A coupled component DCS-EGARCH model for intraday and overnight volatility

Abstract

We propose a semi-parametric coupled component exponential GARCH model for intraday and overnight returns that allows the two series to have different dynamical properties. We adopt a dynamic conditional score model with t-distributed innovations that captures the very heavy tails of overnight returns. We propose a several-step estimation procedure that captures the nonparametric slowly moving components by kernel estimation and the dynamic parameters by maximum likelihood. We establish the consistency, asymptotic normality, and semiparametric efficiency of our semiparametric estimation procedures. We extend the modelling to the multivariate case where we allow time varying correlation between stocks. We apply our model to the study of Dow Jones industrial average component stocks and CRSP size-based portfolios over the period 1993-2017. We show that the ratio of overnight to intraday volatility has actually increased in importance for Dow Jones stocks during the last two decades. This ratio has also increased for large stocks in the CRSP database, but decreased for small stocks in CRSP.

Observation-driven models for realized variances and overnight returns applied to Value-at-Risk and Expected Shortfall forecasting

“…This paper deviates from previous approaches in that we no longer treat the proportionality parameter c as a static parameter. This is in line with findings of Ahoniemi et al (2015), Linton and Wu (2020), among others, who show that the daytime and overnight volatility processes have their own dynamics, and that, hence, the ratio c might not be constant over time. Instead of a static c, we propose a dynamic parameter c t that we endow with the score-driven (GAS) dynamics of Creal et al (2013); see also Harvey (2013).…”

“…For many stocks, our model improves significantly upon the HEAVY model of Shephard and Sheppard (2010) or a model with a static volatility ratio. We also confirm Linton and Wu (2020) by showing that modeling the overnight volatility separately leads to low values of the degrees of freedom parameter, with subsequent potential problems for the existence of sufficient moments (see also Berkman et al, 2012).…”

“…We allow the ratio of open-toclose volatility to close-to-close volatility to vary over time; compare Hansen and Lunde (2005). Considering a time-varying ratio of volatilities is also in line with the work of Ahoniemi et al (2015) and Linton and Wu (2020). They show that daily and overnight volatility can be distinct processes with their own dynamics and, thus, with a time-varying ratio between the two.…”

Estimating Risk in Illiquid Markets: A Model of Market Friction with Stochastic Volatility

“…The non-linear update in Equation ( 11) is robust to fat-tails since it downweights volatility estimates when extreme price movements occur. For this reason, it has been extensively applied in the finance and economic literature; see also Creal et al (2011), Linton and Wu (2020). The score of a zero return in the Beta-t-EGARCH reduces to u t = −1.…”

Overnight volatility, realized volatility, and option pricing

“…Blanc et al (2014) applied a quadratic ARCH framework and found that the intraday return and overnight return behave completely differently as expected. Linton and Wu (2020) proposed a semiparametric coupled component exponential GARCH model for intraday and overnight returns, with t distribution innovation that captures the fat tail of overnight return.…”