Functional-Coefficient Quantile Cointegrating Model And Its Application In PPP

As most economic variables in the macro economy are non-stationary variables(Nelson and Plosser,1982),and the modeling of non-stationary time series can easily lead to “spurious regression” problems(Granger and Newbold,1974).Therefore,since Engel and Granger(1987)proposed the cointegration theory,it has gradually developed into one of the most important research tools in time series analysis.Moreover,it has always been the focus of time series analysis.In this paper,a new co-integration model,the Functional-Coefficient Quantile Cointegrating Model is established,which makes up the deficiency of the cointegration theory in the literature,and further advances its development.Compared with the existing co-integration models,most of which can only depict simple relationships between variables,the model proposed in this paper can capture the non-linear,time-varying and asymmetrical relationships between variables at the same time.Therefore,it can better capture and depict the co-movements of different economic variables in a rather complex economic environment.At the same time,our model is also robust to non-Gaussian condition and could avoid the "curse of dimensionality" problem widely encountered in nonparametric estimation.Secondly,this paper systematically studies the issues of estimation,asymptotic theory and statistical inference of the model.While local polynominal method is employed to estimate the model,Cross-Validation is utilized in choosing the optimal bandwidth.The convergence rate of the estimators is shown to possess the characteristics of both non-stationary time series analysis and nonparametric regression analysis.Additionally,similar to E-G cointegration test method,a two-step method is also proposed to test cointegration.Finally,this paper uses this new proposed model to test purchasing power parity(PPP)between three east Asia's major countries: China,Japan and South Korea and the United States.Compared with the traditional unit root test and co-integration test results,the test results based on the newly proposed method are more likely to support the PPP theory in the above countries.