Studies in Risk and Uncertainty – serie

Economic values are increasingly used in policy analysis and legal settings. With the growing recognition that many of the things that benefit or harm people are outside the market system, have come increasing efforts to develop nonmarket valuation techniques. One such technique is the contingent valuation method (CVM). CVM seeks to value environmental and other nonmarket goods and services by asking individuals about their values using survey methods. These procedures are different from the "revealed-preference" methods that economists have historically employed to estimate economic values. The text questions the reasons for departing from well-established revealed-preference procedures and applying a "stated-preference" method like CVM. It argues that for nonmarket goods and services, revealed-preference methods have two shortcomings that those applying CVM hope to avoid. First, revealed-preference methods involve econometric problems that have yet to be fully overcome. The second shortcoming of revealed-preference methods is that such methods, when applied to environmental amenities, are likely to be only partial measures of value.Given the tremendous interest that exists in economic values and the limitations of revealed-preference methods, it is not surprising that interest in CVM has grown rapidly. This study reviews the application of CVM and compares American experiences in nonmarket evaluation with those in other countries.

CLAUDEJEANRENAUD NILS SOGUEL Smoking is a very common habit all over the world. The prevalence rate ranges from 20% - 40% in industrialised countries, and is dramatically increasing in the developing world. Smoking is risky and there is ample scientific evidence to support this statement. We know that smoking is a major cause of disease and premature death, in view of the fact that 3 million people die each year worldwide as a result of their smoking habit. Twenty years ago, the U. S. Surgeon General identified smoking as the single most important cause of morbidity and premature death (USDHEW, 1979). Tobacco consnmption reduces life expectancy vastly. Epidemiological research shows that people who have died from a smoking-related disease would, on average, have lived for an additional 15 years had they not been smokers (Warner, 1987). The economic analysis of tobacco consumption is a complex and challenging issue, which entails addressing many different questions. What is the economic burden of smoking and do smokers "pay their way"? How do individuals perceive their own health risks? What is the effect of the addicting properties of nicotine on the behaviour of a rational, utility­ maximizing individual? Lastly, what is the most effective way to discourage tobacco consumption? In this context, the assessment of the social burden of smoking using a cost-of-illness framework has played a central role since the beginning of the seventies.

Growth of the electric industry Electric generating stations and the associated transmission and distribu­ tion systems are high ticket items as are the costs of fuel and of operating the industry. The country presently spends about $150 billion annually on electricity. It is the country's largest industry, with assets of over $400 billion (Edison Electric Institute Statistical Year book, 1985). The country's electric generating capacity and the end use of electricity grew exponentially for about 70 years, starting at the beginning of the century, with a doubling period of roughly seven years (see figurelO-1). Over much of this period electric utility planning could simply consist of laying a ruler on such a logarithmic plot. The utilities could then, know­ ing plant construction times, write out purchase orders. Improvements in power generating technology allowed electric rates to decline, which permitted the market to expand to accommodate the additional supply. The prospective growth was a heady vision and provided much of the stimulus for the supporters of nuclear power. An example of such extra­ polation made in 1971 is shown in figure 10-2 (The U. S. Energy Problem, 1971). Another forecast (Electrical Power Supply and Demand Forecasts for the United States Through 2050, 1972) projected an installed capacity of 1. 5 million MV(e) in the year 2000, of which 45 percent was to be nuclear. For the year 2050 the installed capacity would have risen to 5. 2 million MW(e), of which 88 percent was to be nuclear.

Traditional utility theory, growing out of the ideas of von Neumann & Morgenstern and Savage, asserts that wise decision makers should maximize some form of expected utility. Decision analysis as a technology implements this prescription. But even after careful thought, people do not necessarily behave that way. The new generalized utility theories attempt to model what people actually do. This book grows out of a NSF-sponsored Conference that brought generalized utlilty theorists and decision analysts together to examine the normative, prescriptive and descriptive implications of the new utility theories. The book begins with a review of the phenomena that the new utility theories are intended to explain and of the theories themselves. It then presents the "old time religion" of utility maximization as a normative and prescriptive theory. It explores how utility maximization needs to be and can be amplified and supplemented for practical prescriptive purposes. The next section of the book looks at what characteristics generalized utility theories would need to have in order to be prescriptively useful. The crucial one turns out to be a form of path independence.Two chapters show that the form of path independence essentially forces the theory embodying it to be a version of traditional utility maximization. The next section of the book looks at the relation between generalized utility theories and the data they are intended to explain. A final section contains an evaluative discussion that weaves the themes of the book together. "Utility Theories: Measurements and Applications" provides a definitive answer to the question about the relation between new utility theories and decision analysis that inspired it. It also brings into focus a number of related questions, and reports a great deal of theoretical and empirical progress on the topics to which it is addressed.

