Calibration of the agent-based continuous double auction stock market by scaling analysis

This paper proposes one calibration method for the agent-based continuous double auction (CDA) stock market by scaling analysis based on the work by Pasquini and Serva (1999) . We design and build an agent-based CDA stock market, which uses the same trading mechanism as the Chinese stock market. We also perform a scaling analysis of the absolute returns in both the artificial and real stock markets. The results show volatility correlations as power laws in all the markets. More importantly, the power-law exponent is not unique, and all such exponents follow a multi-scale behavior. All exponents image trend to the theoretical value 0.5 with increasing scaling index γ. Scaling character is an important intrinsic quality of the stock market, and this method can be used in calibrating the agent-based stock market model.