CIRJE-F-431. Hirukawa, Masayuki, "A Two-Stage Plug-In Bandwidth Selection and Its Implementation for Covariance Estimation", June 2006.

To improve two existing bandwidth choice rules for kernel HAC estimation by Andrews (1991) and Newey and West (1994), this paper proposes to estimate an unknown quantity in the optimal bandwidth (called the normalized curvature) with a general class of kernels and derives the bandwidth that minimizes the asymptotic mean squared error of this estimator. The theory of the two-stage plug-in bandwidth selection and a reliable implementation method are developed. It is shown that the optimal bandwidth for the kernel-smoothed normalized curvature estimator should diverge at a slower rate than the one for the HAC estimator with the same kernel. Finite sample performances of the new HAC estimator are assessed through Monte Carlo simulations.