible about who will gain and who will lose, and let decision makers take
24      distributional considerations into account in whatever way they see fit.
25
26      4.7 Conclusions
27
28            What are we to conclude from this discussion? Several points bear repeating.
29      First, and perhaps most importantly, knowing the risks that arise from current and future
30      exposures to pollutants in the environment is certainly important, but it is an insufficient
31      basis, on its own, for planning future regulatory or other forms of action. That is, it does
32      not naturally follow that the best thing to do once one knows which risks pose the
33      biggest threats to human health or the environment is to take actions to reduce the
34      biggest risks first. While perhaps superficially counterintuitive, it may make much more
35      sense to attack problems farther down on the list.

                                             4-28

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote


 1
 2           This is because a more sensible basis for action is asking which problems most
 3     lend themselves to amelioration.  That is, we ought to ask where the resources we have
 4     available for environmental improvement (be they financial, human, or other) can best
 5     be deployed. This in turn requires that we know how much risk reduction we can get
 6     and for how much money in each of the risk categories. This is no more than the
 7     environmental application of "the biggest bang for the buck" aphorism that originated
 8     when analytical thinking was applied to defense spending.
 9
10           Although economics can provide only parts of the answer to the question, "How
11     can we best spend our environmental dollars?", those are important parts nonetheless.
12     Economics can tell us, for instance, how much it would cost to attempt to reduce
13     average ambient ozone concentrations from 0.10 parts per million to 0.08 parts per
14     million. It can tell us how much it might cost to reduce heavy metal deposition into
15     aquatic ecosystems by 30 percent. And it can tell us what we could expect to spend to
16     reduce pesticide residues on fresh fruits and vegetables.
17
             Economics can do more than inform us about the costs of various types of
19     control actions.  It can also illuminate  the value that individuals place on these and
20     other types of risk reductions, and how these values compare to the costs of taking
21     actions. This latter role of economics-benefit-cost analysis-is controversial, poorly
22     understood, and often misused. This is no reason, however, to throw the baby out with
23     the proverbial bathwater. If carefully used,  with its basic assumptions clearly and
24     openly described, benefit-cost analysis can be an invaluable aid to informed decision
25     making.
26
27           If economics is used carefully,  it can contribute significantly to the nation's
28     environmental protection efforts. Specifically, along with quantitative risk assessment
29     and other analytical tools, it can help us identify places where we can accomplish a lot
30     for a little, and avoid spending a great deal to do  relatively little good. It is harder
31     making decisions in cases where both the benefits and the costs seem large, or those
32     cases where both are on the small side. Here, finer distinctions must be drawn. This
33     means in turn that the assumptions that underlie  a benefit-cost assessment may make
34     the difference between a favorable and an  unfavorable outcome. So long as those
35     assumptions are made explicit, however, and so  long as distributional and other

                                            4-29

-------
      IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote


1     considerations are also factored into decision making, BCA can be a most valuable
2     input into the formulation of an appropriate environmental policy.
3
4
                                           4-30

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote


 1     4.8 References Cited
 2
 3     Arrow, K., M.L Cropper, G.C. Eads, R.W. Harm, L.B. Lave, R.G.Noll, P.R. Portney, M.
 4           Russell, R. Schmalensee, V.K. Smith, and R.N. Stavins. 1996. Is there a role
 5           for benefit-cost analysis in environmental, health, and safety regulation?
 6           Science. 272:221-222.
 7
 8     Baumoi, W.J. and W.E. Gates. 1988. The Theory of Environmental Policy. Second
 9           Edition. Cambridge, United Kingdom: Cambridge University Press.
10
11     Cropper, M.L and W.E. Gates. 1992. Environmental Economics: A Survey. Journal
12           of Economic Literature 30:675-740.
13
14     Freeman, A. M., III. 1993.  The Measurement of Environmental and Resource Values:
15           Theory and methods. Resources For the Future. Washington, DC. 516 pp.
16
17     Freeman, A. M., III.  1997.  On Valuing the Services and Functions of Ecosystems.
             Chapter 11 in Simpson, R. D. and N.L. Christensen, Jr. (Eds), Ecosystem
19           Function and Human Activities: Reconciling Economics and Ecology, New York.
20           Chapman and Hall.
21
22     Hammitt, J.  1997.  Are the Costs of Proposed Environmental Regulations
23           Overestimated? Evidence from the CFC Phaseout. Working Paper, Harvard
24           Center for Risky Analysis, January.
25
26     Jaffe, A.B., S.R. Peterson, P.R. Portney, and R.N. Stavins. 1995. Environmental
27           Regulation and the Competitiveness of U.S. Manufacturing:  What Does the
28           Evidence Tells Us?  Journal of Economic Literature 33:132-163.
29
30     Palmer, K., W.E. Gates, and P.R. Portney.  1995. Tightening Environmental
31           Standards: The Benefit-Cost or the No-Cost Paradigm? Journal of Economic
32           Perspectives 9:Number 4, pp. 119-132.
33
34     Porter, M.E. and C. van der Linde.  1995. Toward a New Conception of the
35           Environment-Competitiveness Relationship. Journal of Economic Perspectives

                                           4-31

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote


 1           9:Number4, pp. 97-118.
 2
 3     Rutledge, G.L. and M.L. Leonard.  1992. Pollution Abatement and Control
 4           Expenditures, 1972-1990. Survey of Current Business, June. Pp. 25-41.
 5
 6     Schmalensee, R.  1994. The Costs of Environmental Protection, jn Balancing
 7           Economic Growth and Environmental Goals. Mary Beth Kotowski (Ed).
 8           Washington, D.C.: American Council for Capital Formation Center for Policy
 9           Research. Pp. 55-75.
10
11     U.S. Congressional Budget Office. 1992. Environmental Regulation and Economic
12           Efficiency. Washington, D.C.: U.S. Government Printing Office.
13
14     U.S. Department of Commerce. 1993.  Pollution Abatement Costs and Expenditures,
15           1991. Economics and Statistics Administration, Bureau of the Census.
16           Washington, D.C.: U.S. Government Printing Office.
17
18     U.S. Environmental Protection Agency. 1983.  Guidelines for performing regulatory
19           impact analyses. Washington, DC.
20
21     U.S. Environmental Protection Agency. 1990. Environmentallnvestments: The Cost
22           of a Clean Environment.  Washington, D.C.
23
24     von Winterfeldt, D. and W. Edwards. 1986. Decision analysis and behavioral research.
25           Cambridge University Press.
26
27
                                           4-32

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
 1
 2     Endnotes
 3
 4     'See definitions of consumer surplus and producer surplus, above, in our discussion of
 5     environmental benefits.
 6
 7     2For a useful decomposition and analysis of the full costs of environmental regulation, see:
 8     Schmalensee (1994).
 9
10     'The notion that environmental regulation can foster economic growth is a controversial one
11     among economists. For a debate about this proposition, see:  Porter and van der Linde, 1995;
12     and Palmer et aL, 1995.
13
14     "For example, if a firm chooses to close a plant because of a new regulation (rather than
15     installing expensive control equipment), this would be counted as zero cost in narrow
16     compliance-cost estimates, but it is obviously a real cost.
17
18     5Note that it is extremely difficult to measure "retarded innovation," because the degree of
19     innovation that would have occurred in the absence of the environmental-protection measure is
20     essentially unobservable.
21
00     6For over twenty years, the PACE survey provided a unique source of information on pollution
       control costs, since it was the only annual, representative survey to cover all of U.S.
&4     manufacturing.  In  1996, the collection of data was terminated for budgetary reasons. Given
25     that the annual costs of the survey were approximately $1 million, this decision itself ought to be
26     questioned on benefit-cost grounds. In April 1999, EPA announced that it would reinstitute the
27     annual PACE survey with the Bureau of the Census.
28
29     7For detailed discussion of environmental compliance cost measurement problems, see U.S.
30     Congressional Budget Office (1985).
31
32,    8An irony of the movement of environmental policy towards more reliance on performance
33     standards (both uniform and market-based) and less on end-of-pipe technological standards is
34     that it is becoming  increasingly difficult to measure even the narrowest notion of pollution
35     abatement costs (Jaffe et al., 1995).
36
37     9 Formally, if bt denotes benefits in year t and ct costs, the discounted sum of net benefits over
38     a horizon of T years is given by:
39
40            T
41            I (bt-ct)/(1+r)t
42            t=0
                                                4-33

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote


 1       CHAPTER 5.  ASSESSING THE VALUE OF NATURAL RESOURCES
 2
 3                             TABLE OF CONTENTS
 4
 5
 6     OVERVIEW	5-1
 7
 8     5.1  Introduction	5-4
 9          5.1.1  Background	5-4
10          5.1.2  Objectives and Approach	5-5
11
12     5.2  Valuation and the Decision Context	5-6
13          5.2.1  The Valuation Process in Regulation	5-7
14          5.2.2  The Decision Context	5-9
15
16     5.3  The Nature of Values	5-14
            5.3.1  Introduction  	5-14
            5.3.2  Values	5-15
19
20     5.4  The Economic Valuation Framework	5-20
21          5.4.1  The Concept of Economic Value	5-20
22          5.4.2  Economic Value and Benefit-Cost Analysis	5-21
23          5.4.3  Economic Valuation of the Functions and Services of Environmental
24                Systems  	5-22
25          5.4.4  Issues and Problems	5-23
26          5.4.5  Conclusions	5-25
27
28     5.5  The Importance of Deliberative Processes to Valuation	5-26
29          5.5.1  Introduction  	5-26
30          5.5.2  Aspects and Recommendations 	5-28
31
32     5.6  Additional Approaches to Valuation of Environmental Systems	5-37
33          5i6.1  Introduction  	5-37
34          5.6.2  Findings	5-37
"<;
       5.7  Summary and Conclusions	5-47

-------
      IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote


1
2     5.8  References Cited	5-52
3
4

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3, 1999-Do Not Cite or Quote


 1        CHAPTER 5. ASSESSING THE VALUE OF NATURAL RESOURCES
 2
 3
 4     OVERVIEW
 5
 6          In its 1990 report, Reducing Risk: Setting Priorities and Strategies for
 7     Environmental Protection, the SAB recommended that the Agency "develop improved
 8     analytical methods to value natural resources and to account for long-term
 9     environmental effects in its economic analyses" because "traditional methods of
10     economic analysis tend to undervalue ecological resources and fail to treat adequately
1 1     questions of intergenerational equity." During 1996, Deputy Administrator Fred Hansen
12     urged the SAB to address the need for improved methods of measuring environmental
13     benefits.
14
15          The Valuation Subcommittee was established as a multidisciplinary group to act
16     on this request. The specific charge was for the  Subcommittee to define better the full
       range of relevant questions that must be considered in ecological and human health
•i w     valuation. It was further to identify conditions where existing economic methodologies
19     seem not to address or monetize adequately ecological endpoints that  may be
20     important contributors to the value society places on ecological resources. The
21     Subcommittee was challenged to use its broad diversity of expertise and perspective to
22     scope out a more complete framework for valuation and to identify the types of
23     methodological developments or research needed to implement the framework.
24     Although it was requested to do so in the charge, the Subcommittee was not able fully
25     to consider valuation of human health issues, and the Subcommittee's conclusions may
26     not apply directly to human health.
27
28          The group discussed the charge and worked on its response in public meetings
29     on two separate occasions. Also, in a three-day workshop during April, 1997 the
30     concept of environmental valuation was discussed from five perspectives: 1) the
31     environmental management decision context for valuation; 2) the nature of value and
32     values; 3) the economic concept of value; 4) the importance of deliberative processes
33     to valuation; and 5) additional approaches to environmental valuation.
34
                                           5-1

-------
        IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1            Many stakeholders feel that existing economic analysis methods undervalue
 2     ecological resources because they miss important issues. The Subcommittee agrees
 3     that valuation of ecological services is a complex process and a still-evolving field of
 4     research and that improvements in both approach and practice are needed. This report
 5     discusses the components of the process that should be considered in valuation efforts.
 6     It concludes that there is no single approach to valuation that can be used in all
 7     situations. The Subcommittee's work confirms that an interdisciplinary perspective is
 8     required to meet the challenges and complexities of valuation.
 9
10            One of the basic premises of welfare economics is that economic values are
11     based on individuals' preferences and that people know their preferences.  Where
12     individuals are ignorant of the roles of ecological functions in contributing to valued
13     service flows, it may be necessary to use experts' knowledge of the functioning  of
14     environmental systems as an input in the valuation process. In principle, the economic
15     valuation framework can be utilized to define and measure the economic values of
16     changes in the functions and services of environmental systems that affect individuals'
17     welfare either directly or indirectly. However, this framework may be difficult to
18     implement in practice where the relationship between the function and the service flow
19     to individuals is indirect or subtle.  Economic approaches to valuation are not
20     mechanisms for producing "the answer" since they may be incomplete, may include
21     some elements which are difficult  or impossible to estimate, and may employ
22     preference elicitation processes that are incomplete.
23
24            Not all benefits or costs can be easily quantified, much less translated into dollar
25     terms. There is a need for qualitative methods to help decision makers understand the
26     hard-to-define values that are important in finding solutions to complex ecosystem
27     problems. When integrating the results from different methods, care must be taken to
28     assure that quantitative factors do not dominate important qualitative factors.  To be
29     most useful, valuation issues and  approaches should be made as explicit as possible
30     and should involve assembling the appropriate available information, clearly stating the
31     assumptions and uncertainties, and ensuring that the application of methods is
32     transparent. These analyses are best used to inform but not dictate, decisions  related
33     to environmental protection policies, programs, and research.
34
35

                                              5-2

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1          There is much uncertainty in our knowledge and understanding of many of the
 2     factors necessary in the environmental decision-making process.  Uncertainty will
 3     persist, yet we cannot wait for certainty in our analytical inputs to make decisions. The
 4     process of adaptive management, discussed in this report, is attractive because it
 5     allows decisions to be made and management to proceed in the face of uncertainty.  In
 6     doing so, it allows for feedback from experience gained during implementation of
 7     environmental management practices, the development of new knowledge by the
 8     research community, and for revisiting and revising past decisions on the basis of the
 9     new insights provided by this research and experience.
10
11           General themes that emerged during the Subcommittee's discussions included:
i
1            1)  For decision-making purposes in a governmental context, ecological valuation
1            is an anthropocentric exercise (people's wishes count; there is no external set of
1            values watting to be discovered for application to decision-making).
1 .
1 '           2)  The value of anything reflects its contribution toward the achievement of
             some goal.  The process of valuation cannot be separated from the need to
j i           reach agreement on goals.
J i
             3)  Environmental valuation requires a diverse and interdisciplinary process
             involving interaction and deliberation among scientists, decision makers, and
  !           other stakeholders to identify goals and to define endpoints to characterize those
             goals.

  i           4)  Existing economic approaches, broadly considered, are consistent and
             coherent frameworks for valuation because they organize a system of trade-offs.
  I           However, they are not mechanisms for producing "the answer" because they
  }           may omit trans-economic values that may be important, may include some
  )           elements that are difficult or impossible to estimate, and may employ preference
  I           elicitation processes that are incomplete.

  3           5)  An expanded, rich, and complex process using multiple approaches is
  I           required to fully encompass ecological  valuation.
                                            5-3

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1           The Subcommittee's work confirms the challenges and complexities of
 2     environmental valuation exercises, and the resultant environmental management
 3     actions based on those values. The Subcommittee recommends that expanded, rich,
 4     and complex processes be employed to fully characterize environmental values. This
 5     process will involve interaction and deliberation among scientists, decision makers, and
 6     other stakeholders in order to identify goals, define endpoints to characterize those
 7     goals, and to implement approaches to achieve those goals. This process must be one
 8     of on-going dialogue and adjustment and it should consider 1) why people care for the
 9     things they do-their preferences; 2) the appropriate use of deliberative processes to
10     elicit preferences and the rationale for them; 3) economic valuation frameworks to
11     define and measure the economic value of changes in environmental systems functions
12     and services affecting individual welfare; and 4) presentation of available physical or
13     other quantitative measures, or qualitative descriptions of the effects of alternative
14     actions when costs and benefits are not fully captured by monetary measures. The
15     Subcommittee recognizes that environmental valuation remains a craft embedded in
16     political processes.
17
18     5.1  Introduction
19
20           5.1.1  Background
21
22           In its  1990 report, Reducing Risk: Setting Priorities and Strategies for
23     Environmental Protection, the U.S. Environmental Protection Agency's (EPA) Science
24     Advisory Board (SAB) recommended that the Agency "develop improved analytical
25     methods to value natural resources and to account for long-term  environmental effects
26     in its economic analyses." This recommendation was based on the view that
27     "traditional methods of economic analysis tend to undervalue ecological resources and
28     fail to treat adequately questions of intergenerational equity." In February 1996, Deputy
29     Administrator Fred Hansen met with the SAB Executive Committee and urged the SAB
30     to address the need for improved methods of measuring environmental benefits as part
31     of its project to update Reducing Risk.
32
33           In response, the Executive Committee agreed to establish a Valuation
34     Subcommittee as part of the Integrated Risk Project.  The Valuation Subcommittee's
35     deliberations were intended to complement those of the Economic Analysis

                                             5-4

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1     Subcommittee so that the final Integrated Risk Project report would include information
 2     relevant to the economic assessment of the benefits and costs of various risk reduction
 3     options, as well as recommendations for improving future assessments of the value of
 4     ecological resources to society.
 5
 6           5.1.2  Objectives and Approach
 7
 8           The Valuation Subcommittee was charged to define better the full range of
 9     relevant questions that must be considered in ecological and human health valuation.
10     The prevailing feeling was that existing economic methodologies, which are used to
11     quantify the monetary value of certain ecological resources, do not address or monetize
12     adequately important ecological endpoints (ecological processes, environmental
13     structures, and many other characteristics) that may be important contributors to the
14     value society places on ecological resources (e.g., intergenerational issues,
15     stewardship  issues, sustainability options, and time scale issues). Similarly, many
16     environmentally mediated quality of life issues of great importance to people  and
1**     society are difficult to quantify (e.g., aesthetic, culture, religion, security, equity).
       Consequently, a new look at the overall approach to valuation was requested: one that
19     considered the utility of current economic methodologies as a point-of-departure, but
20     which went beyond these and identified other approaches to supplement them and
21     provide a more comprehensive basis for valuing ecological systems and their
22     relationships to human health and quality of life. The Subcommittee was challenged to
23     use its broad diversity of expertise and perspective to scope out a more complete
24     framework for valuation and to identify the types of methodological developments or
25     research  needed to implement the framework.
26
27           In  order to meet this charge, a diverse group was assembled with individuals
28     from disciplines such as public policy, economics, ecology, philosophy,
29     communications, psychology and sociology. Institutional affiliations represented in the
30     group were also diverse; for example, members came from academia, industry,
31     independent research institutions, and consulting firms.  Some members were formerly
32     associated with government organizations.
33
34           The Subcommittee discussed the charge and  planned for its response in public
35     meetings on two separate occasions. These meetings led to a three-day workshop in

                                             5-5

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1     Baltimore, Maryland during April, 1997. At this workshop, the members were assigned
 2     to one of five panels and led Subcommittee discussions about valuation from five
 3     perspectives. The discussions and recommendations resulting from these activities are
 4     the basis for this chapter.
 5
 6           The Subcommittee discussions fell into the following broad areas: 1) the
 7     environmental management decision context for valuation; 2) the nature of value and
 8     values; 3) the economic concept of value; 4) the importance of deliberative processes
 9     for eliciting values and the goals on which they are predicated; and 5) additional
10     approaches to environmental valuation. Although requested to do so in the charge, the
11     Subcommittee was not able to consider valuation of human health issues.
12     Consequently the Subcommittee's conclusions may not apply directly to human health.
13
14           The focus of Subcommittee activity was the consideration of alternatives and
15     enhancements to economic analysis techniques that would allow the Agency to
16     characterize the benefits and costs of environmental protection actions more
17     comprehensively and accurately and to ensure that environmental decision-making
18     takes account of the values that people attach to ecological systems, and system
19     functions and components.
20
21     5.2 Valuation and the Decision Context
22
23           The need to conduct environmental valuation exercises arises because people,
24     as they exercise their prerogative of free choice, sometimes carry out activities that
25'    result in changes to environmental systems that can be detrimental.  For some of these
26     situations, society has established laws to control the behavior that leads to the
27     detrimental effects.  Government agents are entrusted with the responsibility and
28     authority to evaluate the  detrimental changes that meet or exceed triggering-criteria
29     contained in these laws,  and to decide whether action is needed to prevent or control
30     these changes. These decisions require,  in turn, a basis for comparing alternative
31     outcomes, and this leads to the need to place a value on the ecological changes
32     envisioned.
33
34
35

                                             5-6

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1           5.2.1 The Valuation Process in Regulation
 2
 3           When a statutory criterion is triggered or a problem is otherwise identified, a
 4     chain of activity is initiated that results in a decision on the need for, and if appropriate,
 5     the type of intervention. As discussed in previous chapters, the I ED framework
 6     provides a way of considering information (both data and professional judgments)  on
 7     risks, control options and practices, economic impacts, and social and quality of life
 8     issues.  Equally important is the consideration of the value ascribed to alternative
 9     outcomes.  Methods to estimate such values are a critical component of the IED
10     framework.
11
12           As discussed in Chapter 4, value is routinely assessed during the environmental
13     management decision-making process  using economic analysis techniques. The  usual
14     approach is to apply economic analyses to estimate the costs and benefits  of each of a
15     suite of potential  environmental management options. The costs and benefits are then
16     compared to determine the net benefit associated with each option, and this
117     information, along with other decision criteria specified by the decision-maker, serves
       as the basis for deciding whether to take action, and if so, on the most suitable action to
19     implement.
20
21           Figure 5-1 reflects a prototypical relationship between increasing degrees of
22     environmental stressor control and the  societal costs and benefits that might be
23     associated with a specific control option. The relationship demonstrated in the figure
24     suggests that at first, toward the origin, the initial marginal costs of control are minimal
25     but gradually increase because a higher degree of effort is needed to bring about
26     increasing reduction in the stressor. At the same time, the associated marginal benefits
27     are high with the initial reduction in the  stressor but decline with additional reduction.
28     The key point shown by the figure is that typical economic analysis techniques, which
29     compare costs and benefits of control options (usually in dollar terms) reveal a point in
30     the stressor reduction process at which the associated costs of additional control  are
31     not "worth it" because the benefits gained are less than those costs. Further reductions
32     in the stressor will bring benefits, but these are won only at costs (expenditures of
33     resources with alternative uses) that on net, in the judgment of the party doing the
34     evaluation, make the overall outcome worse. In a public policy context, the government
35     decision-maker stands as the agent for the people as a whole  and acts on  the

                                             5-7

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
         ($) A
         D
         o
          I
          I
         a
         r
         s
                                 Percent Reduction  in Stressor
100%
         Figure 5-1. Typical relationship of marginal costs and marginal benefits associated
                     with a proposed environmental control option.

 1'    valuations that they place on stressor reduction and on the associated use of resources
 2     to accomplish it.  An example is the distribution of the heavy metal, lead in the
 3     terrestrial environment at a hazardous waste site. Initially high concentrations of lead
 4     pose significant environmental risks which can be sharply reduced during a site clean
 5     up program. These initial reductions in lead concentrations can be achieved at
 6     relatively low costs; however, a point of excess control can be reached where the costs
 7     of additional lead removal exceed the marginal benefits and associated  reduction in
 8     risks realized.
 9
10           While accepting the stylized approach to trade-offs embodied in this diagram, the
11     Subcommittee noted that its effectiveness in practice as a guide to action depended on
                                             5-8

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1     the completeness and correctness of the elements included in the calculation of the
 2     prototypical curves.  As illustrated in the discussion below, with respect to
 3     completeness, the Subcommittee found that in practice some elements valued by
 4     people could be systematically omitted, others could be poorly quantified, and others
 5     might evolve only in  a social context and not surface from commonly used individual
 6     preference measures. With respect to correctness, the Subcommittee noted that
 7     selection of the elements to be valued was determined by the social goals presumed to
 8     be relevant, and that those goals were a complex mixture that included matters
 9     sometimes overlooked such as fairness, long-term sustainability, and stewardship.
10     Further, the Subcommittee noted that in practice, estimation of the values that
11     individuals placed on these elements is subject to error, and that aggregation of such
12     values involves methodological uncertainties. These considerations formed the basis
13     for Subcommittee discussions of enhancements to economic analysis techniques to
14     allow a more comprehensive and accurate characterization of environmental values.
15     The results of the discussions are captured in the sections that follow.
16
1"7            5.2.2 The Decision Context

19            Many environmental regulations have been positively received by the public and
20     these actions have produced noticeable improvements in the quality of our air, land,
21     and waters. In addition, the regulation of certain chemical products and residuals have
22     brought improvements in-human health and the state of certain wild species. However,
23     some environmental decisions have met with concern and conflict and have led to an
24     erosion of trust in governmental institutions. Common criticisms of environmental
25     decisions are that they are yielding inadequate returns for the resources used, often
26     focus on perceived,  rather than actual, risks, and define problems too narrowly  (e.g.,
27     see Sexton, 1996).  In addition, there is a growing sense that decision-making should
28     be more open, allowing greater understanding and participation by stakeholders and
29     the public.
30
31            In considering the source of controversy in environmental management decision-
32     making, Brown (1997) suggested that "...disputes about environmental policy, while
33     often seeming to be about the facts, are-at least as much-about an underlying value
34     framework that legitimates a particular decision." Dietz et al. (1989) noted that
35     participants in risk policy debates differ in their perceptions about the role of facts and

                                             5-9

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1     values in such conflicts.  The norms and goals of society, in effect, provide the
 2     dominant contextual grounding for environmental decision-making.
 3
 4           The values that people place on certain health and environmental states,
 5     conceptions of their rights and obligations, and outcomes associated with
 6     environmental perturbations and responses, reflect their norms and social objectives.
 7     The values underlying these debates and decision-making processes are not always
 8     clear. These values may be complex and in competition with other values, including not
 9     only the values of other people, but even values held by the same person. Because
10     value frameworks are so diverse and complex, there is a need to refine methodologies
11     for eliciting and comparing values quantitatively and qualitatively, both interpersonally
12     and individually.
13
14           "Values" as used in the two preceding paragraphs is a term which describes the
15     set of underlying factors that, taken together, cause people to hold the opinions that
16     they hold and to make the choices that they make when presented with real situations.
17     Unfortunately, the English language also uses the term "value" and its derivatives in an
18     operational sense as a descriptor of the  "worth" of outcomes, measured in terms of
19     what would be willingly sacrificed in exchange.  Thus the confusion in terms between a
20     person "valuing" health in the abstract, the first definition of the term, and a person
21     placing an inferred 'Value" of so many dollars, but no more, on avoiding a day of limited
22     activity and discomfort due to a cold.  The exchange of economic "value" is therefore
23     derived from the vector of ail the underlying abstract "values" held by the individual in
24     concert with the situation presented.  When it comes to making environmental (and
25     other) decisions, "value" is used in an operational sense as a measure of what one
26     outcome is worth in comparison with alternatives. This is the sense in which values are
27     reflected in, for example, benefit/cost analysis.  It is with this meaning that "value" is
28     used  in most appearances in this Chapter. Other uses of the term also occur, however,
29     and are made clear by the context.
30
31           The Valuation Subcommittee discussed the issues of scientific  uncertainty,
32     competing values, and the completeness of individual valuation approaches. It
33     concluded that all of the approaches used to develop summary information on risks,
34     costs, and benefits for decision-makers reflect to some degree the disciplinary
35     background and values held by those individuals who conduct the assessments.  This

                                             5-10

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1     highlights the importance of knowing who participates in data collection and evaluation
 2     and the methods they use to do so. It further suggests the importance of making sure
 3     that individuals and groups having a stake in the outcome of decision-making
 4     participate in the process of describing and valuing changes in environmental systems,
 5     and in estimating the costs and benefits of intervention to control or redress these
 6     changes. Some form of interaction and deliberation among agencies and among those
 7     who hold direct or indirect interests in decisions is often recommended as a way of
 8     identifying and addressing the disparate values held by individual stakeholders (NRC,
 9     1996; Risk Commission Report, 1997).  The Subcommittee has endorsed this
10     recommendation.
11
12           Valuation and value judgments are important in all phases of the IED framework
13     (see Figure 5-2). Initial definition of values, and the  goals to which they give rise, is
14     required in the early problem formulation phase to help define the scope of the
15     problem, its context, and the breadth  of evaluation required in the analysis phase.
16     Value judgments are also made during the analysis  and characterization of risks.
17     During the Analysis and Decision-making Phase, valuation input is again needed as
       alternative management or control options (e.g., technologies and practices) are
19     evaluated to help identify and measure the relative costs and benefits of different
20     degrees of risk reduction and of alternative ways of accomplishing it.
21
22           Values considered during environmental management decision-making are
23     those associated with changes to both human dominated (e.g., farms, forests) and
24     natural environmental systems. Substantial variation exists within natural ecosystems
25     and to be socially significant, a change caused by a stressor must exceed the natural
26     variability of the environmental system and involve impacts affecting significant social
27     values. However, since no precise standards exist for measuring stressor reduction
28     costs and associated societal benefits at the system level, it is inevitable that the
29     decision on which values to weigh most heavily will require deliberation and negotiation.
30     Therefore, it is clear that the articulation of these values, and the resultant regulatory
31     actions based on them, must necessarily involve deliberation including scientists, the
32     public, and policy makers in a process of on-going dialogue  and adjustment.
33
34
35

                                            5-11

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
2F
26
27
28
29
30
31
32
33
34
35
                        Environmental Valuation
                        Environmental Valuation
       Figure 5-2. Valuation in the IED Process

      Although current valuation techniques can deal satisfactorily with some
environmental system changes (e.g., the impact of air pollution on managed agricultural
systems), they may deal inadequately or not at all with others.  When one thinks of
considering the benefits of environmental systems, it is important to recognize that
some things can be quantified and some not.  Of those that can be quantified, some
can be monetized and some not. For example, it is possible to quantify concern with
biodiversity without being able to monetize it, or otherwise express it in units that can be
compared to, say, changes in unemployment. Therefore, environmental system
                                         5-12

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1     valuation efforts must include both quantitative and qualitative perspectives in an
 2     integrated approach to capture all components of things valued.
 3
 4           Quantitative methods include those approaches designed to measure human
 5     attitudes, preferences (i.e., choices when presented with alternatives, desires,
 6     operational proximate goals) and values. As discussed in the previous chapter, these
 7     methods include: 1) direct measures of preferences and values using survey methods;
 8     2) inferring preferences and values from observed behavior and choices; 3)  inferring
 9     preferences and values from choice experiments; and 4) using estimates of  values of
10     similar commodities that were derived from using one of the above methods, or what
11     economists refer to as benefits transfer.
12
13           Since not all environmental costs and benefits can be easily quantified for use in
14     benefit-cost analyses, there is a need for qualitative methods to help decision-makers
15     understand the hard-to-define values that are important in finding solutions to complex
16     environmental system problems. Qualitative methods must involve at least the detailed
17     narrative description of effects in complete terms such that comparisons with other
1      ecological attributes can be made.  Although this qualitative information cannot be
19     directly compared to quantitative information in common terms, qualitative information
20     should be clearly described and it should be readily available to assist during decision-
21     making (e.g., see Florida Risk-Based Priority Council, 1996).
22
23           This information is used to guide decisions. At times, the path forward seems
24     clear and uncomplicated, and decisions are straight forward.  At other times, however,
25     uncertainties exist — in how the interventions chosen will work, how the systems to be
26     altered will react, and in how the outcomes will be valued. The process of adaptive
27     management has developed as an innovative technique to assist in environmental
28     decision-making in the face of such uncertainty.  Rather than waiting until complete
29     data are available that allow participants to understand all aspects of the at-risk system,
30     adaptive management allows action to proceed on the basis of the best understanding
31     initially available. The response to action is monitored, and the information  gained is
32     used to modify or design the next stage in the program. In adaptive management there
33     is direct feedback between science and management such that policy decisions can
34     make use of the best available scientific information in a sequential fashion. This idea
35     has been well-developed in  many contexts for nearly three decades (Campbell, 1969)

                                            5-13

-------
        IRP Draft Integrated Report-Peer Review Draft, May 3, 1999-Do Not Cite or Quote

 1      and has been recently articulated around environmental problems by Lee (1993). As
 2      the IED framework implies, valuation is not a one-time activity, it is an iterative part of
 3      the risk management process responding to the development of new information over
 4      time. That is, changes in the perception of the value of actions toward reaching goals is
 5      part of adaptive management as well.
 8
 9            In the following sections we discuss four topics relevant to environmental
1 0     valuation: 1 ) the nature of values; 2) the traditional economic valuation framework; 3)
1 1     the importance of the deliberative process to valuation; and 4) alternatives for the
12     determination of environmental system values. Each of these sections includes a
1 3     discussion and provides some recommendations for the consideration of valuation
14     issues during environmental decision-making.
15
1 6     5.3 The Nature of Values
17
18            5.3.1 Introduction
19
20            Controversies surrounding environmental decisions are often expressed in terms
21     of disagreements about values. These controversies are difficult to resolve in part
22     because the word "values" is used in different ways and with different meanings.
23     "Values" can refer to different goals of environmental decisions or to different means for
24     achieving those goals.  But "values" can also be used to express concerns that are not
25     about goals or outcomes at all. Such concerns can be about the appropriateness of
26     different procedures for reaching decisions, the expressive functions of different
27     actions, or other things. People sometimes use the word "values" to refer to the
28     strength of preferences for different outcomes, and it may be possible to interpret these
29     values quantitatively, in monetary terms, without distortion. But people also use
30     "values" to refer to more qualitative and difficult-to-quantify aspects of environmental
31     outcomes. The challenge for the IED concept, then, it to include appropriate means for
32     both interpreting values and incorporating them into the decision-making process.
33
34            Valuation exercises should be designed to explore the nature, meaning, and
35     importance of these different environmental values.  They should help to reveal what

                                             5-14

-------
        IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

  1      people value, why, and (at least for some values) how much. They should challenge
  2      those people to think critically and clearly about their value judgments and the reasons
  3      for making them. And they should provide satisfactory ways of bringing together
  4      different kinds of values or concerns, while taking account of the costs and benefits of
  5      different alternatives, in ways that help parties reach agreement on values, where this is
  6      possible, or agreement on decisions where disagreements about values persist.  The
  7      section that follows explores some aspects of the nature and meaning of environmental
  8      values, and it examines a few of the complex factors that must be understood and
  9      reconciled in the process of environmental decision-making.
 10
 11           5.3.2  Values
 12
 13           a)    EPA should base its decisions on human interests and values. In
 14                 doing so, it is important that EPA realize that  people care for things
 15                 for different reasons and in different ways.  Valuation must be
 16                 sensitive to these features of human values.
 17
 1«.           Philosophers have debated whether moral values are fundamentally
 19      anthropocentric (i.e., whether values refer to or are based solely on human needs and
 20      interests) or whether some values may be understood independently of human nature.
 21      EPA must base its decisions on human interests and an understanding of the reasons
 22      why people care for the things they do, and for this reason it must focus on human
 23      values. At the same time,  it is important to realize that people  care for things for
 24      different reasons and in different ways. Some things are valued for themselves, or as
 25      ends, and some things instrumentally or as means for the satisfaction of other ends.
 26      People care about some things simply because they desire them or because they
 27      contribute to their happiness. But other things have value because people judge them
 28      to be good or important in their own right, and therefore care about these things
 29      because they believe them to be valuable.
 30
-31           The valuation of environmental resources must be sensitive to the basic
 32      character of human values. Most people value the aspects of  nature and
 33      environmental systems both instrumentally and as ends. They want to protect the
 34      environment as a thing to be valued in itself, as well as managing it for the other
 35

                                            5-15

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1     services it provides.  Environmental management involves multiple goals, which differ in
 2     kind as well as in intensity. The decision problems of environmental management are
 3     thus multi-attribute problems. They involve bringing together different kinds of values
 4     that may conflict or be incompatible in a number of ways. Values from all these
 5     sources are involved in forming preferences.
 6
 7           b)    Preferences need to be examined. Some preferences may be more
 8                 important than others in ways that are not captured by
 9                 measurements of strength of preference alone. Preferences should
10                 be elicited in ways that reveal which preferences individuals endorse
11                 in light of information and the reasons that bear on them.
12
13           Preferences refer to a person's ordering or ranking of alternative states, or in the
14     economic sense, alternative bundles of market and non-market goods and services.
15     Preferences are the basis for the choices that people make; and preferences may be
16     revealed by these choices. As described in the previous chapter, economists have
17     developed two categories of techniques for assessing values: indirect measurement
18     (revealed preference) and direct questioning (expressed preference). The discussion
19     and recommendations that follow describe a supplementary approach, wherein
20     stakeholders are directly involved in the deliberative process to collectively define
21     ecological values.
22
23           EPA needs to examine the preferences people express for different alternatives
24     and their reasons for holding those preferences. People can be asked to reveal their
25     preferences, but the process often  should not end there.  Some preferences may be ill-
26     considered; some may be based on erroneous or insufficient information; and some
27     may be inconsistent with some other deeply held values.  And some preferences may
28     be less important to some individuals than they are to others-tor clear and convincing
29     reasons. The process of revealing preferences needs to explore and examine these
30     reasons, and it should lead people  not only to express their preferences but to endorse
31     the preferences they have expressed after they have been examined in the light of
32     information and reasons that bear on them.  Further, the process can lead others to
33     better appreciate reasons they may not have considered or to formulate opposition to
34     some expressed preferences that may form the basis for rational deliberation and
35     consensus-building.

