December 1996





Henry M. Levin

Stanford University

December 1996 Draft—For Comments


Prepared for presentation at Annual Meetings of the American Economics Association, New Orleans, January 4, 1997. The author is a Visiting Scholar at The Russell Sage Foundation, 1996‑97 and the David Jacks Professor of Higher Education and Economics, Stanford  University.




Most of the policy discussion on the effects of educational vouchers has been premised on theoretical or ideological positions rather than evidence. In recent years a substantial amount of empirical evidence has accumulated on achievement differences between public and private schools; on who chooses and its probable impact on educational equity; and on the relative costs of public and private schools and on a voucher system. The purpose of this paper is to provide a summary of that evidence. (1) Present results among numerous studies suggest little or no difference in student achievement between public and private schools for a given student, but some evidence of higher rates of graduation, college attendance, and college graduation for Catholic high school students. (2) Evidence is consistent that educational choice leads to greater socioeconomic (SES) segregation of students as the more advantaged have historically been the families most likely to take advantage of choice. Families tend to choose schools where other students are from higher SES backgrounds, resulting in probable effects of rising inequality in achievement because of ³peer² and other effects associated with high SES schools and the withdrawal of such students from schools with lower SES students. (3) Despite assertions that costs of private schools are considerably lower than those of public schools, a comparison of costs in the Milwaukee Public Schools with private, voucher schools suggests that the costs of comparable services at public and private schools appear to be similar. But, the shift from the existing system of providing public education through school districts to a market system of educational vouchers would require considerable additional public resources for a supportive infrastructure that would provide an effective system of choice and competition. In particular, costs would rise because of public subsidies for existing private school students, record‑keeping for each voucher student, school monitoring, school accreditation, student transportation, information systems on alternatives, and adjudication of disputes. Preliminary estimates suggest an excess public cost on the order of $75 billion per year nationally, about 25 percent of present spending, about $ 1,500 per year per student.







Since Milton Friedman proposed his original voucher plan some four decades ago (1955) with a wider dissemination in his important book on Capitalism and Freedom (1962), the idea has taken on more and more credibility. Frustration with public schools in the inner cities has been a particularly important reason for emerging support of vouchers. Yet, both advocates and detractors tend to argue more from theoretical and ideological grounds than empirical ones on the consequences of vouchers. The purpose of this article is to consider empirical evidence concerning three issues on which there have been strong views expressed in the policy arena. (1) Will vouchers improve student achievement? (2) Who will choose and what are the educational consequences? (3) What is the evidence on comparative costs of public versus private schools and on the costs of a voucher system?


It is only fair in addressing these types of issues that I clarify where I stand on vouchers. Almost thirty years ago (Levin 1968) I argued that the situation of inner city students is so dire that we ought to be willing to design good experiments with vouchers or voucher‑type mechanisms to ascertain their effects on both individual and societal outcomes. In subsequent publications (e.g. Levin 1980, 1987) I have argued that the specific design of a voucher system with respect to finance, regulation, and information will be crucial in determining specific outcomes rather than leaving the discussion at a generic and abstract level, a point that is also stressed by Moe (1995) and Murnane (1986). More recently (Levin 1991) I have suggested that the private benefits of vouchers are likely to be positive relative to the present system in terms of satisfying narrow consumer preferences, but the social consequences will be worse because of greater inequality and the further deterioration of a common educational experience as social goals of schooling are sacrificed to consumer sovereignty. In what follows I will not take a stand on vouchers as much as try to read the present evidence on the three issues set out above.




To a large degree the arguments for educational vouchers have been premised on whether they will improve the educational achievement of students, particularly students from poverty backgrounds and inner‑cities where school results are considered to be particularly woeful. Because student achievement is considered to be a universal goal of schools, it has become the sine qua non for evaluating school reform. We should begin this section by explaining how limited this focus is in the context of market choice. The rationale for market choice in education is to give families the freedom to pursue their own educational preferences. For some families academic achievement will be the prime goal; for others it will be a school environment that is safe and supportive; for others yet it will be a quest for educational reinforcement of religious or philosophic values. Although most families may have some concern for the academic dimension, it may not be the prime dimension and may even be overwhelmed completely by other school and family goals as systematic studies have shown (Echols and Willms 1995). For example, in evangelical Christian schools it appears that preparation for the Kingdom of God far outweighs concerns about academic achievement (Peshkin 1986). So, comparing the effectiveness of schools only on student achievement is not fully consistent with measuring the impact of vouchers on educational outcomes where families may choose schools according to many criteria. And, even as a measure of social outcome, achievement tests are a limited and highly imperfect sample of the range and depth of knowledge and skills, values, attitudes, and other behaviors that we expect schools to inculcate in the young (Inkeles 1966).


Comparisons of Public/Private Student Achievement


Nevertheless, the primary focus in comparing public and private schools—even in the absence of vouchers—has been to ascertain whether either sector has an advantage in achievement, net of differences attributable to differences in student characteristics. It should be noted that controls for self‑selection are problematic in that even when controlling for race and indicators of social class of students, families who choose private schools and make a financial effort to pay for them are likely to be more educationally motivated than those who do not. Therefore, we would expect students from such families to have higher achievement than similar students who do not make the efforts to switch from a public to a private school. Whether one can control statistically for this self‑selection effect is problematic (Witte 1992).


The first major study by Coleman et al. (1982) compared a cross‑section of 10th grade students in public and private (mainly Catholic) schools, controlling for race and socioeconomic background. They found that students in private schools had slightly higher achievement, from .12 to .29 standard deviation units, depending upon the test. But, their results were criticized as overstatements of the private school effects because of inadequate controls for selection bias and other problems in the statistical design (Goldberger and Cain 1982). Purported adjustments for some of these problems reduced considerably or eliminated the private school advantages (Willms 1983).


