Heterogeneous effects of economic cycles across the income distribution: A common factor model approach

Abstract

We extend the work of Bitler and Hoynes (2015), which uses state-level panel data and a two-way fixed effects (TWFE) model to examine the heterogeneous effects of economic cycles across the household income distribution. Using a more flexible common factor model that allows common shocks to have differential effects across states, we find statistically and economically significant differences from TWFE estimates, particularly at the lower tail of the income distribution. While TWFE estimates suggest a hump-shaped pattern — where effects are smaller at the lower and upper tails — our estimates indicate that cyclical impacts are largest at the lowest income levels and decline monotonically as income rises. These results suggest that conventional TWFE models may understate the vulnerability of lowest-income households to economic downturns.