Why Milwaukee’s Voucher Program Isn’t the Litmus Test for School Choice Everywhere (Despite Its Positive Results)

Milwaukee, Wisconsin has a 25-year-old school voucher program (MPCP) restricted to low-income students, district-run chartered public schools and some privately funded vouchers. Because the MPCP is the oldest publicly funded U.S. voucher program, its results have attracted a lot of attention.  Sadly, much of what has been said and written, from school choice expansion proponents and opponents alike, is wrong, misleading or irrelevant. 

Here’s why.

Even before the 1998 expansion of the MPCP, some prominent advocates of school choice expansion announced that the restriction-laden Milwaukee program would provide the evidence needed to determine whether “school choice works” and also whether we can rely on market-driven accountability to drive public school system improvement.

Opponents of school choice expansion warned of horrific effects that never happened and shamelessly painted with a very broad brush when the assessments of the restriction-laden MPCP began revealing that the gains generated by school choice in Milwaukee had fallen far short of the unwarranted high expectations school choice advocates had set.

Even with the modest 1998 expansion, which left the ban on co-payment (price control) and low-income eligibility criterion in place, the MPCP still had far too many restrictions to qualify as a market accountability experiment.

Some of those prominent claims arose during MPCP’s first eight heavily studied years (1991–98), when a program “designed to fail” provided vouchers to just 1 percent to 1.5 percent of Milwaukee Public School (MPS) students. And the lucky few, early-year, low-income voucher applicants could use the vouchers, worth much less than the MPS per-pupil expenditure, only at non-sectarian schools, which at the time were just three of the six really shaky, low-tuition, non-sectarian schools operating at the time. The MPCP rules said the eligible private schools could cash the vouchers only if they accepted them as full payment, which amounts to price control—a devastating, anti-market ‘pricelessness’ restriction that persists to this day with just one negligible exception.

A 1998 Wisconsin Supreme Court decision allowed expansion of the number of vouchers and permitted sectarian schools to enroll voucher users, but it did not eliminate very difficult circumstances for the MPCP to be deemed successful. Even with the modest 1998 expansion, which left the ban on co-payment (price control) and low-income eligibility criterion in place, the MPCP still had far too many restrictions to qualify as a market accountability experiment.

That was true whether the standard is success as a school system transformation catalyst—the correct standard for a low-performing school system—or with the still widely perceived ‘gold standard’ comparisons between voucher users and unsuccessful voucher applicants, meaning those who did not win the lottery to receive a voucher.

Such ‘gold standard’ comparisons are much more appropriate for testing medicines than for the social sciences, especially for effects of school choice expansion, where the effects of policies relevant to low-performing school systems—rather than isolated schools—are not limited to participants. The ‘gold standard’ comparisons amounted to asking whether the private part of a low-performing (“Nation at Risk” low) school system could outperform politically crippled, but still financially advantaged, public schools. The ‘gold standard’ approach assesses whether student transfers among existing choices will greatly improve student outcomes, information that does not provide much insight on the potential for school system reform driven by expansion of school choice.

After first struggling in their assigned public school, voucher users must out-perform unsuccessful voucher applicants with voucher funding of a little more than half the amount supporting the MPS students.

Still, the voucher users did outperform the Milwaukee Public School students, who were denied vouchers, in some subject areas, but only by small margins. In some subject areas the voucher users fared no better. And studies of school choice programs have also found just small, positive effects on public school users.

The small effects have become an opposition talking point. They say school choice doesn’t make much difference. Even President Barack Obama said, “Actually, every study that’s been done on school vouchers says that it has very limited impact if any.” Diane Ravitch’s widely read blog asserted that the Milwaukee experience, especially MPCP, should be a poster child for the wrongheadedness of school choice as a reform catalyst.

But the small effect on MPS performance should not be surprising. The voucher program rules give the MPS schools little tangible reason to behave competitively.

Milwaukee public schools suffer few, if any, funding losses when students leave with a voucher. Lacking a tangible basis for competitive behavior, I have argued that sorting effects are a more likely basis for the small MPS improvements attributed to the voucher program. Sorting effects occur in MPS when some parents move their children to different schools; schools they think will work better for them. That leaves more instructionally homogenous, teachable classrooms behind thus less need for difficult differentiated instruction.

The children who leave are the ones for whom the futile, assigned public schools’ attempt at a one-size-fits-all instructional approach is least effective. The voucher users are likely outliers in most of their classrooms. Why would they leave a good fit?  Even if the voucher users are neither sufficiently numerous nor empowered to drive school system change, movement to better-fit existing schooling options helps the voucher users and the children that stay behind in the assigned public school.

