The Political Economy of School Choice

INTRODUCTION

By the “regulated competitive delivery” of educational services, or more briefly “school choice,” I mean a system with the following characteristics:

choice: parents/guardians choose the schools their children will attend;

competitive delivery: private firms, public institutions, associations of teachers, and local community groups are entitled to offer such educational services;

public funding: accredited schools are financed from public funds at a fixed rate per eligible student per year, the amount being adjusted for such educationally relevant characteristics as grade level and special learning needs, as well as such economic factors as the local price level ;

accreditation: schools receiving public funds must meet acceptable standards concerning physical plant, staffing, curriculum, admissions, and financing;

accountability: quantitative measures of the performance of participating schools are maintained and disseminated to the public.

School choice has gained considerable popular and academic support in recent years.1 Much of the debate over school choice has been based on evidence concerning the relative performance of public and private schools. But such evidence gives a very partial picture, since a competitive system for delivering educational services, for reasons discussed below, is likely to be regulated in much the same manner as are current markets in such areas as health care, financial services, transportation, and communication. Private alternatives to state-run schools today are not subject to such regulation, and their performance may be either enhanced or impaired by this lack.2

Moreover, it is a mistake to evaluate the competitive delivery of educational services on the basis of traditional educational performance measures alone, since consumers value the ability to choose, independent of any measurable effects of such choice on standard measures of educational performance. For example, consumers value the right to choose their personal health care providers, although there is no strong evidence that the right to choose per se materially improves the quality of health care in the aggregate. Nor is this preference for choice irrational. Consumers want respect and solicitous concern in their interactions with service providers, whatever the characteristics of the good or service they receive. A key source of respect is power, and parents have power only to the extent that they can penalize their service providers when displeased. The most direct way of penalizing a service provider is to switch to an alternative provider, and this is effective only in a competitive delivery system in which the welfare of service providers depends on the size of their clientele.3

For these various reasons a theoretical model of the competitive delivery of educational services is indispensable for educational policy purposes. The model sketched in this article is informed by the following considerations. First, while educational services are part of the economic system, and while economic efficiency and consumer sovereignty are usually enhanced when producers are subject to the pressures of competition, in many areas markets are subject to “market failures” of a sort that their performance is improved when competitive systems are regulated, or when services are supplied in a noncompetitive manner.

Second, in evaluating a system of school choice we must not treat the government as capable of intervening optimally to correct market failures. For one thing, majority rule and coalition-formation, the standard tools of electoral politics, produce optimal policies only under stringent conditions.4 For another, legislators and administrators are normally subject to the influence of particular individuals and groups whose interests diverge from those of the general public.5 Finally, the costs of effective and flexible regulation may be unacceptably high.6 Ultimately, we must decide whether the benefits of the competitive delivery of educational services outweighs the costs of regulation, the costs of misregulation, plus the costs of the market failures that remain after regulation, when compared with the current system of delivery or another proposed alternative.7

Third, and intimately related to the previous two points, it is traditional in discussions of public policy to treat the market as the sphere of private life relevant to the allocation of private goods and services, and the state as the sphere of public life relevant to the constitution of our collective selves and the satisfaction of our common needs. In terms of educational policy, this dichotomy takes the form of viewing the desirability of school choice as dependent on whether education is seen as a “private” or a “public” good. The dichotomy is specious. Education is a public good in that it affects the nature of social relations in society as a whole. Yet properly regulated markets in educational services may be effective instruments for the satisfaction of our common, socially constituted educational goals. If the basic “social good” character is determined by the democratic process, the fact that its form of delivery involves the use of competitive markets does not undermine this determination.

For instance, we can decide collectively that all members of society have a right to publicly funded eyeglasses, but that they may be purchased from private vendors. The “common purpose” lies in the recognition of the right of people to have eyeglasses, the specification of what is included in the category “eyeglasses” (does it include sunglasses? ski goggles? designer sunglasses? battery-heated neon-lit glasses? ritually blessed oculars?), and who has the right to receive them (is a documentable vision impairment required? how often can they be replaced?). The public purpose is set by the various regulations surrounding the market in eyeglasses, and the market itself is merely a means of executing this purpose.