Longitudinal results based on sophomore‑to‑senior changes found smaller private school advantages, from a range of no difference to .1 of a standard deviation in achievement (Alexander and Pallas 1985; Haertel et al.., 1987; Hoffer et al., 1985; Willms, 1985). This effect is statistically significant, but small, amounting to only about 10 points or less on the Scholastic Aptitude Test for college admissions, a trivial advantage. Further, it means that the achievement overlap between the two sectors is so great that 46 percent or more of public school students have higher achievement than the average private school student who is statistically similar (Levin 1987:634). Using earnings equations for 1976 data (the achievement data were collected in 1980), such an achievement advantage translated into earnings gains of less than 5 cents an hour for high school graduates some 4 years after graduation and about 1 day less unemployment a year among a cohort that experienced about 50 days of annual unemployment.


More recent statistical studies have also found no differences in achievement or only minimal differences. The most sophisticated studies from a modeling and statistical perspective are those of Goldhaber (1996) and Gamoran (1996). Goldhaber uses the (NELS 88) data set and finds no difference in achievement between comparable students in public and private schools (Goldhaber 1996). Gamoran¹s use of the same data set with a different statistical technique, hierarchical linear modeling, also finds no achievement difference or a very slight private school advantage, depending upon which statistical formulation is used. In the few cases where differences are found in favor of private schools, the advantage is not even as large as the trivial differences cited earlier.


Even when differences are found in such public‑private achievement studies, they are often questionable. For example, Sander (1996) found no difference in achievement between public and Catholic schools for students who attended Catholic schools for 1‑7 years, but an advantage only for those who had attended Catholic schools for 8 years. Not only is it puzzling that the putative Catholic school advantage takes eight years to ³bloom² with nary a hint of a bud in the earlier years, but even this result is questionable because it is not based upon an equivalent public school comparison group. When restricting the finding only to those who have attended Catholic school continuously for eight years, it is necessary to compare achievement with an equally stable public school group of students who have not been mobile. School stability has been found to be an important correlate of school success in the general literature (Rumberger and Larson 1996). But no attempt was made to compare the achievement of students with 8 years of Catholic school (presumably most attending the same school) with a comparable group of public school students who attended the same public school. An appropriate comparison would have been to compare students with the same stable attendance patterns between the two sectors to net out school effects.


Another recent study that models existing data sets to estimate effects of vouchers, Hoxby (1996), finds that a statistical proxy for private school subsidies was associated with both greater school competition and both higher private and public school achievement, findings that reinforce the textbook version of competition. However, as it turns out her model is based upon arbitrary assumptions which, when relaxed, can yield exactly the opposite conclusions (Kane 1996). That is, the results are not robust under a range of plausible assumptions on model construction and interpretation.


The Milwaukee Experiment


Of course, none of the public‑private comparisons can be as instructive as the direct evaluation of a voucher intervention. There are a handful of voucher‑type mechanisms funded by private sources, but none has been subjected to a careful evaluation of achievement effects (Moe 1995a). The only attempt to assess directly the impact of vouchers on student achievement has been the Milwaukee Voucher experiment. That experiment allowed students from families with incomes no more than 1.75 times the poverty line to attend private nonsectarian schools in Milwaukee with public funds. The numbers of participants were limited to no more than 1 percent of Milwaukee Public School enrollments except for the fifth year of the program when the limit was raised to 1.5 percent. Some seven schools participated initially, rising to 12 in the last two years. September enrollments in the private school program rose from 341 in 1990‑91 to 830 in the 1994‑95 school year, considerably below the maximum number eligible to participate which varied from 931 in the first year of the program to 1450 in the fifth year. Attrition rates from year‑to‑year were considerable, varying from 46 percent in the first year to 28 percent in the fifth year, so relatively few students participated for three years or more (Witte, Sterr and Thorn 1996:Table 1)..

Witte and his associates compared student achievement and found no systematic differences between voucher students in private schools and statistically similar students in the Milwaukee Public Schools. Recently their findings were challenged by Jay P. Greene, Paul E. Peterson, and Jiangtao Du (1996). These authors argued that since oversubscribed schools had to randomly choose students from their applicant pool, these conditions ³allowed for a natural randomized experiment.² They then compared students who had been chosen to participate with those in the applicant pool who had not been chosen. In short, they found that private school voucher students in their first two years had achievement levels that were not different from non‑accepted applicants who were in the Milwaukee Public Schools. However, they found that voucher students in the third and fourth years of participation scored higher than the general pool of non‑selected students. They concluded that ³Students benefit in measurable ways from the choice experience only after participating in the program for three or more years (Greene, Peterson, and Du, 1996:13).²


Although it can be argued that the students who entered the voucher schools were equivalent for comparative purposes to the non‑selected students, it cannot be argued that third and fourth year students were equivalent to the control group of non‑selected students. In fact, attrition rates were approximately 30 percent annually. Attrition students had lower test scores than those who continued to participate in the voucher schools (Witte et al., 1994, Table 1.8, p. 23). That is, consistent with the general literature on school mobility, in that students who persisted in the same school were superior to those who moved back to the Milwaukee Public Schools. The Peterson et al . analysis, then, compares the stable group who persisted for three or four years in the same school (superior in achievement to those who did not persist) with all non‑selected students. The persistent voucher students were a superior subset, not a random subset, of the original applicant pool. Therefore it is invalid to compare them with the original non‑chosen group and to conclude that the higher achievement scores of voucher students in their third and fourth year were due to a schooling effect. In fact, it is probable that Witte has overstated the comparative achievement of the choice students in the third and fourth years by not providing a statistical control for mobility to compare only with public school students who had been in the same school for three to four years. A more elaborate analysis of this comparison using instrumental variables and Heckman adjustments for self‑selection finds no difference in achievement, and the ³effect² that Peterson, Greene, and Du report is created by a dramatic loss in both numbers of students and achievement scores of remaining students in the control group rather than a comparative rise in achievement of those in choice schools (Witte 1996).


Differences on Other Outcomes


My own reading of the body of studies comparing student achievement in public and private schools is that there is no difference for equivalent students or that differences are trivial. In other respects there may be a private school advantage for some groups. Although Sanders (1996) found no difference in achievement between public and private schools for Hispanics and African‑Americans with statistical controls for family background, Neal (1995) found that urban minority students attending Catholic secondary schools have considerably higher graduation rates than comparable public school students and higher college graduation rates. He attributes these results to the particularly poor public schools that are available to this group of students. It should be noted that Bryk and Lee (1988) found that students in Catholic high schools are more likely to be assigned to an academic track than in public high schools. Neal estimated that the greater educational attainments of Catholic School minority students in urban settings lead to earnings that are eight percent higher than for comparable students in urban public schools. Evans and Schwab (1993) also found greater high school graduation rates and college enrollment rates for Catholic school students. Finally, as expected, parental satisfaction with the schools their students are attending seems to be consistently higher for parents of students enrolled in private schools relative to prior public schools attended (Witte et al. 1995; Moe 1995a).






