Investigating China's Stock Market Efficiency and Forecasting China's Stock Price Volatility

Student thesis: Doctoral ThesisPhD

Abstract

This thesis aims to investigate the efficiency of China's stock market, detect the macroeconomic and financial drivers of China's stock price volatility, and assess the power of the most significant drivers for predicting China's stock price volatility. The data used in this thesis starts from 2005 since this is when the non-tradable reform was implemented, significantly influencing on the efficiency of China's stock market. The stock price volatility has been obtained using the GARCH-MIDAS model and a comprehensive set of drivers is considered to try to exploit more information. Using a variety of unit root tests, this study shows that China's stock market is weak-form efficient. In addition, this thesis also finds that the US stock market and the development and openness of China's stock market has a significant effect on China's stock price volatility using several novel significance tests and the time-varying VAR model. However, this information is shown to have no additional ability in predicting China's stock volatility after controlling for the past information contained in the stock prices using the penalized regression models and Support Vector Regression model. This thesis contributes to the current literature in several aspects. First, a panel unit root test allowing for the smooth structural breaks and accounting for the cross-sectional dependence, which has not been applied in the current literature, is used to examine the weak-form efficiency of China's stock market; Second, it considers a comprehensive set of potential drives, especially the drivers related to the development and openness of China's stock market which the current literature has not put an emphasis on. Moreover, a number of new significance tests based on the penalized regression models are used for the first time to detect the impact of these drivers on the stock price volatility; Third, it uses machine learning techniques and penalized regression models to assess the power of the macroeconomic and financial variables in predicting China's stock price volatility, which can provide more robust evidence than those provided by the current literature.
Date of Award4 Nov 2020
Original languageEnglish
Awarding Institution
  • University of Bath
SupervisorBruce Morley (Supervisor) & Haibao Wen (Supervisor)

Cite this

'