I I .': '••Afjril:l983/.. I i I .w Benefit-Cost Assessment . U.S. EPA Headquarters Library •;•''"'. . Mail code .3201.' '•"'••; -' • ^ 1200'Pennsylvania Avenue NW s.;" : Washington DC 20460; . Programs Volume I \A f I ------- I I I I if*0 April 1983 I 0- I I I I I I I I I I I I Benefit-Cost Assessment Handbook for Water Programs Volume I Prepared for U.S. Environmental Protection Agency Economic Analysis Division Washington, DC 20460 U.S. EPA Headquarters Library Mail code 3201 1200 Pennsylvania Avenue NW Washington DC 20460 Prepared by Dr. William H. Desvousges Research Triangle Institute Research Triangle Park, NC 27709 and Dr. V. Kerry Smith University of North Carolina Chapel Hill, NC 27514 The information in this document has been subject to the U.S. Environmental Protection Agency's (EPA) Peer and Administrative Review, and it has been ap- proved for publication as an EPA document. Mention of trade names or com- mercial products does not constitute endorsement or recommendation for use. ------- I I I I I I I I I I I I I I I I I i i April 1983 Benefit-Cost Assessment Handbook for Water Programs Volume I Prepared for U.S. Environmental Protection Asency • Economic Analysis Division Washington, DC 20460 Prepared by Dr. William H. Desvousses Research Triansle Institute Research Triansle Park, NC 27709 and Dr. V. Kerry Smith University of North Carolina Chapel Hill, NC 27514 The information in this document has been subject to the U.S. Environmental Protection Agency's (EPA) Peer and Administrative Review, and it has been ap- proved for publication as an EPA document. Mention of trade names or com- mercial products does not constitute endorsement or recommendation for use. ------- I I I I I I I I I I I I I I I I I I I PREFACE As one component of the decision process, water program offices at the Federal, State, and local levels along with advisory citizen groups have per- ceived the need to consider the beneficial and detrimental effects of policies that affect water quality. In response, the U.S. Environmental Protection Agency's Office of Policy Analysis sponsored the development of this Benefit- Cost Assessment Handbook for Water Programs. This experimental handbook will not short-circuit any Federal, State, or local water quality laws. Rather, It provides suggestions about how to evaluate the economic aspects of a pro- posed policy as a regular part of the decision process. These evaluations can identify water quality policies that have highest priority, so that society's resources can be directed to the areas that will have the greatest benefits. They also can serve as one of several analyses that support the decision proc- ess for any specific water quality policy. Such ah approach can help assure the attainment of our Nation's water quality goals with a minimum expenditure of resources. The Benefit-Cost Assessment Handbook for Water Programs is a primer. It assumes only a limited familiarity with economics. Throughout the handbook, case studies help to clarify points. Data needs, key assumptions, and other relevant points are covered for different ways of determining the relationship between desirable and undesirable effects of a program decision. Since water program offices have begun to move toward the use of benefit-cost concepts, the scope of the handbook is broad enough to explain how to conduct benefit-cost assessments in diverse applications. The costs and health benefits of drinking water policies are covered elsewhere, so this handbook concentrates on the benefits and costs for other water programs. Although most of the examples in this volume are for hypothetical water quality standards decisions, the tools can be applied to a broad spectrum of water quality decisions, and even to environmental issues in general. If there is sufficient interest in this experimental approach, a second volume may be developed to provide more in-depth discussion of the benefit- cost assessment for use by practitioners. It also may be desirable to tailor a similar document for specific water programs. The Office of Policy Analysis welcomes comments and suggestions, which may be directed to: Ann Fisher U.S. Environmental Protection Agency 401 M Street, S.W. Office of Policy Analysis (PM-220) Washington, D.C. 20460 (202) 382-2783 iii ------- I I I I I I I I I I I I I I I I I I I ACKNOWLEDGMENTS Why are the acknowledgments always the last section to be written in any report? This question, which often .occurs on the way to work toward the end of a project, is likely one without a definite answer. We could surmise that it comes from an unwillingness to let go at the end of any project (not likely) or simply that we have saved the best for last. "i We are indebted to many people who have given of their time to assist us in writing this handbook. Yet there is one person whose commitment to the project stands out. That person is our project officer at the Environmental Protection Agency (EPA), Dr. Ann Fisher. The simple truth is that without Ann Fisher this handbook would not have been written. With the many hours she has given to the project she is more author than anything else. Since she would never have allowed this acknowledgment to appear .in the report had she read it, perhaps this is the reason for leaving it until last. As this handbook has evolved, it has drawn from many sources. Chapter 4 pertaining to the costs of regulatory actions draws heavily on an earlier chapter prepared by Metasystems, Inc., and particularly Tze-Wen Chi and Peter Morgan. Indeed, most of the examples are taken from this earlier draft. Assistance on the cost chapter came from the Office of Analysis and Evaluation under Louis Dupuis. John Kukulka and Joe Yance from the staff suggested many helpful revisions in support of the project. Office of Policy Analysis staff members provided valuable detailed critiques of early drafts. Reed Johnson, Bob Raucher, Peter Caulkins, and Skip Luken gave good counsel. Joan O'Callaghan's valuable edi- torial suggestions on an early draft have been incorporated into this draft. The EPA staff in the water program area have also played an important role in this project. Patrick Tobin, Dave Sabock, xand Marjorie Pitts of the Office of Water Regulations and Standards, Myron Temins and Charles Moar of the Office of Water Programs Operations (OWPO), and Jerry Manorola, formerly of OWPO, gave extensive comments. In addition, helpful comments were received from Dan Huppert, MNFS Southwest Fisheries Service; Robin Gottfried, University of the South; John Luensman, Department of Planning and Development, Chautauqua County, New York; John Loomis, U.S. Forest Service; and Mike Piette, University, of Hartford. v ------- Finally, the RTI staff has been most supportive. Tayler Bingham brought his experienced view of environmental regulations to bear on all of the drafts of this handbook. Tayler's insightful comments are valued , highly. Hall Ashmore's editorial efforts mark every page of this handbook. Hall's economy with words and his ability to make an author's prose bear fruit are assets for any handbook or report. Finally, Jan Shirley and the Institute's Word Processing Specialists have done yet another outstanding job of producing a pleasing report. Jan's devotion to quality and her ability to know when to coax or cajole a reluctant writer are most appreciated. VI ------- I I I I I I I I I I I I I I I I I I I CONTENTS Chapter Page 1 BENEFIT-COST ASSESSMENT: A COMMON SENSE APPROACH TO DECISIONS 1-1 1.1 introduction 1-1 1.2 Regulation: An Overview 1-2 1.3 Benefits: An Economic View 1-4 1.4 Costs: An Economic View 1-5 1.5 Benefit-Cost Assessement: What Is It? 1-6 .Is Benefit-Cost Assessment Different from Cost-Benefit Analysis? 1-6 Benefit-Cost Assessment: ' A Step-by- Step View 1-7 1.6 Key Considerations in an Assessment 1-11 1.7 Summary 1-12 1.8 Guide to Handbook 1-13 2 ISSUES IN A BENEFIT-COST ASSESSMENT 2-1 2.1 Introduction 2-1 2.2 Including Intangibles in a Benefit-Cost Assessment 2-1 Introduction 2-1 . Example 2-2 2.3 Economic Impact Measures 2-5 Assessing Household Impacts 2-5 .Ability to Pay 2-5 Financial Capability 2-6 • Assessing Industrial Impacts 2-6 Assessing Changes in Employment, Output, and Prices 2-6 2.4 What to Do About Distribution: Problems in Adding up Over People 2-7 VII ------- I I I I I I I I I I I I I I I I I I I CONTENTS (continued) Chapter Introduction 2-7 Example I 2-8 Example II. 2-9 Summary: Distribution 2-9 2.5 Discounting Future Benefits and Costs: Adding up Over Time . 2-10 Time Preference: What Is It? 2-11 How To Determine the Discount Rate in a Less Than Ideal World 2-12 Including Risk in an Assessment of Discount Rates 2-14 What Are the Empirical Implications of the Discount Rate Issues? 2-15 The Simple Mechanics of Discounting . 2-16 Discounting: A Summary Review 2-18 2.6 Summary 2-19 3 MEASURING THE BENEFITS OF WATER QUALITY PROGRAMS 3-1 3.1 .Introduction 3-1 3,2 Categories of Benefits: Ah Overview 3-1 3.3 Practical Concerns: Selecting a Benefit Estimation Method 3-4 3.4 Household Benefits : 3-5 Theory: The Demand Function " 3-5 Practice: Methods for Measuring Household Benefits • 3-6 Travel Cost Method .' 3-6 Data Needs and Key Assumptions and Features 3-8 Case Study: Alternative Approaches for Estimating Recreation and Related Benefits of the Monongahela River . 3-9 viii ------- I I I I I I I I I I I I I I I I I I I CONTENTS (continued) Chapter Survey-Contingent Valuation -. . . . 3-14 Data Needs, Key Assumptions/ Limitations, and Features 3-16 Case Study: Recreation and Related Benefits of Water Quality Improvements in the Mondngahela River 3-18 Survey—Recreation Participation Models 3-20 Data Needs, Key Assumptions/ Limitations, and Features 3-21 Case Study: A Participation Survey Approach to Valuing Water Quality Improvements 3-22 Hedonic Property Value Method 3-25 Data Needs, Key Assumptions/ Limitations, and Features 3-26 . Damage Function Method 3-27 Data Needs, Key Assumptions/ Limitations, and Features 3-27 3.5 Business Benefits 3-27 Theory: The Supply Function 3-27 Practice: Cost Savings Function Method for Measuring Business Benefits 3-28 Data Needs, Key Assumptions/ Limitations, and Features . 3-29 Case Study: Irrigation Benefits . . . 3-30 3.6 Public Water Supply Benefits 3-32 3.7 Summary 3-32 4 MEASURING THE COSTS OF WATER QUALITY PROGRAMS 4-1 4.1 Introduction 4-1 4.2 Measuring Costs: The Basic Concepts 4-1 4.3 Measuring Costs: Two General Approaches ....... 4-3 4.4 Types of Costs 4-4 Capital Costs '. 4-4 IX ------- I I I I I I I I I I I I I I I I I I I .CONTENTS (continued) Chapter Operation and Maintenance Costs 4-5 Reinvestment Costs 4-6 Salvage Value '. 4-6 Summary Cost Measures 4-8 4.5 Practical Issues in Measuring Costs 4-8 Sources of Cost Data 4-8 The Use of Cost Indexes 4-9 Major Factors Affecting Cost Estimates 4-11 Sample Data Form 4-12 4.6 Cost Estimating Techniques 4-12 Components of Treatment Systems 4-12 Estimating treatment Costs 4-15 4.7 Examples Using Cost Techniques 4-17 Example 1: Pulp Mill Using EPA Development Document 4-17 Example 2: POTW Using CAPDET 4-21 Example 3: POTW Using Technology Assessment Manual . 4-25 Example 4: What to Do When There Is "No Information" 4-25 4.8 Final Cautions 4-27 4.9 Summary 4-27 5 COMPLETING THE BENEFIT-COST ASSESSMENT 5-1 5.1 Introduction 5-1 5.2 Sensitivity Analysis: A Gauge to Believability .... 5-1 Introduction 5-1 Example 5-3 Step 1: Translate the Benefits and Costs into Present Values 5-3 Step 2: Perform Sensitivity Analysis for Discount Rate and Key Assumptions 5-4 Step 3: Interpret Sensitivity Analysis 5-4 ------- CONTENTS (continued) Chapter Page 5.3 Displaying the Assessment Results 5-4 Narratives 5-6 Arrays 5-6 Graphs •. 5-6 5.4 Benefit-Cost Assessment Checklist . . . .- 5-9 5.5 Summary 5-9 6 BENEFIT-COST--SAMPLE SCENARIOS 6-1 6.1 Introduction 6-1 6.2 Simple Case Scenario 6-2 Introduction 6-2 Simple Case Scenario Format . 6-2 6.3 Medium Case Scenario 6-4 .Introduction 6-4 Medium Case Scenario Format. . . . ' 6-5 6.4 Complex Case Scenario 6-6 Introduction 6-6 Complex Case Scenario Format .' . . 6-7 6.5 Summary 6-26 7 References 7-1 XI ------- I I I I I I I I I I I I I I I I I I I FIGURES Number 1-1 Effects and responses to water quality regulatory actions 1-2 1-2 Key steps in a benefit-cost assessment .'.... 1-9 1-3 Key steps in a complex benefit-cost assessment . 1-10 3-1 A spectrum of water quality benefits 3-2 3-2 The demand function and consumer surplus 3-5 3-3 Illustration of the treatment of water quality with travel cost demand model 3-7 3-4 General travel cost demand model for a water quality improvement 3-13 3-5 Supply .function and the producer surplus 3-28 3-6 Irrigation benefits ' 3-30 4-1 Measurement of opportunity/costs 4-2 4-2 Sample model plant data form 4-13 5-1 Tradeoff curve. 5-7 5-2 Checklist for a water quality standards benefit-cost assessment 5-8 6-1 Travel cost demand function and consumer surplus with boatable water 6-10 6-2 Travel cost demand function and change in consumer surplus with fishable water 6-12 6-3 Travel cost demand function and change in consumer surplus with swimmable water . . 6-16 XIII ------- I I I I I I I I I I I I I I I I I I I Number 2-1 2-2 2-3 2-4 2-5 2-6 TABLES Benefits and Costs of Attaining Fish and Wildlife Use: Array 1 2-3 Benefits and Costs of Attaining Fish and Wildlife Use: Array 2 . 2-3 Benefits and Costs of Attaining Fish and Wildlife Use: Array 3 2-4 Summary of Industrial Impact Measures 2-7 Distribution of Benefits and Costs 2-9 Summary of Final Description Factors Influencing Shadow Price of Capital 2-13 3-1 Classification of Benefit Categories 3-3 3-2 Site Attributes Considered in Monongahela River Study 3-11 3-3 Estimated Equations for Site Demand Parameter Estimates 3-11 3-4 Specifications for the Dissolved Oxygen (DO) Levels Associated With Use Designations 3-12 3-5 Benefit Estimates from Generalized Travel Cost Model With the Monongahela Survey Respondents 3-14 3-6 Summary of Biases in Contingent Valuation Experiments 3-17 3-7 Willingness to Pay for Three Levels of Water Quality 3-19 3-8 Vaughan-Russell Model—Predicted Effects of BPT Regulations on Participation . 3-24 4-1 Components of Capital and .Operation and Maintenance Costs 4-5 4-2 Variables and Definitions for Measuring Costs 4-7 4-3 Example 1: Cost Summary-Market Bleached Kraft, Subcategory 4-18 xv ------- I I I I I I I I I I I I I I I I I I I TABLES (continued) Number 4-4 Example 2: Current and Upgraded Treatment Trains for CAPDET . 4-22 4-5 Example 2: Price and Cost Inputs to CAPDET 4-23 4-6 Example 2: POTW Upgrading Cost Summary 4-24 4-7 Example .3: Development of Capital Costs (Nitrification) 4-26 4-8 Example 3: Summary of Costs for Nitrification Upgrade 4-26 5-1 Key Elements of Benefits and Costs . .' 5-2 5-2 Sensitivity Analysis Calculations: Discount Rate at 4 percent . . 5-5 5-3 Sensitivity Analysis 5-5 6-1 Demand for Recreation for River 1--Water Quality at Level Suitable for Boating 6-9 6-2 Demand for Recreation for River 1--Water . Quality at Level Where Gamefish (Bass) Can Live in River 6-11 6-3 Capital Equipment and Cost for Model Plant to Meet Regulation • 6-17 6-4 Estimated Capital Costs 6-18 6-5 Cost of Treatment Plant, City B 6-18 6-6 Present Value of Benefits . 6-22 6-7 Present Value of Operating Costs 6-23 6-8 Total Project Costs ' 6-23 6-9 Benefits and Costs of Attaining Fish and Wildlife Propagation Use Designation 6-24 6-10 Distribution of Benefits and Costs 6-25 XVJ ------- I I I I I I I I I I I I I I I I I I I CHAPTER 1 BENEFIT-COST ASSESSMENT: A COMMON SENSE APPROACH TO DECISIONS 1.1 INTRODUCTION Should a State change the designation of impaired uses for a river? Will a sewer overflow project .provide benefits in excess of costs? Is advanced waste treatment necessary to attain a river's designated uses? Are treatment plants more desirable on some rivers.than on others? This handbook—a primer on benefit-cost assessment—shows how economic principles can help decisionmakers make these difficult choices. Its primary intent is to demonstrate the common sense inherent in benefit-cost assess- ments of alternative choices. In addition, this handbook shows how to add an economic dimension to scientific and technical analyses in considering the full range of impacts from a proposed water quality action. Since they focus on the alternatives available to society, economic prin- ciples are especially relevant to water quality program problems. Specifically, the economic viewpoint recognizes that the scarcity of society's resources forces choices among alternatives. However, choosing among alternatives creates tradeoffs—i.e., one thing must be given up to attain another. Thus, water quality decisions produce both desired and undesired effects for society. Benefit-cost assessment simply uses economic principles to help the decisionmaker make these choices. Water quality programs implement regulatory mandates or provide assist- ance either to those adversely affected by the regulations or for specific projects. Principles covered in this handbook could be applied to: Effluent guidelines issues that require limits on specific indus- trial discharges. Water quality standards issues where States designate uses for water bodies and develop criteria to achieve the uses. Advanced treatment issues where the Federal Government pro- vides financial assistance to construct municipal treatment plants that require advanced technologies. Combined-sewer overflow issues where Federal assistance is provided to deal with municipal runoffs that create pollution problems. 1-1 ------- I I I I I I I I I I I I I I I I I I I This chapter discusses the basic principles in a benefit-cost assessment. Specifically, Section 1.2 provides an overview of regulation, using linkages between regulatory actions, effects, and changes in behavior. To highlight the basic principles, Section 1.3 describes .benefits from an .economic perspec- tive, and Section 1.4 details a similar discussion for costs. The basic con- cepts of benefit-cost assessment are described in Section 1.5, along with a step-by-step view of an assessment. Section 1.6 gives some groundrules for performing an assessment and Section 1.7 summarizes the key points in the chapter. Finally, Section 1.8 provides a guide to the remainder of Volume I of the handbook. 1.2 REGULATION: AN OVERVIEW Understanding how benefit-cost assessments can be used is easier with some knowledge of how a regulation affects economic activities. The key to this understanding is the linkage (shown in Figure 1-1) between (1) a change in a regulation (an action), (2) its technical effects, and (3.) the behavioral responses to it. Water Quality Regulatory Action(s)" Change Designated Use(s) Modify Criteria to Provide for Designated Use(s) Technical Effects of Water Quality —- Rcgulatory Action(s) Changes in Effluents Changes in Water Quality Change in Ecological Habitat Effect* on Economic Agents Behavioral Effects I of Water Quality _j Regulatory Action(s) | Behavioral Responses of Economic Agents Figure 1-1. Effects and responses to water quality regulatory actions. 1-2 ------- I I I I I I I I I I I I I I I I I I I One example of an action represented by the first two blocks in Figure 1-1 is a change in the uses designated for a water body and the associated modifications of technical water quality criteria to accommodate these uses. The action changes effluent levels and the resulting water quality and ecolog- ical habitat—all of which affect households and businesses, the primary eco- nomic agents. A change in effluent levels simply means more or fewer pollut- ants will be discharged into the water body, thus altering overall water qual- ity. The changes in water quality alter the diversity of microorganisms, fish, or flora and fauna and can noticeably change the local ecological habitat. The magnitude of the technical effects depends on specific water body characteris- tics, the nature of the pollutant being controlled, and the extent of control. For example, river depth, flowrate, and riverbed geology will influence the technical effects of changing the designated uses of a river to include a warm- water fishery. Not all water quality programs are regulatory programs. For example, the combined sewer overflow (CSO) program aims at directly enhancing water quality by reducing the surge of pollutants following a severe storm. Even in these programs, a determination of the linkages between the project and its technical effects is essential. Equally important to benefit-cost assessment is how businesses and house- holds are affected by the action. For example, if improved water quality will support a warmwater fishery in a watercourse, more gamefish will likely inhabit the river, thus enabling fishermen 'to catch more fish--the technical effect on the household. However, to achieve the level of dissolved oxygen necessary to support gamefish, regulation might be required so that firms clean up their discharges into the river. From society's viewpoint, therefore, actions have ' both beneficial and detrimental effects. Maximizing the public good requires consideration of both types of effects. Determining how beneficial and detrimental effects balance out requires consideration of the final linkage—how primary economic agents change their behavior in response to technical effects. For example, if the technical ef- fects of a water regulation (e.g., an increase in gamefish populations) allow fishermen currently using a watercourse to use it more, new users may be attracted to the site. In economic terms, this situation is described as in- creased demand for a site's recreation services. The amount of the demand increase will be determined both by site attributes (features) and by the site users. Important site attributes include the proximity of substitute fishing streams, the number of access points, and the quality of local natural fea- tures, such as the surrounding countryside. The incomes of the fishermen, the price of fishing equipment, and how badly the users and potential users want to fish—their preferences for fishing—will also affect the ultimate behav- ioral response to increased gamefish populations. However, just as households (or fishermen, as in the example) respond to the technical effects of regulatory actions, firms also respond. Specif- ically, they may decide to close down operations, alter waste treatment proc- esses, or alter product mixes to meet the technical standard required by the decision. Clearly, each of these behavioral responses has different conse- 1-3 ------- I I I I I I I I I I I I I I I I I I I quences, but the magnitude of the regulation's technical effects is determined by the range of feasible responses and the market conditions for the goods produced by the affected businesses. Thus, firms facing more favorable market conditions find a wider range of choices open to them, and those facing strong pressures from competing firms are more limited. The actual regulatory process is considerably more complex and less cer- tain than indicated above. For example, businesses using water in their production process could be adversely affected by the regulation if higher dissolved oxygen levels corrode their water pipes, resulting in higher operat- ing costs. Similarly, individuals who are not users of the site may be af- fected if they view general increased ecological diversity as a beneficial ef- fect. In the end, therefore, practitioners will have to determine which ef- fects are relevant for inclusion in the benefit-cost assessment. More importantly, the linkages discussed in this section do not attach values to the positive and negative effects. Rather, they merely suggest a way of viewing the regulatory process to help determine what the effects are. In some cases, however, the decision process is eased if effects are converted to values. Unfortunately, the attachment of values to the effects is a trouble- some process for many potential users of benefit-cost assessment. 1.3 BENEFITS: AN ECONOMIC VIEW ' The economic approach to defining and measuring regulatory benefits is unfamiliar to many noneconomists. Quite simply, however, economics ap- proaches benefits from society's perspective, assigning values based on indi- viduals' willingness to pay for particular regulatory effects.* In essence, economics implicitly assumes individuals are best suited to value the effects of water quality programs. Once both the beneficial and detrimental effects of a proposed action have been identified, the practitioner may heed to weigh their relative importance before a final decision is made. Of course, the economic valuing process described above can help determine relative importance--e.g., area fishermen's willingness to pay for an action to increase gamefish populations vs. the costs incurred by a local industrial plant whose discharges the action will require be cleaned up—but it has limits. Indeed, no approach—economic or otherwise—is a substitute for the judgment that decisionmakers must exercise to make choices among alternatives representing various types and degrees of well- being to a variety of population subgroups (fishermen, plant owners, etc.). Benefit-cost assessment is a framework for identifying and organizing informa- tion to ease the decisionmaking process, not a decision rule. *lt should be recognized, however, that, added up over all persons, individual willingness to pay is influenced by the income, or wealth, available to each person. 1-4 ------- I I I I I I I I I I I I I I I I I I I One frequently asked question is, "Why do people have to 'pay' for the beneficial effects of water quality programs?" The answer is that they do not; alternatively, they could accept payment to forgo the effects. In effect, the two measures—willingness to pay and willingness to accept—are equally good, but different, measures. While the "accept" measure implicitly assumes the individual "owns" the rights to the beneficial change, the "pay" measure assumes the opposite. Volume II of the handbook will discuss how benefits based on these two different measures can be related. The important distinc- tion is the equity question--!.e., whether individuals own rights to the pro- gram benefits. Although willingness to accept is an equally good measure, willingness to pay is normally used to discuss regulatory benefits because it can be revealed in markets, when they exist, through purchases of good's/or services affected by the program. Then, benefits can be measured empirically. Although these markets clearly do not exist for the effects of water quality improvements, willingness to pay is still a useful way of valuing benefits. An assessment may describe benefits only in qualitative terms, but the description can be written from the perspective of willingness to pay. The benefits measurement approaches discussed in Chapter 3 are ways that economists have approached the benefits problem when markets do not exist. None of these approaches gives precise estimates of willingness to pay. Each is a blunt tool, capable only of giving rough estimates, which are sufficient in most cases. 1.4 COSTS: AN ECONOMIC VIEW Opportunity cost measures the cost of any resource--e.g., labor, ma- chinery, environmental resources—in terms of its next best alternative use. That is, the value of forgone alternative uses of any resource provides the basis for estimating the cost of any specific use. As a result, opportunity cost considers tradeoffs—i.e., how much must be given up of one thing to have more of another. For example, assume a proposed project would improve a lake's water quality to permit recreational fishing, boating, and swimming. The lake can- hot now support any of these activities, but it would if quality were improved by constructing a waste treatment plant along a river that feeds the lake. In this example, the opportunity costs of the action would be the forgone op- portunities of all the resources used in improving water quality. In the ab- sence of market imperfections, the opportunity cost of construction inputs— equipment, materials, labor, land, etc.--would be valued by their market prices. In addition, if the action precludes use of the river or the lake for other activities (such as industrial or agricultural uses), the values of these forgone alternatives would also be part of the opportunity costs. Many practitioners consider cost estimation an easier task than benefit estimation. Perhaps a more accurate view is that many find it less objection- able to value the labor, materials, and equipment used as a result of an action. However, difficulties can arise when the full social costs of the in- vestment alternatives are considered, or when effects on rates of technological change are included. Cost estimation is likely to involve as many judgments 1-5 ------- I I I I I I I I I I I I I I I I I I I as benefit estimation and is subject to the same general cautions. Indeed, caution is advised in making too great a case for precision in the measurement of economic well-being when less precision than desired exists in the linkage between the regulatory action and its effect on economic activities. The prob- lems in establishing linkages do not imply these technical issues should .be ignored, only that they be considered in their proper perspective—as a part of the problem of evaluating the benefits and costs associated with a change in water quality. 1,5 BENEFIT-COST ASSESSMENT: WHAT IS IT? Benefit-cost assessment is a way of organizing information—a method for identifying all the favorable and unfavorable outcomes of a proposed action. Where necessary for complex decisions, many of these outcomes can be con- verted into a common set of units (usually dollars) to permit consistent com- parisions of benefits and costs. Monetization may be impossible for outcomes which defy measurement. Even in these cases, the benefit-cost assessment framework can organize information associated with an action. In short, benefit-cost assessment is a practical method for including basic economic principles in the decision process. Although benefit-cost assessment is a guide to decisionmakers, it does not provide the final answer to a public policy decision. Other factors such as the public's view of appropriate uses for a particular water body or the fairness of cost impacts on particular groups are important considerations. What a .benefit-cost assessment does do is provide an organizing framework for information the public and rulemaking body can use in making more in- formed decisions. It is important to recognize that value judgments are a part of all deci- sions. Benefit-cost assessments supplement scientific and technical informa- tion with economic information that may help decisionmakers make these judg- ments. Very simply, a well-structured benefit-cost assessment can reduce the complexity of what needs to be considered, making the decision process more manageable. Is Benefit-Cost Assessment Different from Cost-Benefit Analysis? One of the first things that comes to mind for potential practitioners of benefit-cost assessment is the past misuse of cost-benefit analysis. These misuses emphasized the search for a ratio—the one "number"--that would ra- tionalize or justify a project. In many instances, the misuses involved an attempt to include benefits that were, at best, marginally related to a project. This is not the case for the benefit-cost assessment suggested in this hand- book. Since benefit-cost assessment requires a consistent, systematic treat- ment of benefits and costs, an outside observer can easily discover when a practitioner tries to stretch the approach beyond the limits dictated by com- mon sense. When the assessment process is carefully conducted, common sense will provide a reasonable guide through most of the decisions. Any- thing not sensible should be scrutinized. Critics who maintain that benefit- cost assessment can be used to justify anything overlook the fact that any approach can be abused. 1-6 ------- I I I I I I I I I I I I I I I I I I I Since it can compare benefits and costs in qualitative terms, in qualita- tive terms with some quantification, in quantitative terms, or in monetized terms, benefit-cost assessment is a more flexible approach than conventional cost-benefit analysis. The key, of course, is the nature of the decision. A qualitative assessment will reveal whether the potential benefits and costs at stake in a water quality decision are clearcut. No further economic assess- ment will be needed, yet the decisionmaker will have a logical, consistent basis for economic considerations. If the situation is more complicated, or if the potential benefits and costs at stake are considerably larger, a more de- tailed benefit-cost assessment can make the decision more manageable. Many ways exist to tailor a benefit-cost assessment to fit the needs of the Issue at hand. For example, monetization can play a critical role in more complicated benefit-cost assessment decisions. Specifically, by blending mone- tization with qualitative judgment, .benefit-cost - assessment can determine whether the benefits bear a reasonable relationship to the costs involved and whether there will be significant impacts on certain parts of the population or the economy. A misconception that arises with benefit-cost assessment involves recent techniques—such as survey techniques—developed to deal with previously immeasurable or nonmonetizable benefits, such as enhanced ecological diver- sity or amenities. These survey techniques are not opinion polls; they rely on carefully designed questionnaires to measure an individual's willingness to pay for these benefits. One fact has been clearly shown by all the survey studies: such previously nonquantifiable benefits are indeed a substantial component of the total environmental benefits picture. That is, the studies have shown these benefits to be large, and an assessment that overlooks them may indeed understate the full benefits. While most water quality decisions will not require a survey to determine these previously nonquantifiable bene- fits, some benefit-cost assessments may adapt the results of recent surveys for specific sites (e.g., see Chapter 6, Section 6.3). A few instances may occur in which, because the potential costs are so large, practitioners may want to use simple surveys to get at least a ball park estimate of such poten- tial benefits. Benefit-Cost Assessment: A Step-by-Step View Each of us makes decisions every day, judging whether the anticipated consequences of an action will be "worth" the "costs." Of course, the mean- ings of "worth" and "costs" vary from one person to the next because differ- ent people evaluate the same action differently. Whatever the outcome, how- ever, the logic underlying the decision process is the same. Based on this decision logic, benefit-cost assessment is a method for defining "worth" and "costs," offering a logical framework for structuring information for decisions in the public sector. Although performing a benefit-cost assessment is not a mechanical task with each step completely known in advance, it is possible to outline the gen- eral steps that are useful in assembling a complete assessment. These steps 1-7 ------- I I I I I I I I I I I I I I I I I I I flesh out the linkages between a policy decision and the behavioral changes, highlighted earlier in Figure 1-1. For instance, an assessment can portray the individuals and firms that will be directly affected by a regulatory action, how they will be affected, and how they will change their behavior in response to the regulation. The logic of a benefit-cost assessment is straightforward, as shown in the following steps: Define the action Determine an appropriate approach based on resources or com- plexity of the action Identify and estimate the incremental benefits of the action Identify and estimate the incremental costs of the action Compare the benefits and costs of the action Assess the plausibility of the results Highlight the distribution of benefits and costs and financial impacts of the action Integrate the assessment into other aspects of the decision- making process. . For example, Figure 1-2 illustrates these steps for a water quality stand- ards action, such as a State's changing the uses designated for a river segment. Steps in the upper portion of Figure 1-1 assemble and organize the available data for the change in designated uses. The optional analyses of the technical aspects of a water quality standards decision—analyses for use attainability, site-specific criteria, and wasteload allocation—can be valuable sources of data on the technical linkages between an action arid its effects. By sorting the data according to whether the action's effects result in a bene- fit or a cost to society, the State can compare, roughly, the benefits with the costs. The level of difficulty in the benefit-cost assessment process is dictated by the complexity of the effects and responses to the program actions. For example, when benefits and costs of an action are clearcut and have values that are comparatively small, a simple qualitative assessment is in order. In these cases, the assessment process merely describes the distribution of bene- fits and costs—i.e., who in society receives the benefits and who bears the costs—presents the results, and organizes them for the water quality deci- sion. However, if a qualitative assessment reveals that potential benefits and costs are substantial or not clearcut, a more detailed and comprehensive as- sessment is in order, as shown by the steps in Figure 1-3. In these cases, the practitioner must measure, value, and discount the benefits and costs and judge the sensitivity of the results. In most instances, staff resources and existing information can be combined for an assessment. In a few situations, 1-8 ------- I I I I I I I I I I I I I I I I I I I Identify the change in use designation for the river segment Obtain results of the use- attainability assessment for river segment Clearcut Case List benefits and costs and make a determination of the complexity of the benefits and costs for change in river segment Determine as much as feasible about quantified and nonquantifiable benefits 1 Determine as much as feasible about quantified and nonquantifiable costs Determine whether benefits bear a reasonable relationship to costs Complex Case Determine plausibility of results Highlight distribution of benefits and costs Present results of benefit cost assessment Figure 1-2. Key steps in a benefit-cost assessment. 