Advocates for vouchers believe that the advent of marketplace choice in education will level the playing field by providing options in education for those who are most disadvantaged by the present educational system. According to this view, children from middle and higher socioeconomic families can choose to live in the best neighborhoods with good schools or to send their children to private schools. In contrast, children from poorer families are captives in neighborhood schools in inner cities or rural areas without the ability to pursue alternatives. If alternatives are provided, large numbers will use their vouchers to choose better schools, requiring neighborhood schools to improve or putting them out of business if they fail to improve. This view is reinforced by the fact that surveys of poor and minority families show that they favor choice even more than other groups (Lee, Croninger, and Smith 1996). But, such a scenario assumes that the poor will take advantage of a choice system to outflank their local public schools in behalf of better education for their students.


In this section I will suggest that the evidence consistently supports the following conclusions. (1) Choosers will be more advantaged both educationally and economically than non‑choosers, those who do not actively choose schools for their students, but relegate them to their assigned school. (2) For choosers an important criterion of choice will be the socioeconomic status (SES) of other students where the most preferred schools will be those enrolling more advantaged students leading to increased segregation. And (3) both peer and contextual effects of higher SES students have positive effects on achievement, leading to the conclusion that inequalities in educational outcomes are likely to be exacerbated by vouchers.


(1) Who Chooses?


Choice systems may lead to two types of ³cream skimming². In the first type, families who are better off may be more likely to take advantage of school choice than those who are worse off because of better access to information, greater ability to afford transportation, a higher penchant to exercise educational alternatives, and greater generic experience with choice and alternatives. A second type of cream skimming refers to the tendency of schools to seek and choose students from families of higher SES and with higher previous educational accomplishments(as modeled by Nechyba 1996 and corroborated empirically in Belgium by Vandenbergh 1996). To some degree, the second of these can be reduced through strong regulation requiring random selection among student applicants, although both the political viability of such a regulation and its practical reinforcement may be problematic. But the first type of cream skimming, which is a consequence of voluntary choice by families, may be endemic to educational choice systems as the empirical literature suggests. In both public choice and voucher‑type systems it appears that those who exercise the choice option are more likely to be of higher SES and to have higher achievement scores than those who continue to attend their assigned schools (Archbald 1988; Rubenstein, Hamar, and Edelman 1992; Martinez, Godwin, and Kemerer 1996; Witte and Thorn 1996). Ambler (1994) found such cream skimming in educational choice participation for both England and France. Willms and Echols (1992) found the same to be true in Scotland as did Vandenberghe (1996) in Belgium. Archbald (1988) and Moore and Davenport (1990) found that magnet schools in the large cities that rely on choice to reduce school segregation tend to attract higher socioeconomic students rather than a random mix. Even when participation was restricted to families with incomes no higher than 1.75 times the poverty level, parents of choice applicants in the Milwaukee voucher experiment had considerably more education and parental involvement than the average parent of children in the Milwaukee Public Schools. In a publication on four private voucher plans in the U.S., Terry Moe (1995 a), one of the most knowledgeable supporters of vouchers, Moe (1995b), concluded that the problem represents a serious challenge:


Skimming is rooted in the calculus of choice itself: in the utility functions of parents, the information they bring to bear, and their income constraints. Some parents put a higher value on education than others and so are willing to give up more to secure quality schooling for their children. Some parents have more information than others and thus know more about what schools are available and how good they are. And some parents have higher incomes than others and so are better able to acquire good information and afford good schools. Unrestricted choice, then, may well lead to selection effects with a class bias (Moe 1995b).


Moe concludes that such skimming can be reduced through restricting choice to those who are most disadvantaged as well as making sure that voucher plans are ³Šsocially engineered through appropriate institutional design² to insure greater social equity (Moe 1995b:24). Presumably such design features would include more comprehensive and interactive systems of information as well as adequate transportation within the educational marketplace. However, such social equity features may have a high cost, an issue that is addressed later, and it is not clear that society is willing to pay these costs.


Moe concludes, correctly in my view, that the issue is not whether there is skimming, but whether the skimming will be worse than the present public system where students tend to be largely segregated in schools with students similar to themselves (Moe 1995b: 24). Moe is also correct in suggesting that by restricting the choice only to those most disadvantaged by the present system, social equity would be likely to improve over the present system, although I have doubts that political dynamics would support that solution over the long run. But, what is the likely impact of a more extensive system of choice on student segregation?


(2) Impact on Socioeconomic Segregation

Many observers have been concerned about the consequences for segregation of educational choice. It has been argued even that one of the direct purposes of choice is to increase segregation according to religious and cultural differences to create communities of human capital through common ³social capital² (Coleman 1988). Understandably, private schools tend to specialize in market niches by creating differentiated rather than generic products in order to appeal to clientele with particular political, philosophical, educational, and religious orientations. This has been evident in Holland where publicly‑funded private schools accounted for almost three‑fourths of all enrollment in 1980, and where over 90 percent of these schools were sponsored by religious groups (Ambler 1994:468‑469). Surely this leads to greater religious segregation than would be found if schools were based strictly on attendance boundaries.


