assignment 4

Policy Evaluation

When it is viewed as a cycle or sequential pattern of functional activities, the final phase of the policy process is evaluation. As much an art as a science, policy evaluation involves the estimation, appraisal, or assessment of a policy, its content, implementation, goal attainment, and other effects. The policy evaluator wants to know whether and to what extent a policy has accomplished its goals or whether it has had other effects, intended or unintended.

Policy evaluation may also seek to identify factors that contributed to the success or failure of a policy. This, in turn, may lead to recycling of the policy process (problem definition, formulation, adoption, and so on) in order to continue, modify, strengthen, or terminate the policy. Put differently, information gained through evaluation feeds back into the ongoing policy process.

As a functional activity, evaluation can occur at any point in the policy process, not simply after some effort has been made to implement a policy. Thus, attempts are made to determine prospectively—to estimate or predict—the likely effects, or the costs and benefits, of policy alternatives prior to their adoption. Typically, however, policy evaluation “looks backward” to what has happened whereas the other stages of the policy process look forward to attaining a goal.1

Before proceeding further in examining evaluation, we need to pause and look at policy impacts, at the general sorts of effects or consequences that may ensue from policies.

Policy impact

To begin, we need to distinguish policy outputs and policy outcomes. Policy outputs are the things actually done by agencies in pursuance of policy decisions and statements. The concept of outputs focuses attention on such matters as amounts of taxes collected, miles of highways built, welfare benefits paid, price-fixing agreements prosecuted, traffic fines collected, or foreign-aid projects undertaken. Outputs usually can be readily counted, totaled, and statistically analyzed. Examining outputs may indicate, or seem to indicate, that a lot is being done to implement a policy. Such activity, however, sometimes amounts to nothing more than what Professor William T. Gormley Jr. calls “bean counting.” Agencies, under pressure from legislators, interest groups, and others to demonstrate results, “may focus on outputs, not outcomes, in or- der to generate statistics that create the illusion of progress.”2 If the percentage of students graduating from universities in a state increases, does this tell us anything about the quality of education that they are receiving?

Policy outcomes (sometimes called “results”), in contrast, are the con- sequences for society, intended and unintended, that stem from deliberate governmental action or inaction. Social-welfare policies can be used to illustrate this concept. It is fairly easy to measure welfare-policy outputs such as amounts of benefits paid, average level of benefits, and number of people assisted. But what are the outcomes, or societal consequences, of these actions? Do they increase personal security and contentment? Do they reduce individual initiative? Did Aid to Families with Dependent Children, now Temporary Assistance to Needy Families (TANF), have the effect of encouraging promiscuity and illegitimacy, or teenage pregnancies, as some alleged? Do welfare programs help keep the poor quiescent, as others contend?3 Questions such as these, which are tough to answer, direct our attention to the societal effects of policies. Among other things, policy students should want to know whether policies are accomplishing their intended purposes, whether society is changing as a consequence of policy actions and not because of other factors such as private economic decisions, and whether it is changing as intended or in other ways. Policy impacts are an amalgam of outputs and outcomes.

The impact of a policy may have several dimensions, all of which should be taken into account either in the conduct of a formal evaluation or in the course of an informal appraisal of the policy.4 They include the following:

1. Policies affect the public problem at which they are directed and the people involved. The target populations whom the policy is intended to affect must be defined, whether they are the poor, small-business people, disadvantaged schoolchildren, petroleum producers, or whoever. The intended effect of the policy must then be determined. If it is an antipoverty program, is its purpose to raise the income of the poor, to increase their opportunities for employment, or to change their attitudes and behavior? If some combination of such purposes is intended, analysis becomes more complicated because priorities should be assigned to the various intended effects. Typically, policies accomplish at least a portion of their goals or objectives.

Further, a policy may produce either intended or unintended (unforeseen or unplanned) consequences, or even both. Unemployment compensation may improve the economic situation of the unemployed, as designed by its proponents, but it may also cause them to delay in finding new jobs. A public- housing program may improve housing conditions for urban blacks, but it may also be so administered as to contribute to racial segregation in housing and neighborhood schools. The agricultural income and price support programs, intended to enhance farmers’ incomes, also may lead to overproduction of supported commodities (only a few such as corn, wheat, and cotton are covered), to higher food prices, and to unearned increases in land values.

The causes of unintended policy consequences are multiple: the complexity of the issue being addressed; insufficient information on which to ground a decision; pressure to act quickly on a matter; too little consultation with other legislators or affected persons; partisan pressures and the politicization of issues; and, finally, human error. Also, statutory language needs mention. A law may be clumsily drafted, or too detailed, or too vague, and thus permit unintended interpretations.5

A good illustration of a policy with negative unintended consequences was financial deregulation (or non-regulation). The Financial Services Modernization Act (1999) permitted banks to combine commercial banking (deposits and loans), investment banking, and insurance services. Huge banking conglomerates developed, some of which became viewed as “too big to fail” because of their potential effects on the economy. The Commodity Futures Modernization Act (2000), which was slipped into a huge omnibus appropriations bill and enacted with no debate, prohibited regulation of credit default swaps (don’t ask) and other financial derivatives. These statutes, and decisions by regulatory officials not to use the authority they did possess, because of a belief in self-regulating markets, contributed greatly to the near financial collapse of the American economy and the Great Recession.

2. Policies may affect situations or groups other than those at which they are directed. These are variously called third-party effects, spillover effects, or externalities.7 The construction of urban interstate highways is of much benefit to motorists and trucking companies. However, they also cause inconvenience, disorder, and social disruption for the neighborhoods through which they run, and they have helped produce urban sprawl. Clear-cutting in national forests, which is of benefit to timber companies and in line with the perspective of those who view forests as tree farms, is profoundly disturbing to environmentalists, nature lovers, and many sportsmen and women because it results in the destruction of wildlife habitat, the loss of aesthetic value, and the siltation of trout streams.

These two examples portray negative externalities, but externalities may also be positive. Public-education programs not only educate students but also provide employers with a more capable workforce and the community with better-informed citizens. A strong line of research supports a correlation between education and support for democracy. Those who contend that only those who have children in public schools should contribute toward their support ignore such positive externalities. Although pollution-control programs impose costs on many industries, they are a boon to the manufacturers of pollution-control equipment. Many of the outcomes of public policies can be most meaningfully understood as externalities that impose costs or provide benefits for third parties.

3. Policies have consequences for future as well as current conditions; for some policies most of their benefits or some of their costs may occur in the far future. Was the Head Start program—a preschool education program for the poor—supposed to improve participating children’s short-term cognitive abilities or their long-range development and earning capacity? Did regulation of the field price of natural gas, a policy initiated in the 1950s and extending into the 1980s, really produce a shortage of gas in the 1970s, as some contended (notably petroleum-industry officials and their supporters, who had long been critics of price regulation)? Are we better off under deregulation? The future effects of some policies may be very diffuse or uncertain. Assuming that patent and copyright policies do indeed stimulate invention and creativity, and that these activities in turn enhance economic growth and societal development, how does one measure their benefits, either qualitatively or quantitatively? Again, how does one appraise (with reasonable objectivity) the effects of the National Foundation on the Arts and the Humanities’ support for literacy, artistic, and museum activities? Would the elimination of such policies as these have deleterious consequences for American society? 4. Just as policies have positive effects or benefits, they also entail costs. Economists seem never to tire of telling us that there is no such thing as a free lunch. Costs come in different forms. First, there are the direct costs for the governmental implementation of a policy or program. These are usually fairly easy to calculate, whether stated as the actual amount of money spent on the program, its share of total governmental expenditures, or the proportion of the gross domestic product devoted to it. Budgeting documents will yield such figures. If, however, a governmental expenditure serves multiple purposes, such as operating the space program and developing new technology, the allocation of costs becomes more perplexing.

