Font Size: a A A

Comparison And Research Between Higher-moment And Conditional CAPM Models In Shanghai Stock Market

Posted on:2007-12-23Degree:MasterType:Thesis
Country:ChinaCandidate:Y D WangFull Text:PDF
GTID:2179360182980790Subject:Quantitative Economics
Abstract/Summary:PDF Full Text Request
Forecasting stock yield is the kernel problem in researching asset price-making. Though more and more scholars admit that stock yield may be forecasted, it is still a puzzled problem for Chinese academic field that how we can find an effective financial asset price-making tool that is proper for Chinese stock market.Capital Asset Pricing Model(CAPM) which appeared in the middle 1960s is the first equilibrium model about financial asset price-making. It is also the first financial asset price-making model that can be tested through metrologic method, and leads to a revolution in financial theory and had an important historic significance. The central character of CAPM is that only systematic risk works in stock price-making and that stock income is in a direct proportion with its systematic risk which is measured by β.Scholars at home and abroad have made a lot of research about CAPM. It is universally accepted that the model is basically in accordance with the market. But still much evidence was found that it is not in accordance with the market. The article tries to make some improvement on traditional CAPM.After analyzing why the precision of traditional CAPM is not high, we made some improvement on traditional CAPM in two aspects. First, we introduce systematic skewness and systematic kurtosis into traditional CAPM, making it to be an even generalized model containing higher-moment. Second, considering the time-varying quality of β , we use Multivariate GARCH (Generalized Autoregressive Conditional Heteroscedastic) model to make time series analysis. Thus, we develop traditional CAPM into a conditional CAPM. The purpose of the article is to find a financial asset price-making tool that is proper for Chinese stock market and guide security investment with it.In data experiment, after strict statistic quality tests of the sample data, we found two improved CAPM, made data regressing experiment and also made short and middle forecasting. According to the result, the author made a comparison between traditional CAPM, higher-moment CAPM and conditional CAPM. Themain results of the article are as follows:1. The introduction of higher-moment is beneficial to the improvement of forecasting precision.2. Financial time series has the auto-correlation character to some extent. In conditional CAPM, GARCH model which we used could describe this character of financial time series well.3. The forecasting precision of GARCH model is limited, but it still has a good ability in short and middle forecasting.4. Conditional CAPM has a higher forecasting precision than higher-moment CAPM.
Keywords/Search Tags:CAPM model, Systematic skewness, Systematic kurtosis, conditional CAPM, Multi-GARCH, time-varying Beta
PDF Full Text Request
Related items