What are we learning about the long-run?

An attempt is made to link together earlier definitions of the long-run found in micro and macro economics with recent developments in econometrics; specifically cointegration. It is suggested that the links are not strong and that most of the previous work in econometric theory has been unnecessarily over-precise. Unit root processes can be replaced by processes that approximate them without loss of interpretation. The possibility of embedding cointegration theory into a very general non linear theory is suggested. An example uses a nonIinear relationship between UK short and long run interest rate proposed by Frank Paish