“Experiments on Strategic Behavior in Two-Sided Matching Markets”
Marco Castillo, Ahrash Dianat, George Mason University
“Matching theory” has informed the design of institutions in areas as diverse as kidney exchange, entry-level labor markets, and school choice. These institutions often operate as centralized clearing houses, in which participants submit rank-order lists of their preferences and a particular algorithm calculates the final outcome, i.e., who is paired with whom in the final matching. There are some environments where agents on one side of the market have an incentive to misrepresent their preferences in order to bring about a more favorable outcome. In these situations the particular outcome that is finally implemented depends on agents’ ability to behave strategically either in isolation or as a group. Our work uses laboratory experiments to investigate whether the ability of agents to coordinate on misrepresenting their preferences in a particular fashion depends on features such as how profitable and safe it is to misrepresent and also the size of the market. We hypothesize that strategic behavior will be more likely to arise where it is more profitable and safer. In addition, we hypothesize that agents will coordinate on strategic behavior in small markets but not in large markets. However, we predict that coordination in large markets can be obtained by starting with small markets and incrementally adding players who are aware of the market’s history.