Direct costs also include private expenditures that are necessary in order to comply with public policies, such as those for industrial health and safety and environmental-pollution controls. These may be more difficult to discover or calculate. Moreover, it is possible that some companies would have installed protective devices in the absence of policy, whether out of a desire to serve the public interest or to protect themselves from lawsuits. Should their costs then be assigned to the policy? In the absence of governmental subsidies, the direct costs of complying with regulatory policies initially fall primarily on the regulated, who have an ideological incentive to inflate claimed costs, deliberately making the policies appear more burdensome. Ultimately, of course, most compliance costs are likely to be shifted to consumers in the form of higher prices for goods and services.

This brings us to indirect costs. Public policies may cause reduced production, higher prices, or mental anguish or distress. Expenditures to meet coal-mine-safety requirements may cause a reduction in mine output. People called for jury services typically receive compensation that does no more than cover commuting costs. The consequence is lost wages, or lost production, or both. Many public policies have indirect or social costs of these sorts. Mental anguish may occur when one’s hometown is flooded by a new impoundment. Financial compensation may be paid for one’s childhood home, but what of the loss that occurs because, being under twenty feet of water, it can no longer be visited? The concept of indirect cost often calls for putting a price tag on intangibles. They tend to defy precise measurement.

Finally, there are opportunity costs, a concept that rests upon the fact that we cannot do everything at once and that resources used for one purpose (e.g., flood control) cannot be used for another purpose (e.g., public housing). “Opportunity cost is a decision making rather than an accounting concept.”8 It focuses attention on what one has to give up (or, alternatively, what one will gain) if resources are used for one purpose rather than another. Thus, economists argue that the United States’ all-volunteer army will not attract needed recruits if the opportunity costs of military service are kept below those of civilian life. When the ranks of the military were largely filled by draftees, less concern had to be given to opportunity costs and the compensation of service personnel.9

5. The effects of policies and programs may be either material (tangible) or symbolic (intangible). The consequences of symbols are both important and hard to measure. Symbolic outputs, according to political scientists Gabriel A. Almond and G. Bingham Powell include “affirmations of values by elites; displays of flags, troops, and military ceremony; visits by royalty or high officials; and statements of policy or intent by political leaders.” Consequently, they “are highly dependent on tapping popular beliefs, attitudes, and aspirations for their effectiveness.”10

Symbolic policy outputs produce no readily discernible changes in societal conditions. No one eats better because of a Memorial Day parade or a stirring speech by a high public official on the virtues of free enterprise, however ideo- logically or emotionally satisfying such actions may be for many people. More to the point, however, policy actions ostensibly directed toward meeting mate- rial wants or needs may turn out in practice to be more symbolic than material in their effect.

This shift in policy tone is well illustrated by the Fair Housing Act of 1968. Enacted by Congress in part because of pressure created by the assassination of Dr. Martin Luther King Jr., the 1968 law prohibited discrimination in the rental or sale of housing because of race, color, religion, sex, or national origin. How- ever, the Department of Housing and Urban Development (HUD), which was assigned primary responsibility for its enforcement, was authorized only to me- diate disputes between a person who thought he or she had been discriminated against and the renter or seller. The Justice Department, in turn, could not act unless it found a “pattern or practice” of discrimination. As a consequence of these weak enforcement provisions, the Fair Housing Act in practice did not live up to its promise. The act, which one member of Congress called a “tooth- less tiger,” was of little use in preventing discrimination in housing.

The Congress in 1988 reached a compromise agreement on legislation to strengthen enforcement of the Fair Housing Act. Now a person who believes heor she has been discriminated against can file a complaint with HUD, which, if it cannot settle the dispute, can issue a charge of housing discrimination. At this point either party to the dispute can choose to have it decided by either a federal district court or an administrative hearing. If either party chooses to go to the federal court, that choice prevails. This procedure was intended to put some real “teeth” into the enforcement of the act and give it material rather than merely symbolic effect. It has been partially successful, but complaints have been far fewer than actual incidents of discrimination.

Studies indicate that presidential level support for fair housing has varied from Johnson to Obama. So have enforcement actions by federal, state, and local agencies. Though some forms of discrimination have lessened, overall housing discrimination remains a problem in many cities and suburbs. Housing segregation and discrimination contribute to other problems—for example, segregated schools, limited job opportunities, and exposure to high crime rates in minority areas.11

Other public policies that appear to promise more symbolically than their implementation yields in material benefits include antitrust activity (especially as that involves “trust busting”), public-utility rate regulation, and equal employment opportunity. Even though the actual effect of a policy may be considerably less than is intended or desired, it nonetheless may have significant consequences for society. An antipoverty program that falls short of its mark may provide some assurance to people that the government cares about poor people and wants to reduce poverty. Legislation on equal employment opportunity informs people that their government, officially at least, does not con- done discrimination in hiring on the basis of race, sex, or nationality. Apart from the effects such policies have on societal conditions, they may contribute to the maintenance of social order, support for government, and personal self- esteem, which are not inconsequential considerations.

The analysis and evaluation of public policy is usually focused upon what governments actually do, why, and with what material effects. We should not, however, neglect the symbolic aspects of government, despite their intangible and nebulous nature. The rhetoric of government—what governments say, or appear to say—is clearly a necessary and proper subject for the policy analyst.

Policy Evaluation

Policy evaluation, as a functional activity, is as old as policy itself. Leg- islators, administrators, judges, pressure-group officials, media commentators, and citizens have always made judgments about the worth or effects of particular policies, programs, and projects.

The Boston Tea Party, for instance, is often portrayed as an unfavorable evaluation of one of King George’s colonial tax policies. In actuality, those most threatened were colonial merchants who sold tea smuggled into the country and enjoyed large profits. They were the organizers of the Tea Party, not consumers.

Many policy judgments have been of an impressionistic or intuitive variety, based at best on anecdotal or fragmentary evidence, and strongly influenced by ideological, partisan, or idiosyncratic valuational criteria. Thus, staunch conservatives may regard public-housing programs as socialistic and in need of repeal, regardless of their causes, necessity, or consequences; or Democrats may support higher taxes on corporations and the rich because they believe this will enhance their electoral opportunities. Unemployment compensation may be deemed a “bad” program because the evaluator claims to know “a lot of people” who improperly receive benefits.

Apocryphal stories about “welfare queens” who drive Cadillacs to collect their welfare checks are commonplace among welfare critics. Most of us are familiar with this style of policy evaluation and have perhaps enjoyed doing a bit of it ourselves. Much conflict results from this sort of evalua- tion, however, because different evaluators, depending upon their values or criteria, may reach sharply divergent conclusions on the merits of the same policy. The food stamp program is a recent example of this.

Another form of policy evaluation centers on process, on the operation or administration of a policy or program.12 Although the terms policy and pro- gram are often used interchangeably, a line can be drawn between them. Policy is defined in this book as a course of action followed by government in dealing with some problem or matter of concern. A policy may spawn a variety of programs, a program being defined as a set of rules, routines, and resources intended to carry out a policy or a portion thereof. Programs, which give “direct, specific, and concrete expression” to the policies on which they are based, sometimes are marked out in legislation. More likely, however, they will be the handiwork of administrative officials. Thus, the Antitrust Division has created several programs to enforce the antitrust laws, such as the competition advocacy and anti-cartel enforcement programs.

