A dynamic model of the limit order book (Q2284921)

The authors consider an equilibrium model of the limit order book (LOB) in a stock market, where a large number of competing agents post ``buy'' or ``sell'' orders. Having observed the prices asked by his competitors, each agent must determine an optimal strategy, maximizing his expected payoff. Because of the presence of the other sellers and of the upper bound on the acceptable price, asking a higher price for an asset reduces the probability of selling it. As a consequence, a unique shape of the LOB is determined, which represents a Nash equilibrium between the various agents. The paper has two main goals. First, a two-sided LOB is studied, where the incoming order can be either a buy or a sell order. In this situation, a large number of agents, each holding a small amount of cash and stock, must determine a pricing strategy. In this setting, two results are proved showing that the shape of the two-sided LOB, in the moment when the incoming buy or sell orders can be possibly executed, can be uniquely determined. More general case, where the maximum or minimum prices acceptable to external agents are random, is also treated. As a second extension, the authors consider a time-dependent problem involving sequence of \(N\) incoming orders and establish conditions ensuring that, at each time step, the shape of the two-sided LOB when the incoming buy or sell orders of time \(1\le i\le N\) are possibly executed, can be uniquely determined by backward induction.