Causality between trading volume and returns: Evidence from quantile regressions

We analyse the causality between past trading volume and index returns in the Pacific Basin countries. OLS results indicate no causal link between volume and returns. However, the quantile regression method reveals strong nonlinear causality: positive for high return quantiles and negative for low ones. Causality in quantiles is not a statistical artefact of causality in periods of high volatility, i.e., causality does not occur in a clustered manner. Causality in quantiles helps to explain the lack of causality between volume and raw returns on the one hand and a strong causal relationship between volume and return volatility on the other.