A program, or several programs, conceivably could be eliminated without impairment of the underlying policy. In 2003, the Bush administration pro- posed eliminating several fairly specifically focused education programs be- cause they were deemed unnecessary, duplicatory, or replaceable. However, the broad, underlying policy of federal support for public elementary and second- ary education remained firmly in place. Most of the targeted programs were retained by Congress.

Questions asked about how a policy or program is being run may include: What are its financial costs? Who receives benefits (payments or services) and in what amounts? Is there unnecessary overlap with or duplication of other programs? Are legislatively prescribed standards and procedures being fol- lowed? Is the program honestly administered? This kind of evaluation, which may involve much monitoring of agencies and their officials, will tell us some- thing about whether there is honesty or efficiency in the conduct of a program, but like the first kind of evaluation, it will probably yield little or nothing in the way of hard information concerning the societal effects (outcomes) of a program. On the other hand, process evaluation is often helpful to program managers wanting to improve the administration of their programs and reduce their potential for political criticism.

A third, comparatively new type of policy evaluation, which has been get- ting increasing attention in the national government in recent decades, is the systematic and intendedly objective evaluation of programs. This form of evaluation, which I will refer to as systematic evaluation, employs social science methodology to measure the societal effects of policies or programs and the extent to which they are achieving their goals or objectives.

The Departments of Labor, Health and Human Services, and Energy, for instance, have assistant secretaries whose responsibilities include program evaluation. Bureaus within these and other departments often include policy- and program-evaluation staffs. Moreover, they enter into many contracts with private research organizations, university scholars, and others for the performance of evaluation studies. More attention appears to be given to evaluating social programs (welfare, education, health, and nutrition) than most other areas of governmental activity. This preference probably arises from the proliferation of social programs in recent decades, their substantial financial costs, the controversies that swirl around them, and the dislike of “welfare.”

Systematic evaluation seeks information on the impact of a policy or program on the public need or problem at which it is directed. Utilizing particularly the talents of social scientists (economists, political scientists, psychologists, and sociologists), it involves the specification of goals or objectives; the collection of information and data on program inputs, outputs, and consequences; and their rigorous analysis, preferably through the use of quantita tive or statistical techniques.

Evaluation researchers use a variety of evaluation research designs, which are often coupled with high-powered statistical and mathematical techniques. The three evaluation designs discussed here—the experimental design, the quasi-experimental design, and the before-and-after study—are basic.13 An ex- amination of them will convey a notion of the problems involved in setting up evaluation studies.

The experimental design is the classic method for evaluating a policy or program. Two comparable groups—an experimental, or treatment, group and a control group—are randomly selected from the target population. The ex- perimental group receives treatment through a policy or program; the control group does not. Pretests and posttests of the two groups are used to determine whether changes, such as improved reading scores or lower incidences of a disease, have occurred in the two groups. If the performance of the experimen- tal group is significantly better than that of the control group, the program is held to be effective. A high level of validity and generalizability is accorded the results of experiments. An example is in order.

Several years ago, the Delaware Department of Labor conducted a field experiment14 to assess the effectiveness of various activities in helping “dis- located workers”—“persons who have lost long-term, stable jobs due to an in- creased international competition and/or changing technology.”15 The goals of the program for dislocated workers, which was funded through the Job Part- nership Training Act, were to help increase workers’ earnings and reduce their need for unemployment compensation benefits. A coterie of 175 workers with comparable characteristics was identified; 65 were randomly assigned to the treatment group, with the other 110 becoming the control group. Members of the treatment group were provided counseling on job-search activities, assistance in locating openings, retraining, and other services. Both groups were monitored for a year. Comparison of their performance revealed that the con- trol group did better than the treatment group in meeting the program’s goals. It was concluded that “the program did not appear to improve participants’ job prospects.”16

Use of the experimental design may not be possible because of costs, time, and ethical or other considerations. The quasi-experiment, then, may be a useful alternative. The process of random selection is not used. Rather, the treatment group is compared with another group (a “comparison group”) that is similar in many respects. Consequently, in the quasi-experiment there is greater likelihood that the performances of the two groups could be influenced by their internal characteristics rather than the program treatment. Quasi-experiments nonetheless are still seen as quite useful for many purposes.

A well-known quasi-experiment involves the Connecticut highway speed- control program.17 Following a record number of highway traffic fatalities, the state initiated a crackdown on speeding. Initial data indicated that the enforcement program had significantly reduced fatalities. It was possible, however, that this resulted from other factors, such as differences in the weather or more safe cars in operation. To control for such possibilities, the Connecticut high- way fatality rate per 100,000 people was compared with that of neighboring states where there had been no enhanced enforcement program. This showed that Connecticut’s fatality rate was lower than those of the other states, thus supporting the inference that the Connecticut crackdown had a positive effect in reducing traffic deaths.

The before-and-after study of a program compares the results of a program after a period of implementation with the conditions existing prior to its inception. Thus, one might compare the quality of water in a river before and after a pollution-control program was put into effect. Before-and-after studies often have low costs and take less time to conduct. A major drawback, however, is that the changes that occur are open to rival explanations. Improved water quality in our imaginary river could be due to increased flow, voluntary action by polluters, or an economic recession’s having caused reduced industrial activity. On the other hand, if a before-and-after study finds little change in the desired direction, then it is likely that the program is not having much effect.18 Given all of this, it is still possible for before-and-after studies to produce useful information about a program that otherwise would be unavailable.

As this discussion indicates, systematic evaluation draws on experience in assessing the effects a policy or program has on the public need or problem at which it is directed. It permits at least tentative, informed responses to such questions as: Is this policy achieving its stated objectives? What are its costs and benefits? Who are its beneficiaries? What happened as a consequence of the policy that would not have happened in its absence? Consequently, systematic evaluation gives policy-makers and the general public, if they are interested, some notion about the actual impact of policy and provides discussions of policy with some grounding in reality. Evaluation findings can be used to modify current policies and programs and to help design others for the future.

Of course, evaluation studies can also be used for less laudable purposes. Professor Carol Weiss comments, “Program decision-makers may turn to evaluation to delay a decision; to justify and legitimate a decision already made; to extricate themselves from controversy about future directions by passing the buck; to vindicate the program in the eyes of its constituents, its funders, or the public; to satisfy conditions of a government or foundation grant through the ritual of evaluation.”19 In short, evaluators may be motivated by self- service as well as public service and by a desire to use analysis as ammuni- tion for partisan or personal political purposes. Thus, the staff of the Federal Paperwork Commission was interested only in evaluations supportive of their goal of eliminating as much governmental regulation of business as possible, albeit under the guise of paperwork reduction.20

We now turn to some of the ways in which policy evaluation is handled in the national government.

Policy Evaluation Process

Much policy evaluation is performed by nongovernmental actors. The communications media; university scholars and research centers; private re- search organizations such as the Brookings Institution, the Urban Institute, and the American Enterprise Institute; pressure groups; and public-interest organizations, such as Common Cause, the Audubon Society, and Ralph Nader and his “raiders” (a collection of public-interest activists) are examples; all make evaluations of public policies and programs that have greater or lesser effects on public officials. They also provide the general public with information, publicize policy success and failure, sometimes act as advocates for unpopular causes, and occasionally provide representation for those unrepresented in the policy process, such as the aged confined to negligently run nursing homes or exploited migratory farm workers.

