Abstract:
|
Unit root test, which checks if a time series from macro-economics or management areas is indeed a random walk, is a significant part of forecasting the future of time series data. Various tests for unit root have been suggested so far but it is known that those tests suffer from size distortion and low power problems. In this article, we propose a new unit root test based on kernel density estimator. Under various models, we conducted a simulation study to compare our method with KPSS(Kwiatkowski, Phillips, Schmidt and Shin)(1992) test. Under the simulation setting suggested by KPSS (1992), we also compared our method with their test. The results indicate that our method is as good as or better than their test.
|