EDUCATION AS A CONSUMER GOOD

Consider an economy populated by firms that produce goods and services, selling them for a price to consumers. Under certain assumptions it can be shown that market competition produces the optimal amount and quality of each good and service, given the preferences and wealth of consumers. These assumptions include the following:

Many producers: The market for each product is large compared to the efficient operating size of a firm that produces it, so each consumer can choose from a number of firms producing the same product.

Products have known quality: The characteristics of each good (quality, durability, safety, and the like) are known to all consumers.

Consumers know their preferences: Consumers know the potential contribution to their welfare of each good and service.

Price is set so that supply equals demand: Prices can be attached to goods and services, and the price is free to move according to market forces.

Products are private goods: One person’s consumption of a good does not affect the welfare of other consumers in the economy.

We can motivate this conclusion, which I will not prove here, with the following argument. By assumption people know what is good for them and their consumption activity does not affect the welfare of others. Hence their choices are the best possible indicators of the value of the goods and services they consume. Since optimal firm size is small compared with total demand for a good or service, the market will sustain many firms in each industry, so competition among these firms will drive prices to their cost of production. Thus goods with the appropriate characteristics are supplied in amounts desired by consumers at prices reflecting costs of production.

If one or more of the underlying assumptions of this model do not hold, then market allocations will not be optimal. We call such violations “market failures.”

To what extent do educational services conform to this model? I will assume we are discussing primary and secondary schooling, where students are wards of their parents (I will henceforth use the term to include guardians as well), living at home, and commuting daily to school. The producers in this case are basically schools, although a single organization could operate many schools, and a single building or set of buildings could house several distinct schools. I will assume that the consumer is the student, and the product delivered has a complex set of attributes including developing physical, cognitive, psychomotor, spiritual, and interpersonal skills, instilling desirable moral values, entertaining, feeding, protecting from harm, and the like.

The assumption that students are the consumers of educational services is critical to my analysis, since it implies that we assess an educational system by its effect on the welfare of students. This “student-centered” assumption is controversial, and there exist at least two plausible alternatives. One is the “parent-centered” view that parents are the consumers of the educational services of their children, in the sense that parents have the right to have their personal preferences dictate the educational environment to which their children are exposed. Another is the “society-centered” notion that some social body, such as the state or the community, holds this right.

I do not want to criticize these alternatives here, although I believe neither is defensible. It appears, however, that many of the misunderstandings between proponents and opponents of school choice, as well as differences among proponents of school choice concerning its proper implementation, flow from different implicit assumptions concerning “whose preferences count” in the educational system. Proponents of the parent-centered view generally accept a relatively laissez-faire approach to educational policy issues, in which public funding, accreditation, and accountability are absent or minimal. By contrast, proponents of the society-centered approach generally oppose school choice on the grounds that social bodies, such as school boards, associations of professional educators, or groups of public-spirited citizens, are better situated to make the appropriate educational decisions.

Under the student-centered assumption taken here, what are the conditions for market optimality?

Many producers: Educational services are provided in an extremely decentralized manner and on very small scale compared with many industries. However the consumer (the student) must travel to the producer, so the effective demand for a given school is spatially limited. Thus the “many-producers” assumption is not reasonable under conditions where it is economical to have at most one or two schools within reasonable commuting distance of each student.

Products have known quality: Many of the services provided by schools are directly observable, including the quality of the school facilities, ancillary programs, transportation services, food quality, average class size, and teacher availability. Other more technical aspects of school quality, such as teacher quality and past student performance, may be provided by the reputation a school develops over time as well as by the quantitative performance measures collected by public agencies.

Consumers know their preferences: The consumers in this market (young students) do not have a systematically well developed knowledge of the effect of schooling on their welfare, nor do they have the maturity to act appropriately on the knowledge they do possess. Children attend school because they are compelled to do so by their parents and by law. The content and other characteristics of the services delivered to students are likewise determined by agents other than the consumer of these services. Thus the assumption that consumers know their preferences is radically violated in this case.

Price is set so that supply equals demand: This assumption is totally absent in the case of school choice. The price of educational services, which in this case would be the per-student tuition, is set by administrative fiat, with the expectation that suppliers will be willing to educate all eligible students at this price.

Products are private goods: There are several plausible external effects of a student’s consumption of educational services. For one, schooling affects the social behavior of the student both in youth and adulthood, and hence impacts on society by influencing the extent to which the student contributes to a common social purpose. In addition, students influence fellow students by affecting their pace of learning, their physical safety, their enjoyment of the educational experience, and their tolerance of cultural, racial, and other forms of social diversity. Educational services thus may have significant externalities.

