Empirical Study On The Measurement Of Stock Liquidity Based On The Analysis Of Ultra-High Frequency Data

Databases recording all the information about trading and related elements have become available since the increased automatization of stock markets and related information technology.The availability of detailed databases on intra-day activity in stock markets has opened the possibility of econometric investigations on the microstructure of such markets. China stock market has developed to the second one in asian, with the whole market value and the number of listed stock increasing greatly. As the market developing, the trading volume has expanded greatly and the volatility of stock index become highly.As we know, the main function of stock market is providing the good chance for investor to trade immediately and efficently at low cost. It means that the market should provide ennough liquidity. In this paper, I propose volume durations, i.e. the waiting time until a predetermined volume is absorbed by the market, to model the intraday liquidity. Since this concept measures the trading volume per time it is strongly related to market liquidity, I also focus on volumes measured independently of the side of the market as well as on buy volumes, sell volumes and volumes measured on both market sides simultaneously. As this measure varies over time, it can be forecast and explained. As for econometric modeling of the different duration concepts, I apply the Log-ACD model to volume durations relative to the trade process of fifty securities listed on the Shanghai Stock Exchange. This paper introduces the related literature and the concept and measure of liquidity. Then the paper introduces the econometric frame for duration, including the linear and non-linear ACD model and the analysis process of duration. As a consequent, the paper constructs and analysis the whole and one-side duration of 15 stocks listed in Shanghai Stock Exchange in the econometric frame. Finally, the paper discusses the intra-day pattern of the liquidity and the estimated results. And by evaluating in-of-sample forecasts, evidence is provided that Log-WACD models are a valuable tool for predicting volume durations.