Collectively, this aggregation of private evaluators of public policies has been designated by Professor Richard Hofferbert as the “policy evaluation industry.”21 We will take a look here at three of its components.

Many university scholars engage in evaluation on their own initiative, as members of university research centers, while under contracts with government agencies, or to gain academic advancement or fame. Many universities have graduate schools of public policy whose faculty have a more “applied” or practical orientation and are inclined to undertake policy research. Some of

the research of university scholars is published in such journals as Evaluation and Program Planning, Evaluation Review, and Journal of Policy Analysis and Management. For the most part, though, university scholars are more interested in making theoretical contributions to their academic disciplines than in getting involved in current policy debates.

A number of private research organizations (“think tanks”) devote all or most of their time to performing evaluation studies.22 The Manpower Development Research Corporation uses funds from national, state, and local governments and foundations to research social welfare and employment programs. The RAND Corporation operates mainly as a contract researcher producing policy reports for the Department of Defense. Research organizations of this sort often can act more independently and be more critical of agency actions than can internal agency personnel. “Of course, contract researchers often face pressures to follow the agency line, especially if they are highly dependent on a single agency.”23

In their reporting of the news when public policies are involved, the media tend to focus on policies and government actions that go awry. Some sort of policy scandal is even more “newsworthy.” The operative assumption seems to be that that sort of reporting will attract more viewers or readers than will information about a successful program or the routine operations of government. Some years ago, one of the television networks featured an occasional series titled “The Fleecing of America” as part of its evening news report. The focus was on governmental blunders and misadventures. Coverage of this na- ture contributes to the (incorrect) notion that “the government can’t do any- thing well.” More positively though, it may also help to focus public attention on some matters that need examination and correction. Television and radio programs that recount public-policy successes are scarce. One does not hear much about soil conservation, brownfield restoration, railroad safety, or pub- lic housing.

Moving on, within the boundaries of the national government, policy evaluation is performed in numerous ways by officials and organizations. Some- times these evaluations are rigorous and systematic; at other times, they are rather haphazard or sporadic. In some instances policy evaluation has become formalized, a regular component of the policy process; in others it remains essentially informal and unstructured. In the remainder of this section, we glance at a few forms of official policy evaluation—congressional oversight, Government Accountability Office (GAO) studies, the activities of presidential commissions, and “in house” evaluations handled within agencies.

Congressional Oversight

Although it is not specified in the Constitution, one of the primary functions of Congress is the supervision and evaluation of the administration and execution of laws and policies, which is commonly referred to as congressional oversight. Some, agreeing with the English political theorist John Stuart Mill, think that this is the most important function a legislature performs.24 Oversight, however, is not a

separable, distinct activity; rather, it is an integral part of almost everything that members of Congress do, including gathering information, legislating, authorizing appropriations, and being of service to constituents. It may be in- tended either to control the actions of agencies, as when they sometimes are required to clear actions in advance with particular committees, or to evaluate agency actions, as when individual members or committees seek to determine whether administrators are complying with program objectives established by Congress. It is the evaluative aspect of oversight that is pertinent here.

Oversight may be performed by a number of techniques, including (1) casework, that is, intercession with agencies as a consequence of constituents’ demands and requests; (2) committee hearings and investigations; (3) the appropriations process; (4) approval of presidential appointments; and (5) committee staff studies.25 In the course of these activities, members of Congress reach conclusions about the efficiency, effectiveness, and impact of policies and programs—conclusions that can have profound consequences for the policy process. Congressional oversight is in essence more fragmented and disjointed than continuous and systematic. Bits and pieces of information, impressionistic judgments, and the members’ intuition and values are blended to yield evaluation of policies and those who administer them. On the whole, however, members of Congress are more likely to be involved with initiation and adoption of policy than with evaluation, at least in any systematic sense.

The congressional practice of enacting laws—for example, the Clean Air Act and the Head Start law—for determinate lengths of time builds evaluation into the legislative process. Periodically such laws have to be reauthorized by Congress, thereby creating opportunities to make changes in them on the basis of experience and available information. At the state level, sunset laws have a similar effect. These laws provide that state agencies will be terminated unless they are periodically evaluated and reauthorized. The programs of a terminated (“sunsetted”) agency may be shifted to another agency. That op- portunities for evaluation are created does not, of course, guarantee that they will be well exploited.

Government Accountability Office

The GAO, usually regarded as an arm of Congress, has broad statutory authority to audit the operations and financial activities of federal agencies, to evaluate their pro- grams, and to report its findings to Congress.26 The agency, which has several thousand employees, has become increasingly involved with the evaluation of programs and now gives only a minor portion of its attention to financial auditing.

The Legislative Reorganization Act of 1970, which revamped the congressional committee system, also directed the GAO to “review and analyze the results of government programs and activities carried on under existing law, including the making of cost-benefit studies,” and to make personnel available to assist congressional committees in handling similar activities. A subsequent statute authorized GAO to establish an Office of Program Review and Evaluation. Because of its expanded evaluation activities, the agency hires many more people trained in the social sciences than it once did.

Evaluation activities may be undertaken by the GAO on its own initiative, because of directives in legislation, at the request of congressional commit- tees, or sometimes at the behest of individual members of Congress. In the course of a year, the GAO produces several hundred evaluation studies, varying in length from several pages to a few hundred. The titles of four evaluation studies issued by the GAO in 2012 indicate the diversity and evaluative thrust of its actions: FDA’s Food Advisory and Recall Process Needs Strengthening; Border Security: Additional Steps Needed to Ensure that Officers Are Fully Trained; Nuclear Safety: DOE Needs to Determine the Costs and Benefits of Its Safety Reform Effort; and School Bullying: Extent of Legal Protections for Vulnerable Groups Needed to Be More Fully Assessed. GAO officials are also frequently called upon to testify before congressional committees.

Copies of GAO studies are sent to members of Congress and the affected agencies. The agencies are required by law to report to Congress and the Office of Management and Budget (OMB) on actions taken in response to the GAO’s recommendations or on why they did not act.

Problems in Policy Evaluation

The most useful form of policy evaluation for policy-makers and administrators, and for policy critics who want a factual basis for their positions, is a systematic evaluation that tries to determine cause-and-effect relationships and rigorously measures the results of policy. It is, of course, often impossible to measure quantitatively the effect of public policies, especially social policies, with any real precision. In this context, then, to “measure rigorously” is to seek to assess policy impacts as carefully and objectively as possible, using the best information available and making careful judgments. There is no reason to assume that only quantitative data and analysis are meaningful.

Determining whether a policy or program is doing what it is supposed to do, or doing something else, is not an easy, straightforward task, as some appear to assume. Snap judgments are easy to make but lack definitiveness. A variety of conditions raise obstacles or create problems for the effective accomplishment of policy evaluation. These include uncertainty over policy goals, difficulty in determining causality, diffuse policy impacts, and others, all of which are reviewed in this section.

Uncertainty over Policy Goals

When the goals of a policy are unclear, diffuse, or diverse, as they frequently are, determining the extent to which they have been attained becomes a difficult and frustrating task.42 This situation is often a product of the policy-adoption process. Because the support of a majority coalition is needed to secure adoption of a policy, it is usually necessary to appeal to persons and groups possessing differing interests and diverse values. To win their votes, commitments to the preferred policy goals of these various groups may be included in the legislation.