The Conference on "Utility: Theories, Measurements, and Applications" met at the Inn at Pasatiempo in Santa Cruz, California, from June II to 15, 1989. The all-star cast of attendees are listed as authors in the Table of Contents of this book (see p. V), except for Soo Hong Chew and Amos Tversky. The purpose of the conference, and of National Science Foundation Grant No. SES-8823012 that supported it, was to confront proponents of new generalized theories of utility with leading decision analysts com­ mitted to the implementation, in practice, of the more traditional theory that these new theories reject. That traditional model is variously iden­ tified in this book as expected utility or subjectively expected utility maximization (EU or SEU for short) and variously attributed to von Neumann and Morgenstern or Savage. I had feared that the conference might consist of an acrimonious debate between Olympian normative theorists uninterested in what people actually do and behavioral modelers obsessed with the cognitive illusions and uninterested in helping people to make wise decisions. I was entirely wrong. The conferees, in two dramatic straw votes at the open­ ing session, unanimously endorsed traditional SEU as the appropriate normative model and unanimously agreed that people don't act as that model requires. (These votes had a profound impact on my thinking; detail about them and about that impact is located in Chapter 10.

This text examines the impact of environmental concerns and their related legislation on the mining industry. Specifically, the book focuses on two issues facing the mining industry: Superfund liability and global warming. The first issue has a tremendous impact on the sale of properties, insuring operations, developing new properties, joint venture and many other aspects of mining firm operations. The second issue is of interest due to the uncertainty associated with the "greenhouse effect". By focusing on these issues, the volume reveals the tradeoffs between the desire to create and maintain a given level of environmental quality while continuing to produce the goods we use to make our everyday lives more enjoyable.

This study explores the dilemma of siting a high-level nuclear waste (HLNW) repository. The authors examine siting conflicts from a variety of perspectives - political, psychological and sociological - and identify the fundamental determinants of public opposition to waste disposal facilities. In assessing the causes of public opposition, the book draws on various surveys of attitudes toward the repository as a function of predictors such as perceptions of risk, benefits and fairness. Exploring the factors that underlie public opposition to a repository should enable the reader to understand why previous siting efforts may have failed.

This text uses numerical simulations to take an integrated quantitative look at how the various institutions influencing workplace safety lead to the observed levels of illnesses and injuries among US workers. It pieces together the mosaic of interactions among workers, employers, state government, and the federal government that is numerically realistic in the sense of using economists' current knowledge of quantitative linkages. The objective has been to map out how the US economic system determines employment patterns, wages and workplace safety levels.

Since the 1970s, economic theory has extended its field of application to non-market goods such as environmental resources and health. Although it is impossible to assign a price to these goods on the basis of market mechanisms alone, the fact that they have no price does not mean that they have no value. One technique in which economists have shown a marked interest is the contingent valuation method (CVM), which has mainly been used to assign a monetary value to environmental goods. It was first applied to natural resources used for recreational purposes. CVM has been applied to health only recently, so that studies in this field are relatively more scarce than those dealing with the environment, including transport. An overview of experiments conducted in Europe is also included.

This study explores the dilemma of siting a high-level nuclear waste (HLNW) repository. It examines siting conflicts from a variety of perspectives - political, psychological, and sociological - and identifies the fundamental determinants of public opposition to waste disposal facilities as a means of designing more effective approaches to solving the typical siting dilemma. In assessing the causes of public opposition, the book draws on various surveys of attitudes toward the repository as a function of predictors such as perceptions of risk, benefits, and fairness. Exploring the factors that underlie public opposition to a repository enables one to understand why current siting efforts have failed. More importantly, the data are useful in defining what new strategies might be effective in obtaining public consent for a HLNW storage facility. Finally, the book focuses on the problems associated with siting a HLNW repository. The analysis of public opposition and the recommendations made for successful siting can be generalized to almost any attempt to site a noxious facility.