But, to what degree does the fact that choosers tend to be from higher socioeconomic (SES) origins lead to greater SES segregation of students? Since 1982 Scotland has permitted parents to request public schools other than those to which their students are designated by public authorities. By the late eighties about 9 percent of entering secondary students attended a school outside of their designated areas with the numbers rising to 11‑14 percent in urban areas according to sources cited by Willms and Echols (1992:340). By the early nineties about 15‑18 percent of pupils in the most urban areas had requests for other schools made on their behalf, with some areas experiencing requests for more than 50 percent of students (Willms 1996: 140). Willms and Echols (1992: 344) found that parents requesting non‑designated schools had significantly more education and higher occupations than those who kept their children in designated schools, as much as .35 of a standard deviation higher. Average SES of pupils in the chosen schools was about .25 of a standard deviation greater than in designated schools. Thus, choosers tended to have higher SES than non‑choosers and to request schools with higher SES than their designated schools. Willms and Echols (1992) conclude that this is a major criterion of selection because higher SES schools tend to have high achievement scores, although not necessarily high value‑added, which should be a more central concern. But, high SES of the student body of a school is easily observable, whereas direct measures of value‑added are not. Overall the effect of this choice process was to increase student segregation by SES of Scottish students between 1985 and 1991 (Willms 1996). In response to Moe¹s (1995b) question on whether choice increases student SES segregation, the answer in the Scottish case seems to be clearly affirmative. An analysis of Belgium shows even greater student segregation under choice (Vandenberghe 1996), but this probably includes both types of cream‑skimming set out above.


(3) Consequences of Increased Segregation


Willms and Echols (1992) proceeded to estimate the effects of schools on student achievement and found that parents tended to choose schools with high achievement scores and student SES, but not schools with high ³value‑added² results after taking account of student intakes. That is, the superior school ³effects² were mainly due to a higher SES student body rather than school effectiveness with a given group of students.


In turn, it appears that increased student segregation by SES will increase inequality of opportunity because aggregate SES of the school seems to have an impact on achievement independent of the impact of the student¹s individual SES on her achievement (Arnott and Rowse 1987; Link and Mulligan 1991; Henderson, Mieszkowski, and Sauvageau 1978; Rumberger and Willms 1992; Shavit and Williams 1985; Summers and Wolfe 1977; Willms 1986). It is not clear whether this effect comes from the influence of peers, school climate, or teaching conditions or differences in teacher expectations and curriculum (Dreeben and Gamoran 1986; Dynarski, Schwab, and Zampelli 1989; Rumberger and Willms 1992). Most of us who teach will naturally raise the level of challenge if we believe our students are well‑prepared than if we believe that they are not. At the same time, students with high educational ambitions may create an atmosphere that supports those norms among their peers. Whatever, the cause, the contextual effect of SES seems well‑established. Further, it suggests rising inequalities in achievement between students of lower and higher SES as they become increasingly segregated in schools with students like themselves. As higher SES students leave lower SES school environments for higher SES schools, their achievement will rise. But, their departure reduces the aggregate SES of the schools that they leave with a resulting decline in the achievement of the remaining students in those schools. It is important to keep in mind that this is a zero‑sum game because there are only a fixed number of high SES student enrollments at any one time. Thus, not all potential choice students can benefit from a high SES school environment, as it will be non‑reproducible beyond a relatively limited number of schools.


Further, the negative effects on low SES students are likely to be greater than the gains of high SES students. The negative impact of segregation on the achievement of students in low tracks (largely low SES) is not offset by the higher achievement of students in high tracks (largely high SES) according to statistical analysis by Gamoran and Nystrand (1994). This is also the conclusion of Henderson, Mieskowski and Sauvageau (1978) whose results suggest that overall achievement is higher in heterogeneous rather than segregated school environments because any loss of achievement by the higher groups is more than made up by the higher achievement of the lower groups. Summers and Wolfe (1977) also found that less able students benefit more from this effect, while higher ability students are less affected. Thus, if choice leads to greater SES segregation, the impact on achievement will be to reduce aggregate student achievement unless gains through school competition will offset the achievement losses due to increased student segregation by SES. However, existing empirical findings comparing public and private school achievement are not promising in this regard.


Increased segregation has other consequences as well, particularly on preparation of students for democratic life. Effective participation in a democracy requires a willingness to tolerate diversity as well as exposure to a common set of values and knowledge. Research on political socialization has shown that tolerance for other points of view is related to the degree to which different children are exposed to different viewpoints on controversial subjects in both home and school (Torney‑Purta 1984). It also requires a common core of experiences to create citizens who can function democratically (Gutmann 1987: pp. 50‑64). But, by segregating students to a greater extent than existing schools according to SES, religion, and other dimensions, the exposure to diversity and to a common core of experiences is seriously undermined.





Of course, to economists and society it is not only the educational effects of vouchers that should be considered, but also their costs. There are two levels of costs in assessing the voucher alternative. First, there is the cost at the school level. That is, for a given result in school effectiveness, what are the relative costs of public versus private schools? Second, there is the public (and overall) cost of the administrative and service infrastructure necessary to support a voucher marketplace relative to the present system which is considerably more centralized at both state and district levels. In this section, we will consider what is known about each. Costs of Private Versus Public Schools


Even if private and public schools are about equally effective in producing student achievement, observers have suggested that non‑elite private schools entail considerably lower costs than public schools. For example, Peterson and Noyes (1996) claim that Milwaukee voucher schools were receiving only half as much for each student as the Milwaukee Public Schools. Therefore, they assert that even if the voucher schools are no more effective than the Milwaukee Public Schools, they are twice as efficient in the use of society¹s scarce resources. A publication of the Cato Institute, a Libertarian organization, makes the same point by comparing the tuition at private schools in several localities with total per pupil expenditure in the public school system in those areas (Boaz and Barrett 1996). This conclusion is also stated in other publications that advocate a market approach to education (e.g. American Enterprise Institute, 1978; West 1981).


But, a comparison of public school expenditures with private school tuition is not a valid approach to comparing costs. The problem is that the finance and service mix of public and private schools is quite different. For example, tuition is a much poorer proxy for the overall costs of private schools than is per‑pupil expenditure as a measure of public school costs. Most private schools rely heavily on supplementing tuition with fund‑raising events, special student fees for extra activities, financial contributions, and in‑kind contributions. In addition, those that are sponsored by religious organizations (the majority of private schools) receive donated or subsidized facilities and many Catholic schools are staffed partially by teaching clergy whose ³salaries² understate substantially their true market value (Bartell 1968). The result is that tuition charges cover only a portion of the overall costs of private education. Although the public sector costs are not a complete measure of the costs of public education, especially because of their treatment of capital expenses, they are far more complete in comprising all of their resource inputs at market prices than is the tuition measure for private schools.


