Monetary Policy Through Production Networks: Evidence from the Stock Market 
A central endeavor in macroeconomic research seeks to understand how shocks are transmitted throughout the economy. Recent findings suggest that microeconomic shocks—those that affect a single firm, industry, or related group—spread through the larger economy and generate aggregate fluctuations via the production network (the economy’s input-output structure). However, there is little evidence about how much of the overall response to macroeconomic shocks can be attributed to network effects, since these effects are difficult to measure. By treating monetary policy changes as demand shocks, and tracing how stock prices are affected during a narrow time window centered around FOMC policy announcements, the authors quantify how a macroeconomic shock is disseminated throughout the real economy via input-output linkages. They employ spatial autoregressions to decompose the overall monetary shock into a direct effect and a network effect.
Key Findings
- Large network effects in the US economy account for 50 to 85 percent of the overall impact that monetary policy shocks have on stock prices. These results are robust to controlling for different sample periods, different event windows, various types of policy announcements, industry-demeaned returns, and other common shocks occurring in the same event window.
- The measured direct effects of monetary policy are larger for industries that sell most of their output directly to end-consumers. The greater importance of direct effects for these industries is consistent with the intuition that monetary policy may directly increase consumer demand for goods and services, which then gets transmitted to these firm’s suppliers and propagates further upstream in the production network.
- Expansionary monetary policy shocks not only directly increase the demand for the goods of firms selling to end-consumers, but also lead indirectly to higher-order demand effects through the increased demand for intermediate inputs. These higher-order effects can rationalize the large and cross-sectionally heterogeneous effects that monetary policy shocks have on stock market prices.
- The authors also find that indirect network effects have an impact on sales and operating income. The indirect response increases up to seven quarters after the monetary policy shock, but then loses statistical significance after eight quarters.
- The network effects documented in firm and industry fundamentals indicate that monetary policy shocks affect the real economy, at least in part, through cash-flow effects, a result that is consistent with findings in Bernanke and Kuttner (2005) and Weber (2015).
Implications
The evidence that the US input-output structure experiences large network effects suggests that policymakers should consider incorporating production networks into their models in order to account for the significant role that these effects appear to play in transmitting monetary policy. These findings also open up novel avenues to develop asset-pricing theories based on the economy’s input-output network/structure.
Abstract
Monetary policy shocks have a large impact on stock prices during narrow time windows centered around press releases by the FOMC. We use spatial autoregressions to decompose the overall effect of monetary policy shocks into a direct effect and a network effect. We attribute 50 to 85 percent of the overall impact to network effects. The decomposition is a robust feature of the data, and we confirm large network effects in realized cash-flow fundamentals. A simple model with intermediate inputs allows a structural interpretation of our empirical strategy. Our findings indicate that production networks might be an important mechanism for transmitting monetary policy to the real economy.