1-9 ------- I I I I I I I I I I I I I I I I I I I Identify the change in use designation for the river segment Obtain results of the use- attainability assessment for river segment List benefits and costs and make a determination of the complexity of the benefits and costs for change in river segment Determine nonquantifiable, nonmonetized benefits Complex Case Value quantifiable benefits using techniques in Chapter 3 Determine nonquantifiable, nonmonetized costs Value quantifiable costs using techniques in Chapter 4 Translate benefits and cost values into common units using a discount rate and appropriate time horizon Conduct sensitivity analysis for key variables and check for less than fully employed resources Highlight distribution of benefits and costs Present results of benefit-cost assessment Figure 1-3. Key steps in a complex benefit-cost assessment. 1-10 ------- I I I I I I I I I I I I I I I I I I I outside assistance may be needed for a more detailed assessment; likely candidates include area universities and consulting firms. Regardless of out- side assistance, following the flow chart can ensure good quality results. Thus, the strength of benefit-cost assessment is its ability to organize material in a consistent manner and yet remain flexible enough to accommodate a wide range of cases. Nonetheless, the practitioner must recognize that each program action will introduce hew complexities requiring judgments that can be made .based only on an understanding of the strengths and weaknesses of the benefit-cost assessment process. 1.6 KEY CONSIDERATIONS IN AN ASSESSMENT . Four practical problems arise in implementing general benefit-cost prin- ciples: determining a baseline, determining the primary effects, avoiding double counting, and using expenditures to measure benefits. The benefits and costs of any water quality action reflect both regula- tions already in place and specific features of the affected water bodies. This means the baseline must be identified before the benefits and costs of a hew action can be determined. For example, technology-based requirements, and any variants of them, usually form the regulatory baseline for additional water quality decisions. In cases where the technology-based requirements have not been met, determining the baseline is difficult due to uncertainties in the predictions of the effects that the in-place regulations will produce. Effective enforcement is geherajly assumed for existing regulations. In an actual assessment, practitioners will have to make judgments about these base- line issues. If uncertainty exists in the determination of the.baseline, this should be clearly stated and addressed later, when the plausibility of the overall assessment is considered. The specific attributes of a site are also important in linking benefits with the effects of water quality decisions. For example, .swimming benefits will not likely be significant for a river that is only a few feet deep in places and has considerable current, no complementary facilities (such as beaches or access points), or large amounts of barge traffic. However, swimming may be important when adjacent parks and facilities are present and pollution is the limiting factor, as is the case for certain river pools in the Mississippi River in the Minneapolis area [Larson, 1981]. Similarly, the costs of achieving a particular level of water quality will depend on site-specific water quality as it existed before the regulatory action took effect. The analyses of use- attainability site-specific criteria, and wasteload allocation, any one of which may be performed as an optional part of water quality standards decisions, can be a valuable source of technical information. Another important distinction is between primary and secondary benefits and costs. Primary benefits and costs arise directly from the action, while secondary benefits and costs follow the impact of the primary ones. Only primary benefits and costs should be included in an assessment, because link- ages are often too imprecise to make even a rough determination of secondary benefits and costs. For example, .while increased recreation activities and 7-11 ------- I I I I I I I I I I I I I I I I I I I enhanced ecological diversity are among the primary benefits of a water quality improvement, the increased revenues to providers of recreation equipment, for example, are simply expenditures—secondary benefits-- not primary bene- fits. When these expenditures are included, the opportunities for double .counting increase substantially. If increased receipts of recreation equipment suppliers are added to willingness to pay, then that part of users' willingness to pay is double counted. In effect, including secondary benefits in an as- sessment opens up the assessment to the same suspicions that plague some applications of traditional cost-benefit analysis. The distinction, between primary and secondary benefits is important in the solution of another problem that arises in a benefit-cost assessment: the use of an expenditure approach to measure benefits. The expenditure ap- proach adds up an area's recreation-related expenditures. This reflects the costs of recreation, not an individual's willingness to pay for recreation. For example, the approach would include the costs of the fishing gear itself— amounts that are costs and not benefits; In addition, it does not count the difference between the maximum an individual would pay and the amount he actually pays—in technical terms the consumer surplus. In effect, the ex- penditure approach includes some costs on the benefits side of the ledger and excludes other benefits entirely. Total recreation expenditures may be useful in identifying some of the effects on a community's economic activity (e.g., increased sales tax receipts or recreation-related employment). Even in this limited use, however, the use of total expenditures omits important flows of funds out of the community to pay for goods externally produced. Confusion on these points often re- sults because it seems logical that expenditures should be benefits. However, expenditures are costs and benefits; they are not al| benefits. Both double counting and miscounting occur when this approach is used. 1.7 SUMMARY Benefit-cost assessment: Applies a formal dose of common sense to evaluating water quality regulations and programs. Provides a flexible approach for organizing the information needed to make water quality decisions. Enhances but does not supplant the value judgments of decisionmakers. Uses.society as the basis for accounting benefits and costs. Focuses on individual willingness to pay and opportunity cost to measure benefits and costs, respectively. Concentrates on primary benefits and costs. 1-12 ------- I I I I I I I I I I I I I I I I I I I 1.8 GUIDE TO HANDBOOK This handbook is organized into six chapters. Chapter 1 is the intro- duction to benefit-cost assessment. Chapter 2 shows how to include intangi- bles arid distribution effects in an assessment and considers the question of discounting benefits and costs. Chapter 3 describes specific methods for esti- mating the benefits of water quality programs, including techniques for devel- oping the monetary components of benefits needed in complex cases. Chap- ter 4 gives the basics of estimating costs, focusing on the incremental costs of water quality regulations. Chapter 5 describes a sensitivity analysis as a guide to a plausible assessment and highlights methods of presenting the re- sults of an assessment. Chapter 6 illustrates benefit-cost assessment prac- tices with simple, moderately difficult, and complex causes to reflect different types of water quality decisions. 1-13 ------- I I I I I I I I I I I I I I I I I I I CHAPTER 2 ISSUES IN A BENEFIT-COST ASSESSMENT 2.1 INTRODUCTION How can intangible benefits and costs be included in a benefit-cost assessment? What does it mean to discount benefits and costs? Are there rules of thumb for discounting? Does benefit-cost assessment ignore the distribution, of benefits and costs? In practice, the positive and negative effects of a program action occur at different points in time, affecting households and firms over a number of years. In many assessments the need or ability to monetize benefits and costs may be small, or limited resources may preclude monetizing. Clearly, a need exists for a logical approach to intangibles and for a convenient way to include them in an assessment. For those assessments where benefits and costs are monetized, two main "adding up" issues arise: discounting and distribution. Discounting pro- vides a consistent basis for adding benefits and costs over time. |t is one of the most complex and controversial issues in an assessment. Similarly, simply adding benefits and costs over people or firms may hide important issues. To explain how a benefit-cost assessment addresses these important Issues, Section 2.2 discusses intangible benefits and costs and uses arrays, or tabular, displays, to feature them in the assessment. Section 2.3 briefly describes how to measure impacts on firms and households, and Section 2.4 illustrates how the distribution of benefits and costs can be included in an assessment. Section 2.5 describes the discount rate, its role in a benefit- cost assessment and key issues in selecting a discount rate. Finally, Sec- tion 2.6 summarizes the chapter's major points. 2.2 INCLUDING INTANGIBLES IN A BENEFIT-COST ASSESSMENT Introduction The uses prescribed in the Clean Water Act for the water quality stand- ards are likely to provide intangible benefits relating to enhanced species diversity and ecological habitats and improved aesthetics. By its incommen- surability, this type of benefit presents problems for determining the net benefits of a use designated under the standards program--or for any water quality program. The types of benefits or costs that comprise the intangibles group change over time with improvements in valuation techniques. For example, the travel cost technique estimates willingness to pay for recreation benefits that initially were treated as intangibles. This section provides a 2-1 ------- I I I I I I I I I I I I I I I I I I I method for including intangibles in a benefit-cost assessment. Volume II will contain more details on intangible benefits. A recommended method uses a system of tabular displays, or arrays, to present both tangible and intangible benefits. These arrays are tailored to fit the nature of the assessment being conducted. The first array simply lists and describes the benefits and costs. A second array presents mone- tized values only for those benefits and costs for which monetization is almost always accepted, with the remaining benefits and costs being listed, de- scribed, and quantified to the extent possible. These values are based on individuals' willingness to pay and opportunity cost as measured by the tech- niques presented in Chapters 3 and 4. The final array presents as many benefits and costs in monetized terms as possible, with the other benefits and costs being listed, described, and quantified. Example Suppose a State is considering adding the fish and wildlife use to a stream that is currently designated for agricultural and industrial uses. A system of arrays for a benefit-cost assessment of this change in use designa- tion is illustrated in Tables 2-1, 2-2, and 2-3. The first array (see Table 2-1) lists and describes all the benefits and costs in qualitative terms. In cases where issues are clearcut, this array would, by itself, provide information sufficient for making the decision. As issues become more complex, additional arrays are essential for information sufficient to make the decision. The second array for this example (see Table 2-2) presents monetary values for benefits and costs for only those categories that almost all practi- tioners agree can be monetized. This supplemental information in the second array clarifies the issues in the assessment by focusing attention on the nonmonetized values. This array shows that the low end on the range of monetized recreation benefits is exceeded by the high end on the range of the costs and that the action produces nonmonetized benefits. The decisionmaker would have to determine how the nonmonetized aesthetic benefits and ecologi- cal diversity influence the net result given the overlap in the range estimated for benefits and costs. i The system of arrays shows how a benefit-cost assessment can reduce the dimensions of a complex issue to focus the decisionmaker's attention on the most difficult aspects. The last array (see Table 2-3) shows the mone- tization of as many benefits and costs as possible. The range of monetized benefits is estimated to be $17 million to $37 million, with additional nonmone- tized benefits attributable to the enhancement of the ecological diversity. The estimated costs of attaining the additional use designated range between $9 million and $14 million. Several features of this example call for additional discussion. The estimated monetary values for aesthetic benefits move the benefit range to a level at which benefits exceed both the minimum and maximum estimates of the 2-2 ------- I I I I I I I I I I I I I I I I I I I Table 2-1. Benefits and Costs of Attaining Fish and Wildlife Use: Array 1 Description of Benefits 1. Provide an additional resource that can support swimming, fishing, and recreation near water in a metropolitan area with only limited substitutes available. 2. Improve the aesthetic value for users of the resources services, such as recreators or property owners near the stream. 3. Improve the aesthetic vatue for residents of the area based on possible use in future or just from knowing the stream is cleaner. 4. Enhance the ecological diversity of the stream area by providing an improved habitat for fish species and wildlife from surrounding areas. However, none of these species is unique or endangered. Description of Costs 1. The residents of the city will require advanced treatment for their wastes. 2. Three industrial dischargers will have to modify their waste treatment operations. Table 2-2. Benefits and Costs of Attaining Fish and Wildlife Use: Array 2 Quantity Monetary value (million $, present values) Types of Benefits 1. Fishing, swimming, recreation near water 2. Improved aesthetics for users--recreators and prop- erty owners near stream 3. Improved aesthetics for nonusers--vatue to residents in area from knowing stream is clean should they use it or from just knowing it is clean 4. Enhanced ecological diversity 1 million visits 10 to 30 Types of Costs 1. Advanced treatment for municipal wastes 2. Advanced treatment for industrial dischargers 10 new fish species, smallmouth bass and others; 1,000 acres of improved wildlife habitat; no unique species are provided 1 new plant 3 additional treat- ment operations 8 to 10 1 to 4 2-3 ------- I I I I I I I I I I I I I I I I I I I Table 2-3. Benefits and Costs of Attaining Fish and Wildlife Use: Array 3 Quantity Monetary value (million $, present values) Types of Benefits 1. Fishing, swimming, recreation near water 2. Improved aesthetics for users--recreators and prop- erty owners near stream 3. Improved aesthetics for nonusers 4. Enhanced ecological diversity 1 million visits 10 to 30 5 Types of Costs 1. Advanced treatment for municipal wastes 2. Advanced treatment cost for industrial dischargers 10 new fish species, Not monetized although no unique species are provided; 1,000 acres of improved wildlife habitat. 1 new plant 8 to 10 3 additional treat- 1 to 4 ment operations costs. By monetizing these benefits with a contingent valuation survey (discussed in Chapter 3), the practitioner can show that the minimum benefits exceed even the highest cost. This makes a strong case for adding the use. In addition, the array in Table 2-3 shows nonmonetized benefits that would increase the total value of the benefits even more. This example illustrates the case for a river segment with large recreation potential that justifies the extra cost involved in carrying out the detailed assessment. River segments that have intermittent flows or whose entire flow is effluent would have low recreation potential and would not require such a detailed assessment. To show how the system of arrays can present assessment results, the discussion of this example concentrates on the efficiency aspects of the use designation. As noted in Section 2.4, however, the decision process .also should consider information on distribution effects. 2-4 ------- I I I I I I I I I I I I I I I I I I I 2.3 COST IMPACT MEASURES This section describes the general approach and some specific measures for assessing the cost impacts on communities and industries of meeting water quality regulations. This handbook does not advocate the use of any one measure; rather, it emphasizes that impacts be evaluated. Overall, the objective is to determine the incremental effect of compliance costs on earn- ings, production, and employment in the affected locality. However, the financial ability of a community or industry to absorb these costs is also important. For each major impact category, the following sections describe one or more measures with varying degrees of sophistication, data requirements, estimation methods, feasibility, and accuracy. By no means :are these meth- ods the.only way to proceed. Assessing Household Impacts The share of publicly owned treatment works (POTW) costs allocated to households in the form of higher sewer or water charges is assumed to be borne directly by those households, reducing annual income by the amount of the total annual costs. This implies that households cannot pass oh these costs by increasing their wages. Together with data on household income, total costs of compliance and community indebtedness are used to develop measures of the ability of households to bear these costs—i.e., how these costs affect income arid indebtedness. The discussion of impact measures is kept brief here because an EPA document, the.Financial Capability Gu|debook,* provides a detailed guide to community financial capability analysis. Household impact measures are divided into two major types: ability to pay and ability to finance. Ability-to-pay measures focus on the ability of the residents to bear the costs of water quality improvements, regardless of the current financial status of the community. Ability-to-finance measures focus on the ability of the community to finance the costs of compliance. The Financial Capability Guidebook develops 11 key indicators used to judge the ability to bear the impacts specified by these measures. Ability to Pay The measure of ability to pay is the ratio of compliance costs to median household income. Compliance costs are defined as total annual costs: the sum of annualized capital costs, operation and maintenance costs, and contri- butions to contingency funds. Focusing on values for the median household is convenient, but particular situations may call for a more detailed examina- tion of the distribution of income or wealth in the community. *Municipal Finance Officers Association and Peat, Marwick, Mitchell & Co., Financial Capability Guidebook (Draft), prepared for EPA, Office of Water Program Operations, Washington, D.C., May 1982. Contact the OWpO at EPA for further information about this document. 2-5 ------- I I I I I I I I I I I I I I I I I I I The Income measure may be used to estimate two kinds of cost impacts: the incremental costs of a particular water quality program and the total costs of compliance associated with all treatment. In some cases, total costs may be useful because the practitioner is interested in a view of cost impacts overall. For specific water programs, however, a focus on the incremental costs of decisions .that move water quality beyond that resulting from the statutory, technology-based controls can reveal the cost impacts due to the individual program. Financial Capability The financial capability of a community has a significant effect on its ability to raise additional funds in the bond market. If a community has high indebtedness or low tax revenues, it will have a. lower bond rating and face higher costs of capital. The impact measure is the total outstanding debt of the community before and after the project being considered. Assessing Industrial Impacts A firm's compliance cost may arise either from direct costs of treatment facilities or process changes or from its share of POTW costs. In turn, these costs lead to changes in profitability, output, and employment and may result in partial or complete closures. Of course, an accurate measure of impact requires some estimate of a business cash flow, but this estimate may be difficult to obtain unless the companies concerned cooperate. Thus, a large tradeoff exists among the measures .with respect to feasibility and accuracy. Table 2-4 summarizes the proposed impact measures, data sources, and their availability and reliability. The measures will be described in greater detail in Volume II. Assessing Changes in Employment, Output, and Prices Changes in output and employment in response to treatment costs are important because they give rise to indirect impacts. If workers are unem- ployed, they reduce their spending; if a plant reduces output, its demand for inputs from supplying firms slackens, with indirect repercussions on commun- ity income and employment. Even if estimating indirect impacts is infeasible, the direct effects of compliance costs on community employment and income are useful for assessing the equity implications of a regulatory action--"who is affected?1' As a rule of thumb for assessing the effect of regulatory actions, firms usually cannot pass through treatment costs by raising prices. This rule normally holds for water quality standard actions because they are site- specific and may affect only certain businesses. Although the rule may not hold for some regulatory actions, it is difficult to predict under what circum- stances firms might be able to pass through costs. 2-6 ------- 1 1 1 1 1 1 1 1 1 IB Table 2-4. Summary of Industrial Impact Measures Measure Profitability (including closure) 1. Cost/sales ratio 2. Cost/production cost ratio 3. Net cash flow 4. Rate of return 5. Net present value 6. Company solvency Reductions in employment and output 1 . Due to closure 2. Due to output reduction Price changes expected to be small Source Production and price estimates or public data bases EPA economic impact analysis Plant financial data Plant and company financial data Plant and company financial data Company financial data • Plant data Plant data, engi- neering report Availa- Relia- bility bility High Low / High Low Medium Medium Low High Low High Depends Depends on size on size Medium Medium Medium Medium 2.4 WHAT TO DO ABOUT DISTRIBUTION: PROBLEMS IN ADDING UP OVER PEOPLE I ntroduction A net benefits estimate does not evaluate projects based oh the distri- bution of net benefits. Rather, the evaluation is based on efficiency criteria, U.S. EPA Headquarters Library Mail code 3201 2-7 120° Pennsy/vanfa Avenue NW Washington DC 20460 ------- I I I I I I I I I I I I I I I I I I I which show how to allocate society's resources to maximize well-being.* Although the weights assigned to individual recipients of the benefits and costs of a project are treated equally, the distribution of net benefits can be described for the 'policy under evaluation. In these descriptions, benefits and costs are separated according to the economic agents affected. For example, one might classify households by income group, or firms by indus- try, and evaluate each group's share of the net benefits. Then, the overall benefit-cost assessment can account for distribution. Some attempts have been made to explicitly include equity (in terms of the effects of the project on different income groups) in benefit-cost assess- ment. They are not uniformly accepted. All of the weighting schemes are based on the premise that the extra utility or satisfaction derived from an extra dollar of income declines as income increases. Thus, redistribution of income will lead to increases in total utility or satisfaction for society. There are flaws inherent in any weighting scheme for benefit-cost as- sessment. The most difficult one to overcome is that it is hard to get society to agree on the appropriate weights. Equity in income is only one possibil- ity; regional and racial equity are others. In the final analysis, the weights are simply attempts to "guesstimate" the decisionmaker's preferences. This does not imply that distribution information should not be developed. Rather, it suggests that the tradeoff between efficiency (as measured by the aggre- gate net benefits) and various types of equity considerations (as reflected in the distributions of these net benefits among economic agents .under different classifications) is unlikely to be capable of being assigned a fixed relation- ship. Ultimately, the importance of distributional issues will depend on the decisionmaker's judgment, ' Two examples can be used to illustrate how distributional information has supplemented the conventional net benefit information in a benefit-cost as- sessment. Example I . Suppose an:improvement in water quality will provide. $10 million a year in net benefits. The distribution issue is to determine who will receive these benefits. The most commonly used method is to array the benefits by the shares that will accrue to different income groups, as shown below: Income ($) Less than 10,000 10,000 to 20,000 20,000 to 35,000 More than 35,000 Percent of net benefits 20 30 35 15 *Efficiency criteria indicate both the cheapest way of achieving a partic- ular level of water quality and what level of water quality makes sense given competing uses of resources. ------- r I I I I I I I I I I I I I I I I I I This change would favor lower income groups: half of the net benefits accrue to people with incomes of $20,000 or less, and over three-fourths accrue to people with incomes of less than $35,000 a year. By quantifying and monetizing benefits and costs, benefit-cost assessment provides a clear picture of the distribution effects of the change. The distribution of any intangible benefits and costs also should be considered in the decision. Example II Consider the same situation as in the previous example with $10 million in net benefits from the change in water quality. Table 2-5 arrays the distribution of benefits and of project costs. Rather than arraying by income groups, the categories in Table 2-5 break down the distribution of benefits and costs over broad groups in society to illustrate another way that distri- bution effects can be highlighted. Table 2-5. Distribution of Benefits and Costs Benefits: Who Receives? Users of river for recreation People who receive enhanced aesthetics values for river Downstream users for municipal water supplies Downstream companies who use water for industrial processes Costs: Who Bears? Residents who incur higher sewer and water bills because of advanced treatment requirements for wastes Stockholders of companies who have to install new equipment or change production processes to meet the standards Consumers who purchase products whose prices are increased as a re- sult of companies' compliance Summary: Distribution Information on distribution effects of water quality programs is an essen- tial ingredient in a benefit-cost assessment. It can be described with either summary measures like income group shares, or simply listed in narrative form. 2-9 ------- I I I I I I I I I I I I I I I I I I I 2.5 DISCOUNTING FUTURE BENEFITS AND COSTS: ADDING UP OVER TIME One of the most crucial issues in a benefit-cost assessment that relies on monetized benefits and costs is the selection of the appropriate discount rate—an interest rate used to translate dollar amounts of benefits and costs occurring in different years into a common unit of comparison, usually a pres- ent value. The discount rate is a positive number because individuals prefer immediate consumption and the associated immediate satisfaction to future con- sumption and the corresponding future satisfaction. Then, to persuade indi- viduals to give up immediate consumption in exchange for future consumption, the level of consumption must be increased at that later date. This increase is an opportunity cost and is .demonstrated, for example, when companies pay interest, or share future profits, to take advantage of current investment op- portunities. Preference for satisfaction now rather than later—technically known as the positive rate of time preference—is demonstrated when someone installs new carpet on a time payment plan or finances a new car. The satis- faction or utility from the carpet or car is gained now at the expense of a financing charge. There are five key concepts in determining discount rates: Social rate of time preference: the rate at which society is willing to exchange present consumption for future consump- tion. Consumption rate of interest: the rate at which individuals are willing to exchange present consumption for future con- sumption. Marginal rate of return on private investment: the incremental return on the last unit of investment by a private firm. Opportunity cost of public investment: the cost of a govern- ment investment measured in terms of forgone private consump- tion or investment. Risk: the degree to which investment in a public project will affect the variation in the outcome of all public investment. While a large share of the costs of meeting a water quality standard occurs in years immediately after a standard is .set (e.g., firms invest in new treatment processes, and cities construct advance waste treatment plants), benefits will not accrue until after the new plants and processes are in place. These benefits may accrue for 50 or 100 years—a period over which dollar amounts of both benefits and costs will vary greatly. The pattern of dis- counted net benefits often will look very different if a high discount rate is used rather than a low one. As this example shows, assessing benefits and costs of any water quality program requires an appreciation of the basic principles underlying the definition and the selection of an appropriate dis- count rate, an understanding of the empirical implications of discounting, and a practical knowledge of how to work with discounting techniques. 2-10 ------- I I I I I I I I I I I I I I I I I I I Time Preference: What Is It?* The use of discount rates in benefit-cost assessment can be explained by viewing discounting issues within the context of an ideal market economy, the characteristics of which include perfect competition in all markets, complete certainty in decisions., no transaction costs, no taxes, and no limitations on any credit market. In such an economy, all goods are priced at the oppor- tunity costs of the inputs used to produce them, and individuals and busi- nesses are able to borrow or lend, subject to their ability to repay, as much as they desire at the market rate of interest, which is determined by the demand and supply of loanable funds. . In hopes of obtaining future earnings, businesses in this ideal world would invest funds to. the point where their extra benefit equals their oppor^ tunity cost. In this economy, the market interest rate will be the opportun- ity cost of capital for the firm. Consequently, optimizing behavior by each business and efficiently working markets will ensure that the market interest rate will be equal to the marginal return on investment. The firm could not rearrange its investments and improve its long-term profit picture. At the same time, individuals would arrange their consumption and savings such that their own marginal rate of time preference (also known as the consumption rate of interest) would equal the market rate of interest. In this ideal case, the final outcome is that market forces create an equilibrium in which an individual's consumption rate of interest and a firm's marginal return on investment are the same because both correspond to the market rate of inter- est. This equilibrium ensures an efficient allocation of resources over time. If these equalities were not maintained (for example, if the consumption rate of interest were less than the marginal return on investment), an individual could improve his welfare (by consuming less, saving, and earning a return that permitted greater consumption in the future). Introducing public investments—such as those mandated by water quality programs—requires that the resources supporting them displace either private consumption or private investment. An efficient allocation of resources means that these investments earn a return at least equal to the marginal return on capital or the consumption rate of interest that would be required for these alternative uses (i.e., private investments or consumption). If it is also assumed that all individuals are alike with respect to factors determining their rates of time preference, society's overall rate of time preference should equal the market rate. The implication for selecting a discount rate for the water quality pro- grams in an ideal society is that either the social rate of time preference or the opportunity cost of capital would be appropriate because they are the same—i.e., both are equal to the market rate of interest. Unfortunately, when the assumptions of the ideal case are relaxed, the two rates diverge. These divergences explain why the selection of a discount rate for public sector investments has been such a difficult and often controversial issue. This section draws extensively oh Lind et al. [1982] 2-11 ------- I I I I I I I I I I I I I I I I I I I How To Determine the Djscount Rate in a Less Than Ideal World The selection of a discount rate for a benefit-cost assessment of a water quality program must be accomplished in a world considerably different from that of the ideal economy. Some of the most important divergences from the ideal economy can be attributed to the following factors: The tax on corporate income, which drives a wedge between the private rate of return and the rate of time preference. A higher rate of return on private investment is required to off- set the effects of income taxes, which cause the divergence between social and private rates. The dependence of future generations on decisions made by present generations. This dependence gives rise to a "public good"--the welfare of the future generations—that may not be included in the decisions of the private market. Private markets, which may be out of long-run equilibrium with an immediate discount rate different from the appropriate long-term rate. The determination of the appropriate private market rate, which is difficult because there are numerous capital markets, each with its own interest rate. » Public investment dollars, which do not necessarily displace private investment dollars but may instead use tax revenues that displace current consumption in the private markets. Attempts to reconcile these divergences have created a complex literature on the criteria for selecting an appropriate discount rate for public invest- ments or regulatory evaluations under different circumstances. Lind [1982] has distilled this literature, concluding that the rationale for discount rate selection should be based on the full opportunity costs of capital. Lind's approach does not ignore the potential divergence between socie- ty's social rate of time preference and the market rate of interest. Rather, he suggests that the social rate of discount be set equal to the social rate of time preference and that the shadow price of capital be used to adjust for the full opportunity costs of capital. The shadow price of capital is defined as the present value of the future stream of consumption benefits associated with $1 of private investment discounted at the social rate of time preference. A benefit-cost assessment to evaluate government investment decisions based on this approach considers the implications that investments have for consumption over time. The basic question to be answered is, "What does public investment displace?" To the extent public investment displaces pri- vate investment, that portion of the costs of the public project should be valued at the shadow price of capital. That is, the costs of this portion of the investment are valued in terms of the consumption forgone. When the 2-12 ------- I I I I I I I I I I I I I I I I I I I forgone consumption due to displaced investment is added to the balance of the costs (i.e., those displacing immediate consumption), all costs have been converted to their equivalent losses in private consumption. Similarly, bene- fits that lead to increased private investment should also be adjusted to reflect their potential to yield future consumption streams. The adjustment is to multiply that share of the private investment by the shadow price of capi- tal. This procedure adjusts benefits and costs at each point in time and expresses them in terms of the equivalent amount of .consumption that could be obtained. Table 2-6 .shows estimates of various parameters so that a practitioner could use this general procedure in a benefit-cost assessment. The estimated value for the marginal return on private investment in Table 2-6 is 10 per- cent. This approximation is based on judgment and the empirical relationship that the estimated average return, adjusted for inflation, is between 10 and 15 percent. More accurate measurements of the costs of capital are hindered by the Inconsistency between accounting data and economic concepts, and by implicit adjustments both for risk in the returns to capital data and for firms' inconsistencies in following established procedures to make capital budgeting decisions. Based on depreciable assets data for 1973, 1974, and 1975, the length of the typical private investment is estimated at 15 years, with a range of 10 to 20 years. Based on empirical work on consumption and savings, the marginal propensity to save is assumed to be 0.2. Table 2-6. Summary of Final Description Factors Influencing Shadow Price of Capital Shadow price of capital 1.62 to 2.57a 1.9b Marginal propensity to save 0.2 0.2 Marginal return .on private capital (%) 10 10 to 15 Social rate of time preference (%) 2 2 to 6 Length of typical private investment (years) 10 to 20 15 This row is based on Lind's Table 4 comparing the shadow price of capital under a range of assumptions for the social rate of time preference, marginal return on private investment, and length of typical private investment. This row is based oh Lind's discussion on pp. 101-102 of the unpublished manuscript. The returns on private capital are pre-tax returns. The shadow price of 1.9 is the central value associated with the variations in each of the parameters involved. The range of values was 1.65 to 2.15. The range of social rates of time preference are reported to indicate that they would be consistent with this shadow price. 