A Test for Dependence and Covariance Estimator of Market Microstructure Noise

There are many approaches for estimating an integrated variance and covariance in the presence of market microstructure noise. It is important to know a dependence of noise to construct the integrated variance and covariance estimators. We study a time dependence of bivariate noise processes in this paper. We propose a test statistic for the dependence of the noises and an autocovariance estimator of the noises and derive its asymptotic distribution. The asymptotic distribution of the autocovariance estimator provides us to another test statistic which is for significance of the autocovariances and for detection whether the noise exists or not. We obtain good performances of the test statistics and autocovariance estimator of the noises in a finite sample through Monte Carlo simulation. In empirical illustration, we confirm that the proposed statistics and estimators capture various dependence patterns of the market microstructure noises.