                                            5-16

-------
        IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1           EPA's objective, therefore, should not be simply to elicit preferences but to try to
 2     uncover or construct critically examined preferences. The process for doing this often
 3     must be deliberative and not be simply limited to gathering and assembling
 4     preferences.
 5
 6           c)    Some important preferences for protecting and managing
 7                 environmental resources are constructed in the elicitation process,
 8                 so it is important to be sensitive to the dependence of preferences
 9                 on the elicitation method and the way the alternatives are framed.
1 o                 This fact has some important consequences: 1) the assessor is
11                 necessarily more an active participant than a neutral observer in this
12                 process; 2) people need to be provided tools to help them think
13                 clearly about their preferences; and 3) the valuation process should
14                 be guided by some operating principles, e.g., questions should not
15                 be posed in only one way.
16
1~           In many cases, values are revealed by the choices people make in their daily
1 v      lives. It might be said that the values underlying these choices were formed by the
19      individual's life experiences. In other cases, the values may not be pre-formed or
20      revealed in the choices people have already made, and they must be constructed
21      during the valuation process: The preferences expressed in the context of a valuation
22      of changes in environmental resources are often constructed at the time the
23      preferences are elicited and cannot be separated from the elicitation process. This fact
24      does not negate the importance of the information gained from preferences constructed
25      in this way. People should continue to be asked to express the values they hold,
26      however they are formed.  But we must be extremely sensitive to the dependence of
27      preferences on the process of elicitation and the way the alternatives are  framed.
.28
29           One consequence of the constructed values perspective is the need to be
30      sensitive to the reasons and causes of the preferences that are  expressed. A second
31      consequence is the need to understand the sensitivity of expressed values to
32      methodological factors. A third consequence is an appreciation that the assessor is not
33      a neutral elicitor but necessarily an active participant in the  process of revealing
34      preferences.
35

                                            5-17

-------
        IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1            Decisions that are made by the individual being questioned in the course of a
 2     valuation exercise will include important values and factors that are not independent of
 3     how information is framed and how preferences are measured. EPA needs to go
 4     beyond some traditionally accepted economic approaches to the valuation of natural
 5     resources.  Simply providing respondents with information may not yield expressions of
 6     coherent and consistent values. People need to be provided with tools to help them
 7     think about their values. Examples of tools that might help people indicate which of
 8     their desires or preferences they would regard as important, and which might be merely
 9     fleeting, are to be found in multi-attribute value analysis and various forms of structured
10     deliberation processes that encourage people to think and reflect upon what they are
11     saying, and to consider whether they really mean what they say.
12
13            Finally, the constructive nature of valuation processes should lead EPA to
14     exercise caution and to take care, where possible, to pose a question in more than one
15     way.  EPA should also be explicitly aware of framing and elicitation effects in order to
16     avoid the biases that they create. Approaches should be developed to attempt to
17     decrease biases from framing, anchoring, etc.  EPA might also consider sponsoring
18     research to determine how accurate direct expressions of preference might  be, under
19     the best deliberative conditions, as indicators of citizens' true environmental values.
20
21            d)     Because  of the qualitative dimensions of values, the valuation
22                  process must emphasize deliberative as contrasted to mechanistic,
23                  predetermined components. Valuation must allow people to express
24                 the different qualities and kinds of values they regard as important,
25                 and it should challenge them to think clearly and usefully about
26                 these values.
27
28            Structuring valuation processes to emphasize deliberative as contrasted to
29     mechanistic or algorithmic ways should help EPA to achieve another important goal,
30     which is to be sensitive to the different qualities and characteristics of values that
31     people often regard as  important. Especially in the areas of environmental protection,
32     many people regard some values as "protected" in ways that lead them to resist
33     conceiving what is important to them as an exchangeable commodity. Also, many
34     people think there are some important social values involved that cannot be reduced to
35     an aggregation of individual preferences.

                                            5-18

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1
 2           The valuation process should allow room for people to express values of these
 3     sorts.  Again, however, even these preferences are not simply to be accepted at face
 4     value. These values, too, need to be challenged and explored.  A sensitivity to the
 5     different characters of values, as well as to the need to explain and defend them with
 6     reason, should become part of the deliberative process.
 7
 8           The nature of values and the choices they support are important topics of
 9     economics, but these subjects have also been studied in depth by other social science
10     disciplines, especially psychology and sociology, and by philosophy. The participation
11     of these other disciplines in the design of valuation instruments/surveys should be
12     encouraged.
13
14           e)     The valuation process should be transparent and explicit. Any
15                 valuation process is subject to possible abuse. Rather than trying to
16                 deny this fact, EPA should guard against possible abuse or bias by
t"7                 making the process explicit and open to review.

19           Any valuation process is subject to possible abuse, and that is certainly true for
20     the kind of more deliberative and constructive process that we are recommending be
21     considered along with more traditional approaches.  Given the nature of environmental
22     values, the agent or elicitor of those values is involved in the process in a way that
23     cannot be detached. Rather than deny that fact, EPA should aim instead at protecting
24     against the potential for abuse in other ways. These involve making as much of the
25     process explicit and open as possible, so that it will be transparent to participants and
26     observers alike.
27
28           Eliciting values is a difficult process in that there are no hard and fast formulas
29     that guarantee success. However, there are better and worse ways of doing these
30     things, and it is important to try to become aware of the lessons that have been learned
31     by able practitioners and to support a culture in the EPA that will encourage Agency
32     officials to develop and use the requisite skills.
33
34                                     	
35

                                            5-19

-------
        IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1           The next section discusses the concept of economic value, economic value and
 2     benefit-cost analysis, economic valuation of the functions and services of environmental
 3     systems, and some issues and problems associated with economic valuation.
 4
 5      5.4 The Economic Valuation Framework
 6
 7           5.4.1  The Concept of Economic Value
 8
 9           a)    The economic value of an environmental change is the increase in
10                 human economic well-being or welfare that it produces.  It is
11                 measured by the amount of some other good (usually money, for
12                 convenience) that can be taken away from those individuals affected
13                 by the change without reducing their welfare below the pre-change
14                 levels.
15
16           The instrumental value of anything stems from its ability to contribute to the
17     achievement of some goal (Costanza and Folke, 1997, p. 49). One kind of value is
18     economic value. The economic value of a thing depends on its contribution to the
19     economic well-being of people. This economic concept of value has its foundation in
20     neoclassical welfare economics.  Each individual's welfare depends not only on that
21     individual's consumption of private goods purchased in markets and of goods and
22     services provided by the government,  but also on the quantities and qualities of
23     nonmarket goods and service flows each receives from the environment. Thus the
24     basis for valuing changes in the flows of goods and services from the environment is
25     their effects on human welfare. This anthropocentric focus of economic valuation does
26     not preclude individuals from having concerns for the survival and well-being of other
27     species.  Individuals can value the survival of other species not only because of the
28     uses they make of them (for example, for food and recreation) but also because of an
29     altruistic or ethical concern. The latter can be the source of nonuse, passive use, or
30     existence values.
31
32           The economic value of a good, that is, its contribution to the welfare of an
33     individual, is measured by how much of some other good is required to make up for its
34     loss. The economic value of a good could in  principle be measured by the quantity of
35     any other good that can substitute for its loss. If the substitute good has a market price,

                                            5-20

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1     then value can be expressed in terms of the money required to purchase the necessary
 2     quantity of the substitute good. Since in principle any other market good can be a
 3     substitute for the good in question, the economic value of the good can be defined as
 4     the amount of money or purchasing power required to substitute for the loss of a unit of
 5     the good, that is, to restore the individual to his/her original level of welfare.
 6
 7          b)    Economic values can be revealed by the choices that people make
 8                as they trade off one good thing for another.
 9
10          Since the source of economic value is the preferences of individuals, the choices
11     that people make reveal something about these preferences and values.  A variety of
12     methods have been developed to infer the values placed on environmental goods and
13     services from the choices that people make. The trade-offs that people make as they
14     choose less of one good and substitute more of another good reveal something about
15     the values people place on these goods. An example is the trade-off between money
16     and air quality that people consciously or unconsciously make  because houses in
17     clean air areas have higher prices, other things held equal,  as has been demonstrated
       in a number of studies. For further discussion of economic valuation methods, see, for
19     example, Braden and Kolstad (1991), Freeman (1993), and Chapter 4.
20
21           5.4.2 Economic Value and Benefit-Cost Analysis
22
23           a)     The economic values of changes in environmental systems should
24                 be included in benefit-cost analyses of environmental policies. Such
25                 benefit-cost analyses will be useful inputs into public policy
26                 decision-making.
27
28           The purpose of economic valuation of the services of environmental systems is
29     to provide decision-makers with information that will help them make choices about
30     policies toward specific environmental systems in the face of scarcity and opportunity
31     cost. In most cases the question is not in the form of "either/or," but about how far to
32     go.  For example, the question usually would not be whether to protect an
33     environmental system or do nothing, but instead what level or degree  of protection or
34     other management intervention should be attempted. Information on the values of
35     environmental systems should take a form that is helpful in these cases.  This means

                                           5-21

-------
        IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1     that the relevant question is the value of the change in the services of environmental
 2     systems that an action would produce, not the total value of all services (Toman, 1997).
 3
 4           If society wishes to make the most (in terms of individuals' well-being) of its
 5     endowment of resources, it should compare the values of what its members receive
 6     from any environmental change or use of a resource (that is, the benefit of the change)
 7     with the values of what its members give up because of the change (that is, the costs).
 8     If the sum of all  benefits exceeds the sum of all costs, the proposed change passes the
 9     "easy" benefit-cost test. When there is more than one proposal or project or degree of
10     change under consideration, only one option will pass the more restrictive benefit-cost
11     test and yield the highest aggregate net benefits. A society could choose to make
12     public policy choices solely on the basis of comparisons of aggregate benefits and
13     costs (maximizing net benefits). However, there may be other things besides economic
14     benefits and costs that a society might want to consider when making policy choices
15     (see, e.g., Arrow et al., 1996).
16
17            5.4.3  Economic Valuation of the Functions and Services of Environmental
18           Systems.
19
20           a)    The economic value of a function or service of an environmental
21                 system can take the form of a direct use value (the value of a service
22                 provided to people), indirect use value (when a function indirectly
23                 supports a service used by  people), or nonuse or existence value
24                 (when an individual values an environmental system even though
25                 he/she does not make any direct or indirect use of it).
26
27           The functions of environmental systems include  photosynthesis,  decomposition,
28     nutrient recycling, and  so forth.  The services of  environmental systems are the
29     materials and other services that they provide that enhance  human welfare and are
30     therefore valued by people. Examples of service flows include wood and fiber from
31     forests, and amenities  such as scenic vistas and wildlife observation. The values of
32     these services are often called direct values (Brown, 1990) or direct use values
33     (Goulder and  Kennedy, 1997) to distinguish them from  the nonuse or existence values
34     mentioned above and the indirect use values discussed below.
35

                                            5-22

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1           The functions of an ecological system can also have an economic value if they
 2     help to support some service flow to people.  If they do support a valuable service flow,
 3     the contribution made by these functions is an indirect use value (Brown, 1990;
 4     Perrings et al., 1995; Goulder and Kennedy, 1997). The indirect use value of a function
 5     can be derived from the change in the value of the service flow that it supports. For
 6     example, if an increase in the rate of photosynthesis in an ecological system results in
 7     an increase in the flow of economically valuable food or fiber, the economic value of the
 8     increase in photosynthesis is the increase in the value of the food or fiber it supports.
 9     What is required to measure this indirect use value is knowledge of the link between the
10     function of the ecological system and the economically valuable service that it supports.
11
12           Sometimes the connection between a function of an environmental system and
13     an economically valuable service flow may be quite direct, as in the case of
14     photosynthesis producing useful plant material.  But in many cases, the connection can
15     be indirect and quite subtle.  For example, photosynthesis by wild flowers may help to
16     support a population of wild bees that also pollinate commercially valuable fruits.  The
17     basic point is that the economic theory of value accommodates changes in the
       functions of ecological systems that affect the well-being of individuals indirectly.
19     However, these values can enter into the decision process only if the links between
20     functions and services are known.
21
22           The functions of environmental systems may also have intrinsic worth or value in
23     the eyes of some. This intrinsic value may be a form of nonuse value.  Such nonuse
24     values have the same standing as direct and indirect use values because they reflect
25     the preferences of individuals. While quantification of nonuse values is difficult, they
26     are of no lesser importance for this reason.
27
28           5.4.4 Issues and Problems
29
30           a)    Aggregation of Individual's Values — Benefit-cost analyses are
31                 based on the summation of individuals' positive (benefits) and
32                 negative (costs) values. This simple summation raises issues of
33                 equity or fairness. For this and other reasons, most economists
34                 suggest using aggregate net benefits as only one of several possible
35                 considerations in public policy decision-making.

                                           5-23

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1           As discussed in Chapter 4, the simple summing of individuals' values to calculate
 2     net benefits is not without its problems. In summary, assessment of net benefits offers
 3     no guidance on the desirability of the resulting distribution of benefits and costs (i.e.,
 4     some projects with positive net benefits may produce benefits for some people but at
 5     the expense of reducing the well-being of others). Another problem is that the
 6     economic values of individuals reflect not only their preferences, but also their
 7     economic circumstances, especially their wealth or income. One who judges the
 8     present distribution of income or wealth to be inequitable has reason also to question
 9     the economic values that emerge from that distribution as a basis for making public
10     policy decisions.
11
12           For reasons such  as these, many writers advocate using aggregate net benefits
13     as only one input into the decision process and allowing consideration of other factors,
14     such as distributional equity, as well (e.g., Arrow et al., 1996).
15
16           b)    Preferences, Values, and Knowledge — One of the basic premises of
17                 welfare economics is that economic values are based on individuals'
18                 preferences and that people know their preferences.  Where
19                 individuals are ignorant of the roles of ecological functions in
20                 contributing to valued service flows, it may be necessary to use
21                 experts' knowledge of the functioning of environmental systems as
22                 an input in the valuation process.
23
24           A basic premise of welfare economics is that economic values are based on
25     individuals' preferences and that people know their preferences. This premise is
26     problematic when it comes to applying welfare economics to the valuation of the
27     services and functions of environmental systems because an individual might act as if
28     he/she placed no value on a function if the individual were ignorant of its role in
29     contributing to a valued service flow from the ecological system (Dasgupta, 1990;
30     Goulder and Kennedy, 1997).  In the practice of environmental system valuation, there
31     are ways to work around the ignorance of individuals about ecological systems. Value
32     measures can be based on ecologists' knowledge of the relationship between the
33     functions of ecological systems and valued services that they provide. Many examples
34     exist in the environmental valuation literature where technical knowledge of physical
35     and biological  relationships coming from experts is incorporated into the process by

                                            5-24

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1     which individuals' preferences are determined. For example, even though people may
 2     not know the relationship between concentrations of air pollutants and the incidence of
 3     respiratory disease, expert knowledge of this relationship can be combined with
 4     peoples' revealed values for avoiding disease to calculate the benefits of improved air
 5     quality. However, when values are elicited through stated preference methods, the
 6     provision of information by experts is more problematical, as is discussed in section
 7     5.3.2 above.
 8
 9           c)    Replacement Cost as a Measure of Value — The cost of replacing an
10                 ecological function with a human engineered system can be used as
11                 a measure of the value of the natural system, but only if substantial
12                 conditions are met.
13
14           Some authors have used estimates of the cost of replacing a function of an
15     ecological system with a human engineered system as a measure of the economic
16     value of the function itself. For example, Gosselink et al. (1974) used an estimate of
17     the cost of a tertiary sewage treatment system as the economic value of the nutrient
       removal function of a wetland.  Replacement cost can be a valid measure of economic
19     value only if three conditions are met (see, for example, Shabman and Batie, 1978):  1)
20     the human-engineered system provides functions that are equivalent in quality and
21     magnitude to the natural function; 2) the human-engineered system is the least cost
22     alternative way of performing this function; and 3) individuals in aggregate would in fact
23     be willing to incur these costs if the natural function were no longer available.
24
25            5.4.5 Conclusions
26
27           In principle, the economic valuation framework can be utilized to define  and
28     measure the economic values of changes in the functions and  services of ecological
29     systems that affect individuals' welfare either directly or indirectly. In principle, the
30     framework can be used to formulate the choices required to incorporate a variety of
31     factors into social decisions. This framework  may be difficult to implement in practice
32     where the relationship between the function and the service flow to individuals is
33     indirect or subtle.  Further, effective application of the framework requires that  all
34     relevant matters be included, which sometimes suggests the use of processes including
35     deliberation, discussed immediately below, to assure that this happens. There is now a

                                            5-25

-------
        IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1     growing number of examples of application of the economic framework to
 2     environmental valuation (see, e.g., Pairings et al., 1995; Freeman, 1997).
 3
 4                                         * * » • »
 5
 6           The next section discusses the need for deliberation in eliciting values. This
 7     reflects the Subcommittee's interest in finding ways to determine people's values
 8     regarding specific environmental problems, in the face of the shortcomings that are
 9     suggested as being associated with traditional environmental valuation exercises.
10
11      5.5 The Importance of Deliberative Processes to Valuation
12
13            5.5.1  Introduction
14
15           Deliberation among interested parties, though not required in all decision-making
16     cases, is often useful for eliciting information on important goals, on "best judgment" on
17     certain relationships, on measures for judging goal attainment,  and on value tradeoffs
18     that must be made in decision-making and implementation. A variety of types of
19     deliberation are available, depending upon the sufficiency of knowledge available to
20     address the problem at hand, and the extent of agreement that might exist on the types
21     of values involved in the issue.
22
23           Appropriate types of deliberation can help EPA select suitable analyses and
24     reach decisions that incorporate broadly acceptable tradeoffs of efficiency, fairness,
25"    and sustainability.  A substantial literature related to environmental decision-making,
26     including a number of recent reports (NRC, 1996; Presidential Commission of Risk
27     Assessment and Risk Management, 1997), stress the importance of deliberation to
28     Agency decision-making.
29
30           By deliberation we mean..."any formal or informal processes for communication
31     and for raising and collectively solving problems" (NRC, 1996; p. 73). Deliberation is
32     seen as the process of bringing together those people whose expertise or perspective
33     is useful in order to craft management and policy options. Deliberation can examine
34     findings,  interpretations, and/or economic and non-economic values in light of each
35     other to achieve greater clarity and to construct and consider management options.

                                            5-26

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1     Thus, deliberative processes differ from but supplement analytic processes which use
 2     "systematic application of specific theories and methods"  (NRC, 1996; p. 215) such as
 3     those of science, social science, and economics. Some types of deliberation primarily
 4     involve disciplinary experts, while others require involvement of stakeholders who bring
 5     disparate perspectives and knowledge to the table.  In some cases, stakeholders may
 6     be disciplinary experts who may have different interpretations of the state of knowledge
 7     than experts in the Agency. Deliberative processes need not be reserved for
 8     interactions with non-federal stakeholders, but may also be used to facilitate intra-
 9     Agency or interagency discussions on an issue.
10
11           Some of the reasons given in the NRC and Presidential Commission reports for
12     using deliberative processes include:
13
14           Clarifying and potentially advancing resolution of issues of fairness. Issues
15           such as distributional equity (Who benefits?) and procedural equity (Who
16           decides?) can be raised explicitly and viewed in the specific context in which an
*7           environmental decision needs to be made (see, e.g., Vaughan, 1995). (When
             we use the word "fairness" we are referring to both distributional and procedural
19           equity, two issues that have been concerns of the  Environmental Justice
20           movement.)
21
22           Informing multi-dimensional tradeoffs among efficiency, fairness,
23           environmental sustainability, and other concerns. In situations that involve
24           conflicts about such tradeoffs,  deliberation can explore potential tradeoffs and
25           craft various options that might reduce the amount of conflict (see, e.g., NRC,
26           1996).
27
28           Increasing credibility.  By discussing various perspectives, the credibility of
29           information and Agency decision-making can be enhanced, although agreement
30           with Agency process does not necessarily result in agreement with Agency
31           decisions (see, e.g., Rosener,  1981; Mazmanian and Nienabur, 1979).
32
33           Informing priorities for research. Studies suggest that more data are not
34           necessarily better for organizational decision-making. Individuals and
35           organizations have limits on the amount of information they can assimilate.  In

                                            5-27

-------
        IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1           addition, some information may be peripheral to the key dimensions of the
 2           decisions to be made and do little to clarify or resolve conflicts (Dietz, 1988).
 3           Deliberation can help determine the research that is most likely to be key to
 4           decision-making and distinguish it from research that may not reduce uncertainty
 5           sufficiently to inform the decision at hand (NRC, 1996). Thus deliberation can
 6           conserve resources.
 7
 8            5.5.2  Aspects and  Recommendations
 9
10           a)     We recommend that the context (characteristics, scope, and
11                  implications) of an Agency decision be taken into account in the
12                  selection of deliberative processes.
13
14           This concept is expressed by our typology of deliberative processes with
15     stakeholders and experts (Figure 5-3). The type of deliberation, the selection of
16     specific deliberative techniques [e.g., scientific meeting, citizen advisory panel, citizen
17     jury, informal meeting, etc.), and the selection of participants all vary by the situation
18     (English etal., 1993).
19
20           There is no one-size-fits-all approach. The Agency should consider two
21     questions when determining the type of deliberation that is most appropriate:
22
23                  /. To what extent is the agreement on values (e.g., fairness, sustainability,
24                  efficiency, etc.) and on appropriate tradeoffs among them sufficient to
25                  reach a decision?
26
27                  The relative importance of efficiency, fairness, sustainability and other
28                  concerns may vary among experts and stakeholders. When Agency
29                  decisions require tradeoffs among and along these dimensions and are
30                  likely to lead to conflict, agencies are often forced to make judgments that
31                  cannot be based solely on knowledge.  The selection of an appropriate
32                  type of deliberation will depend on where the extent of agreement falls on
33                  the continuum between high and low.
34
35
36

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999—Do Not Cite or Quote
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17

19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
      STATE OF
     KNOWLEDGE

        Insufficient
DELIBERATION TYPE
         Sufficient
                      EXPERT
                   DELIBERATION
           INTEGRATED
          DELIBERATION
                     OVERSIGHT
                   DELIBERATION
          STAKEHOLDER
          DELIBERATION
                    STATE OF VALUE AGREEMENT
Figure 5-3.  Typology of Deliberation Processes With Stakeholders
  and Experts (adapted from Chess, Dietz, and Shannon, 1998)

 //. To what extent is the state of knowledge sufficient to address the
 problem at hand?

 By knowledge we mean information and understanding from the biological
 and physical sciences, engineering, economics, and the other social
 sciences.  The answer to this question depends on the extent of
 knowledge about information critical to making a particular decision.  In
 many situations, knowledge from environmental sciences, such as
 information about environmental system processes and the nature of
 potential threats, may play a decisive role. In others, knowledge about
 economic costs and benefits may be decisive. Social science knowledge
 may also be decisive, for example by providing an understanding of
 communities where demographics, ethnicity, or racial composition must
                                            5-29

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1                 be given serious consideration. Agency diagnosis of a situation, and the
 2                 form of deliberation needed, will depend, in part, on its assessment of the
 3                 extent to which available information is adequate for making a decision.
 4                 When the state of knowledge is insufficient or controversial, or when there
 5                 is lack of agreement about the state of knowledge, more extensive
 6                 deliberation will be needed.
 7
 8           The Subcommittee appreciates that Agency decisions may be constrained by
 9     regulations, resources, and court decisions. These limitations should be made clear to
10     participants in any deliberative process. Yet, situations in which agencies have no
11     latitude are rare (Pflugh and Shannon, 1990).  Without reflection on the above two
12     questions, agencies may fail to appreciate the likelihood that a decision, made without
13     appropriate deliberation, might become hopelessly stuck in environmental gridlock.
14
15           b)    When agreement about values (economic and  non-economic) is high
16                 and the state of knowledge (relevant science, economics, and social
17                 science) is sufficient and/or non-controversial, Agency decision-
18                 making is likely to be routine.  Deliberation will only be needed
1 g                 periodically, if at all, for oversight (Oversight Deliberation).
20
21           Most Agency decisions are routine administrative ones that conform to existing
22     regulations and policies. Such decisions may include non-controversial permitting,
23     changes in labeling, and minor shifts in administrative procedures.  In such situations,
24     oversight deliberation, the periodic conferring of experts to assess a program and
25     potential modifications, is appropriate. However, if conflict develops around multi-
26     dimensional tradeoffs or the state of knowledge, the type of deliberation will need to
27     move toward another quadrant.
28
29           c)    When agreement about values is low, but the state of knowledge is
30                 sufficient and/or non-controversial,  Agency decision-making will
31                 require multi-dimensional tradeoffs based on knowledge.
32                 Stakeholder Deliberation is needed.
33
34           In such situations, the state of knowledge is sufficient to  inform multi-dimensional
35     tradeoffs, but there is little agreement about which tradeoffs to make. Because the

                                             5-30

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1     conflict is usually based fundamentally on values, not knowledge, the deliberation can
 2     involve primarily stakeholders who will evaluate tradeoffs in light of their priorities.
 3     Stakeholders, informed by available knowledge, can craft options with varying tradeoffs.
 4     Experts may provide information about the potential impacts of various options, but they
 5     need not be as extensively involved as in expert deliberation, as noted in the following
 6     section.
 7
 8           d)    When agreement about values is high, and the state of knowledge is
 9                 insufficient and/or controversial, Agency decision-making is likely to
10                 be experimental and iterative. Expert Deliberation is needed.
11
12           In such situations, making decisions is difficult primarily because of the state of
13     knowledge.  For example, there may be limited knowledge about the impact of human
14     management on a particular environmental system. Expert deliberation — on-going
15     conferring among experts (often from different disciplines) — can be needed to develop
16     appropriate monitoring processes and to interpret results. Based on the results, expert
•f     deliberation may result in recommendations for changes in management of the
       environmental system; i.e., adaptive management with expert deliberation at intervals
19     determined by the nature of the experiment.  For example, monitoring of the impact of
20     reducing water flow to an environmental system may require experts to confer at regular
21     intervals to review monitoring data and determine if the water flow should be changed.
22     However, if value-based conflict arises over the results of such iterative decision-
23     making, the situation will require integrated deliberation, involving experts and outside
24     stakeholders working together to make multi-dimensional tradeoffs on the basis of
25     limited knowledge.
26
27           e)    When agreement about values is low and the state of knowledge is
28                 insufficient and/or controversial, Agency decision-making is likely to
29                 require multi-dimensional tradeoffs based on insufficient knowledge.
30                 Integrated Deliberation involving both experts and outside
31                 stakeholders is needed.
32
33           These decisions are usually the most difficult for agencies because there is little
34     confidence in the state of knowledge about the impacts of tradeoffs on economic
35     efficiency, fairness, sustainability and other concerns. In such situations integrated

                                            5-31

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1     deliberation may be needed. By integrated deliberation we mean on-going interaction
 2     among experts and stakeholders during problem formulation, collection of information,
 3     and development of options (NRC, 1996; Commission on Risk Assessment and Risk
 4     Management, 1997).  Integrated deliberation may also be needed during
 5     implementation and performance evaluation, and may take the form of adaptive
 6     management with stakeholders and experts reviewing the results and suggesting
 7     iterative changes. The nature of integrated deliberation depends on the situation, but,
 8     in general, the greater the conflict (or potential conflict), the more extensive the
 9     deliberation needed.
10
11           f)     EPA can identify stakeholders by asking the following questions
12                 (adapted from Chess and Hance, 1994):
13
14                 -Who has information and expertise that might be  helpful?
15                 -Who has been involved in similar decisions before?
16                 -Who has wanted to be involved in similar decisions before?
17                 -Who may be affected by the decision?
18                 -Who may reasonably be angered if not included?
19
20           The process for selecting stakeholders should be perceived as fair, lead to a
21     broad representation of stakeholders, and should ensure that representatives of
22     stakeholder groups are perceived by the groups as acceptable representatives (NRC,
23     1996; English, 1993).
24
25           Stakeholder involvement processes need not preclude involvement by a greater
26     number of participants. In fact, in conflicted situations it may be useful to complement  a
27     stakeholder involvement process with survey data  (see, e.g., Kathlene and Martin,
28     1991). Researchers are also exploring forms of deliberative processes such as citizen
29     juries, which involve random selection of citizens (Brown et al., 1995; Crosby et al.,
30     1986).
31
                                            5-32

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1           g)    Involving stakeholders before extensive development of Agency
 2                proposals may generate constructive options and trust. Integrated
 3                and stakeholder deliberation should include stakeholder involvement
 4                in problem formulation and options generation.
 5
 6           Early involvement facilitates agencies' ability to define problems and craft
 7     solutions that may meet satisfactorily the goals of both agencies and stakeholders.
 8     When participation is on-going, the Agency can identify problems before they become
 9     crises (Thomas, 1995). As the Government Accounting Office (1994) has suggested in
10     the context of Superfund, earlier outreach can improve participatory efforts and
11     minimize dissatisfaction.
12
13           h)    We recommend that the existing literature on group processes,
14                ongoing research on deliberation, experienced practitioners, and
15                stakeholder input be used to provide guidance on the selection of
16                deliberative methods.
.»^
             Just as selection of the type of deliberation is determined by the situation, the
19     selection of deliberative methods is dependent on the context (English, 1993).  There is
20     no single guide that can cover all situations, and care must be taken in designing a
21     process so it is appropriate to the problem being addressed. As with all policy analysis
22     tools, high quality results depend on careful application of the methods that are
23     grounded in both theory and practical experience. The sources of guidance for
24     selection of deliberative processes include:
25
26                 Existing Literature. There is a substantial and sophisticated body of
27                 research on small group processes and a smaller but important body of
28                 research on deliberative processes.  Key entry points to this literature
29                 include Mazmanian and Nienaber, 1979; Rosener, 1981; Fiorino, 1990;
30                 Dietz, 1994; Renn et al., 1995; NRC, 1996; and U.S. Department of
31                 Energy, 1996.
32
33                 Ongoing Research. Increasing effectiveness in the use of deliberation
34                 will require ongoing research on deliberative processes and their use in
35                 policy contexts. Some research of this type is proceeding, including a few

                                            5-33

-------
        IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1                  projects funded by the recent EPA/National Science Foundation joint
 2                  initiative. But compared to other policy analysis methods, very little has
 3                  been invested in research to support deliberatiorv(NRC, 1996).
 4
 5                  Experienced Practitioners. Like many methods of policy analysis, the
 6                  translation of theory and empirical research into practice is a craft that
 7                  depends on practitioners who are both knowledgeable and experienced.
 8                  Such practitioners can also be an important source of systematic
 9                  information.  However, deliberation rarely can simply be turned over to a
10                  consultant. Both integrated deliberation and stakeholder deliberation
11                  require Agency involvement.
12
13                  Stakeholder Input: Those participating in a deliberation are a vital source
14                  of information on how to proceed.  Further, systematic post-hoc studies of
15                  stakeholder perceptions are critical to building cumulative knowledge (see,
16                  e.g., Balch and Sutton, 1995; and NRC, 1996). For further examples of
17                  such studies, see also "Existing Literature" above.
18
19            i)     EPA should build additional institutional capacity for deliberation.
20
21            Effective decision-making depends upon institutional capacity appropriate to the
22      challenges posed by the situations we have described. That institutional capacity is
23      inadequate, both in the larger community and in EPA, and the Agency should move to
24      develop it in both arenas.  Research and training support, for example, would deepen
25      the resources available to EPA, while greater use of deliberative processes would call
26      forth additional interest in the outside community.  EPA can utilize external resources to
27      make more effective use of deliberation, but it also needs to increase its internal
28      capacity by adding expertise and making other organizational changes such as:
29
30                  Expertise. Deliberative processes depend on experienced, trained
31                  practitioners and knowledge and skills provided by those with familiarity
32                  with social science research in areas such as psychology, sociology,
33                  anthropology, political science, management, and economics. EPA may
34                  need to hire personnel with this expertise.
35

                                             5-34

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1                 Organizational Structure. Currently, community relations, public
 2                 information, and other personnel serve as technicians (carrying out
 3                 instructions) rather than being part of organizational decision-making
 4                 (Grunig, 1992). Thus, key decisions may be made without the presence
 5                 of a manager with expertise in stakeholder values and concerns. We
 6                 suggest that those with expertise in deliberation should serve not only as
 7                 technicians, but as both mid-level and senior managers, who can assist
 8                 the Agency to develop appropriate deliberative processes and to make
 9                 critical decisions that involve conflicts over values.
10
11                 Organizational Learning.  To improve deliberative processes, agencies
12                 must leam from experience (Chess et al., 1995).  Mechanisms are
13                 needed to collect, analyze, and interpret  data on deliberative processes.
14                 Just as records of environmental monitoring or risk assessments help to
15                 improve those practices, learning about deliberation will require record
16                 keeping and means to institutionalize what is learned.
17
             j)     Because environmental systems are arranged in hierarchies,
19                 effective deliberation requires institutional arrangements that reflect
20                 this structure, with appropriate authority, institutional capacity, and
21                 information available at local, regional, and national levels.
22
23           Such institutional arrangements allow  citizens who are directly affected by
24     tradeoffs among fairness, economic efficiency, and sustainability to deliberate over their
25     resolution and to bring to bear critical information about local conditions, particularly
26     those involving spatial and temporal distribution of resources.
27
28           Relationships among institutions at the same level and with  those at other levels
29     should be relatively flexible, allowing stakeholders and experts the leeway to match
30     institutional capacities with the specific ecological problem. To enable community-
31     based institutions to participate in making multi-dimensional tradeoffs, care should be
32     taken to avoid severe imbalances among different levels of institutions.
33
34           When community-based institutions have already developed management of
35     resources and made multi-dimensional tradeoffs, other levels of government should

                                            5-35

-------
        IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

  1      recognize the local management systems that are in place and, to the extent possible,
  2      work with rather than supplant those systems (see, e.g., Schlager, 1994).  While
  3      community institutions can draw on local knowledge about distribution of resources over
  4      space and time, they are often at a disadvantage in involving additional scientific
  5      expertise. Conversely, higher level institutions have greater capacity and resources to
  6      draw on scientific expertise but may be at a disadvantage in their appreciation of local
  7      knowledge about the functioning of an environmental system or of a resource within it.
  8      Thus, successful  resolution of sustainability problems and multi-dimensional problems
  9      may be more likely if scientific capacity of local institutions is increased and higher level
10      institutions draw on local knowledge.
11
12            k)     Deliberation should be protected from undue manipulation by
13                  special interests.
14
15            Conflicts among interest groups vying for power is inherent in democracy.
16      Deliberative processes can be protected from undue manipulation through: a) open  and
17      transparent processes, b) rules and mechanisms to ensure due process; c) sufficiently
18      broad participation so that deliberation is not dominated by special interests, d)
19      structured group processes that reduce the possibility of manipulation, and e) expertise
20      of those leading the processes so that manipulation is seen quickly and constrained.
21
op                                      *****
23
24            The following section presents information on additional approaches to valuing
25'     benefits of environmental systems. The additional approaches are discussed in
26      response to the need to improve this ability to value environmental benefits in ways  that
27      address the limitations of traditional economic analysis techniques. The following
28      considerations, and others, need to be the subject of continuing research as society's
29      needs to address environmental issues become increasingly more complicated and
30      exceed the adequacy of current methodologies.
31
32
33
34
35

