GARCH(1,1)모형에 관한 베이지안 연구

Abstract

The objective of this study was to estimate the fluctuation of anearning rate and risk management using the price index of Korea stocks. After an observation of conception of fluctuation, GARCH model which is presumed heteroscedasticity with the price index of stocks was used. Based on Bayesian method, the Volatility clustering and fluctuation phenomenon was found to be lasted using GARCH model. In addition, the effects of fluctuation on the time-series was evaluated, which showed the heteroscedasticity. Winbugs, Bayesian software, and MCMC method were used for analysis. The basic of risk management for finance theory, VaR, was presumed by monte carlo simulation with Simple Moving Average (SMA). SMA is one of the fluctuation methods that could be used operationally to predict the fluctuation and fluctuation on GARCH (1,1) model. In addition, VaR was compared with actual earning rate. With these methods, it can be concluded that GARCH (1,1) model could reflected the recent status of market and was more suitable to estimate the fluctuation of finance time-series data than SMA.