Non‐trading day effects in asymmetric conditional and stochastic volatility models

Abstract

Summary It is well known that non-trading days (or holidays) can have significant effectson the returns in financial series. In this paper, we analyze three models of non-trading dayeffects in stochastic volatility models with leverage effects, namely (i) the approach basedon the dummy variable in conditional volatility models; (ii) the approach based on a discretetime approximation of a continuous time stochastic volatility model and (iii) the twin non-trading day stochastic volatility model which nests the above two models. The three modelsare also estimated and tested within the asymmetric and exponential conditional volatilityframeworks. All the models within the stochastic, asymmetric conditional and exponentialconditional volatility frameworks are estimated and compared using a selection of financialreturns series. Key words: Symmetric and asymmetric conditional volatility , Exponential conditionalvolatility , Stochastic volatility , Non-trading day , Non-nested , Monte Carlo likelihood