Sáez-Marti, Maria; (1995) Evolutionary Game Theory, Markets and Conventions. Doctoral thesis (Ph.D), UCL (University College London).

Text 10017243.pdf Download (3MB)

Abstract

The Thesis applies evolutionary game theoretic ideas to the modelling of economic behaviour. The traditional approach is to assume that economic agents are perfectly rational. The rationality assumption requires that agents be very sophisticated in their decision making. This creates a gap between the behaviour postulated by economic models and real people's behaviour. Recently, theorists' attention has turned to models of bounded rationality in which agents are assumed to find their way to equilibrium by trial-and-error methods. The most widespread approach to modelling boundedly rational behaviour is the use of evolutionary game theory which also provides useful insights into the equilibrium selection problem. The first chapter is a critical survey of evolutionary game theory. In the second chapter we endogenize the learning rules in a modified version of Young's bargaining model which provides an evolutionary explanation for the asymmetric Nash bargaining solution. The Nash Demand Game is played by two different populations. Players choose their strategies in the light of some limited information about the strategies players from the other population have used in the past. An interesting result is that the better informed population has higher bargaining power. The main drawback in Young's model is that the amount of information, and therefore the bargaining powers are fixed exogenously. We endogenize players's learning rules and test for evolutionary stability. We study whether one population using a particular learning rule can be invaded by a mutant learning rules. We show that, when information is costless, the only evolutionarily stable learning rule maximizes players' information. If both populations follow the same learning rule, the equilibrium which is selected is the symmetric Nash bargaining solution. When information is costly there is a trade-off between costly learning and the rewards of being well informed. Finally we show that an economy populated by players who follow very simple imitative rules is socially more efficient than an economy of rational players. In the third and fourth chapters we introduce models which endogenize the equilibrium selection problem. We show that, in a model of a credit market an equilibrium is selected in which the market is fully developed, although the model has another equilibrium with interesting stability properties. The agents are assumed to follow very simple behavioural rules and sometimes to experiment with unplayed strategies. Similar results are obtained after the introduction of a small proportion of rational players. The model predicts interesting dynamics which are consistent with some evidence about the Great Depression. Real shocks trigger episodes of credit-crunch and temporary financial collapse which is observed in the process of adjustment towards the post shock equilibrium.

Download activity - last month

Export as CSVJSONXML

Download activity - last 12 months

Export as XMLJSONCSV

Downloads by country - last 12 months

Export as CSVJSONXML

Archive Staff Only

View Item