The discussion that follows relates the characteristics of effective market education to voucher and tax-credit programs. In doing so, it points to specific features that both sorts of programs must have if they are to produce the results expected of them.

Budgetary Considerations

Properly designed market education programs should impose lower costs on state budgets than do current public school systems. Holding the kind and condition of school facilities constant, private schools are, on average, less costly to operate than public schools.[45] That means voucher, personal use tax credit, and donation scholarship amounts can all be set lower than existing per-pupil spending in public schools. The savings that accrue from choice programs can be assigned to whatever use voters and state legislators deem best. In the present climate of state budgetary crises, these savings will be most welcome.

From a political standpoint, it is tempting to redirect at least some of the savings from a choice program back into the traditional public school system. This is widely believed to increase public support for market education programs, but choice proponents should proceed with caution if they choose to follow this path. There is empirical evidence from the Chilean pseudo-voucher program that giving extra money to poorly performing public schools only makes them less efficient and less responsive to competition. Public schools that do not receive this extra funding tend to be more efficient and responsive to competition, though they are still significantly less efficient and responsive to competition than are private schools. This, of course, makes perfect sense. The schools with the most to lose from failing to meet their students’ needs tend to make the greatest effort to satisfy those needs. Schools that receive extra funding regardless of how well they are performing are under less pressure to perform well, and hence their performance deteriorates.[46] So while redirecting some choice program savings back into the traditional public school system may improve the prospects for passage of a choice program, it is likely to diminish the effectiveness of that program.

An Absence of Mandatory Curricula and Testing

In practice, parental choice and school autonomy are both dependent on education regulations being kept to a minimum. One of the most damaging sorts of government intervention in the education market is the imposition of a mandatory curriculum. No single official curriculum can adequately reflect the diversity of parental demands and preferences. A mandatory curriculum undermines the value of parental choice by homogenizing the options from which parents can choose. It limits the ability of schools to innovate, respond to variations in parental demand over time, and target particular niche clienteles. In other words, it takes the market out of market education.

Government imposed testing regimes are similarly harmful, for a number of reasons. First, they tend to drive the curriculum implicitly, as schools focus on the high-profile tested subjects and sideline those that are not tested. This brings about the same market-stifling conditions as an explicit official curriculum, and it is precisely what occurred, for example, under the English “Payment by Results” testing program of the 1870s.[47]

Many education reformers favor mandatory state testing because they believe it is necessary for (and capable of) improving academic achievement and ensuring that schools use effective pedagogical methods. These beliefs are misplaced. As already noted, parents historically have made wiser educational decisions for their own children than education officials have made for the children of others.[48] Education bureaucracies are staffed by the same flawed human beings who comprise education markets, but those bureaucracies lack the superior incentive structure that guides decision-making in a competitive market. Tests designed at state expense by education experts and officials are no less likely to encourage dubious and faddish teaching methods than they are to promote proven and effective ones.

The Program for International Student Assessment (PISA) is one of the most recent, lavishly funded, and widely talked-about international tests, and it measures exactly the kind of “fuzzy” mathematics practices that are popular among most government-school-affiliated mathematics educators, but which are less highly regarded by actual mathematicians, concerned parents, and pro-mandatory-testing school reformers.[49] Unlike the National Assessment of Educational Progress (NAEP), which allows the use of calculators only on selected items, and unlike the Third International Mathematics and Science Study (TIMSS), which does not allow the use of calculators, PISA allows calculators to be used freely on all questions. PISA was expressly designed, however, to not be computationally intensive. It has 8 questions requiring numerical computation compared to 44 such questions on the NAEP and 55 on the TIMSS.[50] Not surprisingly, pedagogically traditional parent-funded private schools that emphasize computation and other “non-fuzzy” mathematical skills did not fare as well on this test as they do on countless other more conventional ones. One study even suggests that they were in fact surpassed by public schools on the PISA, after controlling for a vast array of student, school, and socio-economic characteristics — an almost unheard of reversal.[51]

There is no question that enshrining PISA as a mandatory government test would force private schools to modify their curricula and adopt the “fuzzy” math curriculum that PISA is designed to test. Schools that tried to resist would appear inferior to prospective customers using the official government test scores as a basis for choosing their children’s schools.

Another argument, that mandatory testing should be imposed because parents want their children tested, is logically invalid. To the extent that parents want test results, they can and will seek testing for their children in an open education marketplace. Mandatory imposition of government tests is thus not only counterproductive, it is unnecessary. The private sector is capable of providing, and in fact already does provide, extensive independent testing and certification services within the field of education and training (such as professional software proficiency certification exams), and these highly-valued services would be expanded under a competitive K-12 marketplace. The vast for-profit unregulated private tutoring industry in Japan provides numerous opportunities for parents to test their children’s proficiency in countless fields of knowledge — no compulsion is required.

Finally, while the imposition of mandatory curricula and testing might make school choice legislation easier to pass into law, choice proponents should be aware that these victories would ring hollow. Legislation that imposed a government curriculum and testing package would not create a market at all, but rather an arrangement more akin to the contracting out of public schools to private managers. The educational goals would be defined by the state, and satisfying the requirements of the state, not families, would become the first priority of private schools—as is now largely the case in the Netherlands.