2-13 ------- I I I I I I I I I I I I I I I I I I I In the procedure, the social rate of time preference is set equal to the consumption rate of interest for individuals. Unfortunately, there is no unique rate of interest in the real economy. But it is possible to gain some insight into individuals' rates of time preference from their savings and investment decisions. For example, the real after-tax average rate of return on treasury bills (a safe investment available to many people) over the period 1926 to 1978 was -0.5 percent for an individual in the 20-percent tax bracket, and the return on a mutual fund containing "average market" stocks totaled 4,6 percent. For individuals, the real rate of return must lie somewhere between the riskless treasury bill rate and the stock market equity returns. , Adjusting for the effects of unanticipated inflation shows that the mar- ginal rate of time preference must be in the range of -2 and 5 percent, with the average close to 0. Regardless of the actual point estimate selected, it will be considerably different from the 10 percent real rate of discount recom- mended by the Office of Management and Budget (OMB) for public, investment projects.* The shadow price of capital can be calculated from the estimates of the other parameters following Lind's procedures. Specifically, Lind estimates that the most likely value is within the range between 1.65 and 2.15. This shadow price is then used as described above to convert the benefits and costs into consumption equivalents. Including Risk in an Assessment of Discount Rates So far the implications of uncertainty for determining the discount rate have been implicitly ignored. Since the levels of benefits and costs are uncertain, there should be adjustments for the probabilities that a particular level of each will occur. Depending on its source and nature, the uncer- tainty can either be addressed in this choice of a discount rate or be directly reflected in the measurement of benefits and costs. Alternatively, some com- bination can be attempted. Each of these alternatives will be considered in Volume II. The main conclusion that can be drawn from the risk studies covered in Volume II is that the characteristics of the public investment project, together with the relation between the variability in the public investment and the variability in national income, are crucial factors in determining whether a riskless or risk-adjusted discount rate should be used in an assessment of public investments. The important relationship that must be explored is whether water quality program investments increase the variability in national income. For most of these applications, the effects of risk will be small, and a riskless rate can be used. *lt is difficult to interpret this rate as an estimate of the social rate of time preference. It may well be the equivalent to a recommendation that all projects be discounted at the marginal rate of return on private investment. Additional confusion is added by the Water Resources Council [1979] guidelines which tie the discount rate to an index. The current rate in this procedure is 7-7/8 percent. 2-14 ------- I I I I I I I I I I I I I I I I I I I What Are the Empirical Implications of the Discount Rate .Issues?* One of the reasons for the controversy concerning the use of discount rates is that the empirical implications of the discount rate can have a sub- stantial influence on the outcome of the benefit-cost assessment. Fox and Herfindahl [1964] reevaluated Federal water projects, previously evaluated at a 2-5/8 percent discount rate, at new discount rates of 4, 6, and 8 percent. Nine percent of the projects that previously had benefits exceeding costs at 2-5/8 percent experienced the opposite result with a 4-percent discount rate; 64 percent experienced the opposite result at a 6-percent rate; and 80 per- cent experienced the opposite result at an 8-percent rate. The implications are quite clear: Most of the projects .had costs exceeding benefits at the higher discount rate, and all had positive benefits at the low 2r5/8 percent rate. Many water program regulations or projects will have benefits that will accrue 10, 20, 30, or even 100 years in the future and costs that could be substantial during initial periods. For example, suppose a water quality project requires a $1 million investment in 1982 and will provide $50 million in recreation .benefits at the end of 50 years. The net present values at differ- ent discount rates would be as follows: Discount rate 5 percent 8 percent 12 percent Net present value 3,336^186 6,606 -826,990 As shown, the discount rate is crucial in determining the ultimate assessment of benefits and costs. This emphasizes the importance of the sensitivity analysis recommended in Chapter 5, which shows that in some cases there will be positive net benefits regardless of the discount rate, while in others the outcome of the assessment is very sensitive to the discount rate applied. Another important empirical distinction in discounting is the difference between real and nominal rates of discount. The difference is the expected rate of inflation. Most benefit-cost analyses are conducted using constant dollar values for the benefits and costs. In these cases 'the real rate of discount should be used. Benefit-cost assessments have employed real discount rates ranging from 0 to 4 percent, while the nominal rates have ranged from 8 to 16 percent. The higher end of the scale for nominal rates represents the influence of recent high levels of inflation and market interest rates. The .difference between the real and nominal rates is quite substantial and indicates why it is important not to mix the two in a benefit-cost assessment. For example, com- pare the implications of a real discount rate of 2 percent and a nominal rate of 10 percent (inflation is expected to be approximately 8 percent). With the *This discussion is adapted from Just, Hueth, and Schmitz [1982]. 3-15 ------- I I I I I I I I I I I I I I I I I I I nominal rate used as the discount rate, society would be indifferent between $1 now and $13,780 in 100 years. However, if a real rate is employed, the difference would be $1 now versus $7.24 in 100 years. Thus, although it seems low, given the high interest the economy is currently experiencing, a real discount rate of 2 to 4 percent may make considerable sense in a benefit- cost assessment where the long-term perspective is essential. The Simple Mechanics of Discounting This section offers a brief review of the mechanics of discounting. The reader is urged to consult a finance text for present value tables and more detailed discussions on discounting. The discounting problem in a benefit-cost assessment is how to translate benefits and costs occurring .in different periods into a common basis for comparison. The most frequently used basis is present value, which is defined as the amount of money at the present time that some future amount is worth. Discounting is the process of computing the present value of a future stream of dollars.* Consider a simple example that might arise in the assessment of a water quality standards program. Suppose a State is considering changing the presently unattained use designation for a river segment from "fish and wild- life propagation" to "limited warmwater fishery." In this case, there will be a loss in potential benefits, as well as cost savings from the forgone pollution control investment for cities or industries. The monetary values for the benefits and costs associated with this decision are as follows: Year Benefits forgone Cost savings 1982 40,000 $100,000 1983 40,000 10,000 1984 40,000 10,000 The discounting problem is: 1. Select the appropriate discount rate. 2. Translate future benefits and costs into present values for comparison. The discounting formula for this procedure is: 1 P.V. = (HO* *For simplicity, assume all dollars accrue at the end of each year so there is no need to account for differences within a year. 2-16 ------- I I I I I I I I I I I I I I I I I I I where P.V. = the present value factor for either benefits or costs i = the discount rate t = the time period. The discount factor P.V. is multiplied by the benefits or costs foK each period in the planning horizon; then the results are summed. Suppose that i = 4 percent with the monetary values for benefits and costs from above. The discounting calculations are the following: (1982) (t=0) an nnn ± (1983) (t=1) (1984) (t=2) Benefits forgone = 40,000 + Cost savings = 100,000 + Benefits forgone = 40,000 + 38,462 + 36,982 Cost savings— 100,000 + 9,615 + 9,246 Net cost savings = $3,317 . $115,444 $118,861 The change of the use designation will yield a small positive cost savings. In this example, the forgone benefits end in 1984. In most cases, they would continue into the future for whatever time horizon is selected for the assess- ment. i If the benefit (or cost) stream is constant each year at A for the full life of the project (assumed to be n years), the formula for the present value can be rewritten as follows, often referred to as the present value of an annuity: 1 P = A 1." (1+0 n where P = present value A = annual amount i = discount rate. Present value may also be determined by using the tables in a finance or accounting text. Another .variation on the discounting mechanics is to trans- 2-17 ------- I I I I I I I I I I I I I I I I I I I late amounts into annual values that can be compared. The formula for this calculation, often referred to as the uniform-series capital recovery factor, is: i(l-H) A = Pj Discounting: A Summary Review While the arguments summarized above require the use of judgment for each new situation, general guidelines do exist. For a wide range of water quality programs, the social rate of time preference would range from 2 per- cent to a maximum of 6 percent. The recommended procedure is to consider the implications of each of these values for the discounted net benefits of the decision.* If the present value of net benefits is positive and the project decision remains unchanged, it is unnecessary to further consider the techni- cal issues affecting the selection of one value in this range. However, for those cases where the value of net benefits (i.e., positive versus negative) is affected by the discount rate, a more refined selection is recommended that considers the practical implications of Lind's analysis. This process requires answering four key questions: 1. What are the sources of the public investment resources? Are they tax revenues that can, in principle,, displace private consumption or investment?! In addition, what are the likely portions coming from each source? Answers to these questions will affect the importance of estimating the shadow price of capital. 2. How large are the private investments required by the water quality action? If they are large, adjustment by the shadow price is likely to be warranted to reflect the full opportunity costs of these investments. 3. What is the nature of the risk associated with the investment and its relationship to overall economic activities? The answer to this question provides a basis for judging whether an adjument for risk should be made in evaluating the project. If most water quality investments increase the variability in eco- nomic activity for a State (or the county as a whole), then the selection will tend to be at the higher end of the range for the social rate of time preference. *ln some cases, legal restrictions mandate the use of a specific rate; e.g., advanced treatment applications require the applicants to use the Water Resources Council's rate of 7-7/8 percent. tUser fees should be regarded as payments for services provided and therefore do not displace private investment. 2-18 ------- I I I I I I I I I I I I I I I I I I I 4. Are the sources of finance for the project known in advance (i.e., Federal sharing of a local project's cost)? Since one objective of benefit-cost assessment is to improve the overall allocation of resources, any Federal share of the costs should be treated in the same way as the local share, with consid- eration given to the full opportunity costs of the funds used. It is impossible to recommend a single rate of discount as relevant for all situations. Each decision may well have special attributes that will need to be reflected in the selection. It is important not to let the technical considera- tions involved in defining the appropriate discount rate become overwhelming. For most purposes, the 2 to 6 percent range of values will be all that is necessary. 2.6 Summary Intangibles .should be viewed from society's willingness-to-pay benchmark even though they are incommensurable. Arrays or tabular displays are useful exposition tools for intangibles. Household impacts can be measured according to ability to pay or ability to finance. One measure of ability to pay is the ratio of compliance costs to median household income. Ability to finance is reflected in a city's bond market rating in the financial community. Industrial impacts can affect profits, output, and employment. Distribution can be highlighted in arrays showing who receives and who bears. Discount rates should be selected carefully. Shadow price of capital should be considered in gauging the full opportunity cost of public investment. A sensitivity analysis should be performed for the effects of selecting the. discount rate. Several discount rates should be tried. The implications of risk should be considered for the discount rate. The relationship between the variability of investment or outcomes of regulatory policy relative to variability in national income should be considered. 2-19 ------- I Net benefits should be determined on a present value basis. Formulae and tables should be used. • * Real and nominal discount rates should not be mixed. I I I I I I I I I I I I I I I I 2-20 ------- I I I I I I I I I I I I I I I I I I I CHAPTER 3 MEASURING THE BENEFITS OF WATER QUALITY PROGRAMS 3.1 INTRODUCTION What are the benefits of water quality programs? What are the basic benefit principles? What methods are available to measure benefits? What are the key assumptions and features of these methods? What case studies are available to illustrate the use of these1 methods? While benefit-cost assessments that use only qualitative information may suffice for clearcut water quality decisions, quantitative information can simpli- fy more difficult decisions, by reducing the complexity of the issues and clari- fying the central issues. This chapter provides a framework for gathering and organizing this information to measure the benefits of water quality pro- grams. This chapter briefly reviews the concept of benefits and practical ap- proaches for measuring them. In particular, it discusses the conceptual issues involved in estimating the benefits of water quality improvements, de- scribes techniques for measuring different types of benefits, and presents case studies that show how different practitioners have employed them. Spe- cifically, Section 3.2 summarizes benefit categories, Section 3.3 highlights general issues in selecting a benefits estimation methodology, and Section 3.4 describes approaches for measuring household benefits. Section 3.5 discusses business benefits by summarizing key aspects in studies of agricultural, indus- trial, and navigational benefits of water quality improvement. Section 3.6 briefly describes public water supply benefits, and Section 3.7 summarizes the issues covered in the chapter. Case studies follow the text in each of the appropriate sections. (The scenarios in Chapter 6 show how the case studies can be applied in new situations.) 3.2 CATEGORIES OF BENEFITS: AN OVERVIEW Since each household or firm undertakes different types of activities, each is differently affected by water quality changes. A natural starting point in appraising the various types of benefits of water quality programs is to place them in broad classes. Figure 3-1 shows the categories of benefits associated with water quality programs. The top part of the figure aligns each benefit type with the uses made of water bodies. More perspective on these benefit types is given by Table 3-1, which lists them according to households or firms (the type of economic agent likely to receive them) and the methods appropriate for measuring them. For example, the user benefits category separates health benefits from recreation. Contingent valuation and hedonic property value models are potential candidates for measuring both 3-1 ------- I I I I I I I I I I I I I I I I I I I Potential Water Quality Benefits Current User Benefits Intrinsic Benefits Direct Use Indirect Use Potential Use No Use 1 — Recreational '- fishing, swimming, boating, rafting, etc. In Stream _— 1 1 — Commercial — fishing, navigation 1 — Municipal — drinking water, waste disposal Withdrawal — Agricultural - irrigation I— Industrial/Commercial — cooling, process treatment, waste disposal, steam generation i — Recreational'— hiking, picnicking, birdwatching, photography, etc. Near Stream _ Relaxation - viewing — Aesthetic*- enhancement of adjoining site amenities i — Near-term potential use Option' 1 1 — Long-term potential use • — Stewardship - maintaining a good environment for everyone to enjoy (including future Existence'^ family use-bequest) ' — Vicarious consumption — enjoyment from the knowledge that others are using the resource. Figure 3-1. A spectrum of water .quality benefits. categories of benefits. In addition, damage functions may be useful for health benefits, while the travel cost and recreation participation survey ap- proaches can measure recreation benefits. Although they are not directly associated with use, option and existence values are potential sources of economic benefits. Option value is the amount that an individual may be willing to pay (over his expected user values) for the right to use a water body--e.g., a river—in the future because uncer- tainty exists either in the river's availability or in the individual's use of it. Specifically, if an individual thinks he may want to use the river, but isn't sure, then he may pay some amount each year for the right to use it. When this payment exceeds the benefit the individual would receive from use, the excess is the option value. Existence value is the willingness to pay simply for the knowledge that a resource exists--!.e., the value an individual places on a resource just because he knows it is there. Thus, because of a steward- ship or related motive, an individual might be willing to pay something to main- tain a river even though he knows he will not use it. 3-2 ------- I I I I I I I I I I I I I I I I I I I Table 3-1. Classification of Benefit Categories Type of economic agent Benefit category Type of measure- ment method Household: User Recreation Household: nonuser Business and munici- pality: user Health Option value Existence value Cost savings Travel cost model Contingent valuation survey Recreation participation survey Hedonic property value model Damage function Contingent valuation survey Hedonic property value model Contingent valuation survey Contingent valuation survey Cost function For businesses and municipalities, classification is more clearcut: The effects on a firm's cost of production are the primary interest because the role of water quality is exclusively reflected through these effects. Irriga- tion, nagivatio'n, and .process uses are examples of water uses where cost savings may arise. Benefit classification should not be misinterpreted. In most'cases it is impossible to separate all the sources of a benefit estimate. For example, while willingness-to-pay estimates for a water quality improvement derived from a hedonic property value model (discussed below) may be based on both direct and indirect uses of the water body, the contributions of each cannot be shown in practice. Equally important, overlaps should be expected be- tween the methods used and the types of benefits derived. A contingent valuation survey's (discussed below) estimate of willingness to pay for im- proved water quality may include health, recreational, and nonuser benefits. The exact composition will depend on the water body under study and the nature of the questions used to elicit the information. The classification scheme offered in this section simply shows possible sources of benefits and may help identify measurement methods. However, because this scheme does not fully define each of the methods or describe how a water quality change is- introduced in each, the following sections offer more detailed discussions of the various benefit estimation approaches, in- cluding specific case studies that summarize previous work. 3-3 ------- I I I I I I I I I I I I I I I I I I I 3.3 PRACTICAL CONCERNS: SELECTING A BENEFIT ESTIMATION METHOD One of the first questions about benefit-cost assessment facing the prac- titioner is how to select a method or set of methods for measuring the bene- fits of water quality programs. This decision will be influenced by the fol- lowing conditions: The time and financial resources available. The types of economic agents affected by the change in water quality regulation. The nature and magnitude of the changes in the water quality regulations themselves. Data availability. These conditions largely determine the appropriate level of detail for a study. In most benefit-cost assessments of water quality programs,.existing data and results available from other studies will be sufficient. In complex cases, new data and case-specific methods may be necessary. However, even in these cases, when neither time nor resources are available, existing literature and, to a lesser extent, ad hoc methods must be used.* In practice, the ideal conditions routinely assumed in theoretical analyses of benefit measures simply do not exist. Indeed, many benefit analyses of environmental resources result in compromises that arise from a poor under- standing of the exact association between water quality and particular activ- ities of economic agents. However, while compromises may be necessary to measure the benefits of water quality regulations for some water bodies, every effort should be made to measure benefits based on willingness to pay and cost savings—the only definitions for economic benefits that have clear theoretical justifications. Finally, because resources are limited, benefit-cost assessments of water quality programs require that resources be wisely used—i.e., that they be closely matched with the complexity of specific cases. Whether or not re- sources are available, however, the practitioner must clearly understand the features of each benefit measurement method to make an intelligent choice among them. *For example, while the Water Resources Council's current guidelines for cost-benefit analysis, recommend use of the travel cost or contingent valua- tion approach for estimating economic benefits of outdoor recreation services, they acknowledge that practitioners may have to use ad hoc approximations such as activity-day values—constant dollar values proposed for days of par- ticular types of outdoor recreation (see Water Resources Council [1979])— multiplied by projections of user-days. 3-4 ------- I I I I I I I I I I I I I I I I I I I 3.4 HOUSEHOLD BENEFITS There are several ways to measure the benefits of water quality pro- grams. However, to compare the accuracy of* these measures, the practitioner must also understand how individuals value different goods. Fortunately, economics provides an objective way to measure these values using the basic concept of a demand function. Theory: The Demand Function The first organizational guidepost that economics provides is the concept of an individual demand function, shown in Figure 3-2. This function de- scribes for any good, X, the maximum quantity of the good an individual would be willing to purchase for each price of X. The downward slope of the curve indicates that individuals are willing to buy more of X at lower prices than at higher prices. The simple diagram in Figure 3-2 assumes all other factors that might influence demand—including income, the prices of related goods, etc.--do not change. Frequently, there is no need to actually meas- ure a .demand curve. What it does is provide a basis from which the benefits to households can be viewed. time Figure 3-2. The demand function and consumer surplus. If the market process establishes a price at P , the individual will pur- chase Q of X and make a total expenditure equal l8 P AQ O. Since the area under tne demand curve measures the individual's maxfrnun? willingness to pay for each unit of consumption, the total willingness to pay for Q is the entire area—total expenditures plus the triangle P P.A. This difference—between o j 3-5 ------- I I I I I I I I I I I I I I I I I I I what the individual actually pays and the amount he is willing to pay--is the consumer surplus, or the dollar measure of the satisfaction an individual re- ceives from consuming a good or service, less what he pays for it. As a dol- lar measure of individual welfare, consumer surplus is not ideal, but most studies have found it to be a good benchmark. Practice: Methods for Measuring Household Benefits Household benefits may be measured by using the travel cost method, a contingent valuation survey, a recreation participation survey, the hedonic property value method, or the damage function method. Advantages and dis- advantages, data requirements, and key assumptions of each method are high- lighted in this section. Case studies show how the methods have been used in recent applications. Travel Cost Method One of the most popular approaches to describe demand for the services of recreation facilities, the travel cost model,* has been used to estimate re- creational benefits in a wide variety of applications. The -logic underlying this model is simple. Recreators at a particular site pay an "implicit" price for using the site's services through the travel and time costs associated with visiting that site.f Since recreators visit a site from diverse origins, their "travel behavior" can be used to analyze the demand for the site's services. That is, all else being equal, any person will continue to travel to the site until the marginal value of the last trip is exactly equal to its full costs (i.e., the travel expenses and the opportunity cost of the time spent travel- ing). As a rule, the travel cost model estimates the demand for the representa- tive individual. Therefore, to estimate the aggregate benefits of water qual- ity improvement at a site, the practitioner needs to estimate how many recrea- tors would use it. The solution to this problem will depend upon the data used to estimate individual demand. For example, the visits made by resi- dents of an origin zone, usually during a season, relative to the population of that origin zone is .the quantity measure—a rate of use--conventibnally used in the travel cost model. Since benefits are for the "representative" rate of use, multiplying by the population of that origin zone will yield its aggregate benefit estimates. Overall benefits would be the sum of the zone benefit esti- mates . *For further details see Dwyer, Kelly, and Bowes [1977], Freeman [1979], Smith [1975], and Desvousges, Smith, and McGivney [1983]. fMost public recreation facilities either have no user fees or have nomi- nal fees that do not reflect the marginal cost of a site's recreation services. Thus, these fees are hot indicative of the equilibrium prices that would arise if conventional market mechanisms allocated the services of recreation sites. 3-6 ------- I I I I I I I I I I I I I I I I I I I Travel cost models also can be estimated by surveying users at a specific recreation site. Then, the benefit estimates are for a "representative" recre- ationist and estimates of total site use estimates, must be obtained independent- ly. Possible sources of these estimates include the Corps of Engineers, the Department of Interior, or parks and recreation departments at the State level. If a travel cost model can be estimated, it.can be used to estimate con- sumer surplus. However, this is only part of the problem that must be solved to appraise the benefits of water quality improvements. In addition, the linkages between the regulations, the changes in water quality, and the recreation decisions of individuals must be known. One possible linkage is a change in demand for the services of a recreation site because individuals wish to use them in one or more activities. A change in the level of water quality may permit a wider range of uses, increase the individual's enjoyment (and hence valuation) of existing uses, or both, which increases the demand for the site's services at each (implicit) price. Therefore, to evaluate the implications of a change in water quality for an individual's economic well-being, water quality must be linked to the vari- ables in a recreation demand function. Three ways for making this association will be discussed in detail in Volume II. The case shown in Figure .3-3 offers the most acceptable approach for linking water quality to recreation site de- mand. It incorporates the effects of water quality as a determinant of the demand in the travel cost model for a site's services (see Freeman [1979], Chapter 8). Since little evidence generally exists on the variation in water Implicit Price E $/visit WQ«>WQ Demand (WQ«) Demand (WO) 0 Q1 Q2 Quantity of Water Quality Treated as a Demand Change v'S't*/Y«»r Figure 3-3. Illustration of the treatment of water quality with travel cost demand model. 3-7 ------- I I I I I I I I I I I I I I I I I I I quality across the same water body, the travel cost model is difficult to imple- ment empirically with a single site. Conceptually, however, a change in water quality is assumed to shift site demand by providing a wider ranger of activities. Thus, in Figure 3-3, the benefits associated with a change from WQ to WQ* would be represented by the area HGFE. The scenario in Chap- ter 6 shows an existing study that can be applied to new situations (see Chapter 6, Section 6.4). Data Needs, Key Assumptions/Limitations, and Features The following checklist outlines the data needs and key assumptions and features of the travel cost model. Data Needs: Origin—county of residence or zip code—for users of the recreation site. These are often available from recreation management agencies for samples of users. Population size and summary measures for features of the popu- lation in each origin zone (e.g., median family income, median age, and median education). Sources include census data, national and State recreation surveys, and site surveys. Round-trip mileage from each origin to site. This information can be calculated from maps. Vehicle costs per mile and implicit time costs of travel. Travel costs should be calculated as operating costs per mile for the vehicle. Time costs can be estimated with the approximate wage rate for the household head. One source is a wage and occupation survey. Key Assumptions/Limitations and Features: The model is site specific. It measures the demand for the services of a site, not total or general recreation demand. The model measures only user benefits. Consistency in the length of stay for each type of trip in ag- gregate data. For example, all trips are treated as day visits or as weekend visits. A site's demand depends on its potential services for the re- quired activities. (For example, a minimum-sized river seg- .ment is necessary for power boating, while a river segment with extensive locks and dams is not conducive to canoeing.) The cost of time spent at the site is excluded. This suggests that "full-cost" may not be expressed in a demand relation- ship. 3-8 ------- I I I I I I I I I I I I I I I I I I I There are no good substitute sites available. If many substi- tutes are available, the simple model will overstate the demand for the site. The travel cost is assumed to capture all the factors that in- fluence the decision to recreate at the site. (For example, this assumption implies that no changes in access, docks, or other site features occur.) The only purpose of the trip is to recreate at the site. If this is not the case, the cost of the trip has a joint cost and benefits are overestimated. * • CASE STUDY: ALTERNATIVE APPROACHES FOR ESTIMATING RECREATION AND .RELATED BENEFITS OF THE MONONGAHELA RIVER* I ntroduction The travel cost model for the Monongahela River study assumes that site features or attributes affect the individual's ability to participate in recrea- tional activities at any particular site, as well as the quality of the recrea- tional .activities undertaken. It considers the demand for a recreation site as a derived demand. That is, a site's services are desired because of the recreational activities that can be undertaken at that site. Common sense suggests that a recreation site's features or attributes will influence the de- mand for its services. Since the level of water quality is a site attribute, a basis is established for relating water changes to shifts in demand for a recreation site's services. Approach The measurement approach examined numerous water-based recreation sites from the Federal Estate Survey component of the 1977 National Outdoor Recreation Survey. This survey provides specific information on the sample recreationist patterns of use during a single season for each site. The sample sizes for each site ranged from approximately 30 to several hundred respondents and included information on individuals' recreation behavior, socioeconomic characteristics, travel time necessary to reach the site, residen- tial location/ and a variety of other factors. This information permitted the estimation of individual travel cost demand models for each of the recreation. sites. x Several advantages of this travel-cost model include: *This discussion is taken from Desvousges, Smith, and McGivney [1983]. 