What Do Labor and Consumption Data Jointly Tell About Labor Income Risk?

This paper estimates a general stochastic process for labor income via indirect inference, by jointly using labor income data together with the information embedded in the dynamics of individual consumption. We extend earlier work in several directions. First, we do not restrict income shocks to follow a random walk, an assumption that has been made in previous studies that use consumption data to estimate income risk (Blundell and Preston (1998), Blundell, Pistaferri and Preston (2005) among others). Second, we use an auxiliary model for the indirect inference method that captures the rich dynamics of household consumption. To this end, we impute household consumption in PSID using the procedure developed in Blundell, Pistaferri and Preston (2005). Third, we estimate a general process that allows for heterogeneity in income growth rates. Thus, our approach allows us to bring both consumption and income data to distinguish between two alternative views of the income process: the random walk model (as in MaCurdy (1982), Abowd and Card (1989)) versus the profile heterogeneity model (as in Lillard and Weiss (1979), Baker (1997), Guvenen (2005))