REGULATING SCHOOL CHOICE

We have presented a formidable list of deviations from conditions for market optimality. What sorts of regulatory interventions could attenuate the market failures that flow from violations of these conditions?

The Assumption That There Are Many Producers

The minimum of schools in a community required for effective competition and consumer choice depends on many factors, and can be estimated only by detailed study. The number is probably greater than two, and may be as large as five or six. It is unlikely that this number will be attained by a free market except perhaps in communities with high population density and efficient transportation. Where these conditions are absent, competition must be encouraged by government intervention. One possibility is to enlarge the potential clientele of schools by subsidizing student transportation costs. Another is to lower startup costs for schools by offering low-interest loans to new schools, or purchasing educational physical plant and making this available to schools on a rental basis. Finally, schools can be required to share physical plant, such as classrooms, athletic facilities, specialized instruction resources, thereby decreasing minimum feasible school size.

In areas of low population density, it is possible that no combination of these strategies is likely to be successful, and an educational monopoly will result. Indeed, special subsidies may be needed to extend educational services at all to students in such areas. Even in such cases a modicum of school choice can be sustained if local governments are free to contract with potential suppliers of educational services, in much the same manner as towns and cities contract yearly for sanitation services and public transportation.

The Assumption That Product Quality Is Known

Some important school characteristics will not be available unless standardized measures of school performance, such as achievement scores, student retention rates, and the like, are maintained and available to the public. Such areas as teacher accreditation, adherence to building and other safety codes, and the use of appropriate instructional techniques can be subject to more or less detailed specification. The more stringent the requirements a school must satisfy to be accredited as a public tuition recipient, the more perfectly will parents know school quality. The down side, of course, is that the more stringent the requirements, the less real choice will be involved, the higher will be the administrative overhead costs, and the greater will be the latitude for political malfeasance and influence-seeking in the setting of the requirements.

The Assumption That Consumers Know Their Preferences

Both the current educational system in the United States and the school choice alternative are organized to reflect the understanding that parents have broad rights in making educational choices on behalf of their children. The “accreditation” requirement that I have included among the institutional foundations of school choice reflects the notion that parents’ rights are not unlimited in this respect, and that children’s rights and welfare are respected only when parental choice is subject to more or less stringent regulation. The current educational system includes safeguards of this type as well.

Nevertheless, the more effectively a competitive educational delivery system expands the range of parental choice, the more critical the assumption of parental competence and good will becomes. Under current conditions, unless parents are relatively wealthy or highly motivated, their choice of educational environment for their children is effectively limited to the choice of a residential community with a desirable school system. Residential location, however, is generally dominated by family concerns of which the quality of public education is merely one. Under school choice, many families are likely to gain opportunities now limited only to the wealthy, and hence the assumption that parents are the agents best able to make educational decisions on the part of their children would have greater impact on social outcomes, and hence must be more solidly founded, than in the case of the current educational system.

The school choice model is reasonable only if under normal conditions parents are more capable than other agents in society to make informed choices concerning the educational needs of their children and are more willing than other agents to expend the time and effort to place their children in the proper educational environment. Many educators doubt that parents possess these qualities, and are unhappy at the prospect of schools being run by profit-making institutions, of educational entrepreneurs plying an uncritical and indifferent public with get-educated-quick schemes and offering modish blandishments incompatible with educational excellence. For instance, one educator told me recently that there is “much evidence that people make choices about all kinds of things, including schools, for all sorts of reasons, only some of which have to do with the assumed purposes of those services or institutions. People choose cars because of their color, doctors because of their office location, schools because of their uniforms. Some people—many people—make critical life decisions with hardly a thought.”8 Many also argue that poor and uneducated families are especially unprepared for making informed choices, so a system of school choice would exacerbate educational inequality.

Proponents of educational choice, by contrast, tend to minimize the importance of the problem of uninformed parents, arguing that mistrust of the consumer reveals a paternalistic attitude toward parents that is unwarranted in light of the ability of people to choose intelligently in other areas of their lives. They argue, moreover, that the idea that there will be an “underclass” of parents who habitually support inferior schools is unlikely, since it is difficult to point to other areas where people so consistently make poor choices concerning their own welfare that the right to choose would best be placed in other hands. Finally, proponents of school choice point out that we would expect even the “poor and uneducated” to follow the lead of prominent community groups, political, cultural, and religious, in making choices for their children. If this is generally the case, instances of parental incompetence or malfeasance are not likely to be systematically related to the demographic or social characteristics of families.