3-9 ------- I I I I I I I I I I I I I I I I I I I Deriving individual " estimates for the time associated with traveling to the site as well as the roundtrip distance for each trip. Using the opportunity cost of time to evaluate travel time and estimating opportunity cost for each individual based on his characteristics, including age, education, race, sex, and occu- pation. Considering for each site the potential effects of individuals' differences in onsite time per visit. A total of 22 individual site demand functions were estimated based on this survey. For example, Equation (3.1) is a general description of one of these site demand models: In V ~ a + bTC + cY, (3.1) where In V = the natural -log of the number of visits by a household to the site in a recreation season. TC = the travel cost per visit to the site, including out-of-pocket vehicle operating costs and the opportunity cost of the time spent traveling. Y = family income. The basic hypothesis of this study is that variation in the estimates of a, b, and c across sites reflects the .effects of those sites' characteristics on the representative individual's demand for each site's services. Thus, each estimate provides the basis for describing how a change in any attribute would affect demand. The second step in the study involved estimating the relationship be- tween variations in the .site-specific estimates of a, b, and c and each site's attributes. The site characteristic information was obtained from records of the U.S. Army Corps of Engineers and the water quality data from the U.S. Geological Survey. Table 3-2 reports the site attributes, including the water quality measures used in the model. Many other attributes, such as boat launches, docks, and recreational facilities, were tried, but none was statis- tically significant. For sites where the information on water quality was incom- plete, the average value for all sites was used. This treatment of missing values means that the estimated relationships will rely primarily on sites with observed readings for the water quality variables. Since precision in the estimates of demand parameters a, b, and c in Equation (3.1) varied, a statistical procedure was used to account for the quality of the estimates. Table 3-3 shows the estimated equations for the demand parameters. As Table 3-3 shows, many of the attributes have statis- tically significant effects on the demand parameters, particularly on the travel 3-10 ------- I I I I I I I I I I I I I I I I I I I Table 3-2. Site Attributes Considered in Monongahela River Study Variable name Description Source SHMILEa ACCESS8 ARSIZE3 D0a'b Number of miles of shoreline for the site Number of developed multipurpose recrea- tional areas plus the number of developed access areas on the site The ratio of the pool surface acreage during the peak visiting period relative to the total project area in acres Dissolved oxygen based on monthly readings U.S. Army Corps of Engineers U.S. Army Corps of Engineers U.S. Army Corps of Engineers U.S. Geological Survey These variables were considered as monthly readings and as 4-month averages in the specification of the demand parameter models. K DOM and DOV correspond to the average value of dissolved oxygen over the four monthly observations and the variance about that average, respectively. Table 3-3. Estimated Equations for Site Demand Parameter Estimates* Variable Intercept term SHMILE ACCESS ARSIZE Average DO Variance in DO A a 1.51 (4.08) 0.0003 (1.25) -0.0059 (-1.50) -0.395 (-1.75) 0.0045 (1.07) 0.0005 (1.86) Site demand parameter 6 -0.0246 (-9.48) -0.00001 (-6.76) 0.00008 (2.81) 0.0033 (2.27) 0.00018 (5.99) 0.00001 (4.08) estmates c 0.000005 (0.308) 9.74 x 10~10 (0.09) 4.69 x 10~7 (2.56) -1.94 x 10~6 (-0.18) -1.22 x 10"7 (-0.60) 9.39 x 10"11 (0.01) a The numbers in parentheses below the estimated coefficients are the asymp- totic (approximate) t-ratios for the null hypothesis of no association—the larger the number, the more likely the hull hypothesis is rejected. These equations show how the parameters of the individual site demand equations vary with changes in the site's attributes. 3-11 ------- I I I I I I I I I I I I I I I I I I I cost parameter. The water quality attribute has a highly significant effect oh travel costs, with the most plausible results obtained using the mean value of dissolved oxygen over the 4-month summer period (June through September) and the variance in dissolved oxygen about that mean. However, the small variation in the water quality measures over the sample suggests these find- ings be interpreted cautiously. The model was used to evaluate the benefits of a water quality improve- ment for users of the Monongahela River in Pennsylvania. This site was not included in the 22. used to estimate the model. The model was applied to an independent data set based on a household survey of residents in the Monon- gahela River Basin (see the case study under Contingent Valuation below for more details). The survey reported sufficient information on the respondents' socioeconbmic characteristics, as well as specific portion(s) of the river used, to construct individual demand curves varying by river site (at a total of 13 different sites) and by individual. The benefit calculations were as follows: Estimates, of consumer surplus loss per user if the river were no longer available for its current use—recreational boating. • ' Estimates of the increment to consumer surplus associated with • improving water quality from the current level that permits boating, to a .level that would accommodate recreational fishing. Estimates of the increment to consumer surplus associated with improving water quality from the current level (boating) to a level that would accommodate swimming. The levels of dissolved oxygen used in the .benefit calculations for each of these use designations were the values selected by Vaughan in Mitchell and Carson [1982] in a water quality ladder developed for Resources for the Future (RFF). The variance in dissolved oxygen was held.constant at levels corresponding to those generally observed at the 22 sites. Table 3-4 provides the RFF ladder thresholds for each activity. Table 3-4. Specifications for the Dissolved Oxygen (DO) Levels Associated With Use Designations Use designation DO level (percent saturation) Beatable Fishable . Swimmable Drinkable 45 64 83 90 These thresholds correspond to those used in RFF's water quality ladder. 3These use designations were considered for benefit analyses. 3-12 ------- I I I I I I I I I I I I I I I I I I I Benefit Estimation Figure 3-4 illustrates the model. DCWQt) corresponds to the representa- tive individual's demand for the site's services at a water quality level of WQj. Since the demand function is specified as semi-log in quantity, there is no maximum price at which visits to the site will be zero. Therefore, the benefit calculation required a maximum feasible price--P*. This was taken to correspond to the largest travel cost incurred by any of the users of the Monongahela River ($22.65 per roundtrip). $/viiit P* - $22.65 D(WQ2) In visit yr Figure 3^4.General travel^ost demand-model lor a water quality improvement. The first benefit calculation involves the computation of the baseline area P.ABP*—the loss in consumer surplus if the .site could no longer provide the services available with a water quality of WQj to the user with a travel cost of P.. Incremental benefit estimates were derived by estimating the addition to consumer surplus associated with the increment to water quality. In Fig- ure 3-4, a change from WQX to WQ2 would be shown as leading to an incre- mental benefit of ACDB for the user at a travel cost of P.. Table 3-5 pro- vides a summary of the average benefit estimate for each change and the range of estimates over the survey respondents. This case study provides a detailed model for incorporating site attri- butes into estimating the demand for water-based recreation sites and, in turn, for evaluating the benefits from changing one or more of those attri- butes. Since It was developed from data primarily on flat-water recreation sites, the model can be used to predict benefits for changes in attributes for a sizable range of recreation sites. However, the actual estimates of the benefits of water quality improvements in the study must be regarded as tentative because of the limited available information on water quality. The approach illustrated by this case study could be used in a wide variety of applications. For one example, see Chapter 6, Section 6.4; 3-13 ------- I I I I I I I I I I I I I I I I I I I Table 3-5: Benefit Estimates .from Generalized Travel Cost Model With the Monongahela Survey Respondents Benefits' Use designations Mean Range Benefits lost as a result of the loss of ability to undertake boating activities Incremental benefits gained as a result of water quality improvement changing use designation from boatable to fishable Incremental benefits gained as a result of water quality improvement changing use designation from boatable to swimmable $53.35 $ 4.52 $ 9.49 0 to $70.80 0 to $8.60 0 to $18.30 The benefits are measured as the consumer surplus per user for the use of the river during a single recreational season. Survey—Contingent Valuation The contingent valuation survey approach for estimating the benefits- as- sociated with a nonmarketed commodity such as water quality improvements involves asking individuals about their willingness to pay for different levels of the commodity involved. Use of the survey approach requires that the practitioner determine the aspects of changes in environmental quality individ- uals value and convey these aspects to the respondent. The approach as- sumes that individuals will accurately reveal their valuation of potential be- havioral responses in hypothetical market experiments. These experiments depend on a survey procedure and a survey instrument. The survey proce- dure determines the appropriate sampling plan and specifies the general re- quirements of the survey instrument. The survey instrument is the question- naire used to elicit the respondents' answers. A survey instrument is the cornerstone of the hypothetical market used in the contingent valuation survey approach. It will generally consist of the following sections [Rowe and Chestnut, 1981.]: Introduction and statement of purpose Nonvaluation questions 3-14 ------- I I I I I I I I I I I I I I I I I I I Scenario development and market definition Bidding or valuation questions. The first two sections are self explanatory.and will not be discussed further. The scenario .development and market definition section is considered the most critical aspect of survey instruments because it must carefully present the alternative levels of environmental quality. In the case of water quality, scenario development describes the linkages between the regulatory action and the resulting.change in water quality. Verbal or written descriptions, supple- mented by visual props, are used in this activity. Scenario development must be informative and realistic. It must portray the probabilities of the effects, as well as the effects themselves. After the hypothetical scenario is developed and the market is defined, the bidding or valuation of the environmental commodity takes place. Several questioning formats can be used: Pi rect Question - - The interviewer directly asks the individual's willingness to pay for a specified change in the amount of a commodity—water quality--that has been carefully defined. No cards or other aids are used to obtain the amounts. Bidding Game—The interviewer defines the change to be evalu- ated, suggests to the individual an amount representing the value of the change (the starting point), and asks whether he would .be willing to pay that amount. Based on the response, the interviewer raises or lowers the suggested value by a fixed amount and repeats the process until the individual agrees no further change is necessary. * Payment Card—This approach also does not directly ask about willingness to pay. The interviewer explains the specified change to be evaluated, provides the individual with a card displaying an array of potential values, and asks him to select a value or give any value for willingness to pay. These num- bers range from zero to values judged to be outside the range .of responses. Some surveys (notably Mitchell and Carson [1981]) have adjusted the upper bounds of values on the cards for higher income respondents. In addition, in an anchored payment card format, some responses have been identified as reflecting the share of an individual's taxes associated with specific public programs, such as education and defense (see Mitchell and Carson [1981])- Bidding Game With Budget Constraint—This approach is a very recent innovation for the bidding game format discussed in Brookshire et al. [1982]. Before requesting a bid in the for- mat explained above, the interviewer asks the individual to estimate his after-tax monthly income and allocate it into ex- penditure categories--for example, electricity, shelter, enter- 3-15 ------- I I I I I I I I I I I I I I I I •. I I tainrrient, savings, and a residual. After this information is acquired, the interviewer conducts a bidding game with an additional question: "Which of the categories of expenditures would be reduced in order to make the proposed payment?" Ranked Choice and Willingness to Pay--ln this approach, intro- - duced by Rae [1981a, T981b], the interviewer provides individ- uals with different hypothetical market outcomes—proposed pay- ments and a specified level or change in water quality, for example, to be ranked. These ranks are then used in a statis- tical analysis to estimate the individual's willingness to pay. The question format and description of the hypothetical market are important determinants of the quality of the estimates derived from a contingent valua- tion experiment. The results of a comparative analysis of the direct question, payment card, and bidding game formats in Desvousges, Smith, and McGivney [1983] suggest that, for questions associated with water quality, the question format has some effect on the average willingness to pay. The starting point for the bidding game appears to influence the average values. Two basic types of biases can arise in designing the format of a contin- gent valuation study. Since Schulze, d'Arge, and Brookshire [1981], Rowe and Chestnut [1981], and Mitchell and Carson [1981] have all discussed these biases in detail, a brief overview of their conclusions is provided in Table 3-6, which defines the bias, identifies the studies that considered its potential effects, and summarizes the current understanding of its effects. Overall, the results seem to suggest that starting point bias may be the most important consideration (aside from the hypothetical nature of the ques- tions, which has not received sufficient testing to fully gauge its implications) in using the contingent valuation framework. Most of the other potential sources of bias can be controlled in the structuring of the instrument and the explanations provided to sample respondents. Several additional technical assumptions are highlighted below in a summary of data requirements and key assumptions of the approach. Data Needs, Key Assumptions/Limitations, and Features The following checklist outlines the data needs for the contingent valua- tion survey approach, along with its key assumptions and features. Data Needs: Survey of individuals designed to be representative of affected population. Clearly defined and pretested survey instrument. In-person interviews are generally more reliable than telephone or mail surveys. 3-16 ------- I I I I I I I I I I I I I I I I I I I Table 3-6. Summary of Biases in Contingent Valuation Experiments Type of bias Definition Studies that have tested for bias Summary of current results General Hypothetical Error introduced by posing hypothetical conditions rather than actual condi- tions to an individual; response may not be a good guide to actual actions individual would take One known test-- Bishop-Heberlin [1979] Strategic Attempt by respondents to influence out- come of study by systematically over- or under-bidding so action favors their true interests; strategic responses depend on how payment scheme is defined and whether it is believed At least eight tests (see Schulze, d'Arge, and Brookshire [1981] for summary; Cronin [1982]) Instrument Related Starting point Contingent valuation experiments using bidding game format have started with suggested payment and use yes or no responses to derive final willingness to pay; suggestion may be perceived as appropriate bid At least five tests (see Schulze, d'Arge, and Brookshire [1981] and Rowe and Chestnut [1981]) Vehicle Information Interviewer Characteristics of proposed mechanism • for obtaining respondent's willingness to pay may influence responses Effect of information provided to respondent on costs of action under study or other dimensions of problem may affect responses Responses vary systematically according to interviewer At least four tests (see Schulze, d'Arge, and Brookshire [1981] and Mitchell and Carson [1982]) At least four tests (see Schulze, d'Arge, and Brookshire [1981] and Mitchell and Carson [1981]) Two tests-- Desvousges, Smith, and McGivney {1983] and Cronin [1982]) Some indication that hypothetical nature of question did influence responses, but could not dis- tinguish this effect from instrument- related biases Very little evidence of strategic bias except for Cronin [1982] Some differences in opinion over impor- tance of starting point bias; Mitchell-Carson feel starting point bias is important, and Desvousges, Smith, and McGivney [1983] provide some support; Schulze, d'Arge, and Brookshire [1981] feel it is more limited Some evidence of effects in at least two studies Limited evidence of effects No evidence of bias Bias present The definitions and results summarized in this table are based on Schulze, d'Arge, and Brookshire [1981], Rowe and Chestnut [1981], and Mitchell and Carson [1981]. 3-17 ------- I I I I I I I I I I I I I I I I I I I Key Assumptions/Limitations and Features: Individuals' responses to hypothetical questions are assumed to be indicative of their actual valuations of the changes de- scribed in the questions. Careful tests are required to determine starting point effects, appropriate mechanisms for payment, and consistency of re- sponses with other budgetary requirements. Careful Control is required over information given respondents so answers are based on the same information in each inter- view. **************** CASE STUDY: RECREATION AND RELATED BENEFITS OF WATER QUALITY IMPROVEMENTS OF THE MONONGAHELA RIVER* Introduction This contingent valuation survey measured the recreation and related benefits of water quality improvements in the Monongahela River Basin in Pennsylvania. It compared alternative question formats for asking individuals' willingness to .pay and measured both user and nonuser values. In a house- hold survey conducted by 9 professional interviewers from the five-county area, an 80 percent response rate was obtained from a clustered random sample of 393 households. Approach In any contingent valuation study, the survey questionnaire is the key element for providing plausible results. By dividing the questionnaire into a version for each question format and distributing each version equally among the interviewers, the Monongahela study compared the techniques, A water quality ladder, developed by RFF (see the travel cost case study for more detail), was used to establish a linkage between levels of water quality and the associated uses for recreation. Tied to scientific meas- ures of water quality, the ladder steps permit the respondent to give his will- ingness to pay for the various levels of water quality. The contingent valuation method requires a way to make the hypothetical payment for water quality improvements. User fees, increases in sales taxes, and increases in water bills are among the alternatives used. This study expressed the additional annual amounts as taxes and higher consumer prices. Also used by Mitchell and Carson [1981], this method corresponds roughly with how a respondent actually pays for water quality improvements. *This discussion is taken from Desvousges, Smith, and McGivney [1983]. 3-18 ------- I I I I I I I I I I I I I I I I I I I Each version of the questionnaire used a different technique to elicit the respondent's willingness to pay. The iterative bidding technique was used in two versions, with the interviewer starting the bidding at $25 in one version and at $125 in another. The direct question techniques and the payment card were also used. The payment card contained values arrayed from $5 to $775, but no other information. All versions of the questionnaire required the respondent to place a value on a degradation in water quality in the Monongahela from its present overall level of beatable to a level where the river was unsuitable for any recreation. Additional amounts were elicited for water quality improvements to support fishing and swimming. These amounts reflected actual use and potential use in the future, with a followup question requiring the respondent to break down the amounts into actual and potential use. . Benefit Estimation For each version of the questionnaire, Table 3-7 presents the average amounts users and nbnusers of the Monongahela River were willing to pay for Table 3-7. Willingness to Pay for Three Levels of Water Quality9 ($/yr) Users Nonusers Combined Mean Standard deviation Number of respondents Payment card 117.9 (47.1) 82.8 117.0 (53.8) 104.7 17 37 93.8 (71.6) 108.9 (92.8) 54 Mean Standard deviation Number of respondents Direct question 98.2 (47.4) 34.5 103.5 (81.5) 66.4 17 34 55.7 (38.8) 85.2 (71.8) 51 Bidding Game: $25 starting point Mean 59.5 (42.4) 51.4 Standard deviation 38.1 (31.9) 53.1 Number of respondents 19 39 54.1 (48.4) 48.5 (47.1) 58 Mean Standard deviation Number Bidding Game: $125 starting point 194.4 (109.4) 79.2 136.5 (129.2) 102.5 16 32 117.6 ( 89.3) 126.0 (111.6) 48 As defined in Section 3.2, numbers in parentheses are individuals' estimated mean option values and corresponding standard deviations. 3-19 ------- I I I I I I I I I I I I I I I I I I I avoiding a decrease in water quality and for raising it to swimmable levels.* Several conclusions can be drawn from the results in Table 3-7. The esti- mates of willingness to pay—regardless of method used to elicit the amount— are quite substantial, ranging from $35 to $195 per year. The results are sensitive to the method used; the payment card with the $125 starting point bidding game gave higher amounts than the direct question with the $25 start- ing point bidding game, there is some indication of a starting point bias, but the evidence is not conclusive. Users of the Monongahela generally were willing to pay higher amounts than nonusers, with their average values ranging from $59 to $194 compared to a nonuser range of $34 to $83. Estimates of option value are about half the user's willingness-to-pay values and are statistically significant for both users and nonusers. The results imply that benefit estimates based solely on recreation use may substantially understate the total benefits of water quality improvements. The approach illustrated by this case study could be used in a wide variety of applications. .For one example, see Larson [1981]. Survey--Recreatioh Participation Models Many State and Federal agencies undertake surveys of the general popu- lation in an effort to identify household participation patterns for recreational activities. As a rule, these surveys provide detailed information on house- hold characteristics and on the types and amounts of participation in outdoor recreation. These surveys have been used to .estimate recreation participation models. Such models are neither demand nor supply relationships but sum- maries of all the determinants of the likelihood that an individual will parti- cipate in recreational activities—for example, boating, fishing, or swimming-- as well as of the level'of participation in these activities. Generally, these models divide the participation decision into two steps: determining whether a person participates in a particular activity and modeling the expected num- ber of days (or trips) he spends at the.activity over a season. These models have been developed from a framework that views the indi- vidual as maximizing well-being by selecting levels of service flows for his consumption. Individuals produce these service flows by using time and/or purchased goods and services. For example, the level of participation in a recreation activity is one measure of recreation service flow that requires the person's time, any equipment associated with the activity, and the services of a recreation site as inputs. Participation models, which describe the final result of the activity, are influenced by each. element in the production of service flows. *These mean amounts are calculated exclusive of the respondents who rejected the approach and those who were shown to be outliers by a statis- tical analysis. For a complete discussion of the procedures used to make these determinations and the small differences that result from the exclusions, see Desvousges, Smith, and McGivney [1983]. 3-20 ------- I I I I I I I I I I I I I I I I I I I This perspective is important because it establishes a natural association between participation and travel cost models. Travel cost models are demand models for the services of a recreation site. These services add to the pro- duction of the recreation service flows. In addition, benefits measured by a participation model should, for consistency, use the demand for the recreation service flow and not for the recreation site, but these demands are difficult to isolate.* This often makes assignments of benefit estimates resulting from a change in water quality arbitrary (see Davidson, Adams, and Seneca [1966]). Instead of identifying the recreation facilities used by the survey, re- spondents, participation surveys have, as a rule, required crude measures of recreation supply characteristics to be merged. In principle, this merging permits the use of summary measures of water quality for regional areas as determinants of the likely participation and the level of participation of the representative household. Yielding crude approximations at best, this prac- tice reflects the paucity of data in this area. However water quality is introduced, it should be emphasized that the results of these models are estimates of the levels of use of recreation activi- ties and not economic benefits. Methods for measuring these values must be developed independently. Data Needs, Key Assumptions/Limitations, and Features The following checklist outlines the data needs for the participation sur- vey method, along with its key assumptions and features. Data Needs: Survey of recreation patterns of the genera! population, with socioeconomic detail and identification of residential location (preferably in more detail than State of residence). Identification of sites used for recreation activities, or at least some measure of the supply of recreation facilities, is highly desirable. . Measures of water quality for sites used by respondents, or linkage between water quality and capacity-related measures for recreational activities. Key Assumptions/Limitations and Features: An independent estimate is required of an individual's willing- ness to pay for a day or a trip spent in each recreational activity. *See Deyak and Smith [1978] and Bockstael and McConnell [1981]. 3-21 ------- I I I I I I I I I I I I I I I I I I I The demand and supply relationships are assumed to remain stable. Model specification (i.e., two-step partition of participation decision and level of participation) is assumed to be correct, and functional forms are assumed to be adequate approxima- tions. Measures should be provided at a general level, not on a site- specific basis. **************** CASE STUDY: A PARTICIPATION SURVEY APPROACH TO VALUING WATER QUALITY IMPROVEMENTS* , Introduction and Basic Logic The Vaughan-Russell [1982] study, the most ambitious and detailed application of a recreation participation model to date, focused on the recrea- tional fishing benefits that arise from a change in water quality. It used the fact that more "desirable" freshwater sport fish--coldwater and certain warm- water species—require better water quality. Improved water quality may alter the types of fish that can be supported in a water body. Assuming the supporting recreation facilities are available, Vaughan-Russell suggest that there will be a change in the type of fish (and perhaps a net increase in the level of fishing participation) from less desirable to the more desirable varie- ties. The sources of benefits from the water quality change arise from: The change in the composition of fishing activities Any net increase in the level of .participation {in fishing. to implement this logic on a national scale, the Vaughan-Russell objec- tive, requires the following: Measuring the availability of freshwater bodies for fishing and their water qualities at a geographically disaggregated level. Modeling and measuring the influence of water quality on par- ticipation in recreational fishing and on fishing .activities by type of fish sought. Measuring the economic benefits according to the type of fish sought. *This study is taken from Vaughan and Russell [1982]. 3-22 ------- I I I I I I I I I I I I I I I I I I I This review highlights some of the key elements in this application of partici- pation models. Approach The first step in the analysis was to estimate the available "fishable" waters. Using dissolved oxygen and suspended solids criteria, Vaughan- Russell projected total fishable water on a pre- and post-policy basis. - For each policy scenario, they used the following steps: 1. Calculate the percentage increase from baseline levels of total fishable water as represented in the RFF water quality network model. 2. Estimate the policy impact by applying the improvement factors from Item 1 above to a national baseline of fishable acres per capita. 3. Estimate the change in composition of fishing activities by using the water quality network model to calculate fishable water by species type. These steps are based on a recreational fishing participation model that divides a person's participation choices into three decisions: 1. "Decide whether or not to fish" 2. "Decide what to fish for" 3. "Decide on a level of participation." The model implies that the amount of fishable water available affects the probability of an individual's being a fisherman. Then, the suitability of water quality to support a class of fish (e.g., coldwater, warmwater, game, and rough) affects the type and level of fishing activity. Policies that change water quality affect the availability of each type of fishable water. In the Vaughn-Russell study, this relationship established the necessary technical linkage between water quality and behavior (see Chapter 1). The empirical analysis of participation was based on the 1975 National Survey of Hunting, Fishing, and Wildlife Associated Recreation. For Parts (1) and (2) of the fishing decision, probability models were estimated using several statistical techniques. The models included a wide array of socio- economic variables (e.g., age, sex, income, region, residency in metropolitan area, residency in State with coastline, and total acres of fishable water in the State per capita). The estimated effect of acres of fishable water per capita on the likelihood of participation in fishing was positive and statisti- cally significant. U.S. EPA Headquarters Library Mail code 3201 3-23 1200 Pennsylvania Avenue NW Washington DC 20460 ------- I I I I I I I I I I I I I I I I I I I Several methods were considered for estimating the second-stage proba- bility models--!.e., for the particular type of fishing selected. The final set of models for these probabilities was based on three overlapping fishing choices--some trout, some bass, and some .-rough fish. The most significant effect of the fishery-specific water quality variable .was the coldwater game- fish category. Nonetheless, the signs of 'the effects of the water quality variables generally agree with a^ priori expectations for all categories of fishing. In the last component of the model, the number of days spent fishing in each category were estimated. The participation model requires the use of all three components to evaluate the implications of a change in water quality on the types of fishing chosen. Table 3-8 illustrates the results of one of the Vaughan-Russell scenarios—adoption of Best Practicable Technology (BPT) for the predicted changes in the mix of fishing .activities undertaken. The last three lines in Table 3-8 provide the "bottom line" implications of the model in physical terms. They are not benefit estimates but, rather, increases in the number of fishing days of various types. If benefit estimates are to be de- rived from the model, these fishing days must be valued. Table 3-8. Vaughan-Russell Model--Predicted Effects of BPT Regulations on Participation Change from base case Change in relevant variable BPT/BASE" Probability of being a fisherman Probability of doing some: Trout fishing Bass fishing Rough fishing Days per capita per year: Trout Bass Rough Total days per year: Trout Bass Rough +0.0001' +0.0076 -0.0142 -0.0039 +0.02 +0.34 +0.51 +7-2 x 106 -1.3 x 106 +5.6 x 106 SOURCE: Vaughan and Russell [1982], Table 6-1. ' aThe logit estimates using a sample size of 5,000 were used for these esti- mates . This may seem an inconsequential change in the probability of being a fisherman, but it implies an>increase of 20,000 fishermen per year. 3-24 ------- I I I I I I I I I I I I I I I I I I I Benefit Estimation To value these fishing days, Vaughan-Russell conducted a separate sur- vey of fee fisheries in the United States and used the results of the survey to estimate a travel cost model, accounting for the effects of a number of characteristics of the individual fisheries.* Separate models were estimated for trout and catfish (as a basis for valuing the rough fishing). The esti- mates of consumer surplus were converted to a per day per person basis. Considering the differential in these values between trout and catfish, the estimates ranged from $1.77 to $8.06 in the final models. These estimates provided the basis for valuing the increments to fishing days projected in the participation model. When the relevant daily per capita consumer surplus is applied to the estimated increments in fishing days, the incremental benefits can be calculated from the Vaughan-Russell model. One of the most important potential limitations to the Vaughan-Russell methodology is the procedure used to value fishing days. The relevant bene- fit measure is the demand for fishing as a recreational service flow, not the measure estimated from the demand model for a site's services. While there is a correspondence (see Anderson [1974] or Carlton [1979]), the relationship between the two will depend on the nature of the other inputs to fishing and the activities undertaken at the site. Since the Vaughan-Russell model relies on a very specific definition of the recreational activity and treats trips to the fisheries as single-day visits, the discrepancies may not be great. How- ever, the transfer of the relationship between travel cost and participation models implied by their framework may not be possible in other applications. There should be little doubt, even with this cursory review, that the Vaughan-Russell model represents an- enormous undertaking and is the best effort available to date for modeling recreation participation. The approach illustrated by this case study could be used in a wide variety of applications. For one example, see Chapter 6, Section 6.4. Hedonic Property Value Method Two types of recent models use market data on either property values or real wages along with quantitative measures of environmental amenities to esti- mate individual willingness to pay for a change in one or more amenities. Re-" searchers have applied both property and wage models to value air quality but have used only property value models in the case of water quality. These models, known as hedonic models, use two assumptions: (1) par- ticipants in a market accurately perceive the characteristics of different hous- *The characteristics did not include water quality because the fisheries were separated by the type of fish. 3-25 ------- I I I I I I I I I I I I I I I I I I I ing sites—including water quality-and, in making their location decisions, will consider them along with the prices for the housing units; and (2) there exists a continuous array of combinations of these characteristics across dif- ferent housing sites within the market. An ideal market process will ensure that equilibrium housing prices and rents will reflect the marginal valuations of the characteristics. Data Needs, Key Assumptions/Limitations, and Features Only a checklist on data needs, key assumptions, and features is pre- sented because the resources required to use this method would exceed those available in most States. A detailed discussion and a related case study will be presented in Volume II. Data Needs: Property values (preferrably sale price) for residential sites around water bodies with different water qualities in the same housing market. Information on other site and neighborhood characteristics that may affect property values. .Information on individuals' perceptions of water quality and relationship to available physical measures of water quality. Key Assumptions/Limitations and Features: Market equilibrium Full knowledge of the implications and effects of water quality Ability to determine extent of market and specify relationships for hedonic price and demand functions Full adjustment and ease of mobility. Damage Function Method The damage function method applied to valuing the benefits from water quality improvement is most relevant for the effects of water quality on human health. In principle, this approach examines all the possible physical effects of each type of emission into a water body. However, usually only the health effects are considered. To use this approach it is necessary to estimate, for each class of effect (i.e., chronic versus acute), health impacts that stem from the relationship between the physical effect and the concentration of the relevant water pol- lutant, as well as any other factors that might influence the pollutant's impact. These relationships are the damage functions. They are used to estimate the physical effects of specified changes in water quality as measured 3-26 ------- I I I I I I I I I I I I I I I I I I by the concentrations of each of the individual pollutants. The method does not provide a way to value the physical effects so that independent estimates of the health benefits must be developed. Data Needs, Key Assumptions/Limitations, and Features The following checklist outlines the data needs for the damage function method, along with its key assumptions and features. A detailed case study will be presented in Volume II. Data Needs: Measures of concentration of relevant pollutants in water used by population over time and over .the geographic location. Measures of features of population and health patterns. Measures of other exogenous factors that may also affect ob- served health patterns of population. Key Assumptions/Limitations and Features: Provides largely statistical summaries of data on existing popu- lation experiences. Assumes no behavioral substitition on the part of populations in response to levels of each pollutant. : Produces results sensitive to the statistical procedures used to estimate the models. Maintains the primary advantage allowing classification of effects according to physical impacts. 