The Model Cities Act, enacted during the Johnson administration, was a laudable attempt to confront urban problems. Goals of the Model Cities Act included rebuilding slum and blighted areas; improving housing, income, and cultural opportunities therein; reducing crime and juvenile delinquency; lessening reliance on welfare programs; and maintaining historic landmarks. No priorities were assigned to the various goals; nor, indeed, were their dimensions well-specified. Substantial funding was provided for Model Cities evaluation research, which was supposed to determine the extent to which the diverse goals of the act were being met. Limited in impact at best, the Model Cities program has been relegated to the dustbin of history.

Determining the real goals of a program can be a difficult or conflictual task. Persons occupying different positions in the policy process, such as legislators and administrators, or national and state officials, or possessing differing ideological or philosophical perspectives, may define goals differently, act accordingly, and reach differing conclusions about a program’s accomplishments or success. Moreover, “because ‘what you measure is what you get,’ choosing the right goals to measure is essential.”43 Later in this chapter, we will see how the multiple goals of the Head Start program, and measurement problems, have complicated its evaluation.

Difficulty in Determining Causality

Systematic evaluation requires that societal changes must be demonstrably caused by policy actions. The mere fact that when action A is taken, condition B develops does not necessarily mean the presence of a cause-and-effect relationship. Other actions (or variables) may have been the actual causes of con- dition B. As we know, many common colds are “cured,” not by ingesting medicines, applying ointments, or using nasal sprays, but by the human body’s natural recuperative power.

Consider this example. Many states require periodic automobile safety inspec- tions in an attempt to reduce highway traffic accidents and fatalities. Research indicates that states with mandatory inspection laws do tend to have fewer traffic fatalities than do other states. Other factors, however, such as population density, weather conditions, and percentage of young drivers might in fact have more power in explaining the difference. Only if such conditions are controlled in the analysis, and if differences remain between states with and without inspections, can it be accurately stated that a policy of periodic automobile inspections reduces traffic deaths. In actuality, such laws do seem to have a modest beneficial effect.44

To further illustrate the problem of determining causality, let us take the case of crime-control policies. The purpose, or at least one of the purposes, of these policies is deterring crime. Deterrence may be defined as the prevention of an action that can be said to have had a “realistic potential of actualiza- tion,” that is, one that really could have happened.45 (This assumption is re- quired to avoid the kind of analysis that holds, for example, that consumption of alcoholic beverages prevents stomach worms since no one has ever been afflicted with them after starting to drink.) The problem here is that not doing something is a sort of nonevent, or intangible act. Does a person’s not com- mitting burglary mean that he or she has been effectively deterred by policy from so acting? The answer, of course, first depends upon whether he or she was inclined to engage in burglary. If so, then was the person deterred by the possibility of detection and punishment, by other factors such as family influence, or by lack of opportunity? As this example indicates, the determination of causality between actions, especially in complex social and economic mat- ters, frequently is a daunting task.46

Diffuse Policy Impacts

Policy actions may affect groups other than those at whom they are specifically directed. A welfare program may affect not only the poor but also others such as taxpayers, public officials, and low-income people who are not receiving welfare benefits. The effects on these groups may be either symbolic or material. Taxpayers may grumble that their “hard-earned dollars are going to support those too lazy to work.” Some low-income working people may indeed try to go on welfare rather than continue working at grubby, unpleasant jobs for low wages. So far as the poor who receive material benefits are concerned, how do benefits af- fect their initiative and self-reliance, family solidarity, or maintenance of social order? We should bear in mind that policies may have unstated intentions. Thus, an antipoverty program might have been covertly intended to help de- fuse the demands of black activists, or a program to control importing of beef may be intended to appease cattle growers politically but not really do much to limit foreign competition.

The effects of some programs may be very broad and long-range in nature. Antitrust policy is an example. Since the policy was originally intended to help maintain competition and prevent monopoly in the economy, how does one now evaluate its effectiveness? We can look at current enforcement activity and find that some mergers have been prevented and many price-fixing conspiracies have been prosecuted, but this record will tell us little about the extent of competition and monopoly in the economy generally. It would be pleasing to be able to determine that the economy is n percent more competitive than it would have been without antitrust policy. Because its goals are general and because measuring competition and monopoly is difficult, this determination just is not possible. Interestingly, after a century of anti- trust action, we are still without agreed-upon definitions of monopoly and competition to guide policy action and evaluation. No wonder those assessing the effectiveness of antitrust policy sometimes come to sharply different conclusions.47

Difficulties in Data Acquisition

As implied in some previous comments, a shortage of ac- curate and relevant statistical data and other information may handicap the policy evaluator, particularly when one’s concern is with policy outcomes. Thus, an econometric model may predict how a tax cut will affect economic activity, but suitable data to indicate its actual impacts on the economy are hard to come by. Again, think of the problems in securing the data needed to determine the effect on criminal-law enforcement of a Supreme Court decision such as Miranda v. Arizona, 48 which held that a confession obtained when a suspect had not been informed of his or her rights when taken into custody was inherently invalid. The members of the President’s Crime Commission in 1967 disagreed about its effect, the majority saying it was too early to determine results. A minority, however, held that, if fully implemented, “it could mean the virtual elimination of pretrial interrogation of suspects ....Few can doubt the adverse effect of Miranda upon the law enforcement process.”49 Absence of data does not necessarily hinder many evaluators.

The use of “Miranda cards” to inform suspects of their rights now has become standard police practice. A consensus exists among criminal-justice scholars and law-enforcement officers to the effect that this reform has had little adverse effect on law enforcement. Various field and quantitative studies support this view. Moreover, it is suggested that the Miranda rule has helped improve professionalism among the police.50 For many social and economic programs, a question that typically arises is “Did those who participated in programs subsequently fare better than comparable persons who did not?” Providing an answer preferably involves an experimental evaluation design utilizing a control group. The difficulty in devising a control (or comparison) group for a workforce program is summed up in this passage:

A strict comparison group in the laboratory sense of the physical sciences is virtually impossible, primarily because the behavior patterns of people are affected by so many external social, economic, and political factors. In fact, sometimes the legislation itself prevents a proper comparison group from being established. For example, the Work Incentive Program legislation required that all fathers must be enrolled in the WIN program within 30 days after receipt of aid for their children. Therefore, a comparison group of fathers with comparable attributes to those fathers enrolled in the program could not be established. Even if all the external factors of the economy could be controlled, it would still be impossible to replicate the social and political environment affecting any experimental or demonstration program. Thus, it is easy for a decision maker to discount the results of almost any evaluation study on the basis that it lacks the precision control group.

