Research On Tail Features And Extreme Value Dependence In Stock Markets

In the year, Mandelbrot (1963) had indicated that the distribution of asset returns is not well approximated by the Gaussian normal, but with fat tails. And then, much research on rare events arose from this contribution and Mandelbrot's points had been proved well. And the existence of fat tails affects many methods and technologies of financial risk management, portfolio selection, investment hedging strategy, and so on. Further more, some financial extreme events in recent years have made great impact over the world. Therefore, it is becoming more important to study the dependence of extreme events. And this paper continues the research on the tail features of asset returns, especially of stock markets. Based on the research, we go deep into the study of extreme value dependence in stock markets, that is, the probability of the large change for stock price such as rise and drop wildly at the same time. And this research may provide the theoretical basis for efficient asset assignment over the world.In this paper, extreme value theory (EVT) is it's mainly research method. Based on EVT, we apply the POT (Peaks over Threshold) and BMM (Block Maximum Model) model to study the series of asset returns in stock markets and their tail features and their fitted distributions. Then, we apply the measures and estimation method of asymptotically independent and asymptotically dependent provided by Tawn (2004) to study the extreme dependence structure between different stock markets, especially between domestic stock market with other countries' stock markets. Through our analysis, we can make following conclusions: the tail distribution of stock returns series is dissymmetrical, left-tail is fatter than right-tail, that is, the frequency of drop wildly of stock price is larger than that of rise wildly of stock price. And from December 16, 1996 when the system of our country's stock market was transformed, the tail feature of the return series of Shanghai complex index is consistent with the point. And the extreme dependence between stock markets is also dissymmetrical; left-tail dependence is stronger than right-tail dependence. Moreover, the extreme dependence structure between developed countries is more stable, and there is strong tail asymptotic dependence between developed countries. And the tail asymptotic dependence between our country's stock market with other countries' is much weaker. Tail dependence between our country's stock market with other countries' appears to increase in more recent time.