Even accepting this argument, it remains true that where choice requires considerable expertise and information on the part of the consumer, government regulation is of critical importance to empower people in making informed choices and thereby avoiding serious market failures. Appropriate regulation increases the effectiveness of consumer choice by lowering the cost of acquiring the information needed to make informed choices. Financial institutions, for instance, are regulated in virtually all economies, because individual investors cannot be expected to make exhaustive investigations of each insurance company, mutual fund, pension plan, or bank with which they do business. Similarly, the deregulation of the airlines in the United States did not leave air safety to the marketplace, depending on the public’s ability to use air safety records as a guide to the choice of carriers. Rather, safety standards are Federal Aviation Agency–mandated. Finally, virtually every successful universal health care system in the advanced economies allows client choice, but because of client inexpertise, the conditions for accreditation as a health care provider are quite stringent.

Similar considerations apply in the area of educational services, where in addition to serious informational problems, simple incompetence and malfeasance in parental choice undoubtedly exist, and would acquire increased importance in a system of school choice. First, parents may be sufficiently dysfunctional that they be unable or unwilling to make responsible choices on behalf of their children. Second, parents may hold educational beliefs that are not conducive to the welfare of their children, at least according to the accepted social wisdom in this area. Finally, parents may use their right to choose schools for their children selfishly, promoting their personal ends rather than their children’s welfare.

In such cases proper regulation can considerably attenuate the market failure involved. Where parents are found legally incapable of making considered choices for their children, guardians must exercise the school choice decision. Unscrupulous schools that are found to mislead gullible parents may be pursued legally, as is currently done in the delivery of financial and insurance services. The extent to which parents have the right to exercise highly idiosyncratic choices is a hotly contested issue, but it is probably not difficult or costly to limit parental choice in whatever manner is decided through the political and judicial process, by the appropriate accreditation of teachers, teaching methods, curriculum content, and educational practices. Finally, schools can be prohibited from pandering to the selfish interests of parents (e.g., by prohibiting kickbacks to parents).

The Assumption That Price Is Set So That Supply Equals Demand

The “price” of educational services in a system of school choice is the tuition fee, which is set by administrative fiat and paid by the government rather than the consumer. A traditional criticism of this procedure on economic grounds is that if the price is set high enough to satisfy consumer demand when the price to the consumer is zero, too much education will be supplied. This may not be a problem when the consumption good is educational services, since it is widely held that the positive externalities generated by education are sufficiently great that the actual consumption of educational services is subobtimal.

The Assumption That Education Is a Private Good

When the private and social return to an educational practice differ, a system of unregulated school choice in which parents act on behalf of their own children’s welfare will favor private over social welfare. Suppose, for example, that the social good involves teaching students to sacrifice their own needs to that of society under some conditions, while the private good is to have your children taught how to make optimal use of the self-sacrificing predilections of others. Teaching self-sacrifice then is a “social good” the benefits of which are not limited to the students so taught, but are available to all members of society who may deal with these students in the future. A system of school choice will not automatically lead to the teaching of social norms of this type, unless they happen to be in the best (private) interest of students themselves. Similarly, social interaction in a diverse society may be better served when students learn in schools that are heterogeneous in terms of race, ethnicity, and social class. These social benefits may however be purchased at the cost of lower educational achievement or fewer educational amenities for some or all of the heterogeneous groups involved. A system of school choice might then result in a socially undesirable amount of social diversity.

Finally, the presence of one student in a school may have an “external effect” on other students, either positive or negative. Students with high achievement levels, with well-developed social skills, with a commitment to cooperative behavior, and with parents who are willing to contribute time, effort, and money to the school will benefit other students and their families. Such students and their families will then prefer to associate with one another rather than with others who have less to offer them in return.

It might be argued that, unlike the previous two examples, this is a source of inequality induced by choice, but is not a market failure, since it deals with the redistribution of benefits among agents, not the total level of benefits. But it is also reasonable to treat such “obligatory sharing” as an insurance policy for students who may in fact be born rich or poor, more or less intelligent, or into more or less socially caring families. An educational environment may have the effect of “ironing out” these arbitrary inequalities among students, just as health insurance partially corrects for randomly generated health differences among individuals.