Because of problems such as those mentioned in the quotation, experimental designs frequently cannot be used. (This reason is apart from their often high dollar cost.) Second-best alternatives must then be utilized, such as a quasi- experimental design using a nonequivalent control group.52

Official Resistance

Evaluating policy, whether it be called policy analysis, measurement of policy impact, or something else, involves re- porting findings and making judgments on the merits of policy. This is true even if the evaluator is a university researcher who thinks that he or she is objectively pursuing knowledge. Agency and program officials will be alert to the possible political consequences of evaluation. If the results do not come out “right” from their perspective, or worse, if the results are negative and come to the attention of decision-makers, their program, influence, or careers may be thrown in jeopardy. Consequently, program officials may discourage or disparage evaluation studies, refuse access to data, or keep in- complete records. Within agencies, evaluation studies are likely to be most strongly supported by higher-level officials who must make decisions about the allocation of re- sources among programs and the continuation of given programs. They may, however, be reluctant to require evaluations, especially if their results may have a divisive effect within the agencies. Finally, organizations tend to resist change, and evaluation implies change. Organizational inertia may thus be an obstacle to evaluation, along with more overt forms of resistance

A Limited Time Perspective

The time horizon of legislators and other elected officials often extends only as far as the next election. Consequently, they, and others who think like them, often expect quick re- sults from governmental programs, even social and educational programs whose effects may take many years to fully appear. This being the case, short- run evaluations of program accomplishments may be unfavorable. A good example is the New Deal’s resettlement program, which provided opportunities for land ownership to thousands of black sharecroppers in the South during the late 1930s and early 1940s. It was judged as a failure and just another New Deal boondoggle by contemporary critics. A decades-later evalua- tion of the program by policy analyst Lester Salamon concluded, however, that it had significant, positive, long-term effects, although not as an agricultural policy.53 At modest cost, it did transform “a group of landless black tenants into a permanent landed middle class that ultimately emerged in the 1960s as the backbone of the civil-rights movement in the rural South.” If the time dimension is ignored in evaluation studies, the results may be flawed and ne- glect important long-term effects. The pressure for rapid feedback concerning a policy can then create a dilemma for the evaluator.

Evaluation Lacks Influence

Once completed, an evaluation of a program may be ignored or attacked as inconclusive or unsound on various grounds. It may be alleged that the evaluation was poorly designed, the data used were inadequate, or the findings are inconclusive. Those strongly interested in a program, however, whether as administrators or beneficiaries, are unlikely to lose their affection for it merely because an evaluation study concluded that its costs are greater than its benefits. Moreover, there is also the possibility that the evaluation is flawed. Governmental programs are not terminated solely as a consequence of an unfavorable systematic evaluation, although such evaluations did contribute to adoption of the Airline Deregulation Act. Of course, evaluations frequently lead to incremental changes or improvements in the design and administration of programs. That is the intent of many program evaluations done by the GAO, for instance, which, perhaps, is all that should be asked or expected of most evaluations.

Policy Evaluation: The Use and Misuse of Cost–Benefit Analysis

Cost–benefit analysis is a formal, quantitative evaluation technique that re- quires identifying the costs and benefits of either a proposed or actual policy and translating them into monetary values for purposes of comparison. It assumes that society will be made better off only by policies (or projects, or pro- grams) whose benefits exceed their costs. Cost–benefit analysis has been most frequently used to evaluate proposed policies.

Sometimes, though, it is employed to appraise existing policies. Thus, economist A. Myrick Freeman III used it to evaluate national air- and water- pollution-control policies. He found that the control of air pollution from stationary sources yielded benefits that were much greater than control costs. On the other hand, the costs of controlling industrial and municipal sources of water pollution were greater than the benefits realized.54 In the following discussion, the focus will be on cost–benefit analysis primarily as a prospective evaluation technique. The major steps in performing a cost–benefit analysis can readily be sum- marized.55

First, one identifies all of the effects or consequences of a policy and categorizes them as costs or benefits for various groups. (Note that this requires establishing which groups are entitled to be considered in determining costs and benefits.) Both direct and indirect effects should be analyzed.

Second, dollar values are placed on various costs and benefits. This will be relatively easy for items that are customarily bought and sold in markets. For such matters as good health, the prolongation of human life, or scenic vistas, it will be much more difficult.

Third, some of the consequences or effects of a policy will be current or near-term, but others will occur many years in the future, or even the next generation. Hence, proponents of cost–benefit analysis call for a discount rate to equate the value of near and long-term effects. The basic assumption under- lying the discount rate is that a dollar today is worth more than a dollar years from now. Inflation may reduce the dollar’s value or purchasing power. Alter- natively, the money could be invested and its value would increase. My guess is that most readers would prefer a thousand dollars today to a thousand dollars seventeen years from now. What is the explanation?

Fourth, the costs and benefits, direct and indirect, current and future, of the policy are compared. If benefits exceed costs, the policy is acceptable; conversely, if costs exceed benefits, it should be rejected, or a better way of doing it should be found.

So presented, cost–benefit analysis appears as a reasonably clear-cut method for appraising policies. In actuality, there are significant problems involved in its application, a few of which are examined here.

Good data on the costs and benefits of a policy are frequently difficult to come by. How, for example, does one calculate the value of the health ben- efits of cleaner air? Or of the esthetic benefits of reducing haze in national parks? How are dollar values to be assigned to such matters? The value of land flooded (a cost) for a reservoir can readily be determined by reference to the value of nearby land. But what of the value of an ancestral home located there? The data and dollar values on which a cost–benefit analysis is based can be of tenuous and arguable nature. Professor Frank Rourke states that “the numeri- cal values assigned to both costs and benefits are more often the product of imagination than mathematical skill.”56

It is, further, no easy task to identify the appropriate discount rate. It can be based on such criteria as the interest rate, the rate of inflation, or the opportunity costs of capital—that is, the rate of return that money would earn if devoted to private investment rather than public purposes. Despite its importance, there is no scientific way to decide on a discount rate. A low dis- count rate preserves the value of future benefits whereas a high discount rate can sharply reduce their value. During the Reagan years, the OMB advocated a discount rate of 10 percent. This discount rate meant that the value of future benefits, such as lives prolonged two or three decades hence by reducing the incidence of cancer, would have very low value. This, in turn, increased the likelihood that a cost–benefit ratio would be unfavorable, and that regulation would not be justifiable.

Cost–benefit analysis is based on the premise that efficiency is the pri- mary, if not the only, value to be realized. Actions are evaluated on the basis of whether resources are used to improve the aggregate public good.57 Little attention is accorded alternative or competing values—equity, human dignity, personal freedom, and equality, to name some. These are important to most people. The American system of criminal justice, for instance, is not very efficient because of our concern with equity and due process.

Another problem that arises in the course of many cost–benefit analyses is the need to place a value on a human life. Some take the position that life is priceless and that attempting to place a dollar value on life reduces it to just another commodity. In response it is argued that many policy decisions, such as industrial-safety standards and highway speed limits, have an impact in terms of human lives. It is better to objectively take into account the value of life. The EPA in 2008 determined that the “value of a statistical life” was $6.9 million in current dollars, nearly a million dollars less than five years earlier. This figure was calculated from opinion surveys asking people what they were willing to pay to avoid specified risks and what employers paid employees to assume additional risks. (Figure 7.1 displays some alternatives for valuing a human life.) The EPA has traditionally put the highest value on human life of any federal agency. Interestingly, all units within the EPA do not use the same figure.58 Under the cost–benefit regimen, the less a life is worth, the less need for regulatory protection, and vice versa.

Finally, let us note that cost–benefit analysis emphasizes the consequences for society as a whole. As we know, however, public policies distribute advantages and disadvantages, or costs and benefits. Those who pay the costs of policies often do not benefit from them, and vice versa. Put differently, policies have distributive consequences that are of importance. People may appropriately be more concerned with who benefits from industrial-safety policies than whether their total costs exceed their benefits.