Surveys of parents who use vouchers for their children have repeatedly shown them to be very pleased with the assigned public school alternative, which means the program has at least worked as an escape hatch. Nearly all studies have shown that they reap small gains in the tested academic areas plus, probably, benefits through choice-making criteria that standardized tests cannot capture.

The escape hatch nature of the small, restriction-laden Milwaukee program is further confirmed by the assumptions and findings of several studies, especially some recent findings that include the purported “last word” on the program’s effects. The earlier findings that the “systemic effects” (the competitive effects of the program on academic achievement in MPS) have been small led the authors of the “last word” to assume that unsuccessful voucher applicants were not affected by the voucher program; an implicit ‘no systemic effects’ assumption. And at least in terms of the standardized test scores, the voucher users were not greatly affected: “the achievement growth of students in the voucher program was [slightly] higher in reading but similar in math.”

Since some school choice advocates repeatedly bet the farm on the Milwaukee voucher programs’ outcomes—a truly lousy experiment in market accountability—the pro-establishment spin doctors are swooping in for the kill.

For example, this from Ravitch: “Anyone who looks at the NAEP reports on urban districts will see that after 22 years of vouchers, charters, and competition, Milwaukee is a poster child for the failure of vouchers, charters, and competition. The students in those schools all perform [at] about the same level. No sector is better.”

Not only is Ravitch guilty of groundless, sweeping generalizations, but her critique rests on the wrong basis. Success or failure of school choice is not a matter of which “sector is better” or whether current private schools, with much less money per child, perform better for similar children than current public schools. We need high-performing school systems regardless of the public-private mix.  In our current school system, because of our governance and funding rules, many schools in both sectors are unacceptably low-performing, and not just urban public schools.  Lance Izumi’s “Not as Good as You Think” multi-state series has shown us that low-performance is also the norm outside the inner city; in well-off areas as well as low-income areas.  The roots of low performance take their toll everywhere. One size fits all nowhere.

It is very sad that Dr. Diane Ravitch has descended from outstanding education historian (I cite her past work a lot!) to another data-twisting protector of the status quo. Her exceptional ability to ferret out important facts quite often does not extend to comprehension of what they mean. She proved that, yet again, with her reaction to the recent Ford-Andersson findings about private school entrepreneurship.  Those findings about Milwaukee private school start-up failure rates that Ravitch thought were so horrible, despite program restrictions that should have made failure rates unusually high, were low compared to failure rates for new businesses in general. She marshaled no evidence to show that the failure rates were disruptive, or to see if the authorities had taken any prudent steps to diminish disruption. She also did not note the high failure rates of the Milwaukee Public Schools, schools that continue to receive funding and have children assigned to them. The private part of our school systems purges poorly conceived and poorly run schools. The public school system does not.

That said, Ravitch’s comparison of Milwaukee to other urban districts with NAEP data needs to be done more rigorously, and she may turn out to be right about the small Milwaukee effects. There should be nothing shocking about a very limited, “designed to fail” program having very limited effects.

Left unsaid—school choice opponents’ ever-present broad brush—is that we have no U.S. examples of genuinely competitive school markets. But, because of unwarranted proponent hype, the broad brush had unearned credibility. Competition as a school system improvement catalyst was widely deemed present in Milwaukee, and now supposedly seen as tested and found ineffective. Small effects from a small, restriction-laden program combined with choice advocates’ bear hug for the Milwaukee program may assure that we never see genuine competition at work in K–12 schooling markets—a scary thought.

What are the appropriate lessons?

With regard to the other U.S. school choice policies, and the vast majority in foreign countries, we need to stop imagining them to be more than they are.  We need to be prepared for the status quo-proponent spin, especially the broad brush attacks with facts and fears of little or no relevance to school choice expansion efforts that could be school system transformation catalysts.

Only one U.S. educational choice policy might turn out to be worthy of all of the hype and hope widely expressed about every school choice expansion. The universal Nevada Education Savings Account has just emerged from the shadow of an inevitable lawsuit. With an adequate funding stream plus philanthropic dollars to finance tuition co-payment for low-income households, we may finally see what careful shopping and entrepreneurial initiative can do for schooling. We have a benchmark from which to document a likely transformation of schooling in the state’s major urban centers.

We need to carefully document and explain everything that happens in Nevada. Some—for example, failed attempts at new schools—will not be good. We need to have policy proposals for softening the sometimes hard-edged effects of markets pushing resources to ever-better uses.

We need to be better prepared for the status quo-proponent spin, especially the broad brush attacks with facts and fears of little or no relevance to school choice expansion efforts that could be the school system transformation catalysts a ‘Nation at Risk’ needs.


*Opinions expressed by our guest bloggers are their own and do not necessarily reflect those of EdChoice.