Such divergences of private and social benefit can be ameliorated by proper regulation. Schools can be accredited to receive public funds only if their curriculum reflects “appropriate” social values. There could be a set of national guidelines specifying specific educational practices prohibited and mandated in accredited schools. Similarly, if ethnic and racial diversity is a social goal, schools can be offered bonuses or subsidies for having a diverse student body, or can be taxed or otherwise penalized for having an excessively homogeneous student body.

WHY GOVERNMENT MIGHT REGULATE AN EDUCATIONAL SYSTEM MORE EFFECTIVELY THAN IT CAN RUN AN EDUCATIONAL SYSTEM

The attractiveness of a system of regulated competitive delivery in education depends on what at first sight might seem to be a curious asymmetry in the treatment of government effectiveness: It posits that government is more successful in regulating the production of goods and services than in producing these goods and services directly. Why might this be the case?

Historical experience indicates that democratically accountable governments can effectively regulate the competitive delivery of goods and services where necessary. The political dynamic leading to effective regulation, however, is often volatile, imprecise, and slow to converge to a stable set of policies. Moreover, a system of regulation, once stabilized, is costly to maintain, is subject to continual pressure by groups who stand to gain by regulatory changes, is systematically subobtimal due to the political influence of groups that stand to lose from the implementation of more efficient policies, and is slow to accommodate technical innovation, changes in consumer preferences, and new social priorities.

Examples of effective industry regulation include agriculture, banking, securities, insurance, consumer products, occupational health and safety, transportation, communication, and health care. In each case some market economies are over regulated, others are under regulated, and all are to a greater or lesser extent misregulated. In each case the process of instituting a system of regulation has been discontinuous, erratic, and imprecise. But it is virtually inconceivable that such industries operate effectively without such regulation. Indeed, just as surely that there exist no successful economies without market competition, there also exist no successful economies that operate without significant and widespread state intervention.

Education is one of the few areas of economic life where the model of “regulated competitive delivery” has not penetrated to an appreciable extent. In most market economies the whole range of educational services is delivered by government monopolies, although families who can afford to do so are permitted to opt for private schools. The United States and Canada are among the few countries that have many nonstate providers of higher education. Despite the effectiveness of this system, it is not yet widely copied by other countries.

There is no historical experience indicating that government has a comparative advantage in the production of goods and services. Most successful economies reserve a minimal role for state production, and where state production does take place (e.g., the U.S. postal system, or state-owned automobile companies in Western Europe), it does so in competition with the private sector, and could doubtless be eliminated without compromising the viability of the economy in question.

It is often argued that the same forces that prevent government from producing effectively also prevent government from regulating effectively. Were this correct, the school-choice issue would reduce to a polar choice between unregulated laissez faire on the one hand and an inefficient government monopoly on the other. One explanation of government failure used to support this polar opposition is that governments are inherently “bureaucratic” and bureaucracy is inherently inefficient. This is hardly convincing, since modern economies are dominated by firms with extensive bureaucracies, from which we may reasonably conclude that bureaucratic organization often contributes to lowering costs and otherwise ensuring competitive success. Another explanation is that the democratic process is inherently incapable of rendering government accountable to the general public. This view is theoretically more defensible, but is ultimately unconvincing, I believe, because of the marked, though imperfect, success of government regulation in so many areas of economic life.

The private versus public dichotomy does not underlie the weakness of governments to produce efficiently. The relevant dichotomy is competitive versus noncompetitive. Public-sector control of an industry is generally ineffective because state-run firms are not subject to the forces of competition. No degree of democratic accountability of government can induce state-run monopolies to produce efficiently, for the following important reason: Only the competitive interaction of firms generates the information necessary to judge firm performance. Hence competition is necessary to render firms accountable to agents outside the firm.9

This fact allows us to explain in a straightforward manner why governments are capable of regulating competitive delivery systems, but incapable of replacing such systems by direct public-sector production: Competition among political decision makers through the democratic political process gives voters the knowledge (albeit incomplete) to assess their relative performance in economic policy, and also gives voters the power to reward and penalize their performance. Where government directly produces goods and services, political decision makers lack the information to reward and penalize public-sector producers, and hence the accountability of these decision makers to the public has no effect on the performance of public-sector producers.