This work is designed to stimulate new developments in decision theory, game theory and general equilibrium theory, as well as in their applications to economics. It is divided into three parts - decision theory, game theory, and the theory of markets. Though its orientation is primarily methodological, some articles are more applied. The consistent use of formal analysis and methodological individualism constitutes the unifying theme of the book. It should be of interest to both students and teachers of microeconomics and game and decision theory.

Can insurance be used as a means to obtain compliance with environmental policy? Answering this question requires examination of a broad mosaic of academic issues, including current systems available for providing compensation and deterrence, use of contracts (including insurance) as substitutes for tort law, limitations of regulatory policy-making by government agencies, pre-conditions for creation of insurance products, and market mechanisms necessary for insurance to be purchased or sold. The purpose of this volume is to highlight the potential role that insurance and performance standards can play in managing environmental risk. Insurance can play a significant role in dealing with one of the most problematic issues facing society today - how to compensate for environmental exposures. This text analyzes the ability of insurance to play a role in managing environmental risk. It begins by outlining the role insurance plays in society in contrast to other societal tools for addressing risk: government benefit programs and imposition of involuntary liability using the court system. By so doing, the book describes the comparative advantages of insurance.The work then analyzes the insurability of the risks. Finally, the book applies the insurability analysis to three concrete environmental examples.

Contingent valuation (CV) measures what is called passive use value or existence value. The CV method has been used to measure the benefits of environmental policy actions. CV measures of economic value rely on choice. In CV studies, choices are posed to people in surveys; analysts then use the responses to these choice questions to construct monetary measures of value. The specific mechanism used to elicit respondents' choices can take a variety of forms, including asking survey respondents whether they would purchase, vote, or pay for a programme or some other well-defined object of choice. It can also be a direct elicitation of the amount each respondent would be willing to pay (WTP) to obtain an object of choice or the amount each respondent would be willing to accept (WTA) in compensation to give it up. This volume is composed of three sections. The first section provides background into the issues underlying the public and academic discussion regarding CV and the reliability of CV estimates of economic value. In addition, this section reviews the theory underlying the measurement of economic value and discusses those aspects of the theory most relevant to CV.The second section focuses on issues that have formed the core of the CV discussions including: sensitivity of WTP estimates to the size of the programme offered, tests for theoretical consistency of CV results, and the sensitivity of results to context and numerous other features of the survey and its administration. The final section addresses the application of CV to actual economic valuation tasks and discusses the types of practical problems the CV researcher will encounter.

For people interested in risk management, medical activity represents a stimulating field of study and thought. On the one hand, progress in medical knowledge and technology tends to reduce the risks to survival that individuals would face in the absence of appropriate diagnostic or therapeutic instruments. On the other hand, new medical technologies simultaneously create their own specific risks, sometimes simply because their effects are less well-known than those of established ones. In a sense any medical progress simultaneously generates new risks while destroying old ones. Moreover, unlike many financial risks that can be either divided or transferred to others (e.g. through diversification, insurance or social security) the personal aspects of medical risks are by essence indivisible and non-transferable. As a result, they are in a sense more threatening than financial risks for risk averse patients. These two facts explain and justify the growing interest in risk economics for the fields of medical decision making and health economics. In Risk and Medical Decision Making, part 1 is developed inside the expected utility (E-U) model and analyses how comorbidity risks affect the well-known "test-treatment" thresholds. Part 2 is devoted to a specific non E-U model with the same purpose: how would one define a threshold in this context and how would one value a diagnostic test? In each of these two parts both diagnostic and therapeutic risks are considered.

The September 11, 2001 terrorism attack on the United States has led government officials to rethink anti-terrorism policies and researchers to assess the implications for the study of risk and uncertainty. This book draws on the expertise of eminent researchers in several risk-related fields to assess three substantive areas of concern - risk beliefs, insurance market effects, and policy responses. The risk belief analyses consider several key questions. How do people think about the risks of terrorism? What are their attitudes toward these risks? To what extent are these low probability and highly dramatic risks overestimated? Several chapters present original survey results analyzing these different aspects of terrorism risk assessments. These studies also begin to explore how people might be willing to sacrifice civil liberties to reduce the risk of terrorism and whether perceived terrorism risks are affected by the severity of the outcome and by proximity to past terrorist attacks. The insurance industry incurred financial losses generated by the terrorism attack. The risks had not been foreseen and were not reflected in insurance pricing. These new terrorism risks generated considerable uncertainty for insurance markets, leading to insurance stock price declines that are documented in this book. Subsequently, a stock price rebound occurred, particularly for the higher quality firms. A third pair of essays deals with policy responses to terrorism risks. A central theme of these analyses is that protective actions by one party have fundamental effects on the risks posed to others. Making airlines immune to terrorist attack may shift the terrorism attacks elsewhere, diminishing the net improvement in security. The papers included here examine how resources should be targeted given these offsetting effects. Contributors to this volume include J. David Cummins, Neil A. Doherty, Baruch Fischhoff, Geoffrey Heal, HowardKunreuther, Cass R. Sunstein, W. Kip Viscusi, and Richard J. Zeckhauser, among others.

