Research On Measuring Unexpected Risk Based On Realized Volatility And Its Application

Under the rapid development and ever changing of financial instruments,the domestic financial market faces numerous risks and challenges.Compared with expected risks,unexpected risks seriously hinder the investment activities and risk management of market participants,and even cause systemic financial risks.In academia and industry,quantitative management of unexpected risks has always been the focus and difficulty.On the basis of reading a large number of literature reviews and brokerage research reports,this paper introduces multiple types of predictors and constructs different datasets based on these predictors.By comparing the prediction performance of various realized volatility prediction models under different datasets,this paper proposes an unexpected risk measurement model based on realized volatility.The model can decompose realized volatility into expected volatility and unexpected volatility,corresponding to expected risk and unexpected risk,respectively.In addition,this paper deeply studies the relationship between unexpected risk,expected risk and portfolio return by grouping individual stocks,and selects the corresponding investment portfolio according to the relationship to construct the optimal quantitative investment strategy.Finally,this paper further analyzes and discusses the constructed quantitative investment strategy.The empirical conclusions are as follows:First,under different datasets,the prediction performances of machine learning methods are similar with that of HAR model.The degree of predictability of realized volatility is between 52% and 58%.Second,the relationship between unexpected risks and return is significantly negative.When volatility is overestimated,significant positive risk compensation will be obtained,and the more sensitive individual stocks are to unexpected risk,the higher the compensation return.The relationship between expected volatility and return can be positive or negative.The less sensitive a stock is to expected volatility,the higher is the return.Third,the quantitative strategy constructed in this paper has significant excess returns.Compared with the CSI300 index in the same period,the strategy is more likely to obtain significant excess returns in a volatile market or bull market,but the maximum drawdown of the strategy is larger in a bear market.Finally,The Fama-French three-factor can explain unexpected risks factor to a certain extent,but there is still 55% of the information that the three-factor model cannot explain.In a word,the research on unexpected risks in this paper explains the mystery of volatility and return,and the constructed quantitative investment strategy has guiding significance for investors to manage unexpected risks,and also provides data and model support for investors in stock selection decisions.