                                             5-36

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1     5.6 Additional Approaches to Valuation of Environmental Systems
 2
 3           5.6.1  Introduction
 4
 5           "Valuation ultimately refers to the contribution of an item to meeting a specific
 6     goal. A baseball player is valuable to the extent he contributes to the goal of the team's
 7     winning. In ecology, a gene is valuable to the extent it contributes to the goal of survival
 8     of the individuals possessing it and their progeny. In conventional economics, a
 9     commodity is valuable to the extent it contributes to the goal of individual welfare as
10     assessed by willingness to pay.  The point is that one cannot state a value without
11     stating the goal being served. Goals are not necessarily independent, i.e. increasing
12     activity toward one goal may result in tradeoffs which decrease other goals" (Costanza
13     and Folke, 1997)
14
15           5.6.2 Findings
16
17           a)     The value of something stems from its ability to contribute to the
                   achievement of some goal.
19
20           b)     Fairness, sustainability, and other values, as well as economic
21                 efficiency, are goals that should be recognized and explicitly
22                 addressed in environmental resource decision-making. Such goals
23                 can be compatible and mutually reinforcing; i.e., sustainability could
24                 be enhanced by improving the efficiency of resource use and
25                 reducing ecological stressors, and fairness can enhance
26                 acceptance of and compliance with environmental decisions.
27
28           Conventional economic value is based on the goal of maximizing individual
29     economic welfare, which includes all the things people want including environmental
30     quality and sustainability of ecological systems. But explicit recognition of such non-
31     private goals, and of their importance, is crucial in environmental valuation.  For
32     example,  sustainability is critical to future generations. It is important to properly
33     ascribe value to outcomes that contribute to achieving sustainability, social equity, or
34     other ends that may be deemed important, as well as the value placed on the goals of
35     producing things people want directly. This broadening is particularly important if the

                                            5-37

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1     goals are potentially in conflict. The narrow view of environmental decision-making as a
 2     choice of one or another action, each representing widely differing value-driven
 3     viewpoints, outcomes, benefits, and costs, is rarely the case. Rather, environmental
 4     decision-making is an activity that involves trade-offs among a number of goals that are
 5     the product of different value choices. Decisions must reflect some combination of
 6     these disparate values. Goals that may be important in environmental decision-making
 7     include those relating to aesthetics, services of nature, recreation, consumption,
 8     culture, biodiversity, and human health. Because of the variety of values-driven goals
 9     that enter the decision-making process, there is a need for a number of ways to
10     conduct valuation exercises that are used to inform decision-makers.
11
12           c)    Ecological systems provide diverse benefits to society, including
13                 marketed goods and services, nonmarket direct and indirect use
14                 values, and nonuse or existence values. Ecological services have
15                 values that can be identified and at least partially quantified.
16
17           d)    Appropriately characterizing the societal value of the full range of
18                 environmental services requires approaches in addition to traditional
19                 individual preference techniques.  Such techniques would inform the
20                 democratic processes that can be used to determine and then
21                 implement policies which enhance the flow of environmental
22                 services that would not be provided by the operation of market and
23                 quasi-market fulfillment of individual choice.
24
25           e)    An  important example of an element that may not be adequately
26                 reflected in valuation exercises is biodiversity. Biodiversity is
27                 positively related to ecological services and is declining principally
28                 because of anthropogenic- induced land use changes and other
29                 habitat alteration. Biodiversity has benefits, and its loss is a cost that
30                 should be factored into decision-making.
31
32            Biodiversity is one of the important attributes of environmental systems which
33     clearly emerges when aggregates of species and populations are considered. The term
34     biodiversity is commonly used to describe the number and types of organisms in a
                                            5-38

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1     given area, as well as the variability (genetic and phenotypic) within populations of
 2     organisms.  In addition, biodiversity includes "the ecological roles that these species
 3     play, the way that composition of species changes as we move across a region, and the
 4     groupings of species (ecosystems) that occur in particular areas (such as grassland or
 5     forest), together with the  processes and interactions that take place within and between
 6     these systems.  It also covers the diversity of ecosystems in landscapes, of landscapes
 7     in biomes, and of biomes on the planet." (Heywood et a!.,  1995).
 8
 9           Because biodiversity reflects the complex interactions among species in an area,
10     and between living organisms and the physical environment, it is reasonable to
11     suppose that the services provided by a specific environmental system are affected by
12     the various levels of diversity contained within the system.  Therefore, when conducting
13     valuation exercises, it is important to consider the biodiversity of the system.
14
15           According to Heywood et al. (1995), The values placed on biodiversity are
16     strongly linked to the human influences on it and their underlying social and economic
17     driving forces.  They are  also dependent on some degree of knowledge of the scientific
       role of  particular elements or processes of biodiversity in the functioning of our
19     ecosystems  and societies.... [W]hile it is undoubtedly true that the multiple values of
20     biodiversity are not adequately captured in its market value, if we want to commit and
21     prioritize resources to its conservation and sustainable use, then applying economic
22     measures to its evaluation is unavoidable."
23
24            It is likely that the  environmental values most often included in economic analysis
25     are those associated with environmental services flows having direct market values
26     (e.g., timber, Pharmaceuticals, fish, wildlife, etc.). However, the importance of
27     biodiversity to the provision of these, and many other service flows, has not yet been
28     adequately recognized or explored. Therefore,  we need to identify critical functions and
29     to focus research efforts on these so that important  measures reflective of the variety of
30     factors associated with biodiversity can be completely reflected in the analysis of
31     benefits from environmental management.
32
33            f)     Environmental systems are organized in the form of nested
34                 hierarchies, having spatial, temporal, and organizational dimensions.
35                 For this reason, decision-making authority should be organized in a

                                            5-39

-------
        IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1                 way that mirrors or parallels the natural systems. A major benefit
 2                 would be reduced information costs and analytical complexity.
 3
 4           There is a need to manage events in the human/environmental system at a
 5     range of different scales (Wilson and Dickie, 1995).  In the management of natural
 6     resources,  the spatial scale inherent in the regulatory apparatus often influences the
 7     nature of the questions that are addressed. For example, marine fisheries are
 8     managed primarily on a regional scale (by the U.S. regional fishery management
 9     councils), and this management has tended to focus on questions of how many fish to
10     harvest rather than on ecological factors such as habitat and population structure that
11     occur at smaller scales.  In addition, large scale management oversimplifies local
12     phenomena that are important to regulatory goals; ecological interactions are complex
13     and the outcomes of human intervention are often uncertain. Because region or
14     ecosystem-specific differences can lead to large transaction costs in centralized
15     management schemes, an allocation of decision-making authority toward a
16     decentralized hierarchical regulatory system can bring timely and accurate information
17     to bear on both the regulatory and valuation problems.
18
19           With respect to valuation, the ideal is to exercise decision-making authority at the
20     scale where individual incentives are most closely aligned with societal objectives. Put
21     another way, individuals, acting in their own self interest, may over-exploit open access
22     resources (often referred to as the "tragedy of the commons"-Hardin, 1968).
23     Individuals acting as parts of groups or communities may adopt norms or rules of
24     behavior that serve to avoid such over-exploitation (Ostrum, 1990; Schlager et al.,
25     1994).
26
27           Another benefit of decentralized management is to improve the equity and
28     credibility of decision-making processes. These are important characteristics that can
29     improve the potential for stakeholder "buy-in" and compliance with decisions.
30
31           g)    Ecosystem management is a promising new approach for making
32                 choices in the face of the uncertainties associated with the risks and
33                 benefits of management decisions, especially in the areas of fairness
34                 and environmental sustainability.
35

                                            5-40

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1            Many of the more straightforward environmental problems that have been the
 2     focus of the last few decades of environmental protection in the U.S. (e.g., air and water
 3     pollution by toxic chemicals) have been addressed in EPA's and other agencies'
 4     environmental decisions and policies, resulting in improved environmental conditions.
 5     However, the human activities that now cause higher risks for ecological systems are
 6     more physical than chemical, involving such stressors as habitat alteration and
 7     hydrologic  modifications that cause effects at regional scales. The next generation of
 8     advances in environmental protection, then, must deal with regional-scale interactions
 9     of humans with natural ecological systems.  Ecosystem management (also termed
10     community-based environmental protection [CBEP] by EPA) is an emerging concept
11     that addresses such regional issues and seeks to develop mutually dependent
12     sustainability for human and ecological systems (U.S. MAB, 1994; Christensen et al.,
13     1996; Harwell et al., 1996).
14
15           Ecosystem management uses a deliberative process for society to make explicit
16     choices about what level of sustainable ecological quality is desired at each specific
17     location within a region. This deliberative  process requires extensive and recursive
       interactions among scientists, decision-makers, and stakeholders to identify spatially
i9     explicit environmental goals, define the ecological and societal endpoints that
20     characterize those goals, and implement management approaches to achieve the
21     goals.  This requires an approach to governance that is commensurate with the spatial,
22     temporal, and organizational hierarchies of the ecological systems. Further, ecosystem
23     management involves making decisions with explicit consideration of the uncertainties
24     associated with environmental risks and benefits.
26
26           Sustainability is a central desired tenet of ecosystem management.
27     Consequently, the time horizon for decisions is by necessity inter-generational in
28     human terms and, thus, issues of sustainability and inter-generational equity (i.e.,the
29     distribution of ecological services over generations of humans) are intimately linked.
30     Since ecosystem management is goal-driven, it is essential that science properly inform
31     the societal goal-setting process by defining what is sustainable and what constraints
32     the natural systems impose on management options.  Just as importantly, however, it is
33     up to society, not scientists alone, to set the sustainability goals.  Once these goals are
34     established, then science can help by translating the societal goals into specific
35     ecological, hydrological, and societal conditions and identifying the specific measures

                                             5-41

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1     that need to be monitored in order to evaluate how well the goals are or are not being
 2     met.  The comparison of how ecological systems respond to decisions and policies with
 3     specified performance criteria constitutes the critical element needed for adaptive
 4     management, which is a key component of ecosystem management. Consequently, for
 5     regional-scale environmental protection, ecosystem management constitutes a driving
 6     force both linking societal values with scientific understanding and guiding the
 7     environmental decision-making and adaptive management processes, forming the
 8     context in which the science and society, analytic/deliberative process develops (MAS,
 9     1996).
10
11           h)     Scientific analyses and models that explain the principles of
12                 operation and the interconnections among socioeconomic and
13                 ecological systems are important to environmental management.
14                 Not all environmental problems require the same level of scientific
15                 analysis and deliberation. Complex issues may need to emphasize
16                 deliberation and use judgment based on qualitative analyses.
17
18           i)     During deliberations and presentations to the Subcommittee, several
19                 interesting directions emerged that, with further research, may
20                 improve our ability to measure some of the emergent values that
21                 environmental management approaches, and environmental systems
22                 in general, confer on the economic system. Results of such
23                 research should help to determine how best to align the regulatory
24                 process with the opportunities for minimum cost compliance.
25                 Research is also needed to determine how dynamic adjustments to
26                 regulatory interventions can be incorporated better into the process
27                 of estimating the prospective costs and benefits of possible actions
28                 so that realistic information is given to decision-makers.  For
29                 example,  research may be fruitful in the following areas:
30
31                 1)    The relationship between biodiversity and environmental
32                      service flows of importance to humans:
33
34                      As discussed above, it seems a good working hypothesis that

                                          5-42

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1                 biodiversity of an ecosystem is related to the level of goods and services
 2                 of value to humans from that system.
 3
 4                 2)     Systems analysis of the relationship between economy and
 5                       environment:
 6
 7                       A number of recent empirical studies in the United States have
 8                 shown that states with better environmental conditions generally have
 9                 better economic conditions. For example, Meyer (1992) demonstrated a
10                 positive statistical correlation between higher state environmental rankings
11                 and better job conditions, e.g., lower unemployment and higher
12                 productivity.  Similarly, Templet and Farber (1994) demonstrated that
13                 surrogate measures of environmental risk (e.g., the ratio of chemical
14                 releases,  as reported by the Toxic Release Inventory (TRI) data base, to
15                 jobs created in the discharging industrial sector) are inversely related to
16                 measures of economic welfare (e.g., unemployment rates and disposable
17                 income per capita), and Cannon (1993) showed that higher economic
                   growth rates prevail where environmental rankings are higher. Hall (1994)
19                 developed state rankings based on economic and environmental
20                 conditions that turned out to be remarkably similar and correlated well.
21
22                       These results do not establish the fact or direction of causation.
23                 While these results suggest that the economy could be enhanced if the
24                 environment is improved, it is also possible that citizens in higher income
25                 states have a higher demand for environmental quality and convince their
26                 governments to deliver it. It also may be that the initial endowment of
27                 natural and man-made resources in some states led to both economic
28                 and environmental conditions being better.
29
30                       A positive correlation between broad, general measures of
31                 environmental conditions and measures of economic production is
32                 frequent in circumstances where rampant industrial growth has occurred
33                 without consideration of the damage done by residuals, a condition that is
34                 familiar in some developing countries, and indeed, occurred historically in
35                 the United States. When such externality conditions are presented in

                                            5-43

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1                 specific situations (i.e., industrial water discharges damaging a fishery),
 2                 existing regulatory and tort remedies are readily available. The issue is,
 3                 however, whether more subtle and general levels of environmental stress
 4                 cause effects such that actions can be taken which would improve both
 5                 the environment and economic production.  Research on this issue could
 6                 be designed to determine whether such circumstances exist, and if so,
 7                 what remedies are available that would improve environmental conditions
 8                 at a cost that did not exceed the  improvement of economic production and
 9                 the increased value ascribed to the environmental improvement itself.
10
11                       The complementarity interpretation can be understood in relation to
12                 a simple model (Figure 5-4) which places the economic system within the
13                 environmental system and reliant on it to provide energy, materials and
14                 information and to accept waste. If the environmental system is
15                 negatively impacted by disruption or residuals, it may contribute less of
16                 these services to the economic system and, as a consequence, the
17                 economy may become less productive and contribute less to public
18                 welfare. In this view, the economy could be enhanced if the environment
19                 is improved and vice versa.
20
21                 3)    Biophysical measures as a surrogate for the value of
22                       ecological systems:
23
24                        One alternative method for estimating ecological values assumes
25                 a biophysical basis for value (e.g., Costanza, 1980; Cleveland et al., 1984;
26                 Costanza et al., 1989). This theory suggests that in the long run, humans
27                 come to value things according to how costly they are to produce, and
28                 that this cost is ultimately a function of how organized the goods are
29                 relative to their environment. To organize a complex structure takes
30                 energy, both directly in the form of fuel and indirectly in the form of other
31                 organized structures like factories.  For example, a car is a much more
32                 organized structure than a lump of iron ore,  and therefore, it takes much
33                 energy (directly and indirectly) to organize iron ore into a car. The amount
34                 of solar energy required to grow forests can therefore serve  as a measure
35

                                            5-44

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16

id
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
             ENVIR   ON  WENT
    Energy
    Information*
    Materials
                           ECONOMIC SYSTEM
                             Purchase Price (S)
\
r
Goods & Services
l
r

                    PRODUCTION  \   /^
                        &        )   (   CONSUMERS
                    DISTRIBUTION  i   \
      Reconstituted
      Materials (low
      entropy)
                        t
      Labor
                                 Wages ($)
Stressors, e g.. Residuals,,
Disruption (high entropy)
                ENVIR   ON  MENT
   Figure 5-4. The Relationship of Economy to Environment
           (modified from Templet, 1995)

of their energy cost, their organization, and hence, according to this
theory, their value.  Various methods have been suggested in the
literature for estimating these direct and indirect energy costs ranging
from methods based on input-output analysis (Costanza, 1980; Cleveland
et al., 1984; Costanza et al., 1989; Ulanowicz, 1986; Mageau et al., 1995)
to methods based on embodied energy (or "emergy") analysis (Odum,
1996). For a comparison of input-output based and "emergy" based
analysis, see Brown and Herendeen (1996). Of these, the input-output
approach, borrowed from economics, appears to offer the most promise
                                            5-45

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1                 for estimating direct and indirect energy costs.
 2
 3                 4)     The efficiencies and environmental benefits that may result
 4                        from the implementation of industrial ecology principles:
 5
 6                        If Figure 5-4 is treated as the basis for a simple model, inputs of
 7                 energy, information and materials are seen to yield outputs from the
 8                 economic system in the form of goods, services, and residuals.  Improved
 9                 processes for converting inputs into desired outputs can yield lower
10                 energy use and/or lower production of residuals.  Both of these results
11                 tend to reduce environmental stresses. Recognition of these interactions
12                 has led to a focus on "industrial ecology," which emphasizes a holistic
13                 approach to organizing production processes as a substitute for the
14                 piecemeal, sub-process decisions which often ignore potential beneficial
15                 connections across the enterprise and even outside of it. Recycling  of
16                 residuals into useful  products, use of waste heat from some operations as
17                 input into others, changes in production processes to reduce residuals or
18                 to yield residuals that are more easily recycled or used, and other means
19                 of more nearly "closing the loop" have been implemented. The implication
20                 of all this is that the relationship between energy and other inputs and
21                 outputs, including residuals that stress the environment, is not immutable.
22                 Hence, environmental policy actions that penalize the production of
23                 residuals need not lead to proportionate reductions in output of desired
24                 goods. Instead, in reaction to regulatory  requirements, the system may,
25                 and most likely will, adjust at multiple points and these adjustments will
26                 lower costs (compared to a response that just dealt with the residual "at
27                 the end of the pipe") and may have other beneficial effects.  Therefore,
28                 regulatory instruments should encourage private sector decision-makers
29                 to look for means to  lower harmful residuals while leaving them with
30                 maximum flexibility in how this is accomplished.
31
32                        Recent experience with industrial ecology suggests that a priori
33                 estimates of the social costs of reducing  residuals that may harm the
34                 environment are often excessive. As the tenets of industrial ecology
35                 become more widespread and widely adopted, the promise is that

                                             5-46

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1                 improved environmental quality can be bought at a lower cost in lost
 2                 goods and services production than previously thought. This suggests
 3                 that estimates of the costs of reducing harmful residuals should
 4                 incorporate realistic opportunities for systemic adjustment. Further,
 5                 estimates of benefits from such action should be holistic as well.  For
 6                 example, they should take into account any expected reductions in
 7                 environmental harm from input or process changes that would result from
 8                 the regulatory action.
 9
10                       The conclusion is that the optimal level of environmental quality
11                 may be underestimated if too narrow and static a picture of response to
12                 regulatory pressures is taken as the basis for regulatory decisions.
13                 Consequently, opportunities for increases in social welfare may be lost.
14                 Research on the efficiencies and cost-savings that may result from the
15                 implementation  of industrial ecology principles would improve our ability to
16                 estimate accurately net benefits for different management scenarios.
17
             In principle, the economic valuation framework can be  used to formulate the
19     choices required to incorporate ecological and other factors into social decisions.
20     However, traditional economic analysis techniques have limitations of methodology and
21     completeness. This section has identified some of these limitations and sketched some
22     of the research directions, that if successful, may relax them.  A key theme of this
23     section is that provision of a rich and diverse set of information to  decision makers can
24     contribute to improved decision making.
25
26      5.7  Summary and Conclusions
27
28           Because many scientists, decision-makers, and other stakeholders feel that
29     economic analysis methods undervalue environmental resources, the EPA Science
30     Advisory Board established a Subcommittee as part of its Integrated Risk Project to
31     explore ways to advance the  dialogue that would lead to  a more complete framework
32     for estimating the benefits of  environmental management actions.  In meeting its
33     charge, the Valuation Subcommittee considered: 1) the environmental management
34     decision context for valuation; 2) the nature of value and values; 3) the economic
35     concept of value; 4) the importance  of deliberative processes to valuation; and  5)

                                            5-47

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1     additional approaches to environmental valuation.  This chapter is the result of the
 2     Subcommittee's discussions on this issue. The Subcommittees conclusions for each of
 3     the areas considered suggest the following:
 4
 5                 Context is critical to valuation. Important contextual issues include: 1) the
 6           type of environmental management action being considered; 2) societal norms
 7           and objectives that underlie the action being considered; 3) the spatial and
 8           temporal scope of the issue; 4) the identity of stakeholders; and 5) the range of
 9           technological and behavioral options available to control the problem.
10
11                 The norms, principles, and objectives of society represent its values,
12           provide the reason for addressing environmental problems, and provide
13           guidance in identifying the goals of environmental management actions. The
14           benefit of an action reflects its contribution toward the achievement of those
15           goals. Determining benefits cannot be separated from the need to reach
16           agreement on goals.
17
18                 Fairness, sustainability, and other values, as well as economic efficiency,
19           form goals that should be recognized and explicitly addressed in environmental
20           resource decision-making.
21
22                 For decision-making in a governmental context, environmental valuation is
23           recognized as an anthropocentric exercise. People care for things for different
24           reasons and in different ways. Things are valued in themselves (as ends) and
25           instrumentally (as a means for the satisfaction of other ends). People care about
26           some things because they desire them or because they contribute to their
27           happiness, and other things have value because people judge them to be good
28           or important in their own right.
29
30                 Some analysis techniques reveal people's preferences for one or another
31           thing through empirical study. Other techniques attempt to elicit preferences
32           through interaction with individuals. Careful and unbiased interactions with
33           individuals permit us to leam the qualitative dimensions of their value
34           statements.
35

                                             5-48

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1                  A basic premise of welfare economics is that economic values are based
 2           on individuals' preferences and that people know their preferences.  Where
 3           individuals are ignorant of the roles of ecological functions in contributing to
 4           valued service flows, it may be necessary to use experts' knowledge of the
 5           functioning of environmental systems as an input in the valuation process. In
 6           principle, the economic valuation framework can be utilized to define and
 7           measure the economic values of changes in the functions and services of
 8           ecological  systems that affect individuals' welfare either directly or indirectly.
 9           However, this framework may be difficult to implement in practice where the
10           relationship between the function and the service flow to  individuals is indirect or
11           subtle.
12
13                 Economic approaches, while consistent and coherent frameworks for
14           valuation, are not mechanisms for producing "the answer" since they may omit
15           trans-economic values that may be important, may include some elements that
16           are difficult or impossible to estimate, and may employ preference elicitation
17           processes that are incomplete.
1
19                 Not all benefits or costs can be easily quantified. There is a need for
20           qualitative methods to provide valuation measures for decision-makers to use in
21           solving complex ecosystem problems. Care must be taken to assure that
22           quantitative factors do not dominate important qualitative factors.
23
24                 To be most useful, valuation approaches, methods, and information
25           should be made as explicit as possible. These analyses are best used to inform,
26           but not dictate, decisions related to environmental protection policies, programs,
27           and research.
28
29                 Deliberation among interested parties, though not required in all or
30           perhaps most decision-making situations, is often useful for eliciting information
31           on important goals, on "best judgment" on uncertain relationships, on measures
32           for judging goal attainment, and on value tradeoffs that must be made in
33           decision-making and implementation. Depending upon the sufficiency of
34           knowledge available to address the problem at hand, and the extent of
35           agreement that might exist on values involved in the issue, one of the following

                                            5-49

-------
        IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

  1            types of deliberation may be useful during decision-making: oversight
  2            deliberation, stakeholder deliberation, expert deliberation, or integrated
  3            deliberation.
  4
  5                  The relationship between energy and other inputs and outputs (including
  6            residuals that stress the environment) is not immutable. The optimal level of
  7            environmental quality may be underestimated if too narrow and static a picture of
  8            response to regulatory pressures is taken as the basis for regulatory decisions.
  9            Consequently, opportunities for increases in social welfare may be lost.
10            Research is needed to determine how best to align the regulatory process with
11            the opportunities for minimum cost compliance. It is also needed to determine
12            how dynamic adjustments can be incorporated better into the process of
13            estimating the prospective costs and benefits of possible actions.
14
15                  Decision-making must proceed in a timely manner.  Even though there is
16            much uncertainty associated with the factors important to decision-making,
17            decision-makers cannot wait for certainty to make and implement decisions. A
18            process of adaptive management can allow decisions to be made and
19            implemented in the face of uncertainty. Knowledge gained from implementation
20            experience and research provides feedback with which to revisit and revise past
21            decisions.
22
23                  It is important that the total U.S. investment in environmental protection be
24            applied effectively and efficiently.  Rather than simply address stressors in order
25            of some scientifically derived risk ranking, multiple risks should be considered
26            simultaneously, using  both quantitative and qualitative factors, to determine the
27            optimal way to address the full suite of problems that are in need of attention.
28
29            The Subcommittee's work confirms the challenges and complexities of
30     environmental valuation exercises and the environmental  management actions based
31     on those values. The Subcommittee recommends that expanded, rich,  and complex
32     processes be employed to characterize environmental values more fully.  These
33     processes should involve interaction and deliberation among scientists, decision
34     makers, and other stakeholders in order to identify goals,  define endpoints to
35     characterize those goals,  clarify important but uncertain relationships, and to implement

                                             5-50

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1     approaches to achieve those goals. Processes should include on-going dialogue and
 2     adjustment, and they should consider: 1) why people care for the things they do; 2) the
 3     appropriate use of deliberative processes to elicit preferences and the rationale for
 4     them; 3) economic valuation frameworks to define and measure the economic value of
 5     changes in environmental systems functions and services affecting individual welfare;
 6     and 4) presentation of available physical or other quantitative measures, or qualitative
 7     descriptions of the effects of alternative actions when costs and benefits are not fully
 8     captured by monetary measures. The Subcommittee recognizes that environmental
 9     valuation remains a craft embedded in political processes and that much additional
10     research is needed in all areas that are important to estimating the benefits and costs of
11     environmental management action.
12
13
                                           5-51

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1     5.8 References Cited
 2
 3     Arrow, K., M.L Cropper, G.C. Eads, R.W. Hahn, LB. Lave, R.G. Noll, P.R. Portney, M.
 4           Russell, R. Schmalensee, V.K. Smith, .and R.N. Stavins. 1996. Is there a role
 5           for benefit-cost analysis in environmental, health, and safety regulation?
 6           Science. Vol. 272:221-222.
 7
 8     Balch, G.I. and S. Sutton. 1995. Putting the first audience first: Conducting useful
 9           evaluation for a risk-related government agency. Risk Analysis. 15(2): 115-125.
10
11     Braden, J.B. and C.D. Kolstad, eds. 1991. Measuring the Demand for Environmental
12           Quality. Amsterdam, The Netherlands: North-Holland.
13
14     Brown, G.M., Jr. 1990.  "Valuation of Genetic Resources," jnG.H. Orians, G.M. Brown,
15           Jr., W.E. Kunin, and J.E. Swierzbinski, eds., The Preservation and Valuation of
16           Biological Resources. Seattle, WA: University of Washington Press.
17
18     Brown, M.T. and R.A. Herendeen.  1996.  Embodied energy analysis and EMERGY
19           analysis: a comparative view. Ecological Economics. 19:pp. 219-35.
20
21     Brown, P. 1997. Analytical and ethical frameworks for thinking about the  environment
22           and soil. Presentation at a Workshop on Human Perceptions of the Soil during
23           the Inaugural Meeting of the International Center for Soil and Society. Univ. of
24           Maryland. July 1997.
25
26     Brown, T.C., G.L Peterson, and B.E. Tonn. 1995. The values jury to aid natural
27           resource decisions.  Land Economics. 71 (2):250-260.
28
29     Brulle, R. J. 1995. Environmental discourse and environmental movement
30           organizations: A historical and rhetorical perspective on the development of U.S.
31           environmental organizations. Sociological Inquiry. 65.
32
33     Campbell, D.T. 1969. Reforms as experiments. Amer. Psychologist. 24:409-429.
34
35     Cannon, F.  1993.  Economic Growth and the Environment, Economic and Business

                                            5-52

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1           Outlook, Bank of America Economics-Policy Research Department, (415)
 2           622-3215.
 3
 4     Chess, C. and B.J. Hance. 1994. Communicating environmental risk: Ten questions
 5           managers should ask. New Brunswick: Center for Environmental
 6           Communications. Rutgers University.
 7
 8     Chess, C., M. Tamuz, and M. Greenberg. 1995. Organizational learning about
 9           environmental risk communication: The case of Rohm and Haas' Bristol Plant.
10           Society and Natural Resources. 8:57-66.
11
12
13     Chess, C., T. Dietz, and M. Shannon. 1998. Who should deliberate when? Human
14           Ecology Review. 5(1):45-48.
15
16     Christensen, N.L., A.M. Bartuska, J.H. Brown, S. Carpenter, C. D1 Antonio, R. Francis,
17           J.F. Franklin, J.A. MacMahon, R.F. Ross, D.J. Parsons, C.H., Peterson,  M.G.
             Turner, and R.G. Woodmansee. 1996.  The report of the Ecological Society of
19           America committee on the scientific basis for ecosystem management.
20           Ecological Applications 6(3): 725-747.
21
22     Cleveland, C.J., R. Costanza, C.A.S. Hall, and R. Kaufmann. 1984. Energy and the
23           United States Economy: A Biophysical Perspective.  Science. 255:890-897.
24
25     Commission on Risk Assessment and Risk Management. 1997. Framework for
26           environmental health risk management. Presidential/Congressional Commission
27           on Risk Assessment and Risk Management.
28
29     Costanza, R. 1980. Embodied energy and economic valuation. Science. 210:1219-24.
30
31     Costanza, R., S.C. Farber, and J. Maxwell. 1989. The valuation and management of
32           wetland ecosystems. Ecological economics. 1:335-361.
33
34     Costanza, R. And C. Folke.  1997. Valuing ecosystem services with efficiency,
35           fairness, and sustainability as goals.  Pp. 49-70 in: G. Daily (Ed.) Nature's

                                          5-53

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1           Services: Societal dependence on natural ecosystems.  Island Press.
 2           Washington, D.C. 392 pp.
 3
 4     Crosby, N., J. Kelly, and Schaefer. 1986. Citizen panels: A new approach to public
 5           participation.  Public Administration Review.
 6
 7     Dasgupta, P.  1990.  "Commentary," in Gordon H. Orians, Gardner M. Brown, Jr.,
 8           William E. Kunin, and Joseph E. Swierzbinski, eds., The Preservation and
 9           Valuation of Biological Resources. Seattle, WA: University of Washington Press.
10
11     Dietz, T. 1987. Theory and method in social impact assessment. Sociological Inquiry.
12           57:54069.
13
14     Dietz. T. 1988. Social impact assessment as applied to human ecology: Integrating
15           theory  and method. Pp. 207-227 in Human Ecology: Research and Applications.
16           Ed. R.  Boarden, J. Jacobs, and G.R. Young. College Park,  Maryland. Society for
17           Human Ecology.
18
19     Dietz. T. 1994. What should we do? Human Ecology and Collective Decision-making.
20           Human Ecology Review. 1:301 -309.
21
22     Dietz, T. and  A.  Pfund. 1988. An impact identification method for development program
23           evaluation. Policy Studies Review. 8:137-145.
24
25     Dietz, T., P.C. Stem, and R.W. Rycroft. 1989. Definitions of conflict and the
26           legitimation of resources: The case of environmental risk. Sociological Forum.
27           4:47-70.
28
29     English, M., A.K. Gibson, D.L. Feldman, and B.E. Tonn.  1993. Stakeholder
30           involvement: Open processes for reaching decisions about the future uses of
31           contaminated sites. Knoxville: Waste Management Research and Education
32           Institute.
33
34     Fiorino, DJ. 1990. Citizen participation and environmental risk: A survey of institutional
35           mechanisms. Science, technology and Human Values. Pp. 226-243.

                                            5-54

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999—Do Not Cite or Quote

 1     Freeman, A. M., III. 1993. The Measurement of Environmental and Resource Values:
 2          Theory and Methods. Washington, DC: Resources for the Future.
 3
 4     Freeman, A. M., III. 1997. "On Valuing the Services and Functions of Ecosystems," in
 5          R. David Simpson and Norman L Christensen, Jr., eds., Ecosystem Function
 6          and Human Activities: Reconciling Economics and Ecology. New York: Chapman
 7          and Hall.
 8
 9     Gosselink, J.G., E.P. Odum, and R. M. Pope. 1974. "The Value of the Tidal Marsh,"
10          Center for Wetlands Research, Louisiana State University, Baton Rouge,
11          LSU-SG-70-03.
12
13     Goulder, L.H. and D. Kennedy. 1997. "Valuing Ecosystem Services: Philosophical
14          Bases and Empirical Methods," in Gretchen C. Daily, ed., Nature's Services:
15          Social Dependence on Natural Ecosystems. Washington, DC: Island Press.
16
1"7     Government Accounting Office. 1995. Superfund: EPA's community relations efforts
            could be more effective. Washington. GAO B-247753.
19
20     Grunig, J. 1992. Excellence in public relations and communication management.
21          Hillside: Lawrence Erlbaum.
22
23     Hall, B. 1994. Gold and Green, Institute for Southern Studies. P.O. Box 531, Durham,
24          N.C. 27702.
25
26     Hardin, G. 1968. The tragedy of the commons. Science. 162:1243-1248.
27
28     Harwell, M.A., J.F. Long, A.M. Bartuska, J.H. Gentile, C.C. Harwell, V. Myers, and J.C.
29          Ogden. 1996. Ecosystem management to achieve ecological sustainability: the
30          case of South Florida  Environmental Management 20(4) :497-521.
31
32     Hettige, H., R.E.B. Lucas, and D. Wheeler. 1992. The Toxic Intensity of Industrial
33          Production: Global Patterns, Trends, and Trade Policy, American Economic
34          Review, V.82, no. 2, May, pp. 478-81.
35

                                           5-55

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1     Heywood, V.H., I. Baste, and K.A. Gardner. 1995.  Introduction, in: Global Biodiversity
 2           Assessment V.H. Heywood, Exec. Ed. Cambridge Univ. Press. Published for the
 3           United Nations Envir. Programme. New York.
 4
 5     Kathlene, L. And J. Martin. 1991. Enhancing citizen participation: Panel designs,
 6           perspectives, and policy formation. Journal of Policy Analysis and Management.
 7           10:46-63.
 8
 9     Kelman, S.  1981. Cost-Benefit Analysis: An Ethical Critique. Regulation, vol. 5, no. 1,
10           pp. 33-40.
11
12     Mageau, M.T.,  R. Costanza, and R.E. Ulanowics. 1995. The development and initial
13           testing of a quantitative assessment of ecosystem health. Ecosystem Health.
14           1(4):201-213.
15
16     Mazmanian, D.A. and J. Nienaber.  1979. Can organizations change? Washington,
17           D.C.  Brookings Institution.
18
19     Meyer, S.M. 1992. Environmentalism and Economic Prosperity: Testing the
20           Environmental Impact Hypothesis, Project on Environmental Politics and Policy,
21           Mass. Inst. of Technology, Bldg./Room E38-628, Cambridge, MA. 02139.
22
23     National Research Council. 1996. Understanding Risk: Informing Decisions in a
24           Democratic Society. P.C. Stem and H.V.  Fineberg (Eds). Washington, D.C.
25           National Academy Press.
26
27     Odum, H.T. 1996. Environmental Accounting: EMERGY and decision-making. John
28           Wiley, New York, NY.
29
30     Ostrum, E. 1990. Governing the Commons.  The Evolution of Institutions for Collective
31           Action. Cambridge Univ. Press. New York.  Pp. 280.
32
33     Pearce, D. And D. Moran. The Economic Value of Biodiversity. Earthscan Publications
34           Ltd. London. Pp. 172.
35

                                            5-56

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999—Do Not Cite or Quote

 1     Perrings, C., et al. 1995. The Economic Value of Biodiversity. Jjr  V. Heywood, ed.
 2          Global Biodiversity Assessment. Cambridge, UK: Cambridge University Press.
 3
 4     Renn, O., T. Webler, and P. Wiedemann. 1995. Fairness and competence in citizen
 5          participation: Evaluating models for environmental discourse. Dordrecht: Kluwer
 6          Academic Publishers.
 7
 8     Rosener, J.B. 1981. User-oriented evaluation: A new way to view citizen participation.
 9          Journal of Applied Behavioral Science.
10
11     Sagoff, M.  1988. The Economy of Earth: Philosophy, Law, and the Environment. New
12          York: Cambridge University Press.
13
14     Schlager, E., W. Blomquist, and ST. Tang. 1994. Mobile flows, storage, and self-
15          organized institutions for governing common pool resources. Land Economics.
16          70(3):294-317.
17
       Sen, A. 1995. Rationality and Social Choice. American Economic Review, vol. 85, no.
19          1.PP-24.
20
21     Sexton, K. and B.S. Murdock, Eds. 1996. Environmental policy in Transition. Making
22          the Right Choices.  The Minnesota Series in Environmental Decision-making.
23          Volume I. Center for Environmental and health Policy. School of Public Health.
24          University of Minnesota.  Minneapolis, MN. 109 pp.
25
26     Shabman, LA., and S.S. Batie. 1978. The Economic Value of Natural Coastal
27          Wetlands: A Critique. Coastal Zone Management Journal, vol 4., no. 3, pp,
28          231-237.
29
30     Templet, P.M. 1995.  The Positive Relationship between Jobs, Environment and
31          Economy: An Empirical Analysis and Review.  Spectrum. Spring 1995 Issue,
32          pp.37-49.
33
34     Templet, P.M. and S. Farber. 1994. The Complementarity Between Environmental and
35          Economic Risk: An Empirical Analysis, Ecological Economics. 9:153-165.

