Conference Agenda

Pricing on two-sided markets such as AirBnB or Uber is typically automated via learning algorithms that adapt to changes in demand and supply. This paper investigates the convergence of deep reinforcement learning (DRL) algorithms towards equilibrium prices in two-sided markets, building on the foundational platform model introduced by Armstrong. We begin by establishing that the pricing game in two-sided markets is strictly monotonous, which proves the convergence of projection algorithms to an equilibrium. These results do not extend to DRL, which is more challenging to analyze theoretically due to their use of neural networks. In extensive experiments, we demonstrate that popular DRL algorithms such as Proximal Policy Optimization (PPO) and REINFORCE reliably converge to price equilibria in these markets as well. Our findings are significant in light of recent concerns regarding algorithmic collusion by AI-driven pricing agents in traditional oligopoly models. Contrary to fears raised by studies in these markets, our results suggest that such phenomena are less of a concern in two-sided market structures.

In this paper we give a brief introduction to dynamic games and their applications to problems of macroeconomic policy. Then we present an application of the dynamic tracking games framework to modeling a monetary union. We use a small stylized nonlinear two-country macroeconomic model of a monetary union to analyze the interactions between fiscal (governments) and monetary (common central bank) policy makers, assuming different objective functions of these decision makers. Using the OPTGAME algorithm for the numerical approximate solution of these games, we study the impacts of an exogenous fall in aggregate demand and the resulting increase in public debt, similar to the economic crisis (2007–2010) and the sovereign debt crisis (since 2010) in Europe. In the union, the governments of participat-ing countries pursue national goals when deciding on fiscal policies whereas the common central bank’s monetary policy aims at union-wide objective variables. The union considered is asymmetric, consisting of a “core” with lower initial public debt, and a “periphery” with higher initial public debt. We analyze several possible solutions to deal with these crises within the dynamic game model under consideration.