| China's stock market started since 1990 and developed by leaps and bounds. However, it is still an emerging market, subjected to many investment non-economic factors interfere and most investors have emotional irrational behavior, making the stock price volatility. Since China's stock market is not mature yet, researches on the Chinese stock market become more necessary. Moreover, since the 21 century, China's stock market has experienced a number of distinct stages, showing obvious circle of bear and bull market. So it is necessary to divide the stock market into different stages and study their performance separately. In addition, the complex network as a research tool for financial and stock markets, having been used and affirmed by more and more scholars.Based on these, we establish complex networks:minimum spanning tree, hierarchical tree and planar maximal filtering graph, and study the networks' topology property and clustering structure during the bull and bear. Details are as follows:First, we select Shanghai 50 Index constituent stocks actual price daily data and divide the stock market into bull or bear stages. Then, we use the DCC-MVGARCH model construct dynamic correlation coefficient, so that we can get the minimum spanning tree hierarchical tree and planar maximal filtering graph. And then, we study the network's topological structure properties from the view of minimum spanning tree and planar maximal filtering graph, including the average path length, clustering coefficient and degree. And then we analyze the clustering effect by the planar maximal filtering graph. Finally, we compare the minimum spanning tree and maximum filtering method, and give recommendations to regulators and investors.By building the complex networks of dynamic stock market, we study the networks' topology property and clustering structure during the bull and bear periods. This may provide a useful reference for stock investment and risk supervision. |
PDF Full Text Request