                                           5-57

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1     Thomas, J.C. 1995. Public participation in public decisions: New skills and strategies for
 2           public managers. San Francisco: Josey Bass.
 3
 4     Toman, M.A.  1997.  "Ecosystem Valuation: An Overview of Issues and Uncertainties,"
 5           in R. David Simpson and Norman L. Christensen, Jr., eds., Ecosystem Function
 6           and Human Activities: Reconciling Economics and Ecology. New York: Chapman
 7           and Hall.
 8
 9     Ulanowicz, R.E. 1986. Growth and Development: Ecosystems Phenomenology.
10           Springer-Verlag, New York.
11
12     U.S. Department of Energy. 1996. Site-specific advisory board initiative: Evaluation
13           survey results. USDOE.
14
15     U.S. Environmental Protection Agency. 1990. Environmental Investments: The Cost of
16           a Clean Environment. Washington, D.C.
17
18     U.S. Man and the Biosphere Program. 1994.  Isle au Haut Principles: Ecosystem
19           Management and the Case of South Florida  (HDS 002), US Department of
20           State, US MAB, Washington, DC.
21
22     Vaughan,  E. 1995. The significance of socioeconomic and ethnic diversity for the risk
23           communication process. Risk Analysis. 15(2): 169-179.
24
25     Wilson, J.  and L. Dickie. Parametric Management of Fisheries: An Ecosystem Social
26           Approach, 1995.  in Property Rights in a Social and Ecological Context, eds.
27           Susan Hanna and Mohan Munasinghe, The Beijer Institute of Ecological
28           Economics and the World Bank, Stockholm and Washington, pp. 153-166.
                                           5-58

-------
PART IV INPUTS TO ENVIRONMENTAL DECISION-MAKING:
          RISK REDUCTION APPROACHES

-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote


  PART IV—INPUTS TO ENVIRONMENTAL DECISION-MAKING: RISK
                       REDUCTION APPROACHES

                                  Preface

      Previous chapters have discussed approaches for assessing and comparing
environmental risks and the implications of various risk reduction goals and choices on
economic or other measures of welfare. These analyses are important tasks in the
Problem Formulation and Analysis and Decision-making phases of the I ED framework.
An additional set of analyses required by the framework, however, relates to the overall
design and selection of risk management scenarios to address "the problem." This
latter set of analyses, including identification and selection of specific risk management
approaches, is the focus of the following chapter. The IED framework calls for
preliminary consideration of risk reduction options during Problem Formulation, but the
most in-depth consideration of options occurs during Phase II when the environmental
problem, or set of problems, and associated environmental goals, have been defined.
Options are analyzed with regard to their potential to reduce single and multiple risks of
concern, associated costs, sustainability, equity, and other criteria specified by the IED
participants.

      The Risk Reduction Options Subcommittee (RROS) was composed of
individuals with broad experience in the various engineering and non-engineering
approaches to risk reduction so that a wide range of management tools would be
considered. The group was charged with developing a process for identifying the most
effective risk management approaches for a variety of types of risk problems that might
confront a decision-maker. To meet this charge, the RROS formed subgroups to
consider options analysis from three different perspectives: 1) for a stressor-based
problem (e.g., ozone); 2) for a geographically based problem (e.g., risks associated with
an urban area); and 3) for a media-based  problem (e.g., contaminated groundwater).

      The resulting 10-step method for identifying,  screening, and selecting risk
reduction options is the subject of Chapter 6, and is summarized in Figure 6-1. The ten
steps are:
      Step 1: Define the problem

                                     IV-i

-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

      Step 2: Develop background information
      Step 3: Identify the spectrum of risk reduction options
      Step 4: Establish screening criteria
      Step 5: Screen  potential risk reduction options
      Step 6: Evaluate remaining risk reduction options
      Step 7: Refine the options
      Step 8: Select an option
      Step 9: Document the process
      Step 10: Quantify option effectiveness

In relation to the I ED framework, Steps 1 through 4 are primarily the province of the
Problem Formulation Phase, whereas Steps 5 through 9 relate primarily to the Analysis
and Decision-making Phase. As noted in the I ED framework, however, some iteration
is inevitable between Phases I and II. The final step, Step 10, occurs during the
Implementation and Performance Evaluation Phase of the IED framework.

      The first step in the 10-step method is to define the problem. The chapter
articulates the importance of a clear problem statement, including specific goals for
what/whose risk is to be reduced and by how much. Clear environmental goals, with
explicit statement  of the relationships and the potential tradeoffs among goals, are the
foundation from which objectives for the risk reduction program can be derived.  In
addition to information on the nature of the risks (e.g., location, severity, specific
subpopulations or endpoints affected, and exposure media), the chapter discusses
other apects of Problem Formulation that are particularly critical for the design of a risk
reduction program.  For example, the desired state or goal must be stated in
measurable terms and methods for measuring improvements toward the goal (i.e.,
indirect measures such as stressor levels or environmental outcome measures such as
reduction in a specific  adverse health effect) must be specified.  It is also essential to
identify likely  constraints on possible solutions (e.g., budget, time, jurisdictiona!
authority, and legality) that will help establish screening criteria for the risk reduction
options.

      The chapter then goes on to describe the remaining 9 steps in the methodology,
from development of background information and identification and screening of
potential risk  reduction options, to selection of an option that best meets the screening

                                      IV-ii

-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

and selection criteria and documentation of the decision.  Various decision analysis
methods are discussed briefly, including benefit/cost analysis (described more fully in
Chapter 4), matrix qualitative ranking, and multi-attribute decision procedures. The last
step in the RROS method, quantifying the effectiveness of the risk reduction program,
is analogous to the Performance Evaluation phase of the IED framework. The topic of
performance evaluation and report cards is discussed further in Part V of this report.

      In addition to describing a comprehensive decision process for selecting risk
reduction options, the chapter places special emphasis on the types of environmental
problems best  or least suited to different risk reduction tools.

      Although the chapter often refers to "the problem solver," integrated
environmental  decision-making requires that many types of participants  communicate
and contribute to problem formulation and solution. In addition, much of the chapter is
written around  the assumption that the problem and the constraints surrounding its
solution have already been specified and that the task at hand is to choose among
approaches for risk reduction. It is worth emphasizing, however, that the IED
framework recognizes the important role of preliminary options analysis  during Problem
Formulation, so that problems are not just defined on the basis of risk and our interest
in reducing risk, but also take into account our ability to reduce risk within likely
constraints.

      The methodology for developing and selecting  risk reduction options is
discussed primarily in terms of reduction of a particular risk, e.g., one associated with a
particular chemical stressor or source.  The chapter notes that although it is often best
to address risks with a combination of risk reduction tools, this is often not done
because of inadequate information on the multiple sources of a stressor and their
relative contribution to total risk. Extending the methodology to an integrated problem
set containing  risks from multiple stressors (e.g., those experienced by a particular
community) will further increase the complexity of the  analysis. Thus, it  will be
important to aggregate or disaggregate the problem set (e.g., using "root cause" or
"common source/common pathway" analysis) so that analysis of risk reduction options
is more manageable. The chapter also notes that the complexity of the  analysis,
including the screening and selection of options, increases greatly with an increasing
number of criteria against which options are to be "optimized."

                                     IV-iii

-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
      In summary, when applying the 10-step process to an integrated problem set,
consisting of multiple stressors, sources, or endpoints, the following issues arise:

      1) Aggregation is helpful in identifying multiple stressors that may have a
      common set of risk reduction options with the objective of selecting a set of
      options that will provide the most risk reduction for the set of risks being
      analyzed. Practically speaking, it will not be possible to optimize risk reduction
      over all stressors of concern considered at once. For this reason, screening and
      aggregation of stressors into manageable subsets should be driven by analysis
      of common aspects of stressors,  roots causes, and/or activities.

      2) In order to compare risk reduction across sets of options and sets of risks,
      and to evaluate risk reduction per unit cost, it is critical to have a common
      measurement of risk or common denominator for all risks.  In many cases,
      comparisons of risks and risk reduction under different scenarios will involve
      unlike risks (e.g., cancer risk in humans vs. chronic health effects vs. effects on
      wildlife populations), even where those risks have a common "root cause" (e.g.,
      a single stressor or source).

      3) Uncertainty of the analysis is likely to increase as a broader set of options is
      considered for more than one stressor. Sources of this uncertainty include the
      relative contribution of different stressors/sources to the total aggregate risk; the
      effectiveness of combined options for reducing aggregate risk; and the
      benefit/cost, equity, or other tradeoffs involved  in addressing groups of risks.
      Some types of options are easier to implement when uncertainty is high (e.g.,
      communication and education and environmental management systems, rather
      than regulations/enforcement); the extent of uncertainty associated with a
      decision may affect the balance of options selected.
                                      IV-iv

-------
   IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
                                  Quote
              CHAPTER 6.  RISK REDUCTION OPTIONS

                          TABLE OF CONTENTS


6.1  Introduction and Approach  	  6-1

6.2  Define the Problem 	  6-6

6.3 Develop Background Information	  6-12

6.4 Identify the Spectrum of Risk Reduction Options	  6-16
      6.4.1 Communication/Education	  6-18
      6.4.2 Enforcement  	  6-21
      6.4.3 Engineering	  6-23
      6.4.4 International and Intergovernmental Cooperation	  6-23
      6.4.5 Management Systems	  6-23
      6.4.6 Market Incentives	  6-23
      6.4.7 Regulation	  6-26

6.5  Establish Screening and Prioritization Criteria	  6-33

6.6  Screen and Prioritize Potential Risk Reduction Options 	  6-40

6.7  Evaluate the Remaining Risk Reduction Options 	  6-48

6.8  Optimize the Options	  6-54

6.9  Select an Option 	  6-59

6.10 Document the Process  	  6-65

6.11  Quantify Option Effectiveness	  6-65

6.12 References Cited  	  6-71

-------
         IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
                                         Quote
 1                     CHAPTER 6. RISK REDUCTION OPTIONS
 2
 3
 4      6.1 Introduction and Approach
 5
 6            The focus of this chapter is on the identification and selection of risk reduction
 7      options to solve an environmental problem or set of problems. The methodologies
 8      and thinking in this chapter are a product of the Risk Reduction Options
 9      Subcommittee (BROS) and are articulated in the 10-step decision process that forms
10      the backbone of the discussion.
11
12            EPA, state, and local regulators have embraced the concept of risk-based
13      environmental decision-making, including the need to prioritize environmental
14      problem-solving efforts.  Nonetheless, the RROS recognized that regulatory decision-
15      makers face many constraints in determining risk priorities and in selecting risk
16      management options, including constraints imposed by implementing statutes and by
       Congress. The Subcommittee defined its goal as helping regulatory officials maximize
18      the risk reduction achieved for any fixed amount of resources, whether those
19      resources are public or private. This chapter, therefore, suggests a methodology for
20      achieving improved risk reduction outcomes, rather than recommending risk reduction
21      solutions for specific environmental problems.
22
23            The goal of the RROS was to develop a methodology that would: (1) identify an
24'     appropriate set of options for any given human health or environmental problem, (2)
25      create a reproducible  and documented decision process where risk reduction
26     decisions are optimized based on explicit decision-maker policy choices, and (3)
27      produce risk reduction decisions transparent to interested parties.
28
29            At the first meeting, the RROS explored different approaches to risk reduction
30     available to regulatory or community decision-makers. The approaches fall into one of
31      the following seven categories which may not be completely independent.
32
33            a) Communication and education approaches, including information or data
34            dissemination to consumers and communities, options that provide technical

                                            6-1

-------
           IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
                                           Quote
 1            assistance and technology transfer to emission producers, and options that
 2            provide information to govemment(s);
 3
 4            b) Enforcement approaches, including government-implemented options to
 5            obtain improved compliance with existing regulations as well as options that
 6            involve non-government parties in enforcement activities;
 7
 8            c) Conventional and innovative engineering approaches, including options
 9            where technology for achieving pollution prevention, waste minimization, or end
10            of pipe control is communicated, encouraged, or mandated;
11
12            d) International and intergovernmental cooperation, including both formal and
13            informal cooperation with other U.S. and non-U.S. government agencies and
14            non-government organizations;
15
16            e) Environmental management systems, including options that focus on
17            communication, regulation, and/or market incentives for improved management
18            systems which will reduce risks/liabilities and improve compliance.
19            Management systems cover the areas of setting performance expectations,
20            defining roles and responsibilities, identifying and managing risks, ensuring
21            communication and training, developing implementing procedures, and
22            assessing/measuring performance.
23
24            f) Market incentives, including a broad range of options that can be
25            implemented independently or in combination with  regulation, communications,
26            or enforcement options. These options can include marketable permits,
27            deposit/refund systems, fees and taxes, subsidies and tax credits, differential
28            regulatory requirements based on performance, government procurement
29            approaches, etc.; and
30
31            g) Regulation including mandated technologies, performance levels, risk levels,
32            specific product composition, specific activity requirements, specific information
33            disclosure requirements, and so forth.
34

                                            6-2

-------
          IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
                                          Quote
 1            Knowledgeable experts in each of the seven areas were invited to share their
 2      experience with these potential risk reduction tools.  Each expert was asked to
 3      address where the option has been used, what lessons have been learned about
 4      where it works well and where it doesn't, existing obstacles to its use, biggest
 5      advantages and disadvantages to the approach, the future directions for the
 6      approach, and risk areas that could optimally benefit from the approach. These
 7      presentations and opportunities for interaction provided Subcommittee members with
 8      a common base of information for methodology development.
 9
10            In order to pursue methodology development, the RROS began by dividing into
11      three subgroups. Each subgroup selected a different approach to a human
12      health/environmental problem.  One subgroup selected ozone, a stressor-based
13      approach. Another subgroup selected Elizabeth, New Jersey, an urban
14      geographic-based or location based approach. The third group selected groundwater,
15      a media-based approach.  All three groups had the same goal; to develop a risk
16      reduction decision methodology that policy-makers could utilize and interested parties
       would find transparent.  By selecting examples of three very different types of problem
18      sets, it was anticipated that a universal methodology that could work for any human
19      health/environmental problem would result.
20
21            Once the three subgroups completed their work, a single common  ten-step
22      methodology was jointly developed by the full RROS.  That methodology was tested
23      against  a fourth environmental area of concern, namely, "How can EPA best
24      implement the tolerance-setting requirements laid out in the Food Quality Protection
25      Act of 1995?"
26
27
                                            6-3

-------
          IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
                                           Quote
 1           The overall risk reduction methodology devised by the RROS contains ten
 2     steps, summarized in Figure 6-1, and discussed in detail in subsequent sections. The
 3     ten steps are:
 4
 5            1. Define the problem,
 6            2. Develop background information,
 7            3. Identify the spectrum of risk reduction options,
 8            4. Establish screening and prioritization criteria,
 9            5. Screen and prioritize potential risk reduction options,
10            6. Evaluate the remaining risk reduction options,
11            7. Optimize the options,
12            8. Select an option,
13            9. Document the process, and
14           10. Quantify option effectiveness.
                                             6-4

-------
          IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
                                           Quote
 2
 3
 4
 5
 6
 7
 8
 9
10
        Define Problem:
              Devetpp Background Information
                                                 External
                                               ^Information^
                    Identify All Risk Reduction Options
                       u
                                                                 Experience]
                     Establish Screening Criteria
                                 .Screen and Prioritize Options
                                    U
Evaluate Remaining Options




Optimize Options
        Rgure 6-1. Risk Reduction Option Selection Methodology
      The RROS methodology has multiple applications. First, if a decision-maker is
faced with a specified environmental area of concern, he/she can apply the
methodology to obtain the best risk reduction outcome for that problem.  A single area
of environmental concern may be expressed in terms of: a) individual stressors, such
as heavy metals or particulates; b) activities, such as energy use or transportation; c)
specific industries; or d) geographic areas, such as the Gulf of Mexico or the greater
Chicago area. The RROS methodology provides a formal approach for evaluating
                                             6-5

-------
           IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
                                            Quote
  1      various risk reduction options so that the decision-maker can select an outcome that
  2      maximally achieves goals while considering existing constraints.
  3
 4            Another application of the methodology is to select a single risk reduction
 5      outcome that will maximize cost-effective risk reduction for a number of different
 6      environmental areas of concern.  For example, the decision-maker may want to
 7      identify which type of risk reduction outcome would maximize risk reduction for a set of
 8      stressors. Or, the decision-maker may want to identify which type of risk reduction
 9      option would maximize risk reduction for certain activities, such as energy utilization,
10      transportation, and agriculture.
11
12            A third application of the methodology is to compare the potential risk reduction
13      outcomes for a several different environmental areas of concern and select the one
14      risk reduction option that maximally achieves goals while considering constraints.
15
16            The comparison of risk reduction options is contingent upon the ability to
17      quantify,  or measure, risk. Qualitative risk evaluations do not work well in this
18      methodology.
19
20       6.2  Define the Problem
21
22            Before attempting to solve a problem it is necessary to develop a clear
23      definition of the problem.  People often minimize the importance of defining a problem
24      for which they seek a solution.  They desire to get right to problem solving, without first
25      making sure that they understand completely and precisely what needs to be
26      corrected or improved. In the risk reduction context, failure to adequately define the
27      problem can result in a sub-optimal solution or possibly a solution that does not work
28      at all.
29
30            The assumption is that at this stage the problem has been identified. The
31      challenge is simply to choose among various approaches to reducing risk for a
32      specific problem.  It would be easy, therefore, to jump directly into problem solving or
33      to a solution.
34

                                             6-6

-------
          IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
                                           Quote

 1            This section describes the need for a clear articulation of the problem.  It also
 2     discusses the four components of a clear problem statement, including the use of dis-
 3     aggregation (breaking the problem apart), the need for stakeholder input, and the
 4     need for coordination among and within governments.
 5
 6           Today's environmental problems are extremely complex and interrelated. For
 7     example, exposure to lead may come from many sources, some of which are
 8     interdependent.  Inhalation of volatilized lead in air, ingestion of lead-contaminated
 9     drinking water, and ingestion by children of paint chips containing lead are but a few of
10     the sources of lead exposure. Exposure to  pesticides presents a similar range of
11     sources: workplace exposure experienced by manufacturing workers, formulators and
12     applicators; incidental exposures in lawns, golf courses or vegetable gardens;
13     exposure to grocery produce; and consumption of groundwater contaminated with
14     pesticides.
15
16            Given this complexity, pursuit of a solution without understanding precisely
       what needs improvement will be ineffective. For example, does one seek to reduce
18     the lead burden overall, or is the interest in  reducing (at least) the lead exposure of
19     children, who are believed to be more susceptible to its adverse effects?
20     Understanding these issues allows one to focus on the issues that will yield the
21     optimum results. It will allow the identification of the root cause of the problem so that
22     solutions can be tailored accordingly.
23
24            Problem definition is also essential in understanding and articulating the
25     objectives, or improvement goals, for the risk reduction program.  These aspects of
26     the issue must be clearly stated as part of the problem definition as well. It is a
27     tautology to say that the over-arching objective of any risk reduction project is to
28      reduce risk and leave it at that.  In any program designed to reduce a risk, or set of
29      risks, there will be complementary and competing objectives. All must be articulated,
30      and competing goals prioritized, before one can design a program to achieve them.
31
32            The program objectives must also be measurable. For example, an objective
33      for a lead reduction program might be: reduce the blood lead level in children in the
34      U.S. to a level at which adverse effects are not likely to occur. Likewise, a consumer

                                             6-7

-------
           IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
                                            Quote
 1      risk reduction program for residual pesticides might seek to reduce the level of
 2      residual pesticides to a level that minimizes adverse effects based upon science's
 3      current definition of what level is acceptable.
 4
 5            Without a clear understanding of the objective and the process for measuring
 6      progress, it will be impossible to determine whether the program is effective. In other
 7      words, these factors are crucial to defining success. Congress, environmental groups
 8      and others frequently berate EPA for failing to show positive results from its programs.
 9      This stems, in large part, from a lack of adequate criteria for defining program
10      success, or, in the case of some programs, using the wrong measurement criteria.
11      The problem-solver must clearly understand the objectives of the program if he/she
12      hopes to design a successful solution.  Without this, developing a solution becomes a
13      game of "pin the tail on the donkey," in which the donkey's location is not known until
14      the tail (the solution) has been pinned on and the blindfold removed.
15
16            The overall objective of the problem definition phase is to develop a specific,
17      narrowly focused statement of what is wrong with the current state, and to provide a
18      clear understanding of the desired state, or goal. The objective of this phase is to
19      move from a version of reality that defies resolution to a version along the pathway to
20      success.
21
22            To accomplish this, one needs to proceed through the six-step process set forth
23      below.
24
25            a) Identify the elements of the problem context.
26
27            b) Understand and, if possible quantify, the magnitude of the harm.
28
29            c) Identify the goal of the risk reduction program or desired state after
30            implementation.
31
32            d) Identify any constraints, or hard limitations on the solution.
33
34            e) Define the affected population.

                                              6-8

-------
           IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
                                            Quote
 1            f) Narrow the problem to its essence, or core elements
 2
 3            The first step is to identify the elements or variables of a problem context. This
 4      might also be characterized as understanding the nature of the problem. What is it
 5      that needs to be improved? Is it reducing environmental harm, improving health, or
 6      improving visibility, for example.
 7
 8            The second step involves an attempt to quantify, or at least understand
 9      qualitatively, the magnitude of the harm. In addition to these fundamental questions,
10      the problem-solver must understand the characteristics of the problem context. Is the
11      problem site-specific, such as a waste site; or is it broad-based, such as nitrate runoff
12      into estuaries.  What are the sources of the risk and the nature of those sources? Do
13      they involve multiple environmental media, or are they limited to a particular
14      environmental media?  Are they large quantity, discrete sources or small quantity,
15      diffuse sources? What are the relationships, if any, among these variables? Answers
16      to these questions will significantly affect the structure of an effective risk reduction
        program. For example, large quantity, discrete sources are inherently more
18      manageable through direct regulatory approaches than are small diffuse sources.
19
20            Third, while gathering information about the nature of the problem, one should
21      also gather information about the desired outcome. Specifically, by looking beyond
22      the solution to the desired state, the goal of the risk reduction program can  be
23      identified.  This understanding must come early in the process if the  problem is to be
24      defined properly. In addition, to be able to measure progress toward the goal, the
25      desired state must be articulated in measurable terms.
26
27            The fourth step is to identify and articulate any constraints on  the solution.
28      These are necessary to establish the screening criteria for risk reduction options.
29      Some likely solution options constraints include: budget, time, jurisdictional authority of
30      the implementing agency and the legality of the proposed solution. For example, if it
31      is desired to achieve the risk reduction within 5 years, that should be identified as a
32      constraint so that solutions that will take more than 5 years will be eliminated in the
33      screening stage. Jurisdictional authority is one screening criterion that must be used
34      cautiously.  One of the key recommendations of this section is that governmental

                                             6-9

-------
           IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
                                            Quote

  1      agencies coordinate and work more closely with one another.  Therefore, one would
  2      not want to screen out options simply because multiple agencies must be involved.
  3      There are, however, circumstances in which Congress will not have delegated
  4      adequate  authority to any agency to implement a particular solution.  In those cases, it
  5      will be necessary to limit the risk reduction options considered to those within the
  6      scope of?? [DFO note: some text missing here?]
  7
  8            The legality of the proposed solution could become an issue in a variety of
  9      ways.  The proposed solution might not be capable of receiving the necessary
10      licenses or permits, or there might be antitrust concerns, particularly in the context of
11      market-based solutions.
12
13            The fifth step in the process is to define the affected population. For risks that
14      affect human health, this requires a clear understanding of the receptor population.
15      Are the risks imposed on a population in a specific geographical area, near a waste
16      site, for example? Or, do the risks cross geographical boundaries, such as in the case
17      of pesticides? For risks that only affect the environment, what is the affected
18      ecosystem? Who are the users of the  resource?
19
20            Properly defining the problem requires divergent thinking, to insure that all of
21      the variables are explored.  This requires solicitation of inputs from diverse
22      stakeholders and experts, including the affected population and the users of  the
23      natural resource, other governments (multi-national and state) and other agencies,
24      and the potentially regulated community.
25
26            One must balance the desire for broad stakeholder involvement with the
27      recognition that there is an associated process efficiency cost. Allowing for public and
28      other inputs to any process necessarily takes time. For problems to which there is a
29      need or desire to move quickly, it may be necessary to allow limited, or in some cases
30      no, public  input.  For example, cleaning up an oil spill, or addressing an imminent and
31      substantial endangerment under the Superfund Program, requires  expeditious action
32      and therefore may not allow for public input because a lack of action may exacerbate
33      the problem.  Even in the case of problems for which longer-term solutions are
34      appropriate, the cost in time and resources must be considered in determining how

                                             6-10

-------
           IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
                                           Quote
 1      extensive a stakeholder input process to devise.  In general, however, appropriate
 2      stakeholder input strengthens the overall decision-making process. Important
 3      considerations for selecting the type and extent of deliberation for a particular decision
 4      are discussed more fully in Chapter 5.
 5
 6            Beyond simple stakeholder input, it is critically important that the agency
 7      leading the risk reduction effort coordinates with and involves other agencies that may
 8      have overlapping or related jurisdiction over the problem.  It is so often the case today
 9      that one federal agency is not fully coordinated with others when developing programs
10      to solve problems. When that happens, at best sub-optimal solutions are the
11      outcome. At worst, the program is a failure. Therefore, in the course of defining the
12      problem, the lead agency should consult at multiple points in the process with any and
13      all relevant agencies. This should include not only federal agencies,  but state
14      agencies, other national government agencies and international bodies as well.
15
16            After developing a broad array of elements of the problem, the problem solver
        must proceed to the sixth step: identifying those elements that constitute the essence
18      of the problem--tf?e rea/prob/em. At this stage, it is necessary to narrow the problem
19      definition so that it becomes one of manageable dimension and to ask if it is effective
20      to deal with many contributing problems, or if they should be separated into multiple
21      problems.  Moreover, one must consider whether the problem should be addressed at
22      a higher level.
23
24            This presents squarely the issue of aggregation and dis-aggregation of
25      problems, terms describing the process of combining common or hierarchical
26      problems, and, alternatively, separating them into smaller parts.  Aggregation requires
27      that one deal with multiple problems, usually at a higher level. This may be
28      appropriate, particularly if the root cause(s) of the problems are common or related.
29      Alternatively, it may be better to dis-aggregate problems to better understand the root
30      cause(s) of each problem more precisely. Dis-aggregation may also  be necessary to
31      create a problem that is resolvable, as opposed to simply a jumble of confusing
32      issues. Problem aggregation and dis-aggregation may need to be an iterative
33      process. The problem-solver may find that after he/she defines the problem and
34      begins preliminary work on solutions, the problem needs to be dis-aggregated. This

                                            6-11

-------
           IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
                                            Quote

  1      takes him/her back to the problem definition stage, where he/she must formulate a
  2      workable problem definition once again.
  3
  4            Problem definition may require the problem solver to circle through the 6-step
  5      process repeatedly. The result, however, will be an optimum problem definition.
  6
  7            The output of this process should be a clear, focused statement of the problem
  8      that contains the following elements.  First, it should contain a clear statement of the
  9      current situation. This should also contain  a clear articulation of the problem, or what
10      is wrong with the current state.  Second, it should contain an explanation of the
11      desired state.  This must be stated in measurable terms. If possible, the problem
12      definition should also describe the method  or methods for measuring improvements
13      toward the desired state.
14
15      6.3 Develop Background Information
16
17            It is prudent to consider the larger problem in which the risk reduction
18      opportunity is imbedded before thinking about information needed to consider the
19      merits of a specific risk reduction option. For example, if the risk reduction option
20      under consideration is subsidizing the construction of radon resistant homes, the
21      process should begin with an analysis of the radon exposure problem. What has
22      been done to date with the problem within the Agency or within other governmental
23      and non-governmental organizations? Why hasn't the problem been solved or at least
24      substantially reduced?  If previous attempts have been made to reduce the risk, how
25      much did different management options contribute toward meeting the goals?   Did
26      these options take longer than anticipated to accomplish the task? How much did
27      scientific uncertainty contribute to undermining management options?  If it is a new
28      problem, why has it only recently come to attention? What are the key elements of
29      the problem that need to be addressed  before a solution can be devised? What might
30      be learned about the problem in the near future that may change the acceptability of a
31      risk reduction option?
32
33            After considering these contextual questions, the next action is to consider what
34      specific information about toxins, engineering, economics, geography, psychology,

                                             6-12

-------
          IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
                                           Quote
 1     law, and other subjects are necessary to start a process that will lead to assessing the
 2     utility of a specific policy or implementation option. Does this information exist? If the
 3     answer is yes, who has the data? How can it be obtained, and in what forms? Is it
 4     reliable?  If it is not highly reliable because of the way it was gathered or recorded,
 5     and/or if the scientific basis supporting it is shaky, are there available time and
 6     financial resources to gather high quality data? Or are there other data that can
 7     substitute for the best data? Can professional experience or the judgment of informed
 8     members of the public substitute for what is uncertain, or at least reduce the level of
 9     uncertainty in the data?
10
11            At the end of the process, one should know:
12
13            a) How the problem came to exist,
14
15            b) How it has changed over time and across space,
16
              c) Who has taken policy or implementation actions to date and what their
18            results have been,
19
20            d) What are the alternative risk reduction options,
21
22            e) Where the most promising places in the casual sequence to introduce risk
23            reduction options exist,
24
25            f) What the opportunity is for a multi-stage hybrid risk reduction option, and
26
27            g) What forces are likely to constrain each potential risk reduction option.
28
29
                                            6-13

-------
           IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
                                           Quote
 1      There are two general categories of information needed to screen options.
 2
 3            Risk Management Information
 4            Spatial and temporal scale of the problem.
 5            Risk reduction objective.
 6            List of identified options.
 7            Screening factors and constraints.
 8            Decision-maker(s) identified and committed to the project
 9            Prospective implementer of selected options.
10
11            Options Performance Information
12            Chance of success of an option within a given set of constraints.
13            Chance of success of composite options within a given constraint (including
14                  synergistic effects)
15
16            Information on risk management scenarios would be collected during the
17      stages that precede screening. This information is integrated into the work done on
18      problem definition, development of key background information, and identification of a
19      full range of risk reduction options.
20
21            Options performance information is required to estimate whether an option is
22      likely to be successful within a given set of constraints. Such a determination can be
23      based on one or more of the following information bases; expert opinion, previous
24      experience, and predictive methods including mathematical modeling.  Using a site-
25      specific case as an example, one may consider various treatment options (nested in
26      the engineering category of options) for reducing contaminant migration source terms
27      at a contaminated site. As determined by the kinetics of physicochemical/biological
28      phenomena and other implementation factors that apply to a given technology, a
29      minimum remediation time is usually required. Thus, the use of quantitative  methods
30      to assess whether or not the time for implementing the technology meets the time
31      requirement stated in the risk reduction objective would be useful.  Quantitative
32      methods for such assessments can be deterministic or probabilistic, but they can be
33      structured to estimate the reliability of an option.
34

                                            6-14

-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
                              Quote
                                6-15

-------
           IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
                                           Quote
 1      6.4 Identify the Spectrum of Risk Reduction Options
 2
 3            The protections delivered by the U.S. environmental management system
 4      derive largely from technology-based or harm-based emission limitations although the
 5      system also relies, to a lesser extent, on other tools such as market incentives and
 6      taxation. The past five years have seen a contentious debate about the structure and
 7      composition of this system.  Critics of the system charge that the tools commonly used
 8      for emission limitations, "command and control regulations," are inefficient from the
 9      standpoint of delivering benefits in an optimally cost-effective way. Specifically, the
10      system has been charged with:
11
12            a) focusing too much on end-of-the pipe;
13
14            b) ignoring the fact that some dischargers can reduce pollution more cheaply
15            than others and that some areas have more serious pollution problems than
16            others;
17
18            c) blocking opportunities available to achieve  emission limitations using
19            cheaper alternatives for compliance; and
20
21            d) imposing more stringent controls on new plants than on existing ones, a
22            dichotomy that creates incentives to keep older, less efficient, higher polluting
23            plants in service (Glicksman and Chapman, 199?).
24
25            "Command and control" approaches have also been faulted with causing a
26      serious misallocation of resources, in that the regulations may be directed at trivial
27      risks while more significant risks sometimes remain unaddressed (FN85 in
28      Glicksman?).
29
30            Critics of conventional regulatory approaches have generally argued for the
31      substitution of economic incentives, such as marketable permits, effluent taxes or
32      charges, or subsidies, to deliver more efficient pollution control (citatation in
33      Glicksman, ??). Although these economic tools have been used occasionally, in the
34      past EPA has been charged with  being generally biased against new approaches,

                                             6-16

-------
          IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
                                           Quote

 1      such as market based approaches, that could conceivably achieve better reductions
 2     more cost effectively (Report on Strategic Options Subcommittee, Appendix C, RR,
 3     1990).
 4
 5            There is little dispute that there are severe inherent limits in the extent of
 6     protection that the existing regulatory structure can deliver. The large number of
 7     processes and substances that must be regulated, the burden of proof to justify
 8     proposed regulations, the information requirements to design the regulations, the time
 9     and cost of issuing and enforcing permits all conspire against solving the still daunting
10     environmental problems that we face with conventional approaches.  Many are
11     concerned that the system leads to excessive delay in addressing important
12     environmental hazards.
13
14            To date, the debate about delivering environmental protection has focused
15     nearly exclusively on the strengths and weaknesses of specific risk reduction tools,
16     such as technology-based standards or economic incentive systems.  It has not
       focused on the sometimes sharp differences in the nature of specific environmental
18     problems to be solved using  these tools.  These differences among types of problems,
19     once properly considered, suggest that no single risk reduction tool is superior across
20     the board in addressing all types of environmental risks. No tool so frequently holds
21     superior promise that the regulator should assume it can be applied to the problem at
22     hand without considerable analysis.  From this perspective, then, other approaches
23     must be considered, including market based incentives, non-technology-based
24     standards, more public disclosure, or less "command and control" - without
25     references to the types of problems remaining for the Agency to address - do not
26     engage the debate where the problem really lies, which is in matching specific
27     environmental risks with the  best possible risk reduction options.
28
29            The Subcommittee firmly believes that each environmental risk should first be
30     examined for certain critical elements prior to selecting a  suitable set of risk reduction
31     options.  In addition, our review of the application of various risk reduction options
32      reveals that the problems EPA has experienced in effectively controlling some
33      environmental risks may stem more from insufficient information regarding types and
34      number of sources of emissions contributing to the problem than  from inherent

                                            6-17

-------
           IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
                                            Quote
 1      limitations of the option selected. This inadequate understanding of the sources can
 2      result in a sub-optimal or ineffective solution.  For example, Amoco Oil Company did a
 3      study of its refinery in Yorktown, Virginia that led to a widely publicized critique of the
 4      limits of the conventional air emissions regulatory system (reference??).  The major
 5      finding of the Amoco study was a very large source of fugitive benzene emissions that
 6      had been unknown both to the regulators and Amoco prior to the investigation.
 7      Without complete understanding of the sources to be addressed at this facility,
 8      however, no risk reduction option in EPA's toolbox would effectively address
 9      emissions at this plant. Thus, a complete understanding of the sources to be
10      regulated is at least as important, if not more important, than the ultimate selection  of
11      the risk reduction option.
12
13            Many environmental risks are best addressed with a combination of risk
14      reduction tools.  Under such an approach, certain contributors to the total risk might be
15      best controlled under one approach, while other sources contributing to the same risk
16      could be best addressed by another.  Because the types and magnitude of
17      contributors to certain risks are so poorly understood, complex, or diverse, the
18      application of a mixture of risk reduction tools to decrease an environmental hazard is
19      uncommon.
20
21            This section describes the full range of risk reduction options currently available
22      to achieve environmental and public health protection.  Unlike several other reviews
23      (OTA: Environmental Policy Tools, 1995; McGarity 46 Law and Contemp. Probs 159,
24,     Summer 1983) addressing this topic, this approach places special emphasis on the
25      types of environmental problems (or elements of problems) best or least suited to a
26      particular risk reduction tool wherever possible.  The chapter attempts to identify the
27      core problems where the tool works well, and where the tool is less effective. Table 6-
28      1 at the end of the section summarizes this information.
29
30       6.4.1 Communication/Education
31
32            Information reporting is a communication/education instrument that requires
33      regulated entities to provide specified types of information to an agency which in turn
34      makes it available to the general public. Typically, the  information concerns facility

