Stock returns volatility in the Tokyo stock exchange

Abstract

This paper examines the stock returns volatility in the Tokyo Stock Exchange in the period 1986 through 1989. Structures of returns volatility are estimated and forecasted. Models of autoregressive conditional heteroscedasticity (ARCH) are fitted to the stock returns. It is found that the returns series exhibit significant ARCH and GARCH effects with nonnormality. Based on fitted ARCH and GARCH models in the period 1986 through 1987 we forecast the volatility of returns in 1988 through 1989. The ARCH/GARCH forecast are compared with a benchmark value, a naive forecast and an exponential weighted moving average (EWMA) forecast. The results show that the EWMA method gives the best forecasts. These findings have implications in forecasting movements of market volatility, with applications to option pricing and control for variation margin risk in stock index futures. © 1991.