Of course if this explanation is correct, then a plausible strategy to improve the performance of public-sector producers is to subject such producers to competitive contexts, either through intra-public-sector competition or through competition with the private sector, or both. In particular, it is at least theoretically possible that regulated competitive delivery in education could take the form of competition within the public schools system, in the absence of a private sector of tuition-receiving schools. There may in fact be powerful political forces operative within the public sector acting to temper and defuse intra-public-sector competition, since potential losers in a competitive struggle, by their proximity to public-sector decision makers, may have the means to lessen competitive pressures. For whatever reason, intra-public-sector competition is relatively rare, and probably does not present a viable alternative to school choice with free entry of private suppliers.

EDUCATIONAL MARKETS AND SOCIAL INEQUALITY

Market economies generate remarkably unequal distributions of income and wealth. Many believe that school choice will lead to a similarly unequal distribution of educational services. I suggest that if the model of school choice sketched above is accurate, this is not likely to be the case.

In assessing the effects of school choice on educational and economic inequality, we must distinguish between the financing and the delivery o f educational services. A competitive private delivery system can be financed in an egalitarian manner. In our model, an equal distribution of services can be achieved by ensuring that children receive tuition vouchers, of equal value for all similarly situated students, redeemable at an accredited school of their parents’ choice. Students with special needs that require costly treatment (e.g., handicapped or learning-impaired students) can be accorded increased tuition vouchers in this system.

The existing degree of resource inequality in education is often justified by the argument that “local schools” should be financed by “local taxes,” ensuring that resource inequalities among communities translate into similar inequalities among schools. This justification would be difficult to support in a system of school choice, since the funding of school choice is national rather than local. This degree of resource equality across racial and social-class categories attained by school choice could thus exceed that of the present system. Of course, if schools are allowed to supplement public tuition with private fees, it is likely that political pressures would lead to a lowering of the size of the public tuition contribution to the point where severe resource inequality reemerges, with middle-class parents supplementing public tuition vouchers and poor families using schools that get by on public tuition alone. To avoid this inegalitarian dynamic, schools receiving public tuition vouchers must be categorically denied the right to charge additional tuition fees, except for certain purposes (e.g., for special recreational events) and in a structly limited amount. Similar provisions are common in health care finance, where insurers not only do not pay, but also do not allow patients to pay, more than a prescribed amount for a particular service. An alternative (and less egalitarian) plan would allow accredited schools to levy additional fees, but only on a means-tested basis, and would prohibit a school from rejecting a student who cannot pay such fees.

The effect of the competitive delivery of educational services may counter the egalitarian effect of public financing, exacerbating existing inequalities of social class and race. We can envision several possible sources of inequality. It has been argued, for instance, that if schools have elective admissions policies, there will be a tendency for schools to become stratified according to ability, with the higher achievers having the advantage of a superior learning environment, as well as more able and committed teachers.

Suppose for the sake of argument that low achievers do sufficiently better when surrounded by high achievers. If high achievers are also helped by a heterogeneous system, then we can expect parents to prefer schools in which a wide range of student abilities are represented, so there is no market failure.

If, by contrast, high achievers are harmed by a heterogeneous school environment, but public policy values the gains to the low achievers above the losses of the high achievers, then we may include in the conditions for accreditation that the student body of a participating school have an adequate degree of heterogeneity in ability level, or at least in admissions. It may in fact be quite difficult to justify such interventions, since regulations of this type require the public recognition that the needs of high achievers are expressedly “sacrificed” to the needs of the less able, violating the principle of equal treatment on which much of public life is based. The current system of course achieves the same end of sacrificing the able to the less able (assuming, of course, that heterogeneity hurts high achievers), but in a manner that appears to satisfy the principle of equal treatment: Nontracking schools admit all legally eligible students and assign them randomly to different classes.

However, it is plausible that allowing elective admissions policies would help all ability levels, since each school could then tailor instruction to the particular needs of its clientele. This occurs in American postsecondary education, where elective admissions is rarely targeted as being unfair or inefficient. It might be thought that such a stratified system would exacerbate inequality by inducing lower aspirations in less successful students. This is not a plausible argument, although I know of no pertinent empirical evidence. It is certainly reasonable to think that schools dedicated to serving students of a given range of ability levels could instill self-confidence and high aspirations in these students.