Financial Responsibility for Parents

As already explained, schools paid for at least in part directly by parents tend to perform better across the board than schools paid entirely by a third party. Direct payment, furthermore, has also been the only reliable means by which parents have remained in the educational driver’s seat, keeping the freedom of choice that is crucial to the effective operation of the market. Both tax credit and voucher programs should thus be designed to maximize direct financial responsibility while still ensuring universal access to the marketplace. The broadest base of evidence supports fully parent-funded schools, so maximizing the percentage of parents who can cover all of the cost of their own children’s education with their own resources is of particular importance.[52]

Though it is tempting to waive co-payments for all low-income families and to give them larger compensatory funding, the Dutch experience suggests that most schools will not use the extra funds to specifically help the children they are meant to serve, but instead use it to improve the working conditions of their own employees by reducing class sizes across the board.

Cost Control

Education spending under public schooling has skyrocketed since the 1920s when the system’s modern bureaucratic form was first taking shape. Per-pupil expenditures in the public schools are up by a factor of 15 in inflation-adjusted dollars if you use the Average Daily Attendance figures (ADA), or by a factor of 19 if you use the average fall enrollment figures.[53] To remain politically viable and practically acceptable over the long term, market education programs will have to be designed to avoid perpetuating this problem.

Universality and Competition

For education markets to work effectively, there needs to be more than just a token level of competition among schools. The largest possible number of families should be able to choose from among numerous education service providers and there should be sufficient new demand for private schooling to justify the creation of new schools. School choice laws that are explicitly limited to serving just a few thousand children, especially those that scatter the eligible children over a large geographical area, do not and cannot create vigorous competition. Instead, eligible children fill some of the empty places in existing schools and few if any new schools are created. This is not a market, it is a modest benefit package for existing private schools, and it will not fulfill the goals that we have set out to achieve.

More specifically, education reformers whose chief concern is the improvement of educational options for low-income children must realize that programs serving only these children are counterproductive. By limiting the number of participating students, targeted or narrowly means-tested programs lower the total number of competing schools; the vigor of the competition among schools; the incentives for quality, innovation and efficiency; and the diversity of specialized services available.

Program Phase-in

No program is going to create a vigorous competitive market overnight, but there are ways of maximizing competition even in programs that are phased in gradually. Introducing a market education program in stages can have practical benefits as well. By targeting the first phase at low-income families it is possible to front-load the program with migrators from government to private schooling rather than taking-in substantial numbers of existing private school students right from the start. This would prevent the situation that arose in Arizona, where the donation credit program has been criticized for doing little to help the poor, concentrating its benefits among middle class families whose children were already enrolled in private schools to begin with.[54] Similarly, front-loading the program with younger students by phasing it in by age cohort can also reduce the system’s initial cost, since elementary schooling is generally less expensive than secondary schooling.[55] The participation of children already enrolled in private schools or homeschools could also be phased-in over a number of years to assure that the program never has a negative fiscal impact on the state budget.

In their book Education and Capitalism, Herbert Walberg and Joseph Bast make a helpful distinction between gradual program phase-ins that are built into the enabling legislation and incrementalism, which they define as the passage of limited trial programs with the expectation that these will eventually be followed by legislation expanding their scope. Incremental expansion has some precedent in the Milwaukee voucher program, which initially excluded religious schools and limited student participation to about a thousand children, but later added religious schools and saw its student cap rise to about 15,000 students. Between phase-in and incremental approaches, however, the evidence seems to favor phase-ins. First, consider that the Vermont and Maine voucher programs have both been around for over a century and neither has been followed by significant scope-enlarging legislation. On the contrary, both have been legislatively circumscribed over the years by the elimination of religious schools from eligibility to participate. Of the world’s two existing universal voucher programs, those of Chile and the Netherlands, neither was built up incrementally. Both emerged, Athena-like, as fully formed national programs.

While it is important to appreciate the political and functional benefits of a gradual phase-in, we must also be aware of some potentially serious pitfalls. The particular risk is that a too-limited initial phase/increment of the program will yield disappointing or simply unimpressive results and thereby undermine support not only for its expansion but for its continued existence. The most obvious way in which this could happen would be for the early phase of the program to fail to bring about a vigorously competitive market. After all, the benefits of market education are not ensured simply by passing a law with the words “parental choice” or “competition” in its title. They are brought about through vigorous competition among numerous relatively unfettered providers.

Consider the enacted, but legally enjoined, Colorado voucher program. The initial stages of this program would scatter very small numbers of eligible students across numerous districts around the state. With only handfuls of students eligible for vouchers in any given neighborhood or district, it would not likely stimulate the creation of new private schools, and would do little to increase market pressures on existing schools. If the courts allow the program to go ahead, it will certainly present a laudable escape hatch for the children who receive vouchers, but the shortcomings just described are a cause of genuine concern. The lack of competitive density and new school creation is likely to minimize its impact on achievement and few new educational options are likely to appear. If such uninspiring results do indeed materialize, even the current escape hatch could well be slammed shut by a skeptical legislature.

Unfortunately, virtually all of the existing tax credit and voucher programs suffer from varying degrees of this same problem. Pennsylvania’s donation tax credit, for example, is capped at $26 million and its beneficiaries can live anywhere in the state. That means that few if any communities may have enough new private school students to create a single new school, let alone to create a vigorously competitive private market. Milwaukee’s trial voucher program is marginally better, supporting about 13,000 voucher students within the district’s boundaries, but it has very nearly reached its cap. With a maximum of just 15% of the city’s students eligible to participate, market forces will necessarily remain limited.