• 内页大图

    内页大图

Testing Instability in A Predictive Regression Model with Nonstationary Regressors

TESTING INSTABILITY IN A PREDICTIVE REGRESSION MODEL WITH NONSTATIONARY REGRESSORS

Zongwu Caia1 c1, Yunfei Wanga2 and Yonggang Wanga3

 

a1 University of Kansas and Xiamen University

a2 Sun Trust Bank

a3 First Tennessee Bank

 

Abstract

 

It is well known that allowing the coefficients to be time-varying in a predictive model with possibly nonstationary regressors can help to deal with instability in predictability associated with linear predictive models. In this paper, an L2-type test statistic is proposed to test the stability of the coefficient vector, and the asymptotic distributions of the proposed test statistic are developed under both null and alternative hypotheses. A Monte Carlo experiment is conducted to evaluate the finite sample performance of the proposed test statistic and an empirical example is examined to demonstrate the practical application of the proposed testing method.

 

Correspondence

c1 Address correspondence to Zongwu Cai, Department of Economics, University of Kansas, Lawrence, KS 66045, USA; e-mail: zongwucai@gmail.com.

Footnotes

  We thank the editor, Professor Peter C.B. Phillips, and the anonymous referees for their insightful comments that greatly improved our paper. Cai’s research was supported, in part, by the National Nature Science Foundation of China grant #71131008 (Key Project).