An Empirical Study On The Predictability Of Excess Returns On Corporate Bonds In China

As a common security,bond has been an important part of the financial market,providing a low-cost financing channel and a low-risk investment option.In recent years,with the promotion of the policy and the deepening of marketization process,corporate bond develops faster and faster,and now the issue size is more than 3.9 billion.Corporate bond is playing a more and more important role.Therefore it’s of great important significance to study the predictability of excess return on corporate bonds in China.This paper selects 283 corporate bonds listed in Shanghai Stock Exchange and Shenzhen Stock Exchange,divides these bonds into groups according to the term and credit rating,and studies the predictability of excess return of each group.The data sample range is from January 2010 to November 2016.On the basis of previous studies,this paper regresses the monthly excess return on corporate bonds and some factors from bond market and stock market.The study finds that factors from bond market such as the credit spread,term spread and liquidity can predict the excess return,and the volatility of stock market works either.As for research methods,this paper chooses the combination prediction method,that is,successively applies VAR model and ARMA model to predict the excess return on corporate bonds,and then combine the two models.The result indicates that the prediction results vary dramatically according to the term and credit rating.Overall,for the same term,the higher the credit rating,the higher the prediction accuracy,for same credit rating,the shorter the period,the higher the prediction accuracy.Compared to VAR model or ARMA model,the combination prediction method can improve the prediction accuracy.