Asymmetric pricing: an agent based model

• Authors: Federico Cecconi;Domenico Parisi
• Venue: MOAS'07 Proceedings of the 18th conference on Proceedings of the 18th IASTED International Conference: modelling and simulation
• Year: 2007

Asymmetric pricing is the tendency of prices to rise readily and fall slowly. Explanations of asymmetric pricing mention monopoly power, costs of changing prices or spreading information, and elasticity differences. In this paper we describe a general dynamic, in a Walrasian competitive equilibrium market, where consumers and producers learn how to behave in the market. For each consumer (producer) we compute a payoff function which has two components: (a) the ability of the agent to actually carry out exchanges; (b) the ability of consumers to minimize, and of producers to maximize, the price actually paid in the exchanges. We study the effects of different learning algorithms, and in particular we compare individual vs. social learning procedures. Consumers arrive to the market with a maximum price each consumer is willing to pay, and producers with a minimum price each producer is willing to accept. We compare two different modalities for price clearing. In the food-market modality the payoff is computed using the exchange price requested by the producer. In the auction-market modality the payoff is computed using the exchange price offered by the consumer. In both modalities an exchange takes place only if the consumer's price is greater than the producer's price. The simulations show that asymmetric pricing emerges only in the food-market modality, as predicted by an analytic study of the model. Finally, we compare the simulation results with a classical asymmetric pricing model, the menu cost model ([1]).