Does volume really matter? A risk management perspective using cross-country evidence

Abstract

This paper examines the impact of volume on conditional volatility and value at risk (VaR) in the context of mixture of distribution hypothesis (MDH). We test whether the support for or against the hypothesis is unconditional and holds true universally irrespective of the time period under study, the stock market under study, and the distributional assumptions so made on the residuals of the returns. We find that the persistence in volatility shows negligible reduction in all the indices across subperiods, thus refuting the claims of the MDH: that volume can explain the heteroscedasticity of returns. However, we do find that volume can act as a proxy for information post the sub?prime financial crisis, and it does impact VaR as the estimates improve significantly for some of these indices, which exhibit a strong correlation between volume and volatility.