Click for next page ( 252


The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 251
Keeping Pace with Science and Engineering. 1993. Pp. 251-262. Washington, DC: National Academy Press. Environmental Regulation and Technical Change: Overview and Observations I. CIarence Davies The case studies in this volume are a rich trove of information related to environmental regulation and technical change. They provide insight into a wide variety of situations where the policy community and the scientific and engineering community meet or fail to meet. The cases present a spectrum of types of decisions, the most important distinction being between determining what problems should be considered and how to deal with a particular problem. The cases present a variety of enviromental problems, ranging from formaldehyde to acid rain and from trihalomethanes to vegetation in Chesapeake Bay. Across the cases, there are many different actors, each with different values, views, and roles. The cases deal with decisions at the state and local level as well as with deci sions made by the federal government. The state of scientific understanding relevant to each decision varies widely, from a relatively well understood problem like acid rain to a situation like disinfecting by-products in drink- ing water where half the by-products are still unidentified. In all cases, the relevant technical data usually suffer from some degree of uncertainty and therefore are subject to interpretation. Despite the multitude of differing factors between the case studies and within each one, it is possible, if difficult, to draw some tentative generali- zations and make some general observations triggered be the case studies. , == INCORPORATING NEW TECHNICAL INFORMATION The problem that was uppermost in the minds of the Steering Commit- tee when it commissioned the case studies was the extent to which environ 251

OCR for page 251
252 J. CLARENCE DAVIES mental regulations take account of new scientific discoveries. This is a basic problem but, as will be discussed later in this paper, not the only type of problem illustrated by the cases. The cases present a broad range of the times clasped between the dis- covery of new relevant technical information and its incorporation into regulat- ing decisions. Only five years elapsed between the discovery of trihalomethanes in drinking water and regulations designed to reduce their levels, but in most cases it took more time to incorporate new information. The Chesa- peake study estimates the typical lag between discovery and action at "ten years or more" (p. 29~. The ozone study estimates that the lag is "five to fifteen years" (p. 79~. The cases cite many different factors that influence how much time it takes to readjust. Three seem to be of particular importance: the extent to which the new information threatens the status quo; the degree of uncer- tainty of the scientific data; and the involvement of the public or legislators in the decision. All other things being equal, the greater the threat posed by new infor- mation to the status quo the longer it takes to incorporate that new informa- tion into decisions. The status quo can be defined both for the public sector and for the private sector. The public sector status quo consists of all the previous decisions made by the government. The process of making regula- tory decisions typically consumes a good deal of time, money, and political capital. Government regulators are therefore loath to upset decisions that have already been made. The private sector status quo is established by the technological and other investments that the regulated parties made to com- ply with previous policy decisions. Naturally, the firms that have made such investments are reluctant to see the regulatory basis for the investment changed in any way. Both the public and private sectors regard new technical information in the context of the capital political as well as financial that has been sunk into past decisions, and both may be reluctant to make new investments to accommodate new information. As between the public and the private sec- tor, it seems likely that the degree of threat to the public sector will gener- ally be more important in accounting for the time it takes to incorporate new information, although the case studies do not provide good evidence on this. The operational corollary to the status quo hypothesis is that the least delay will occur in regulating new problems because these represent less of a threat to the status quo. An example is formaldehyde, which had not been regulated (except for acute effects) before the Chemical Industry Institute of Toxicology (CIIT) rat study. Formaldehyde as a carcinogen was placed high on the regulatory agenda even before the CIIT study was complete, and it was regulated almost immediately by the Consumer Product Safety Com -

