Working Paper: NBER ID: w31337
Authors: Nir Jaimovich; Stephen J. Terry; Nicolas Vincent
Abstract: Heterogeneous firm models are ubiquitous in modern macroeconomics. We revisit a central feature of these models: the idiosyncratic shock process faced by firms. Using a large representative firm-level dataset, we document nonparametrically that the common assumption, a Gaussian AR(1) shock process, is at odds in important ways with observed fat-tailed firm dynamics. We embed these findings within a standard quantitative general equilibrium heterogeneous firm dynamics model and show that the nature of firm-level shocks has a sizable quantitative effect on the economy’s responsiveness to aggregate shifts.
Keywords: firm dynamics; macroeconomic implications; idiosyncratic shocks; Gaussian AR(1); quantitative general equilibrium
JEL Codes: E0
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| Gaussian AR(1) assumption for idiosyncratic shocks (C22) | firm-level distributions that diverge from empirical observations (D39) |
| empirical data shows a leptokurtic distribution of revenue growth (D39) | incompatible with Gaussian AR(1) assumption (C22) |
| firm values are clustered at lower revenue levels (L25) | higher likelihood of exit (J63) |
| stochastic structure of shocks (C69) | macroeconomic outcomes (E66) |
| nonparametric model (C52) | higher sensitivity of exit rates to economic changes (J63) |
| Gaussian AR(1) model (C22) | misestimations in policy impacts and economic predictions (D78) |