                                            6-18

-------
          IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
                                            Quote
 1     emissions or operating characteristics, e.g. spill plans, pollution prevention plans, etc.
 2     Information reporting (for public use, as opposed to compliance use) is based on the
 3     theory that disclosure of polluting activities by regulated entities will raise public
 4     concern, which will then lead them to respond to this concern by changing behavior.
 5     The scheme takes advantage of a regulated entity's desire to be a good neighbor and
 6     responsible corporate citizen, as well as their fear of adverse publicity or possible loss
 7     of sales. In addition, because the public's heightened awareness of polluting activities
 8     increases the possibility of new regulations, it provides another type of incentive for
 9     regulated entities to pursue pollution reduction strategies.
10
11            Information reporting provides less direct assurance than many other tools that
12     goals will be met, because it does not mandate overall pollution limits or place an
13     explicit price on pollution. Furthermore, information reporting will change the behavior
14     of different regulated entities differently. Even if a high percentage the regulated
15     community reports, it is hard to know how many will change behavior to reduce
16     pollution and whether there will be some locations with disproportionately few
       reductions.
le
19            Information reporting programs are being used with increasing frequency. They
20     were sparked by the Bhopal, India disaster, which alerted many in the United States to
21     the need to know more about the chemicals used and stored at industrial facilities.
22      Now, the major federal disclosure requirements are found in Section 313  of the
23      Superfund Amendment and Reauthorization Act (SARA), the Toxics Release
24      Inventory (TRI), which calls for certain manufacturing  facilities to submit annual reports
25     on the amounts of listed toxics chemicals released (routinely or accidentally) into the
26      environment.  At the state level, California has created Proposition 65, California's
27      safe Drinking Water and Toxic Enforcement Act.  It requires public warning of the
28      potential cancer or reproductive effects of 542 listed chemicals either emitted or
29      present in products. This policy has resulted in the installation of warning signs
30      throughout the State of California, including in grocery stores and restaurants.  There
31      remain questions as to whether all the signage has truly reduced exposure of
32      Califomians to toxins.
33
34            Information reporting is most likely to be effective in situations where:

                                              6-19

-------
           IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
                                           Quote
 1            It is possible to convey the information in a simple and accurate manner
 2
 3            The regulatory target is a product where consumer preferences are driven by
 4            environmental concerns, as opposed to exclusively price, quality, performance,
 5            and convenience or other resource(s) for purchasing the product;
 6
 7            The link between the product and environmental damage is clear;
 8
 9            The risk, though relatively easy to mitigate, is not on a scale to have been
10            prioritized by government programs; or
11
12            Hazards are not due to multiple exposures or toxic hot spots, or where hazard
13            is due to combined factors.
14
15            Technical assistance, a communication/education approach, consists of
16      government programs designed to educate private entities and the public to make
17      better environmental choices. It seeks to achieve environmental goals by increasing
18      the understanding of pollution problems and potential solutions.  It may take the form
19      of manuals and guidance, training programs and materials, information
20      clearinghouses, hot lines, facility evaluations, and technology R&D (OTA, p. 138).
21      Most technical assistance services are provided at no, or minimal, cost to the user,
22      and participation in programs is typically voluntary.
23
24            Some technical  assistance programs have been developed in response to
25      congressional mandates, while EPA and other agencies have  initiated others. Section
26      507 of the Clean Air Act,  for example, requires states to establish Small Business
27      Stationary Source Technical and Environmental Compliance Assistance Programs.
28      Others have been initiated by EPA or other agencies to help implement mandated
29      environmental programs. For example, section 319 of the Clean Water Act calls for
30      states to manage diffuse non-point sources of water pollution. EPA and the U.S.
31      Dept. of Agriculture have developed extensive guidance documents describing best
32      management practices that non-point sources might use to control their pollution.
33      Other programs do not respond directly to statutory mandates but are derived from the
34      general objective of improving environmental quality. For example, the EPA Green

                                            6-20

-------
          IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
                                           Quote

 1      Lights Program conducts energy audits and makes specific recommendations for
 2     energy efficiency in exchange for an agreement from participants to install the
 3     recommended lighting systems.
 4
 5           Technical assistance is most effective where polluting entities are not
 6     knowledgeable about the environmental consequences of their actions or alternatives
 7     available to improve their current environmental practices.
 8
 9       6.4.2  Enforcement
10
11           Liability provisions require that polluters pay for the damage they cause by
12     requiring compensation for environmental damage. It is enforced two ways: by
13     common-law theories such as negligence or nuisance, or by statute, particularly the
14     Comprehensive Environmental Response, Compensation, and Liability Act
15     (CERCLA). For either type of liability, a successful claim typically requires an
16     established causal link between the harm and the pollution, which has been traced
       back to its source.
18
19            Liability differs from regulation in that it engages the responsible parties
20     after-the-fact. Liability rules establish the price of environmental damage, while
21      regulations seek to control the activities that create such damage. As such, liability is
22     a crude instrument for affecting behavior; by itself it cannot ensure that the appropriate
23     amount of care is exercised in all circumstances (Shavell, 1984, in Percival, p. 134).
24      Despite this, the specter of potential liability is widely thought to encourage pollution
25      prevention and responsible waste management ethics, because the dollar amounts
26      involved can be huge, much larger than the costs of preventive behavior in the first
27      instance.  As of September 1994, for example, Exxon had already spent $3.4 billion to
28      clean up the spill and settle suits from the Prince William Sound oil spill.
29
30            Liability also provides incentives for environmental auditing and other
31      self-appraisals, in order to gauge the potential financial exposure and correct
32      problems before they grow (OTA, p. 125). As a policy tool, liability is particularly
33      attractive when a private actor is in a better position than the government to assess
34      the risks of its activity and to  determine the level of care to exercise.  On the other

                                             6-21

-------
           IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
                                             Quote
 1      hand, regulation is more efficient than liability for preventing accidents caused by
 2      those likely to escape responsibility for their actions; this factor usually weighs in favor
 3      of regulation as a means for deterring environmental damage because of the difficulty
 4      of proving causation under common law.
 5
 6            Several statutes provide liability schemes to address environmental problems.
 7      CERCLA provides for strict  retroactive liability for cleanup of hazardous waste dumps.
 8      The Oil Pollution Act provides strict liability for natural resource damages, and third
 9      party damages caused by petroleum spills. The Clean Water Act makes responsible
10      parties liable for cleanup costs for a spill of hazardous substances into surface waters,
11      capping liability at $50 million unless the discharge was the result of negligence  or
12      willful misconduct.
13
14            Liability is an effective incentive for environmentally beneficial behavior only to
15      the degree that the decision-maker is concerned that he/she will be held liable.
16      Difficulties involved in proving a causal link between a particular action and damage to
17      a particular plaintiff has limited the effectiveness of this tool for addressing many
18      environmental hazards. Liability works best where:
19
20            The contamination is traceable back to one source (harm that is widely
21            dispersed or difficult  to trace will not spark adequate incentive to reduce risks);
22
23            There is sufficient information on the extent of harm that is caused by a
24            substance or activity to establish a causal link between the harm and the
25            pollution;
26
27            The impacts of the contamination are sufficiently concentrated to make a claim
28            worthwhile to the injured  party; or
29
30            The party responsible for the damage  has the capacity to provide
31            compensation for the full amount of harm the actions produce. (If an activity
32            can cause more damage than the actor is capable of repaying, fear of liability
33            will not provide sufficient incentive for private investment in an efficient level of
34            precautions.)

                                              6-22

-------
          IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
                                           Quote
 1           If liability insurance is available (as a matter of fact, it is often required to ensure
 2     that compensation will be available for victims of environmental damage), the
 3     existence of this insurance can reduce responsible party incentive to prevent damage
 4     in the first instance.
 5
 6       6.4.3 Engineering
 7
 8           Engineered solutions are often the first ones considered.  The propensity to
 9     jump to this conclusion and the proven versatility of engineered solutions has led to
10     the proliferation of end-of-pipe treatment technologies.  Only the recent development
11     of pollution prevention strategies has lessened their popularity. Engineered solutions
12     are often very costly and take years to implement.
13
14       6.4.4 International and Intergovernmental Cooperation
15
16           International and intergovernmental cooperation lends itself to those issues that
1      have large  geographical impact.  Global warming is an excellent example whereby the
18     cooperation of many world organizations is required to control the emission of
19     greenhouse gases that eventually leads to global warming.
20
21       6.4.5  Management Systems
22
23           Management systems integrate options to create effective solutions.  It starts
24     with establishing organizational principles and ends with the attainment of corporate
25     goals. This option is gaining favor as entities realize that it establishes the basis for
26     getting the most from existing resources.
27
28       6.4.6  Market Incentives
29
30           Under tradable emissions , a market incentive approach, the government first
31     sets a level of aggregate emissions over a specified time period, consistent with
32     environmental goals by issuing only the number of permits corresponding to that level.
33     The total allowable emissions are then allocated to individual sources through
34     government-issued permits. Unlike conventional permit systems, however, each

                                            6-23

-------
           IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
                                            Quote
 1      regulated entity can buy and sell permits from others.  An entity might choose to do so
 2      if the relative costs of emissions control make it less expensive to buy (or profitable to
 3      sell) a permit from another entity. In theory, trading would continue until the cost of
 4      controlling another pound of pollution is the same for all entities and is equal to the
 5      cost of a permit.
 6
 7            The most well known applications of emissions trading include: Sulfur dioxide
 8      (SO^ reduction for acid rain control and oxides of nitrogen (NOJ reduction for urban
 9      ozone in Los Angeles. It was considered but rejected to control volatile organic
10      carbon (compounds?) (VOC) emissions (ozone precursors) in the Los Angeles basin.
11      Trading has also been proposed under the NPDES system for effluent discharge
12      trades, in particular for phosphorus reductions in the Dillon Reservoir (Colorado) and
13      the Tar-Pamlico River Basin (North Carolina) and for biochemical oxygen demand
14      (BOD) reduction in the Fox River (Wisconsin). These watershed trading programs are
15      under-utilized to date because there are not disproportionate costs across  sources
16      associated with the reductions. Wetland mitigation banking, where agencies authorize
17      distribution of wetlands in return for promise of future enhancement of other wetlands,
18      is also another trading venue. A lack of certainty that  the appropriate regulatory
19      agency will approve a "trade" is thought to be the "rate limiter" to trading.
20
21            The reductions must be  quantifiable to an acceptable level of certainty.  This
22      generally means that the regulatory agency must have the capacity to measure
23      pollutant levels in question at the source or as they apply across the system. This
24      includes measuring the baseline pollution level and changes from baseline to allow a
25      source to generate tradable credits. Because establishing actual emission reductions
26      for mobile sources is more difficult than for most stationary industrial sources, it is less
27      clear that emissions trading schemes will achieve intended results where mobile
28      sources are involved.
29
30            There must be a large number of sources with  significant variations in control
31      costs.  If everyone's cost to control emissions is the same, then emission credits will
32      not be a bigger bargain than reducing emissions at the facility. The  pollution problem
33      must not involve hot spots. Otherwise, the trading system may result in inadvertently
34      disproportionate exposures. The program must be evasion proof. Actions that

                                             6-24

-------
          IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
                                           Quote

 1      produce credits must be enforceable. The program is especially valuable where
 2     expected control costs are very high; this will enhance interest in a more
 3     cost-effective approach. The emissions should have a well-defined relationship to the
 4     problem being solved.
 5
 6           With pollution charges, a regulated entity is required to pay a set dollar amount
 7     for each unit of pollution emitted or disposed. Sources are free to choose whether to
 8     emit pollution and pay the charge, or to pay for the installation of controls to reduce
 9     the charge.  Pollution charges do not set a limit on emissions or production. However,
10     if they are set high enough, and if they vary according to the amount of pollution
11     produced, they can provide significant financial incentives to reduce or eliminate
12     environmentally harmful behavior.
13
14           Although charges have commonly been levied against sources as a
15     revenue-raising measure, such as with  permit fees, pollution charges set at a level
16     sufficient to change behavior have not been used extensively in the United States.
       Two exceptions are solid waste management and the charge that was levied on
18     chlorofluorocarbons (CFCs) during their phaseout. Under typical solid waste
19     schemes, companies pay charges that rise as waste volume rises, while most
20     households pay flat fees unrelated to the amount of waste generated. Volume-based
21     charges have been applied to household waste in about 100 jurisdictions across the
22     country as well, however, and they are believed to have been successful in reducing
23     volumes of waste generated in many of these instances (OTA, p. 123).
24.
25            During the CFC phaseout, users were required to pay a charge per pound of
26      CFC, multiplied by an ozone-depleting  factor. The charge was increased in
27      subsequent years. CFC production decreased much more rapidly than originally
28      anticipated, and many attribute this rapid decline to the effectiveness of the CFC tax
29      (OTA, p. 122). Because the tax was used in conjunction with a mandatory phase-out,
30      it is difficult to ascertain the relative importance of each tool in achieving the  overall
31      success of this program. Clearly, however, a combined approach of charges and
32      phase-out can be effective in reducing environmentally harmful behavior.  It is
33      important to note that the CFC taxes collected by the U.S. government went to general
34      revenue and were directly allocated to  environmental programs/improvements.

                                             6-25

-------
           IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
                                            Quote
 1            Subsidies are policy instruments that provide financial assistance to entice
 2      entities to change their behavior or help entities having difficulty complying with
 3      requirements.  Generally, subsidies take the form of grants or low- no-interest loans to
 4      municipalities and public entities and preferential tax treatment for private entities.
 5      The Clean Water Act, for example, had a long-standing construction grant program
 6      that enabled municipalities to comply with Clean Water Act requirements for
 7      wastewater treatment plants. This grant program was phased out by the 1987
 8      amendments to the Act and replaced with a state revolving loan fund. The Clean Air
 9      Act also authorizes several grant programs. For example, section 105 provides for
10      grants to state and local governments to implement air pollution control programs.
11
12            For many years, private companies have been allowed to take accelerated
13      depreciation on investments aimed at reducing water pollution.
14
15        6.4.7 Regulation
16
17            Harm-based standards are established on the basis of what is required to
18      achieve health or environmental protection goals.  EPA typically establishes these
19      standards by determining the amount of the pollutant in the ambient environment that
20      will meet a health or environmental goal set by Congress. This determination involves
21      scientific judgments as to the extent to which different concentrations of the pollutant
22      cause harm. After the Agency establishes an acceptable concentration, a
23      mathematical model is used to calculate an overall allowable pollution load for the
24      region with which this acceptable concentration will not be exceeded. EPA or the
25      state then apportions an acceptable pollutant concentration or loading among the
26      individual sources that it has identified. Harm-based standards are expressed in
27      facility permits as emission rates for the source (mass per unit time period), as a
28      concentration of pollutant in a source's discharge, or as a percentage reduction in
29      emissions from a source.
30
31            As a practical matter, EPA's ability to craft a policy that meets a harm-based
32      goal is contingent on a thorough understanding of the sources contributing to the
33      pollution load from the facility or region. Often, the Agency has insufficient knowledge
34      of the number, types, and magnitude of contributors to the problem. Bubble emission

                                             6-26

-------
          IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
                                           Quote

 1     and integrated permitting initiatives have been developed to overcome the lack of
 2     knowledge in this area by providing flexibility to regulated entities to alter control levels
 3     for various sources within a facility as long as overall limitations are achieved.  (The
 4     3M integrated permit for its Hutchinson, Minnesota facility, for example, allows the
 5     company to shift emission controls among sources within the facility as long as
 6     aggregate VOC control levels are satisfied.)  However, neither bubbling not integrated
 7     permitting have been widely used to date, again because of the inadequate
 8     knowledge of sources to be "bubbled" or because of monitoring and accountability
 9     difficulties.
10
11           Harm-based standards are  notoriously difficult and time-consuming to set
12     because of analytical uncertainties and gaps in available data about both the scientific
13     basis for concern and the demography of the sources contributing to the problem.
14     Uncertainties inherent in predicting the effects of different patterns and levels of
15     environmental releases also  pose  big problems.  Harm based standards also require
16     extensive data on existing ambient pollutant concentrations and health effects, which
       often are not available.
18
19            Nonetheless, harm-based standards, along with technology/design standards,
20     are the most heavily used environmental policy tools in the U.S. Often, the two are
21     used together. The Clean Water Act provides harm-based standards for "water
22     quality limited streams", where industrial sources must comply with stricter pollution
23     control than would otherwise be necessary in order to achieve improved water quality
24,    in a highly negatively impacted area.  The Clean Air Act National Ambient Air Quality
25      Standards (NAAQS), also exemplify harm-based standards.
26
27            In contrast to harm-based standards, technology-based standards derive from
28     the level of control technology Congress expects pollution sources to implement, such
29      as "reasonably available control technology." Technology based standards can be
30      viewed as the opposite of health-based standards, because instead of asking what is
31      needed to protect health or the environment, they ask what is possible to do (Percival,
32      p. 146). Most commonly, these standards are set as a performance level for the
33      regulatory target to meet, without specifying how this should be done (performance
34      standards). They also can specify how a certain facility or piece of machinery should

                                             6-27

-------
           IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
                                             Quote
 1      be designed and engineered (design standards). EPA develops both design and
 2      performance standards, although some of its performance standards become de facto
 3      design standards.  This can occur where EPA's performance regulation is established
 4      on a specific technology and, in responding to the target, regulated entities decide the
 5      safest course to compliance is to install a specific design technology. EPA's New
 6      Source Performance Standards under the Clean Air Act and its various effluent
 7      limitations under the Clean Water Act are examples of performance standards, but
 8      they are often criticized for limiting acceptable technologies in practice to those that
 9      EPA has used as a basis for developing the performance standards  (Percival,  p. 150).
10
11            Disadvantages technology-based standards  are that they: a) may cause the
12      expenditure of more or less resources than necessary, b) assume state of technology
13      is knowable, and c) discourage innovation.
14
15            Fixed emissions standards based on current assessments of technology are
16      widely faulted for the absence of incentives for further innovation.  Laws that force
17      industry to develop innovative solutions by adopting standards more stringent than
18      those attainable by current-available technology overcome this important limitation.
19
20            Both the Clean Water Act and the Clean Air Act call for technology-forcing
21      regulations. The effluent limitations in the Clean Water Act, which are based on "best
22      available technology economically achievable," look to transfer state-of-the-art
23      practices in one industry to another  in order to achieve reductions beyond what is
24      currently in practice. The  Clean  Air  Act Amendments? of 1990 contains a broad effort
25      to force technology developed for mobile sources, including better tailpipe controls,
26      cleaner fuels, and cleaner engines than existed at the time of passage of the Act.
27
28
29
30
31
32
33
34
35
36
"Congress's initial approach to technology-forcing regulation was to engage automakers in a
high-stakes game of chicken. Title II of the 1970 Clean Air Act mandated emission standards
for automobiles that were "a function of the degree of control required, not the degree of
technology available today.' The Act required automobile manufacturers to slash vehicle
emissions by 90 percent in order to be able to continue selling cars in the United States after a
specified deadline, subject to a one-year extension. Not surprisingly, the automobile
manufacturers fought hard to convince EPA to extend the deadline on  the ground that the
necessary technology for reducing vehicle emissions was not yet available. After EPA
Administrator William Ruckelshaus shocked the industry by refusing to extend the deadline,
                                              6-28

-------
   IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
                                    Quote
the auto industry convinced a court to order him to reconsider. In May 1973, EPA agreed to
relax the standards for two years. Meanwhile, GM had a breakthrough in the development of
catalytic converter technology, which promised to be a boon to fuel efficiency and
performance but would also meet the emission standard required in the Act.  Six weeks after
EPA extended the deadline, GM announced that all of its 1975 models would be equipped
with the devices. "I came out of the whole exercise with catalytic converters as a
technological optimist in what the industry can do," said Walsh, the EPA's top auto pollution
expert in the 1970's and now a private consultant. "When you give them a challenge, they
meet it." (Percival, p. 168-171)	
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11      Disadvantages of technology-forcing regulations are that they a) may compel too
12      much monies to be spent in a narrow area and b) require a tremendous amount of
13      information.
14
15            A product ban either prohibits or restricts the manufacture, distribution, or use
16      of a substance identified as a problem to human health or the environment. It is most
17      seriously considered where product use is sufficiently damaging that zero use is a
1"      desirable outcome. Product bans may be imposed prior to the product's sale  and use
        in commerce, through various pre-market product approval programs that seek to
20      prevent excessively risky products from reaching the marketplace. More publicly
21      visible bans concern themselves with substances already on the market, such as lead
22      or asbestos. Whereas the burden of proof lies with the manufacturer to show that a
23      product is safe for pre-market approval, the burden of proof lies  with EPA to show that
24      a product should be banned once it is already on the market.
25
26            Product bans and restrictions focus on the commodity itself rather than on
27      polluting byproducts from its manufacturing. As a result, they are used primarily
28      where the hazard is the commodity itself.  The U.S. has banned lead in gasoline and
29      paint, asbestos, PCBs, CFCs, and numerous pesticides, including DDT, Aldrin, and
30      Dieldrin.
31
32            Product bans are most effective where a single substance or activity is causing
33      a particular environmental problem. They are not effective where the environmental
34      problem has multiple or complex causes,  such as low dissolved oxygen
35      concentrations in urban rivers.
                                     6-29

-------
           IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
                                            Quote

  1             A product ban is best used where all of the uses of a product pose
  2      unacceptable risks. Otherwise, considerable analytical resources must be set aside to
  3      enforce a narrower ban or prevent unauthorized use.
  4
  5             Product bans have superb potential to induce technological innovation by
  6      stimulating rapid research on substitutes. However, they are risky when the
  7      consequences of a failure to develop an alternative are very high (such as, if we were
  8      to contemplate the banning of the internal combustion engine to address urban smog).
  9      Their success hinges on the development of genuinely safer alternative products;
10      otherwise the substitute will pose its own hazards.
11
12             With a challenge regulation, responsibility for solving a problem is shifted from
13      government decision-makers to the sources themselves. Under this approach, the
14      government establishes targets necessary to solve an environmental/public health
15      problem, with a timetable for implementation.  The targets are defined for multiple
16      sources, at the industry sector rather than the individual facility level. These sources
17      are given the collective responsibility for designing and implementing a program that
18      meets the targets.  The government specifies an alternative program or sanction that
19      will be triggered if progress towards the targets is not achieved.
20
21             Challenge regulations have not yet been extensively adopted in any country
22      and have not been used to date in the United States.  EPA's voluntary 33/50 Program
23      (reference?), however, is an example of a pilot program to test this approach.  The
24      most widespread use abroad has been in establishing producer responsibility for
25      various forms of wastes to encourage source reduction and recycling. In the German
26      Green Dot program, the government established a regulatory approach outlining
27      industry's obligations to take back packaging from customers. However, it then gave
28      industries the opportunity to establish an alternative recycling program of their own for
29      meeting the targeted rates.
30
31             Since industry designs the implementation plan, this tool is not well suited to
32      problems with localized concerns about the effects of the pollution.  Thus, it should not
33      be used to  reduce pollutants or solve problems for which exposures vary widely
34      across locations.

                                             6-30

-------
          IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
                                          Quote

 1            The challenge is unlikely to be met if it is difficult to fairly allocate responsibility
 2     across the target group. Competition among firms may make it difficult to develop an
 3     implementation plan.
 4
 5           Some identified options may not be based on technology. They could be policy
 6     options. Performance data are rather sparse on non-technological options, and
 7     models constructed to estimate very high uncertainties and controversies may
 8     characterize their reliabilities within a given constraint. Issues related to the
 9     performance of a variety of both technological and policy-level options have been
10     discussed by Davies and Chou (1992), Levin (1990), IINEPA (1996), Norman and
11     Keenan (1996), C and EN (1996), CHEM NEWS (1996), OTA (1995), USEPA (1990)
12     and UN (1993). Performance information on composite options (i.e	) is very
13     sparse.
14
                                            6-31

-------
            IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
                                                 Quote
 1
 2
 3
 4
Table 6-1.  Conditions Under Which Options Are Most and Least Effective
 5

 6
 7
 8
 9
10
11
12
13
14'
15
16

17
18
19


20

21
22
    • Communication/Education
   SaSiiKi
      -Tech. Assist.
                               • Regulated entity is not
                                 knowledgeable
     Engineering
     Management Systems
                                                              Non-materiel issues
                            B SiS^^^K^^^^Sl
                                                           Weak organizational structure
—££._*- ,—^y^, g .*K..-f.irv^-i/.ffs. ^ ^..-...j^-.^.


   -Tradable Emissions
                                 ! •"^•<*S-;vj:r"--. • •- ^v~^ly;~'™~"~JVT"*Ts*''i?L"*v-'""";^'~;*! •^-*~^T-y^^*-v^,t?;T:yy" T* ^*)yt' "' ly^T^^^~^^ff?TS^'*^y|*^

                                  • High control cost
                                  • Multiple sources w/ variable
                                    control capability
                                  • Quantifiable emissions
                                  • Fixed sources
                                  • Enforceable

                                   • Single substance is cause
                                   > All uses have unacceptable
                                    outcome
                                                           • Multiple sources
                                                           • Few or no alternatives
                                                  6-32

-------
          IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
                                           Quote
 1      6.5  Establish Screening and Prioritization Criteria
 2
 3           The simplest way to evaluate risk reduction options would be to reduce the
 4      consequences of specific options to a single measure of effectiveness. (In fact, this is
 5      precisely what a classic cost-benefit analysis of a regulatory option seeks to do.)
 6      However, it is not easy to achieve consensus on reducing to a single measure the
 7      wide range of views and legitimate perspectives that come to bear in formulating risk
 8      policy. An alternate approach is to reduce the consequences of a risk reduction
 9      option to performance on as small a set of criteria as possible.  The Subcommittee
10      concluded that five basic criteria could be used to evaluate risk reduction options:
11      environmental effectiveness, cost, equity, workability, and flexibility. With regard to
12      the first two criteria, we have concluded that it is more transparent to separately
13      consider environmental benefits and environmental costs, rather than to combine
14      them into a single measure of net benefits.  Ideally, performance in each of these
15      groupings should be summarized along a single dimension, such as lives-saved or
1P      dollars-of-cost.

18      Environmental effectiveness encompasses the variety of criteria that have to do with
19      the environmental impacts/physical benefits. For the most part, the issue here is to
20      what extent does a regulation accomplish the environmental protection goals that
21      originally motivated the regulation?  The focus in on the meaningful environmental
22      consequences of a regulation, rather than simply meeting some arbitrary ambient
23      concentration goal.  Another consideration  in this category is the likelihood that the
24      environmental goals will be met. As an example, the NAAQS and associated State
25      Implementation Plans for urban ozone would suggest that environmental goals are
26      well served by this regulatory approach. In  reality, however, in many cities the goals
27      are often not met. The economic version of this criterion (discussed in Chapter 4) is
28      efficiency, defined as that level of risk reduction which maximizes the surplus of
29      benefits over costs. Therefore, an aggregate measure of environmental benefits is
30      desirable to help compare the performance along this dimension with performance on
31      other evaluative criteria.
32
33      Cost is the second grouping and includes a number of factors. Of primary interest are
34      the direct and indirect costs associated with a regulation, excluding the monetized

                                            6-33

-------
           IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
                                            Quote
 1      value of the environmental benefits that may result from the regulation (which are
 2      accounted for under the previous category?).
 3
 4      Equity deals with the distribution of costs and benefits from the regulation; i.e., who
 5      bears the cost and who benefits. This might involve the distribution of costs and
 6      benefits to economic groups (e.g., rich vs. poor), ethnic groups (e.g., blacks vs.
 7      whites), or different regions of the country (e.g., Midwest vs. west). Clearly any
 8      regulation that involves concentrated costs and benefits will have very different
 9      political characteristics than a regulation with more dispersed costs and benefits.
10
11      Workability deals with how easy it might be to initiate and implement a regulation.
12      Some regulations have a great deal of administrative burden.  Some regulatory
13      approaches are so new as to make their success highly uncertain. Some regulatory
14      approaches are very familiar to regulators and polluters so as to be very workable.
15
16      Flexibility refers to how easy a regulation is to change when new information emerges
17      or to adapt to new circumstances.
18
19            There are two purposes for the evaluative criteria discussed here.  One is to
20      help fine-tune a given option.  For instance, if an emission charge has been proposed,
21      there are many choices as to exactly how it will be implemented, what pollutants will
22      be included, what the level of the charge will be, and what kinds of enforcement will be
23      included. Using the five criteria, the option in question  can be adjusted, tailored, and
24      honed to do as well  as possible against each.  This results in an option that performs
25      as well as possible.
26
27            The second purpose for the five evaluative criteria is to compare several
28      different approaches to solving a particular environmental problem.  Each approach
29      must be fine-tuned.  The only question remaining is which approach is the top
30      candidate for implementation.
31
32            In order to facilitate the comparison of options, it is important to summarize
33      performance for the criteria along as few dimensions as possible. This may be most

                                             6-34

-------
          IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
                                           Quote
 1      easily done for cost, using a dollar metric. It can also be done for the other criteria,
 2     though perhaps not so easily.  For instance, the performance of an option on equity
 3     grounds could be measured on a scale of 1 to 10, with guidelines to govern the use of
 4     particular ratings. On environmental effectiveness, results could be summarized in
 5     terms of statistical lives saved or the willingness-to-pay of the population to achieve
 6     the level  of environmental performance, relative to some baseline.
 7
 8           A  more extensive discussion of each of the five evaluative criteria follows.  This
 9     discussion is not intended to cover all aspects of each but to give an idea of the
10     breadth of issues that are embodied in the criteria.
11
12           Environmental effectiveness summarizes how well a risk reduction option
13     performs in terms of its environmental outcomes, which obviously is an important
14     aspect. It is preferable to define environmental performance in physical or health
15     terms (lives saved, species protected), rather than in terms of administrative process
•"*     measures (e.g., the likelihood of meeting a regulatory target such as no more than
       three violations of an ambient standard in a year). Outcome measures are not always
18     possible, however, since the physical effects may be difficult to quantify.
19
20     Consider the following issues with respect to environmental effectiveness:
21
22           a) How does expected performance compare to environmental goals?  If the
23           goal is to reduce morbidity and mortality by a certain  amount, has this goal
24           been achieved? If the goal is to reduce stressor concentrations by a certain
25           amount, has this goal been achieved? Often there will be no explicit or
26           quantitative environmental goal. One may simply wish to reduce pollutant
27            levels as much as possible (or justified). In this case, there is no explicit end-
28            point.
29
30            b) What are the quantifiable physical gains from the regulation? For example,
31            what is the reduction in morbidity and mortality? What are the goals in
32            recreation opportunities (with cleaner rivers, for example)?  What are the
33            maintenance goals from a cleaner ambient environment? What is the nature of

                                            6-35

-------
          IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
                                           Quote

 1            species preservation or ecosystem health?  The point is not to measure gain in
 2            arbitrary units such as how much BOD reduction has been achieved, but in
 3            more meaningful terms (if possible).
 4
 5            c) What is the value of the environmental gains?  It is sometimes, though not
 6            always, possible to quantify the physical gains. This should be in a single
 7            metric such as dollars or the number of statistical lives saved and should
 8            summarize the diverse environmental consequences of the regulation.
 9
10            d) What is the probability of success in obtaining the expected environmental
11            gains? Some risk reduction options are more proven than others.
12
13            e) Can performance be monitored? One must be able to conclude whether
14            aggregate environmental goals are being met. It is not always easy to monitor.
15            For instance, VOC emissions from small businesses may be very difficult to
16            monitor.
17
18            f) When can results be expected? Is the regulation designed for immediate
19            results or will it take a long time to be effective?
20
21            g) What are the associated risks?  For instance, regulations to reduce sulfur
22            dioxide emissions may contribute to accelerating global warming.
23
24            h) Is pollution prevented rather than abated? Although the environmental
25            effects may appear to be the same, it is more desirable to reduce the
26            generation of pollution through process change rather than simply appending
27            an abatement technology (end-of-pipe treatment).
28
29            The criterion of cost includes all of the non-environmentaf costs and benefits of
30      a risk reduction option.  The specific way in which costs are measured is discussed in
31      Chapter 4. Questions relevant to this criterion include:
32
                                            6-36

-------
          IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
                                           Quote
 1            a) What are the direct costs of implementing the risk reduction option? This
 2            would include the costs incurred by a polluter in reducing emissions, which
 3            might be costs of abatement equipment or costs from adopting more costly
 4            processes.  Also included would be administrative costs - costs associated
 5            with applying for permits, participating in review procedures, and costs
 6            associated with uncertainty of what might emerge from the regulatory process.
 7            A third direct cost would be the consumer losses associated with higher product
 8            prices (because of higher production costs). It is also important to include the
 9            regulatory costs, incurred by the EPA.  These  would include the costs of
10            formulating and administering the regulation as well as enforcement and
11            monitoring costs.
12
13            b) Are costs of residual damage reflected in product prices?  In order for
14            consumers to receive an adequate signal regarding how damaging the
15            production of goods might be, the price of those goods should reflect all
16            environmental costs, including the costs of abatement but also including the
              environmental damage associated with residual pollution. If consumers see
18            product prices that are excessively low, this will generate too much
19            consumption of the polluting product and consequent inefficiencies.
20
21            c) Are there incentives to innovate? Some options provide little incentive to be
22            creative and innovate. Often, the option spells out what to do and innovation
23            only results in tightening the requirements. It is important that an option provide
24            incentives for innovation by polluters and by equipment vendors. Often this
25            requires that all or part of the gains from the innovation accrue to the innovator.
26
27            d) How much regulatory commitment is involved with the option? If there is little
28            commitment, does that increase costs substantially? When a regulator
29            commits to a risk reduction option, and cannot retreat,  then polluters make
30            least cost investments. Without commitment, polluters will take more flexible,
31            but also more costly, approaches to meeting regulations. A lack of commitment
32            on the part of regulators can only increase costs, often with no associated
33            environmental gain.
34

                                            6-37

-------
           IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
                                            Quote
 1            Equity: Problems arise when any identifiable group disproportionately bears
 2      the costs of a selected option.  One might also be concerned with benefits falling
 3      disproportionate on particular groups; there seem to be fewer political problems when
 4      small groups benefit from a governmental action than when small groups bear costs
 5      (substantiating reference?).
 6
 7            Basically, what is at issue here is the distribution consequence of a risk
 8      reduction option. Several issues arise:
 9
10            a) What is the incidence of costs? Do any groups disproportionally bear the
11            costs? For example, if a benefit is distributed equally in a population and the
12            cost is applied equally to each individual, lower income families will pay a
13            higher percentage for equal benefit.
14
15            b) What is the incidence of benefits?  This is a very similar issue as the
16            incidence of costs, although perhaps not as important. Nevertheless, it is
17            important to identify any groups in the population that may disproportionally
18            gain the benefits.
19
20            c) Do benefits and costs accrue to the same group? If one group
21            disproportionally bears costs but also reaps benefits, there is less concern than
22            if different groups bear the costs than receive benefits.
23
24            d) Are the regulations confiscatory?  Polluters  may be more opposed to a
25            marketable permits system where permits are  auctioned off initially  than a
26            system with permits freely distributed.  Any regulation that is viewed as
27            excessively confiscatory is more likely to be opposed in the political process.
28
29            e) Are benefits and costs concentrated or diffused?  The more diffuse the costs
30            and benefits, the less likely it is that specific groups will emerge to oppose the
31            regulation in the political process.
32
                                              6-38