OCR for page 251
OVERVIEW AND OBSERVATIONS 253 mission (CPSC). The CPSC regulation was subsequently overturned by the Court of Appeals, the court stating that CPSC had erred in relaying solely on the CIIT study. The longest delays in incorporating new information will occur in rela' lion to methodological information that threatens a large number of previ- ous decisions. Research on dioxin and later scientific findings about form- aldehyde threaten to require a change in the way the federal government models exposure-response relations for carcinogens. In turn, such a change in modeling threatens all of the standards that have previously been estab- lished for carcinogens. As of this writing, the federal agencies have not accepted any change in the way that they model carcinogenicity, although a large-scale review of dioxin toxicity is under way. The degree of threat to the status quo may be the most important factor in explaining the interval between the discovery of new information and incorporating it in regulatory decisions, but the degree of certainty of the new scientific data and how the data are communicated are also important variables. In general (and not unexpectedly), the more uncertain the deci- sion maker is about the validity of the new information, the longer will be the delay in using the data. Much of the regulatory action (or inaction) described in the case studies takes place against a background of scientific disagreement. Whether it is the interpretation of the formaldehyde epidemiological studies, the dioxin animal data, or the health risk from chlorination, a high degree of uncer- tainty characterizes the state of knowledge. In a case such as dioxin, the primary disputes seem to be between one set of scientists and another rather than between scientists and regulators. In this context the Chesapeake case documents an interesting shift of initiative from managers to scientists as the complexity of protecting the Bay from excess nutrients was increasingly recognized. Although scientific uncertainty may increase the discretion exercised by regulators, in the cases involving compliance with ambient standards (Chesapeake Bay, tropospheric ozone) the complexities of the sources, transport, and effects of various pollutants may lead to increasing reliance on models and on the scientists who construct and run the models. A third significant variable in explaining the time required to adjust regulations to new understanding is the degree of public or legislative in- volvement. However, the relationship between this variable and the length of time to incorporate information is not straightforward. A high degree of involvement seems to produce more rapid regulatory action, but regulators may overreact and the action may be too rapid or the involvement too intense to allow use of the best available scientific information. This seems to have been the case, for example, in the proposal for recycling require- ments as a condition for approving new municipal waste combustors. Other examples not covered in the cases readily come to mind. -

OCR for page 251
254 J. CLARENCE DAVIES Delay does not seem to be influenced by whether the new information would make a regulation more stringent or less. In the dioxin and formalde- hyde cases, information that would make regulation less stringent has not been incorporated. But there have also been tong delays in using new information about tropospheric ozone precursors that would result in more stringent controls. and the same is true of controls for limiting nitrogen _ ,. At, ~ ~ , _ inputs in Chesapeake Bay. The formaldehyde case presents us with a situa- tion where the regulated industry, confronted with a choice between less stringency and the status quo, chooses the latter. ACTING WITHOUT ADEQUATE SCIENCE Delay in incorporating new technical information can result in regula- tory action (or inaction) based on bad science, but regulators may take actions poorly supported by science because information is simply not available or is deliberately ignored. The case studies implicitly use two quite different models of the regula- tory process. One model characterizes decision makers as reactive they delay doing anything until new information pushes them into action. This model highlights the problem of delay in incorporating new information. However, the second model characterizes decision makers as more impul- sive. In this model decision makers need to move ahead, need to make decisions, and they will use whatever science good or bad is available at the time. The authors of the ozone case study describe this problem suc- cinctly: "Information sufficient to resolve key technical uncertainties sel- dom becomes available within the time frame in which policymakers feel compelled to act.... 'tA]n adequate knowledge base' that required to support truly informed decision-making is the pot at the end of the rain- bow: while we may try, we never seem to get there" (p. 85~. Two general factors account for this problem. The first is that decision makers are subject to an array of forces that compel them to take action- statutory deadlines, court mandates, pressure from legislators or constitu- ents. These forces can frequently lead to regulatory timetables that are unconnected to the process of developing needed information. The munici- pal waste combuster case provides a good example of such pressures at work. The other factor involves the limitations on the science base for makin environmental decisions and the uncertainties inherent in our technical un- derstanding of environmental problems. Almost every scientific discipline has some relevance to environmental problems, but environmental problems are relatively new concerns for almost every discipline. Even when ad- equate resources have been made available, it simply takes a long time for new technical information to be developed and verified. Thus, as each of -

