We propose new cointegration tests using signs of the regressors as instrumental variable. Our tests have the asymptotic standard normal distribution and are free from the dimension of regressors under the null hypothesis of no cointegration. A Monte-Carlo simulation shows that the proposed tests have a stable size and an improved power. Particulary, the tests have better power for small numbers of observations.
Cointegration;test statistic;sign;instrumental variable;standard normal distribution;
Dickey, D. A. and Fuller, W. A. (1979). Distribution of the estimators for autoregressive time series with a unit root, Journal of the American Statistician Association, 74, 427-431
Engle, R. F. and Granger, C. W. J. (1987). Co-integration and Error correction: Representation, estimation, and testing, Econometrica, 55, 251-276
Kremers, J. J. M., Ericsson, N. R. and Dolado, J. J. (1992). The power of cointegration tests, Oxford Bulletin of Economics and Statistics, 54, 325-348
So, B. S. and Shin, D. W. (1999a). Cauchy Estimators for autoregressive processes with applications to unit root tests and confidence intervals, Econometric Theory, 15, 165-176
So, B. S. and Shin, D. W. (1999b). Recursive mean adjustment in time series inferences, Statistics and Probability Letters, 43, 65-73