Article ID Journal Published Year Pages File Type
5095984 Journal of Econometrics 2017 14 Pages PDF
Abstract
The data generating process (DGP) for generic dynamic panel data consists of a law of state dynamics g, a selection or attrition rule h, and an initial condition F. I study nonparametric identifiability of this complete DGP (g,h,F) using short unbalanced panel data, allowing for nonseparability between observed states and unobserved heterogeneity in each of g, h and F. For T⩾3, the DGP is identified by using a proxy variable. For T⩾6, the three additional periods construct a proxy, and thus the DGP is identified without an auxiliary variable.
Related Topics
Physical Sciences and Engineering Mathematics Statistics and Probability
Authors
,