GARCH-based Empirical Analysis Of Volatility In Shanghai Securities Composite Index

Volatility research bears great significance in asset pricing. Whether mod-el applications and investment strategies work smoothly or fail in vain depends on the accuracy of volatility estimation. The history of volatility research is relative long, yet in1982Engle articulated the idea of ARCH, which brought about a fresh start for researchers thereafter. In1986, based on ARCH Boller-slev added historical volatilities in the ARCH model, designed a generalized ARCH model, named GARCH, which better described the long memory fea-ture of stock volatility. Then, GARCH became a very foundation of quantita-tive measure for volatility. In this paper, GARCH and EGARCH are utilized for empirically analysing observations of everyday close price in Shanghai Se-curity Composite Index, with a sample size of3942, dated from1997.1.2to2013.6.7. By comparison of GARCH and EGARCH results, our conclusions are drawn.There are5sections in this paper. In the first section, research back-ground, meanings, methods, purposes and frameworks are introduced. The second section mainly introduces ARCH, GARCH and EGARCH models, in-cluding their definitions and ideas behind, focusing on the advantages and shortcomings of each model. The third section tests and analyses the samples, including testing autocorrelation, stationarity and heteroscedasticity, and con-cludes that ARCH models can be applied in analysing volatilities in Shanghai Security Composite Index. In the fourth section, numerous empirical analy- ses are done by both GARCH and EGARCH. In the fifth section, conclusion are drawn. The main results are that though both GARCH and EGARCH can be utilized in sample analyses, the outcomes deviate; With this sam-ple, EGARCHs generally do better than GARCHs; Within EGARCHs, E-GARCH(5,6) does best in capturing the fat-tail feature in the stock market; the entire sample may not have a significant leverage effect, yet the segmented sample does have, in2001-2008, an reversed leverage effect is found, which may off set some leverage effect in the samples as a whole.Volatility research is crucial among regulators, practitioners, scholars and investors. I hope this paper can help market stakeholders for reference, and contribute to building a financial market which is more functionalized, flexible and resilient.