Bayesian state-space modeling for analyzing heterogeneous network effects of US monetary policy

TL;DR

This paper introduces a Bayesian state-space model to analyze how US monetary policy effects vary across industries and over time, accounting for dynamic network structures and heterogeneity in responses.

Contribution

It extends existing econometric models by incorporating time-varying industry responses and network structures, capturing non-linearities and heterogeneity in monetary policy transmission.

Findings

Impacts of monetary policy vary significantly over time and industries.

Higher-order effects are prominent during periods of financial stress.

Industry responses relate to proximity to end-consumers.

Abstract

Understanding disaggregate channels in the transmission of monetary policy is of crucial importance for effectively implementing policy measures. We extend the empirical econometric literature on the role of production networks in the propagation of shocks along two dimensions. First, we allow for industry-specific responses that vary over time, reflecting non-linearities and cross-sectional heterogeneities in direct transmission channels. Second, we allow for time-varying network structures and dependence. This feature captures both variation in the structure of the production network, but also differences in cross-industry demand elasticities. We find that impacts vary substantially over time and the cross-section. Higher-order effects appear to be particularly important in periods of economic and financial uncertainty, often coinciding with tight credit market conditions and financial stress. Differentials in industry-specific responses can be explained by how close the respective industries are to end-consumers.