A Study On Volatility Characters Of Securities Market Return Based On Change Point Detecting

It is an important research field in finance to do systemic and in-depth theoretical analysis and empirical research of the volatility characters of securities market return. However, as an emerging market, the Chinese securities market shows great fluctuations and engenders sudden changes due to the imperfection of market mechanism, the infection of the national policies, the economic environment and the investors’ anticipation. This paper analyzes the volatility characters of the Chinese securities market return from the view of change point.This paper firstly reviews the relevant literature of change point and securities market return volatility. It summarizes the application of the change point methods in the field of the volatility of securities market return. It also does some theoretical investigation on the methods to detect change point and the volatility characters of securities market return. On this basis, this paper does some relevant tests of the basic statistical characters of securities market return. Then it uses modified ICSS algorithm to detect change points. And it looks for the events near by the change points. Finally, it divides time period of volatility based on the detected change points. By constructing dummy variables, this paper puts the structure change factors into the GARCH family models to study the effect on the volatility characters of securities market return when there are change points.The results show that there are30change points respectively in the return time series of Shanghai Securities Exchange Composite Index and Shenzhen Securities Exchange Component Index. As an emerging market, the Chinese securities market abnormally fluctuates more intensely and frequently. Most of the events corresponding the change points are policy events. The Chinese securities market is a "policy driven market’". There are spurious persistence and asymmetric effect in the return volatility. The TARCH model with dummy variables is much better for reflecting the volatility characters of securities market return.