Asymmetric Volatility and Cross Correlations in Stock Returns under Risk and Uncertainty

Abstract

Capital market efficiency is a matter of great interest for policy makers and investors in designing investment strategy. If efficient market hypothesis (EMH) holds true, it will prevent the investors to realize extra return by utilizing the inherent information of stocks. They will realize extra returns only by incorporating the extra risky stocks in their portfolios. While empirical tests of EMH and risk-return relationship are plentiful for developed stock markets, the focus on emerging stock markets like India, Pakistan, Sri Lanka, etc., began with the liberalization of financial systems in these markets. With globalization and deregulation, the enormous opportunities of investment in South Asian stock markets have attracted the domestic and foreign institutional investors in general, and to reduce their portfolio risk by diversifying their funds across the markets in particular. The efforts are made in this study to examine the cross-correlation in stock returns of South Asian stock markets, their regional integration, and interdependence on global stock market. The study also examines the important aspects of investment strategy when investment decisions are made under risk and uncertainty. The study uses Bombay stock exchange listed index BSE 100 for India, Colombo stock exchange listed Milanka Price Index for Sri Lanka, Karachi stock exchange listed KSE 100 for Pakistan, Dhaka stock exchange listed DSE-General Index for Bangladesh, and S & P Global 1200 to represent the global market. It carries out a comprehensive analysis, tracing the autocorrelation in stock returns, cross correlations in stock returns under risk and uncertainty, interdependency among the South Asian stock markets, and that with the global stock market. The research methodology applied in the study includes application of Ljung-Box to examine the cross-correlation in stock returns, ARCH and its generalized models for the estimation of conditional and asymmetric volatilities, and Ljung-Box as a diagnostic testing of fitted models, and finally correlation to examine the interdependency of these markets in terms of stock returns and expected volatility. The results bring out the following: L-B statistics suggests the presence of autocorrelation in stock returns in all Asian stock markets; however, for the global market, autocorrelations are significant at 15 lags, and thereafter they are insignificant. The significant autocorrelations in stock returns report volatility clustering in stock returns, reject the EMH, and hold that current stock returns are significantly affected by returns being offered in the past. ARCH and its generalized models significantly explain the conditional volatility in all stock markets in question. The study rejects the relationship between stock returns and expected volatility; however, the relationship is significant with unexpected volatility. It brings out that investors adjust their risk premium for expected variations in stock prices, but they expect extra risk premium for unexpected variations. With their entry into the liberalization phase, South Asian stock markets have reported regional interdependence, and also interdependence with the global stock market.