Could not ethnic and racial segregation be exacerbated through elective admissions? Racial and ethnic discrimination is already illegal under law in the United States in most cases, so explicit discrimination is probably a problem that could be routinely and effectively handled by the judicial system. De facto discrimination on the basis of academic performance will doubtless occur, but it is unlikely to be more extensive than de facto segregation on the basis of residential community, as is the case in the current educational system. Moreover, a school choice system would offer disadvantaged and minority communities the resources to form schools catering directly to the needs of their particular constituencies. Such local initiative could be a potent force for improving the educational opportunities of disadvantaged and minority communities. Finally, as we have seen, if student body diversity is deemed a sufficiently important social value, accreditation could require schools to meet specific diversity requirements, or schools could receive bonuses for meeting diversity goals. As in other cases of government intervention in the private economy, however, regulations favoring diversity are vulnerable to attack for violating the principle of equal treatment.

EDUCATIONAL MARKETS AND THE TRANSMISSION OF CULTURE

Markets support diversity by tolerating the disparate preferences of consumers. Subject to minimum startup costs for forms, and subject to the consumer’s ability to pay, a market can satisfy whatever pattern of wants, common and arcane, base and refined, solemn and frivolous, that the consuming public happens to exhibit. Indeed, this capacity to support diversity is considered to be among the more attractive features of a market allocation: rather than fighting over what we are to consume in common, markets allow us all to get our own way.

It follows that insofar as schools are responsible for instilling a common cultural heritage in students, it is possible that an unregulated system of school choice could lead to progressive breakdown in shared cultural resources, as groups of parents form educational environments conducive to promoting their distinct beliefs, practices, and conceptions of the good. Of course our current educational system allows parents to act in this manner, but does not facilitate this practice by subsidizing it, thus ensuring that it have a relatively limited impact on social life. School choice could lead to an explosion of alternative school cultures.

Educators promoting a “multiculturist” approach to social values might well applaud such a situation, and might be expected to be among the strongest supporters of school choice. Others frustrated by the lethargy with which bureaucracies respond to new ideas will welcome it no less warmly.

Others will demur, however, on one of at least two grounds. First, many continue to believe that a harmonious and cohesive community must be based on shared values, and that extreme cultural heterogeneity would degrade community life. Second, there is no assurance that aggregating individual choices through the marketplace produces as socially desirable a result as the collective choices people make through the political process. For instance, commercial television has given us, many will say, more sex, violence, and inanity than would have been generated by a system governed by a state agency following legislated guidelines.

In effect, these criticisms recognize that the culture generated in education is a social good that is not necessarily produced optimally through competitive markets and individual choice since, as we have seen, a parent’s choice may impose costs on others that the parent need not take into account. The solution is to prohibit accredited educational institutions from crossing the boundary of acceptable educational environments. This could be accomplished through national guidelines specifying educational practices prohibited in accredited schools (e.g., inculcating racial intolerance, teaching creationist biology, espousing particular and/or divisive religious beliefs), and another set that schools are obligated to perform, including teaching a core curriculum, adhering to a code of ethics in treating students, and fostering values consistent with a pluralist democracy. It is likely that a set of such prohibitions and goals could be developed that would be approved by a considerable majority of the voting public, despite the fact that there would be some areas in bitter dispute that hence would be omitted from such a set. While I cannot speak authoritatively on legal questions, I doubt that issues of freedom of speech or expression would apply to such a set of guidelines, since governments regularly dictate content in the current educational system (including publicly supported higher education).

Indeed, there is nothing in the concept of competitive educational delivery systems that prevents the government from specifying exactly which books are used for which courses, and which pedagogical techniques are used under which conditions. I would not endorse such extensive government intervention, of course; my point is merely that such measures are compatible with competitive delivery.

The “democratic control of culture” could also be decentralized to local government. As an extreme case, the national accreditation process could require that each local community (e.g., city, town, metropolitan area, or county) have a school board with the right to approve educational content for schools located within their jurisdiction. Again, I would not approve of the use of such measures in more than a limited way, but they are feasible in a system of school choice.

THE PROBLEMS OF TRANSITION TO SCHOOL CHOICE

Probably the most challenging problems involved in implementing an effective system of school choice involve the problems of transition to the new system. One such problem is the danger that the political compromises that are required to achieve a mandate for change might substantially undermine the goals of change. One example of a “fatal compromise” would be the funding of partial as opposed to total vouchers, and another would be an excessively loose accreditation process. Many educators reject school choice on the grounds that the probability of a poor implementation is unacceptably high.

