| The global financial crisis has made the world economy stuck in recession,and China’s economic structure has been changing since the crisis,with a decline in the economic growth rate.Having in-depth knowledge of the financial market and maintaining the financial stability are of great importance for the stable growth of China’s economy under the complex external and internal conditions.On the other hand,the “Big Data” trend is stretching to various aspects of the economy thanks to the development in the information technology.While bringing convenience,this trend has provided new perspectives on learning about economic phenomena.Based on the methodologies of high-dimensional factor analysis,this dissertation investigates into fluctuations in China’s stock market in this circumstance,serving as reference for learning about the pattern of financial fluctuations as well as financial risk management.This thesis introduces the theory of volatility and high-dimensional factor models and then reviews research on financial fluctuations through high-dimensional factor analysis.Based on the existing research,this thesis studies fluctuations in China’s stock market from four perspectives empirically.The empirical analysis first considers the aggregate market,investigating the predicting performance of sparse components in macro fundamentals on aggregate volatility,in which the method of high-dimensional factor analysis is used to summarize information form the subset of high-dimensional observed macro variables.Then the thesis focuses on the individual stock level,and explores the grouped patterns in individual share volatility,with the group-specific factors driving the within-group co-movement in volatility.The thesis goes a step further to extend the analysis to the tail systematic risk,with the common component extracted from the conditional quantiles of stock returns describing the change of the extreme systematic risk of individual stock.Finally the thesis investigates the treatment effects of the expiration of share lockups on individual stock price characteristics based on a policy evaluation framework for panel data,using the method of high-dimensional analysis to estimate the counterfactuals.The results from the analysis on the aggregate market show that the predicting power of sparse factor structures in macro fundamentals on aggregate volatility is stronger than the subsets of sparse characteristics which are the variables themselves,and that the volatility predicting patterns of sparse factors and characteristics are different.The investigation into the individual shares emphasizes the latent factor structures in individual volatility,through which the group membership is determined differing from the industry classifications of stocks.The common factors are closely related to macro fundamentals and have predicting power on individual share volatility.The results about the extreme risk indicate that the upside and downside tail systematic risk of individual stock is asymmetric,and that firm characteristics can account for the heterogeneous impact of macro risk factors on tail systematic risk.The political analysis reveals that the treatment effects of the expiration of share lockups on volatility are stronger than returns and that individual share characteristics can hardly explain the heterogeneous treatment effects.Overall,the application of methods related to high-dimensional analysis in this thesis can help discover the risk characteristics in China’s stock market which reflected from the co-movement of economic and financial variables and serves as reference for the management on and the resolution of financial risk. |