3.5 BUSINESS BENEFITS Benefits from the water quality programs can accrue to firms as well as to households because many provide for a wide spectrum of uses for rivers and streams, including industrial/commercial, agricultural, navigation, and municipal water supply uses. Measurement of business benefits are often easier because market prices are usually available to value these benefits. Theory: The Supply Function* In addition to the demand function, discussed earlier, economics provides a second organizational guidepost for measuring benefits—the supply function, *This discussion is a summary of the discussion in Just, Hueth, and Schmitz [1982]/Chapter 4. 3-27 ------- I I I I I I I I I I I I I I I I I I I shown in Figure 3-5. The supply curve shows the maximum quantity of out- put of good X the firm is willing to supply at each relevant price. If the market establishes the price at P , the firm will produce the quantity Q . The upward slope of the curve in Tigure 3-5 indicates that the firm is willing to sell more at higher prices than at lower prices, assuming that factors in- fluencing the supply function—the prices of inputs, such as labor, energy, machinery, and technological improvements—do not change. Price $/X Quantity X time Figure 3-5. Supply function and the producer surplus. The concept of producer surplus is used as the general measure for a change in the welfare of a firm. In Figure 3-5, producer surplus is shown as an area equal to the area above the supply curve and below the price line for the firm or industry. Whether a producer is better off can be determined by examining the change in producer surplus. Producer surplus provides a measure of a change in welfare for a firm because its welfare is measured directly in dollars of cost savings. This view of a firm is a simplified model that does not include important differences among firms or the distribution of profits among owners and resource suppliers. Practice: Cost Savings Method for Measuring Business .Benefits The estimation of a firm's benefits from water quality has been much less sophisticated than the estimation of household benefits. The primary focus has been on estimation of the cost savings associated with the water quality change. The estimates are derived largely from engineering cost estimates. In principle, economic cost functions could provide the basis for these esti- mates, but, in practice, they have not. The hypothetical example for business benefits is for irrigation, but the same method could be adapted to the other situations. The incremental pro- ducer benefits that might arise from water quality programs are: 3-28 ------- I I I I I I I I I I I I I I I I I I I • Reductions in industrial cost—firms using water in processes may have reduced production costs because either less treat- ment of water is required or less maintenance is required for pumps, pipes, and other parts of the production process. Reductions in agricultural costs—farmers using water for irri- gation may have reduced production costs because less mainte- nance of irrigation equipment is required or less fertilizer per bushel of crop is needed. Reductions in navigation costs—barges and other water trans- port conveyances may have reduced risk of accident. Ship maintenance may be reduced. From .these simple examples two major points arise. The first is that the focus on incremental benefits of water quality decisions will mean these benefits are considerably smaller than they would be if meas- uring the total benefits of all water regulations were the primary objec- tive of the benefit-cost assessment. That is, the baseline is important in .measuring the benefits of a particular decision. The second point Is that the relative orders of magnitude of these benefits will be very speci- fic to the Individual water bodies evaluated. For example, the comple- mentary attributes necessary for recreation, such as access and overall surroundings, might be at very low levels in some instances, while the potential for producer benefits from other designated uses is very large. This point emphasizes the importance of the focus in .the proposed water quality standards program on selecting key segments and considering each on a specific basis. Data Needs, Key Assumptions/Limitations, and Features The following checklist outlines the data needs of the cost savings approach, along with its key assumptions and features. Data Needs: Cost data for firm (see Chapter 4 for details) Demand information such as market prices and responsiveness of sales to price changes. Key Assumptions/Limitations and Features: All products and inputs (labor, machines) are bought and sold In markets that are perfectly competitive; that is, no buyer or seller has influence over market prices. The supply curve reflects the marginal social cost of producing the product or service. This implies that neither external costs nor subsidies are present in the market. (This is un- likely for the irrigation example because of the various legal and regulatory influences in the market, but it is a useful as- sumption to simplify exposition.) 3-29 ------- I I I I I I I I I I I I I I I I I I I A shift in the supply of a :producer's services will not affect the price at which they are sold in the market. * . * CASE STUDY: IRRIGATION. BENEFITS Suppose that a State is considering a change in the designated uses.of a river segment to provide for fish and wildlife propagation. As a byproduct of this use designation, the quality of water available for irrigation is as- sumed to improve, thus shifting the supply of irrigation services—because more high-quality water is available for irrigation—outward from Sj to S2, as shown in Figure 3-6.* In an application, the practitioner may have limited data on some costs, but seldom enough to estimate the entire supply function. By supplementing the available data (e.g., from the Bureau of Reclamation or the U.S. Department of Agriculture [USDA]) with assumptions based on com- mon sense, a rough cut at the problem can be obtained. For example pur- poses, the entire supply curve is drawn. $/103 gat US o> 1 <*•» o I 1 a. Change in ~ Producer Surplus irrigation services 100 200 Quantity of Irrigation Service Figure 3-6. Irrigation benefits. 300 103 gal/yr *The stream's flow is assumed to be strong enough that the increases in irrigation will not noticeably reduce it. Prior to the change in use designa- tion, the limiting factor is assumed to be water quality rather than flow. 3-30 ------- I I I I I I I I I I I I I I I I I I I The task is to calculate the irrigation benefits from this shift in supply-- the change in producer surplus attributed to the implementation of the new designated use. The crosshatched portions of Areas a and b in Figure 3-6 are the two components of the change in producer surplus. Area a is the new producer surplus resulting from reduced cost of the original volume of irriga- tion services (100). Area b is the new producer surplus on the additional volume of irrigation attributable to the reduced cost resulting from the water quality standards. As drawn, Area a is a parallelogram showing the cost savings on the original irrigation volume. Using the formula for the area of a parallelogram: .Area a = (side) x (perpendicular distance to parallel side) Area a = $(2-1) x (100-0) Area a = $(1) x (100) Area a - $100/year. Area b is the additional irrigation induced by the decrease in costs due to the water quality standards. Using the formula for the area of a triangle: Area b = ^(base) x (height) Area b = ^$(2-1) x (300-100) Area b= %$(D * (200) Area b - $100/year. Thus, the change in producer surplus—the measure of firm benefits— attributable to the cost savings from the additional river uses in the water quality standards is $200 a year. There are several important caveats to the forgoing simple example for estimating firm benefits: The costs are not quite as simple as in this example. However, the basic measurement concept still applies. The assumptions required for the example are stringent ones, but they do provide a workable approximation for many indi- vidual river segments. In the cases of agriculture and nagivation, institutional factors in those markets may distort the true social cost. For ex- ample, the subsidization of waterway activities and the regu- lated rates in railway and highway transportation may violate the assumptions of perfect competition in those markets. 3-31 ------- I I I I I I I I I I I I I I I I I I I If the assumption of perfect competition in the input markets does not hold, the producer surplus may actually accrue to providers of labor or capital services. Market power resulting in control over market prices and out- puts will distort the supply relationships and make prices higher than in competition. 3.6 PUBLIC WATER SUPPLY BENEFITS Reductions in treatment requirements for municipal water supplies consti- tute another potential source of benefits from water quality program decisions. By having a use designation that provides for fish and wildlife propagation, a city that uses a river as a water supply source may be able to provide water with less treatment than if the use designation were agricul- tural or industrial. Once again, it is essential to remember that the focus of the benefit measurement .should be the incremental benefits attributable to the particular policy, not total benefits from all water regulations. Since calculation of these benefits could proceed exactly as in the case .of business .benefits, no case study is provided. The same key assumptions apply as with business benefits. The critical issue of toxic substances or toxic pollutants, which would apply to a public water supply, is not considered in this handbook. As more information becomes available on the extent and effects of toxics and their relationships to the water quality regulations, a change in this focus may be warranted. This is an issue that bears future scrutiny because of .potential health benefits from reducing toxic pollutants. 3.7 SUMMARY This chapter has reviewed basic benefits concepts and the approaches used to measure them. An individual's willingness to pay is the centraKtenet underlying all the methods discussed in this chapter. Even though all assess- ments may not require the practitioner to use measurement methods, willing- ness to pay provides an organizing principle for even qualitative assessments of benefits. Major points developed in the chapter include: The assessment should be tailored to balance the complexity and importance of the policy action to the available resources. • The change in consumer surplus should be used as the meas- ure of willingness to pay for improvements in the well-being of households. 3-32 ------- I I I I I I I I I I I I I I I I I I I The travel cost method can provide willingness-to-pay-based measures of water quality changes under many conditions. It measures only user benefits and is sensitive to the treatment of time costs and substitute sites. The travel cost method estimates demand for site services. •The contingent valuation survey method also provides wifling- ness-to-pay-based measures of water quality changes under many conditions. It can measure both user and intrinsic bene- fits and is sensitive to questionnaire design and administra- tion. The participation survey method provides measures of changes in level of use—visitor days—for a recreation activity and often requires ad hoc valuation of use to develop benefits. Cost savings can provide estimates of willingness-to-pay-based measures (producer surplus) of changes in the economic well- being of firms. 3-33 ------- I I I I I I I I I I I I I I I I I I I CHAPTER 4 MEASURING THE COSTS OF WATER QUALITY PROGRAMS 4.1 INTRODUCTION Why is opportunity cost the basis for measuring costs? What types of approaches are available to measure costs? What practical problems arise in measuring costs? Are there any examples of cost measurement? This chapter focuses on determining the incremental costs of a water quality program. It defines various cost categories, discusses their relation- ships, and reviews costing methods relevant for users of this handbook. It covers costs for both industry and publicly owned treatment works (POTWs) and presents several case studies to illustrate how the general principles are applied. Although the examples are oriented toward potential water quality standards decisions, they are.general enough for use.in other applications. Throughout this chapter, opportunity costs—the value to society as a whole of a resource's best alternative use—provide the measurement basis for costing in a benefit-cost assessment. Engineering and accounting cost esti- mates may differ from opportunity costs because of cost-sharing .mechanisms such as taxes and subsidies. In these cases, the practitioner faces the dif- ficult task of determining the value of opportunity costs. This chapter suggests practical approaches to costing and highlights some of the more difficult issues, which will get .more detailed treatment in Volume II. Specifically, Section 4.2 presents the basics underlying the meas- urement of costs, and Section 4.3 describes two general approaches, engineer- ing and econometric, to measuring costs. Section 4.4-defines types of cost and major cost categories to be used in an assessment. Section 4.5 discusses the practical aspects of determining costs—including data, sources, the use of indexes, and major factors influencing cost estimates—and concludes with a sample data form. Section 4.6 describes the engineering methods for estimat- ing costs, and Section 4.7 provides examples using the engineering methods. Section 4.8 offers some general cautions for measuring costs. Finally, Section 4.9 summarizes the chapter's major points. 4.2 MEASURING COSTS: THE BASIC CONCEPTS This section reviews the fundamental economic principle of cost: oppor- tunity cost, which measures the cost of any resource in terms of its next best alternative use. That is, the value of forgone alternative uses for any resource provides the basis for estimating the cost of any specific use of that resource. 4-.1 ------- I I I I I I I I I I I I I I I I I I I For a water quality decision, opportunity costs include both explicit costs (e.g., wages and salaries, or payments for materials and energy) as well as implicit costs (what self-owned and employed resources could have earned in their best alternative uses). For example, the implicit cost of labor is the highest wage the owner could receive for his labor services. For firms and households, the opportunity costs are the private costs of a regulatory action. If the action also negatively affects others, either households or firms, additional costs are incurred—in technical terms, exter- nal costs. In a benefit-cost .assessment, the opportunity cost to society is the relevant measure of cost, the sum of private and external costs. The economic guideposts of supply and demand functions described in Chapter 3 can be used to view the opportunity cost concept. These functions are combined in Figure 4-1, which illustrates a market for good X. The de- mand curve, D, shows the amount demanders are willing to buy at each of several prices, while the supply curve, S, reveals the amounts suppliers will provide at various prices. Market forces'will cause the price to settle at P0, with the resulting quantity at Q0. Price $/Ox Deadweight lots Q! QQ Quantity/time Figure 4-1. Measurement of opportunity/costs. One way to view the cost of a regulatory action is to suggest it will im- pose costs on firms, resulting in a shift of the industry supply curve from S to S1. In this case, the regulation causes a shift in the curve by an amount that recovers the costs of compliance—shown by the hatched area in Figure 4-1. These costs constitute one element in the opportunity cost of the regu- lation—the additional opportunity costs (e.g., extra operation and mainte- nance) required to meet the regulatory standard. The second element of op- portunity cost is shown by the shaded triangle—the loss to society because the produced and purchased quantity of X is reduced from Q0 to Qt when supply shifts. In technical terms, some producer and consumer surplus is lost (dead weight loss). 4-2 ------- I I I I I I I I I I I I I I I I I I I In practice, it is usually impossible to construct the supply curve for an industry. Instead, the practitioner can use the available cost information, often for an "ideal" plant, and assume the supply curve is horizontal, in most cases/ the loss in accuracy caused by making this assumption is not a severe limitation because the estimated compliance costs, the largest compo- nent of opportunity - cost, are less affected than the dead weight loss to society. Under most circumstances, market prices of goods and services provide a good estimate of their social opportunity costs. However, in some cases, there can be significant differences between market price and social cost. As .noted in Chapter 2, the most significant case relevant to water quality pro- grams is the difference between social and private rates of return on capital due to corporate income taxation. In addition, subsidies and grants may cause the financial costs to participants to differ from social costs. 4.3 MEASURING COSTS: TWO GENERAL APPROACHES Evaluating the costs of regulatory actions requires the collection and analysis of relevant industry and municipal cost data. A thorough accounting of costs for potentially affected firms and cities is desirable. There are two main approaches to estimate these costs: econometric cost estimation and engi- neering cost estimation. Econometric cost estimation, sometimes referred to as the statistical ap- proach, uses cost-output relationships that are identifiable through empirical testing. Sometimes, statistical cost estimation offers a way to determine the costs of proposed alternatives. It is possible that data can be gathered for actual firms, most of which have identifiably different production processes and some of which already meet regulatory alternatives of the type under con- sideration. Given sufficient data, production relationships representative of complying and noncomplying firms can be estimated. Then, if the prices of labor and equipment are known, it is possible to establish cost-output relation- ships known as cost functions. These cost functions, in turn, can be used to evaluate the cost of regulatory alternatives. Unfortunately, empirical studies of this type are rarely practicable for regulatory analyses. The major difficulty usually is that sufficient technical data are .not available. In particular, published data are usually scarce, and potentially useful data, available only from firms, are typically considered proprietary. The second approach to cost estimation is the engineering cost approach. This approach offers a viable alternative to statistical cost estimation because it does not rely on the availability of a firm's actual data. Rather, engineers familiar with relevant industrial processes use a wide variety of information to establish relationships between inputs, outputs, and costs. These relation- ships are presented for hypothetical facilities both with and without proposed regulatory controls. The practitioner uses the facility data to determine the costs of the regulatory alternatives. This general approach is the focus of this chapter. 4-3 ------- I I I I I I I I I I I I I I I I I I I 4.4 TYPES OF COSTS This section defines the major categories that can be used in assessing costs. The basic types of cost are capital costs, operating and maintenance costs, reinvestment, and salvage costs. The discussion is oriented toward the incremental costs of a regulatory action that affects water quality. In many situations, this orientation will mean upgrading an existing facility rather than constructing a new facility. Incremental costs may include both additional end-of-pipe treatment units and modification of existing production or treatment units. Most documents relating to treatment costs are concerned with the former because the set of unit treatment processes is fairly well defined, allowing cost estimation to be more standardized. Changes in production processes (or manufacturing) are more difficult to analyze and are not discussed. Capital Costs Capital costs (K) represent initial costs associated with the construction or upgrading of a facility to meet the treatment requirements plus periodic reinvestments as individual components wear out and must be replaced. Table 4-1 shows the kinds of costs that should be included in a capital cost estimate. They are divided into three main categories: component installed construction costs, noncomponent construction costs, and nonconstruction costs. The first category includes physical treatment units (e.g., activated carbon, chemically assisted clarification) and miscellaneous structures. The second category includes construction items not necessarily associated with individual structures, such as site preparation. The last category includes all the miscellaneous costs in addition to construction costs, including contrac- tor fees and interest payments. Care must be taken to identify incremental costs associated with the particular water quality decisions. For example, up- grades of existing facilities may not require any additional or miscellaneous structures. In addition to initial capital costs, replacement costs or reinvestment costs are required over the life of the project as individual pieces of equip- ment reach the end of their useful life. Although the line between replace- ment and repair can be a fuzzy one, the definition is quite operational. Items that are depreciated over a number of years rather than expensed imme- diately as costs are considered to be capital items and are treated as reinvest- ment. In most cases, no adjustment is required to use .engineering cost esti- mates as measures of social cost as long as they are based on market prices. This includes the proper allocation of interest and contractor's fees. Some caution is required in dealing with reinvestment because the available meas- ures are based on what tax laws allow and not on the actual social cost. Volume II will cover this in more detail. 4-4 ------- 1 1 Table 4-1 . Components of Capital and Operation and Maintenance Costs 1 1 1 A. Capital Costs (1) . Component installed construction costs unit processes miscellaneous structures (2) Noncomponent construction costs site preparation piping electrical instrumentation (3) Nbnconstruction costs land costs engineering and construction supervision contingencies .administrative and legal miscellaneous nonconstruction labor (testing, etc. design fees interest during construction B. Operation and Maintenance Costs . (1) Variable operating costs • labor materials chemical energy (2) Byproduct, other credits (3) Overhead items insurance taxes administrative and other allocations Operation and Maintenance Costs : Operation and maintenance (O&M) costs represent the annual running and maintaining the facility after its construction (see Table are divided into three groups: variable operating costs (labor, energy, etc.)/ byproduct and other credits, and overhead items. .) costs of 4-1) and materials, Control efforts resulting from in-plant process changes may also affect revenues or 1 1 1 production costs; these are included in the byproduct credits. 4-5 ------- I I I I I I I I I I I I I I I I I I I To be consistent with the definition of incremental costs, the overhead items should include only the additions required by the project, not a pro rata allocation based on overall costs. The treatment of taxes in a benefit- cost assessment raises some questions. Although taxes are not costs from the point of view of the nation as a whole, tax receipts that leave a particular State may be regarded by that State as a real cost in its benefit-cost assess- ment. State and local taxes should not be counted as social costs in any case; they are simply transfers. Annual O&M costs may vary over the life of the facility. For example, growth in population or water use will increase the flow to a POTW over time, causing O&M costs to increase. If the growth rate is high, using the first year's O&M costs may significantly underestimate average annual O&M costs over the life of the project. Reinvestment Costs Reinvestment costs (RC) represent periodic replacements of individual units whose lifetime is shorter than that of the overall project. ; Depending on data availability, it may be more convenient to represent this process as either the replacement of particular units at discrete intervals or as a con- stant fraction of initial investment costs each year. Table 4-2 shows the for- mula for each. If the first approach is used, capital costs may be broken down into groups with different average ages and the formula applied sepa- rately to each group using the lifetime applicable to that group. Salvage Value Salvage value (SV) is the market value of the facility at the end of the planning period. A wide range of values is possible depending both on what that alternative use is and on what assumption is made about reinvestment (see the previous subsection). One extreme case is that the facility is ex- pected to continue operating beyond the end of the planning period in the same fashion as before. In that case, the salvage value depends on the ini- tial investment cost and the remaining useful life after the planning period. If the reinvestment process is best characterized as a series of periodic rein- vestments, the value of the facility is proportional to the ratio of its useful life at the end of the planning period and its total useful life. - If the reinvestment process is better represented as an average reinvest- ment of amount dK each year (see Table 4-2), the expected lifetime of the facility has no end as long as the annual reinvestments are made. Therefore, the salvage value at the end of the planning period is still K if the facility is expected to continue in the same use as before. If the facility is not expected to continue in its present use after the planning period, its scrap value must be determined. Any permanent fixtures 4-6 ------- 1 1 1 1 1 1 1 • 1 1 1 1 1 • Table 4-2. Variables and Definitions for Measuring Costs 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. Variable Planning period Real social discount rate Investment costs O&M costs Investment lifetime Physical depreciation rate Number of replacements in planning period Growth rate of O&M costs Reinvestment costs Fraction of salvageable value Salvage value Present value of stream of payments for N years at discount rate s Present value of rein- vestment costs Present value of salvage Present value of O&M costs Present value of all capital costs Present value of all project costs Total annual cost of project aFirst definition in the right-hand ments, where RC = K; the second Symbol N s K OM L d M 9 RC q SV PV(N,s PVRC PVSV PVOM PVK TPV TAG Source Parameters Parameters Project costs Project costs Project costs Project costs Greatest integer N/L Parameter K every L years; or dK every year Project costs qK ) (1-(1+s)"N)/s RC I (1+s)~J ; or j=1 RC-PV(N,s) SV(1+s)~N OM-PV(N,s) K + PVRC - PVSV PVK + PVOM TPV/PV(N,s) column represents periodic reinvest- definition represents annual reinvest- ments, where RC = dK. 1 1 4-7 ------- I I I I I I I I I I I I I I I I I I I such as concrete tanks or structures are likely to have ho salvage value.* Machinery and equipment items may have 25 to 50 percent of their original value. Cash or natural resources on hand can be valued at 100 percent of current value, and other current assets may .be valued at 70 to 100 percent of their current value, depending on their salability. Summary Cost Measures Discounting allows the four types of costs to be combined to obtain a single overall cost estimate for a project. Based on the discussion given in Chapter 2, the use of the discount rate is summarized below. Specifically/ discount rates are used to construct two kinds of summary measures: pres- ent value (PV), where O&M costs are capitalized, and total annual costs (TAG), where capital costs are annualized. Table 4-2 shows these two methods along with the variables used to develop measures of social cost. In addition, Table 4-2 summarizes the formulas for the components of the total present value of the costs of a project. The present value of all capital and O&M costs (TPV) is obtained by adding the present values of the cost components (salvage value is sub- tracted ): TPV = K + PVRC - PVSV + PVOM (4.1) The total annual cost (TAG) of the project is a constant amount whose present discounted sum over the project period is equal to the present value of project costs TPV. By the definition of the present value factor (PV) given in Table 4-2, TAG = TPV/PV(N,s) 4.5 PRACTICAL ISSUES IN MEASURING COSTS (4.2) This section discusses the practical aspects of measuring costs. It covers sources of data, the use of cost indexes, and major factors affecting cost estimates. A sample data form, one method of organizing the costing process, concludes the section. Sources of Cost Data There are three commonly used sources of cost data: vendor informa- tion, estimating manuals, and industry information. In the past, many prac- *Firms can write off such assets as tax losses. However, these writeoffs are not appropriate for measuring social costs because society still bears the full cost of the resources. The distribution of who in society bears the cost is different. Taxpayers and the firm bear the costs when they are written off, and the firm and consumers of its product bear the cost when they are not written off. 4-8 ------- I I I I I I I I I I I I I I I I I I I titioners have been successful in obtaining cost estimates for both installed capital costs and annual operating costs from equipment vendors. These esti- mates are generally solicited in writing along with operating features of the model plants for which cost estimates are needed. Estimating manuals have also been useful in developing cost estimates. Two popular manuals are published by Richardson Engineering Services, Inc. [1977] -and R. S. Means Company, Inc. [1981]. Richardson's manual is useful in preparing cost estimates for process industries, and Means' manual is geared to the construction industry. Both manuals are written to provide costs for individual components rather than total systems and thus require some skill and care in their use. Estimates can also be obtained by the use of cost factors. In this procedure, major equipment costs are multiplied by appropriate factors to estimate other cost elements. The factors are derived from experience with previous plant construction costs. Some procedures use a single factor to estimate total capital investment, but greater accuracy can be achieved from a method such as Guthrie's [1974], which separates labor and material costs and applies individual factors to each major process item. Information supplied by the controlled industry may be useful in estimat- ing costs. If plants can be identified that are already controlled to the level under investigation, both total installed costs and annual operating costs can be obtained from this source. Obtaining estimates of each cost element from more than one source is a way to validate estimates. The Use of Cost Indexes Treatment cost indexes allow cost estimates from different years to be converted to dollars of a single year to yield a valid comparison. Costs of various components included in both capital and O&M costs change over time due both to overall changes in the price level (inflation) and to changes in relative prices. The problem of forecasting general price inflation can be avoided by expressing all costs in constant dollars of a given year, but changes in relative prices still need to be predicted. Unless there are com- pelling reasons to do otherwise, it is simplest and reasonably safe to assume constant real costs in future years. Of course, this assumption must be applied to all components of the benefit-cost assessment to ensure consist- ency. Cost indexes are usually represented as a number showing the ratio of the cost of a unit in dollars of a given year to the cost of the same unit in dollars of the base year multiplied by 100. In adjusting these costs to con- stant dollars in a given year, the practitioner must take account of the change in the index and the overall change in the price level. For example, if the cost of a treatment plant is given for 1975 in current (i.e., 1975) dollars and must be converted to 1977 costs in 1977 dollars, the U.S. Environ- mental Protection Agency's (EPA) Sewage Treatment Plant Construction Cost (STPCC) Index [Michel] can be used as follows: 4-9 ------- I I I I I I I I I I I I I I I I I I I .Cost (77) = Cost (75) = Cost (75) x STPCC (77) STPCC (75) 278.3 250.0 = Cost (75) x 1.11 . If the costs for a given year are to be expressed in the constant dollars of another year, then the gross national product (GNP) deflator or another general price index must be used. For example, costs shown above for 1977 may need to be converted to 1976 current dollars. Using the convention that PGNP (N) is the GNP deflator* in year N and that the notation cost (m, n) refers to costs corresponding to year m expressed in constant dollars of year n gives: Cost (77, 76) = Cost (77, 77) = Cost (77, 77) PGNP (76) PGNP (77) 1.321 1.398 = Cost (77, 77) x 0.945 . Cost indexes are available for both capital and O&M costs. One index to use for O&M costs is the U.S. Environmental Protection Agency's (EPA) Municipal Wastewater Treatment Plant Operation and Maintenance Cost Index [Michel]. This index is a weighted average of cost indexes for labor, chemi- cals, power, maintenance, other costs, and a "quality added" factor. Al- though developed primarily for secondary treatment plants, the mix of inputs for O&M costs of advanced treatment plants should not differ much. Several construction cost indexes are available for adjusting capital costs. These include EPA's STPCC index used earlier, the Engineering New-Record Cost Index (ENR),t and the Chemical Engineering Plant Con- struction Cost Index (CE).f The nature of the treatment system being exa- mined determines which of these indexes is most suitable to use. 