Working Paper: CEPR ID: DP15614
Authors: Yuriy Gorodnichenko; Lilia Maliar; Serguei Maliar; Christopher Naubert
Abstract: We study a heterogeneous-agent model with sticky-prices in which total factor productivity and individual productivity are subject to stochastic volatility shocks. Agents save through liquid bonds and illiquid capital and shares. To construct equilibrium, we use a deep learning algorithm. Our method preserves non-linearities, which is essential for understanding portfolio choices. With rich heterogeneity at the household level, we are able to quantify the impact of uncertainty across the income and wealth distribution. We find that persistent high levels of uncertainty increase wealth inequality, and that in response to a contractionary monetary policy shock, illiquid wealth inequality decreases and liquid wealth inequality increases
Keywords: Machine Learning; Deep Learning; Neural Network; HANK; Heterogeneous Agents
JEL Codes: E21; E31; E44
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| persistent high levels of uncertainty (D80) | increase wealth inequality (D31) |
| innovations to aggregate uncertainty (D89) | decrease wealth inequality (D31) |
| contractionary monetary policy shock (E49) | decrease illiquid wealth inequality (D31) |
| contractionary monetary policy shock (E49) | increase liquid wealth inequality (D31) |