Purpose
– The aim of this paper is to measure and evaluate the relationship between returns-volatility and trading volume and returns and volatility of financial market indexes using time-varying copulas.
Design/methodology/approach
– The time dynamic dependence parameter is allowed to evolve according to a restricted ARMA-type equation which includes a constant term that is driven by a hidden two-state first-order Markov chain.
Findings
– In using this time dynamics in conjunction with non-elliptical distribution functions and tail dependence measure, the authors are allowing for (and focusing on) non-linearities in the returns-volume-volatility relationship. The results support the assumption that current trading volume provides information about future volatility as well as that there is a negative relationship between returns and their volatilities in financial market indexes.
Originality/value
– The authors provide an interesting empirical interpretation for the regimes the authors have identified: in the high dependence regime the sequential information arrival hypothesis and/or noise trading hypothesis are valid, consequently future volatility prediction is possible and persistent but does not last indefinitely; in the low dependence regime, the future volatility prediction is more unlikely to occur, since both trading volume and return negatives have a low (near zero) relation with future volatility.
