On the use of cross-sectional measures of uncertainty

In this paper, we provide a coherent theoretical investigation of the relationship between cross-section and time series measures of uncertainty, which are often employed as perfect substitutes in empirical applications. The main Önding of our analysis is that there exists an ambiguous sign in the discrepancy between the two measures of uncertainty arising from the presence of cross-sectional dependence amongst individuals. Thus our study underpins the importance of accounting for cross-sectional dependence, in line with recent inferential theory of panel data models.