Discovering Traders’ Heterogeneous Behavior in High-Frequency Financial Data

Abstract

This paper develops a utility-based heterogeneous agent model for empirically investigating intraday traders’ behaviors. Two types of agents, which consist of fundamental traders and technical analysts, are considered in the proposed model. They differ in the expectation of future asset returns and the perceived risk. This paper incorporates the unique characteristics of high-frequency data into the model for the purpose of having a reliable and accurate empirical result. In particular, a two-test procedure is developed to test the market fractions hypothesis that distinguishes the heterogeneous agent model from the representative agent model. The proposed heterogeneous agent model is estimated on the Taiwan Stock Exchange data. The results suggest that fundamental traders expect the correction of over- or under-pricing in the future. Technical analysts act as contrarian traders. Technical analysts also believe that buyer-initiated (seller-initiated) trading will further raise (lower) future prices. The bid-ask spread has a crucial effect on the investment risk for the technical analysts. Moreover, technical analysts are short-sighted, have less market fraction, but perform slightly better.