Evolutionary stability in first price auctions (Q367477)

In this paper, the authors introduce a new approach to the study of evolution in first price auctions. They analyze the stability properties of the Nash equilibrium under the replicator dynamic and general payoff-monotonic dynamics. The paper consists of six sections (with the introduction, the brief presentation of the model, and the conclusion as Sections 1, 2 and 6, respectively) and 3 appendices with proofs and simulation results. In Section 3, the authors introduce the replicator dynamic, the payoff monotonic dynamic (that generalizes the replicator dynamic), and the Brown-von Neumann-Nash (BNN) dynamic which belongs to the family of innovative dynamics. While under the replicator dynamic, only strategies that are already present in the population can be adopted, in the BNN dynamic, strategies are adopted proportionally to their excess payoffs, even if they are not present in the population. Section 4 contains the main (theoretical) results on dynamic (in)stability. The first result is a negative result and states that the unique Nash equilibrium in first price auctions is not asymptotically stable under payoff monotonic dynamics, in particular, the replicator dynamic, for arbitrary initial populations. The authors also find a positive result by imposing some conditions on the initial population. More precisely, they show that for some classes of bid functions, when the initial population includes a continuum of strategies around the equilibrium, the replicator dynamic converges to the Nash equilibrium. In Section 5, simulations are conducted for the replicator and the BNN dynamics which suggest that the latter converges much faster than the replicator dynamic. As there are very few theoretical results for the BNN dynamic, simulations allow seeing the behavior of this dynamic.