Structural relations, cointegration and identification: some simple results and their application

摘要: This paper presents and applies some results on the interpretation of cointegrating regressions. The key concept is irreducible (IC) relation, one from which no variable can be omitted without loss cointegration property. Extending earlier results, it shown that under certain circumstances, IC relations are identified structural forms. It possible, at least in principle, to learn about structure simultaneous long-run directly analyses, contrast with well-known fact such knowledge obtained correlations between stationary variables. also estimated by asymptotically mixed Gaussian median unbiased estimators, permitting standard inference. MINIMAL, an algorithm for extracting subsets a data set, applied variety artificial actual data.