Correcting microstructure comovement biases for integrated covariance

Finding a precise variance-covariance matrix is the building block of empirical finance. While microstructure-noise-robust methods for realized volatility are in the mainstream of financial econometrics, little if any attention has been devoted to estimating a noise-free realized covariance for overlooking the well-documented manifestation of commonality in market microstructure factors such as order flows, liquidity or herding. By documenting and recognizing this fact, we propose a microstructure-noise-free nonparametric covariance estimator to uncover the virtual integrated covariance. The estimator is easy to implement and performs admirably.