OCR for page 251
OVERVIEW AND OBSERVATIONS 255 the cases demonstrates, decision makers frequently must operate with a major gap between what they need to know and what the technical commu- nity can tell them. An interesting and important situation arises when the technical under- standing is available but is incompatible with the existing regulatory frame- work. For example, it is well established scientifically that nitrogen oxides simultaneously contribute to the problems of tropospheric ozone, acid rain, and visibility. However, the regulatory process does not take account of this simultaneity and treats nitrogen oxides separately in the context of each problem because each problem is addressed through a separate regulatory program, which in turn is authorized by separate legislative provisions. It could be argued that the regulatory approach of dividing the world of chemicals into two sharply distinct categories carcinogens and noncarcinogens is also a result of the need for government agencies to try to simplify a com- plex world. POLICY AGENDA NOT BASED ON SCIENCE If actions unsupported by science and engineering are taken because the regulatory agenda is unrelated to the scientific agenda, it becomes necessary to investigate how the agendas are established. The case studies shed a good deal of light on this question, and they also highlight two sets of additional problems: the extent to which the political agenda is not based on science and the extent to which the scientific agenda is distorted by politics or is not responsive to the needs of policymakers. Two major recent EPA reports, Unfinished Business and Reducing Risk, concluded that the enviromental problems that receive the most attention are not the environmental problems that pose the greatest risks to public health or the environment. The political agenda the issues that receive priority attention from government officials has little relationship to an agenda based on good technical information about risks. The case studies prepared for this symposium do not directly address the priority-setting problem, but they do contain numerous insights into how the political agenda is established. They suggest that it is established in a two-step process. First, some new technical information appears; then the new information receives a push from some political force an interest group, a congressional committee, the President. The lesson is that new technical information by itself does not significantly influence the political agenda it must be assisted by some type of political propellant. The case studies contain examples in which scientific information by itself has had some influence. For example, the acid rain case states (p. 171), "The public release of . . . three reports with prestigious scientific imprimaturs was a major reason the Reagan administration felt compelled

OCR for page 251
256 J. CLARENCE DAVIES to initiate limited planning for a national strategy to reduce acid deposition and shifted away, at least symbolically, from exclusive reliance on its re- quirement for further research."2 However, the fact that the Reagan shift was only symbolic may support the two-step hypothesis. The Chesapeake Bay case suggests that physical events can influence the political agenda. lIere, Hurricane Agnes led to the Chesapeake Bay study, which in turn led to political action. Whether a dramatic physical event can substitute for political sponsorship or whether the dynamics of agenda-setting are different at the regional level than at the federal level are unanswered questions. The major role played by the media in interpreting both physical events and technical findings also must be considered. The state of scientific understanding also influences the political agenda by serving as a reality check. In a broad sense, of course, the policy- maker's picture of physical reality is largely determined by scientific and engineering information, although the information may be highly filtered or actually erroneous by the time it reaches the policymaker. A more concrete example of this function of technical information is given in the ozone study, where government agencies were forced to face a variety of policy questions after empirical studies showed that automobile inspection and maintenance systems did not reduce automobile emissions to the degree that had been predicted. Technical information, or its absence, also can keep issues off the po- litical agenda. For example, even though long-range transport of ozone was required to be considered under the provisions of the Clean Air Act, it was not considered or regulated because of the lack of "specific data suitable for air quality modeling and analysis of long-range transport" (p. 68~. THE INFLUENCE OF POLITICS ON THE SCIENTIFIC AGENDA Just as technical information influences the political agenda, so the reverse is true. The scientific agenda is composed of the subjects and types of research undertaken by the technical community. Politics affects the state of technical understanding through the authorization and funding of research and testing programs. Political influence on the scientific agenda at times can be excessive or insufficient. If it is insufficient, the scientific agenda will be nonresponsive to policy needs and the technical information needed by decision makers will be unavailable. If it is excessive, necessary research may not be done, new problems may go unidentified, and the integrity of the scientific process will suffer. The responsiveness (or lack thereof of the scientific community to policy needs was brought sharply to attention by the very large investment made in acid rain research under the National Acid Precipitation Assess

