Spuriousnonparametricregressionineconometrics.docVIP

Spurious nonparametric regression in econometrics* Li Songchen, Chen Haiyan College of Mathematics and Computational Science, Shenzhen University, Guangdong Shenzhen, 518060, China School of Management, Tianjin University, Tianjin, 300072,China Abstract This paper develops the asymptotic theory for the Nadaraya-Watson kernel estimator and local polynomial estimator when two independently integrated processes are used in a nonlinear regression. It is shown that the Nadaraya-Watson kernel estimator and the local polynomial estimator do not possess limiting distributions in this context but actually diverge at rate as the sample size , and this is slower than that of parameters in linear regression. In spite of the difference in the rate of divergence between the parametric and nonparametric cases, they all can induce spurious regression. Keywords: Nadaraya-Watson kernel estimation; Local polynomial estimation; Spurious regression; Integrated processes; Local time; Quadratic variation. Introduction The spurious regression was first studied by Granger and Newbold 1974 using simulation, and the simulation results are supported by Phillips’ 1986 theoretical analysis. Phillips proves that the usual test statistic in a spurious regression does not have a limiting distribution but diverges as the sample size approaches infinity. He also shows that has a nondegenerate limiting distribution while the statistic converges in probability to zero. These results has been generalized by Marmol 1998 , Tsay and Chung 2000 to cases with fractionally integrated processes. Most studies of the spurious regression concentrate on linear parametric regression and the asymptotic distributions of parameters and related statistics. By contrast, spurious regression in nonparametric cases and the asymptotic distributions for Nadaraya-Watson NW, henceforth kernel estimator and local polynomial estimator of nonlinear relationship of independently integrated processes are presently unde