A Short-Panel Identification of Lifetime Income Process
24 Pages Posted: 13 Jan 2016
Date Written: December 22, 2015
Without long panel structures, the persistence of individual income risks cannot be identified. In this paper, we derive bounds on age-profile persistence levels with short panels where there are only 2-year panel structures. These bounds remarkably capture a low persistence in the early ages and the gradual rise of persistence as an individual in the U.S. gets older. Using intentionally constructed short and long panels from Panel Study of Income Dynamics (PSID), we then verify how much the short-panel estimates might deviate from the long-panel estimates, provided that long-panel persistence levels are known. We find that despite its higher volatility, the short-panel estimated variance of permanent and transitory shocks mostly falls within the 95% confidence intervals of long-panel estimates and life-cycle patterns are quite in line.
Keywords: Short panel data, Idiosyncratic income risk, Persistence levels, Partial identification
JEL Classification: C33, D31, J31
Suggested Citation: Suggested Citation