Problems such as those sketched here have not prevented cost–benefit analysis from being used as a tool in governmental decision-making for several decades. The Flood Control Act of 1936 specified that flood-control projects could be undertaken by the Army Corps of Engineers only “if the benefits to whomsoever they may accrue are in excess of the estimated costs.” This standard must also be used for water projects handled by the Natural Resources Conservation Service and has been voluntarily employed by the Bureau of Reclamation. In the 1960s, cost–benefit analysis was first used in evaluating defense programs and then domestic programs as part of PPBS.

In the 1970s, Presidents Ford and Carter directed executive-branch regulatory agencies to prepare “inflation impact statements” and “regulatory analyses,” respectively, in developing some proposed regulations. These statements involved analyzing their expected economic consequences. The Carter administration made it clear, however, that although regulatory agencies should con- sider the burdens and gains of proposed regulations, a cost–benefit test was not to be used in appraising them.59

A goal of the Reagan administration when it took office was to substantially reduce governmental regulation of private economic activity. People who were critical of the programs under their jurisdiction were appointed to regulatory positions. A second action involved issuing Executive Order 12291 in February 1981,60 which drew heavily upon the Carter administration’s experience. The order required that proposed major regulations issued by executive- branch agencies (the independent regulatory commissions were exempt) must be accompanied by regulatory impact analyses assessing the potential bene- fits, costs, and net benefits of the regulations, including effects that could not be quantified in monetary terms, unless such calculations were prohibited by law. Some statutes ban the use of cost–benefit analysis for the programs they establish.

Major regulations were defined as those likely to have an annual impact on the economy of $100 million or more, to lead to major cost or price increases, or to have “significant adverse effects on competition, employment, investment, productivity, innovation, or the ability of U.S.-based enterprises to compete with foreign-based enterprises in domestic or export markets.” The OMB was authorized to make the final determination of what was a major rule, to supervise the evaluation process, and to delay the issuance of proposed or final rules if it found the regulatory analyses were unsatisfactory.

Rules could be issued only if their estimated benefits exceeded their esti- mated costs. If a choice was available, the less costly alternative was to be selected. The burden of proof that this standard was met rested with the agency. An action by the OMB holding up a rule could be appealed to the President’s Task Force on Regulatory Relief, which was staffed by the OMB and comprised several executive officials under the leadership of Vice President George Bush. (The word relief in the task force’s title indicates its orientation.) Although the task force was phased out in 1983, all of this planning was intended to ensure, among other things, that “Regulatory Action shall not be undertaken unless the potential benefits to society for the regulation outweigh the potential the potential costs to society.” Thus, cost–benefit analysis was to be more than an analytical technique; it became a decision rule with a conservative bias. The Reagan regulatory-analysis program was a center of controversy. Critics contended that it was used improperly to reduce the extent of regulation and to delay the issuance of rules rather than to improve the quality of regula- tions by encouraging better analysis. The OMB was also accused of improp- erly interfering in the regulatory process by usurping authority vested in the regulatory agencies. The administration denied such accusations. In practice, though, administration officials demonstrated much more vigilance about the costs than the benefits of regulation in trying to reduce the burden of regula- tory activity on businesses.

The George Bush administration continued the regulatory-analysis pro- gram and in time created the Council on Competitiveness, an interagency committee chaired by Vice President Dan Quayle, to work with the OMB in perpetuating the use of cost–benefit analysis.61 In the final two years of the George Bush administration, the Council on Competitiveness acted vigorously to represent the business community in the regulatory process and to reduce the number and strength of new regulations.62 For the most part, it avoided publicity and sought to leave few “fingerprints.”

The Clinton administration quickly abolished the Council on Competitive- ness and later replaced Executive Order 12291 with its Executive Order 12866, entitled “Regulatory Planning and Review.” The Clinton regulatory-review pro- gram continued to make use of cost–benefit analysis for major rules. Regulatory review, however, was conducted more openly and less stringently and intrusively than under the Reagan and George Bush administrations. Regula- tory agencies experienced little to complain about.

The George W. Bush administration retained Clinton’s executive order. The Office of Information and Regulatory Affairs (OIRA), however, now came under the direction of a true believer in cost–benefit analysis. Much of his activity at a university policy center had been funded by business corporations. In its first few months under the Bush administration, OIRA rejected twenty- one proposed regulations, more than had been turned down by the Clinton administration in eight years.63 This was not surprising given the strongly conservative, antiregulation stance of the George W. Bush administration. OIRA continued for the next seven years to take a tough stance on proposed rules.

As the record here presented indicates, the uses of cost–benefit analysis, and presidential review of proposed regulations, have become regularized fea- tures of the regulatory process. They will continue under the Obama administration, albeit with softening modifications.64 How they are implemented depends substantially upon the ideological leanings of a presidential administration. Clearly, cost–benefit analyses is a malleable evaluation instrument.

Fairly used, cost–benefit analysis may contribute to rationality in the decision- making process by aiding in the identification and appraisal of alternatives, helping to identify impacts or consequences, and otherwise developing information and insights that will assist persons in making reflective, well- considered decisions. A careful appraisal of the likely costs and benefits of a proposed action, and of the persons and groups upon whom they will fall, is certainly useful, regardless of whether all are converted into dollar figures, and without converting cost–benefit analysis into a decision rule.

Cost–benefit analysis, however, is open to manipulation to support the values and preferences of its users. In the instance of Executive Order 12291, be- cause of the stoutly antiregulatory orientation of its implementers, it became a form of partisan political analysis dressed up as regulatory rationality.65 Again, it is doubtful that the Army Corps of Engineers has ever been unable to under- take a rivers-and-harbors project that its officials really wanted to construct because a favorable cost–benefit analysis could not be contrived.

Policy evaluation, as our discussion indicates, is more than a technical or objective analytical process; it is also a political process. In the next section, a case study of the Head Start program illustrates how political factors can affect the conduct and results of an evaluation of a social program. The case also demonstrates that such evaluations, even when intended to be neutral or objective in form, become political because they can affect allocation of resources.

Policy Termination

As noted in the preceding section, the evaluation and appraisal of a policy, dissatisfaction with its costs and consequences, and the development and expansion of political opposition may produce a variety of responses to it, including termination. Policies are only one set of targets for termination. Others are programs (e.g., the rural abandoned-mine program), projects (e.g., the cross-Florida barge canal), and agencies. More than half of the states, for example, have enacted sunset laws, which require the legislature to renew periodically the authorization for administrative agencies. If this is not done, agencies are automatically abolished. The termination of an agency, however, may mean only that its programs are shifted elsewhere.

Another possibility is that an abolished agency may be replaced by one or more new agencies. When the Interstate Commerce Commission, which had been established in 1887 to handle railroad regulation, was ended in 1997, its place was taken by the Surface Transportation Board. By then many of the duties of the ICC had been eliminated. Following the Deepwater Horizon disaster in April 2010, the Minerals Management Service, which many viewed as the captive of the petroleum industry, was terminated. In its stead appeared the Bureau of Safety and Environmental Enforcement, in charge of permits, inspections, and safety matters; the Bureau of Ocean Energy Management, to handle leasing, review development plans, and look after renewable energy; and the Office of Natural Resources Revenue, to take care of financial matters.

The focus here is on policies. Most of us can readily identify a number of government policies that we regard as wasteful, unnecessary, or inappropriate because they offend our ideological inclinations. Others, however, may not share our beliefs and instead may view these same policies as necessary and desirable, perhaps needing some change or improvement, but on the whole worth keeping. Perhaps these people directly benefit from the policies. Or their ideology may inform them that such policies are laudable uses of governmental power. Just as most policies arise out of conflict, so too there will be dis- agreement over their worth and retention.