This second edition of Stochastic Dominance is devoted to investment decision­ making under uncertainty. The book covers four basic approaches to this process: a) The stochastic dominance (SD) approach, developed on the foundation of von-Neumann and Morgenstem^ expected utiHty paradigm. b) The mean-variance approach developed by Markowitz^ on the foundation of von-Neumann and Morgenstern's expected utility or simply on the assumption of a utility function based on mean and variance. c) The "almost" stochastic dominance (ASD) rules and the "almost" me- variance rule (AMV). No matter whether one employs objective or subjective probabilities, the common stochastic dominance criteria and the mean variance rule may lead to paradoxes: they are unable to rank prospect A w^hich yields $1 with a probability of 0.01 and a million dollars with probability of 0.99, and prospect B which yields $2 with certainty. This is an absurdity as in any sample of subjects one takes, 100% of subjects choose A. The "almost" stochastic dominance criteria and "almost" mean variance rule, which have been recently been developed by Leshno and Levy in 2002^, suggest a remedy to such paradoxes.

The September 11, 2001 terrorism attack on the United States has led government officials to rethink anti-terrorism policies and researchers to assess the implications for the study of risk and uncertainty. This book draws on the expertise of eminent researchers in several risk-related fields to assess three substantive areas of concern - risk beliefs, insurance market effects, and policy responses. The risk belief analyses consider several key questions. How do people think about the risks of terrorism? What are their attitudes toward these risks? To what extent are these low probability and highly dramatic risks overestimated? Several chapters present original survey results analyzing these different aspects of terrorism risk assessments. These studies also begin to explore how people might be willing to sacrifice civil liberties to reduce the risk of terrorism and whether perceived terrorism risks are affected by the severity of the outcome and by proximity to past terrorist attacks. The insurance industry incurred financial losses generated by the terrorism attack. The risks had not been foreseen and were not reflected in insurance pricing. These new terrorism risks generated considerable uncertainty for insurance markets, leading to insurance stock price declines that are documented in this book. Subsequently, a stock price rebound occurred, particularly for the higher quality firms. A third pair of essays deals with policy responses to terrorism risks. A central theme of these analyses is that protective actions by one party have fundamental effects on the risks posed to others. Making airlines immune to terrorist attack may shift the terrorism attacks elsewhere, diminishing the net improvement in security. The papers included here examine how resources should be targeted given these offsetting effects. Contributors to this volume include J. David Cummins, Neil A. Doherty, Baruch Fischhoff, Geoffrey Heal, HowardKunreuther, Cass R. Sunstein, W. Kip Viscusi, and Richard J. Zeckhauser, among others.

For people interested in risk management, medical activity represents a stimulating field of study and thought. On the one hand, progress in medical knowledge and technology tends to reduce the risks to survival that individuals would face in the absence of appropriate diagnostic or therapeutic instruments. On the other hand, new medical technologies simultaneously create their own specific risks, sometimes simply because their effects are less well-known than those of established ones. In a sense any medical progress simultaneously generates new risks while destroying old ones. Moreover, unlike many financial risks that can be either divided or transferred to others (e.g. through diversification, insurance or social security) the personal aspects of medical risks are by essence indivisible and non-transferable. As a result, they are in a sense more threatening than financial risks for risk averse patients. These two facts explain and justify the growing interest in risk economics for the fields of medical decision making and health economics. In Risk and Medical Decision Making, part 1 is developed inside the expected utility (E-U) model and analyses how comorbidity risks affect the well-known "test-treatment" thresholds. Part 2 is devoted to a specific non E-U model with the same purpose: how would one define a threshold in this context and how would one value a diagnostic test? In each of these two parts both diagnostic and therapeutic risks are considered.

