Analysis of Upstream, Downstream, and Common Firm Shocks Using a Large Factor-Augmented Vector Autoregressive Approach

Abstract

We evaluate the roles of upstream (supplier-to-user), downstream (user-to-supplier), and common factor shock transmission across firms by deriving interfirm networks and common factors from US equities over 1989–2017. We overcome the econometric challenges of estimating the large factor-augmented vector autoregressive (FAVAR) system by developing two alternative approaches: one prioritizing computational efficiency and the other providing the full posterior distribution of all model parameters and factors. We find that (i) common factors drive an increasing variance share of returns, (ii) supplier shocks are more evident in equity price movements than downstream exposures, and (iii) removing the impact of common factors is increasingly important for revealing interfirm connections.