Research
A beliefbased theory of homophily (with Alvaro Sandroni)
Homophily, the tendency to interact with similar people, is a welldocumented phenomenon that has important economic implications. We provide a beliefbased explanation of homophily. Building on the dual process account of Theory of Mind in psychology, we assume that players with a similar background have similar impulses telling them which action is appropriate. Players are introspective and recognize that their own impulses are a good predictor of the impulses of players that are like them, and realize that others are likewise introspective. We show that if players can sort (e.g., by choosing certain occupations) to meet people like them, then high levels of homophily are possible even with limited asymmetries in preferences.
Intuitively, the introspective process amplifies small asymmetries in preferences, allowing for effective coordination.
Finite Depth of Reasoning and Equilibrium Play in Games with Incomplete Information (submitted).
Online appendix.
If players have a finite depth of reasoning, can their equilibrium behavior be described by the standard framework? Intuitively, one might expect that “simple” Harsanyi type spaces, in which the beliefs of types are determined by the first few orders of beliefs,
would be good “asif models” models for finitedepth type spaces. However, I show that this is not the case: for any finitedepth type space, there is no Harsanyi type space (simple or not simple) that gives the same equilibrium behavior in all games.
Robust Multiplicity with a Grain of Naiveté (with Aviad Heifetz; submitted).
Online appendix.
In global games and other games with sufficiently rich payoffs, types typically have a unique rationalizable action (Weinstein & Yildiz, '07). This result, however, presumes that players have an infinite depth of reasoning and that this is common belief. We
show that multiplicity of rationalizable outcomes is a typical and robust phenomenon if we relax that assumption only slightly, i.e., if players have an infinite depth and this is almost common belief.
Ranking Friends (with Yossi Feinberg; submitted).
Say you need a friend to help you move. You call your friends one by one to see whether one of them can help. If a friend does not or cannot help you (for example, because she can't get
the day off, or because she doesn't even try), you call the next friend on your list. If everyone has such a list, or ranking, of friends, and needs help on a regular basis, what is the optimal social structure?
Ambiguous Language and Consensus (with Joe Halpern; submitted).
Language, and communication more generally, is often ambiguous, i.e., can be interpreted in different ways. In a series of papers, we develop a logic to model this, and investigate
its consequences for belief formation and consensus. In one application, we show that in the presence of ambiguity, the claim of Aumann (1987) that ``there is no rational basis for people who have always been fed precisely the same information to [entertain different probabilities]'' holds only under very restrictive assumptions: players' beliefs can be consistent with a common prior only if all information is public, not just common, that is, all players receive the same signals, and this is common belief.
Publications
Robustness of equilibria in anonymous local games, Journal of Economic Theory, 146, pp. 300–325, 2011.
Online appendix. (Working paper version.)
Inequality and network structure (with Garud Iyengar, Rajiv Sethi and Sam Bowles), Games and Economic Behavior 73, pp. 215–226, 2011.
(Working paper version.)
Learning to be prepared (with Mark Voorneveld),
International Journal of Game Theory 37, pp. 333–352, 2008. (Working paper version.)
An axiomatization of minimal curb sets (with Mark Voorneveld and Henk Norde),
International Journal of Game Theory 33, pp. 479–490, 2005.
(Working paper version.)
Publications in Other Fields
Betting Strategies, Market Selection, and the Wisdom of Crowds (with David Pennock, Rajiv Sethi and Nisarg Shah); TwentyEighth AAAI Conference on Artificial Intelligence (AAAI14), 2014; forthcoming.
A logic for reasoning about ambiguity (with Joe Halpern); Artificial Intelligence, 2014; forthcoming.
Ambiguous language and differences in beliefs. Proceedings of KR2012, 2012. Ray Reiter Best Paper Award.
Random
intersection graphs with a tunable degree distribution and
clustering (with Mia Deijfen), Probability in the Engineering and
Informational Sciences 23, pp. 661–674, 2008.
Zn and
Cdinduced features at the GaAs(110) and InP(110) surfaces
studied by lowtemperature scanning tunneling microscopy (with Randy de Kort, Maurice van der Wielen,
Ari van Roij, and Herman van Kempen), Physical Review B 63, 125336, 2001.
A lowtemperature scanning tunneling microscopy study
on the Sn and Zndoped InP(110) surfaces (with Randy de Kort and Herman van Kempen),
Surface Science 482, pp. 495–500, 2001.
PeerReviewed Surveys
Learning with fixed rules: The minority game, Journal of Economic Surveys. Forthcoming.
Free
trade and its enemies (with Paul Tang), De Economist, pp. 152–153, pp. 427–437, 2004.
