Which vintage of data to use when there are multiple vintages of data?: Cointegration, weak exogeneity and common factors

Multiple vintages of data on the same variable are a pervasive feature of published data. This creates problems for empirical research, which requires a single observation series for each variable. This paper adopts an explicitly multivariate approach to solve this and related problems. Based on common factors and weak exogeneity, I show which vintage, or linear combination, of vintages should be used and how to test for cointegration with multiple vintages of data