-------
          IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
                                          Quote
 1            Some risk reduction options work better than others.  Workability addresses
 2      how easy it will be to implement an option.
 3
 4            a) How large is the experience base with a particular risk reduction option? If
 5            an option is a variant of one that has been used many times before, then both
 6            the EPA and the polluter will have experience in making it work.  In contrast, a
 7            new option may be more difficult to implement and work out the problems.
 8
 9            b) How much information does the government need in order to implement the
10            option?? The primary criticism of command-and-control regulation is that the
11            EPA requires an unreasonable amount of information.
12
13           c) What are the demands on government? An option that has less demands
14           on government is to be preferred to one that has heavy demands on
15           government, all other things being equal.
* -
 ,.            d) Is the public involved in the decision process?  The more the public is
18           involved as a risk reduction option is implemented, the less likely it will be that
19           unexpected problems develop. Often when the public is excluded, factors that
20           would have been brought to light through public hearings end up causing
21            significant problems during implementation.
22
23           e) Are enforcement and monitoring easy? Clearly some regulations are easier
24           to enforce than others. Product bans are straightforward, provided leakage
25           from unregulated areas is not a problem.  If there is significant reliance on
26           unobserved  actions by polluters (e.g., maintenance), then it will be difficult to
27           ensure that these actions have been taken.
28
29           f) Are the financial burdens on government excessive? Some options are
30           cheaper than others to administer. The larger the burden on government, the
31           more difficult to sustain.
32
                                            6-39

-------
           IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
                                            Quote
 1            g) Are options viewed as excessively intrusive? Some options are politically
 2            unpopular because they are viewed as too intrusive.  For instance, proposals to
 3            use intelligent vehicle systems to automatically charge for road use have been
 4            met with privacy objections. Some citizens do not like the idea of their
 5            movements being centrally monitored.
 6
 7            No matter how well an option is designed, changes will undoubtedly be needed
 8      making flexibility important.  It is  important to recognize the ease with which an option
 9      can be modified to tailor to specific conditions.  Several issues emerge.
10
11            a) Is the option adaptable to new information? If new information requires that
12            Congress re-draft the appropriate legislation, the option will be more difficult to
13            modify than if changes are automatic. The more readily an option can adapt to
14            changed circumstances, the better it will perform.
15
16            b) Does the option promote innovation and diffusion of innovation? This issue
17            was mentioned in the cost-effectiveness grouping. It is important that an option
18            be flexible enough to allow innovation (this is not always the case). It is even
19            more important that the option reward innovation. Innovation is a primary
20            means for increasing environmental effectiveness while reducing costs.
21
22      6.6 Screen and Prioritize Potential Risk Reduction Options
23
24            Options screening is the process by which a broad range of risk reduction
25      options is reduced to a set of unranked options for more detailed comparative
26      analyses. An option can be single or composite. Screened-in options are those that
27      can be implemented to meet the risk reduction objective within  all constraints specified
28      in the problem formulation.  A constraint is a specified magnitude of a screening
29      factor. For example, if the budget limitation is $1,000,000, then budget is a screening
30      factor, and the constraint is $1,000,000.
31
32            An important utility of the screening stage is that it enables the assessment of a
33      wide array of identified options. The probabilities of option targeting and arbitrary

                                             6-40

-------
          IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
                                           Quote

 1     elimination of options are minimized. Option targeting is a common problem in
 2     remedy selection.  For example, engineering remedies are more likely to be selected
 3     as a remedy for a contaminated site than communication options, regardless of their
 4     relative attributes.  The screening process provides a systematic method for
 5     determining whether each identified option meets specified performance criteria.
 6     Depending on the constraints, both categories of risk reduction options may be
 7     screened-in for more detailed comparative analysis during subsequent stages of the
 8     options selection process.
 9
10           The application of multi-criteria, multi-objective analyses to a large number of
11     identified options may unduly complicate the options selection process. The
12     screening step provides an opportunity for the identification of potentially unfeasible
13     options, thereby eliminating these options from further consideration. The number of
14     options suitable for comparative analyses in the post-screening stages of the options
15     selection process is then reduced to a more manageable level.  In essence, data
16     requirements on options screened-in are proportional to the number of options. The
       generation of a prioritized (and reduced) list of options for comparative analyses
18     reduces the costs and time required for acquiring relevant data. For example, if the
19     direct financial cost constraint specified for  reducing ecosystem risk due to
20     contaminated sediments at a site is $1,000,000 and an effective treatment technology
21     can not be implemented under $3,000,000, that option would be screened out as
22     written. This would eliminate the need to acquire additional data that are required to
23     compare it with other options in subsequent analysis. It should be noted that this
24     option may be restructured and re-screened later in the process.
25
26            The screening process satisfies the  need to give assurance to stakeholders
27     that a narrow set of options was not pre-targeted.  Transparency is an important issue,
28      especially in cases where the implementation of an option is associated with shared
29      risks and costs. Transparency is particularly important when different parties within
30     the impacted communities prefer different options, some of which are rejected for not
31      meeting specified  selection criteria. In the  selection of a site for a waste management
32      facility, the potential impacts on the surrounding communities vary both spatially and
33      temporally. Consequently, communities proximal  to selected sites may feel unfairly
34      treated if the site selection process is not transparent. Because the issue of

                                             6-41

-------
           IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
                                           Quote
 1      environmental equity has gained importance within the past decade, several
 2      systematic techniques for site screening have been developed (Wright et al., 1993;
 3      Smith, 1996; Frannis, 1993; Knight, 1985; Schwartz et al., 1989; Zeiss and Paddon,
 4      1992).  The screening process provides the opportunity for decision-makers to show
 5      that no party or segment of the community had been pre-selected to bear a
 6      disproportionate hardship.
 7
 8            Some options that are individually undesirable may exhibit increased utility in
 9      combination with other options. The screening process provides the decision-
10      maker(s) with the first opportunity to identify options that could be potentially
11      combined.
12
13            The specificity of an option is an important determinant of the uncertainties
14      associated with the potential performance of the option in risk reduction.  Diffuse and
15      imprecise options have high uncertainties. A distinction should be made between a
16      composite option that comprises distinct options and a genetically stated option. The
17      generic option can be regarded as a category of options. Specific options within the
18      category may screen in while others may screen out. Thus, the performance
19      uncertainty is higher for the option category than for any of the options within it.  As an
20      example, Table 6-2 shows various option categories and specific options that are
21      nested within them.
22
23
24
                                             6-42

-------
           IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
                                             Quote
 1
 2
 3
 9
10
11

12
13
14
15
16
17
18
19
Table 6-2. Examples of option categories and specific options (option categories have
         greater performance uncertainties than specific options).
           ENGINEERING
                                                            H^
                                   •  Site cleanup
                                   •  Waste containment
                                   •  Technology development
                                   '•*•; fGbmmunfty-
                                   ••'^Format education
           REGULATION
                                      Control of production process
                                      Banning of products
                                      Facility siting controls
                                      Facility operational controls
           MARKET INCENTIVES
                                   •  Tax relief
                                   •  Relaxation of controls
                                   •  Subsidies
                                   •  Direct purchase of friendly products
    ENVIRONMENTAL
    MANAGEMENT SYSTEMS
Application of technology and techniques to mitigate
risks, and direct use communities for operations and
planning
       If an identified option is merely described as "Engineering," high uncertainty is
associated with its effectiveness as a risk reduction option. If, however, a specific
engineering measure is identified, the expected performance can be stated with a
higher level of confidence. A specific option such as site cleanup, can be further
                                              6-43

-------
           IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
                                             Quote
 1      desegregated into specific cleanup technologies: electrokinetics, pump-and-treat,
 2      incineration, etc. Generically stated options should be refined when possible.
 3
 4            Another important factor is the relaxation or tightening of constraints. As
 5      discussed in the next section, the amount of money or other resources allocated in the
 6      budget for use on a risk reduction project is a constraint. Some potentially effective
 7      options may screen out because they cannot be implemented within the budget
 8      (constraint). If identified options are to be left unmodified, one alternative is to relax
 9      the constraint by increasing the budget resources. The greater the gap in the
10      professional hierarchy between the project implementer and the budget allocators, the
11      more difficult it is for the risk reduction project implementer to relax a budgetary
12      constraint (i.e., it may not be within the implementer's authority to do so).
13
14            The objective of the risk reduction project needs to be defined as explicitly as
15      possible. The objective is a condition or an acceptable level of service at which any of
16      the identified options is considered potentially effective. The objective should not be
17      confused with a  constraint.  Examples of objectives are:
18
19            a) Reduction of particulate matter in the ambient air in Washington, DC by
20            30%;
21
22            b) Reduction of metals concentration in River Brown to Drinking Water Levels;
23            or
24
25            c) Reduction of cancer risk due to radon in U.S. homes by 50%.
26
27            The operational rule for options screening is that if an option is deemed
28      impossible to implement within limits specified on critical factors, that option is
29      eliminated from further analysis unless it is refined or combined with other options to
30      satisfy the limits set by the project constraints. These critical factors are herein called
31      screening factors. Examples of screening factors are:
32
33            a) Budget,

                                              6-44

-------
          IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
                                           Quote
 1            b) Time,
 2
 3           c) Jurisdictional Authority, and
 4
 5           d) Legality.
 6
 7           If a minimum or maximum level allowed on each screening factor is specified,
 8     then this threshold level is considered as a constraint. For example, a regulatory
 9     agency may seek to reduce paniculate matter in the ambient air in Washington, DC by
10     30% with a maximum budget of $13 million within the next 3 years. For this risk
11     reduction objective, budget and time are screening factors. More specifically, for any
12     option to screen in, it must be implementable within budget and time constraints of
13     $15 million and 3 years, respectively.
14
15           Essentially, a "yes" or "no" answer is required, and each screening factor is
16     taken one at a time. It may also be possible to estimate the level of confidence
       associated with the "yes" or "no" answer. For example, members of a city council may
18     state: "We are 80% confident that increasing parking fees in a central business district
19     of Washington, DC will reduce air pollution by 40% within a 3-year period, because car
20     pooling or greater use of the metro  system will occur at the new parking fee level."
21     Thus, "increase in parking fees" would screen in as an air pollution risk reduction
22     option at the level of confidence of 80%. The confidence level estimated is an
23     expression of uncertainty. It is worth noting that an increase in parking fees is not
24     necessarily the best option or set of options.  Such a determination is not the focus of
25     the options screening stage of analysis. Rather, the purpose of this effort is to
26     determine how many options or combinations are available for implementation within
27     the constraints established.
28
29            Risk reduction objectives, screening factors and associated constraints have a
30     hierarchical but interwoven structure.  The decision-maker on constraints depends on
31     the scenario.  The scenario could be general or site-specific with respect to the spatial
32     coverage of the desired results of actions associated with risk mitigation objectives.
33     An example of a general scenario is an effort to implement measures that are


                                             6-45

-------
           IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
                                            Quote
  1
  2
  3
  4
  5
  6
  7
intended to reduce skin cancer in the United States. The problem may not be
confined to a particular location with respect to causes and/or effects.  An example of
a site-specific scenario is groundwater contamination beneath a gasoline storage tank.
The relative level of involvement of the decision-maker as a function of the scope of
the scenario is illustrated in Figure 6-2.
Site Specific Coverage
g- High
*
|0 >
Private Corporations (ABCD)
B
E
Ti
G IK D
t

State/Cou
Agencies (EF
International Agenc
p
nty
-GH)
'M
es (IJKL)
O H

N
National
Regulatory
Agencies (MNOP)
* 	 3!
L: ^
/ High
Decision Making Role
 8
Figure 6-2.   Decision-Makers on Constraints on Risk Reduction Objectives Under
            Various Scenarios
 9
10
11
12
13
14
      National regulatory agencies such as the EPA tend to be involved as decision-
makers, mostly in general-coverage scenarios.  In some cases, they may also be
involved in risk reduction programs at a site. Private companies are involved as
decision-makers almost exclusively at specific sites. Generally, the decision-maker
specifies the overriding constraints. Sometimes, the decision-maker may not be the
eventual implementer of some of the options that have high screen-in potential. The
                                            6-46

-------
           IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
                                            Quote

 1      roles of regulatory agencies and private companies in constraint specification is
 2      illustrated by the following examples.
 3
 4            a) A regulatory agency identifies constraints for its risk reduction policy
 5            development approaches.
 6
 7            b) Regulatory agencies usually specify constraints (generally time and
 8            performance level) associated with a requirement that the private sector
- 9            implement risk reduction projects.
10
11            c) A private company specifies constraints for various screening factors that it
12            has selected for evaluation of internally generated options for solving an
13            environmental issue.
14
15             It is important to note that options screening involves the use of one screening
Y      (evaluation) factor at a time. Constraint levels on other screening factors are held
1.      constant.  An option is considered as being fully screened after it has been evaluated
18      with respect to each of the set of screening factors. Screened-in options are those
19      that are likely to meet the specified risk reduction objective under the constraint levels
20      specified on each of the screening factors. During screening an effort is not made to
21      rank the option(s) relative to others. In contrast, subsequent analysis (broad analysis
22      and optimization) of screened-in options involve explicit ranking. Furthermore, pooled
23      or aggregate ratings or scores (with or without weights) are not used to express the
24      utility of each option as in the screening stage. The sequential steps in the options
25      screening methodology are illustrated in Figure 6-3.
26
                                              6-47

-------
           IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
                                              Quote
         Information Sources
                                        Methodology
         Input from Problem
         Definition Stage
         Input from Option
         Identification Stage"
                          STEP1: Identification of Specified
                            Objectives and Constraints
                        STEP 2: Identification or Refinement
                         of Screening Factors and Options
         Input from Collected
         Background Information.
         Expert Opinion.
         and Models
                        STEP 3: Compilation of General and
                         Performance Information on Each
                           Individual or Composite Option
                             STEP 4: Analysis of Options with Respect to
                               Constraints on Each Screening Factor
                                                              Output
 Itemized List of Scenario
^Factors: Objective. Screening
 Factors, Constraint Levels.
 and Identification Options
 Summary of Input Data to
 be Used in Analysis of
 Each Option
                                                          ListofScreened-ln
                                                          Options for
                                                          Subsequent Analysis
                     STEP 5a: Refinement
                    of Screened-Out Option
                                               STEP 5b: Inclusion of
                                            Options into Screened-ln List
         Figure 6-3.  A Screening Methodology for Risk Reduction Options
2
3
4
5
6
7
8
6.7 Evaluate the Remaining Risk Reduction Options


       The primary focus of this effort is to develop a common base of information for
comparing and selecting among the various sets of options, however different they
may be. For example, this might include comparing  the use of consumer education
programs on pesticide residues, to the imposition of  an environmental charge on
certain pesticide uses, to combining use restrictions  with a trading program, all as
                                               6-48

-------
          IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
                                           Quote
 1     options for reducing the consumption of pesticide residues in food.  A secondary, but
 2     extremely important, purpose is to continue to lay the groundwork for explaining the
 3     selection process to the wide range of interested parties.
 4
 5           The first concern is to be clear about the inputs to the assessment process.  In
 6     addition, data will be necessary on:
 7
 8           a) the past or current conditions relevant to understanding and controlling the
 9           identified risks (patterns of pesticide use, likely alternatives, patterns of food
10           consumption);
11
12           b) the performance of individual risk reduction options in other, hopefully
13           similar, circumstances (consumer labeling, liability schemes); and
14
15           c) the characteristics of each option relevant to the evaluation criteria.
 - ^
 .,           Some of this data will have been generated on the way to this point in the
18      process.
19
20           The most problematic set of inputs, however, are the estimates of the future
21      results of implementing any particular risk reduction option. For example, estimating
22      the reduction in exposure levels resulting from a food labeling program is extremely
23      difficult.  No matter how much data on past experience has been collected, these
24      estimates will  reflect a significant degree of subjective judgment by the estimating
25      parties.  This judgment will reflect their past experience, as well as their approach to
26      the available data.
27
28            As such, it is important to identify clearly the people who are providing the
29      estimates, hence, making these judgment calls. They can include:
30
31            a) the parties selecting the priority risks to be addressed;
32
                                             6-49

-------
           IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
                                            Quote
 1            b) outside experts on the characteristics of those risks or individual risk
 2            reduction options (including those from other program offices in EPA or other
 3            regulatory agencies);
 4
 5            c) the assessors themselves; and
 6
 7            d) other interested or affected parties (e.g., businesses, workers, consumers,
 8            and environmental groups).
 9
10            The likely accuracy of their estimates on any particular factor (along with their
11      approach to the underlying assumptions) should be the focus when assigning tasks
12      among these different parties.
13
14            Since application of all the assessment criteria requires that estimates of
15      performance be made, all include a degree of subjectivity. That fact tends to be
16      obscured, however, when an estimate is expressed in numbers, rather than in words.
17      Many people have a tendency to accord greater weight and certainty to quantified
18      data (whether justified or not). The task for decision-makers is to recognize and
19      account for this effect during the evaluation process.
20
21            The effort to be consistent across evaluation criteria is further complicated by
22      the fact that some measures more readily lend themselves to quantified expression.
23      Examples include: expected reductions in pesticide use; expected reductions in
24      pesticide residues; and the costs of applying alternative methods of pest control.
25
26            At the same time, all of the evaluation criteria can draw from quantitative data.
27      Population statistics are a key part of equity considerations (in what regions are the
28      pesticides of concern used and by whom; how do patterns of food consumption vary
29      across different population groups).  Results of prior experience with particular risk
30      reduction options includes percentage changes in baseline conditions (such as
31      through the lead phase-out program under the Clean Air Act) and the levels of
32      incentives created in the past (such as the barriers to innovation which can arise from
33      too rigid an application of technology based controls).

                                             6-50

-------
          IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
                                           Quote
 1            In fact, all of the measures and results of the assessment process can be
 2     expressed in either a quantitative or qualitative manner.  For example, the result of an
 3     assessment of environmental effectiveness for any particular option can be expressed
 4     either as a number (1,2 or 3) or a word (low, medium or high). Results can  also be
 5     aggregated across criteria for any particular option in either qualitative or quantitative
 6     terms. It should be remembered that, even though numbers can be applied, the
 7     accuracy of the analysis is the same as using low, medium, and high. However, the
 8     general belief is that using numbers makes the decision  more accurate or quantitative.
 9     Avoid falling into the "analysis trap."
10
11           Choosing a method of expression requires consideration of a number of
12     competing factors, most of them going to the question of how important to the
13     decision-maker is consistency in the options' scores across particular criteria.
14     Converting all of the estimates into the same scale of numbers, or words, helps the
15     comparison of different options across the evaluation criteria,  but may obscure
16     important bits of data relevant to an option's score on any particular criterion (such as
       hiding a specific estimate of the costs to be borne by particular populations in an
18     overall score of "high" or "3" on the equity criterion). Similarly, aggregating  a particular
19     option's scores across all criteria make for easier comparisons among risk reduction
20     options, but obscure the relative strengths and weaknesses of any particular option
21      (an overall score of "medium" or "2" might result from either scores of "medium" on all
22     criteria, or a mix of only "lows" and "highs" across the criteria).
23
24-           However expressed, the basis for the results of individual assessments should
25      be explained. The goal in doing so is to illuminate the policy choices and assumptions
26     that are behind the analysis.
27
28            The analysis of risk reduction options includes considerable amounts of
29      uncertainty. This is true in both the data that are available and the estimates that are
30      made.  Data may be lacking because they have not been or cannot be collected
31      (actual pesticide residues in food consumed in all different regions of the country, with
32      any seasonal variations). Theoretically, policy makers operating under time or cost
33      constraints  (experience of other regulators with trading programs) may not have
34      reviewed available data. Data may be sought on processes that are inherently

                                             6-51

-------
           IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
                                            Quote
 1      variable and incapable of accurate measurement or prediction (actual levels of illness
 2      resulting from chronic exposure to pesticide residues in the U.S.).
 3
 4            At the same time, the manner in which these uncertainties are expressed
 5      affects peoples' decision-making in unusual ways.  As with quantified expression, the
 6      more certainty that people believe is associated with a piece of data or an estimate,
 7      the more weight they give to it, whether it deserves that weighting or not.
 8
 9            Decision-makers have several options for addressing these uncertainties. Least
10      productive is to abandon the assessment process for particular risk reduction options
11      just because the relevant inputs are uncertain. Another approach is to identify the
12      data needed, along with the time and costs necessary to obtain them, before deciding
13      how best to proceed.
14
15            The most useful approach  is explicitly to reflect the uncertainty in the
16      assessment process. Any predictive process faces differing levels of uncertainty.
17      Policy action often needs to be taken in the face of those uncertainties. To the extent
18      policy makers can be explicit about the level and type of uncertainty in the different
19      parts of their decision-making process, their judgments can be understood and
20      evaluated directly by concerned parties. This is true not only for those criteria
21      expressed in a quantitative manner, but for each criterion and the assessment as a
22      whole.
23
24            Having collected the information and considered the issues described above,
25      the process of analyzing particular risk reduction options against the evaluation criteria
26      is a relatively straightforward  process of applying the decision-maker's best judgment.
27      The estimated measurements developed for any particular option are reviewed and
28      scores determined for each criterion.  For example, an option would have a high score
29      for cost-effectiveness if its direct implementation costs were low  and a high level of
30      risk reduction was expected to be achieved.
31
32            One question that does arise is who should conduct the assessment.  While the
33      information collection process and the explanation of the results should reach the

                                            6-52

-------
          IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
                                           Quote
 1     widest possible audiences, optimally, the assessment process itself should be
 2     conducted by the ultimate decision-maker(s).  Such an approach means that less
 3     specific expertise is brought to bear on evaluating any particular criterion than would
 4     be the case if a group of specially retained experts were asked to undertake the
 5     assessment and report their results.  It has the advantages, however, of providing
 6     greater consistency to the review across criteria and options, as well as making the
 7     policy judgments clearer and more readily explained.
 8
 9            Two major products result from completion of the assessments for each risk
10     reduction option. One is a series of scores across the evaluation criteria for any
11     particular group of options. As discussed above, these scores can be expressed in
12     qualitative or quantitative terms. The level of uncertainty associated with the score for
13     any particular criterion should be considered and reflected appropriately.
14
15            A major issue here is whether to aggregate the separate scores on different
16     criteria into a single, overall score for each risk reduction option.  The major advantage
       of doing so is to make comparison of the results easier across different options. For
18     example, if one option has an aggregate score of "high" and another an aggregate
19     score of "medium", one could arguably have a basis for choosing the first over the
20     second without further analysis.
21
22            The major disadvantage of an aggregated score is that the total  hides the
23     relative strengths and weaknesses of particular options on particular criteria. For
24     example, the aggregated score of "high" for one option may include a lower score  on
25     equity or flexibility than the aggregated score of "medium" for another option.
26     Assigning different weights to different criteria as part of the aggregation process can
27     make some adjustment. For some policy makers, this will not be sufficient and the
28     entire matrix should be used to compare the strengths and weaknesses of different
29     options.
30
31            The second major product of the assessment process is the identification of
32      issues that should be fed back into earlier steps in the methodology, including:
33      problem definition; data collection; risk reduction options identification;  as well as the
                                             6-53

-------
           IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
                                           Quote
 1      choice, design and measures of the evaluation criteria. Within the time available, the
 2      use of such feed-back loops will improve considerably the ultimate outcome.
 3
 4            The final step in the assessment of risk reduction options is to document the
 5      process used. What were the inputs and how were they assembled? What measures
 6      were used and how were they evaluated/assessed? How were scores assigned and
 7      by whom?
 8
 9            With the baseline of assessment results in hand for a number of different risk
10      reduction options, the decision-maker can then move to the next steps of optimizing
11      the  remaining option sets and selecting one or more for implementation.
12
13      6.8  Optimize the Options
14
15            Any given set of risk reduction options is but some small subset of the virtually
16      infinite set of possible options. For this reason, we propose that any subset that has
17      passed a preliminary screening and back-of-the-envelope evaluation should be refined
18      further before full scale option evaluation is undertaken. Thus risk reduction option
19      refinement begins with a set of options that have been at least cursorily evaluated on
20      multiple criteria and screened on one or more criteria.  Some of the  options
21      considered have already been redefined in this process before this stage.
22
23            The goal of this step in the analysis is to improve good alternatives, with the
24      subsidiary objectives of improving and reducing the option set under consideration.
25      Given that the options have been evaluated on a set of criteria, ideally refinement will
26      improve their ratings on each of these criteria.
27
28            The simplest way to view refinement is to think of improving a risk reduction
29      option  relative to a single criterion. For example, if quality adjusted life years (QALYs)
30      lost were the criterion and the number of QALYs lost could be reduced without
31      changing other attributes of the option, the option that resulted in fewest QALYs lost
32      would be preferred, ceteris paribus. When the criteria are quantifiable and it is easy to
33      agree on appropriate measures, such refinements could, in theory,  be treated as

                                             6-54

-------
          IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
                                           Quote

 1      optimization problems. One could think of the task as that of maximizing risk
 2     reduction subject to constraints imposed by other criteria.
 3
 4           The choice of criteria will obviously determine what changes are made in
 5     options.  Less obviously, how criteria are interpreted and considered will also influence
 6     how options are refined. Some criteria will be candidates for maximization.  Risk
 7     reduction, flexibility, and technological innovation are examples of these. Others, such
 8     as cost, will be candidates for minimization.  Depending on how criteria are
 9     formulated, assessment of an option relative to a criterion may be qualitative or
10     quantitative.  It is assumed that most will be subject at least to some kind of ranking.
11     When this is not the case, it is likely that criteria incorporate or reflect multiple values
12     that might usefully be discussed separately (see Table/Figure??-list of proposed
13     criteria, including explicit consideration of competing risks: ref, OTA report)
14
15           Treating risk reduction option refinement as an optimization problem introduces
16     several problems, however, because many criteria that might be used are not easily
       quantifiable on comparable scales. For example, flexibility and technological
18     innovation could be difficult to measure on a scale that all interested and affected
19     parties would agree is comparable to any risk reduction measure chosen (e.g.,
20     QALYs).  Measuring such attributes on subjective scales is a possibility (e.g., asking to
21     what extent the interested and affected parties (Commission on Risk Assessment and
22     Risk Management, 199?) agree or disagree that the option promotes technological
23     innovation).  Such an approach would introduce respondent variability that could be a
24     basis for discussing further option refinements, but might be difficult to predict and use
25     in any consistent approach to optimization.
26
27            Not only will evaluation of an option on any given criterion pose a challenge, but
28     predictions of changes on multiple criteria subject to specific refinements in an option
29     or option set are likely to be highly uncertain and difficult to quantify.
30
31            Optimization generally assumes that a single, summary, quantitative measure
32     can be used to capture and represent the overall value or effect of the option. As
33     discussed elsewhere, we do not recommend such summary measures, for procedural
34      reasons among others. They involve establishing specific tradeoffs between criteria,

                                             6-55

-------
           IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
                                            Quote

 1      and, by representing the results as a summary measure, can conceal changes vis-a-
 2      vis specific criteria. Hence optimization may be a useful way of conceptualizing partial
 3      evaluations of options (e.g., relative to cost, or time), but we do not recommend it as a
 4      way of thinking about the entire refinement process.
 5
 6            A wide range of policy tools can be used to define an option, including various
 7      forms of incentives, education, regulation, enforcement (add to list) and mixes
 8      thereof. Because intervening at many points in a hazardous process can reduce risks,
 9      very different options may achieve similar risk reduction goals.  For example,
10      education about use of emergency medical services or alternatives may achieve as
11      much risk reduction as education about proper use of specific protection devices (e.g.,
12      with application of pesticides). Banning a given product may achieve the same risk
13      reduction as providing incentives for development of alternative products.  In such
14      cases, however, the options' other attributes may differ considerably (e.g., flexibility, or
15      cost).
16
17             If those  involved in refining options find at this point that they are having trouble
18      determining whether a change improves an option, it may be worthwhile for the
19      interested and affected parties to return to a discussion of the values at stake
20      (Keeney, Value Focused Thinking) and consider redefining the criteria used to
21      evaluate options.  For this  reason, it is likely to be important that interested and
22      affected parties are represented in the option refinement  stage as well.
23
24             Expertise and creativity are key components of structuring good solutions to
25      difficult policy problems. Expertise is only acquired through many years of experience
26      and training, and is likely to change the way people think about the task at hand
27      (Anderson; Van Lehn refs).  Such differences often create communication problems
28      between experts and non-experts involved, which should be addressed explicitly.
29      However, systematic consideration of a variety of ways of improving options, along
30      with good communications between decision-makers with expertise in different areas
31      (including interested and affected parties other than EPA), should foster creativity.
32
33             Options can be refined by redefinition in at least the following ways:


                                             6-56

-------
          IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
                                          Quote
 1            Aggregation - combining options,
 2           a) Increase the scope of a proposed action (e.g. the actor, time-frame,
 3           resources allocated, kinds of tools used, or intervention points in the hazardous
 4           process)
 5           b) Combine complementary options to create option sets - with due
 6           consideration to what exactly is to be done, as well as how much of each action
 7
 8           Dis-aggregation - changing or eliminating parts of option sets, and
 9           a) Reduce the scope of a proposed action (e.g. the actor, time-frame,
10           resources allocated, kinds of tools used, or intervention points in the hazardous
11           process)
12           b) Eliminate parts of option sets
13
14           Some combination of aggregation and dis-aggregation
15
•»"           Each of these approaches raises further questions, such as what the optimal
        scope of a proposed action is, or what mix of actions and in what proportions is likely
18      to be optimal or best on the most criteria. In any proposed change, risk tradeoffs
19      should be reconsidered explicitly, as they are likely to change as well.
20
21           As  mentioned above, how criteria are considered is also likely to influence
22      option refinement. Such influences are not likely to be obvious, but may be serious.
23      For example, criteria that are easier to measure may be weighted more heavily
24      implicitly (i.e., receive more attention, be considered more thoroughly) simply because
25      they are easier to evaluate and discuss.  Similarly, if a criterion is usually considered
26      under a given set of constraints that set of constraints may be  imposed implicitly.
27
28           Other kinds of cognitive traps are discussed at length in the judgment and
29      decision-making literature (Kahneman and Tversky refs; Pious; Arkes and Hammond).
30      Some examples:
31
                                            6-57

-------
1
           IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
                                            Quote

              a) Context effects generally (e.g., framing effects) - an option described in
 2            terms of losses may be evaluated differently than the same option described in
 3            terms of gains;
 4
 5            b) Use of heuristics such as availability (a risk that comes more easily to mind
 6            may be considered riskier than a comparable risk that is not as "available");
 7
 8            c) Prominence - people may prioritize or weight an option implicitly based on
 9            the dimension that is most important to them; and
10
11            d) Uncertainty - people may weigh less ambiguous attributes of an option more
12            heavily, in implicit tradeoffs, because of an inherent  preference for avoiding
13            ambiguity (Hsee ref).
14
15            Knowing when to stop refining an option set is as important as at least
16      attempting to refine the option set originally considered. Several guidelines for
17      stopping rules are proposed, both in terms of the refinement process, as well as in
18      terms of its original goals.
19
20            Process rules could include stopping once each option has been examined at
21      least once; considered input from participants with differing perspectives (all interested
22      and affected parties); and/or examined historical precedents, approaches used in
23      other countries,  and approaches used for other risks (by analogy).
24
25            Achievement-oriented stopping rules will be focused on whether the resulting
26      options comprise a manageable or workable set. This might be determined by
27      considering resource limitations, by undertaking back-of-the-envelope value of
28      information analyses, or by achieving some set that can easily be screened to a
29      predetermined size.
30
31            This step in the risk reduction option ranking procedure is a quality check.
32      While elements  of the refinement process discussed above are likely to have been
33      incorporated in earlier phases of the procedure, including this step explicitly helps

                                             6-58

-------
          IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
                                           Quote
 1     ensure that those creating the options and ranking them will not become locked into
 2     an unnecessarily restricted way of thinking about options. The goal of this step is to
 3     improve the options being considered, by explicitly considering whether they could be
 4     redefined to better meet the criteria. To keep this step manageable, specific stopping
 5     rules are recommended.
 6
 7     6.9  Select an Option
 8
 9           The decision-maker is now faced with selecting the "best" risk reduction
10     option(s) from a set of options or group of options that have been screened and
11     optimized to address the defined environmental problem.  Thus, the goal of this step
12     in the process is to actually select the risk reduction option for the defined problem.
13     As with other components of the process, this selection process should be as
14     transparent as possible to those outside the selection process.  In many ways, it is
15     even more important for this selection process to be made transparent since this
16     activity will ultimately result in the risk reduction option that will be carried forth. The
       decision-maker will ultimately have to make the decision based upon the current
18     policies of the Agency but it is a basic tenant of this effort that any decision be fully
19     delineated and documented, including any factors taken into account and how they
20     are weighted relative to one another. This  section discusses decision analysis
21     methods that are appropriate for use in the selection of a risk reduction option by
22     being rigorous enough to provide the necessary transparency to the selection process,
23     by being robust enough to deal with the uncertainties of complex environmental risk
24     reduction option selection, and generally being compatible with environmental risk
25     reduction option selection.
26
27            In previous steps, there has been a  "down selection" of risk  reduction  options
28     and potentially some changes or aggregation were made to the original options in
29     order to ensure that the best mix of options or groups of options are being considered.
30     Now the optimized, aggregated, and screened list of risk reduction options will be
31     further evaluated in an attempt to select the best option or set of options. This
32     selection among the various risk reduction  options will be based upon a comparative
33     evaluation of the options versus a set of criteria. Using the evaluation of the alternative
34     options relative to these criteria as input, one or more decision  analysis methods,

                                             6-59

-------
           IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
                                             Quote

  1      described below, will be used to make the final selection of the risk reduction option or
  2      set of options. As previously noted, the evaluation criteria span the range of issues
  3      appropriate for consideration when making an environmental decision from those
  4      dealing with technical performance and cost factors to those dealing with social
  5      factors. As was the case in the screening steps, the goal is to quantify the score or
  6      value for each evaluation criteria for each risk reduction option.  The degree to which
  7      the scores of each evaluation criteria can be quantified and with what degree of
  8      certainty, will  influence the type of decision analysis method used, as discussed
  9      below.
10
11            There  are numerous decision analysis methods that can be used for assessing
12      the alternative risk reduction options and several good literature sources that deal with
13      the multi-attribute decision analysis. The selection of the appropriate decision
14      analysis procedures will depend on many factors including the degree of quantitation
15      that can be brought to each of the evaluation criteria and the need for a transparent,
16      well delineated, selection process.  The general types of decision analysis methods
17      that could be used for this analysis range from those that are most qualitative to the
18      those that are most quantitative:
19
20            a) Holistic,
21
22            b) Cost Benefit,
23
24            c) Matrix qualitative (High to Low) Ranking,
25
26            d) Decision Factor Analysis (numerical rating factors with weighting on the
27            factors), and
28
29            e) Optimization (multi-attribute decision procedures).
30
31            In the end, the decision-maker will likely choose a hybrid of these processes in
32      making the decision on the best mix of risk reduction options in order to take
33      advantage of the special features of the different decision analysis tools for the

                                              6-60

-------
          IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
                                           Quote
 1     particular environmental problem.  In the following paragraphs we discuss the basis of
 2     each of these approaches and the potential utility of the decision approach for the
 3     types of selection encountered relative environmental risk reduction options.
 4
 5           The holistic decision process refers to the completely (or almost completely)
 6     qualitative procedure of making the decision among the various risk reduction options
 7     based upon intuition, professional judgment and certain known facts about the various
 8     evaluation criteria. This procedure can be used for consideration of the other criteria
 9     when one dominant criteria such as environmental performance has been quantified
10     more rigorously. This approach suffers from the fact that it does not adequately allow
11     delineation of the decision process due to the use of personal judgment.
12
13           Cost benefit analysis is a decision analysis procedure that can be used to
14     evaluate the relative costs and benefits of the risk reduction options. This procedure,
15     for example,  would allow a comparison of the costs for each pound of pollutant
16     removed by the various risk reduction options. If each  of the evaluation criteria  could
       be cast as a monetary cost or benefit to society then a  comprehensive quantitative
18     evaluation of costs and benefits could be made.  For complex environmental problems
19     with numerous risk reduction options and a large number of criteria, this approach is
20     limited by the large informational requirement.
21
22            Matrix comparisons involve the development of a matrix of the assessment of
23     each of the evaluation criteria for each risk reduction option. A quantitative (numerical
24     value) or a qualitative (high, medium, low) ranking is developed for each criteria for
25     each risk reduction option. The decision-maker then must choose the risk reduction
26     option to implement that provides the best balance in simultaneously meeting all of the
27     criteria.  It is of paramount importance that the decision-maker delineate the tradeoffs
28     that were made in the final decision and provide the rationale for the final selection
29     given the different criteria considered to be important.  Some potential decision rules
30     that are useful in the final selection are provided in the following section.
31
32            Decision factor analysis is similar to the matrix approach but differs in that each
33      of the evaluation criteria is weighted and scored with a numerical value during the
34      evaluation of each risk reduction option. In this approach, each option is evaluated

                                             6-61

-------
           IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
                                            Quote
 1      either qualitatively or quantitatively relative to each evaluation criteria and given a
 2      quantitative numerical score. The weight placed on each criterion reflects the relative
 3      importance of this criteria to the decision-maker for this particular environmental
 4      problem. The selected risk reduction option is thus the option with the highest
 5      cumulative score given the numerical score and weighting for each evaluation criteria.
 6      The advantage of this approach is that it clearly delineates the factors used in the
 7      evaluation and the weighting applied to the criteria. The disadvantage of this
 8      approach is the need to develop quantitative scores that discriminate between the
 9      options that may be difficult to quantify for complex environmental problems and
10      complex criteria such as social equity.  Furthermore, not all preferences are consistent
11      with this approach (e.g. lexicographic preferences).
12
13            Optimization refers to the decision-making process that attempts to quantify the
14      decision into a formulated relationship  between an objective function that is to be
15      maximized (or minimized) relative to a  series of constraints (or criteria).  The objective
16      function can be a function of cost, benefits, and detriments and can account for
17      tradeoffs among these by incorporating relative weighting factors and functional
18      relationships. There are a number of methods for determining the optimum set of
19      decision variables and commercial software is available to carry out the procedures.
20      While this procedure is the most quantitative approach it  is often not suitable for
21      complex environmental problems due to the difficulty in establishing functional
22      relationships between decision variables especially with qualitative factors and due to
23      the nonlinear relationships that often exists among the various factors.
24
25            We conducted a number of example evaluations involving the actual selection
26      of risk reduction  options for a series of hypothetical environmental problems. The
27      environmental problems were oriented around different contextual viewpoints,
28      specifically environmental stressor (e.g.,  VOC emissions as related to urban ozone),
29      the media (water, air, solid waste) context and location (place based) context. In
30      these examples, the selection criteria included both factors that were quantifiable and
31      those that were unlikely to be quantifiable for the set of different risk reduction options.
32      The qualitative nature of some of the inputs to the decision, tends to move the
33      selection to be of a subjective nature.  However, a purely holistic approach does not
34      provide the appropriate degree of delineation and therefore transparency.

