Free to Choose

[Jim Manzi]

Ramesh has a great series of posts at The Corner commenting on a debate at City Journal on school choice.  Sol Stern wrote an article recanting his support for markets as the key enabling reform for school improvement, and arguing that we should instead focus on making the current system work better.  He calls this the “instructionist” approach in contrast to the “incentivist” market-oriented approach.  There is an amazing series of long letters written to City Journal in response to this article, presenting viewpoints at various points on the instructionist-to-incentivist spectrum.  The article plus the letters collectively comprise an excellent introduction to the school reform debate.

I could write a lot about this, but I’ll try to limit myself to four points that I think are most important.

1.  School choice improves the performance of participants today.  As per my earlier post on evidence of causality in social science, one can make this statement with about as much certainty as one can make any non-trivial statement about causality relevant to public policy because we have multiple replications of true random assignment trials.  Further, there are natural experiments (normally called “differences-in-differences” analysis in education research literature) indicating that school choice probably also improves performance for students who do not personally participate in it, but are in school systems where choice is introduced, presumably because competition forces improvement to their schools.  Throwing sand in the air about aggregate levels of system performance, overall trends and other rhetorical techniques can’t really get past these facts.

2.  These results have been achieved within almost comically artificial “markets”.  Conditions vary, but one fundamental issue is that in many localities schools don’t lose the funding for a student if the student leaves the school by participating in a choice program.  This would be like your local hardware store having its annual revenues set by a “hardware board” based on the number of people who live in the area that need flashlights and screwdrivers, independent of now many flashlights and screwdrivers it actually sold.  My guess is that they wouldn’t be staying open late on Saturdays.  Another issue is that when, as in many systems, some tiny fraction of the population is allowed to participate (normally by lottery, which is why we have so many replications of random assignment trials), there is insufficient demand to stimulate the creation of much alternative school capacity, achieve scale effects that would induce large enterprises to start new school chains and so forth.  In other words, these are markets with huge limitations on the demand side and the supply side.

3.  The instructionist vs. incentivist debate is kind of silly.  For one thing it confuses means and ends, and for another it represents an entirely static view of instruction.  It’s like two guys in 1980 in the Soviet Union arguing about whether they need an open consumer market for PCs in order to compete with the Western computer industry, or if it’s smarter to just get their factories to implement known improvements to chip design and fab processes.  Of course, one of the goals of creating a real market is to force the implementation of such (painful, and hence politically resisted) design and production reforms that won’t happen without the incentives created by a market.  Beyond this, markets are required to invent new reforms (or at least to discover what innovations are working) and apply them broadly, as well as to match non-uniform educational alternatives with the, to put it mildly, non-uniform needs among tens of millions of students.

4.  Advocates of school choice do need to move beyond current efforts if they are serious about materially changing American K – 12 school results.  A Republican Party that wanted to win would make this a major initiative spanning administrations.  I summarize what I think needs to be done in an article in the current National Review:

We need competition for students among public schools in which funding moves with students, and in which schools are far freer to change how they operate. As we have seen in the private economy, only markets will force the unpleasant restructuring necessary to unleash potential. Conservatives have long had this goal, but are unprepared to win the fight. Achieving it would be at least a decade-long project.

We should start at the state and local level by scaling up the numerous existing demonstration projects to district, city, and state levels. This can be used to build sufficient pressure to force national action, using the federal education budget as a lever. The model to follow is the successful campaign to reform welfare. The biggest difference in this case is the presence of the extremely powerful teachers unions. Ultimately, political leaders will have to break them as decisively as Thatcher broke the coal-miners union in Britain.

The role of the federal government could be limited, but crucial. Suppose it established a comprehensive national exam by grade level to be administered by all schools and universities that receive any federal money, and required each school to publish all results, along with other detailed data about school budgets, performance, and so forth, each year. Secondary, profit-driven information providers, analogous to credit-ratings agencies and equity analysts, would arise to inform decision-making. The federal role would be very much like that of the SEC for equity markets: to ensure that each school published accurate, timely, and detailed data.

Once success is obvious, this can act as a platform for further change. This approach will build a constituency with an entrenched interest as powerful as the one that protects Social Security, and eliminate a major prop of the statist coalition. Inevitably, parents will demand and win the right to apply these vouchers to private schools. This would also become a model for other reforms of entitlement programs from retirement accounts to medical care.