The Research On The Price-volume Relationship Of Stock Based On Agent-based Computational Finance

In summary, the empirical research on price-volume relationship focused on four areas: trading volume and absolute price changes, trading volume and price changes itself, dynamic causal relationship between trading volume and price changes, trading volume and price volatility. This paper mainly focused on the relationship between trading volume and absolute price changes, price volatility.Large numbers of empirical studies conclude that there is a strong positive correlation between absolute price changes and trading volume,which is the price-volume relationship. Until now, the explanations of the relationship are mostly based on two models, the mixture distribution hypothesis and the asymmetric information theories. Among them, the former study can’t give mechanism explanation from the perspective of market microstructure. At the same time, the latter study has not reached the consistent conclusion now. The main reason is that we can’t get the data of the information-driven trading volume from the stock market.The agent-based computational finance has not involved in the study of the price-volume relationship. This paper argues that it’s feasible to develop a simple artificial stock market model, taking into account its’ advantages of bottom-up micro modeling and the controllable subjects. It not only is good for obtaining the data of information-driven trading volume, but also is good for studying the price-volume relationship from the market microstructure level.The works we have done are as follows in this paper:(1) From the start, the mechanism of the trading volume is discussed, then different trading characteristics between informed agents and uninformed agents are distinguished, which lay the foundation for the modeling of agents; based on this, the hypotheses about why and how the price-volume relationship generated are put forward.(2) An artificial stock market model is developed from three aspects of assets, agents and market price formation mechanism. The artificial data is obtained, for example the returns and the information-driven trading volume. The hypotheses are verified.(3) Five research models are established using the GARCH models; based on this, price-volatility relationship is studied with the artificial time series data.The main conclusions are as follows:(1) The cause of the price-volume relationship is the information-driven trading volume.(2) The prices quoted by the informed agents are more aggressive impacts the produce of price-volume relationship, but is not the essential condition.(3) Trading volume of the informed agents are more large is the essential condition for the produce of price-volume relationship.(4) Compared with the trading volume, the positive correlation between informationdriven trading volume and price volatility in the short term is more significant, the information-driven trading volume can explain the persistence of price volatility well. There is a negative correlation between the non-information-driven trading volume and price volatility in the short term. The information-driven trading volume can take the place of trading volume. Then we can get the conclusion that the leading role of the relationship between trading volume and price volatility is the information-driven trading volume.The artificial data, the condition hypotheses and the conclusions of this paper are innovative to some extent. The conclusions also have some realistic significance, for example, they are conducive to grasp the rules of price volatility for the stock market researchers.