« Back to Results

Practical Considerations in Deploying Matching Mechanisms

Paper Session

Saturday, Jan. 5, 2019 8:00 AM - 10:00 AM

Atlanta Marriott Marquis, International 9
Hosted By: American Economic Association
  • Chair: Alex Rees-Jones, University of Pennsylvania

How Well Do Structural Demand Models Work? Counterfactual Predictions in School Choice

Parag Pathak
,
Massachusetts Institute of Technology
Peng Shi
,
University of Southern California

Abstract

Discrete choice demand models are widely used for counterfactual policy simulations, yet their out-of-sample performance is rarely assessed. This paper uses a large-scale policy change in Boston to investigate the performance of discrete choice models of school demand. In 2013, Boston Public Schools considered several new choice plans that differ in where applicants can apply. At the request of the mayor and district, we forecast the alternatives' effects by estimating discrete choice models. This work led to the adoption of a plan which significantly altered choice sets for thousands of applicants. Pathak and Shi (2014) update forecasts prior to the policy change and describe prediction targets involving access, travel, and unassigned students. Here, we assess how well these ex ante counterfactual predictions compare to actual outcome under the new choice sets. We find that a simple ad hoc model performs as well as the more complicated structural choice models for one of the two grades we examine. However, the structural models' inconsistent performance is largely due to prediction errors in applicant characteristics, which are
auxiliary inputs. Once we condition on the actual applicant characteristics, the structural choice models outperform the ad hoc alternative in predicting both choice patterns and policy relevant outcomes. Moreover, refitting the models using the new choice data does not significantly improve their prediction accuracy, suggesting that the choice models are indeed “structural.” Our findings show that structural demand models can effectively predict counterfactual outcomes, as long there are accurate forecasts about auxiliary input variables.

Reducing Congestion in Matching Markets Using Informative Signals

Itai Ashlagi
,
Stanford University
Mark Braverman
,
Princeton University
Yash Kanoria
,
Columbia University
Peng Shi
,
University of Southern California

Abstract

Learning and communication of preferences are activities that create congestion in matching markets. Early results by Segal (2007) show that a high level of communication may be inevitable under arbitrary preferences. We discuss models, in which unobservable component of an agent preferences satisfy natural assumptions. We identify certain types of informative signals that allow to reach an equilibrium outcome with a low amount of communication. These signals prevent agents to reach out to partners for which they have a non-negligible chance to match with. Out findings further provide a justification for why stability is a plausible solution concept for large matching markets.

Obvious Mistakes in a Strategically Simple College Admissions Environment: Causes and Consequences

Ran Shorrer
,
Pennsylvania State University
Sandor Sovago
,
Vrije University Amsterdam

Abstract

Although many centralized school admission systems use strategically simple mechanisms, applicants often make dominated choices. Using administrative data from Hungary, we show that many college applicants forgo the free opportunity to receive a tuition waiver. We provide causal evidence that applicants make more such mistakes when applying to programs where tuition waivers are more selective. A non-negligible share of these mistakes are consequential, costing applicants approximately 3,000 dollars. Costly mistakes transfer waivers from high- to low-socioeconomic status students, and increase the number of admitted students. Our results suggest that mistakes are more common when their expected utility cost is lower.

An Experimental Investigation of Preference Misrepresentation in the Residency Match

Alex Rees-Jones
,
University of Pennsylvania
Samuel Skowronek
,
University of Pennsylvania

Abstract

The development and deployment of matching procedures that incentivize truthful preference reporting is considered one of the major successes of market design research. In this study, we test the degree to which these procedures succeed in eliminating preference misrepresentation. We administered an online experiment to 1,714 medical students immediately after their participation in the medical residency match--a leading field application of strategy-proof market design. When placed in an analogous, incentivized matching task, we find that 23% of participants misrepresent their preferences. We explore the factors that predict preference misrepresentation, including cognitive ability, strategic positioning, overconfidence, expectations, advice, and trust. We discuss the implications of this behavior for the design of allocation mechanisms and the social welfare in markets that use them.
JEL Classifications
  • D4 - Market Structure, Pricing, and Design
  • I2 - Education and Research Institutions