The Efficiency of an Artificial Double Auction Stock Market with Neural Learning Agents

Abstract

The goal of this paper is to investigate the convergence property in a double auction market and test the robustness of the results. We construct an artificial equity market where agents trade a risky asset that pays a stochastic dividend each period. Artificial Neural Networks take on the role of traders, who form their expectations about the future return and place orders based on their expectations. Market prices are set endogenously by trading among agents in a double auction market. The efficiency of this artificial market is measured by the convergence of the price to the Rational Expectations Equilibrium (REE). We find that market dynamics under double auction converge to the REE in in some ex-periments. This convergence, however, is sensitive to the deviation from rationality among the agents. In the experiment where we introduce noise trading, convergence becomes unattainable. Minimal rationality is not sufficient to generate convergence in a double auction market when the market price is endogenous.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 39 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.