OCR for page 251
OVERVIEW AND OBSERVATIONS 257 ment Program (NAPAP). The acid rain case study concludes that NAPAP had some but not much impact on the 1990 Clean Air Act. The author reminds us that, "the political process, not science per se, dictates how much information is enough as well as the conditions under which informa- tion guides a given policy decision" (p. 183~. Nevertheless, to the extent that NAPAP failed to provide an adequate technical basis for the 1990 amendments to the Clean Air Act there is clearly a problem because this is the purpose for which the NAPAP was established. On a smaller and less visible scale, the NAPAP experience may be a common occurrence. The author of the ozone case comments, "Excessive uncertainty . . . may be a consequence of governmental research programs not being sufficiently long term and consistently focused to provide the information needed to reduce uncertainties to acceptable levels" (p. 83~. The Chesapeake Bay case contains a variant of this problem dealing with large-scale quantitative models: "As the cost of the model increases (in terms of time and money) and the corporate memory is lost, the model begins to take on a life of its own and the predictions become reality. Thus, there is a tendency for the management community to reach the conclusion that additional scientific information is no longer needed." (p. 32~. Two types of political impact on the science and engineering agenda are very important but are not dealt with extensively in the case studies. First is the neglect of monitoring and studies to support the evaluation of govern- ment programs. Government officials tend to place a low priority on evalu- ative studies and this is reflected in the science agenda. For example, the ozone study notes that VOC and NOx levels are not measured and that, "This paucity of information severely limits the ability to evaluate the ef- fectiveness of emissions control programs" (p. 779. The second type of impact, one on which the Steering Committee had hoped to collect information, is that of regulations on the development and adoption of new technologies. The case studies do not shed much light on this question, and what limited light they shed is different in different cases. The authors of the ozone case study state, "regulations generally prompted the development of technologies required for compliance" (p. 86~. How- ever, in the municipal waste case (p. 131), the Clean Air Act standards based on "best demonstrated technology" prevented EPA from forcing the development of new technology. Similarly, the Chesapeake Bay case (p. 30) notes that "Clearly, reliance on a particular technology (secondary treat- ment) as the basis for regulating nutrient inputs has inhibited the develop- ment of alternative (less costly, more effective?) approaches and technolo- gies." One can conclude from these observations in the case studies that per- formance standards (those that just establish the level of compliance to be achieved) encourage new technology while standards that specify a particu -

OCR for page 251
258 J. CLARENCE DAVIES tar type of technology are an impediment to the diffusion, and perhaps to the development, of new technologies. Such a conclusion may well be valid, but the whole area of the relationship between regulation and technol- ogy deserves more detailed and intense exploration. IMPROVING THE PROCESS The authors of the case studies provide a number of suggestions for improving the way the regulatory process relates to scientific and engineer- ing information. Before reviewing these suggestions, it is worth keeping in mind Robert White's basic point that the goal of any improvements must be achieving a balance between often competing forces. In his introduction to this volume, White states (p. 5-6), ". . .as desir- able as it is to have regulations that are based on the best, most current technical understanding, it is also desirable to have a stable regulatory re- gime within which the affected parties can intelligently plan to come into compliance and implement their plans.... We have then two characteris- tics that we all would agree the environmental regulatory system should exhibit: it should respond dynamically to changes in our understanding of the technical aspects of the issues, and it should remain stable on a time scale sufficient for regulated parties to comply with some measure of eco- nomic efficiency. It is evident that these two normative characteristics can be, and frequently are, in conflict." A related point that needs to be made is that the regulatory process involves many considerations besides "science" and "politics." A Manichean view of the process that sees it as a contest between the bad politicians and the good scientists is no more helpful than a view that puts all the emphasis on changing to adapt to new information or a view that only emphasizes stability. There are many other factors that must be taken into account when evaluating or analyzing the regulatory process economics, problems of implementation, legal requirements, public opinion, to name just a few. The process is characterized by numerous actors, and recommendations to improve it must consider multiple values. More research. There can be little dispute that the general state of knowledge regarding environmental problems is often woefully inad- equate. The case studies are, in one sense, simply a series of examples supporting this general statement. Furthermore, the missing knowledge is sometimes fairly elementary for example, the effect of recycling require- ments on emissions from municipal incinerators. An obvious conclusion from the cases is that more research is needed. Because enviromental regu- lations tend to involve information at the cutting edge of technical under- standing, more research is almost always desirable.

