Nonparametric Estimation and Specification Testing in Nonstationary Time Series Models

Abstract

In this paper, we consider both estimation and testing problems in a nonlinear time series model with nonstationarity. A nonparametric estimation method is proposed to estimate a sequence of nonparametric “distance functions”. We also propose a test statistic to test whether the regression function is of a known parametric nonlinear form. The power function of the proposed nonparametric test is studied and an asymptotic distribution of the test statistic is shown to depend on the asymptotic behavior of the “distance function” under a sequence of general semiparametric local alternatives. The asymptotic theory developed in this paper differs from existing work on nonparametric estimation and specification testing in the stationary time series case. In order to implement the proposed test in practice, a computer-intensive bootstrap simulation procedure is proposed and asymptotic approximations for both the size and power functions are established. Furthermore, the bandwidth involved in the test statistic is selected by maximizing the power function while the size function is controlled by a significance level. Meanwhile, both simulated and real data examples are provided to illustrate the proposed approach.