Predicting The Market Excess Return By Illiquidity Basing ARFIMA Model

Liquidity is a complex concept that usually means the ability to make large transactions quickly and cheaply without changing prices.Illiquidity,the opposite of liquidity,is often closely related to transaction costs and market friction.High degree of illiquidity is often accompanied by market depression and financial crisis,which makes it possible to use illiquidity index to predict market excess return.The purpose of this paper is to build various proxies of illiquid from price reaction,effective spread,market volatility,the return reversal and so on.Through the linear aggregation method and principal component analysis,we build a comprehensive market illiquidity measure.Finally,predicting market excess returns from short,medium and long term.This paper found that the long memory and contemporaneous shock of illiquidity proxies and market excess return,which may cause the prediction error.Therefore,combining with ARFIMA model,we build a Monte Carlo simulation.According to the results of Monte Carlo simulation,we decide if establish Bootstrap sampling robustness of predicted results or not.We can correct slope and filter out the proxies that can pass the Bootstrap test.And we explore the impact of the reform of non-tradable shares on the illiquidity proxies in Chinese stock market.By the real-time adjustment method and stochastic detrending method to obtain new illiquidity proxies,we predict the market excess return and compare with the raw illiquidity prediction model.In order to further explore the impact of market value of stock on predictive power of illiquidity proxies,the article select the data from Shanghai and Shenzhen 300 Index,and divide it into three groups based on the market value.With raw illiquidity measures and adjusted measures,we predict the excess market returns of different market value groups from short,medium and long-term aspects,in order to provide a reference for investors investment decisions.In view of the significant correlation between market volatility and illiquidity proxies,we use orthogonal method and regression method to eliminate the market volatility factors of illiquidity proxies,and test the robust prediction ability of illiquidity proxies.The main research conclusions of this paper are as follows: long memory and contemporaneous correlation of variable will cause bias of estimation,and the degree of error depends on the parameter of long memory and contemporaneous correlation coefficient.Illiquidity index can make short,medium and long-term forecast for the return of Shanghai and Shenzhen 300 Index.Among them,the effect of medium-term forecast is the best;The forecasting ability of illiquidity proxies prediction model adjusted by sudden change can be improved.The coefficients between raw illiquidity measures and the expected return are basically positive,which is consistent with the expectation of liquidity premium theory,proving the existence of liquidity premium in Shanghai and Shenzhen 300 Index.The ability of illiquidity index to predict market value decreases with the increase of target market value.The return of small stocks is more sensitive to shocks of market illiquidity proxies.