Modeling Fee Shifting with Computational Game Theory
Abstract
While modern mathematical models of settlement bargaining in litigation generally seek to identify perfect Bayesian Nash equilibria, previous computational models have lacked game theoretic foundations. This article illustrates how computational game theory can complement analytical models. It identifies equilibria by applying linear programming techniques to a discretized version of a cutting-edge model of settlement bargaining. This approach makes it straightforward to alter some assumptions in the model, including that the evidence about which the parties receive signals is irrelevant to the merits and that the party with a stronger case on the merits also has better information. The computational model can also toggle easily to explore cases involving liability rather than damages and can incorporate risk aversion. A drawback of the computational model is that bargaining games may have many equilibria, complicating assessments of whether changes in equilibria associated with parameter variations are causal.