This fear may be a self-fulfilling prophecy. If educators refuse to develop and diffuse a reasonable model of school choice on grounds that the actual implementation will be inadequate, if and when school choice is implemented it very likely will be inadequate. It is incumbent on educators to make clear that there are prerequisites to an efficient and egalitarian choice system, and to specify exactly what these prerequisites are. Where agreement among educators is impossible, the public should know whether disagreements are political, ethical, or scientific, and if the latter, the source of the disagreement and the empirical evidence on which the disagreement could be resolved, or at least on which the range of disagreement could be significantly narrowed.

A second transitional cost is the potential losses to the “educational establishment,” including public school teachers and their unions, and other associations of individuals working within the public school system. It is well known that teachers and their unions are virtually unanimous in opposing school choice. How might a private educational delivery system injure teachers?

First, teachers’ unions are likely to be harmed. The per-member costs of union organization decline as the size of the bargaining unit increases. It is likely that the average size of private educational providers would be smaller than in the current system. Thus unless labor legislation is initiated, for instance obliging accredited schools to form regional associations for the purposes of bargaining with teachers, the current union structure would undoubtedly decline. In the current political environment, it is not likely that voters would approve of measures of this type.

Second, without strong unions or other comparable organizational support, teachers may be required to work longer hours, and their job security might be compromised. It is unclear whether wages would rise or fall. Many consider a move to school choice that excessively weakens the status of teachers unfair. To prevent this, a set of “teachers’ rights” could be implemented as part of a comprehensive school choice system and used in adjudicating between the needs of teachers and the consumers of education.10

Another cost of transition involves the funding of the system of private schools currently in operation. To fund such schools would increase the revenues needed to run the school system by some 10 percent. Most of the current private schools are parochial, but it is quite possible that a system of school choice would accredit such schools forthwith, or would allow such schools to reorganize in a manner acceptable to all parties. Again, this is not a real “cost,” but rather a redistribution from current public school to current private school users. Moreover, a 10 percent increase in costs is about what the system incurs now every two or three years. Doubtless the system could be phased in sufficiently slowly to accommodate this increased cost.

Finally, there is a host of legal issues concerning civil rights and separation of church and state that remain to be worked out. Hardly a week passes without some serious social issue surfacing in the United States concerning de facto and de jure segregation, busing, prayer in the schools, tracking, dress codes, single-gender classrooms, choice of textbooks, and the like. Doubtless the courts will face these problems when the time comes, but help from policy analysts who deal with legal issues surrounding education would be useful.

CONCLUSION

The government must provide some services monopolistically, because the market failures involved in competitive delivery are excessively costly. Examples include tax collection, police protection, national defense, and regulatory agencies. In each case we can provide compelling reasons why competitive delivery would not work. No such reasons can be given in the case of educational services. Indeed, unless there are structural forces prohibiting the emergence of effective regulation, or the costs of efficient regulation are excessively high, competitive delivery of educational services should better meet the private needs of parents and children, while fulfilling the educational systems traditional social functions as well.

People have rather prosaic goals for schools: reading, writing, history, math, and science, punctuality and self-discipline. When they are dissatisfied with what they are getting, they would doubtless benefit from having the power to induce the school to change, using the threat of taking their “business” elsewhere. The existing educational system disempowers parents by obliging them to initiate a complex political dynamic (influence the school board, affect the outcome of a local election, initiate a court battle) against great odds to induce their providers to change. The competitive delivery of educational services, properly funded and regulated, might succeed in short-circuiting this cumbersome process.

Educators often have rather more elevated goals for education: promoting equality and tolerance, teaching artistic, aesthetic and spiritual values, creating community. The idea that these could be promoted by the marketplace is anathema to many, but I have tried to show that this need not be the case. The choice of educational goals can still be debated in the political area, and the results could be implemented through the proper choice of policy tools. They would be codified in the rules for funding and accrediting schools. The use of the market is in this sense an instrument of rather than an alternative to democratic policymaking.

Too many individuals helped in the improvement of this paper to list individually. Many of the ideas expressed here were refined and modified through exchanges with educators on the EDPOLYAN (EDucational POlicY ANalysis internet discussion list, moderated by Gene V. Glass). I would like to thank them for their help.