Friday, November 25, 2016

Remaining Questions and Challenges for Market Design

Very few economic sub-fields have done so well in such little time as has market design. From redesigning school choice to establishing matching markets in kidney to introducing combinatorial auctions of airwaves no other field has so rapidly taken theory to practice and seen it bear immediate fruit. The rest of the economic community took notice and two market designers were awarded the Nobel Prize for their early and profound contributions. One can't help but think that more accolades are due in the next 5-15 years.

However, as I checked back into the field, surveyed the literature and attended an NBER workshop after having taken a multi-year hiatus I couldn't shake the feeling that there are too many unexploited opportunities still out there.....proverbial $20 dollar bills being left on the sidewalk. The picture above is my back-of-the-envelope (whiteboard) sketch of where market design research could go in the next 20 years. 

I was thinking about the effects on agents under the regimes of different market institutions and wonder where are the papers that show those economic effects and trade-offs. I imagined several small rivers (representing agents making decisions before entering the market) that slowly flow downstream, then converge into a lake (representing the market mechanism) that feeds into larger rivers (representing the different agents and the results of their transactions). My questions centered around what happens downstream or after the transaction and how does behavior change upstream prior to deciding when and if to enter the market. 

Stepping into the political landmines of school choice, market designers long ago implemented mechanisms in several large school markets that are stable and safe for parents to express their true preferences without requiring specialized knowledge about the process. But, are these children performing any better after having been placed in schools that are more preferred by their parents? Did the mechanism induce parents to switch from private schools back to public schools or vice versa? Making parents happy and building trust in a public school system is not nothing but the administration of the mechanism that matches students to schools isn't free either. Is it worth it in economic terms? What's the payoff? 

Combinatorial spectral auctions of airwaves vastly increased sales revenue for the government and helped firms narrowly target consumers but what was the downstream effect after the auction Were media companies better able to develop and grow their markets? How did t.v. and radio advertisements change before and after the introduction of the combinatorial auction?

Looking upstream, it seems important to think about how people's investments in human capital differ knowing they will enter a labor matching market under one set of rules versus another. How do different mechanisms affect an agent's decision to participate and when to enter? One gap in the research is the case of dynamic markets where the market clears over time or is continuously attempting to clear. For school choice or doctors matching to a residency program the problem is essentially one-and-done from the perspective of the student and the doctors. Once they match, they presumably go on their way and never deal with the market mechanism again. However, in some markets, like the military, agents will enter the internal market for a new job every three years.  

Looking at all these unasked and unanswered questions I realize that part of the reason for this is simply a resource allocation issue. The number of researchers and grad students thinking deeply about market design is small in absolute terms (probably ~100) and infinitesimal compared to finance, labor or development (10-1 or 100-1?). Still, almost all of the papers that I've read in the past and present focus exclusively on developing the theoretical understanding of market allocation mechanisms under different conditions such as preferences, information and supply constraints. Only time will tell if those deeper lines of research are valuable or not but it seems that there is still a lot of lower-hanging fruit left on the tree of knowledge.