                                              6-62

-------
          IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
                                           Quote
 1
 2            In order to accomplish this selection for an environmental problem with a wide
 3      array of potential solutions, it may be necessary, due to time and resource constraints,
 4      to further narrow the field with the qualitative matrix approach in order to generate a
 5      manageable set of highly ranked options and then conduct a more rigorous analysis of
 6      the criteria for each of the highly ranked options. The combination of techniques
 7      represents a hybrid of the selection processes discussed in the previous sections but
 8      may improve the overall selection process efficiency and ultimate result.
 9
10            The final selection of the optimal risk reduction option or set of options can be
11      made by applying different decision  rules. The use of decision rules is particularly
12      important when the matrix approach is used for the selection process due to the
13      complexity of the particular environmental problem under analysis. For example in the
14      special case called "dominance" where one single risk reduction option is highest
15      ranked relative to every criteria it would clearly be selected. Other possible decision
16      rules include using a lexicographic rule (just looking at a single criterion), selecting the
        option(s) with  the highest ranking on a specific criterion, or excluding options with the
18      lowest ranking on any one criterion.  The policy decision maker could also select the
19      option(s) whose lowest ranking on any criterion is higher than the lowest rankings for
20      other options, or use some model to combine rankings on individual criteria into one or
21      more scores (as discussed above).  Again, the exact application of the decision rule(s)
22      can not be a priori specified. The policy decision maker must, however, document any
23      decision rule employed and provide the rationale for that decision rule in the context of
24      the particular  environmental risk reduction options considered.  It must be possible to
25      follow the thought process through the decision-making to the final selection of the risk
26      reduction option or set of options.
27
28            One of the greatest challenges in selecting a risk reduction option, or set of
29      options, in addressing a complex environmental problem is dealing with the
30      uncertainties  in  the performance of the options relative to the criteria.  In order for the
31      selection process to be transparent to those on the outside of the process, the
32      process must document the uncertainties in the evaluation of each criteria and deal
33      with the uncertainties relative to the final selection in an open manner. As discussed
34      above, the presence of uncertainties for more complex problems and options will tend

                                             6-63

-------
           IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
                                             Quote
  1      to push the decision analysis process to more qualitative methods such as matrix
  2      approaches with low, medium, and high rankings as opposed to decision factor or
  3      optimization approaches with quantitative measurement. In any case, it is important to
  4      deal with the uncertainties within the selection process framework.
  5
  6            There are several potential methods of dealing with the impacts of uncertainties
  7      in the scores of the evaluation criteria.  One method is to have a separate criteria for
  8      each risk reduction option that scores the overall uncertainty of the performance of the
  9      option relative to all of the other evaluation criteria. A risk reduction option with the
10      same overall score relative to the other criteria but a high degree of uncertainty would
11      be less favorable relative to those with higher degrees of uncertainty but equivalent
12      scores on other criteria. It is important not to reject a better option solely on the basis
13      of the certainty of the evaluation criteria. Nonetheless, uncertainty could  be used as
14      one more criterion in the evaluation  of the overall acceptability of an option versus
15      other options.
16
17            A second approach is to attempt to quantify the uncertainty range relative to
18      each criteria. Given this range of criteria scores, a sensitivity analysis of uncertainties
19      for the highest ranked risk reduction options can be undertaken to determine if the
20      rank order of the risk reduction options changes over the range of uncertainty
21      assumptions. This is essentially a form of sensitivity analysis on the options.  The
22      most comprehensive method is to conduct the process in an iterative manner by
23      collecting new data for those parameters that are so sensitive as to change the
24      selection of risk reduction options. This last method can also have a significant impact
25      on time and costs due to the resources required to collect the additional data.
26
27            Ideally the selection  process will  lead to a single option or single set of options
28      that are considered to be "best" by the decision-maker and the decision process,
29      including criteria and criteria weighting,  will be clearly delineated for those outside the
30      process. Under certain situations there may no clear winner, but rather a set of
31      options may be very close with no clear discriminators among the evaluation criteria.
32      In this case, it may be necessary to analyze other criteria or further quantify special
33      criteria to help delineate the better case.
34

                                              6-64

-------
          IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
                                           Quote
 1     6.10 Document the Process
 2
 3           The final decision should be transparent, i.e., the motivation and criteria used in
 4     reaching the decision should be obvious. Therefore, it is important to document not
 5     just the final decision, but also the process leading to the decision.
 6           Both the process and the documentation need to contain sufficient detail to
 7     substantiate decisions, but should not be so detailed as to be unusable. Therefore,
 8     we recommend simple documentation, with tables emphasizing the breadth of options
 9     considered, and the values and judgments used in evaluating those options, and
!0     ultimately in selecting the best option. The process of making judgments is inherently
11     complicated, especially in a public forum, because of the need to elicit different values
12     and input from a broad spectrum of people.  Thus, the need to document, and be
13     consciously aware of biases and decisions when making evaluations.
14
15     6.11 Quantify Option Effectiveness
1P
}             Measurement is a critical component in risk reduction evaluation. The bottom
18     line in implementation of risk reduction options is to determine whether or not the goal
19     is achieved. Once implemented, it is critical that the effectiveness of the risk reduction
20     option or set of options be quantified. If we cannot measure the success of our
21     performance, than we probably cannot determine whether or not we are achieving our
22     goal. It is critical that both what we measure and how we measure are built in before
23     we begin implementing selected risk reduction options.
24
25            The following are the four basic approaches to measurement:
26
27            Direct - Measures that directly quantify the current status against a set goal or
28            goals for improvement (also referred to as outcome measures).
29
30            Indirect - Measures a component that has a direct impact on the specific area
31            for improvement, but does not specifically provide quantitative measurement of
32            the specific improvement goal (including administrative process or stressor
33            measures).

                                            6-65

-------
           IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
                                            Quote

  1            Leading Indicator -  Measurement of a specific parameter that provides an
  2            indication of the probable success to achieve a goal.
  3
  4            Lagging Indicator -  Measures after the fact a specific parameter that may
  5            directly or indirectly measure the performance against the improvement goal or
  6            goals.
  7
  8            To better understand these four approaches, the discussion below describes
  9      measurement approaches using two examples.  The first example discusses a typical
10      set of risk reduction options to achieve a reduction in cases of childhood asthma (the
11      goal). The second example describes a set of risk reduction options to improve the
12      biodiversity in a receiving stream. We selected these examples to describe
13      measurement approaches in both public health and ecosystem environmental
14      problems.
15
16            For this example, the goal of a specific set of risk reduction options is to reduce
17      the number of childhood asthma cases. Therefore, measurement of the number of
18      childhood asthma cases on an annual basis is a direct measurement of the goal. This
19      is also a lagging indicator because the cases have already occurred and, therefore,
20      the measurement system using this indicator tracks the result from the output of the
21      system as it currently exists.
22
23            For purposes of this example, we will assume that there is a correlation
24      between air quality and childhood asthma cases. Therefore, air quality can be
25      monitored within the study area that provides an indirect measurement of the goal.
26      The assumption is that improvements in air quality will result in reductions in the
27      number of childhood asthma cases.  Therefore, measurements of air quality indirectly
28      predict improvements in the number of childhood asthma cases and also represent a
29      leading indicator.  Expanding the community's rideshare program is one of the
30      mechanisms that could be implemented to assist in improving  air quality.  Therefore,
31      the measurement  of the rideshare program (decreases in single vehicle travel miles)
32      represents an indirect indictor on air quality and is also a leading indicator for air
33      quality.
                                            6-66

-------
          IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
                                           Quote

 1           As can be seen from this example, a number of parameters can be
 2     quantitatively measured to help define the system and predict whether the risk
 3     reduction option(s) will achieve the overall goal.  It is important to understand the
 4     interrelationship of what is being measured and how it relates to the specific
 5     improvement goal.
 6
 7           A typical goal for an aquatic ecosystem is to improve both the quantity and
 8     diversity of the biota in a stream, river, or lake. Measurement of biota diversity is a
 9     direct measurement and a lagging indicator. A number of techniques have been
10     described in the literature to perform biodiversity measurements. Clearly, the water
11     quality has direct impact on the biota diversity. Therefore, measuring dissolved
12     oxygen, temperature, salinity, and other water quality parameters are critical in
13     defining the water environment in which the biota live. These measurements of water
14     quality parameters are  indirect measurements of biota diversity and are leading
15     indicators. Improvements in the water quality parameters typically lead to
16     improvements in biota. A secondary indirect measurement can  be reduction in
       combined sewer overflow to the receiving stream.  Typically, combined sewer overflow
 16     adds pollutants to the receiving stream and negatively impacts water quality, and
19     subsequently negatively impacts biodiversity. Projects can be implemented to reduce
20     combined sewer overflow and the reductions in the quantity of sewer overflow as a
21     function of time can be measured, which represent a leading indicator of water quality.
22
23
24            Reduction in the quantity of combined sewer overflow is just one example of a
25      risk reduction option that could be implemented to improve biodiversity.  Clearly, water
26      quality is one of the key controlling conditions within the ecosystem impacting
27      biodiversity. Therefore, other risk reduction options that improve water quality also
28      need to be monitored.  Those risk reduction options that have the greatest impact on
29      controlling the key water quality parameters must be monitored  to determine their
30      impact on the ability to achieve the overall goal.
31
32            In addition to water quality, the physical conditions (including water velocity, low
33      flow water volume, and bottom substrate) within the aquatic ecosystem also play a
34      major role in determining biodiversity.  The purpose of this discussion is to alert the

                                             6-67

-------
           IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
                                           Quote
 1      reader to the fact that ecosystems are complex and that there are a number of
 2      parameters and conditions that impact biodiversity. It is, therefore, critical to
 3      determine the key controlling parameters that impact the specific goal to be achieved
 4      and to quantitatively monitor the risk reduction options which are implemented to
 5      improve those conditions in the eco-system parameters which lead to achieving the
 6      goal.
 7
 8            In order to evaluate the effectiveness of the risk reduction options implemented,
 9      it is necessary to establish a monitoring system/program that provides sufficient data
10      to predict performance and effectiveness. Measurement program plans should
11      describe:
12
13            a) Location and types of samples to be collected;
14
15            b) Frequency of sampling;
16
17            c) Analytical  methods;
18
19            d) Quality control; and
20
21            e) Management review.
22
23            The location, type,  and frequency of sampling is critical to ensure that sufficient
24      data are collected to predict the performance of a specific risk reduction option or of
25      the system as a whole. In the previous example of the childhood asthma cases, it
26      may be beneficial to collect the number of cases on a monthly basis to see if there is
27      any variation in frequency with time of year. Since there is frequently a direct
28      correlation with air quality and there are seasonal variations in air quality, one would
29      expect seasonal variations in the number of childhood asthma cases.  If the number of
30      cases is only collected on an annual basis, the evidence of seasonal variability would
31      not be available.  Collecting the data on a daily basis would be too frequent and
32      provide little added benefit over collection on  a monthly basis.  In the case of the
33      aquatic ecosystem, it is critical to perform the biodiversity analysis at the same time

                                             6-68

-------
          IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
                                          Quote

 1      each year to eliminate seasonal variations. For example, in a river ecosystem, the
 2     biodiversity is very different in the summer from the winter. For biodiversity, it is
 3     important to compare data from the same time of year. However, water quality data
 4     should be collected year-round to understand whether there are seasonal variation
 5     which may have an impact upon the biota. For example, spring runoff events may
 6     transport pollutants into the receiving body of water that may subsequently impact the
 7     biodiversity.
 8
 9           Data collection is extremely resource-intensive. It is, therefore, critical to select
10     the number of locations, appropriate monitoring parameters, and sampling frequency
11     to optimize the cost impact to the project. Sampling needs to be frequent enough to
12     be representative of the study program.  Where appropriate, the frequency should be
13     sufficient to perform proper statistical analysis.
14
15            Most samples that are collected of physical systems, such as water quality
16     samples from a receiving stream, will require laboratory analysis. The analytical
       methods employed should  be consistent with the nature of the samples and with the
18     most recently approved method of analysis. For example, the analytical method for
19     preparing stream water samples for measuring metals is different from the analytical
20     procedures used to prepare samples for analysis of stream sediment materials.  There
21     are a number of existing references that specify current and approved procedures,
22     including Standard Methods for the Examination of Water and Wastewater and
23     procedures published by American Standards for Testing and Materials (ASTM).
24-
25            In all cases, sampling programs require proper quality control. This includes
26     the collection of field blanks, trip blanks  and laboratory blanks and spikes to ensure
27     that there has been no unknown contaminant added as a result of the sample
28     collection methodology, transportation or the analytical procedure.  Depending on the
29      nature of the samples and the type of analysis being performed, there are general
30      recommended procedures on quality control.  It is recommended that before a
31      sampling program be implemented, a formal quality control program be written and
32      approved.
33
                                            6-69

-------
           IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
                                           Quote

 1            Management review of the data is critical to ensure that sampling and analytical
 2      methodologies are properly followed, quality control procedures fully implemented,
 3      and performance tracked.  As required, the sampling program should be modified to
 4      ensure efficient and cost effective sampling and analysis are performed. Frequently,
 5      sampling programs are modified over time to focus on specific areas of interest and
 6      reduce the sampling frequency and analysis for parameters that show little variation.
 7
 8            To improve the overall implementation of risk reduction options, it is critical that
 9      the performance be communicated to the public. The monitoring data should be
10      assembled in a simple to understand format and presented periodically at public
11      hearings/meetings.  This is one of the best methods to ensure acceptance of the
12      program being implemented and to receive feedback from interested parties on the
13      performance of the program.
14
15
                                           6-70

-------
          IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
                                          Quote
 1      6.12 References Cited
 2
 3     C and EN. 1996. Criminal enforcement efforts by EPA continue to grow apace.
 4           Chemical and Engineering News. July 29, pp. 33.
 5
 6     CHEM ENG. 1996. EPA enforcement continues at full throttle. Chemical Engineering,
 7           July, pp. 47.
 8
 9     Davis, M.L. and Chou, G.P. 1992. RCRA Land disposal restrictions for contaminated
10           debris. Hazardous Materials Control, March/April, pp. 30-33.
11
12     Frantzis, 1.1993. Methodology for municipal landfill sites selection. Waste
13           Management and Research, Vol. U, pp. 441-451.
14
15     Keeney, R.L. 1992.  Value-Focused Thinking. Harvard University Press.
 • ^
 . /     Knight,  M.J. 1985. Public acceptance and siting of hazardous waste disposal facilities.
18           Bulletin of the International Association of Engineering Geology, No. 32, pp. 83-
19           89.
20
21     Levin, M.ll. 1990. Implementing pollution prevention: incentives and irrationalities.
22           Journal of Air and Waste Management Association, Vol. 40, No. 9. pp. 1227-
23            1231.
24
25     McGarity, T.  1994.  Radical Technology Forcing in Environmental Regulation. Loyola
26           of Los Angeles Law Review, v.27, p. 943.
27
28     Norman, M.E. and J.D. Keenan. 1996. Market incentives to reduce non-point
29            agricultural nutrient pollution: a theoretical and implementational discussion.
30           Journal of Environmental Systems, Vol. 24, No. 2, pp. 13167.
31
32     OTA 1993. Environmental policy tools: a user's guide. Office of Technology
33            Assessment,  United States Congress, Washington, DC, 212 pp.

                                           6-71

-------
           IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or
                                           Quote

 1      Schwartz, S.F., R.A. McBride, and R.L Powell. 1989. Models for aiding hazardous
 2            waste facility siting decisions. Journal of Environmental Systems, vol. 18, No.
 3            2., pp. 97-122.
 4
 5      Smith, P.M. 1996. Environmental evaluation: fuzzy impact aggregation.  Journal of
 6            Environmental Systems, vol. 24, No. 2, pp. 191-204.
 7
 8      UN 1995. Work programme on indicators of sustainable development of the
 9            Commission on Sustainable Development. Division for Sustainable
10            Development, Department for Policy Coordination and Sustainable
11            Development, United Nations, New York.
12
13      U.S. EPA. 1990.  Reducing Risk: The Report of the Strategic Options Subcommittee,
14            Relative Risk Reduction Project, Appendix C. EPA SAB-EC-xxx, Science
15            Advisory Board, U.S. Environmental Protection Agency, Washington, DC 140p.
16
17      Wright, E.G., H.I. Inyang, and V.B. Myers. 1993.  Risk reduction through regulatory
18            control of waste disposal facility siting. Journal of Environmental Systems, Vol.
19            22, No. 1, pp. 27-33.
20
21      Zeiss, C. and B. Paddon. 1992. Management principles for negotiating waste facility
22            siting agreements. Journal of the Air and Waste Management Association, Vol.
23            42, No. 10, pp. 1296-1304.
24'
                                            6-72

-------
PART V  IMPLEMENTATION AND PERFORMANCE EVALUATION

-------
PREFACE

-------
      IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

           CHAPTER 7.  PERFORMANCE EVALUATION: THE USE OF
                 ENVIRONMENTAL DECISION REPORT CARDS

                             TABLE OF CONTENTS
 1
 2
 3     7.1 Introduction 	7-1
 4
 5     7.2 Types of Performance Measures  	7-3
 6
 7     7.3  Improved Report Cards 	7-8
 8          7.3.1 Employing a Broader Range of Measures 	7-8
 9          7.3.2 Reporting on Environmental Outcomes  	7-10
10               7.3.2.1  Environmental Health	7-10
11               7.3.2.2  Ecological Health	7-13
12          7.3.3 Additional Design Considerations	7-15
13
1     7.4 Implications for Existing Monitoring Systems	7-17
15
16     7.5 Summary and Recommendations	7-19
17
18     7.6 References Cited	7-22
19

-------
      IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1      CHAPTER 7.  PERFORMANCE EVALUATION: THE DESIGN AND USE
 2                     OF ENVIRONMENTAL REPORT CARDS
 3
 4
 5    7.1 Introduction
 6
 7          The lack of a widely accepted and commonly used method for evaluating the
 8    effectiveness of governmental (including environmental) programs represents one of
 9    the great challenges facing agencies such as the EPA as the millennium draws to a
10    close. With the passage of the Government Performance and Results Act (GPRA) in
11    1993, the Congress has mandated that an agency must do much more to answer the
12    bottom-line question "How are we doing?" if it expects continued, let alone expanded,
13    support. In the context of EPA, the answer to this question is even more important now
14    than in the past because of the growing number of complex environmental issues, the
15    evolution of institutional roles from narrowly focused mandates to broader public
16    demand for high quality protection of human health and ecological integrity, and the
" "*    intensified scrutiny of the cost of environmental programs.
  >
19          In the same way that GPRA demands accountability for performance at a
20    government agency-wide level, the IED framework seeks accountability for
21    performance at the level of individual integrated environmental decisions. In the former
22    case, the question is "What is the state of the environment, and what is the impact of
23    Agency actions on that state and the trends in the environment?"  In  the latter case, the
24    question is "What is the state of that portion of the environment that a specific
25    management decision was designed to affect, and what information is available that
26    might suggest adaptive management options that could obtain more desirable results
27    more efficiently?"
28
29          The term "environmental report card" has been used to refer to the evaluations
30    that are applicable to both the more global GPRA-like and the more specific lED-like
31    situations.  For example, recently several groups have recommended that a "report
32    card" be developed for national environmental programs (e.g., Enterprise for the
33     Environment, 1997; Vice President Al Gore, in his direction to the National Science and
34    Technology Council, year??). In many respects the Annual Reports from the Council
35     on Environmental Quality served as prototypes for a national environmental report card.

                                           7-1

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1     In addition, several regional initiatives have generated report cards, including reports of
 2     the U.S. Forest Service and Bureau of Land Management in conjunction with the
 3     Northwest Forest Plan, the Chesapeake Bay Program's annual State-of-the-Bay
 4     publication, the State of the Great Lakes report (Environment Canada and EPA, 1997),
 5     the State of Boston Harbor report, and others (Young, Harwell to provide cites).  Each
 6     of these reports provides a periodic update of governmental agency activities and, to a
 7     limited extent, related environmental improvements.
 8
 9            In contrast to these environmental report cards that assess the environment
10     more broadly, the IED uses the term "environmental decision report card (EDRC)" to
11     refer to an evaluation of specific decisions.
12
13            As depicted in the framework diagram (Figure 1-3), the EDRC is the formal
14     mechanism for feedback and course correction. Such feedback on the extent and
15     distribution of environmental effects is helpful in determining whether the relative
16     seriousness of risks was  accurately characterized in the first place and whether specific
17     risks have changed as a  result of an implemented risk reduction program.  Report cards
18     should also provide the basis for evaluating the performance of specific risk reduction
19     programs or decisions, as judged against decision criteria such as efficacy at reducing
20     aggregate risk, cost,  equity, and time required to achieve risk reduction goals.
21     Therefore, environmental report cards should provide the information needed to a)
22     identify opportunities for course correction and adaptive management, i.e., modification
23     of risk reduction approaches in light of performance  information or new information on
24     risks; and b) assign specific accountability-to individuals, programs, or organizations-
25     for environmental results. A reporting system that focuses on environmental outcomes
26     associated with individual decisions will provide greater focus and discipline by explicitly
27     expressing the relationship between an environmental investment (e.g., money,
28     information, skills, and time) and the results achieved.
29
30            The enhanced accountability provided by a well-designed report card
31     mechanism is particularly important in the context of the greater flexibility espoused in
32     the IED framework for both government agencies and private interests to approach
33     problems with innovative solutions. As the system of integrated environmental
34     decision-making evolves, with  its focus on reducing aggregate risk through flexible
35     approaches to risk reduction, reporting on the rate of achievement of particular single-

                                              7-2

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1     stressor standards (or other individual characteristics affecting environmental health)
 2     will fail to tell the whole story. In its stead, there needs to be a mechanism to gauge the
 3     collective impact of sets of risk management approaches on actual outcomes.
 4
 5          A well-designed and appropriately implemented outcomes reporting system also
 6     can engender long-term public support for environmental programs. By providing
 7     periodic objective reports that link decisions to outcomes on the health of ecosystems,
 8     human  health, and quality of life — in other words, information about the things people
 9     really care about — rather than specific chemical risks whose impacts may not be
10     obviously relevant to most people (e.g., mercury or 1,3-butadiene), the reporting system
11     has the capacity to generate a strong sense of ownership and investment on the part of
12     the public. (For a more detailed discussion of the benefits from performance-based
13     reporting systems, see Enterprise for the Environment, 1997.)
14
15           While interest in performance evaluation is growing, EDRCs and the broader
16     environmental report cards will be only as useful as the information contained in them.
17     In the following sections, we discuss various types of performance measures and some
       issues  associated with their use in environmental reporting systems.
19
20     7.2 Types of Performance Measures
21
22           With the growing interest in performance evaluation, a number of closely related
23     terms have emerged to refer to different performance measures; i.e., those measures
24     that are used  in constructing the report card. For the sake of clarity, the Committee
25*    defines the following types of performance measures:
26
27            a) Process Measures, i.e., measures of administrative effort or program actions
28           that are presumed to result in environmental or health improvements (e.g.,
29            numbers of permits issued, number of enforcement cases pursued, number of
30            contaminated sites cleaned up to standards).
31
32            b) Stressor Measures (levels): measures (levels) of stressors in the
33            environment that are used to determine attainment or non-attainment of desired
34            reductions in stressor levels (e.g., total emissions of a pollutant, levels of
35            dissolved oxygen or turbidity levels in a stream, and density of roads in a

                                             7-3

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1           watershed).
 2
 3           c) Exposure Measures: measures of the co-occurrence or contact between an
 4           individual or population and environmental stressor(s) over a defined time period.
 5           The term exposure is traditionally associated with chemical stressors (e.g.,
 6           stressor levels in food, concentrations of stressors in tissues, time-activity
 7           measures, and total exposure to a contaminant via all routes), whereas the term
 8           co-occurrence is often used as a broader term applicable to chemical, physical,
 9           and biological stressors. Exposure measures are more directly related to effects
10           than ambient levels of stressors in environmental media because they address
11           direct contact with the stressor.  These measures can also be more readily linked
12           to risk management decisions than effects measures since causes of adverse
13           effects are often multi-factorial.
14
15           d) Effects Measures: measures of human and/or ecological effects, the
16           changes in which can be used to assess the impact of an environmental risk
17           reduction program (e.g., asthma rates,  cancer rates, acres of wetlands gained or
18           lost, local extinctions of important species). Another application of effects
19           measures is for condition assessment, in which a suite of effects measures are
20           evaluated and reported in combination  to characterize the health or condition of
21           an entire population or ecosystem. Condition assessment provides a baseline
22           against which to evaluate the success of specific environmental decisions or
23           multiple decisions impacting a population or geographic region.  In addition,
24           reporting on condition provides feedback on whether the problem formulation
25           has correctly identified the set of stressors associated with the environmental or
26           health consequences of concern.
27
28           The Committee considers measures of effects or condition to be environmental
29     outcome measures.  We note that this definition differs from the Agency's definition of
30     "environmental outcomes," which also includes measures of stressor levels (EPA
31     Strategic Plan, 1997). We recommend, however, that stressor measures be kept
32     distinct from environmental outcome measures because a) changes (increases or
33     decreases) in stressor levels do not necessarily translate into changes (increases or
34     decreases) in risk, and b) the public's environmental goals are typically in terms of
35     desired states of health or condition, rather than desired stressor levels.

                                             7-4

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1           The different types of measures can be thought of in terms of a spectrum of
 2     measures ranging from least directly to most directly related to actual environmental
 3     outcomes, which are of primary interest to the public.  In many cases, the spectrum of
 4     measures will also correspond to a spectrum of response times, ranging from shorter
 5     term to longer term measures. For example, it is relatively quick and easy to document
 6     the number of permits issued in a year,  but it is more difficult and time-consuming  to
 7     determine the actual impact of that process measure on the condition of the
 8     environment. Figure 7-1 illustrates this spectrum and relates the Committee's
 9     terminology to other commonly used terms for environmental performance measures.
10
       Figure 7-1. Spectrum of Performance Measures	
                                                                Environmental Outcomes

        Process           Stressor           Exposure           Effects/Condition
        Measures         Measures          Measures           Measures

        activity measures'      pressure indicators3     co-occurrence4         adverse effects5
        output measures^      release measures                         health outcomes9
        response indicators'    emission measures                        state indicators3
        'beans'                                                   ecological indicators7
        <_____	>
        Least directly related                                           Most directly related
        to Environmental Outcomes                                to Environmental Outcomes


        'EPA Strategic Plan; 2GPRA; 3OECD Framework; 'Chapter 2; 5Ecorisk Guidelines; 'Understanding Risk; 7EMAP
11            Examples of the various types of performance measures and their use in concert
12     to evaluate the success of an IED project is illustrated in Figure 7-2 using a watershed
13     restoration example.
14
15
                                              7-5

-------
IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
Figure 7-2.  Evaluating the Success of an Integrated Environmental Decision: A
Watershed Example                                                	
         To illustrate the use of each of the types of performance measures to report on a specific
  Integrated Environmental Decision, consider a watershed management or restoration program.  In
  this example, the ultimate goal of the program is to maintain or re-establish an ecosystem that
  supports a diversity of habitat types along with their resident communities of plants and animals,
  supports essential ecological functions, and is self-sustaining. Here, the  Integrated Environmental
  Decision will consist of a set of interrelated actions, many of which will be designed to address
  multiple stressors in order to achieve a reduction in the aggregate nsk from those stressors.  Other
  IED program actions will focus on restoring damage from past stressors (such as the restoration of
  riparian zones damaged by livestock operations in order to decrease sedimentation downstream,
  provide shade and cooler temperatures for aquatic species, and provide additional nutrients to the
  system from dropping leaves).

         The evaluation criteria used to judge overall IED program results will include measures of
  habitat quality (such as length of intact corridors of natural riparian vegetation), water quality and
  temperature, hydrology that mimics natural variations, the extent of connectivity between floodplains
  and the river, nutrient balance, presence of sustainable native populations, and the like. Taken
  together,  these effects measures effectively describe the condition of the watershed, i.e., whether
  the watershed, in fact, can sustainably support native populations and their habitats and maintain
  ecological functions - and they therefore report on the aggregate results of the IED program.

         Each of these effects measures can also be used as to evaluate the success of individual
  actions within the IED program. Following the example above, measures of length of intact riparian
  comdor, water temperature, and nutrient balance would be used to assess the success of specific
  actions taken to restore riparian zones.

         The time frame required to see changes in the effects measures will vary. Some, such as
  population levels of short-lived species of interest, may be detected after only one year of a
  management regime that alters pollutant inputs and water releases to the system. Other
  environmental responses may take more than ten years to be detectable.

         In addition to  the effects measures and condition assessment, direct reporting on the
  decreased pressure from various stressors will be useful. In the example above, such stressor
  measures might include the decrease in the  rate of new riparian damage and increases in the
  release of cool water  from dams in the summer.  Stressor measures relating to other actions that are
  also part of the IED program might include reductions in ambient pollution levels, decreases in the
  number of unscreened pumps that injure fish, and the restoration of periodic flood flows.

         Finally, process measures will provide insight into the level of effort expended and provide a
  shorter-term indicator that the program is proceeding as planned. Examples of process measures
  might include conservation easement contracts signed for the management of riparian corridors,
  changes in the regulations governing water releases from dams, and numbers of water pollution
  permits updated with  new effluent limits.
                                              7-6

-------
      IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

1           Although the focus of this chapter is on evaluating decisions on the basis of
2     environmental outcomes, integrated environmental decision-making may seek to
3     achieve other goals such as environmental justice, quality of life, inter-generational
4     equity, or public "right to know."  In such cases, it will be important to define
5     performance measures to evaluate these other categories of outcomes, referred to in a
6     recent NRC report (1996) as social, economic, and ethical outcomes.
7
8
                                            7-7

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1     7.3 Improved Report Cards
 2
 3       7.3.1 Employing a Broader Range of Measures
 4
 5           In order to provide a meaningful assessment of environmental program
 6     performance, it is critical that EDRCs provide accurate information not only about
 7     process measures (such as how many industrial units now have permits and are
 8     meeting air pollution requirements, or how much money has been spent controlling
 9     water pollution), but also on environmental and health outcomes. Moreover, the
10     specific measures selected should include those that are related to desired states of
11     human health or ecological condition (such as the proportion of at-risk children that are
12     free from pollutant-related respiratory diseases, or the maintenance of native
13     biodiversity in a stream).
14
15           In cases where outcome measures are not available, or not measurable in the
16     short term, the selection of the most appropriate measures of impact of an
17     environmental decision should be influenced by the strength of the association between
18     the measure and the desired environmental outcome. If, for example, exposure to a
19     chemical is identified as the causative agent for a particular disease, and exposure is
20     closely correlated with ambient concentrations of a stressor, then a stressor measure
21     may be an adequate predictor of resultant health effects. In other cases (e.g.,
22     persistent bioaccumulative toxics), environmental concentrations in a particular medium
23     may not be related clearly to human dose or exposure. In such cases, direct measures
24     of exposure (e.g., fish tissue contaminant levels) or of the incidence rate of adverse
25     effects in populations at risk (e.g., subsistence fishers or fish-eating wildlife)  may be
26     needed to evaluate progress toward the desired environmental outcome.
27
28           The SAB urges the Agency to place increased emphasis on outcome measures,
29     developing new ones where required, because the ultimate goal of integrated
30     environmental decision-making is the actual reduction of risks and adverse effects. We
31     recognize that environmental outcomes, whether measures of health, ecological
32     condition, or quality of life, may not be observable over short time frames,  so that
33     process, stressor, and exposure measures will continue to be important in
34     environmental decision-making and management. Outcome measures, however, are
35     an important means of verifying the theoretical basis for the control or abatement

                                             7-8

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1      options chosen. In other words, if the postulated relationship between stressors and
 2     effects is inaccurate, then stressor or exposure reductions might not produce the
 3     expected improvements in adverse effects or condition. A robust reporting system will
 4     include a mix of process, stressor, exposure, and outcome measures.
 5
 6           Even the Agency's traditional list of process measures will need to be expanded
 7     in order to evaluate the performance of some of the new approaches to environmental
 8     management. In recent years, the Agency has added more tools to the environmental
 9     toolbox, bringing to bear such new approaches as economic incentives,  negotiated
10     agreements, and the like (see Chapter 6). Therefore, the collection of data to evaluate
11     these new approaches should evolve as well. There are several important inputs
12     currently missing from most reporting systems that would provide valuable information
13     for EDRCs that assess how well environmental programs have worked and what
14     changes or adjustment might be made. In the case of a marketable permit system, for
15     example, systematic reporting about the number of transactions, the average price per
16     unit of the traded commodity, the number of market participants, the net reduction in
17     pollution output, and average marginal cost per unit of pollution reduced should be
       compared with initial predictions. To the extent that a risk management program
19     involves more than one tool, the entire panoply of actions should be reviewed for
20     efficacy, cost, and other measures of performance. Separate information should also
21     be provided on the "non-quantitative" inputs to decision-making, including the effects of
22     the risk reduction program on sustainabilrty and equity.
23
24           While all of these parameters are often reported somewhere, seldom are they
25     reported in the same place, in terms understandable to the public, and in a format that
26     provides useful input into the next iteration of decision-making on a particular issue.
27     This feedback loop is essential to the functioning of the proposed IED framework.
28
29           In summary, over the years the Agency has developed a traditional set of
30     systems to track administrative process measures and stressor measures. However, in
31     addition to these measures, the IED framework  emphasizes the need to strengthen
32     reporting by a) expanding the range of measures used to evaluate the performance of
33     new (non-traditional) environmental management approaches, and b) shifting from the
34     traditional focus on process and stressor measures to a more  balanced suite of
35     measures, including measures of exposure and environmental outcomes.

                                            7-9

-------
        IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
  7.3.2 Reporting on Environmental Outcomes

       7.3.2.1 Environmental Health

       Setting up an outcomes-based reporting
system requires an understanding of the chain of
events linking human activities to stressor levels to
effects of concern (whether health, ecological, or
quality of life); hence the need for process, stressor,
exposure and especially effects measures in an
EDRC. The expected response time for various
measures along this chain of events will vary. Thus,
the time frame for monitoring and reporting on suites
of measures  needs to be established based on the
time expected for changes to occur.

