Research On The Influence Of High-frequency Realized Skewness On Stock Returns In Chinese Stock Market

Chinese stock market,as an emerging financial market,is always influenced by multiple factors in different aspects.The asymmetric distribution of stock returns would be especially amplified while the financial market foam changes with the stock prices sharply rising and collapsing.Therefore,using variance to identify the returns volatility wouldn’t be enough.It’s really necessary to form another proxy variable,as known as the skewness,to manifest the asymmetry of the distribution.With more and more applying of the high-frequency data,it turns to be possible to calculate the high-frequency realized skewness.And it will do great in analyzing the distribution characteristics,risk management and assets allocation.From the collections of researches on high-frequency realized moments,the reasons why skewness has become essential and applied,and discovery on the factors impacting the expected returns,we found there was none considering the analysis of high-frequency realized skewness.In addition,the high-frequency realized skewness is the efficient,unbiased and consistent estimation including more information than the low-frequency realized moments.Thus,this thesis would compute the realized skewness based on high-frequency data according to the existing research.After that,we sorted the realized moments of each stock into seven groups and recorded the group returns during the next week.Through generating the long-short portfolios,we tried to make clear the relations between realized moments and stock returns.Meanwhile,we further assess the pricing ability of the high-frequency realized skewness by carrying out Fama-MacBeth regressions adding different variables.With empirical methods used in the whole sample period and specific periods,and various robust tests like changing the computation of the realized moments and extending the assets holding periods,we recognized stock high-frequency realized skewness impacted expected returns most among different realized moments and it’s totally significantly negative.The returns of whole groups sorted by the realized moments also implied that the expected returns would decrease due to the realized skewness.Furthermore,the long-short portfolios based on the skewness also showed the highest returns.Though we still found the different outcomes on realized kurtosis from existing research,the realized skewness did have the most significant effects on expected returns while the relations with expected returns were not robust using realized volatility,and the returns using realized kurtosis were not that much as skewness did.