Year of Graduation
Extraction of a Stochastic Global Trend on the Daily Stock Market Indices
This paper proposes approaches to the volatility modeling and extraction of the global stochastic trend from daily closing values of the stock market. A Kalman-filter type model is suggested, which splits the daily returns of each index into two independent components — local and global. The model takes into account the asynchrony of the stock indices and possible autocorrelation in the global trend. The return-volatility is estimated by the autoregressive conditional heteroskedastic model. The performance of the Kalman-filter type model for the estimation of local and global components of return-volatilty is explored for three markets: Japan, UK, US. Ideas are proposed for further research.