EXAMINING PREDICTABILITY IN THE DISTRIBUTION OF STOCK MARKET RETURNS USING A QUANTILE-BASED APPROACH: A STUDY ON ASEAN STOCK MARKET INTEGRATION

Abstract

This paper conducts a nonparametric Granger causality test in quantiles to investigate the interdependence of ASEAN stock markets. We find ASEAN stock markets have significant bilateral Granger causality in many of the considered quantiles. We find stronger evidence of Granger causality in the tail quantiles, and this result is more pronounced after adjustment for the common factor. The paper also investigates whether evidence of Granger causality in quantiles imply better forecastability of the features of the stock return distribution, using Thailand daily returns. We find better predictability mainly in the lower quantiles, a result similar to the Granger causality test after accounting for the common factor. However, there are many instances where a significant Granger causality result in the quantile did not translate to better predictive ability. Nonetheless, we find that the point forecasts made from an equal-weighted combination of the different country forecasts performed better than the prevailing median.