Liyu Dou recently joins the school of management and economics as an Assistant Professor at the Chinese University of Hong Kong, Shenzhen, and graduated from Princeton University with a Ph.D in Economics. He is an econometrician specializing in time series econometrics, econometric theory and applied econometrics.          

His finished work tackles two key aspects of time series econometrics: Appropriate computation of standard errors, and modelling and measurement of persistence of macroeconomic time series. His ongoing (joint) projects span over a range of topics: Global misspecification analysis in moment condition models; Neural network estimation and inference in conditional moment condition models under the presence of an infinite dimensional parameter; Quantifying externatilities in airline networks.

This paper considers the problem of deriving heteroscedasticity and autocorrelation robust (HAR) inference about a scalar parameter of interest. I derive finite-sample optimal tests in the Gaussian location model, under nonparametric assumptions on the underlying spectral density. The optimal test trades off bias and variability, and requires an adjustment of the critical value to account for the maximum bias of the implied long-run variance estimator. I find that with an appropriate adjustment to the critical value, it is nearly optimal to use the so-called equal-weighted cosine (EWC) test, where the long-run variance is estimated by projections onto q type II cosines. The practical implications are an explicit link between the choice of q and assumptions on the underlying spectrum, as well as a corresponding adjustment to the usual Student-t critical value. Simulations show that the suggested new EWC test also performs well outside the Gaussian location model.