Comparison of Residual based Cointegration Tests: Evidence from Monte Carlo

Abstract

In this article ten cointegration tests based on residuals of cointegrating equation are compared on basis of stringency criterion: a robust technique for comparison of tests using Monte Carlo simulations. Two tests i.e. Phillips and Ouliaris’ Pu and Choi Durbin-Hausman statistic are the leading performers and are recommended for any sample size. The remaining eight tests are recommended for only large sample sizes of 200 or greater. The use of all these ten tests is not recommended when presence of both intercept and linear time trend is assumed in cointegrating equation unless the sample size is very large i.e. greater than 200.