But, beyond this the service mix is very different between public and private schools. For example, few private schools provide special educational services for the handicapped, while public schools are required to do so by law. Average costs of education for each special education student have been estimated to be almost 2.5 times the average cost for the non‑handicapped student (Chaikind, Danielson, and Brauen 1993). In New York City it was estimated that the cost was four times that of regular students in 1993 (Lankford and Wyckoff 1996: 231). Moreover, special education students represent about 12 percent of all students nationally, but are concentrated almost completely in public schools (with the exception of those in very high‑cost, specialized private schools mainly subsidized by government).


Further, the comparison of average per‑pupil expenditures for public schools includes other services not provided by private schools. For example, most private secondary schools do not provide vocational education, a course of study which varies from two to more than five times the cost of regular education, depending upon the specialization (Hu and Stromsdorfer 1979). Transportation services and food services are included in the total for public school expenditures, but private schools normally charge extra fees for these services. Finally, the tuition charges that are usually compared with public school expenditures are those for elementary schools (typically parochial, Catholic schools), while the public school figures comprise both elementary schools and the more costly secondary schools.


To test the assertion that the Milwaukee voucher schools had a per‑pupil cost that was half of that of comparable public schools, I contacted the Milwaukee Public Schools to obtain per‑pupil expenditure breakdowns (Haselow 1996). Voucher schools were receiving $ 4,373 per student in 1996‑97. The Milwaukee Public Schools had an estimated budget for the same year of $ 7,628 per student, but this amount included many services not required or provided by the voucher schools. For example, the voucher students were enrolled in kindergarten to eighth grade schools, while the Milwaukee Public Schools total included the more expensive high school students as well. Voucher schools did not enroll expensive special education students and did not provide vocational education, transportation or the extensive food and health services provided by the Milwaukee Public Schools.


A more appropriate comparison is to compare site‑based expenditures in Milwaukee Public Schools with the voucher schools. Milwaukee provides such school‑based allocations according to pre‑determined personnel ratios and other factors. (Secondary schools and middle schools have the largest allocation per student, about $3,815 for each middle school student and $ 3,635 for each high school student.) The Milwaukee Public Schools budgeted $ 3,469 per student in K‑8 schools and $ 3,042 per student in elementary schools. If we compare the voucher with the per‑pupil amount for K‑8 schools, the voucher schools received about $ 1,000 more per student than the comparable Milwaukee Public Schools for the 1996‑97 school year. On the basis of costing experience for public schools, it is estimated that facilities costs on an annualized basis are about 10 percent of total expenditures in what is a labor‑intensive enterprise, closing about half of the gap, but still favoring the voucher schools. The most reasonable conclusion is that voucher schools in Milwaukee are receiving at least comparable allocations per student to those of the Milwaukee Public Schools, once the service mix is accounted for.


Of course, this raises the question of what accounts for the other costs of the school district that are not allocated specifically to the individual school sites. The costs of special education services are budgeted at the central level; with over 12 percent of the students in these categories and excess costs averaging about 150 percent more than regular education, this accounts for almost $1,400 per student when averaged among all students. Transportation costs including those for carrying students to voucher schools, which are paid by the district, average about $565 averaged across all the students of the district, and much more per transported student. In addition, there are the higher costs of secondary schools, food services, health services, and capital costs. Overall costs of central office administration are only about 3 percent of per‑student expenditures. While this does not constitute a precise cost‑accounting for the two sets of schools, it appears that the costs of similar services at the school site may actually be higher at the Milwaukee voucher schools, although the most prudent conclusion at this level of detail is to suggest that they are comparable. Claims that the public schools cost twice as much as comparable private schools in other settings should also be subjected to careful scrutiny. My guess is that such cost comparisons would show that even in the least efficient school districts, costs for similar services are not found to be even close to the two‑fold figure that is commonly cited.


Cost of a Voucher System


A shift from the prevalent system of state finance and governance of education to one based upon educational vouchers will require a profound transformation of institutions required to support the schooling system. In particular, it will require far more transaction costs as states must deal with individual schools and students rather than districts. For example, in California a system of vouchers would require state authorities to keep records and administer vouchers to almost 6,000,000 youngsters in place of dealing with about 1,000 local school districts. In order to assure adequate access to alternatives, it is probable that information centers would need to be established to enable parents to make informed choices, and an expanded system of publicly funded transportation would need to be incorporated. In addition, some type of system of adjudication would need to be provided for parents who wanted a partial refund of vouchers in order to change schools during the academic year. Finally, a state system of monitoring and assessment would be needed to establish voucher eligibility of both students and schools.


The estimation of the costs of a voucher system to replace existing systems of schooling cannot be done without specification of the particular voucher plan that is being considered; the system that it will replace; the setting where it will be applied; and assumptions about the behavior of schools and families under the voucher approach. Cyrus Driver and I (Levin and Driver, 1996; Levin and Driver, forthcoming) have attempted to address these issues and to estimate illustrative costs in five areas associated with a voucher system. These include: (1) accommodating additional students; (2) record keeping and monitoring; (3) transportation; (4) information; and (5) adjudication of disputes.


Only a summary of results will be shown here, so the two source documents should be reviewed for the details underlying the calculations. Cost estimates are generally for 1995 with a few exceptions.


(1) Accommodating Additional Students


If all private school students were to participate with the full range of services provided by the public schools at the average per‑pupil expenditure nationally, the added cost would be about $ 33 billion annually. If only 75 percent would be eligible because some schools would not wish to participate in a plan with government oversight, the cost would be almost $ 25 billion annually. Or, if the voucher were set at 80 percent of public school costs because existing private school students are less likely to need services for special education, compensatory education, bilingual education and so on, these amounts would be about $ 26 billion for all existing private school students and $ 20 billion for a 75 percent transfer rate into the voucher system.