If criticism of and opposition to a policy become sufficiently strong so that the policy-makers feel impelled to take action, a policy is more likely to be altered than terminated. An effort may be made to strengthen the policy to make it more effective, or portions of it that appear especially ineffective or offen sive may be pruned away. This sort of adjustment is illustrated by the conversion of the Comprehensive Employment and Training Act of 1973 into the Job Training Partnership Act of 1982. Alterations were made in the administration of job training, and the public-employment program was jettisoned. Several years ago, the tobacco price-support program was separated from the general farm bill, and much of its cost was shifted to the private sector in order to save it; but only for a while. It later bit the dust.

Policy termination is difficult to accomplish for a number of reasons. Policies come into being because they have political support, and they typically retain some or all of that support. Though they may be few, the supporters of a policy or program probably will be strongly committed to it and may intensely resist change and ignore contrary evidence. The U.S. Army did not eliminate the horse cavalry until World War II, even though the cavalry had been obsolete for years because of military reliance on weapons such as the machine gun and rapid-fire artillery.75 Some army officials could not comprehend an effective fighting force without the horse cavalry, reality to the contrary. The Tea Board consisted of three tea-tasters whose task was to ensure the quality of imported tea. Since its establishment late in the nineteenth century, it survived several efforts to abolish it, finally succumbing in 1996 after a U.S. senator from Nevada made its termination a personal cause, thereby saving the government a whopping $200,000 a year.7

The critics and opponents of a policy may be less intense in their feelings and may be both somewhat disorganized and diverse in their interests as well. It may also be quite difficult to weld together a coalition of sufficient strength to repeal a policy. The Tea Board was like a sitting duck. Some potential oppo- nents of a policy may be most interested in preserving their own favored policies, and thus an attitude of “live and let live” may prevail. An intense minority often prevails over an indifferent majority in the democratic political process.

Within Congress, with its fragmented and dispersed power structures, those with jurisdiction over a policy are more likely than not to be its friends and supporters. They can then use their committee or subcommittee positions to protect the policy against attack, to fend off or stifle unwanted changes, and to block its termination, should that be proposed. Governmental structure favors those seeking to retain policies, just as it once favored those opposing their enactment. There is perhaps a rough equity in this arrangement.

Termination, moreover, is a severe action with unpleasant and negative connotations.77 It has an undertone of admitting failure. Unpleasant consequences may ensue when a policy is terminated: people may suffer lost income and jobs, prices for services or products may increase, and communities may decline. Ill will and other political costs may be entailed. Most public officials thus prefer to be involved in creating new or better policies rather than terminating the old.

Although these factors may make policy termination controversial and difficult, successful termination does occasionally happen. Here are some terminated policies and their dates of demise.

Although these factors may make policy termination controversial and difficult, successful termination does occasionally happen. Here are some terminated policies and their dates of demise.

Fair-trade legislation (1975). This legislation, adopted during the 1930s, permitted manufacturers of trademarked or brand-named products to set mandatory minimum resale prices for their products. It was intended to help small businesses. Over the years fair trade had become a tired, worn-out pol- icy whose time had passed. Little support for it remained, and it was easily repealed.

Commercial airline regulation (1978). Almost all economic (but not safety) regulation was eliminated by the Airline Deregulation Act, the first major victory of the deregulation movement that began in the 1970s. Many policy- makers became convinced, partly as a consequence of a multitude of policy studies, that market forces would more effectively control the airlines and protect the interests of users than would regulation.

Regulation of petroleum prices (1980). This policy, which always had much opposition, came into being as a consequence of the energy crisis in the 1970s. Difficult to administer, it was intended to prevent domestic oil companies from unduly profiting from high world oil prices. Its elimination in preference for market prices was coupled with a windfall-profits tax (see later discussion).

Synthetic-fuels research (1985). This policy was another product of the energy crisis. A costly program intended to develop commercial synthetic fuels as a substitute for fossil fuels, it had accomplished little by the time of its elimination, partly because of the length of time needed to get complex developmental and the Reagan administration wanted more reliance on the market.

Revenue sharing (1986). Adopted during the Nixon administration, revenue sharing channeled billions of dollars annually to state and local governments, with few strings attached, partly to encourage them to be more creative in dealing with public problems. Large federal budget deficits were the ultimate reason for its termination, although congressional opposition to revenue shar- ing was always considerable.

Crude-Oil Windfall Profits Tax Act (COWPTA) (1986). Enacted in 1980, COWPTA was the price the petroleum industry reluctantly paid for oil-price deregulation, which permitted domestic oil prices to rise. From 30 to 70 per- cent of the windfall, the difference between the selling (or market) price of oil and a specified base price, was taxed away. A phaseout of the tax was to begin in January 1988, if $227 billion in revenue had been collected, or within one month following the collection of that amount, but in any event, no later than January 1991.

The price of oil, however, fell in the mid-1980s, and the tax ceased to pro- duce revenue. An industry-supported effort to repeal the act failed in 1986. Success finally came in 1988, when a repeal provision was included in the Om- nibus Trade and Competitiveness Act to pick up or solidify votes for that leg- islation. Time and events had thus made the windfall tax symbolic and readily expendable.

Agricultural production controls (1996). Restrictions on the production of many farm commodities had been a prominent feature of farm programs since their inception in the early 1930s. They had long drawn criticism from Republicans, conservatives, and some farm groups. Most production controls were terminated by the Federal Agricultural Improvement and Reform Act (FAIR Act), also called “Freedom to Farm,” pushed through by the new Republican majorities in Congress. However, generous subsidies to enhance farm incomes were continued.

Glass-Steagall Act (1999). A 1935 response to the Great Depression, this act created bank deposit insurance and provided for the separation of investment and commercial banking. Banking deposit insurance became widely accepted; however, as the decades passed, the separation of investment and commercial banking came under increasing attack as an unnecessary hindrance to the financial business. Repealed in 1998, this action contributed to the financial crisis of 2007–2009.

Farm price support program for tobacco (2004). Put in place in 1938, the tobacco program used marketing quotas (mandatory limits on production) and nonrecourse loans to boost prices. Tobacco growers, worried about falling demand and lower prices for tobacco, and anti-smoking forces long opposed to the tobacco program, advocated its demise. A ten-billion-dollar buyout pro- gram for quota owners and tobacco growers helped gain their support. Many of those who were bought out ceased to grow tobacco.78 Now anyone, wherever located, can grow and sell as much tobacco as they choose.

As these examples indicate, a number of factors may contribute to the termination of policies. A short list includes ideology, the urge to economize, altered political conditions, and clear policy failure. Systematic evaluation played a major role only in airline deregulation. Evaluators (mostly econo- mists) over time were able to gather substantial evidence on the shortcomings of airline rate and route regulation and to effectively portray market regulation as an effective and satisfactory alternative. Most commonly, however, system- atic evaluation has not been a critical element in policy termination.

Indeed, to emphasize a point made earlier, evaluation is more likely to rein- itiate the policy sequence. Problems that emerge or become more intense dur- ing the implementation of a policy may be identified, alternatives for change or improvement may be formulated and debated, and so on, until perhaps the policy is modified in some fashion. It is also possible, of course, that those responsible for implementing a policy will act to make it more acceptable to complaining groups, as by speeding up the issuance of licenses or cracking down on certain kinds of law violations. Policy change, whether legislative or administrative in origin, is more likely to occur than policy termination.