OCR for page 251
OVERVIEW AND OBSERVATIONS 259 This conclusion should be accompanied by the caveat, also illustrated by the cases, that more research may not always help in a particular situa- tion. The massive epidemiological study of formaldehyde conducted by Blair, involving 600,000 person-years of data, did not help much in settling the controversy over formaldehyde. Similarly, many of the studies of di- oxin did little to settle that controversy. Several participants in the symposium raised the question of what type of institution should do the necessary research. The basic point made was that the short-term demands of the regulatory process provide an inhospi- table context for conducting long-term high-quality research, and that there- fore it may be useful to separate organizationally the regulators from the long-term programs. Although not raised in the discussion, ensuring that technical information is incorporated into the regulatory process is a com- peting value. The closer the researchers are to the regulators organization- ally, the more likely it is that their research results will be incorporated in regulatory decisions. Balancing these values and exploring organizational innovations to reconcile them is high on the list of issues that warrant further examination. Communication. Several of the cases comment on the importance of communication between scientists and engineers on the one hand and deci- sion makers on the other. For example, the authors of the Chesapeake Bay study state (p. 31) that the lack of communication (in addition to other differences such as differing time perspectives) had the result that "~1) the management community tends to question the relevance of environmental research conducted by an independent science community, and (2) the sci- ence community tends to question the integrity of the management pro- cess." The communication problem takes several distinct forms. Communica- tion may not take place at all; information may be communicated but be- come distorted; or it may be communicated and understood but ignored by the decision maker. Ways of dealing with communication problems may help in all these situations or may help in dealing with only one of them. In-house capacity. It seems likely that a basic variable in facilitating communication is the extent to which the regulatory agency possesses an in-house scientific capability. If no one in the agency understands the technical issues, communication with the technical community obviously will be difficult. Although the federal and state agencies covered by the cases varied widely in their scientific capabilities, the cases themselves do not empha- size this difference. It is interesting that in the formaldehyde case the agency that responded most rapidly to the first evidence of formaldehyde carcinogenicity was the Consumer Product Safety Commission, the agency -

OCR for page 251
260 J. CLARENCE DAVIES with the least in-house technical capability. There may be times when in- house capability is an impediment to action. However, it is also relevant that the CPSC decision was overturned in the courts for lack of scientific support. Science advisers are used by many agencies to facilitate communication of technical information. Richard Morgenstern, the acting deputy adminis- trator of EPA, observes elsewhere in this volume (p. 247) that the type of science advisory mechanism that has served EPA in the past "should be expanded to encourage scientists and engineers to make best judgments on unresolved technical issues." Peer review is another mechanism to both facilitate communication and provide some quality control over the technical information used to make policy decisions. The EPA Science Advisory Board for a number of years has provided some peer review of the science used in regulatory decisions, and the agency recently has moved to stricter requirements for peer review of scientific data. The cases seem to indicate that most of the regulatory agencies make little use of various formal techniques for dealing with uncertainty or other aspects of decisions. In effect, they seem to lack in-house capability in the decision sciences despite the fact that they are repeatedly faced with the situations and problems for which the methods of decision science were developed. The intergovernmental dimension. In the U.S. federal system, one fac tor affecting the relationships between the technical community and the regulators may be the level of government that makes the regulatory deci- sion. The cases contain a good deal of information about interactions among the different levels of government but, typically, the cases deal with differ- ent types of interactions in different circumstances and suggest different conclusions. In the municipal waste combuster case, the authors conclude: "Our findings suggest that the type of technical and scientific information used may be different at each level of government . . . local regulatory agencies have the most to gain by seeking and requiring the most stringent air pollu- tion control and the highest standards for operation of facilities, regardless of actual risks associated with these facilities" (pp. 128-129~. The authors of the Chesapeake Bay case note the same disregard of scientific data but draw an opposite conclusion with respect to stringency: "In contrast to the perspective of federal officials and reports by local scientists and citizens' groups, state officials in Maryland insisted that the Bay was doing just fine .... Thus, the governing body responsible for the implementation of nutrient control plans, the State, was least receptive to scientific evidence indicating the early stages of Bay-wide eutrophication" (p. 15~. In both cases, roles are reversed at different stages sometimes the federal govern