Economic values are increasingly used in policy analysis and legal settings. With the growing recognition that many of the things that benefit or harm people are outside the market system, have come increasing efforts to develop nonmarket valuation techniques. One such technique is the contingent valuation method (CVM). CVM seeks to value environmental and other nonmarket goods and services by asking individuals about their values using survey methods. These procedures are different from the `revealed-preference' methods that economists have historically employed to estimate economic values. Why depart from well-established revealed-preference procedures and apply a `stated-preference' method like CVM? For nonmarket goods and services, revealed-preference methods have two shortcomings that those applying CVM hope to avoid. First, revealed-preference methods involve econometric problems that have yet to be fully overcome. The second shortcoming of revealed-preference methods is that such methods, when applied to environmental amenities, are likely to be only partial measures of value. Given the tremendous interest that exists in economic values and the limitations of revealed-preference methods, it is not surprising that interest in CVM has grown rapidly. Environmental Resource Valuation reviews the application of CVM and compares American experiences in nonmarket evaluation with those in other countries.

Decisions, Games and Markets is designed to stimulate new developments in decision theory, game theory and general equilibrium theory, as well as in their applications to economics. The book is divided into three parts - Decision Theory, Game Theory, and the Theory of Markets. Though its orientation is primarily methodological, some articles are more applied. The consistent use of formal analysis and methodological individualism constitutes the unifying theme of the book. Decisions, Games and Markets will be of considerable interest to both students and teachers of microeconomics and game and decision theory.

People want more from the government. One thing they want more of is a sense of personal safety, at home and at work (Regulation, Fall 1991). People also want the government to quit wasting money. The objective of having the government provide a safer life for us and our children at minimum cost leads logically to looking at policy within the system involving the private sector plus governments at the federal and sub federal levels. Using numerical simulations our book takes an integrated quantitative look at how the various institutions influencing workplace safety lead to the observed levels of illnesses and injuries among U.S. workers. Our innovation is piecing together the mosaic of interactions among workers, employers, state government, and the federal government that is numerically realistic in the sense of using economists' current knowl­ edge of quantitative connections. Our objective has been to write a Gray's Anatomy, if you will, of how the U.S. economic system, as tempered by government policy, jointly determines employment patterns, wages, and workplace safety levels.

Over the past two decades, economic theory has extended its field of application to non-market goods such as environmental resources and health. Although it is impossible to assign a price to these goods on the basis of market mechanisms alone, the fact that they have no price does not mean that they have no value. One technique in which economists have shown a marked interest is the contingent valuation method (CVM), which has mainly been used to assign a monetary value to environmental goods. It was first applied to natural resources used for recreational purposes. CVM has been applied to health only recently, so that studies in this field are relatively more scarce than those dealing with the environment, although several valuation methodology surveys are available. There has hitherto been no book which has drawn together and analyzed recent contingent valuations in the field of transport accidents. Contingent Valuation, Transport Safety and the Value of Life provides an overview of the experiments conducted in Europe (Denmark, France, the U.K., Sweden and Switzerland). While a number of contributions are critical, others show how the estimates obtained in the area of road transport can be used to assess the adverse health effects of other causes, such as public transport accidents or air pollution.

Superfund liability). This is an issue that is currently having an dramatic impact on the industry. The impact is being felt in transactions involving the potential sale of properties, insuring operations, development of new properties, joint ventures, or more generally, practically every phase of the mining firms operation. The second issue focuses on an environmental topic that has not been specifically addressed in federal legislation, although it has been indirectly considered, that is global warming or the "greenhouse effect". One of the interesting aspects to this environmental problem is the uncertainty associated with it at every phase of the analysis. The predictions of the general circulation models of climatologists are questioned due to the uncertainty of ocean effects, urbanization, etc. (see Burness & Martin, Chapter 5). The economic models are criticized for the uncertainty associated with the benefit estimates from reducing greenhouse gases, particularly carbon dioxide (COJ, concentrations in the atmosphere as well as estimates of the cost of reducing GHG concentrations and/or emissions. This raises the interesting question of what is the optimal policy and what will be the impact of this policy(s) on the mining sector, given the uncertainty. The first of these two topics is addressed by V. Kerry Smith and Ronald G. Cummings, et al. Professors Smith and Cummings were chosen due to their pioneering work in the area of valuation of nonmarket goods, particularly involving the use of survey methods.

