Nonlinearities and Heterogeneity in Firms Response to Aggregate Fluctuations: What Can We Learn From Machine Learning?

Abstract

Firms respond heterogeneously to aggregate fluctuations, yet standard linear models impose restrictive assumptions on firm sensitivities. Applying the Generalized Random Forest to U.S. firm-level data, we document strong nonlinearities in how firm characteristics shape responses to macroeconomic shocks. We show that nonlinearities significantly lower aggregate responses, leading linear models to overestimate the economy's sensitivity to shocks by up to 1.7 percentage points. We also find that larger firms, which carry disproportionate economic weight, exhibit lower sensitivities, leading to a median reduction in aggregate economic sensitivity of 52%. Our results highlight the importance of accounting for nonlinearities and firm heterogeneity when analyzing macroeconomic fluctuations and the transmission of aggregate shocks.