From the latest issue of the journal Economics in Education Review comes a fascinating paper in which author Metin Akyol creates mathematical models that simulate the effects of private school vouchers on the overall education system. It is not a study of an actual voucher program, but instead a thought experiment meant to test whether both universal and targeted voucher programs can increase the efficiency of the education system as a whole. As strange as this may seem to lay readers, there is in fact a long history of such econometric analyses—and their findings are often worthy of consideration.
Akyol’s complex model can’t be fully explained in this short review, but some features are worth noting. It incorporates the findings of empirical voucher studies to increase its reliability. It simplifies the real world in an effort to find the signal in the noise. Every household therefore has only one child, and the hypothetical school district has neither magnet schools nor charters. And one of its defining assumptions is that more efficient public school spending is an effective proxy for increased educational quality. In other words, it presumes that the money saved by greater efficiency can be reinvested in ways that improve outcomes.
Regardless of how one feels about all this, the model ends up producing outcomes that are very similar to empirical findings regarding actual programs. In one important example, the positive effects on voucher-eligible students who do not opt to leave their district school (found empirically by David Figlio in Ohio) are predicted in Akyol’s targeted-voucher model. Public schools in the model are observed to “up their academic game” to retain students when voucher competition is introduced. Additionally, the model predicts that students lowest on the income scale will be least likely to use vouchers, even in a model where vouchers are universally available and not means-tested. This stands to reason, considering that real-world vouchers often fall short of full private school tuition. (To some extent, it was also borne out in Figlio’s research.)
Also interesting is the difference in effects between a universal voucher program and a targeted one. Akyol ran models that replicate the prime goal of vouchers—make private schools affordable for more children—in two ways: by manipulating the voucher availability and by simply changing the family income distribution. The results were not the same. The universal voucher model led to an observable decline in “peer group quality” for those students at the lowest end of the income spectrum who did not take the vouchers. This decline in quality was also present to a lesser extent in the model where vouchers were targeted at low-income students. But it was absent in a model that gave high-ability students lower voucher amounts than their lower-ability peers. This appears to be the theoretical sweet spot that results in the most favorable overall outcome: more students were able to access private schools, and public schools felt the competition keenly enough to improve their academics for the students who remained.
None of this means we must redesign real-world voucher programs based on any one of the mathematical models presented in this paper. But to the extent that modeling can predict real-world outcomes, choice advocates and policymakers ought to consider the results, which can elucidate the potential benefits and challenges of particular voucher designs. If the incoming Trump/DeVos education department is going to prioritize vouchers as a means for improving education, Akyol’s mathematical models have at the very least led him to offer some sage advice: “…the outcomes of a voucher program hinge on its design.”
SOURCE: Metin Akyol, “Do educational vouchers reduce inequality and inefficiency in education?” Economics of Education Review (December, 2016).
