Learning dynamics with social comparisons and limited memory

Change log
Block, Juan I 
Fudenberg, Drew 
Levine, David K 

We study models of learning in games where agents with limited memory use social information to decide when and how to change their play. When agents observe only the aggregate distribution of payoffs and recall only information from the last period, aggregate play comes close to Nash equilibrium for generic games, and pure equilibria are generally more stable than mixed equilibria. When agents observe both the payoff distribution of other agents and the actions that led to those payoffs, and can remember this for some time, the length of their memory plays a key role: With short memories, aggregate play may not come close to Nash equilibrium unless the game satisfies an acyclicity condition. When agents have sufficiently long memory, generically aggregate play comes close to Nash equilibrium. However, unlike in the model where social information is solely about how well other agents are doing, mixed equilibria can be favored over pure ones.

38 Economics, 3801 Applied Economics, 3803 Economic Theory
Journal Title
Theoretical Economics
Conference Name
Journal ISSN
Volume Title
The Econometric Society
Institute for New Economic Thinking (INET) (INET12-00032)
Institute for New Economic Thinking (INET) (INET 12-00032)
Cambridge-INET Institute, the EUI Research Council, and NSF Grant 1643517