Trading duration, mutual funds behavior and stock market shock : Based on ACD model to mine mutual funds investment behavior

Purpose – The purpose of this paper is to explain what information is contained in mutual funds' trading behaviors and to try to further assess the impact on the stock market. Design/methodology/approach – The objective is achieved by an empirical examination using the high‐frequency intraday data. The main methods used for the research are the autoregressive conditional duration model and the UHF‐GARCH model. Findings – This paper gives an empirical study of mutual funds' behavior on two aspects. The first aspect is the direct impact on micro variables. The results show that mutual funds changing their positions will have different influences to the spread, adding position broadens the spread, while decreasing position makes the spread narrow; behaviors of funds change the clustering characteristic of the duration. The second aspect is the impact on the relationships among micro variables. The results indicate that trading started by liquidity buyers will make volatility larger. Research limitations/implications – This paper supposes funds as informed traders and individual investors as liquidity traders in China's stock market. If it is not true, some interpretations of empirical results would be wrong. The authors' results may help researchers to understand the information content of funds' trading behaviors in the microstructure aspect. Originality/value – The paper is an original work, which will be interesting to scholars in market microstructure and to practitioners in the Chinese stock market. The main contributions of the paper are: the use of high‐frequency data to study funds' behaviors and combine the trading duration and investors' trading behavior to analyze the information content of trading behaviors; second, the use of 14 stock samples in the Shanghai Stock Exchange to do the empirical study, which ensures the reliability of the results.