Cointegration and Alternative Specifications

Abstract

The particular topics considered in this chapter can be viewed as being at least in part determined by this study’s stated focus on misspecification tests in the context of Ordinary Least Squares. However, the chapter’s subject matter can also be seen to reflect the way in which econometric theory has developed during the past 30 years. In important respects, there has been a broadening in the range of econometric applications, epitomized by such things as the development of microeconometrics so called, the greater use of survey and panel data in economic research – often in the form of microdata – and other such expansions in the application of econometric techniques, which have not always been limited solely to economics and economists. But notwithstanding this “rippling out,” involving a coincident growing interest in such things as “non-standard” dependent variables, cross-section techniques, and “spatial econometrics” – which has certainly affected the development of econometric software and would be evident from any more general survey that specifically attempted to probe these and related aspects – the dominant themes in the development of econometric theory, or at least the most obvious areas of that development, have continued to be associated with the properties of time series, what can be called specification search, and the characterization of the disturbance term and its possible distribution, jointly and severally.