Research On The Heterogeneous Change And Dynamic Forecasting Of Volatility Driven By Liquidity

Liquidity and volatility have played extremely important roles in the financial markets and are becoming two key factors that characterize the quality and efficiency of the market.With the rapid development of the financial market,global financial market problems caused by liquidity have become more and more serious.Pástor and Stambaugh(2003),as well as Brunermeier and Pedersen(2009),have regarded a sudden depletion of liquidity as the direct factor for the market to fall into a ?crisis? state.Increasingly scholars and practitioners are paying more attention to the liquidity problems.The level of liquidity with time-varying has become a risk factor in financial asset pricing,portfolio income forecasting and risk management,especially the volatility changes driven by liquidity which are even more concerned.In view of this,our thesis starts from the mechanism of volatility changes driven by liquidity,and studies the heterogeneous distribution of volatility changes which are driven by the liquidity.It is found that the more "turbulent" of the market,the more drastic of the changes in liquidity-driven volatility.Secondly,considering the significantly difference of the volatility changes driven by the liquidity between the "turbulent" state and the "steady" state of the market,this thesis proposes the liquidity-adjusted HAR model under the framework of Markov-switching(MS)(the fixed transition probabilities and the time-varying transition probabilities)which can capture the dynamic changes of the volatility driven by liquidity,and forcast the future volatility in China's stock market.The main empirical results are as follows.(1)Based on the quantile heterogeneous regression model,the first empirical result presents the distribution of volatility which is driven by the liquidity regarded as an important risk factor of the market in China's stock market for the first time,and obtains the heterogeneous non-linear characteristic of liquidity impact on volatility,that is,an decrease in liquidity increases the volatility significantly,especially at the up-tails,and leads to an unique J-type distribution,which is different from the U-type distribution of the U.S.futures market.We explain the difference by the majority of retail traders who have very obvious irrational behavior in China's market.Furthermore,the heterogeneous effect of liquidity on the volatility is asymmetric with significantly stronger on ?bad volatility? than on ?good volatility? in China's stock market.Finally,we find that the strongest impact of liquidity on volatility is in the last half-hour before closing,followed by the first half hour after opening for each trading day when we focus on intraday pattern.(2)Economic and financial markets are always switching in different states of?turbulence? or ?stability? because of numerous complex factors such as economic cycle,major events and important policies.The market microstructure theory points out that the influence of liquidity on financial market changes along with the different economic states(?turbulence? or ?stability?).With liquidity being regarded as market risk factor,this thesis proposes the Markov-switching(MS)liquidity-adjusted HAR model to capture the dynamic change characteristics of volatility driven by liquidity,and uses the ?Model Confidence Set?(MCS)test to evaluate various realized-range volatility forecasting models.Furthermore,considering that many liquidity proxies contain market volatility information,we further propose the MS-liquidity-adjusted HAR model by eliminating the market fluctuation information from the liquidity measures and find that it can not only capture the dynamic changes of volatility driven by liquidity,but also performs better in forecasting the volatility.(3)The liquidity with time-varying is considered to be an important risk in asset pricing,and the sudden depletion of liquidity has been becoming the core factor leading to the financial crisis(Pástor and Stambaugh,2003;Brunermeier and Pedersen,2009).So this thesis discusses whether sharply deteriorating liquidity propels the stock market into a?crisis? state and investigates the dynamic impacts of the market liquidity on volatility forecasting by using the Markov-switching(MS)liquidity-adjusted HAR model based on time-varying transition probabilities(TVTP).Empirical evidence suggests that a sharp deterioration in liquidity increases the probability of a ?crisis? state for China's stock market.Out-of-sample forecasting results show that the model we proposed(i.e.,the TVTP-MS-HAR-CJ-LIQ model)substantially improve the predictive performances relative to both the simple HAR model and other generalized HAR-type models.And it can provide some theoretical guidance for market traders and policy makers.