CLAUDEJEANRENAUD NILS SOGUEL Smoking is a very common habit all over the world. The prevalence rate ranges from 20% - 40% in industrialised countries, and is dramatically increasing in the developing world. Smoking is risky and there is ample scientific evidence to support this statement. We know that smoking is a major cause of disease and premature death, in view of the fact that 3 million people die each year worldwide as a result of their smoking habit. Twenty years ago, the U. S. Surgeon General identified smoking as the single most important cause of morbidity and premature death (USDHEW, 1979). Tobacco consnmption reduces life expectancy vastly. Epidemiological research shows that people who have died from a smoking-related disease would, on average, have lived for an additional 15 years had they not been smokers (Warner, 1987). The economic analysis of tobacco consumption is a complex and challenging issue, which entails addressing many different questions. What is the economic burden of smoking and do smokers "pay their way"? How do individuals perceive their own health risks? What is the effect of the addicting properties of nicotine on the behaviour of a rational, utility­ maximizing individual? Lastly, what is the most effective way to discourage tobacco consumption? In this context, the assessment of the social burden of smoking using a cost-of-illness framework has played a central role since the beginning of the seventies.

Can insurance be used as a means to obtain compliance with environmental policy? Answering this question requires examination of a broad mosaic of academic issues, including current systems available for providing compensation and deterrence, use of contracts (including insurance) as substitutes for tort law, limitations of regulatory policy-making by government agencies, pre-conditions for creation of insurance products, and market mechanisms necessary for insurance to be purchased or sold. The purpose of Managing Environmental Risk Through Insurance is to highlight the potential role that insurance and performance standards can play in managing environmental risk. Insurance can play a significant role in dealing with one of the most problematic issues facing society today - how to compensate for environmental exposures. This book analyzes the ability of insurance to play a role in managing environmental risk. It begins by outlining the role insurance plays in society in contrast to other societal tools for addressing risk: government benefit programs and imposition of involuntary liability using the court system. By so doing, the book describes the comparative advantages of insurance. The book then analyzes the insurability of the risks. Finally, the book applies the insurability analysis to three concrete environmental examples.

Contingent valuation (CV) measures what is called passive use value or existence value. The CV method has been used to measure the benefits of environmental policy actions. CV measures of economic value rely on choice. In CV studies, choices are posed to people in surveys; analysts then use the responses to these choice questions to construct monetary measures of value. The specific mechanism used to elicit respondents' choices can take a variety of forms, including asking survey respondents whether they would purchase, vote, or pay for a program or some other well-defined object of choice. It can also be a direct elicitation of the amount each respondent would be willing to pay (WTP) to obtain an object of choice or the amount each respondent would be willing to accept (WTA) in compensation to give it up. This volume is composed of three sections. The first section provides background into the issues underlying the public and academic discussion regarding CV and the reliability of CV estimates of economic value. In addition, this section reviews the theory underlying the measurement of economic value and discusses those aspects of the theory most relevant to CV. The second section focuses on issues that have formed the core of the CV discussions including: sensitivity of WTP estimates to the size of the program offered, tests for theoretical consistency of CV results, and the sensitivity of results to context and numerous other features of the survey and its administration. The final section addresses the application of CV to actual economic valuation tasks and discusses the types of practical problems the CV researcher will encounter.

A management agency --such as a publicly or privately owned electric utility -- must, if it is to be efficient in carrying out its day-to-day tasks, have a means of monitoring its performance to assess the efficiency of its operations and the effectiveness of its planning. For example, how did the demand for electricity compare with that assumed in planning? How effective were the incentives applied to induce energy conservation by users? Such ex post analyses are essential for improving the planning process and hence for improving decisions with respect to efficiency and resource allocation. Unfortunately, it seems to be very difficult for public agencies to make such ex post evaluations an integral part of agency activities, whether the agencies are "producers," e. g. , the Corps of Engineers or the Bureau of Reclamation with respect to water resources management, or are regulatory agencies such as the Environmental Protection Agency or the Food and Drug Administration. Here and there a few ex post analyses of agency programs have been done, but rarely by the responsible agency itself. These analyses have attempted to compare the results actually achieved with the results estimated in planning, either in terms of project outputs or in terms of effectiveness of regulatory and/or economic incentives in inducing desired changes in behavior.