'Both the STPCC and ENR indexes are more oriented to secondary treatment plants, .where large concrete tanks play an important role, but they are also appro- *The GNP deflator is published by the U.S. Bureau of Economic Analysis. The wholesale price index (WPI), another useful index, appears in the U.S. Bureau of Labor Statistics, Producer^ Pnces and Price Indexes. tAppears weekly in the Engineering News-Record, published by McGraw Hill. ^Appears in Chemical Engineering, published by McGraw Hill. 4-10 ------- I I I I I I I I I I I I I I I I I I I priate for some advanced treatment (AT) processes (such as chemically assisted clarification), which rely on large tanks. The CE index is more suitable for processes where equipment plays a large role in costs, such as use of activated carbon. However, it is difficult to make hard and fast rules. The cumulative percent increases of the STPCC and ENR indexes are close, while that of the GE index is somewhat lower. To arrive at a cost figure more accurate than one that results from the use of the national average indexes alone, locality factors can be applied to an estimated cost or cost index. The use of locality factors, which have been calculated from generally available statistics, permits the localizing of national average cost data on various cost items. Locality factors are available from the U.S. Environmental Protection Agency's (EPA) Sewage Treatment Plant and Sewer Construction Cost Index [Michel]. Major Factors Affecting Cost Estimates The cost of a treatment process is influenced by a multitude of factors. Among the important ones are wastewater flow rate, pollutant loadings, plant location, and performance (i.e., amount of pollutant removed and effluent concentrations). Most treatment technologies show economies of scale; i.e., costs increase at a slower rate than flow size. Thus, the cost per gallon of wastewater treated in a large plant is less than that treated in a smaller plant that meets the same treatment performance criteria. In general, the scaling factor* varies with size of flow and technology. For preliminary planning- purposes, the following values are- useful for extrapolating the treatment cost of a given treatment plant size to others: Cost Item Scaling Factor Capital 0.6 to 0.9 O&M 0.7 to 0.9 Labor 0.5 to 0.7 Utilities and 1.0 chemicals The upper end of the range for the scaling factors (implying less economy of scale than the lower end) is associated with the more advanced treatment tech- nologies, such as carbon adsorption and electrolysis, to which this handbook is oriented. Economies of scale- also appear in the relationship of cost to waste loading but are less prevalent than in the cost-to-flow-size relationship. *As used here, scaling factor refers to x In the following equation: where Cj is a cost for a treatment plant with flow Qj and C2 is the estimated cost for a plant with flow Q2. 4-77 ------- I I I I I I I I I I I I I I I I I I I Because wastewater flows and loadings tend to show greater variability in a POTW than they do in an industrial plant, equalization is commonly used to smooth out fluctuations. Even so, the POTW customarily is designed with some excess capacity to accommodate variations in flow and waste loading. In some cases, more intensive treatment effort in the treatment plant— such as additional chemicals or energy input—can result in enhanced perform- ance or accommodation of a greater volume of wastewater without sacrificing design performance. However, the incremental cost usually is high. There- fore, .it is less costly in the long run to upgrade or expand the treatment facility if actual flow, loading, or performance is expected to change appre- ciably from initial design conditions. It is generally more costly to upgrade an existing plant by retrofitting than it is to use the same treatment train in a new plant. The treatment in place at a plant may limit the choice of higher treatment technologies that can be selected to upgrade the plant. Upgrading usually requires engineering effort, rewiring, and additional piping at the existing facility. The additional cost attributed to retrofitting a plant is sensitive to the specific features of the plant and the site, so these factors should be considered in developing retrofit .cost factors. Based on limited observations, costs of retrofitting an existing plant can range from 1 .to 15 percent higher than the cost of incor- porating the same treatment train in the initial design of a new plant [U.S. EPA, 1976]. The added cost of retrofit in percentage terms is inversely pro- portional to the capacity of the treatment facility. Sample Data Form A "model plant form" can be used to compile parametric and cost esti- mates in an orderly fashion. The form can be continually revised to meet the needs of the specific data requirements of an assessment. A model plant questionnaire of this type is applicable only for regulatory alternatives that involve engineering controls. Figure 4-2 presents a version of the model plant questionnaire. 4.6 COST ESTIMATING TECHNIQUES This section on cost estimating techniques defines the components of treatment systems and describes several variations on the basic costing tech- niques. Finally, the question of how to proceed when no directly relevant cost sources are available is discussed. The techniques described here are appropriate for costing both POTWs and industrial dischargers. Components of Treatment Systems To analyze the costs of a particular proposed treatment system, it is useful to break it down into its components. Although each treatment system has its unique aspects, the individual components are more standardized and hence more easily costed using standard references, where costs are defined in terms of a few major parameters. The detail of the breakdown depends on the accuracy of the estimate required. It is useful to understand the relation- ships of the following three levels of treatment units: 4-12 ------- I I I I I I I I I I I I I I I I I I I MODEL PLANT FORM 1. Model Ham Number and Description Number Description 2. If new, ran design tif* (yeari) If existing, itata plant age (yean) and remaining life (yean) 3. Legal Depreciation Period (yean) and Usual Depreciation Method Plant Compliance Equipment 4. Discuss what relationship exirti, if any, between existing plant age and model type. Discuss age distribution of exist- ing plants. Alternative 1 (Baseline) Alternative 2 Alternative n 5. Product Names, Annual Capacity Outputs Primary Product Name Annual Physical Output Capacity Units Specified Producer Price, 19_ per unit .dollars Secondary Product Name Annual Physical Output Capacity Units Specif ied Producer Price, 19 dollars per unit Marketable Byproduct Name Annual Physical Output Capacity Units Specified 8. Annual In-Plant Consumption of any of the Products Lilted in (5L Name Quantity Name ' Quantity 7. Installed Capital Cost 19 dollars OF NEW PLANT 8. Installed Capital Cost of Compliance Equipment for Existing Plant 19 dollars XXX Figure 4-2. Sample model plant data form. 4-13 ------- I I I I I I I I I I I I I I I I I I I Model Plant Form (con.) Alternative 1 (Baseline) Alternative 2 Alternative n 9, Expected Life of Compliance Equipment for Existing Plant (years) 10. Lott Productive Time While Compliance Equipment It Installed in Existing Plant (days) 11. Salvage Value, if any, of Compliance Equipment When Plant Is Closed 19 dollars 12. Total Annual Operating Costs 19 dollars 13. Annual Fixed Operating Costs Total Fixed Operating Cost 19 dollars Labor Number of labor hours Annual Wage per labor hour 18 dollars Labor Cost 19 dollars Energy Costs 19 dollars Materials Costs 19 dollars Overhead Costs 19 dollars Other Fixed Costs 19 dollars 14. Annual Variable Operating Costs Total Variable Operating Costs 19 dollars Labor Number of labor hours Annual Wage per labor hour 19 dollars Labor Cost 19 dollars Energy Costs 19 dollars Materials Costs 19 dollars Overhead Costs 19 dollars Other Variable Costs 19 dollars 15. Current Salvage Value of Plant 19 dollars XXX XXX XXX 16. Salvage Value of Plant at End of Life 19__ dollars 17. Residuals Discharged to Environment To Atmosphere Name '. Quantity To Water Name _ Quantity To Land Name Quantity Figure 4-2. (con.) 4-14 ------- I I I I I I I I I I I I I I I I I I I Unit process: A unit process corresponds to a single treat- ment operation. Examples of unit processes are secondary clarification, filtration, and cyanide destruction. Treatment process: A treatment process is a sequence of one or more unit processes linked together to support a particular pollutant-removal process. For example, the activated sludge process involves the decomposition of organic pollutants by microorganisms. This operation requires a number of unit proc- esses including aeration, sedimentation, and sludge reactiva- tion. • . Treatment train: A treatment train is a sequence of treatment processes. For example, an advanced treatment train may con- sist of the following treatment processes: preliminary screen- ing, primary settling treatment, secondary biological treatment, and nutrient removal by chemical addition. As discussed below, cost estimates for benefit-cost assessments of water qual- ity programs should be broken down at least to the treatment process level and, in some cases, to the unit process level. Estimating Treatment Costs Cost estimation requires a specification .of the treatment train . to be used. A waste treatment train can be described by a flow diagram showing the relation and function of the various treatment and unit processes. One way to proceed is to (1) specify the .important design parameters (such as flow, influent and effluent concentrations, and contact time) of each unit in the treatment process, (2) calculate the resource requirements (for example, for site preparation and construction and for purchased equipment, energy, and labor) of each unit, (3) estimate the indirect costs, and (4) sum to ob- tain a total cost. This calculation is carried out separately for a total capital and a total O&M cost before they are combined into a present value or annual - ized cost. Before illustrating how different sources of cost information can be applied in particular situations, it is useful to distinguish among four differ- ent approaches that are employed to make cost estimates: A total system estimate A planning level estimate An engineering estimate A contractor estimate. The cost estimates produced by these techniques range from gross to refined, depending on the different stages of a project, which range from project con- ceptualization to request for contractor bids. The four costing techniques 4-15 U.S. EPA Headquarters Library Mail code 3201 1200 Pennsylvania Avenue NW Washington DC 20460 ------- I I I I I I I I I I I I I I I I I I I provide a convenient frame of reference for discussion purposes .but do not imply that rigid distinctions can always be made in using one technique versus another in a specific situation. The use of site-specific information should result in more accurate results than use of generalized information. While it is, of course, desirable to develop accurate estimates, the practitioner is always confronted with decid- ing how much accuracy is needed for a particular phase of project planning and what level of effort to commit to the development of the estimates. For the purposes of water quality standards planning, the planning level estimate is generally the most appropriate one to use, and the examples focus on that level. Total System Estimate. In contrast to the other three costing techniques, the total system approach does not attempt to par- tition a treatment train into treatment or unit processes. Usually only one parameter of the treatment system—for example, plant capacity expressed as daily flow—is used with a set of cost curves to obtain total capital and annual O&M costs. The accuracy is ±40 percent. Planning Level Estimate. This approach is based on prior analyses of treatment system components or unit processes in which costs of the units have been related to important design parameters. The purpose is to allow recombination or syn- thesis of total costs resulting from any combination of the unit processes using specified values ,of the design parameters. This level of estimate is appropriate for most water quality pro- gram planning purposes. Application of this technique re- quires that the practitioner identify the major components in the wastewater treatment train, the associated design param- eter values, and the access to generalized cost functions for the components. The accuracy of this costing technique is within ±30 percent. The practitioner may be able to improve the accuracy of the estimate if judgments can be made about how site-specific characteristics differ from average conditions embodied in the generalized cost functions, Engineering Estimate. Like the planning level estimate, the engineering estimate is calculated using unit process data but goes into more detail on the unit processes in the system to adjust specific costs. This technique should yield a cost esti- mate within ±15 percent. Contractor Estimate. The contractor estimate is based on specific engineering designs—or design approaches coupled with specified performance requirements--for the treatment system and it's unit processes. The precision of the contrac- torfs estimate should be within ±5 percent. 4-76 ------- I I I I I I I I I I I I I I I I I I I 4.7 EXAMPLES USING COST TECHNIQUES Detailed engineering cost estimates are unlikely to be available for analy- ses of the type described in this handbook. Therefore, this section explores a range of techniques, from fairly sophisticated costing models (CAPDET) to process handbooks and EPA development documents, and presents specific examples to show how they are applied. Their applicability depends on both the amount of information and the t'ime and resources available to the analyst. States may have their own techniques that are equally applicable. Even if specific information is not available, there are basic similarities in the ap- proaches to water treatment problems taken by different industries. There- fore,' the experience of other industries with similar processes or pollutants may provide a reasonable guide for estimating costs. A library of cost information* is available to the practitioner, including documents and computer programs. The purpose of the following examples is to illustrate the use of various information sources to estimate costs, particu- larly planning level estimates. JExampjeijiii1_:__rfiPijlp__M}|.| Using EPA peyelojpmejTt^Dqcument This example illustrates the use of an EPA Development Document [U.S. EPA, 1980b] for a specific industry and the planning level costing approach. Assume that more stringent effluent standards will be imposed on a 900-ton- per-day kraft pulp mill to meet water quality standards. Based on a waste- load allocation, the effluent concentrations required to meet the water quality standards are 15 mg/L for both biochemical oxygen demand (BOD) and total suspended solids (TSS). The costs are to be estimated for October 1981.t The practitioner determines that the mill belongs to the "Market Bleached Kraft" subcategory (described as one of the industry subcategories in the Development Document) and that concentrations of BOD and TSS in the plant's effluent stream currently are 20 mg/L and 30 mg/L, respectively, which meet Best Practicable Technology (BPT) standards.^ The Development Document identifies three treatment options that might possibly be used to meet the new BOD and TSS targets, but only Option 3, which consists of additional in-plant process controls together with chemically assisted clarifica- tion of the final effluent, can achieve the BOD and TSS targets of the water quality standards. *lmportant sources include: (1) EPA Development Documents for effluent limitations guidelines and standards (issued by the Effluent Guidelines Divi- sion of EPA to provide the technical background for the development of waste treatment rules for particular industries); (2) Areawide Assessment Proced- ures Manual [U.S. EPA, 1976]; (3) Innovative and Alternative Assessment Manual [U.S. EPA, 1980a]. fThe examples in this guidance use historical values of treatment cost indexes. For planning purposes it may be necessary to estimate future values. standards for the pulp, paper and paperboard industry have not yet been promulgated. 4-17 ------- I I I I I I I I I I I I I I I I I I I The next step is to estimate the cost for Option 3. Table 4-3 shows the costs of meeting Option 3 for three model mills with sizes of 350, 600, and 1,600 tons per day. Note that energy costs are presented separately, so they must be combined with O&M costs to obtain the definition of O&M costs used here. Two costs are shown for each category of cost based on alum concentration. As a first approximation, an average cost for the two alum concentrations is used. Next, because none of the three model mill sizes is Table 4-3. Example 1: Cost Summary— Market Bleached Kraft, Subcategory Mill size (tons/day) Capital cost 350 600 1,600 Annual O&M cost 350 600 1,600 Annual energy cost 350 600 1,600 Incremental compliance costs . from BPT: Option 3 ' 6,662 8,974 9,446 , 16,590 17,410 947 1,327 1,327 1,953 2,974 4,550 212 517 351 358 897 917 SOURCE: U.S. EPA [1980b], p. 468, Table IX-7. aFirst quarter 1978 thousands of dollars. Dollar value shown above the line is based on chemical assisted clarifica- tion dosage of alum at 150 mg/L; the value below the line is for dosage at 300 mg/L. 4-18 ------- I I mm . I I I I I I I I I I I I I I I I for a 900-ton-per-day plant, a linear interpolation based on mill size is made between the values for the 600- and 1,600-ton-per-day plants.* This yields a capital cost for Option 3 of $11.5 million. Annual O&M cost is $2.3 million, and energy cost is $0.5 million; these are added for a total O&M cost of $2.8 million. The following steps show the results for various measures of social cost derived from the basic costs, which are used as part of the benefit-cost .assessment. The following parameter values are used: ENR index (1st Qtr 1978) =2,683 ENR index (Oct 1981) = 3,679 EPA O&M index (1st Qtr 1978) = 2.30 EPA O&M index (Oct 1981) = 3.34 s = real social discount rate = 0.10 or O.OSf L = average equipment lifetime = 15 d = depreciation rate = 1/L = 1/15 N - planning .period = 30 years The indexes are historical values, the other parameters are assumed values. The results are (all quantities in thousands of 1981 dollars unless otherwise noted): 1. Capital treatment costs: K = 11,500 (1st Qtr 1978 $) K = 11,500 xfigf K = 15,800 (Oct 1981 $) 2. O&M treatment costs: OM = 2,800 (1st Qtr 1978 $) *Another method for estimating the value for the 900-ton-per-day plant size is to fit a cost curve to the three data points to determine if there are economies of scale of treatment costs instead of assuming a linear relationship between mill size and cost for the two larger model mills. TThese are just sample values chosen to show the effects of different values on the results. See Chapter 2 for discussion of which discount rate to use. 4-19 ------- 1 1 ^^B I • 1 1 1 1 • • 1 1 1 1 1 1 1 1 3 34 OM = 2,800 x |^|2 2.30 OM = 4,100 (Oct 1981 $) 3. Present value factor (equals present value of payment of one dollar per year for N years at discount rate s): PV(30, 0.10) = 9.43 PV(30, 0.05) = 15.37 4. Present value of O&M costs: PVOM = OM • PV(30, 0.10) = 4,100 • 9.43 - OM • PV(30, 0.05) = 4,100 • 15.37 = 38,700 = 63,000 5. Reinvestment cost (assume the entire facility is .replaced in 15 years at the same original real cost):* RC = K = 15,800 6. The present value of the reinvestment cost is found by discounting at rate s over 15 years: PVRC = RC • (1+s)"L = 15,800 • <1.1)"15 = = 15,800 • (1.05)"15 7. Salvage value (assume zero salvage value) SV = 0 8, Present value of salvage (salvage occurs N): PVSV = SV (1+s)~N = 0 9. Present value of salvage (salvage occurs N): * *Not all structures would need to be replaced to represent the average lifetime of the facility. 4-20 = 3.800 (s = 0.10) or = 7,600 (s = 0.05) * at end of planning period, at end of planning period, . Fifteen years is assumed ------- I I I I I I I I I I I I I I I I I I I PVSV = SV (1 + s)"N = 0 10. Total present value of investment cost: PVK - K + PVRC - PVSV = 15,800 + 3,800-0 = 19,600(5 = = 15,800 + 7,600 - 0 = 23,400 (s = 0.05) 11. Total present value of project cost: TPV = PVK + PVOM = 19,600 + 38,700 = 58,300 (s = 0.10) = 23,400 + 63,300 = 86,400(s = 0.05) 12. Annualized capital costs: KANN = PVK/PV = 19,600/9.43 « 2,080 (s = 0.10) = 23,400/15.37 = 1,520(5 = 0.05) 13. Total annualized costs: TAG = KANN + OM = 2,080 + 4,100 = 6,180 (s = 0.10) = 1,520 + 4,100 = 5,620 (s = 0.05) The total present value of project costs calculated in Item 11 above is the amount that would be included in the benefit-cost assessment. Alternatively, the total annualized cost (Item 13 above) would be used if the rest of the assessment were also expressed in terms of annualized costs. Example 2: POTW Using CAPDET Over the past decade a number of computer-based treatment cost esti- mating models have been developed; CAPDET is one of the more widely used models [U.S. EPA, 1981].* CAPDET is not a mathematical optimization model; the CAPDET approach is to prepare cost estimates for alternative treatment trains specified by the user. With reference to the four costing techniques discussed earlier, the CAPDET method is probably best described as an inter- mediate method between the planning level approach and the engineering esti- *To obtain access to CAPDET programs and documentation, contact the Systems Analysis Group of EPA regional offices or the Facilities and Require- ments Division, Office of Water Program Operations, EPA in Washington, D.C. 4-21 ------- I I I I I I I I I I I I I I I I I I I mate approach. CAPDET contains cost and performance equations for-35 unit processes such as activated sludge, carbon absorption, lagoons, incineration, etc. For some unit processes, alternative design approaches are included; for example, the activated sludge process is described by 13 different meth- ods of aeration. Two separate cost estimating methods are incorporated in CAPDET. First, parametric cost estimating is based on statistical analysis of the cost of facilities of similar size and characteristics at other locations. Second, unit cost estimating is based on identification of cost elements to which input prices are applied—e.g., cubic yards of concrete in a clarifier are quantified and an input cost value for reinforced concrete is applied to obtain a construction cost. After recent revisions, CAPDET can be used to estimate costs of upgrad- ing an existing POTW even though the program was developed originally to estimate new plant costs. Its capabilities have also been expanded to provide estimates of financial impacts on households from the construction of a POTW facility. In this example, an existing municipal POTW has an average daily flow of 5 Mgal/d, a maximum flow of 10 Mgal/d, and effluent concentrations of 15 and 20 mg/L for BOD and TSS, respectively. Current secondary treatment uses plug flow for the activated sludge effluent, and the practitioner needs to determine the cost of upgrading the treatment by chemical addition and filtra- tion to meet the water quality standard that requires reductions of BOD to 10 mg/L and TSS to 1 mg/L or less. To use the CAPDET computer program, the unit processes in the current plug flow system and the upgraded treatment are specified as shown in Table 4-4. The CAPDET program can accept any number of wastewater treatment trains that are suitably described by the user. Table 4-4, Example 2: Current and Upgraded Treatment Trains for CAPDET Current treatment sequence Upgraded treatment sequence Raw sewage Preliminary treatment Primary clarifier Plug flow Secondary clarifier Chlorination Gravity thickener Anaerobic digestion Vacuum filtration Hauling and land fill Raw sewage Preliminary treatment Primary clarifier Plug flow Secondary clarifier Coagulation Filtration Chlorination Gravity thickener Anaerobic digestion Vacuum filtration Hauling and land fill 4-22 ------- I I I I I I I I I I I I I I I I I I I Table 4-5. Example 2: Price and Cost Inputs to CAPDET Cost analysis input parameters Interest rate: 10.000 percent Planning period: 30 years Year of dollars used: 1980 Cost Indexes Buildings Excavation Wall concrete Slab concrete Marshall and Swift Index Crane rental EPA Construction Cost Index Canopy roof Labor rate Operator class 11 Electricity Chemical costs Lime Alum Iron salts Polymer Engineering News Record Cost Index Handrail Pipe Cost Index Pipe installation labor rate 8-in. pipe 8-in. pipe bend 8-in. pipe tee 8-in. pipe valve Unit prices 55.00 $/ft2 7.00 $/yd3 207.00 $/yd3 91.00 $/yd3 577.00 67.00 $/h 163.00 15.75 $/ft2 13.40 $/h 9,00 $/h 0.04 $/kWh 0.03 $/lb 0.04 $/lb 0.06 $/lb 1.62 $/lb 2,886.00 25.20 $/ft 295.20 14.70 $/h 9.08 $/ft 86.82 $/unit 128.49 $/unit 1,346.16 $/unit 4-23 ------- I I I I I I I I I I I I I I I I I I I Table 4-5 shows unit costs and prices in 1980 dollars that the practi- tioner provides to CAPDET, as well as the planning period and interest rate to be used in computing project costs. In calculating the capital costs, CAPDET recognizes that each equipment item has a service life associated with it that may be less than the planning period and that the facility has some salvage value at the end of the planning period. As a result, the capita! cost incorporates three components, initial and reinvestment costs and salvage value. Typical useful life periods are as follows: wastewater conveyance structures, 50 years; other structures, 30 to 50 years; process equipment, 15 to 20 years; and auxiliary equipment, 10 to 15 years. In addition, O&M costs can be specified to vary over the planning period. Table 4-6 summarizes the costs of the two systems analyzed by CAPDET, a new treatment plant using the current treatment sequence and a new plant with the two new treatment steps included. The table also shows the unad- justed cost increments obtained from CAPDET for upgrading the POTW and the adjusted costs which are increased by 15 percent to reflect the costs of retrofitting an existing system. The capital costs of upgrading (as shown in Table 4-6) are $1.01 million; total project costs are $1.81 million. Additional O&M costs (which vary be- tween the first and final year) range from $370,000 to $200,000 annually. Present worth is.$5.46 million; annualized cost is $580,000 based on a 30-year planning period and 10 percent rate of interest. Table 4-6. Example 2: POTW Upgrading Cost Summary (millions of 1980 dollars) Incremental New New POTW with POTW with plug flow costs for upgraded POTW plug flow and chemical ** secondary addition and CAPDET Adjusted for treatment filtration unadjusted retrofit Capital cost Total construction 3.60 4.40 4,48 5.46 0.88 1.06 1.01 1.22 cost Total project cost O&M cost 6.53 8.10 1.57 1.81 First year Final year Present worth 0.31 0.46 9.98 0.63 0.63 14.76 0.32 0.17 4.78 0.37 0,20 5.46 Adjusted costs are 15 percent higher than unadjusted costs. 4-24 ------- I I I I I I I I I I I I I I I I I I I Example 3: POTW Using Technology Assessment Manual In this example, planning level estimates of costs for upgrading an exist- ing facility are developed from generalized cost curves that have been devel- oped [U.S. in the 'EPA's Innovative and Alternative Technology Assessment Manual EPA, 1980a]. The manual has been designed specifically to aid Federal and State review authorities in the administration of innovative and alternative requirements of the Construction Grants Program and to provide basic method- ological and technical information to individuals involved in facility plan de- velopment. A municipality is assumed to have secondary wastewater treatment in place at the POTW using the activated sludge process. Wastewater flow is 10 Mgal/d, and the effluent has a monthly average value of 30 mg/L for both BOD 'and TSS. To meet an ultimate oxygen demand (UOD)* of 85 mg/L as part of waste-load allocation, it has been determined that the advanced waste treatment process described in the Manual as "nitrification, separate stage, with clarifier" is required.. The construction and O&M costs are obtained from Fact Sheet 2.1.14 in the Manual using a 10-Mgal/d flow .rate. Construc- tion costs are $1,6 million and O&M costs are $70,000 in 1976 dollars. Several other adjustments to the values read from the cost curves are necessary. The referenced Fact Sheet is based on an ENR construction cost index of 2,475 (for September 1976). Adjusting the construction costs for a first quarter 1980 ENR index of 2,886 yields a construction cost of $1.9 mil- lion. The O&M costs must also be adjusted to first quarter 1980. The EPA O&M index for the third quarter of 1976 is 2.06 and for the first quarter of 1980 is 2.83, yielding an adjustment factor of 1.37. Therefore, O&M costs are $100,000. The Fact Sheet in the Manual directs the user to estimate other capital expenditures that have not been included in the construction cost curves. Table 4-7 replicates Table A-2, which is provided for that purpose in the Manual. Following the directions incorporated with Table A-2, the practi- tioner estimates nonconstruction capital expenditures. As shown in Table 4-7, total capital cost is $3.5 million. The costs of the system are summarized in Table 4-8 together with efflu- ent information. The annualized costs in the table are based on a capital recovery factor of 0.106, reflecting a 10 percent social discount rate and a 30-year plant life. Example 4: What To Do When There Is "No Information" The practitioner may find that no Development Document or other techni- cal data have been published that are directly applicable to a particular type of industrial plant. Nevertheless, it is possible to identify treatment options and develop preliminary cost estimates, suitable for at least the early phases of water quality program analysis. *UOD + BOD5 x 1.5 + NH3 x 4.5. 4-25 ------- I I I I I I I I I I I I I I I I I I I Table 4-7. Example 3: Development of Capital Costs (Nitrification) (All costs in millions of 1st quarter 1980 dollars) Component installed construction costs Unit processes Nitrification Miscellaneous structures (Administrative offices, laboratories, shops and garage facilities) Subtotal 1 $1.9 $0.0 $1.9 Noncomponent costs Piping Electrical Instrumentation Site preparation Subtotal 2 Nonconstruction costs Engineering and construction supervision @ 15 percent. Contingencies @ 15 percent Subtotal 3 Total capital costs Average 10% •8% 5% 5% Range 8-15% 5-12% 3-10% 1-10% $0.2 0.4 0.1 0.1 $0.4 $0.4 $0.8 $0.8 $3.5 Range due to level of complexity, degree of instrumentation, subsoil con- dition, configuration of site, etc., percentage of subtotal 1. ^Percentage of subtotal 1 plus 2. Table 4-8. Example 3: Summary of Costs for Nitrification Upgrade Costs (millions of 1980:1 dollars) Construe- Total tion capital O&M Annual ized costs . costs costs • cost ($) ($) ($) ($) Effluent (mg/L) BOD UOD* 1.9 3.5 0.10 0.46 10 19.5 UOD = BOD5 x 1.5 + NH3 x 4.5. 4-26 ------- I I I I I I I I I I I I I I I I I I I A reasonable approach is to first identify the pollutants that are of major concern at the particular plant, their concentration, and the effluent flow. The next step is to identify waste treatment unit processes that may be appli- cable for the identified pollutants. This can be done by identifying one or more analogous industries--e.g., an industrial plant where the specific pollut- ants of interest are being effectively treated, or have been studied. In all likelihood, the method of treatment of a specific pollutant in one industry will be applicable to the particular plant of interest to the practitioner. The next step is to acquire Development Documents and other reports on the analogous industries. Using the information on the analogous plants, the practitioner should search out the cost-versus-flow-size relationships for the one or more unit processes used to treat the pollutants of interest. This will allow the practitioner to synthesize a wastewater treatment train and estimate a total cost for the process units in the train. This approach is preferable to using a total system cost estimate, in which the costs for each unit process are not explicitly identified. 4.8 FINAL CAUTIONS This section describes several important factors that suggest caution is necessary when measuring waste treatment costs, particularly industrial wastes. The costs of treating wastes are only one element of the entire produc- tion process for a firm. This element is the management of .the residuals that accompany production (see Ayres and Kneese [1969]). The costs of treating wastes can be affected by changes in the level of production, which alters the volume of wastes and perhaps the type of treatment required. Changes in the type of product produced, in the processes used to make it, or in the means of recovering the waste all can affect the cost and type of treatment required. For example, a technology change in pulp production for the paper industry from sulfite to kraft lowered the volume of suspended solids. When the practitioner considers the cost of a regulatory action, it is important to keep this total system view in mind, 4.9 SUMMARY This section summarizes the major points from the chapter. Costs are measured on the basis of opportunity cost—the value of the next best forgone alternative. The two main approaches to estimating costs are the economet- ric and engineering approach. The engineering approach is most often used because data needed in the econometric ap- proach are seldom available. Two major cost categories are capital costs and operation and maintenance (O&M) costs. Capital costs are initial costs of 4-27 ------- I I I I I I I I I I I I I I I I I I I construction or upgrading. Operation and maintenance costs are annual costs of running and maintaining a facility after its construction. Three useful sources of cost data are vendors, estimating manuals, and industry information. Discounting links the four major components of project costs: initial investment, Q&M, reinvestment, and salvage value. Cost indexes allow cost estimates from different years to be compared on a common basis. Useful index are EPA's Sewage Treatment Plant Construction Index, the GNP price deflator series, and Chemical Engineering Plant Construction Cost Index. Important factors affecting costs of waste treatment are flow rate, pollutant loadings, plant location, and performance cap- ability. A model plant data form .can organize information .required for costing. Three levels of waste treatment are unit process (single treat- ment), treatment process (sequence of unit processes), and treatment train (sequence of treatment processes). Planning level estimates of costs based on prior analyses of treatment cost and having an accuracy of ±30 percent are ap- propriate for many water quality program decisions. EPA Development Documents provide a valuable .source of planning level costs for specific industries. CAPDET is a computerized model that prepares cost estimates of alternative treatment trains, estimates the costs of up- grading, and computes a financial impact statement for publicly owned treatment works. CAPDET provides accuracy between ±15 and ±30 percent. Technology assessment manuals provide basic technologies for publicly owned treatment works. Treatment cost estimates are sensitive to many factors in firms' overall production operation, including output levels, types of products, or manufacturing processes. 4-28 ------- I I I I I I I I I I I I I I I I I I I CHAPTER 5 COMPLETING THE BENEFIT-COST ASSESSMENT 5.1 INTRODUCTION What is a sensitivity analysis? How is it used in a benefit-cost assess- ment? What methods are appropriate for displaying the results of a benefit- cost assessment? Is a checklist of elements possible in a benefit-cost assess- ment? In answering these questions, this chapter discusses three sets of prac- tical problems encountered in benefit-cost assessments: establishing plausible results, displaying those results, and organizing the assessment's elements. The complexity of these practical problems varies directly with the complexity of the particular assessment. For example, in a qualitative assessment, a sensitivity analysis need be conducted only in very general terms. However, in a complex assessment, such as one presented in the example below, the plausibility of individual variables is specifically considered. In other words, the resources used in a sensitivity analysis can be tailored to the importance of the decision. The following sections of this chapter highlight these practical aspects of a benefit-cost assessment. Specifically, Section 5.2 presents a sensitivity analysis for a water quality standards example involving monetized benefits and costs. Section 5,3 describes narratives, arrays/ and graphs as alterna- tive ways of displaying the results of an assessment. Section 5.4 presents a checklist for organizing an assessment. Finally, Section 5.5 summarizes the chapter's main points. 