Abstract: In this paper,through extending the existing Johansen cointegration test model,allowing coefficients in the cointegrating regression model having time varying characteristics,and using orthogonal Chebyshev time polynomial to model time varying coefficient,we propose a time varying error correction model which can capture the smoothing time transition,and the maximum likelihood method is used to estimate the time varying error correction model.A likelihood ratio test is constructed to test the Johansen cointegration,and its asymptotic distribution is deduced,gradually being consistent with the chi square distribution.At last,we use the proposed time varying coefficient cointegration regression model to test the purchasing power parity of RMB.The results show that the RMB against the U.S.dollar nominal exchange rate,domestic consumer price index and the American consumer price index have time varying cointegration relationship,but the sign of the coefficient is not consistent with the theoretical expectation.

Key Words: Time Varying; Cointegration Model; Chebyshev Polynomials; RMB Exchange Rate Purchasing Power Parity