(2) Record‑Keeping and Monitoring


A voucher system will require extensive record keeping and monitoring systems for several reasons. First, every child required to be in school under compulsory attendance laws and those continuing their education through high school graduation will need to be monitored to ensure that they are in a school approved to use the voucher. Second, children with different educational needs will be eligible for different vouchers (e.g. students at each level of schooling, handicapped students, disadvantaged students, language minority students). Students will need to be evaluated in terms of needed services and the appropriate magnitude of the voucher. Third, only schools that meet ³approval² standards will redeem vouchers, so schools must be evaluated, certified, and monitored for eligibility. (In 1995‑96 two of the Milwaukee voucher schools closed in mid‑year, stranding the students and relegating their involuntary return to the Milwaukee Public Schools. At the time of the writing of this article, criminal charges were pending because of alleged financial manipulations by the schools¹ operators.) The costs of monitoring and accreditation would be likely to be particularly high because we would expect about twice as many schools under a voucher plan, given that private schools tend to be about half the average size of existing public schools (Chambers 1981). Using the social security system as an analogy, it was estimated that even with cost savings from dismantling the present system, there would be a net additional cost for record‑keeping and monitoring of about $ 2.5 billion nationally. This figure does not include the costs of accrediting for eligibility and monitoring the approximately 200,000 schools we would expect under a market approach, a serious omission and understatement of costs, because we lacked an analogous data base that might be used for such cost estimation.


(3) Transportation


Transportation costs would be expected to be higher under a voucher system than the present system for two reasons. First, the advent of choice should lead to more students attending schools outside of their immediate neighborhoods. Second, the routes are likely to be of lower density and regularity in terms of pickups and deliveries. About 60 percent of U.S. public school students are bussed at present, and we assumed that this would rise to about 80 percent of public and private school students. After scrutinizing a large number of travel modes and examining existing costs for school transportation, we estimated that additional transportation costs would be about $ 42 billion based upon an additional 13.3 million students being bussed and a rise in costs from about $ 415 per student in 1992‑93 to about $ 1,500 per student. It should be noted that bussing costs for desegregation purposes in the St. Louis area are at a level of about $ 2,000 per student per year, a level that was also reported by Milwaukee for its interdistrict bussing program (Haselow 1996). It should also be noted that a telephone survey of private schools in New York found transportation fees to be in this range and higher, with such transport often undertaken through competitive bidding by private firms. And the use of smaller buses and larger catchment areas is even more costly as evidenced by both commercial cost estimates and the experiences in transporting children needing special education.


(4) Information Costs


In order to make informed choices, parents need information on alternatives. At a minimum, families need to know what choices are available and the appropriateness of particular choices for their children. They also need information on such matters as school philosophy, curriculum, personnel, facilities, test scores, student placements after graduation, registered complains and their nature, and turnover rates among students. Using a very modest approach such as one used for a choice program in Massachusetts, we estimated the per pupil cost at about $ 38 per year or about $ 1.8 billion nationally. It should be noted that this probably understates seriously the cost for a highly effective information system that would engage the poor, minorities and immigrants, groups that have been least likely to participate in choice systems, but we do not have a knowledge base for estimating the cost of a more ambitious system.


(5) Costs of Adjudication


Some families will choose schools that they later find are inappropriate for their students. Schools may also wish to suspend or discharge students who do not meet certain standards. In cases like these there may be issues of due process as well as the right of a student to obtain an additional partial‑voucher to use for another school if the original one has been redeemed. There may also be conflicts about whether a student is getting an appropriate voucher for the educational services that the family believes are required and other challenges to the voucher agency. In all of these cases a means of adjudication must be available to quickly resolve the dilemma so that a child¹s education is not seriously interrupted. Using cost data from mediation and due process hearings for special education and assuming that only 1 percent of students will require adjudication in any given year, we estimate the costs of adjudication at about $ 1.8 billion.


Total Costs of a Voucher System


These are first estimates of public costs of a voucher system, and they total almost $ 73 billion, about $ 1,500 per student or an additional 25 percent of the public educational budget nationally under a mid‑range set of assumptions (Levin and Driver, forthcoming). Well over half of this represents a shift in the cost burden from the private to the public sector rather than higher social costs. For example, accommodating present private school students with vouchers and transportation of existing private school students who are presently transported represents a shift, not a higher cost. The net costs of record‑keeping and monitoring may be slightly over‑stated if we have not fully accounted for the savings from existing practices, but this category of costs accounted for less than 4 percent of the total. Information costs are surely too low, and we have not included, at all, the costs of accrediting and monitoring schools to be eligible to redeem vouchers. We conclude that the shift from the existing system to a voucher system with a well‑functioning school marketplace in which adequate transportation and information is provided will demand considerable additional public resources for education beyond those that will need to be allocated for educational vouchers and instructional services.




During the last five years we have come a long way in acquiring evidence that is pertinent to the consideration of educational vouchers, although there are still many gaps in our knowledge base. Unfortunately, policy debates on vouchers are largely devoid of references to the available evidence or are limited to citing only a ³favored² study that supports a particular perspective. I have suggested here that there is a considerable consensus arising from the available corpus of evidence on the first two issues set out in this paper, and at least a first approximation on the cost issues. I want to conclude by stating that nothing in this paper should give much comfort to those who might wish to defend the status quo. In my view, considerable gains in educational efficiency are possible, whether vouchers are the answer or some other type of system reform. Evidence of this claim can be found in a school reform movement that has extended to about 1,000 public, Catholic, and charter schools in some 40 states, where we have demonstrated that substantial improvements in educational results can be obtained (Levin forthcoming).






Ambler, John S. (1994) ³Who Benefits from Educational Choice? Some Evidence from Europe,² Journal of Policy Analysis and Management, Vol. 13, No. 3, pp. 454‑476.


Alexander, Karl L. and Aaron Pallas (1985) ³School Sector and Cognitive Performance: When is a Little a Little? Sociology of Education, Vol. 58, pp. 115‑128.


American Enterprise Institute (1978) Tuition Tax Credits and Alternatives (Washington, D.C.: American Enterprise Institute).


Archbald, Douglas (1988) Magnet Schools, Voluntary Desegregation, and Public Choice Theory: Limits and Possibilities in a Big City School System (Madison, WI: School of Education, University of Wisconsin‑Madison).