OCR for page 251
OVERVIEW AND OBSERVATIONS 261 ment is most receptive to new information and urges the most stringent standards, sometimes the state is most receptive and stringent. The authors of the ozone case study observe that, "where California has assumed a leadership role, it seems to have encountered fewer barriers to action, acceptance, and implementation than has the federal government" (p. 86~. It seems doubtful that state officials in Maryland or Virginia re- sponsible for the Chesapeake Bay would agree. About the only certain conclusion that can be drawn is that the intergovernmental dimension is an important variable in understanding the regulatory process. Courts. Surprisingly, the cases deal very little with the role of the courts, although many environmental regulatory decisions are finally de- cided by litigation, as are the schedules for administrative action. The courts have their own difficulties dealing with technical information, and in recent years experiments have been performed using scientific advisers to judges, panels of expert technical witnesses that can be used by a judge, and other methods to facilitate the objective use of technical knowledge in the courtroom setting. Sequencing. Robert White notes that, "We need to build into the struc- ture of the regulatory system means for reconsidering earlier decisions if and when our understanding changes sufficiently to call earlier decisions into question" (p. 51. The same theme is sounded in the ozone case (p. 85~: "Where circumstances require action, such as smog conditions in the South Coast Air Basin, the waiting time for research results exceeds the time practically available for taking the actionts). Resolution of this dilemma is exceedingly difficult. One option is to design actions that can be carried out in sequence, instituting more stringent controls with time, as needs warrant. The results of research can then influence the 'action sequence' as they become available." Policies can be structured to allow the introduction of new technical information. Although the requirement that National Ambient Air Quality Standards be reviewed every five years has not been adhered to, neverthe- less it has served as a prod for the regulators to consider new technical findings. The provisions of the Vienna Convention on Stratospheric Ozone have worked effectively to adjust regulatory requirements rapidly to chang- ing scientific understanding. Regulatory negotiation. Two of the case studies suggest that formal negotiation among interested parties to develop a regulation can facilitate incorporation of technical information. The authors of the municipal waste combustor case state, "If negotiations were to become a standard compo- nent of a regulatory process, stakeholders would have a more controlled and focused opportunity to provide EPA with information in an atmosphere of cooperation" (p. 1351. However, the cases do not actually show whether -

OCR for page 251
262 J. CLARENCE DAVIES this is true in practice. Regulatory negotiation was not used in the combus- tor case. It was used to deal with a rule on the by-products of water disinfection, but the case study was completed before the accomplishments of the negotiation could be evaluated. Several other groups are examining proposals to improve the use of technical information in environmental regulation and policy. Organiza- tions such as the National Academies of Sciences and Engineering are a continuing presence in this area while groups such as the Carnegie Commis- sion on Science, Technology, and Government and the World Wildlife Fund's National Commission on the Environment have recently focused on the connection between technical information and policymaking. The attention that the issue is receiving should result in concrete steps toward improving the science-policy relationship. CONCLUSION Robert White writes in his introduction: "The primary question of policy is this: Does the current environmental regulatory system strike an appro- priate balance between dynamic change and stability?" (p. 6~. The case studies generally indicate that the answer may be 'no,' although as I have tried to show, the question is in reality very complex and does not lend itself to simple answers. As important as any overall judgment is an understanding of how the regulatory process actually deals with technical information. Such analysis leads to other questions. For example, the authors of the formaldehyde case state: "Perhaps the ultimate issue . . . is who decides when new science should be incorporated into the regulatory process" agency risk assessors, risk managers, external scientific panels, Congress, the courts? (p. 217) The case studies provide a good starting point for considering this and related questions. They move us away from rhetoric and toward an empiri- cal base for addressing some of the most important underlying questions of environmental policy. NOTES 1. See U.S. Environmental Protection Agency, Office of Policy Analysis, Unfinished Busi- ness (Washington, D.C., February 1987); and U.S. Environmental Protection Agency, Science Advisory Board, Reducing Risk (Washington, D.C., September 1990). 2. Editor's note: The three reports referred to here are William A. Nierenberg et al., Report of the Acid Rain Peer Review Panel (Washington, D.C., Office of Science and Technology Policy, 1984); National Research Council, Atmosphere-Biosphere Interactions: Toward a Bet- ter Understanding of the Ecological Consequences of Fossil Fuel Combustion (Washington, D.C., National Academy Press, 1981); and U.S. Congress, Office of Technology Assessment, Acid Rain and Transported Air Pollutants: Implications for Public Policy (Washington, D.C., Government Printing Office, 1984).