Integrated Covariance Estimation using High-frequency Data in the Presence of Noise

Pre-averaging estimators of the ex-post covariance matrix in noisy diffusion models with non-synchronous data

“…More recently, Zhang (2008) extended the two-scale RV to integrated covariance estimation in the simultaneous presence of noise and non-synchronicity, while in concurrent and independent work (Barndorff-Nielsen et al, 2008b) proposed a multivariate realised kernel. Additional work in this growing line of research includes Malliavin and Mancino (2002), Martens (2003), Renò (2003), Bandi and Russell (2005), Griffin and Oomen (2006), Large (2007), Voev and Lunde (2007) and Boudt et al (2008), among others.…”

Correcting microstructure comovement biases for integrated covariance

“…Recent academic work on realized covariance has focused mainly on the issue of asynchronous trading (Hayashi and Yoshida, 2005;Bandi and Russell, 2007;Voev and Lunde, 2007;Yeh et al, 2007;Barndorff-Nielsen et al, 2008;Zhang, in press;Griffin and Oomen, in press). Nonetheless, relative to the extensive literature on the bias-correction for univariate RV and adjustment for the effect of asynchronous trading, comparatively little has been written on the biases in the covariance attributable to commonality in microstructure noise.…”

“…Moreover, when more than one asset is involved, the effect of non-synchronous trading among financial assets induces a severe bias towards zero in the covariance constructed from high-frequency data, and some recent treatments of the effect on realized covariances can be found in Hayashi and Yoshida (2005), Bandi and Russell (2007), Voev and Lunde (2007), Yeh et al (2007), Barndorff-Nielsen et al (2008), Zhang (in press) and Griffin and Oomen (in press) among others. Nonetheless, to our knowledge, the contemporaneous comovement in market microstructure noises has been overlooked in constructing realized covariance in the literature.…”

Estimating integrated co-volatility with partially miss-ordered high frequency data

“…Quantitively, with the wide availability of high frequency data, in recent years, the estimate methodologies of volatility, or quadratic covariation have been improved substantially. Many papers document the progress of the volatility estimation, see Fama and French (1992), Voev and Lunde (2007), Zhang (2011a), Aït-Sahalia et al (2010), Bandi and Russell (2005), among others. However, the coin has two sides.…”

“…Opposite to the blowup of realized covariance under noise, the asynchronicity between assets causes realized covariance to be biased towards zero as the sampling frequency increases, the figure has been documented in Epps (1979), and hence is called Epps effect. In literature, several methods can explicitly fixed the effect of asynchronicity in high frequency data, we refer to Hayashi and Yoshida (2005), Hayashi and Yoshida (2008), Zhang (2011a), Voev and Lunde (2007), Jing et al (2013), Christensen et al (2010), among others.…”