The Exogeneity Question(s)

Abstract

Abstract Exogeneity assumptions are critical to applied time series analysis, but the topic is under-discussed and under-explained in econometric texts and often assumed away in applied work. The central misunderstanding is that there is not an exogeneity assumption; there are many exogeneity assumptions. Weak exogeneity, strong exogeneity, and super exogeneity are more or less likely to hold in different applications and the type of exogeneity that can be assumed tells us what we can learn from a given model. This chapter explains the importance of exogeneity in applied time series analysis, defines the different exogeneity concepts, highlights the different forms of time series analysis they are relevant for, and provides a strategy for evaluating the different exogeneity assumptions. Theory and data must support the exogeneity assumption necessary given the purpose of an analysis and analysts must be transparent in the claims they make about exogeneity.