Eastern Great Lakes Theory Workshop Talk

No Regret Learning in Oligopolies: Cournot vs Bertrand

Georgios Piliouras, Cornell University

Saturday, October 3, 5:45-6:00pm

ABSTRACT

We analyze the two most prevalent models of firm competition in disequilibrium under minimal behavioral hypotheses. Specifically, the only assumption that we make is that firms adapt their strategies over time, so that in hindsight their average payoffs are not exceeded by any single deviating strategy. Although the Nash equilibria of both models are well understood, unfortunately, their predictions about the stable states of the economy are starkly different. More alarmingly, in the case of Bertrand oligopolies, the analysis of Nash equilibria predicts that merely the existence of two firms is enough to push prices down to the marginal cost level. This suggestion is commonly referred to in the economics literature as the "Bertrand paradox". In our work, we show that under the assumption of adapting players these inconsistencies can be resolved, as both models arrive to the same conclusion that increased competition is necessary in order to achieve perfect pricing.

These results are based on joint work with Uri Nadav.

Speaker Bio

Georgios received a diploma in Electrical Engineering and Computer Science from the National Technical University of Athens in 2004, and a M.Sc in Logic, Algorithms and Computation from the National University of Athens in 2005. He is currently a PhD student at Cornell University under the advisement of Eva Tardos.

Georgios is a recipient of Cornell's Olin Fellowship.

< Schedule < EaGL-II homepage