       In looking at current reporting systems, it is
clear that the nation has much more fully developed
systems for gathering and analyzing information on
outcome measures for human health than it does for
ecological condition. The Department of Health and
Human Services (HHS) -the second largest
Department in the federal government - oversees
an extensive  effort to fund and coordinate
information collection and analysis at local, state,
and national  levels, which provides input on
measures of  effects and condition for human health.
The Department has introduced a number of useful
approaches in this regard.  For example, the HHS
effort to establish national human health goals—Healthy People 2000—and to track
progress towards these goals (HHS, 1990; 1995) using outcome measures has served
as a model for the Agency's own  effort to articulate environmental goals (EPA, 1996).
In addition to HHS, there are many other groups - at various scales - that also
periodically release studies of health effects/trends in different geographic areas; e.g,
World  Resources Institute, World Watch, and the World Health Organization.
Reporting on Quality of Life

Human welfare and quality of life
(QOL), defined in terms of non-
health related impacts from
environmental exposures or
ecological change, is inextricably
linked with both ecological condition
and human health issues.  Progress
toward achieving QOL goals can
and should be measured in an
integrated performance evaluation
system. In some cases, QOL
measures will relate to exposures to
chemical agents (e.g., nuisance
odors), and thus might be included
in an  overall report on health. Other
indicators of QOL will denve from
the consequences of a change in
ecological condition (e.g., changes
in biological diversity, availability of
recreationally important species),
and might be reported in concert
with measures or indicators of
ecological effects. Yet other QOL
measures may relate to economic
consequences of environmental
exposures or ecological changes
(e.g.,  changes in property values,
job impacts of changes in the
availability of harvestable natural
resources).
                                              7-10

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1            However, current systems often provide fragmented health information, such as
 2     data on only the most serious outcomes; e.g., deaths from chronic obstructive
 3     pulmonary disease, rather than total incidence of respiratory problems. Many important
 4     health outcomes, such as reproductive health or neurologic health of children, are not
 5     well captured by these systems. In many cases, environmental data are collected
 6     without consideration of how they might be used to understand the role of the
 7     environment in reported health outcomes or how the data might be used to attribute risk
 8     from exposures to environmental stressors. As a result, these data are of limited utility
 9     in assessing the relative contribution of environmentally mediated factors relative to
10     other major drivers, including nutrition, education/lifestyle, economic status, and
11     genetics.
12
13           We recommend, therefore, that the Agency work with HHS and other
14     governmental and non-governmental organizations to develop an assessment that
15     examines the impact of exposures to environmental stressors on human health, with a
16     particular - but not exclusive - focus on those stressors that fall within the Agency's
17     sphere of influence. Emphasis should be placed on measuring total or aggregate
       exposures to various stressors, i.e., by all routes of exposure — inhalation, ingestion,
19     and dermal contact. Such an "environmental health" report card could serve as a
20     component of a larger "human health report card" released by some of these larger
21     organizations which could eventually attribute health outcomes to specific
22     environmental risks or combinations of such risks. Any system that is developed should
23     be systematically scanned for changes that may indicate problems. Reviews at present
24     are episodic and so may not recognize early changes in health indicators.
25
26           Efforts to report through EDRCs on the effectiveness of environmental protection
27     programs in reducing human health risks would benefit from the development of a
28     conceptual framework describing the relationship between goals, interventions, and
29     desired outcomes.  Evaluating the effectiveness of an environmental intervention on
30     human systems will require consideration of the same factors that are used to
31     determine risk; namely the measure of a health effect (outcome), the measure of
32     exposure to identify potential stressors of concern, and the establishment of human
33     dose. The health outcome would be the endpoint of choice for evaluation.  However,
34     when an effect takes many years to occur (e.g., cancer), measures of early steps along
35     the path to the outcome or even exposure measures should be used for evaluating the

                                            7-11

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1     effectiveness of interventions.
 2
 3           At the same time, we need to be aware that implementation of an EDRC
 4     reporting system for human health risk reduction will be hampered, in many cases, by a
 5     lack of data in key areas; e.g., exposures, differential susceptibilities among
 6     subpopulations, and incidence of non-fatal diseases. It is important to recognize that
 7     there are critical gaps in currently available data (discussed further in section 7.4). At
 8     the same time, these gaps should  not limit efforts to collect the data that do exist to
 9     evaluate specific programs today or constrain the consideration of desirable
10     performance measures for which data might be collected in the future.
11
12           Even where data on environmental health are available, the challenge of relating
13     information on exposures and effects remains. Current data often are not sufficient to
14     link exposures in populations to health outcomes.  Occupational/workplace exposures
15     are more easily dealt with in this regard than residential exposures.  Geographic
16     mapping systems are a potential tool for linking residences to exposures.  However, the
17     usefulness of mapping approaches is still dependent on the precision possible for
18     assessing exposures by exact location.  Obtaining such precision is frequently a difficult
19     task in rural and suburban areas.  However, efforts are underway to improve the
20     recording of location for both exposures and, in some cases, outcome data, which will
21     improve the ability to match and model exposure and outcome data in the future.
22
23           To further complicate matters, the susceptibility of populations is known to vary,
24     and science is moving rapidly to determine what constitutes these susceptibility
26     characteristics. At present, we can judge susceptibility by general characteristics  such
26     as age, race, and sex, but even these have not been used extensively to judge the
27     differences in risks. An important challenge lying ahead is to develop methods and
28     reporting systems that use traditional exposure information, population characteristics,
29     and rapidly expanding data on biomarkers of susceptibility and exposure to expand our
30     ability to be more specific in  identifying subsets of the general population that are at risk
31     of disease.
32
33
                                             7-12

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1           7.3.2.2  Ecological Health
 2
 3          The focus of the I ED framework is on improving the environment by making
 4     individual - yet integrated - environmental decisions and tracking their impact on
 5     human and ecological health through appropriate measures consolidated and reported
 6     in an EDRC.  In this regard, discussions of EDRCs for ecological health and human
 7     health are quite similar. However, in the case of ecological health, the particular
 8     outcomes of concern are less obvious, certainly to the public. Whereas a human health
 9     outcome of reduced childhood asthma is easily grasped and enjoys a broad consensus
10     as a desired goal, some ecological outcomes (e.g., increased biodiversity) may be more
11     esoteric. Thus, the Agency, working with the scientific community, should provide
12     leadership in defining specific ecological outcome measures that can be used to
13     quantify progress toward the widely held goal of protecting ecological integrity.
14
15          This challenge of clearly articulating ecological outcomes should be confronted
16     early, in the Problem Formulation Phase of the IED.  Through open and frank
17     discussion among assessors, managers, stakeholders, and the public, common
       agreement should be reached on the desired outcomes and the risk reduction options
19     of choice. In addition, the selection of ecological outcome measures should be guided
20     by a conceptual model, also developed in the Problem Formulation Phase, that relates
21     expected changes in specific environmental outcomes to anticipated changes in
22     stressors resulting from specific management approaches or interventions. The
23     Problem Formulation Phase should also settle on the measures or indicators  of those
24     changes that will be monitored and reported. Generalized conceptual models of
25     various ecosystem types and landscape units may be used to provide a scientific
26     foundation for the conceptual framework.
27
28           While easy to give, in the past this advice has been  difficult to follow. But there
29     is reason to believe that much more is possible today than was possible just a few
30     years ago. For instance, the idea of a conceptual model is working its way into the
31     mainstream (Ref: EPA Eco RA GLs).  The notion of appropriate monitoring endpoints
32     is being integrated into ecological risk management plans (Ref: South Florida).  The
33     concept of adaptive management to address ecological problems is becoming more
34     commonplace (Ref: SETAC on adaptive management).  The practice of stakeholder
35     involvement has gained wide acceptance (Ref: Comm on RA/RM)

                                            7-13

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1           Therefore, the Agency should vigorously pursue the development of ecological
 2     EDRCs to evaluate the performance of its integrated environmental decisions. At the
 3     same time, the Agency should be aware of and support - by contribution to and
 4     utilization of the results of - efforts to generate report cards to assess the overall
 5     condition of the nation's environment.  These efforts can incorporate the results of
 6     individual EPA risk reduction actions in the process of addressing the broader questions
 7     of ecological health; e.g., "Are our landscapes and ecosystems functionally and
 8     structurally intact, supporting a broad range of native animal and plant communities, or
 9     not?"  The National Science and Technology Council recently concluded that
10     comprehensive assessment of the nation's natural resources will require an integrated
11     national strategy for environmental monitoring and research (CENR, 1997). An ongoing
12     project by the Heinz Center, funded by the White House's Office of Science and
13     Technology Policy, is intended to produce a national environmental report card,
14     integrating environmental program results across the federal government, by  the year
15     2001  (Heinz Center, 1998).
16
17           The choice of appropriate ecological outcome measures should be guided by a
18     conceptual model that relates specific environmental goals and sub-goals to
19     management approaches or interventions. The conceptual model should, in  turn, link
20     expected changes in stressor levels to expected changes in human or ecological
21     condition associated with specific risk  reduction actions. In order to assess ecological
22     condition, it will also be important to select measures of ecological attributes for which
23     cause-and-effect relationships with stressors (at least those being managed)  have not
24     been  postulated. The model should also indicate the measures or indicators of that
25     change that will be monitored and reported. In addition, it is important to evaluate
26     comprehensively the essential characteristics of landscapes and ecosystems across
27     levels of hierarchy and time scales in a way that will allow aggregation and
28     disaggregation of information for a variety of geographical boundaries.  In order to
29     provide a scientific foundation for the conceptual framework, it would be useful to
30     employ generalized  conceptual models of various ecosystem types and landscape
31     units.
32
33
34
35

                                            7-14

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1       7.3.3 Additional Design Considerations
 2
 3          Environmental Decision Report Cards (EDRCs) should exhibit the following
 4     general characteristics:
 5
 6          a) Transparent: First and foremost, each EDRC should be derived in a clear
 7     and methodical fashion from objective measurements. Whenever expert judgment is
 8     required to aggregate components of the report card (such as aggregating disease
 9     information on separate subpopulations or aggregating information regarding the quality
10     of different habitats within an ecosystem), the method of aggregation should be laid out
11     explicitly.  If algorithms are used to combine different types of measurements, the
12     values assigned to the elements in the algorithm should be documented and justified.
13     Ideally, EDRCs should include the results of sensitivity analysis and uncertainty
14     analysis.  Clear documentation of the algorithms used also helps to assure continuity
15     over time, so that report card results from one decade can be fairly compared to the
16     next since for some outcomes, such comparisons will need to be made over many
17     years, even decades or generations.

19           b) Relate closely to environmental goals: Measures included in the report
20     card should be directly related to the specific goals of the  risk reduction program.  In
21     other words, the design of the report card should allow one to understand how changes
22     in performance measures relate to the attainment of environmental goals and vice
23     versa. Although different versions of the EDRC (with differing amounts of technical
24     detail and aggregation) will be needed by the various audiences and users of
25     performance information, the underlying scientific framework must be consistent.
26     Ideally, the logic upon which the EDRC is based should be both systematic and
27     hierarchical so that the information can be a) aggregated- at least qualitatively~to a
28     national or international scale, and b) disaggregated to meet regional and local needs in
29     other report card exercises.
30
31           c) Build a continuing historical record: The EDRC should set forth a frame of
32     reference for each important ecological, human health, or quality of life goal, indicating
33     its past and current status, desired long-term condition, and shorter term milestones.
34     To be truly effective, the EDRCs must be maintained and compared from time to time in
35     order to (1) determine the trend, (2) become aware of new factors that might be

                                            7-15

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1     impacting on the goal, and (3) learn the lessons that adaptive management has to
 2     teach.
 3
 4           d) Provide an integrated picture: The report should be organized to convey a
 5     more integrated picture of health (e.g., a healthy terrestrial ecosystem), rather than just
 6     progress toward stressor/pollutant-specific goals; e.g., a target effluent level. In
 7     commenting on the draft EPA document, Environmental Goals for America (EPA,
 8     1996), the SAB noted that the structure of the draft report was not ideal for this purpose
 9     because it was "centered on the current program office- or media-specific issues" (SAB,
10     1997).
11
12           When the IED framework is applied to multiple environmental risks, the
13     performance measures selected during Phase II and monitored during Phase III may be
14     tied to specific risk reduction tools.  Indeed, designers and implementers of risk
15     reduction programs require information on the relative effectiveness of selected risk
16     reduction approaches in meeting risk reduction targets, as assessed against a number
17     of decision criteria (e.g., timeliness, cost-effectiveness, and equity). Such
18     reassessment requires that risk reduction be attributed to specific risk reduction tools, to
19     the extent possible.  At some point, however, it will also be important that performance
20     measures be reported in combination in an integrated report card so that overall
21     progress toward achieving a goal can be assessed.
22
23           e) Credible:  An additional consideration in the design of report cards is the
24     question of who will generate them. It is valuable, from the standpoint of learning, for
25     those implementing the risk reduction measures (whether in the public or private sector)
26     to generate such reports. It is possible, however, that information will be gathered and
27     reported by some group not directly involved in the implementation. Such a scheme
28     has the advantage of independent reporting but reduces the opportunity for learning by
29     those implementing the plans. A third possibility is for the report cards to be prepared
30     by those implementing the plan (self-reporting), combined with an outside audit
31     program by auditors not involved in the implementation. This latter combination has the
32     advantage that those implementing the plans leam as they produce their reports; they
33     also keep their eyes on the yardsticks set up for measurement as they implement
34     plans; at the same time, the problems that can frequently arise with self-reporting
35     schemes are counteracted by the periodic audits.

                                             7-16

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote


 1
 2     7.4 Implications for Existing Monitoring Systems
 3
 4          Decisions on the design of a report card system have implications for information
 5     collection systems, both those that track administrative processes and those that collect
 6     environmental data (e.g., ambient monitoring programs). In selecting performance
 7     measures, there will often be a gap between what it would be desirable to know and
 8     what it may be feasible to know. This gap may exist as a result of several factors,
 9     including the limited state of knowledge about the relationship between stressors and
10     effects, the costs involved in obtaining certain types of information, and the willingness
11     of affected people to provide information. Nonetheless, it is important that careful
12     consideration  of the most desirable performance measures, including those based on
13     an established chain from process and stressor measures to exposure and outcome
14     measures, influence the types of information that are actually collected.
15
16           Existing monitoring systems, both those geared toward environmental condition
17     and those intended to assess human health and welfare, provide invaluable
       information, particularly regarding the implementation of current legal mandates.
19     Compliance monitoring and ambient air and water quality monitoring networks, for
20     example, provide significant process as well as some outcome data. The EDRCs and
21     overall report  cards designed to answer broader questions, however, will likely require
22     changes to existing monitoring systems.
23
24           Ecological Health Outcomes:  Assessments of ecological integrity, for
26     example, require information not only on water quality, air quality, soil quality, water
26     flow, and populations of certain species, which are commonly monitored today, but also
27     measurements of biological community structure, the presence of successional states,
28     diversity of habitat types across the landscape, connectivity, altered topography,
29     hydrodynamic patterns, and so forth. This latter set of parameters is not typically
30     monitored by EPA, but has received greater attention by other federal agencies (USGS,
31     USFS). This  fact emphasizes the importance of strengthening and maintaining the
32     collaboration  among the many agencies that conduct monitoring, an effort begun
33     several years ago under the auspices of the Committee on Environment and Natural
34     Resources (CENR, 1997).  Linking monitoring efforts into a coherent framework will
 35     improve the quality and robustness of the EDRCs.

                                            7-17

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1
 2           Human Health Outcomes: As noted above, much of the data, both on
 3     environmental exposures and on health outcomes, needed to evaluate the
 4     effectiveness of a risk reduction program are not currently being collected. The
 5     measurement of disease for the total population exists only for diseases recorded on
 6     death certificates. Mortality data do not provide information on the incidence or the
 7     occurrence of disease, but only death from a specific disease. The incidence of cancer
 8     is now being recorded in many states (ref CDC Cancer Registry Program), but the
 9     cases cannot be tied to exposures because of the long latency of the disease. Still
10     more important, many health outcomes today are associated with environmental
11     exposures but records are not routinely kept on these diseases for the general
12     population. For example, asthma episodes, behavioral changes in children, and minor
13     neurologic deficits are known to be important effects of various exposures, but there is
14     no complete database regarding these outcomes that could serve to link these effects
15     to specific exposures. Testing for respiratory function, intelligence, hearing, and vision
16     is done on many segments of the population, but the information resources are
17     fragmented and have not been used to evaluate the effects of environmental change.
18     The available data linking environmental exposures to health outcomes have come
19     from special studies on limited and often selected populations.  Further, these
20     outcomes are usually evaluated using surveys designed to detect trends, rather than to
21     look for potential causes.  For many problems, only certain subpopulations are at risk,
22     either because of their susceptibility or special lifestyles that may increase exposures.
23     Little information is available specifically on these subsets to allow for evaluation.
24
25           Exposure: The databases on human exposure have often been limited by
26     infrequent national sampling sources, e.g., the Center for Disease Control's National
27     Health and Human Nutrition Examination Survey (NHANES),  and the Total Exposure
28     Assessment Methodology (TEAM) study (EPA, 1987). States have far more data
29     available but these data often are not in electronic format and available for easy use.
30     Air and water monitoring data are only indirect measures of exposure and reflect only a
31     limited number of monitoring stations. In some situations, only non-compliance data
32     are recorded. This practice limits the usefulness of the information for modeling
33     exposures of individuals over space and time. Research to develop a better
34     understanding of the relationships between centrally located monitoring sites and the
35     actual human exposures could make some of these data more useful. Ultimately,

                                            7-18

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1     however, more direct monitoring of human exposures will be required to demonstrate
 2     the effectiveness of various risk management strategies.
 3
 4          Measurements of pollutants in environmental media are less intrusive than
 5     human exposure measurements.  However, the knowledge needed to link exposure of
 6     individuals to measures of the pollutant in various environmental media has not been
 7     well developed. For example, we have monitored outdoor air for many years.
 8     However, to use such data for exposure assessment, we must have information on
 9     time-activity patterns and the penetration of outdoor air pollutants into buildings since
10     most of the population spends about 90% of the time indoors. In addition, exposures
11     to indoor pollutant sources (e.g., environmental tobacco smoke) will add to those from
12     outdoor air, and will also contribute to adverse health effects. Similarly, measurements
13     of a pollutant in food supplies may not capture changes, both losses and gains, that
14     occur during food preparation in homes and restaurants.  Ultimately, considerably more
15     measurement of exposures will be required, as well as measurement of changes in
16     exposure over time, in order to meet the challenges of performance evaluation.
17
             In short,  the concept of report cards will challenge assessors to make the most
19     out of the data that are available and, perhaps, to re-define and re-design the systems
20     in place to gather environmental data. The goal is to answer questions that people
21     want to have answered.
22
23     7.5 Summary and Recommendations
24
25           The extent of environmental data collected (e.g., on contaminant levels in
26     various environmental media, on tissue burdens, on human exposures, and disease) is
27     expansive and holds the promise of improved measures to generate effective
28     Environmental Decision Report Cards (EDRCs). At the same time, in order to achieve
29     this promise, the current information reporting systems would benefit from additional
30     development and,  in some cases, restructuring.  An improved performance evaluation
31     system will be characterized  by clear articulation of objectives during the Problem
32     Formulation (Phase I), careful selection of performance measures in Analysis and
33     Decision-Making (Phase II), and refinement, integration, and possible expansion of
34     existing data collection efforts in Implementation and Performance Evaluation (Phase
35     III).

                                           7-19

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1           In closing, the SAB would like to direct the Agency's attention to the promise and
 2     challenge of environmental reporting as discussed in this report. That is, reporting in
 3     the context of the IED framework would move the Agency further along the spectrum of
 4     measures illustrated in Figure 7-1, increasing the attention to observable environmental
 5     outcomes from specific management programs. In addition, the challenge of combining
 6     information from a number of integrated decisions into an overall environmental report
 7     card remains.  With respect to this next generation of performance evaluation, the SAB
 8     offers some initial observations:
 9
10           a)  Implementation of a more integrated form of environmental decision-making
11     will require an improved environmental report card system that provides reliable and
12     accurate information about the issues people care about: how healthy are the nation's
13     citizens? is the integrity of our ecological systems intact? is the quality of life getting
14     better or worse? In order to answer these questions effectively, government agencies
15     and the public need to develop a reporting system that focuses on the desired states of
16     human health and ecological condition, rather than program-specific objectives.
17
18           b) Such a system should build on the report cards devised to evaluate specific
19     integrated decisions, along with information from other sources, to report on the overall
20     human health and ecological condition.
21
22           c) In this regard, the Agency, in concert with others in the National Science and
23     Technology Council, can play a leading role in designing the improved reporting system
24     by developing  a conceptual framework that: (1) defines ecological integrity and desired
25,    human health  results in terms of measurements that can be used at the local, state and
26     national levels, and (2) provides a systematic and transparent methodology for
27     aggregating these measurements into an overall picture of health and integrity.
28
29           d) Research on selected report card  components, such as methods of accurately
30     measuring certain ecological characteristics at the landscape scale, methods of
31     identifying susceptible human subpopulations at risk for disease, and methods for
32     combining these into more aggregated measures of performance will be required and
33     should proceed concurrently with the development of an overall report card framework.
34
35           e) Specifically, current data collection and analysis systems should be upgraded

                                             7-20

-------
      IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

1     so that information can be assembled in a format that relates directly to the conceptual
2     framework, as well as the EDRC, and that lends itself to aggregation along a variety of
3     geographical boundaries.
4
5           f) The Agency should provide a results-based environmental report card regularly
6     to the public and to policymakers, and use such a report card — along with specific
7     EDRCs — as a tool to support re-evaluation of program implementation within the
8     Agency.
9
                                            7-21

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

  1     7.6 References Cited
  2
  3     Committee on Environment and Natural Resources. 1997. Integrating the Nation's
  4           Environmental Monitoring and Research Networks and Programs: A Proposed
  5           Framework. National Science and Technology Council, Executive Office of the
  6           President, Washington, DC.
  7
  8     Enterprise for the Environment. 1997. The Environmental Protection System in
  9           Transition: Toward a More Desirable Future.
10
11     Environment Canada and U.S. Environmental Protection Agency.  1997. State of the
12           Great Lakes: 1997-The Year of the Nearshore. Great Lakes National Program
13           Office, Chicago, IL (EPA 905-R-97-013).
14
15     National Research Council. 1996.  Understanding Risk: Informing Decisions in a
16           Democratic Society. National Academy Press, Washington, DC.
17
18     Science Advisory Board.  1997. Review of the Environmental Goals for America: With
19           Milestones for 2005. (EPA-SAB-EC-97-007)
20
21     U.S. Department of Health and  Human Services.  1990. Healthy People 2000.  Office
22           of Disease Prevention and Health Promotion, Washington, DC.
23
24     U.S. Department of Health and  Human Services.  1995. Healthy People 2000
25           Midcourse Review and 1995 Revisions. Office of Disease Prevention and Health
26           Promotion, Washington,  DC.
27
28     U.S. Environmental Protection Agency. 1987. The Total Exposure Assessment
29           Methodology (TEAM) Study: Summary and Analysis: Volume 1. U.S. EPA Office
30           of Research and Development, Washington, DC, EPA/600/6-87/002a, June
31           1987.
32
33     U.S. Environmental Protection Agency. 1996. Environmental Goals for America: With
34           Milestones for 2005 (December 1996 draft).
35

                                           7-22

-------
PART VI CONCLUSIONS AND RECOMMENDATIONS

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 i            CHAPTER 8: CONCLUSIONS AND RECOMMENDATIONS
 2
 3
 4          Over the course of the Integrated Risk Project, the Steering Committee and the
 5     project subcommittees came to a number of conclusions about the need for and
 6     benefits of improved methods of assessing and managing aggregate risks, broadening
 7     benefit/cost analyses to include difficult-to-monetize values, explicitly considering the
 8     role of values in decision-making, and so forth. The preceding chapters include
 9     suggestions and recommendations for specific improvements in the methodologies and
10     approaches that underlie the IED framework. In addition, however, the Steering
11     Committee came to a number of broad conclusions as a result of the discussions and
12     thinking that led to the proposed IED framework itself. These overall conclusions and
13     recommendations, which are the focus of a companion document, Integrated
14     Environmental Decision-making in the 21st Century:  Summary Recommendations, are
15     summarized below:
16
17          Recommendation 1: EPA should accelerate the transition to integrated,
            outcomes-based environmental protection and apply the an integrated
i*          environmental decision-making framework in selected cases, while
20          maintaining the safeguards afforded by the current system.
21
22          Previous environmental management approaches, both regulatory and non-
23     regulatory, have resulted in substantial improvements in human health and ecological
24     condition, particularly in regard to chemical risks. While maintaining the current
25     regulatory and management structure, the IED framework offers an approach for going
26     beyond current protections by taking a truly integrated look at risks, opportunities for
27     risk reduction, and the consequences of addressing (or failing to address) those risks.
28     The focus of this next generation of environmental decision-making should be on
29     environmental results; that is, on demonstrable outcomes (improvements) in the
30     environment resulting from integrated action. In order to further incorporate integrated
31     environmental decision-making into EPA programs, the Agency should select a few test
32     cases where the  IED framework can be applied explicitly. Such test cases offer the
33     best means for demonstrating the potential of the IED framework, identifying needed
34     improvements in  its underlying methodologies, and realizing its benefits with real-world
35     problems.
                                           8-1

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1
 2           Recommendation 2: Because science plays a critical role in protecting the
 3           environment, EPA should commit the resources necessary to expand the
 4           scientific foundation for integrated decision-making and outcomes-based
 5           environmental management.
 6
 7           An important theme of the Integrated Risk Project is the need to integrate
 8     scientific and technical information on risks and risk reduction opportunities with
 9     information on societal values, aspirations, and goals. The call for greater inclusion of
10     multiple disciplines and points of view, however, must not obscure the fact that science
11     and scientific methods are critical to sound environmental decision-making. Science
12     has a unique and critical role to play in protection of the environment. It is through
13     scientific investigations that many environmental problems are first discovered (e.g., the
14     depletion of stratospheric ozone, hypotheses about environmental endocrine
15     disrupters). Science also is instrumental in  developing, testing, and evaluating risk
16     reduction options, and social sciences offer techniques for assessing societal
17     preferences and wants. Implementation of the 1ED framework, particularly its
18     application to multiple-stressor problems, will in many cases require new science, both
19     empirical and theoretical. In order to gain the greatest benefit from the IED approach,
20     the Agency will need to invest in research designed to support integrated risk
21     assessment and management.
22
23           Recommendation 3: EPA should apply and encourage the broader use of
24           risk comparison methodologies, such as those described in this document,
25           that clearly identify how scientific  information and judgment are
26           incorporated into risk comparisons.
27
28           Scientific information on risk, such as quantified risk assessments and
29     scientifically demonstrated relationships between stressors and effects, provides the
30     essential basis for making objective risk comparisons. However, desired scientific
31     information often is incomplete or absent, and scientists have to use their best
32     professional judgment to bridge important gaps in the data.  In other cases, scientific
33     information and analysis, by themselves, are not sufficient for comparing risks; e.g.,
34     when comparing  human healths risks vs. ecosystem risks or cancer risk in adults vs.
35     neurologic risk to children.  In these instances, it is public values that come into play in
36     making the comparisons. Thus, methodologies used to compare environmental risks-

                                            8-2

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

      whether human health risks, ecosystem risks, or both-should incorporate not only the
 2     most up-to-date scientific information, but also should identify explicitly where
 3     professional judgment and values have influenced the results. The SAB has developed
 4     two prototype methodologies for risk comparisons that meet these criteria. These
 5     approaches should be points of departure for the Agency as it seeks to develop and
 6     use science-based methodologies for comparing relative environmental risks.
 7
 8            Recommendation 4: EPA should use a broader range of risk reduction
 9            options in combination to manage environmental risks.
10
11            In 1990, the SAB recommended in Reducing fl/sfr that the nation make greater
12      use of all the tools, including market forces, information, and product specifications,
13      available to reduce risk. Now, almost a decade later, many of those tools are being
14      used to a greater extent than ever before. The challenge for EPA is not only to expand
15      the use of those various tools, but also to use them in creative, coordinated ways to
16      reduce multiple risks to multiple receptors in communities and ecosystems across the
17      country. The SAB has developed a risk reduction options methodology that can be
1 °      applied fruitfully to single stressors. However, the methodology should be especially
       useful for identifying multi-dimensional strategies to control complex environmental
20      problems involving many sources, stressors, and receptors.
21
22            Recommendation 5: When evaluating risk reduction options, EPA should
23            weigh the full range of advantages and disadvantages, both those
24            measured in dollars as costs and benefits and those for which there may
25            not be a comprehensive dollar measure, such as sustainability and equity.
26
27            The valuation framework that underlies environmental decisions is the simple
28      formulation of whether the gains that accrue from protective actions are worth what is
29     given  up to attain them. There is a subsidiary question: would other possible actions be
30     preferable? This question highlights the importance of taking all effects, including long-
31      term effects, into account.  It also raises the sometimes hidden issue of how people
32     value  different aspects of the environment.  Some of society's environmental values
33     can be measured directly in monetizable terms, and others can be inferred and
34     translated into monetizable terms with some confidence.  But other things that people
35     value, such as sustainability and equity, often may be expressed only qualitatively, yet
36     are of no less importance for that reason. As a part of this project, the SAB has

                                            8-3

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1      described the applicability and limitations of the benefit/cost framework, and suggested
 2      areas where new approaches to characterizing values are essential. The SAB also has
 3      attempted to define better the full range of relevant questions that must be considered
 4      in ecological valuation.
 5
 6            Recommendation 6: EPA should make fuller use of the scientific methods
 7            available to characterize public values and incorporate those values into
 8            goal-setting and decision-making.
 9
10            Community and national values have been and will continue to be a primary
11      driver of community-level and national efforts to protect the environment. However,
12      values usually are not weighed transparently in the decision-making process. Rather,
13      they are usually implicit in the judgments made by decision-makers. Thus, they
14      influence decisions in ways that are not clear to, or renewable by, the public.
15
16            Because public values undoubtedly help shape environmental decisions, it is
17      important to understand and document their role in and influence on decision-making.
18      It is also important to elicit public values systematically, differentiate values from
19     technical information as a part of decision-making, and include their effects on
20     decisions as part of the public record. Individual and community values should be
21      solicited systematically  by social scientists and other appropriately trained individuals.
22     The deliberative processes that are used in arriving at decisions should involve
23     professionals trained in fields like consensus-building and dispute resolution. EPA
24     should make more extensive use of existing expertise in the areas of behavioral
25     science and decision logic so that a more complete representation of community values
26     is incorporated into the Agency's decisions.
27
28           Recommendation 7:  EPA should identify, collect, and disseminate
29           scientifically-based environmental metrics organized in new ways to
30           support a more integrated approach to managing environmental risk.
31
32           The transition to and effectiveness of integrated, outcomes-based environmental
33     protection will depend to a large extent on the availability and utilization of appropriate
34     information in the area  of exposure, human health, ecological health, and quality of life.
35     Current information collection  efforts, however, often are insufficient, inadequately
36     organized, or focused on inappropriate endpoints. In the area of human health, for

                                             8-4

-------
      IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

      example, data on non-fatal outcomes of environmental exposures, such as asthma, are
2     not being collected except in a rudimentary manner or at certain sites. Only mortality
3     information is collected systematically at all locations and can be related to some
4     limited information on selected exposures.  With regard to ecological data, more
5     comprehensive information is needed on the current status of ecosystems such as
6     wetlands, lakes, forests, and grasslands. This information should include the extent to
7     which each ecosystem is exposed to  and affected by non-chemical stressors such as
8     habitat conversion, habitat fragmentation, and invasions of exotic species.
9
10            The challenge facing EPA and the nation is not only one of collecting the right
11     environmental data,  but of finding new ways to manage and use that data. Working
12     with other federal and states agencies, EPA should take a leadership role in:
13     identifying critical environmental data gaps, including data on exposures and health and
14     ecological outcomes; integrating the largely fragmented data collection efforts already
15     underway; and disseminating integrated environmental information to decision-makers
16     and the public.
17
i g            Recommendation 8:  EPA should develop a system of "report cards" to
             organize and disseminate information on the status of ecological and
20           human health and the quality of life in order to assess the effectiveness of
2i            its environmental decisions and to guide future environmental
22           management.
23
24           One of the most valuable uses of environmental data is to measure the results of
25     the actions society takes to reduce environmental risk. However, even if such data did
26"     exist, through a vigorous implementation of Recommendation 7, there would still be a
27     need to develop widely accepted and commonly used methods for evaluating a) the
28     state of our environment and b) the success of national environmental protection efforts
29     .. where success is measured in terms of demonstrable outcomes in the health of
30     humans and ecosystems. Regarding the state of the environment, EPA should work
31     with federal and state entities to develop indicators of ecological and human health
32     conditions in our country. Such a environmental report card would be readily
33     comprehensible to the public and policy makers alike, serving to galvanize them on
34     action toward specific, measurable goals.  The concept of a report card should also be
35     applied to evaluating the impact and effectiveness of individual decisions made under
36     the IED methodology; i.e., an environmental decision  report card (EDRC).

                                             8-5

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 1           Reporting on measures of progress towards ecological, human health, and
 2     quality of life goals will provide a number of benefits. First, the exercise of defining
 3     performance measures and reporting on them will bring more focus and discipline to
 4     the Agency by expressing the relationship between investments (measured in time,
 5     money, or information) and environmental results. Second, shorter term measures of
 6     progress (including process measures or measures of stressor reductions) will be
 7     useful for accountability and course correction. Third, longer term measures of
 8     progress (such as improvements in overall human and ecosystem health and quality of
 9     life) will be most helpful in determining whether goals are being met, whether further
10     actions are  needed to control well-recognized stressors, or whether new actions are
11     needed to control new stressors.
12
13           To strengthen their credibility and utility, environmental report cards should
14     contain information derived from objective measurements, be transparent and clearly
15     documented, and provide integrated information on progress towards multiple inter-
16     related environmental goals.
17
18           Recommendation 9:  EPA should expand and develop new collaborative
19           working relationships with other federal and non-federal governmental
20           agencies and others who also will be involved in integrated environmental
21           decision-making.
22
23           Inherent in the I ED framework is the idea that problem formulation and decision-
24     making must match in scale and location. In some cases, decisions may be most
25     effective when local or state governments play the primary role. In others, coordinated
25     action across several levels of government or among a number of state and/or local
27     governments will be required. In still others, the current emphasis on centralized
28     decision-making may be the preferred approach. In any event, integrated thinking
29     about environmental problems will tend to drive decision-making to the agency or level
30     of government where decisions are most appropriately made. As a consequence,
31     EPA's role  will evolve to one in which the depth of control gives way to broader
32     involvement in partnership with other agencies. EPA will continue to be responsible for
33     implementing and enforcing federal environmental laws and statutes, conducting
34     environmental research and development, and conducting stressor-specific risk
35     assessments. At the same time, the Agency can exert national leadership to bring
                                             8-6

-------
       IRP Draft Integrated Report-Peer Review Draft, May 3,1999-Do Not Cite or Quote

 i     together appropriate agencies and stakeholders to explore integrated approaches to
 2     environmental problems.
 3
 4           Recommendation 10:  EPA should aggressively explore options for
 5           reducing risks from significant stressors that currently are addressed
 6           inadequately by the nation's environmental institutions.
 7
 8           Over the course of the Integrated Risk Project, it became clear to the SAB that a
 9     number of important human health and ecological risks are not being addressed
10     adequately by the nation's environmental institutions. In many cases this is because
11     risk management responsibility is not clearly assigned to any one government entity, or
12     is scattered over many agencies and/or levels or government. This fragmented
13     approach results in uncoordinated and incomplete efforts to identify cause-and-effect
14     linkages and to manage those risks. With regard to ecological risks, the SAB has
15     concluded that many of the highest ranking risks (e.g., hydrologic alterations, harvesting
16     of living marine resources, habitat conversion, climate change, and introduction of
17     exotic species) are associated with physical and biological, rather than chemical,
1?     stressors, which do not fall clearly within the purview of any single federal agency.
1      Important human health risks that remain unaddressed include those for which the
20     environmental exposure link is suspected but not certain (e.g., asthma, brain cancer,
21     and non-Hodgkins lymphoma) and risks associated with susceptible and/or
22     compromised human populations.
23
24           To control many of these inadequately addressed risks will require the kind of
25     integrated decision-making envisioned in  this report. It will also required a  new kind of
26     integrated institutional leadership. When EPA determines that serious risks are not
27     being addressed effectively by existing environmental institutions or decision-making
28     systems, the Agency has a responsibility to inform the public about those risks and
29     bring together the appropriate federal, state, and local agencies to address them.
                                            8-7

-------