Cross country comparison and performance of model selection techniques for bounds test of cointegration: Evidence from brıcs countries

Abstract

Only unstructured single path model selection techniques i.e. Information Criteria are used by Bounds test for plausible model selection. This paper evaluates the performance of these five routinely used information criteria (AIC, AICC, BIC, BICC and HQC) along with three structured single path model selection approaches (Forward Selection, Backward Elimination and Stepwise) to be used in Bounds test for plausible model selection by examining the long-run relationship between three pairs of macroeconomic variables i.e. (Energy Consumption (EC) and GDP), (Oil Price (OP) and GDP) and (Broad Money (BM) and GDP) for BRICS countries. In addition, the existence of long run relationship for the three pairs is also assessed and compared across five BRICS economies. It is found that BIC and Backward Elimination procedures are better. Moreover, significant evidence of level relationship between (EC and GDP) and (OP and GDP) is found for all five economies except Brazil..