The Intetremporal Relation Between Expected Returns And Risk

This paper explores the time-series relation between expected returns and risk for alarge cross section of industry and size/book-to-market portfolios. I use a bivariategeneralized atuoreressive conditional heteroskedasticity (GARCH) model t estimatea portfolio‘s conditional covariance with the market and then test whether theconditional covariance predicts time-variation in the portfolio‘s expected return. Fora further discussion, I analysis the applicability of bivariate GARCH model in Chinastock market by changing the samples and portfolios. I also study the intertemporalrelation between expected returns and risk, and the time-series in this relation, andthe cross-sectional character, along with a simple discussion on the market riskpremium. I am satisfied with the empirical results:1. To use the sample interval from Jan2000to Dec2011to estimate the conditionalcovariance and conditional beta, the results find the fact that they all vary with thetime, and also explain that the bivariate GARCH model could be used for estimatingconditional covariance and conditional beta, and it is reasonable.2. To change the sample interval to estimate the parameters, I find that less statisticalvalidity in one parameter or two, just owing to the shock and reform in China stockmarket in1997. However, the estimates still reflect the time-vary character and thereasonable model building. All of above can be say as the following: the bivariateGARCH is applicable in China stock market.3. Restricting the slope to be the same across the equation system to estimate therelation between risk and returns, I find pleased results, the slope is above zero andhave statistical power. Although it is based the given model, the results could explainthe ICAPM model is valid and sufficient in China stock market. Based on the above,when consider the investment opportunity that is stochastic, I make a simple analysison the risk premiums induced by conditional covariance with SMB and HML, theresults are excellent power in statistics, which give a tip for the investors to stare atthe assets with good SMB‘.4. The cross-sectional analysis on ICAPM is to compare with other study methods infact. At luckily, it unites the results in the above, and the applicable, reasonable ofusing conditional covariance to explain the relation between risk and returns, andcross-sectional consistency, that is to say, the consistent cross-sectional relationbetween expected returns and risk. Especially, it is reasonable in economic.