演講者簡介 : Professor Hong received his Ph.D. in Economics from University of California at San Diego. He is currently the Ernest S. Liu Professor of Economics and International Studies in Department of Economics at Cornell University. He is also a distinguished research fellow at Academy of Mathematics and Systems Science and Center for Forecasting Science, Chinese Academy of Sciences (CAS). His research interests include model specification testing, nonlinear time series analysis, financial econometrics, and empirical studies on Chinese economy and financial markets.
演講摘要 : We propose Cramér-von Mises and Kolmogorov-Smirnov type tests for structural changes in linear time series models, possibly with endogenous regressors via a discrete Fourier transform (DFT) approach. If structural changes exist, the conventional OLS and 2SLS estimators are inconsistent for the unknown coefficients. Consequently, the estimated residuals contain the time-varying features of the model parameters. By DFT, we can capture such information in the frequency domain. The proposed tests are powerful against smooth structural changes and abrupt structural breaks and can detect a class of local alternatives at the parametric rate, which is asymptotically more efficient than the existing nonparametric tests. Simulation studies demonstrate the excellent finite sample performance of our tests. In an application to Taylor rules, we find significant evidence of structural changes during the post-1979 period, which is treated as a stable period by Clarida et al. (2000).