Arnott, R. and J. Rowse (1987) ³Peer Group Effects and Educational Attainment,² Journal of Public Economics, Vol. 32, pp. 287‑305.


Bartell, Ernest S.J. (1968) Costs and Benefits of Catholic Elementary and Secondary Schools. (Notre Dame IN: Notre Dame Press).


Boaz, David and R. Morris Barrett (1996) What Would a School Voucher Buy? The Real Cost of Private Schools, Cato Briefing Paper # 25 (Washington, D.C.: The CATO Institute).


Chambers, Jay G. (1981) ³An Analysis of School Size Under a Voucher System,² Educational Evaluation and Policy Analysis, Vol. 3, pp. 29‑40.


Coleman, James S. (1988) ³Social Capital in the Creation of Human Capital,² American Journal of Sociology , Vol. 94, Supplement: pp. 95‑120.


Coleman, James S., Thomas Hoffer, and Sally Kilgore (1982) High School Achievement: Public, Catholic, and Private Schools (New York: Basic Books).


Dreeben, Robert and Adam Gamoran (1986) ³Race, Instruction, and Learning,² American Sociological Review, Vol. 51, pp. 660‑669.


Dynarski, Mark, R. Schwab, and E. Zampelli (1989) ³Local Characteristics and Public Production: The Case of Education,² Journal of Urban Economics, Vol. 26, pp. 250‑263.


 Echols, Frank H. and J. Douglas Willms (1995) ³Reasons for School Choice in Scotland,² Journal of Educational Policy, Vol. 10, No. 2, pp. 143‑156.


Evans, William N., Wallace E. Oates, and Robert M. Schwab (1992) ³Measuring Peer Group Effects: A Study of Teenage Behavior,² Journal of Political Economy, Vol. 100, No. 5, pp. Pp. 996‑991.


Evans, William N., and Robert M. Schwab (1993) ³Finishing High School and Starting College: Do Catholic Schools Make a Difference?² Department of Economics, University of Maryland.


Friedman, Milton (1955) ³The Role of Government in Education,² In R. A. Solo (ed.) Economics and the Public Interest (New Brunswick, NJ: Rutgers University Press).


(1962) ³The Role of Government in Education,² In Capitalism and Freedom (Chicago: University of Chicago Press), Chap. VI.


Gamoran, Adam (1991) ³Schooling and Achievement: Additive Versus Interactive Models,² In Steve W. Raudenbush and J. Douglas Willms (ed), Schools, Classrooms, and Pupils:International Studies of Schooling from a Multilevel Perspective (San Diego: Academic Press), pp. 37‑51.


(1996) ³Student Achievement in Public Magnet, Public Comprehensive, and Private City High Schools,² Educational Evaluation and Policy Analysis, Vol. 18, No. 1 (Spring), pp. 1‑18.


_____________ and M. Nystrand (1994) ³Tracking, Instruction, and Achievement,² International Journal of Educational Research, Vol. 21, No. 5, pp. 217‑231.


Goldberger, Arthur S. and Glen G. Cain (1982) ³The Causal Outcomes in the Coleman, Hoffer and Kilgore Report,² Sociology of Education, Vol. 55, pp. 103‑122.


Greene, Jay P., Paul E. Peterson, and Jiangtao Du (1996), ³The Effectiveness of School Choice in Milwaukee: A Secondary Analysis of Data from the Program¹s Evaluation,² Paper presented at Annual Meetings of the American Political Science Association, San Francisco (August 30).


Gutmann, Amy (1987) Democratic Education(Princeton: Princeton University Press). Haertel, Edward H., Thomas James, and Henry M. Levin, eds. (1987) Comparing Public and Private Schools: School Achievement, Vol 2 (New York: The Falmer Press).


Haselow, Douglas (1996) ³Responses to Letter from H. M. Levin, November 20, 1996² Milwaukee Public Schools. Hoffer, Thomas, Andrew Greeley, and James S. Coleman (1985) ³Achievement Growth in Public and Catholic Schools,² Sociology of Education, Vol. 58, pp. 74‑97.


Hoxby, Caroline Minter (1996) ³The Effects of Private School Vouchers on Schools and Students,² in Helen F. Ladd (Ed.), Holding Schools Accountable (Washington, D.C.: The Brookings Institution, 1996), pp. 177‑208.


Inkeles, Alex (1966) ³Social Structure and the Socialization of Competence,² Harvard Educational Review, Vol. 36, pp. 265‑283.


Kane, Thomas J. (1996) ³Comments on Chapters Five and Six,² in Helen F. Ladd (Ed.), Holding Schools Accountable (Washington, D.C.: The Brookings Institution), pp. 209‑217.


Lankford, Hamilton and James Wyckoff, ³The Allocation of Resources to Special Education and Regular Instruction,² in Helen F. Ladd (ed.), Holding Schools Accountable (Washington, D.C.: The Brookings Institution, 1996), pp. 221‑257.


Lee, Valerie E., Robert G. Croninger, and Julia B. Smith (1996) ³Equity and Choice in Detroit,² In Bruce Fuller and Richard F. Elmore (eds), Who Chooses? Who Loses? (New York: Teachers College, 1996), pp. 70‑91.


Levin, Henry M. (1968) ³The Failure of the Public Schools and the Free Market Remedy,² The Urban Review, Vol. 2, pp. 32‑37.


_____________ (1980) ³Educational Vouchers and Social Policy,² In Care and Education of Young Children in America, Ron Haskins and James Gallagher (eds) (Norwood, NJ: Ablex Publishing Col, 1980).


_____________ (1987) ³Education as a Public and Private Good,² Journal of Policy Analysis and Management, Vol. 6, pp. 628‑641.


_____________ (1991) ³The Economics of Educational Choice,² Economics of Education Review, Vol. 10, No. 2, pp. 137‑158.


_____________ (forthcoming) ³Raising School Productivity: An X‑Efficiency Approach,² Economics of Education Review (forthcoming).


_____________ and Cyrus E. Driver (1996) ³Estimating the Costs of an Educational Voucher System,² In William J. Fowler, Jr. (ed) Selected Papers in School Finance 1994, NCES 96‑068 (Washington, DC: U.S. Department of Education, National Center for Educational Statistics), pp. 63‑88.


