Behavioral Biases and Informational Inefficiency in an Agent-Based Financial Market

The role of competitive markets as efficient aggregators of decentralized information is a fundamental problem in economic theory. This paper studies the informational efficiency of a market with a single traded asset, in which agents expectation formation about future price has two kinds of deviations from rationality. First, traders have adaptive expectations, i.e. they give more importance to the past price than a rational agent. Second, the agents are subject to the confirmatory bias, i.e. they tend to discard new information that substantially differs from their priors. Taken separately, each deviation worsens the informational efficiency of the market. However, for some ranges of parameters, when the two biases are combined, they tend to mitigate each other effect (thus increasing the informational efficiency). We also study the robustness of these findings to alternative specifications concerning market participation, entry of new agents, and the amount of liquidity that agents hold.