In my remarks the other day about this year's Roth/Shapley Nobel, I said that I thought the work was important and useful because it led to the implementation of better institutions in situations in which prices are unlikely to emerge organically. A post from Andrew Coulson at Cato at Liberty fleshes out some reasons why his enthusiasm is limited. In his post he describes some of Roth's work in school-pupil matching, and he observes that

The problem is that this approach to “school choice” correctly assumes that the better public schools have a fixed number of places and cannot expand to meet increased demand. So it’s about finding the least-awful allocation of students to a static set of schools—a process that does nothing to improve school quality.

Meanwhile, there is something called a “market” which not only allows consumers and producers to connect, it creates the freedoms and incentives necessary for the best providers to grow in response to rising demand and crowd-out the inferior ones. It also provides incentives for innovation and efficiency. But instead of advocating the use of market freedoms and incentives to improve education, some of our top economists are spending their skill and energy tinkering with theincreasingly inefficient, pedagogically stagnant status quo. That's a good point to remember when thinking about matching algorithms and institutional design -- if you have a problem, like kidney matching, that is fairly static in its overall nature (although probabilistic in its actual incidence), then yes, matching algorithms are likely to yield higher total benefits/welfare/surplus. But Andrew's right; when there's a possibility of increasing the supply of schools or teachers, or to change the quality of them, that's not a problem where applying a matching algorithm is going to yield as many potential benefits as a market ... if you can overcome the political barriers to school competition and choice. Markets yield better outcomes in general in dynamic situations.

In that case, it's a transaction that's politically repugnant.

Knowledge Problem

Matching and a Static Environment