_____________ and _____________ (forthcoming) ³Costs of an Educational Voucher System,² Educational Economics. Available in pre‑publication form from authors, CERAS 109, Stanford University, Stanford, CA 94305‑3084.


Link, Charles and James Mulligan (1991) ³Classmates¹ Effects on Black Student Achievement in Public School Classrooms,² Economics of Education Review, Vol. 10, pp. 297‑310.


Martinez, Valerie, Kenneth Godwin, and Frank R. Kemerer (1996) ³Public School Choice in San Antonio: Who Chooses and with What Effects?² In Bruce Fuller and Richard Elmore (eds), Who Chooses? Who Loses? (New York: Teachers College Press), pp.50‑69.


Moe, Terry M. (1995a) Private Vouchers (Stanford, CA: Hoover Institution Press). ___________ (1995b) ³Private Vouchers,² in Terry M. Moe (ED), Private Vouchers (Stanford, CA: Hoover Institution Press), pp. 1‑40.


Moore, Donald and Susan Davenport (1990) ³School Choice: The New School Sorting Machine,² In William Boyd and Herbert Walberg (eds) Choice in Education: Potential and Problems (Berkeley, CA: McCutcheon Publishing Co.), pp. 187‑224.


Murnane, Richard (1986) ³Comparisons of Private and Public Schools: The Critical Role of Regulations,² in Daniel Levy (ed.), Private Education: Studies in Choice and Public Policy (New York: Oxford University Press).


Neal, Deal (1995) ³The Effect of Catholic Secondary Schooling on Educational Attainment,² NBER Working Paper Series, Working Paper 5353 (Cambridge, MA.: National Bureau of Economic Research).


Nechyba, Thomas (1996) ³Public School Finance in a General Equilibrium Tiebout World: Equalization Programs, Peer Effects and Private School Vouchers,² (Department of Economics, Stanford University).


Alan Peshkin (1986) God¹s Choice: The Total Fundamentalist Christian School (Chicago: University of Chicago Press).


Peterson, Paul E and Chad Noyes (1996) ³Under Extreme Duress, School Choice Success,² Occasional Paper 96‑1, Program in Education Policy and Governance (Cambridge, MA: Harvard University).


Rubenstein, M., R. Hamar, and Nancy Adelman (1992) Minnesota¹s Open Enrollment Option, U.S. Department of Education Contract No. 89089001, (Washington, DC: Policy Studies Associates).


Shavit, Yehuda and R. A. Williams (1985) ³Ability Grouping and Contextual Determinants of Educational Expectations in Israel,² American Sociological Review, Vol. 50, pp. 62‑73.


Summers, Anita and Roberta Wolfe (1977) ³Do Schools Make a Difference?² American Economic Review, pp. 639‑652.


Rumberger, Russell and Katherine A. Larson (1996) ³The Impact of Student Mobility on High School Completion², Paper presented at Annual Meetings of the American Educational Research Association, New York, (April).


_______________ and J. Douglas Willms (1992) ³The Impact of Racial and Ethnic Segregation on the Achievement Gap in California High Schools, Educational Evaluation and Policy Analysis, Vol. 14, No. 4, pp. 377‑396.


Torney‑Purta, Judith (1984) ³Political Socialization and Policy: The United States in a Cross‑Cultural Context,² In H. Stevenson and A. Siegel (eds), Child Development Research and Social Policy (Chicago: University of Chicago Press).


Vandenberghe, Vincent (1996) Functioning and Regulation of Educational Quasi‑Markets, Nouvelle serie‑No. 283 (Louvain‑la Neuve, Belgium: Faculte Des Sciences Economiques, Sociales Et Politiques, Universite Catholique de Louvain).


West, Edwin G. (1981) The Economics of Educational Tax Credits (Washington, D.C.: The Heritage Foundation). Willms, J. Douglas (1983) ³Do Private Schools Produce Higher Levels of Academic Achievement? New Evidence for the Tuition Tax Credit Debate,² In Tom James and Henry M. Levin (ed.), Public Dollars for Private Schools (Philadelphia, PA: Temple University Press).


_________________ (1985) ³Catholic School Effects on Academic Achievement: New Evidence from the High School and Beyond Follow‑up Study,² Sociology of Education, Vol. 58, pp. 98‑114.


________________ (1986) ³Social Class Segregation and Its Relationship to Pupils¹ Examination Results in Scotland,² American Sociological Review, Vol. 51, pp. 224‑241.


 ________________ (1996) ³School Choice and Community Segregation: Findings from Scotland,² in Alan C. Kerckhoff (ed.) Generating Social Stratification: Toward a New Research Agenda (Boulder, CO: Westview Press).


________________ and Frank Echols (1992) ³Alert and Inert Clients: The Scottish Experience of Parental Choice of Schools,² Economics of Education Review, Vol. 11, No. 4, pp. 339‑350.


Witte, John F. (1997) ³Achievement Effects of the Milwaukee Voucher Program,² Paper presented at the 1997 Annual Meetings of the American Economic Association (New Orleans: January 4).


___________ (1992) ³Private School Versus Public School Achievement: Are There Findings That Should Affect the Educational Choice Debate,² Economics of Education Review, Vol. 11, pp. 371‑394.

Witte, John F., Christopher A. Thorn, Kim M. Pritchard, and Michelle Claibourne (1994) The Fourth Year Report: The Milwaukee Parental Choice Program. (Madison, WI: Wisconsin Department of Public Instruction).


Witte, John F. and Christopher A. Thorn (1996) ³Who Chooses? Voucher and Interdistrict Choice Programs in Milwaukee,² American Journal of Education, Vol. 104, pp. 186‑217.


Witte, John F., Troy D. Sterr, and Christopher A. Thorn (1995) Fifth Year Report: Milwaukee Parental Choice Program, (Madison, WI: Robert M. La Follette Institute of Public Affairs, University of Wisconsin—Madison). December 1996




Copyright 1996. Readers may redistribute this article to other individuals for noncommercial use, provided that the text and this notice remain intact. This article may not be resold, reprinted, or redistributed for compensation of any kind without prior written permission from the author.