Sankhya: The Indian Journal of Statistics

2002, Volume 64, Series B, Pt. 2, 214--233

LM TEST FOR THE CONSTANCY OF REGRESSION COEFFICIENT WITH MOVING AVERAGE INNOVATION

By

MEIHUI GUO

and

C.C. SHEN, National Sun Yat-Sen University, Taiwan

SUMMARY. Assume that the time series can be decomposed into the sum of a deterministic trend, a random walk and a MA(1) innovation. We are interested in testing the constancy of regression coefficient against the random walk alternative. In this paper, we derive the exact Lagrange multiplier test for the above hypotheses and proposed an asymptotically equivalent test statistic. Furthermore, the null and alternative limiting distributions of the test statistic are derived. Finally, we compare the finite sample size and power of the proposed test statistic with that of Kwiatkowski et al. (1992). An example of testing the global warming hypothesis is also demonstrated.

*AMS (1991) subject classification}. *Primary 62H15, 62P20; secondary 62M10.

*Key words and phrases. *Lagrange multiplier test, moving average*.*