Upstream, Downstream & Common Firm Shocks

We develop a multi-sector DSGE model to calculate upstream and downstream industry exposure networks from U.S. input-output tables and test the relative importance of shocks from each direction by comparing these with estimated networks of firms’ equity return responses to one another. The correlations between the upstream exposure and equity return networks are large and statistically significant, while the downstream exposure networks have lower — but still positive — correlations that are not statistically significant. These results suggest a low short-term elasticity of substitution across inputs transmitting shocks from suppliers, but more flexible ties with downstream firms. Additionally, both the DSGE model and simulations of our empirical approach highlight the importance of accounting for common factors in network estimation, which become more important over our 1989-2017 sample period, explaining 11.7% of equity return variation over the first ten years and 35.0% over the final ten.