5.2 SENSITIVITY ANALYSIS: A GAUGE TO BELIEVABILITY Introduction One ingredient in a good benefit-cost assessment is a sensitivity analysis of its key variables and assumptions. The most common variables considered in a sensitivity analysis are the parameters that determine benefits and costs, the discount rate, and the time horizon of the assessment. For example, the effectiveness of a particular treatment process may be uncertain so that the anticipated water quality may not be fully achieved. As previously discussed, the discount rate can be among the most important of these features because it affects both benefits and costs. A sensitivity analysis establishes a range for the net benefits in the assessment rather than simply portraying a single estimate. In principle, this is similar to the procedure in statistics that establishes interval estimates to 5-1 ------- I I I I I I I I I I I I I I I I I I bound the range of possibilities. In a benefit-cost assessment, greater uncer- tainties in the estimates of benefits and costs yield larger bounds to include most of the possible outcomes. A sensitivity analysis will show when the assessment is affected by the assumptions made. In instances where similar benefit-cost assessments are expected in the future, a sensitivity analysis can serve as an agenda for re- search by highlighting assumptions that influence the estimated net benefits. Even when an assessment is sensitive to the assumptions employed, its results are not invalidated. Rather, this sensitivity calls for more care in interpret- ing the results and in determining whether the assumptions are reasonable. A sensitivity analysis employs high and low estimates for both benefits and costs and estimates net benefits for a range of discount rates. Although there are no formal procedures in a sensitivity analysis, the following example highlights the decisions required to implement the various steps. Table 5-1. Key Elements of Benefit and Costs Line item (and key assumption) Cost savings3 Likely estimate (and range), $ million Benefits forgone Line item Likely estimate (and key (and range), assumption) $ million Capital costs New advanced waste treatment plant (size of plant) 4,8 (3.0 to 6.0) Process changes at meat processing plant (extent of changes) 3.2 (1.2 to 5.0) Additional fish- ing (5 percent growth rate) Additional swim- ming (probability of swimming is constant) 1.75 (0.50 to 3.0) 0.50 (0.2 to 1.25) Additional near- water activities (no new activities) 0.50 (0.2 to 1.50) Operating costs Advanced treatment Meat processing 1.0 0.5 Cost savings are the investment and operating costs forgone by not meeting the fish and wildlife propagation use for river. Likely estimate is listed first with the range in parentheses. Benefits forgone stem from the recreational activities had the fish and wildlife propagation use been achieved. Likely estimate is listed first with the range in parentheses. 5-2 ------- I I I I I I I I I I I I I I I I I I I Example Suppose a State is considering changing the use designation for Segment 30 of a river from fish and wildlife propagation (which is not being attained) to agricultural uses. The need for a sensitivity analysis is demonstrated by the data in Table 5-1, which shows the key elements in the benefit-cost as- sessment--cost savings and benefits forgone. Uncertainty both in the ulti- mate size of a new municipal waste treatment plant and in the estimation of the cost savings for the process change at a meat packing plant dictates the bounds for the cost estimates. In this case, the capital cost savings occur in the current year, so they are not affected by the selection of the discount rate. The estimates of the benefits forgone by the change in designated uses show even more uncertainty. The value of the estimated loss in fishing activ- ities depends on a 5 percent per year increase in fishing expected under the previous designated uses. By varying the assumed increase in fishing activi- ties under the new designated uses, the forgone benefits range from $500,000 per year to $3,000,000 per year. The estimation of forgone swimming activ- ities is based on the assumption that the level of swimming would not change. If adjustments are made for the uncertainty of the assumption, the estimated forgone swimming benefits range from $200,000 per year to $1,250,000 per year. The range of forgone benefits for near-water activities depends on the assumption that no new activities are developed for the river. When this assumption is relaxed, benefits for near-water activities range from $200,000 to $1,500,000. Step 1: Translate the Benefits and Costs into Present Values The first step in sensitivity analysis is to translate the benefits and costs into present values to make the net benefit calculation. This example simplifies the calculations by assuming a real rate of 4 percent and a project -life of 50 years. Only quantified and monetized benefits are considered in this example, and all capital costs are spent in year zero. The problem; then, is to translate the annual ,costs and annual benefits that occur each year into their present value equivalents, which requires the use of present value tables. The information needed for using present value (PV) tables is P/A,* 4 percent, 50 years. For the most likely case: Operating cost savings = P = $1.5 (21.482) = $32.2 million Total cost sayings = 32.2 million plus 8.0 million capital cost for- gone, or $40.2 million Benefits forgone = P = $2.75 (21.482) = $59.1 million *P/A is a heading found in most present value tables for translating annual costs into their present value equivalents. 5-3 ------- I I I I I I I I I I I I I I I I I I I Net benefits calculated for the most likely estimate benefits and costs are: Net benefits = PV benefits forgone - PV cost savings $18.9 million = $59.1 million - $40.2 million The assessment of net present value benefits shows the cost savings from the change would be outweighed by the benefits forgone if the most likely esti- mates are used. Table 5-2 summarizes -the results of calculations for the ranges of benefits and costs in addition to the most likely case. * Step 2: Perform Sensitivity Analysis for Discount Rate and Key Assumptions The sensitivity analysis shown in Table 5-3 describes the bounds for the net benefit estimates, using alternative discount rates and ranges of benefits and costs. Part A of Table 5-3 shows the most likely estimates of net bene- fits calculated with three different discount rates. Part B of Table 5-3 shows the outcomes that would result for the worst case expected to occur by esti- mating the cost savings at the lowest end of their range and the benefits fore- gone at the highest end for three different discount rates. Part C of Table 5-3 presents the estimates that correspond to the most optimistic case, with cost savings at the highest end of their range and benefits forgone at the lowest end of their range. Step 3: Interpret Sensitivity Analysis The sensitivity analysis shows that the net present value is sensitive to the discount rate in that the magnitude of the net cost savings estimates vary over a large range. However, changes in the discount rate alone over the range employed are not enough to change the direction of the net benefit as- sessment. In both the most likely and the worst cases the forgone benefits exceed the cost savings for changing the use designation. Only in the most optimistic case, where the cost savings are at the highest estimate and bene- fits forgone are at their lowest do the cost savings exceed the forgone bene- fits . The recommendation that could be made from the sensitivity analysis is that the change is likely to produce forgone benefits greater than the costs, with only a small chance that the results would be otherwise. To the extent there are effects that cannot be expressed in dollars, this range can also be used to indirectly define what the dollar value of these effects would need to be to change the evaluation. In the determination of the appropriate use classification, the decisionmaker could then weigh this small chance. 5.3 DISPLAYING THE ASSESSMENT RESULTS This section discusses three methods for displaying the results of a benefit-cost assessment: narratives, arrays or matrices, and .graphical dis- plays. Each method is described briefly, along with its advantages and dis- advantages. 5-4 ------- I I I I I I I I I I I I I I I I I I I Table 5-2. Sensitivity Analysis Calculations: Discount Rate at 4 percent Cost savings, $ Benefits forgone, $ Capital cost: 8.0 million in 1982 Range: 4.2 to 11.0 Operating cost: 1.5 per year Range: None Present value of operating costs: 32.2 million Present value of total (most likely case): 40.2 Range of present values of total cost savings: 36.4 to 43.2 million Recreation: 2.75 per year Range: 0.9 to 5.75 per year Present value: 59.1 million Range: 19.3 to 123.5 million Total (most likely case): 59.1 Range of total benefits forgone for all cases: 19.3 to 123.5 million Table 5-3. Sensitivity Analysis Discount rate Net present value of cost saving minus benefits forgone (million $) A. Most Likely Levels of Benefits Forgone and Cost Savings 2 4 6 -31.3 -18.9 -11.7 B. Cost Savings at Lowest Estimate—Benefits Forgone at Highest 2 4 6 -129.4 -87.1 -62.8 C. Cost Savings at Highest Estimate—Benefits Forgone at Lowest 2 4 6 29.9 23.9 20.5 5-5 U.S. EPA Headquarters Library Mail code 3201 1200 Pennsylvania Avenue NW Washington DC 20460 ------- I I I I I I I I I I I I I I I I I I I Narratives A narrative uses words to describe the results of an assessment. It can be used to describe either qualitative or quantitative information and is a simple, straightforward approach to displaying the results of an assessment. Its main disadvantage arises when there are several types of beneficial and detrimental effects to be weighed in the water quality decision. The evalua- tion of these diverse benefits and costs may be aided by expressing some of them with numerical estimates of benefits and costs. Combining narrative information with the array method discussed below can ease the comparison of benefits and costs. Examples of narratives used in combination with matrices are shown throughout Chapter 6. Arrays ' An array, or matrix, is a tabular display that contains written and nu- merical descriptions of the outcomes of an assessment. Arrays are most ef- fective when combined with the narrative display method discussed above. An array organizes information in a simple yet visually effective manner. By including qualitative information, the practitioner can easily describe the nature of the benefits and costs and the degrees of confidence in the esti- mates of either. If quantitative and monetized information is included, it can be supplemented with descriptions of any benefits and costs that are not monetized or cannot be quantified. Arrays may be readily adapted to the wide variety of cases likely to be encountered in an assessment of water quality programs. They do not require assumptions about the relationships among the variables presented in the array, and they make it easier for the practitioner to describe relation- ships known only in various degrees of accuracy. Arrays are particularly well-suited for displaying intangible benefits and costs and are used in both Chapter 2 and Chapter 6 to highlight these issues. Most effective arrays are used to organize information and to display substantive descriptions of the information presented. Arrays that present too much information can be divided into several arrays, but care is required to avoid unnecessary con- fusion. Inadequate descriptions may be worse because they require that users invest their own time. Graphs Graphs are an effective way of presenting information in a benefit-cost assessment, but considerable caution and scrutiny are advised when using them. Graphs can effectively show relationships between two variables, but often the information required to draw them is simply not available. These problems are less important for pie charts and bar graphs but are .prevalent for tradeoff curves, which show relationships between two well-defined objec- tives. Figure 5-1 presents an example graph that can be used for two well-de- fined objectives for a reservoir: flood control and recreation. Flood control is measured on one axis in thousand acre-feet of water Impounded annually, 5-6 ------- I I I I I I I I I I I I I I I I I I I 103 acre impounded/yr 8 2 3 Recreation 103 visits/yr Figure 5-1. Tradeoff curve. and recreation is measured on the other in visits per year to the site.. In this case, the shape of the curve shows that as more of the reservoir is used for a single objective, larger and larger amounts of the other must be given up. The tradeoff curve shows a frontier of alternative combinations of flood control and recreation that can be attained with a given reservoir in any year. A point inside the frontier, like M, represents an inferior, or less attractive, combination because more recreation can be attained while maintaining at least the same amount of flood control. However, the tradeoff curve provides no information about the relative attractiveness of points such as A through E. All are oh the frontier, and society must choose which allocation of the two objectives is most desirable. The extreme points A and E clearly show that substantial amounts of either flood control or recreation must be given up for an exclusive use. However, to know whether they are efficient from society's viewpoint requires that the value of each good be known. The economic prin- ciples behind the demand curve (which indicates that people will buy more as price is lowered) imply that there may not be a simple one-to-one relation between extra units of a commodity and the extra value from each unit. While a useful concept, tradeoff curves are inappropriate in many bene- fit-cost assessments. Many assessments for water quality policies will involve more than two objectives that would be extremely difficult to express in a sim- ple tradeoff relationship. For example, it is unlikely that suitable quantitative units could be derived to meaningfully express such objectives as enhanced ecological diversity, which may involve complex relationships that can easily confuse the practitioner about tradeoff relationships. 5-7 ------- Key Elements * Results of use attainability assessment • Other background on river segment, water quality 1. Advanced treatment application 2. Previous standards documents • List of benefits • List of costs • Data on benefits - Primary 1 . Recreation survey— State 2. Recreation survey— U.S. and region — Secondary 1. Published Federal, State, and local studies 2. Census • Data on costs — Primary 1. Engineering studies of treatment costs 2. Economic cost estimates 3. Financial cost estimates — Secondary 1. Published Federal, State, and local studies 2. Financial/investment sources • Techniques selected for measuring benefits 1. Direct survey 2. Participation survey 3. Travel cost 4. Hedonic • Techniques selected for measuring costs 1 . Engineering estimates 2. Computer-assisted model ing techniques 3. Economic cost estimates • List of key assumptions in measuring benefits and costs • Discount rate • Time horizon • Sensitivity analysis • Present value of net benefits - • Assessment of distribution • Cost impact assessment • Final assessment document ( .. _..... Status of Data for Assessment Not Applicable Available Requested Unavailable •v Figure 5-2. Checklist for a water quality standards benefit-cost assessment. 5-8 ------- In summary, tradeoff curves are convenient display techniques but are inappropriate when the objectives are not well defined or easily expressed in quantitative or monetized units. The practitioner should use this technique with a great deal of caution and only when substantial data are available. 5.4 BENEFIT-COST ASSESSMENT CHECKLIST This section describes a checklist that can be used in organizing a bene- fit-cost assessment. The. example is for a water quality standards, case, but the checklist is a general one. The checklist in Figure 5-2 provides a means, of tracking the various steps involved in the assessment process. This checklist can be tailored to fit the needs of each assessment by varying the types of benefits and costs included and by using the columns at the right-hand side of the list. These columns allow the checklist to accommodate the range from simple qualitative to complex quantitative assessments by designating the status of the data required for the assessment. A qualitative assessment can be conducted even with such status categories as "not applicable" and "unavailable." The com- plex assessment is made easier when the categories are "available" or "requested." The larger the "unavailable" category, the more strongly a State should consider using additional resources, either inside or outside State government, to obtain the needed information, especially if benefits or costs are likely to be sizable and if the decision is unclear after a qualitative .assessment. For a complex case involving potentially large benefits and costs, the extra value of acquiring the necessary data can be substantial. Data on a specific water body and the use of the techniques described in this hand- book can greatly simplify the complex quantitative assessment and assist in evaluating a proposed water quality action. 5.5 SUMMARY • • A sensitivity analysis for key variables and assumptions employed in an assessment is essential for a plausible assess- ment. Sensitivity establishes a range of outcomes possible for the assessment and will show when the assessment is sensitive to its assumptions. Three key components of an assessment that will be essential in the sensitivity analysis are the discount rate and estimates of benefits and costs. A narrative describes the results of an assessment in words. Its main advantage is its simple straightforward nature, while its disadvantage is in presenting results for more .complex assessments. Arrays or matrices are tabular displays that contain written and numerical descriptions. Narratives are often combined 5-9 ------- with the arrays as an effective display technique and are suited for a wide range of assessments. Graphs can effectively illustrate relationships between objec- tives that are quantified or monetized. However, the informa- tion necessary for these display techniques is frequently un- available. Tradeoff curves show a frontier of alternative combinations of quantities for two objectives. Caution and scrutiny are ad- vised in using these curves in an' assessment because the necessary information is often unavailable or it is impossible to apply to water bodies with a wide range of uses. A checklist is one way of organizing the procedures and infor- mation in a benefit-cost assessment. 5-10 ------- I I I I I I I I I I I I I I I I I I I CHAPTER 6 BENEFIT-COST--SAMPLE SCENARIOS 6.1 INTRODUCTION In water quality programs, the benefit-cost assessment practitioner must evaluate a wide range of policy actions. These could include both minor changes in designated uses for an intermittent stream and major changes .in designated uses or in advanced treatment or combined sewer overflow to provide recreation and intangible benefits that require substantial investments by cities and firms. Thus, the framework for evaluating these policy actions must be capable of comparing Incremental benefits and costs for a diversity of cases yet provide for a consistency in the application and presentation of the assessment. Even though scenarios are presented only for potential water quality standards decisions, the range of issues is broad enough to be useful to other water quality programs. To illustrate the type of assessments that might arise in water quality programs, this chapter develops three sample scenarios—simple, medium, and complex. Each scenario is designed to build on the preceding one, as new dimensions are added. Each scenario is introduced by a brief description of hypothetical sample cases. Each scenario refers the reader to the relevant handbook chapter(s) that provide more detailed discussion on specific issues. Although the step-by-step framework illustrated in the following scena- rios is both systematic and flexible enough to accommodate most of the benefit- cost assessment needs of various water quality programs, the values it assigns to benefits and costs should be regarded as approximations rather than abso- lutes. This note of caution has nothing to do with the framework itself, but with the specific information--e.g., poor quality data on linkages between water quality, fish propagation, and recreational fishing—used to estimate either costs or benefits. The following sections of this chapter present the three scenarios for benefit-cost assessment. Specifically, Section 6.2 describes a simple case scenario that uses only qualitative information, Section 6.3 presents the medi- um case scenario, providing some quantitative benefits and costs, and Sec- tion 6.4 contains a complex case scenario involving multiple benefits and costs. Finally, Section 6.5.summarizes the key issues in the scenarios. 6-1 ------- I I I I I I I I I I I I I I I I I I I 6.2 SIMPLE CASE SCENARIO Introduction The strength of benefit-cost assessment lies in its ability to organize ma- terial in a consistent manner yet remain flexible enough to accommodate a wide range of cases. This simple scenario illustrates a qualitative benefit-cost as- sessment and demonstrates both the consistency and flexibility in this evalua- tion procedure. In many water quality applications, the potential benefits and costs do not justify anything more than a qualitative benefit-cost assessment. Although a qualitative assessment does not quantify the information it organizes, it does provide a framework for presenting the character of the individual problem and for describing the judgment employed to make the assessment. For example, water quality standards decisions where qualitative assessment may be appropriate include stream-specific standards, such as the following: Public water supply designations for streams that have never been so used or that—because of low cost alternatives—are not likely to be so used in the future. Possible changes include removing the water supply use and adding recreation or agri- cultural uses that the water supply use might have precluded. Primary contact recreation uses for a stream that currently has few access points, as well as water quality limitations. Pos- sible changes include limiting types of recreation or changing uses to accommodate agricultural and industrial activities. Simple Case Scenario Format I. Define the Action .A* State is reviewing the designated uses for a specific segment of a river currently designated for use as a public water supply, although it has never been used in this capacity, and for primary body contact recreation. The action to be assessed is the removal of the water supply use and the addition of an irrigation use. The segment is located in a portion of the State that produces a substantial amount of agricultural products. The seg- ment is a primary source of water recreation in the area and supports fish- ing, swimming, and limited boating. II. Translate the Effects into Beneficial Outcomes and Costs Although the use change will cause slightly lower levels of dissolved oxygen and small increases in the levels of several other biological and chemi- cal water quality parameters, recreation activities will be unaffected. The primary benefit will be the increase in high-quality water available for irriga- 6-2 ------- I I I I I I I I I I I I I I I I I I I tion, resulting in cost savings for farmers who presently have irrigated farm- land or who will be able to irrigate additional farmland.* In terms of forgone benefits, the cost of the action is the loss of a poten- tial water supply source. However, groundwater sources exist, and residents are presently using them at lower costs and are expected to continue their use in the future. Nothing (e.g., shortages or contamination) is expected to threaten these groundwater supplies. . III. Calculate the Value of the Beneficial Effects Based on Willingness to Pay The value of the benefits of the change would equal the farmers' willing- ness to pay to obtain the water for irrigation rather than do without it. The benefits are expected to be positive on this basis, but not necessarily large because some alternative irrigation supplies exist. IV. .Calculate the Value of the Detrimental Effects Based on Opportunity Costs The opportunity cost of this action is loss of the potential source of drinking water. The alternative supplies of drinking water make this oppor- tunity cost near zero. If future demand and supply of drinking water should change substantially, the State could reconsider the designated uses because the long-term physical effects on the water will be limited in nature. V. Compare the Total Benefits and Costs The benefits of the action are the amounts the farmers would be willing to pay from this new irrigation source to irrigate their farmland without the action. The costs of the action, however, are zero, because they consist entirely of the lost opportunity for an alternative supply of drinking water, which can be obtained from existing groundwater resources. Thus, the total benefits of the action are greater than the costs. Therefore, the action will provide positive net benefits for the State and country. Recreation activity will be maintained, and cost savings for farmers will be greater than the loss of a potential source of drinking water. VI. Assess the Plausibility of the Results The results are not likely to change under almost all possible circum- stances. Only if dramatic changes occur in the drinking water situation-- i.e., if groundwater supplies become contaminated or suddenly in short supply-would the outcome be different. *This particular segment has ample flow but farmers have not been per- mitted to use if for irrigation. Withdrawal for irrigation will not noticeably affect recreation activities or fish populations. 6-3 ------- 6.3 MEDIUM CASE SCENARIO Introduction As a group, medium cases can be distinguished from simple and complex cases. While simple cases usually require only a qualitative assessment of .benefits and costs, medium cases normally involve some quantitative measures of benefits and costs. Further, while medium cases usually involve only one type of benefit—such as recreation—and the application of measurement ap- proaches and results from other studies, complex cases often require develop- ment of case-specific measurement approaches. One practical way to distinguish between simple and medium cases is to conduct the qualitative assessment and then to judge whether it provides enough information for a clearcut decision. If ;the outcome is not clear, the more quantitative medium case assessment may be needed. Water quality standards decisions that might require medium assessments include the follow- ing: Fish and wildlife propagation use for a stream that is not being attained because of industrial dischargers, the treat- ment options are limited to high-cost land-treatment solutions, and the river is not an important recreation source. Change could be to provide for a less restrictive fish arid wildlife use or for agricultural/industrial use. Fish and wildlife propagation use for a major river tributary with industrial dischargers. A change in designated use to accommodate existing industrial cooling would maintain rela- tively low levels of dissolved oxygen in a large segment of the river, primarily affecting fishing. The assessment would com- pare the loss in potential fishing benefits with cost savings to industrial firms. Fish and wildlife propagation use for a stream whose hydro- logical equilibrium has been affected by irrigation. A change in designated use to accommodate agricultural irrigation would cause the loss of a potential warm water fishery in the stream. The assessment would compare loss in potential fishing benefits with cost savings to agricultural irrigation users. Attaining a limited warmwater fishery (sunfish, carp, catfish) use for a channelized stream would require an advanced treat- ment plant for municipal wastes. One change would be to provide roughfish passage. The assessment would compare the loss in potential fishing benefits with the cost of advanced treatment. 6-4 ------- I I I I I I I I I I I I I I I I I I I Medium Case Scenario Format . . I. Define the Action A State is reviewing the designated uses for a specific segment of a river currently designated for agricultural arid industrial use and secondary recreation, primarily boating. The action to be evaluated is a use change to provide a warmwater fishery in addition to other uses. The river segment is capable of sustaining a warmwater fishery if discharges .from a municipal treatment plant are reduced to improve the level of dissolved oxygen. II. Translate the Effects into Beneficial Outcomes and Costs The effects of the change will be an improvement in dissolved oxygen levels to sustain the warmwater fishery under all flow conditions. This change will provide a new source for smallmouth bass recreational fishing while maintaining the existing uses. The primary cost of the change will be additional treatment of municipal wastes, but advanced treatment is not antici- pated. Ml. Calculate the Value of the Beneficial Effects Based on Willingness to Pay A travel cost model can be used to estimate the recreational fishing bene- fits. These benefits are estimated to be $70,000 per year and are assumed to remain at that level for the next 20 years.* This amount represents fisher- men's willingness to pay for the water quality improvement, but does not meas- ure nonuser values or intangible benefits associated with the water quality improvement. IV. Calculate the Value of the Detrimental Effects Based on Opportunity Cost The municipal discharger will be required to upgrade the quality of its treatment plant. From engineering estimates, this upgrading is expected to require an initial investment of $500,000 but no increases in operating ex- penses. There Will be a small economic impact on households who pay for the treatment services. (See Chapter 4 for details on measuring costs.) V. Compare the Total Benefits and Costs To compare them, benefits and. costs must be converted into present values. Costs incurred in the present year are already in present value terms ($500,000), but the benefits must be converted into a present value equivalent from a stream of annual dollars over 20 years. Using a social rate of time preference as presented in Chapter 2, these benefits are discounted *See the complex scenario for the use of a travel cost model to estimate recreational fishing benefits. The results are summarized for the medium case to minimize duplication among the scenarios. See Chapter 3 for general dis- cussion of the travel cost approach. 6-5 ------- I I I I I I I I I I I I I I I I I I I at 2 percent for a total 20-year value of $1,148,000, Thus, total benefits resulting from the change are $1,148,000, and total costs are $500,000, result- ing in a positive net benefit of $648,000. Adding the new use will therefore provide benefits in excess of the required investment costs. VI. Assess the Plausibility of the Results The key element influencing the sensitivity of this scenario is the selec- tion of the discount rate. To determine the plausibility of the results, net benefits should be calculated with different discount rates and compared. Chapter 2 recommended a high estimate of 5 percent for the social rate of time preference, and the Office of Management and Budget (OMB) requires the use of 10 percent for major regulations. Net .benefits calculated with these rates are listed below. . Discount rate (percent) 2 5 10 Net benefits ($) 648,000 372,340 95,980 As indicated, the use change would .produce positive net benefits under any of the three discount rates, implying that benefit estimates are not sensitive to the discount rate selected. Sensitivity analyses also may be performed for the estimates of benefits and costs--for example, with benefits and costs esti- mated at ±30 percent of the average value. More details on using sensitivity analysis are provided in Chapter 5. In assessing the plausibility of the results one might .also consider the distribution effects discussed in Chapter 2. In this sample case these effects are an inconsequential part of the assessment because no one group is adv- ersely affected. However, distribution effects vary from case to case and should be addressed.* 6.4 COMPLEX CASE SCENARIO Introduction The complex case scenario is distinguished from the simple and medium case scenarios by several characteristics, including its consideration of multi- ple types of tangible benefits, intangible benefits, and investment in waste treatment beyond the technology-based requirements. In addition, benefits and costs in complex cases are likely to be an order of magnitude, or more, above those in medium cases. In fact, if benefits and costs are not signifj- *The complex scenario that follows provides a detailed assessment of dis- tribution effects. 6-6 ------- I I I I I I I I I I I I I I I I I I I cantly greater than those of the typical medium case scenario, use of the com- plex case scenario is not recommended because the effort it requires would not be justified. . Measurement and valuation of benefits and costs may require several methods discussed in Chapters 3 and 4. The types .of benefits and costs, data availability, and available staff resources will influence the level of detail in the assessment. For controversial decisions, a State may decide to obtain assistance from outside the agency to conduct the assessment. Complex cases may include the following: Adding fish and wildlife propagation and primary contact recreation use designations for a stream that will require advanced treatment for municipal discharges. Benefits will include multiple types of recreation activities, intangible benefits such as enhanced ecological diversity, and nonuser benefits. Costs will include .investments by firms in waste treatment beyond investment required for technology-based requirements. Providing public water supply from a stream that will require advanced treatment for municipal wastes. Water supply bene- fits could be for more than one downstream municipality. Complex Case Scenario Format I. Define the Action A State is reviewing the designated uses for a 10-mile segment of River 1, located between Cities A and B. The river is currently designated for secondary contact recreation, but the State is considering adding swimming and fish and wildlife propagation. These uses cannot be attained by imposing the technologies required under the Clean Water Act (Section 30i(b)(2)), but they could be attained with more stringent controls on municipal and industrial point dischargers. River 1 drains an area of 7,386 square miles, and most of its length is characterized by steep banks and rugged terrain. The major point sources of discharges on the river segment of interest are iron and steel facilities and a municipal sewage discharge just outside City B and upstream from City A. The river is navigable the entire year, and a considerable amount of coal is barged through several locks and dams operated by the U.S. Army Corps of Engineers. River 1 is currently used for boating recreation and for activ- ities near water at several parks located along its banks. The most notable water quality problems limiting swimming and fish and wildlife propagation have been associated with the dissolved oxygen and ammonia levels, with fre- quent violations of current standards at low flow. 6-7 ------- I I I I I I I I I I I I I I I I I I I II. Translate the Effects into Incremental Beneficial Outcomes, Costs, and .Economic Impacts A. Benefits Recreational fishing Swimming Activities "near water" Navigation Ecological diversity Nonuser/nonuse benefits. B. Costs Capital, and operating costs for steel plants Capital and operating cost for City B C. Distributional Impacts Reduced profitability of steel plants Effects on operations (will shutdowns occur?) Increased sewer charges for municipal taxpayers Employment Price Impact on firms within industry III. Calculate the Value of Beneficial Outcomes Based on Willingness to Pay •A. Recreational Boating Benefits Step 1 Determine what data are available for recreation on or near River 1. Likely sources include the State recreation plan, the EPA 208 management plan, the 1977 Department of Interior outdoor recreation survey, and the U.S. Army Corps of Engineers. Step 2 Use the Corps of Engineers recreation survey data on River 1, which contains information on users and where they are from. This informa- tion will allow estimation of recreation benefits with the simplest version of the travel cost model. (See Chapter 3 for more details on travel cost models.) A number of implicit assumptions are used in this scenario. For ex- ample, the State is assumed to have the information necessary to estimate the current and potential demand for fishing and swimming for River 1. The simplest measure of travel cost is used even though it excludes the cost of travel time, time spent onsite, and the influence of substitute sites on the demand for River 1 recreation. Also assume that the travel cost to the site 6-8 ------- is capable of capturing all the factors that influence the decision to fish or swim along the river. This implies that no changes in access, docks, and other site features are occurring. (See the case study of the travel cost model in Chapter 3.) Equation (6.1) illustrates the travel cost demand equation for these as- sumptions. The equation omits several origin-specific variables (such as in- come, age, education, etc.) that determine recreational demands for a site be- cause they would only change the position of the demand curve for each origin zone. The subscript i refers to these omitted variables: A) (6.1) where A = activities that would be permitted by the water quality use classi- fications (A = B for beatable, F for fishable, S for swimmable), V. = number of visits to the site from origin zone i, and POP. = population of origin zone i. Step 3 Calculate recreation benefits for existing water quality level. These calculations are needed to obtain the baseline and must be netted out from the benefits of the additional use designations. Use the data in Table 6-1 to esti- mate travel cost model. Remember: these are user benefits only. Table 6-1. Demand for Recreation for River 1—Water Quality at Level Suitable for Boating Limits of zone of origin (miles) 1 to 20 21 to 40 41 to 60 61 to 80 81 to 100 100+ '1980 population (1,000) 1,000 1,500 2,000 2,500 3,500 5,000 Total No. party visits 1,650,000 600,000 200,000 50,000 25,000 2,500 Visits per 1 ,000 population 1,650 400 100 20 7 0.5 Consumer surplus per individual ($) 2.40 1.80 1.40 1.00 0.60 0.20 6-9 ------- I I I I I I I I I I I I I I I I I I I Calculation of Benefits n Benefits = TO* , .'/VP.' TC - B)dp. (6.2) Benefits = (number of individuals) x (benefits per individual) where TC* = the travel cost at which there would be a zero visit rate, and TC. = travel cost for ith origin zone. Figure.6-1 illustrates the consumer surplus estimates that result from this calculation. Two assumptions are implicit for the estimation. First, as with XYZ in Figure 6-1, calculate the representative individual's consumer sur- plus in an origin zone, and then assume that all individuals in the origin zone have the same consumer surplus. This procedure is possible because use (i.e., the visit measure) is considered to arise from the population as a whole. (The calculation would be different for travel cost models estimated from survey data. See Chapter 3 for specifics.) Travel Cost $/visit Consumer surplus fi(Pi,B> Pop Figure 6-1. Travel cost demand function and consumer surplus with boatable water. For example, in Table 6-1, the fifth column reports an example of the estimated individual consumer surplus by origin zone. Multiplying each con- sumer surplus times the population and adding across origin zones gives the aggregate consumer surplus for the site with boatable water: 6-10 ------- 1 1 1 ^^p 1 1 • 1 1 1 1 • 1 •P_ 1 1 •• 1 M \ 6 Benefits for Boatable Water, 1980-81 = \ population x (consumer surplus ^•4 per individual) Zone 1 A- = '$(1,0.00,000 x 2. Zone 4 + ' 2,500,000 x 1 = $(2,400,000 + 2, + 2,100,000 + 1, Benefits for Boatable Water = B. Recreational Fishing Step 1 Use the results on (see travel cost case study, zone i=1 Zone 2 Zone 3 A *. 40' +f 1,500, 000 x 1.80" +'2, 000, 000 x 1.40' Zone 5 Zone 6 .00"+ 3,500,000 x 0.60' +'5,000,000 x 0.20)* 700,000 + 2,800,000 + 2,500,000 000,000) $13.5 million for the site with current water quality levels. Benefits fishing benefits in the Monongahela River study Chapter 3, for details). The results from the Moripngahela study can be used to calculate benefits analogous to those re- ported in Table 6-2. Note: determining the plausibility River 1 and the Monongahela practitioners might consider three questions in of results: What similarities exist between River? What are the problems in adapting the results? How significant are they? Table 6-2. Demand for Recreation for River 1 --Water Quality at Level Where Limits of zone 1980 of origin population (miles) (1,000) 1 to 20 1,000 21 to 40 1,500 41 to 60 2,000 61 to 80 2,500 81 to 100 3,500 100+ 5,000 a Gamefish (Bass) Can Live in River Consumer Visits per surplus per Total No. 1,000 individual party visits population ($) 3,400,000 3,400 0.10 1,500,000 1,000 0.10 800,000 400 0.10 500,000 200 0.10 350,000 100 0.10 50,000 10 0.10 These calculations assume a parallel shift in demand. It need not be parallel in particular applications (see travel cost case study, Chapter* 3). 6-11 ------- I I I I I I I I I I I I I I I I I I I Step 2 . For this framework, the travel cost model provides the relationship between the demand for the recreation site and the water quality use designa- tion. Thus, an improvement in water quality from boatable (B) to fish- able (F) can be expected to shift the demand for the site's services. This conclusion follows from the. site's ability to support both types of recreational activities under the higher designated use. Figure 6-2 illustrates the type of shift involved. ^Change in consumer surplus with fishable water Figure 6-2. Travel cost demand function and change in consumer surplus with fishable water. To calculate the incremental benefits associated with this change in use designation, we need to estimate YXWV. In terms of the travel cost demand model, YXWV would be given as: Benefits (incremental to fishable) = n - E P°Pi7 TC** TC*. F)dp. - TC . 5=1 /" j J TC. B)dp. (6.3) The second term in Equation (6.3) simply repeats the benefit estimate for the site with the existing designation which allows boating. Thus, the increment to benefits because of the change in water quality is being calcu- lated. This example assumes that individuals from the same origin zones are using the site, and no users from new origin zones so Equation (6.3) can be rewritten as the sum of the increments to the individual benefits realized in each origin zone, as in Equation (6.4): 6-12 ------- I I I I I I I I I I I I I I I I I I I Benefits for increment to fishabfe water n = E popi 'TC** /' TC*i f. (p., F)dp. - J f.(Pi, B)dP] TC (6.4) number of individuals increment to each individual's benefits from use designation For the representative individual, this increment is YXWV in Figure 6-2. Table 6-2 reports some illustrative increments in the fifth column, and the calculation process is then similar to that described above,; Benefits for Increment to Fishable Water, 1980-1981 Zone 1 visits x travel costs Zone 2 Zong 3 = $(1,000,000 x 0.10 +1,500,000 x 0.10 + 2,000,000 x 0.10 Zone 4 Zone 5 Zone 6 __ _ _ .^^^^^^^^^^^^^^^^^^^fc. .^^^^^^^^^^^^^^^^•••••k '2,500,000 x 0.10 V 3,500,000 x 0.10V5,000,000 x 0.10 x = $(100,000 + 150,000 + 200,000 + 250,000 + 350,000 + 500,000) Benefits for increment .to fishable water = $1.6 million per year. C. Swimming Benefits Step 1 Use the results on swimming benefits in the Monongahela River study (see travel cost case study, Chapter 3, for details) to calculate infor- mation such as that in Table 6-2. Step 2 To calculate the benefits from swimming use, the frame of reference must be defined. Specifically, which incremental benefits are of more interest — those associated with moving from a fishable to a swimmable use des- ignation, or those associated with moving from boatable to swimmable? The method used to estimate these incremental benefits will depend on the refer- ence point. For example, in Figure 6-3, which shows all three demand func- tions, movement from fishable to swimmable leads to incremental benefits (per individual) of VWUT. A change from boatable to swimmable includes this in- crement along with the increment associated with the improvement to fishable (YXWV). 6-13 ------- I I I I I I I I I I I I I I I I I I I ^Change in consumer surplus with swimmable water V Pop Figure 6-3. Travel cost demand function and change in consumer surplus with swimmable water. The calculations would follow the same outline used in the preceding two examples. Important in this case is the point of reference used to calculate the incremental benefits. If procedures from previous section are followed: Benefits for the increment from fishable to swimmable water = $750,000. D. , Recreational Benefits of Activities Near Water Improvements in water quality will make additional acres usable for activities .near River 1, including, for example, hunting, birdwatching, hik- ing, photography, and sightseeing. This example shows what can be done to estimate the benefits when demand information is not available. In particular, a participation model can be used to predict the increase in recreation activ- ities near water and a recreation day value estimate can be used from other sources. (See participation survey case study, Chapter 3, for details.) An important issue that arises with this approach is the consistency of benefits derived from the travel cost model with those from the participation model. Specifically, the travel cost model describes the demand for recrea- tional site services which can be used in a variety of recreational activities, including those "near water." Since the participation model for predicting demand for recreation near water may be measuring benefits reflected in the travel cost model, some double counting can occur. The following five steps can be used to estimate the value of the in- crease in near-water activities. Step 1 Determine the availability of data on activities near water for River 1. 6-14 ------- I I I I I I I I I I I I I I I I I I I Step 2 Estimate the increase in near-water activities resulting from the fish and wildlife propagation use designation. Analysis of River 1 determined that 8,931 acres along the river are currently suitable for these activities. Achieving the use designation was predicted to make another 2,700 acres suit- able for near-water activities and to improve the recreational facility rating by 1 unit (on a scale of 1 to 5). Step 3 Estimate the change in the probability of participation (AP) in the near-water activities among the population as a result of the improvements. Using a model (see Chapter 3, participation survey case study for details) that correlates participation in near-water activities with available water recre- ation area and a rating of the recreational facilities, the increased probability of participation can be calculated as: AP = 0.38485 /change in acreage of recreationY+ 0.03142/change in recrea- ^ I water available per capita j \ tional facility rating \ AP = 0.38485 AP = 0.0319. 2,700 new acres 12,000,000 population in Cities A and B) + 0.03142 /I unit improvement I recreational facility \rating Step 4 Estimate the change in participation days attributable to River 1 between Cities A and B. Specifically, using the national average of 2.0 near- water activity days per participant per year, additional near-water activity days can be calculated as: • Change in near- water activity days /change in \ / Regional\ I probability \ [population \ I of. near-water yparticipation (0.03I9) I Cities A I \and B / x (2,000,000) 'Near-water' days per participant (2.0) Change in near-water activity days = 127,600/year. Step 5 Estimate the value of additional near-water activity days. Re- searchers have estimated values per near-water activity day .to range from $12.00 to $18.50 (Volume II will provide more details on these values). Multi- plying these values by the number of additional near-water activity days gives the following range of values for the additional activities: 6-15 ------- I I I I I I I I I I I I I I I I I I I (12.00/near-water activity day)(127,600 days) = $1.5 million (18.50/near-water activity day)(127,600 days) = $2.4 million Note: Remember the limitations of this approach: 1. The valuation step is not based on willingness to pay. 2. Limitations are caused by transferring the model from one study to another. E- Benefits of Improved Ecological Diversity Attaining the fish and wildlife .propagation use designation also means enhanced ecological diversity of River 1 between Cities A and B. These benefits are not quantifiable at the present time but would include more diverse flora .and fauna, enhanced diversity of fish species, and other related elements. Note: These benefits will be listed in the policy array as "positive nonquantifiable" (see Chapter 2 on intangible benefits), allowing the decision- makers to better focus on these questions. In some cases, in fact, it may be possible to list specific species that the new use will add or preserve. F, Benefits of Improved Navigation Although achieving the .designated use may also benefit navigation in River 1, indications are that the magnitude of these benefits is negligible. Thus, no attempt will be made to quantify them (see Chapter 3 on measuring benefits to firms). G. Nonuser/Nonuse Benefits In addition to recreation in or near the water by current users, in- trinsic benefits—option values, existence values/ and aesthetics--may account for important benefits of improved water quality. Option value refers to the value current and potential users place on having the option to use the water resource at some future time. This value is assigned to the resource because there is some uncertainty regarding its future availability or regarding future demand. Existence benefits are measured by the value people place on ac- tions that ensure a resource is (or will be) there, regardless of their actual or potential use. This is sometimes termed vicarious enjoyment or attributed to a bequest motive. Aesthetics refer to beauty that may be appreciated by users and/or those who reside or travel nearby. Many unresolved issues exist concerning the inclusion and valuation of intrinisic benefits. For example, there is not yet any agreement oh whether aesthetics should be included in intrinsic benefits or measured separately or indeed, whether they be measured at all. Thus, intrinsic benefits are in- cluded in this assessment as "positive not quantified" (see the contingent valuation case study in Chapter 3, which highlights an approach for quantify- ing and monetizing these benefits,) 6-16 ------- I I I I I I I I I I I I I I I I I I I H. Estimate Benefits That Will Accrue In Future Years Assume that River 1 recreation benefits will remain constant for the 25-year lifetime used for the analysis. This is probably a conservative as- sumption because, with visit rates staying the same, any population growth in the area would lead to more recreation (unless congestion becomes a problem). This growth is offset, however, by other important factors such as the in- fluence of substitute sites that are excluded from the estimation models. Thus, each benefit estimate can be considered to be a constant stream over the project lifetime. IV. Value the Detrimental Effects Based on Opportunity Costs and Calculate Economic Impacts A. Calculate Increased Capital and Operating Costs for Model Steel Plant Step 1 Determine availability of data sources (see Chapter 4 for details on likely sources of cost data). Step 2 Use the model plant technique discussed in Chapter 4 along with the wasteload allocation to estimate the capital costs of achieving the fish and wildlife propagation use designation. A related study of pollution controls for steel plants suggests that the technologies shown in Table 6-3 will be needed to meet the regulation. Two steel plants correspond closely with the model plants; two are about 3/4 the size assumed for the plant; and 1 plant is 1-1/4 times larger than the model plant. Table 6-4 shows capital cost estimates for these plants. Table 6-3, Capital Equipment and Cost for Model Plant to Meet Regulation Capital equipment Wastewater discharge treatment pond Water integrated flow system Pretreatment and handling system 5 segments 30" piping Total Cost (million $) 1.2 1.6 1.2 2.0 6.0 Step 3: Calculate Operating Costs Assume for the model plant analysis that operating costs are roughly 0.05 of capital costs per year for the life of the plant, regardless of the size of the plant. Using this estimate and the capital costs from Table 6r4, calculate: 6-17 ------- I I I I I I I I I I I I I I I I I I I Table 6-4. Estimated Capital Costs Number of. plants Total 2 2 I 5 Percent of model plant 0.75 1.00 1.25 Total estimated capital cost (million $) 9 12 7.5 28.5 Operating costs/year = 0.05 x $28.5 million = $1.4 million/year. The problem of converting costs into comparable measures is addressed in Chapter 4. B. Calculation of Capital and Operating Costs for City B To meet the requirements of the designated use of fish and wildlife propagation, City B will have to add an advanced waste treatment plant for its sewage wastes. Capital and operating costs can be calculated using the model plant method for advanced waste treatment (see Chapter 4). Step 1 Determine availability of data on costs of advanced treatment plants (see Chapter 4 for details on availability of data on advanced treatment plants). Step 2 Determine additional capacity needed for the advanced waste treat- ment plant. Assume that the facility is to provide for.maximum daily flow equal to two times expected average daily flow. (This information would come from a City B sewer study, for example.) The additional advanced waste treatment capacity required is 60 Mgal/d. Step 3: Table 6-5 shows the capital cost estimate resulting from the model plant approach. Table 6-5. Cost of Treatment Plant, City B Plant capacity Capital cost 60 Mgal/d $5 million 6-18 ------- I I I I I I I I I I I I I I I I I I I Step 4 Use the model plant to estimate operating costs. Based on other studies 10 percent of capital costs can be used to represent operating costs for City B: Operating cost = Capital cost x operating cost factor $500,000/yr = 5 million x 0.10. C. Perform Plant Closure Analysis to Assess Impact of Regulation on Plant Operations The plant closure test provides a straightforward assessment of the impact of the regulation following the rules of thumb referred to in Chapter 2 for plant closures. There is considerable uncertainty involved in actual plant closure decisions, so this appraisal will give a benchmark rather than a com- .plete determination of plant closures. Step 1 Determine data availability (see Chapter 2 and Chapter 4 on finan- cial data sources that are needed for this test). Step 2 Determine opportunity cost of capital and average liquidation value to the two steel companies that own the five steel plants along River 1. Use these estimates in the following formula to determine the critical rate of return for closures: Average liquidation value (m) x opportunity cost of capital (r) = critical rate of return for closure American Steel: m x r = critical rate of return 0.50 x 5 percent =2.5 percent. Riverton Steel: Step 3 m x r - critical rate of return 0.60 x 6 percent =3.6 percent. Compare critical rate of return with plant's rate of return on re- placement cost with the regulation to determine potential Closure candidates: 6-19 ------- I I I I I I I I I I I I I I I I I I I American Steel: Rate of return vs. critical rate of return 5.0 percent vs. 2.5 percent. Riverton Steel; Rate of return vs. critical rate of return 4.8 percent vs. 3.6 percent. Based on this simplified test using rules of thumb, there are no closure impacts. If closure were indicated, .more detailed calculations and data would be needed to assess the issue more thoroughly. Volume II will address these more detailed comparisons. D. Perform Profit Reduction Test Another test to assess the magnitude of the impact of the regulation on companies is the profit reduction test. A simplified version of this test is illustrated below. Step 1 Determine available data (see Chapter 2). Step 2 Estimate reduction of profits as a percent of current dollar returns for the two steel companies: American Steel: 5 percent reduction in current dollar returns. Riverton Steel: 8 percent reduction in current dollar returns. These reductions are both relatively small impacts, so no additional calcula- tions are needed. E. Determine Impact on Households in City B Another issue in the impact assessment is who bears the cost of ad- vanced treatment. Issues that will need resolution include the following: Does City B receive EPA assistance for the advanced treatment plant? If so, the citizens of City B and all U.S. taxpayers bear the cost impacts. If not, only residents of City B bear them. Will future costs be incurred by the residents of City B when these costs are passed along? For purposes of illustration, no advanced treatment assistance from EPA is assumed. 6-20 ------- Step 1 Determine available data (see Chapter 2 on municipal waste treat- ment cost impacts for more on available data). Step 2 Determine impact of increased cost on residents of City B. Sewage costs will increase an average of $15 per year per household for residents of City B. V, Compare Total Benefits and Costs A. Check for Less Than Fully Employed Resources Volume II will present guidelines on determining whether money costs overstate true social costs due to the presence of significant unemployed resources. Assume the review shows that unemployment equals 5 percent in both City A and City 'B and that a major percentage of materials comes from outside the area. There are no problems with overstating costs in this case. B. Discounting Benefits and Costs This is a crucial step in the process of assembling the benefit-cost assessment. There are several key steps to. be performed (see Chapter 2 for detailed treatment of these steps). Step 1 Select discount rate for converting future benefits and costs into present values. OMB guidelines for regulatory impact analyses recommend a real discount rate of 10 percent. This seems high for a real rate that re- flects either the opportunity cost of capital or society's preferred rate. Thus, a range of rates should be used, with 10 percent on the upper end, and the sensitivity of results should be compared to the discount rate. A social rate of time preference procedure suggested in Chapter 2 is illustrated in Step 2 below. Step 2 Discount annual monetized benefits into present values. Annual Monetized Benefits Fishing Swimming Near water Total million $ 1.55 0.75 1.5 to 2.4 3.8 to 4.7 6-21 ------- I I I I I I I I I I I I I I I I I I I Since the benefit amounts are constant across time periods, to convert mone- tized annual benefits into present value equivalents, use the following formu- la. It calculates the present value (P) of a stream of annual benefits (A) for n years (assumed project life is 25 years) and interest rate (i): 1 - 1 P = A (1 * i) n Table 6-6 shows the present values of benefits calculated with discount rates ranging from 2 to 10 percent. Monetized benefits range from roughly 35 to 92 million 1981 dollars. Table 6-6. Present Value of Benefits Discount rate (percent) . Present value of benefits (million $) 2 4 6 10 74.2 to 91.8 59.4 to 73.4 48.6 to 60.1 •34.5 to 42.7 Step 3 To simplify matters, assume that all the capital costs are incurred in 1981, so that they are already present values. Based on the Lind proce- dure, these displaced private investments by the steel companies are multi- plied by the shadow price of capital--1.9. The capital costs for City B treat- ment plant are assumed to displace consumption (so they do not need adjust- ing by the shadow price of capitai). The next task is to translate the stream of operating costs into a pres- ent value equivalent. The operating costs are assumed to be displaced ex- penses for the firm and therefore do not require adjustment by the shadow price of capital. These expenses represent displaced consumption and are dis- counted at the social rate of time preference. This can be done with the same formula as above: 6-22 ------- I I I I I I I I I I I I I I I I I I I 1 - P = A 0 * On| , where A is how a constant stream of annual operating costs. Table 6-7 shows the present value of operating costs calculated with the same range of dis- count rates used in the benefits calculation. Table 6-7. Present Value of Operating Costs Discount rate (percent) Present value of operating costs (million $) 2 4 6 10 37.1 29.7 24.3 17.2 Table 6-8 shows the present value of total project costs, which range from 76 to 96 million 1981 dollars. Table 6-8. Total Project Costs Discount rate (percent) 2 4 6 10 Capital3 59.2 59.2 59.2 59.2 Operating 37.1 29.7 24.3 17.2 Total (million $) 96.3 88.9 83.5 76.4 Private capital costs of the steel companies are adjusted by shadow price of capital. Note: Even if EPA assistance is received for plant construction, the society still incurs roughly the same costs because these EPA funds would be di- verted from their next best alternative use. There would perhaps be some 6-23 ------- I I I I I I I I I I I I I I I I I I I difference because Federal and local dollars might have different forgone op- portunities, but the difference is probably minor for purposes of this assess- ment. table 6-9. Benefits and Costs of Attaining Fish and Wildlife Propagation Use Designation Cost/benefits Quantity Monetary value (million $) Benefits Recreation (fishing, swimming, near water) Enhanced ecological diversity User/nonuse—intrinsic benefits to users and nonusers Costs Capital .and operating (Firms; City B) Environmental Not quantified Not quantifiable Not quantified for this study Not relevant None 35 to 92 Not monetizable Not monetized for this study 76 to 96 0 VI. Assess the Plausibility of Total Benefits and Costs To .assess the plausibility of benefits and costs, construct an array and compare them. For example, Table 6-9 is an array showing both the benefits and the costs of attaining the fish and wildlife use designation. The ranges of monetized benefits and monetized costs overlap, nonmonetized benefits are positive, and there are no nonmonetized costs. The upper end of the monet- ized benefits is slightly less than the upper end of the range of monetized costs. The lower end of the monetized benefits is considerably less than the lower end of the cost range. Thus, the use designation decision would have to consider whether the intrinsic benefits and nonquantified, enhanced ecolog- ical diversity are greater than the difference between lower benefits and cost bounds. When the suggested 2 percent social rate of time preference is used, the monetized benefits are approximately equal to costs. Table 6-10 shows expected benefits from the use designation change and their distribution among area residents and visitors. In some instances, de- scribing the distribution of benefits among more narrowly defined groups may be desirable. For example, assessments involving a river segment that forms the boundary between two States may require more precise distribution infor- mation because the political issues are more complex when more than one State is involved. f 6-24 ------- I I I I I I I I I I I I I I I I I I I Table 6-10. Distribution of Benefits and Costs Benefit/cost Distribution Recreation Enhanced ecological diversity Nonuse/nonuser benefits Navigation Increased operating and capital cost for steel plants Capital and operating costs of advanced treatment plant Users of River 1--primarily recreators from City A and City B To some extent same as above, but probably accrues to society as a whole including future generations. Residents of river basin. Visiting recreators who have bequest motive for future generations. Shippers, boat owners, purchasers of materials shipped along river. Small magnitude in this study. American and Riverton stockholders to the extent the costs are not passed on to customers. Given the elasticity of demand for steel and market shares of these firms, some costs will be passed on to users of steel products. U.S. taxpayers may also bear some costs because of special depreciation treatment for pollution control expenditures. $15 per user of water and sewer service in City B. This may be lower if construction grant Is obtained; then U.S. taxpayers will share in the burden. To gain additional perspective on the distribution of benefits and costs, the breakdown of recreational benefits by different income groups Is shown below: Income ($) 0 to 10,000 10,000 to 20,000 20,000 to 30,000 30,000 to 40,000 40,000+ Benefits (percent) 15 25 35 20 5 In this example, 75 percent of the benefits accrue to households with incomes of less than $30,000/year. Decisionmakers may view this use change as more 6-25 ------- I I I I I I I I I I I I I I I I I I I desirable because of its larger share of benefits going to low and middle in- come groups. The expected costs of the programs also are arrayed in Table 6-10. The basic economic principle underlying the distribution of the industrial operating and capital cost is the responsiveness of demand for the industry's or firm's products—the larger the responsiveness of quantity demanded to price changes, the smaller will be the ability of the firm or industry to pass along cost increases to its customers. The stockholders, workers, and other re- source suppliers to the firm will have to bear the cost increases that are not passed along through lower dividends or reduced wages. For the steer firms in this example, the degree of sales responsiveness to price changes will be influenced by their share of the market, the demands for products which use the type of steel produced, competition from foreign sources of steel,-and competition from substitutes (aluminum). The municipal waste treatment costs likely will be borne by the residents of City B. Residents will include current residents and future residents who move, or are annexed, into the city's water or sewer service system. If the city receives a Federal construction assistance grant, then the largest share of the capital costs is distributed among a much larger group—the Federal taxpayers. •* This example shows how distribution issues are integrated with the other steps in the assessment process. They are described and listed but do not affect the measurement of either benefits or costs. Finally, the exact nature of the distribution descriptions will depend on the complexity of the issues In the assessment itself. 6.5 SUMMARY The strength of benefit-cost assessment is its ability to consist- ently organize information for a wide range of applications. Qualitative assessments, using primarily descriptive informa- tion, are sufficient in many water quality applications because of the sizes of potential benefits and costs. Defining the water quality action to be evaluated in the benefit-cost assessment determines the baseline and suggests the level of complexity in the assessment. Translating effects of a water quality action into beneficial outcomes and costs requires an understanding of the linkages between each element. Valuing beneficial effects should be based on individual's will- ingness to pay for them. Valuing detrimental effects should be based on opportunity cost to society of the effects. 6-26 ------- I I I I I I I I I I I I I I I I I I I Comparing total benefits and total costs may involve qualita- tive, quantitative, or monetized information. Assessing the plausibilty of benefits and costs may involve test- ing the sensitivity of assumptions made in estimating benefits and costs, or in selecting a discount rate. Describing the distribution of benefits and costs may provide valuable data for the decisionmaker. 6-27 ------- ------- I I I I I I I I I I I I I I I I I I I CHAPTER 7 REFERENCES Anderson, J. E., 1974, "A Note on Welfare Surpluses and Gains from Trade in General Equilibrium," American Economic Review, Vol. 64, No. 4, September 1974, pp. 758-62. Ayres, R. V., and A. V. Kneese, 1969, "Production and Consumption Exter- nalities," American Economic Review, 59, No. 3, June 1969, 282-97. Bishop, R. C., and T. A. Heberlein, 1979, "Measuring Values of Extra-Market Goods: Are Indirect Measures Biased?" American Journal of Agricultural Economics, Vol. 6, December 1979, pp. 926-30. Bockstael, Nancy E., and Kenneth E. McConnell, 1981, "Theory and Estimation of the Household Production Function for Wildlife Recreation," Journal of Environmental Economics and Management, Vol. 8, September 1981, pp. 199-214. Brookshire, David S., Mark A. Thayer, William D. Schulze, and Ralph C. d'Arge, 1982, "Valuing Public Goods: A Comparison of Survey and Hedonic Approaches," American Economic Review, Vol. 72, March 1982, pp. 165-77. Carlton, D. W., 1979, "Valuing Market Benefits and Costs in Related Output and Input Markets," American Economic Review, Vol. 69, No. 4, Septem- ber 1979, pp. 688-97. Chemical Engineering, published by McGraw-Hill, New York Cronin, Francis J., 1982, Valuing Njanmarket Goods Through Contingent Mar- kets, prepared for U.S. Environmental Protection Agency, Pacific North~ west Laboratory, Richland, Washington, September 1982. Davidson, P., F. G. Adams, and J. Seneca, 1966, "The Social Value of Water Recreational Facilities Resulting From an Improvement in Water Quality: The Delaware Estuary" in Allen V. Kneese and Stephen G. Smith, eds., Water Research, Johns Hopkins Press for Resources for the Future, Inc., 1956T " Desvousges, William H., V. Kerry Smith, and Matthew P. McGivney, 1983, A Comparison of Alternative Approaches for Estimating Recreation and Related Benefits of Water Quality Improvements, report prepared'for U.S. Environmental Protection Agency, Research Triangle Institute, Research Triangle Park, North Carolina, March 1983. 7-1 ------- I I I I I I I I I I I I I I I I I I I Deyak, Timothy A., and V. Kerry Smith, 1978, "Congestion and Participation in Outdoor Recreation: A Household Production Approach," Journal of Environmental Ecgn£rnics and Management, Vol. 5, March 1978, pp. 63-80. Dwyer, J, F., J. R. Kelly, and M. D. Bowes, 1977, Improved Procedures for Valuation of the Contribution of Recreation to National Economic Develop- ment, Urbana-Champaign: University of Illinois, 1977. Engineering News- Record, published by McGraw-Hill, New York. Fox, I. K., and O. C. Herfindahl, 1964, "Attainment of Efficiency in Satisfy- ing Demands for Water Resources," American Economic Review, Vol. 54, No. 2, May 1964, pp. 198-207. "v Freeman, A. Myrick, Ml, 1979, The Benefits of Environmental Theory and Practice, Baltimore: Johns Hopkins Press for Resources for the Future, Inc., 1979. Guthrie, K. M., 1974, Process Plant Estimating, Evaluation, and Control , Solana Beach, California: Craftsman Book Company of America, 197T! Just, Richard E., Darrell L. Hueth, and Andrew Schmitz, 1982, Applied Wel- fare Economics arid Public Pqljcy., Englewood Cliffs, New Jersey: Prentice Hall, 1982. Larson, S., 1981, Economics .Analysis of Water Quality Improvementsi Missis- sippi River, Minjiea|jgljs-St.^PauJ Twin Cities Area, Minnesota Pollution Control Agency, Roseville, Minnesota7~T§8T. Lind, R. C., K. J. Arrow, G. Corey, P. Dasgupta, A. Sen, t. Stauffer, J. E. Stiglitz, J. Stockfish, and R. Wilson, 1982, Discounting for Time and Risk in Energy Policy, Baltimore: Johns Hopkins, 1982. R. S. Means Company, Inc., 1981, Building Construction Cost Data, Duxbury, Massachusetts: R. S. Means Company, Inc. , 1981. Michel, Robert L., "EPA Municipal Wastewater Treatment Plant Operation and Maintenance Cost Index," U.S. Environmental Protection Agency, Washington, D.C., issued periodically. /" Michel, Robert L., "EPA Sewage Treatment Plant and Sewer Construction Cost Index, U.S. Environmental Protection Agency, Washington, D.C., issued periodically. Mitchell, R. C., and R. T. Carson, 1981, Ah Experiment in Determining Will- ingness to Pay for National Water Quajjty I improvements , draft report prepared for U.S. Environmental Protection Agency, Resources for the Future, Inc., Washington, D.C., June 1981. Municipal Finance Officers Association and Peat, Marwick, Mitchell and Co., 1982, Financial Capability Guidebook, draft, prepared for U.S. Environ- mental Protection Agency, Washington, D.C., May 1982. 7-2 ------- I 1 1 i i i i i i i '—! i i i i i I i i Rae, Douglas A., 1981a, Visibility Impairment at Mesa Verde National Park: An Analysis of Benefits and Costs of Controlling Emissions jn_ the Four Corners Area, prepared for the Electric Power Research Institute, Charles River. Associates, Cambridge, Massachusetts, 1981. Rae, Douglas A., 1981b, Benefits of Improving Visibility at Great Smoky National Park, draft report prepared for Electric Power Research Insti- tute, Charles River Associates, Cambridge, Massachusetts, December 1981. Richardson Engineering Services, Inc., 1977, Process Plant Construction Esti- mating Standards: The Richardson Rapid System. Solana Be€ch, Cali- fornia: Richardson Engineering Services, Inc., 1977. Rowe, R. D., and L. G. Chestnut, 1981, Visibility Benefits Assessment Guide- book, prepared for U.S. Environmental Protection Agency, Abt West, Denver, Colorado, March 1981. X^ Schulze, W. D., R. C. d'Arge, and D. S. Brookshire, 1981, "Valuing Envi- ronmental Commodities: Some Recent Experiments," Land Economics, Vol. 57, No. 2, May 1981, pp. 151-73. Smith, V. Kerry, 1975, "Travel Cost Demand Models for Wilderness Recreation: A Problem of Non-Nested Hypotheses," Land Economics, Vol. 51, May 1975, pp. 103-11. U.S. Department of Labor, Bureau of Labor Statistics, 1982, Producer Prices and Price Indexes Data for August 1982, Washington, D.C., 1982. U.S. Environmental Protection Agency, 1976, Areawide Assessment Procedures Manual, Vol. Ill, EPA-600/9-76-014, U.S. Environmental Protection Agency, Cincinnati, Ohio, 1976. U.S. Environmental Protection Agency, Office of Water Program Operations, Office of Research and Development, 1980a, Innovative and Alternative technology .Assessment Manual, LD-53, U.S. Environmental Protection Agency, Washington, D.C., February 1980. U.S. Environmental Protection Agency, Effluent Guidelines Division, 1980b, Development Document for Effluent Limitations Guidelines and Standards for the Pulp, Paper and the Builder's Paper and Board Mijjs^ Point Source Categories, EPA 44Wl-80/025-b, U.S. Environmental Protection Agency, Washington, D.C., December 1980. U.S. Environmental Protection Agency, 1981, Process Design and Cost Estimat- ing Algorithms for the Computer Assisted Procedure for Design and Eval- uation of Wastewater Treatment Systems (CAPDET), U.S. Environmental Protection Agency, Washington, D.C., January 1981. Vaughan, W. J., and C. S. Russell, 1982, "Freshwater Recreational Fishing: The National Benefits of Water Pollution Control,'1 Washington, D.C.: Resources for the Future, Inc., November 19§2T 7-3 ------- |