The sectoral channel: Not all uncertainty is created equal
Matteo Cacciatore and Giacomo Candian examine the sectoral and spatial dimension of uncertainty propagation through production networks. They construct industry-level uncertainty measures using option-implied volatility data from U.S. publicly traded firms at the NAICS 4-digit level, covering over 100 industries from 1995 to 2024. This forward-looking market-based measure captures investor expectations about sectoral risk.
Empirical and theoretical approach
They estimate dynamic responses using panel local projections that trace effects within affected industries and along supply chains, weighted by input-output linkages. A multi-sector New Keynesian model with production networks, nominal rigidities, and time-varying sectoral uncertainty rationalizes the findings.
Findings
- Asymmetric network propagation: Uncertainty’s effects depend critically on where it originates:
– Upstream uncertainty (chemicals, steel) acts like a supply shock: raises prices and reduces employment across the network
– Downstream uncertainty (automotive, insurance) resembles a demand shock: lowers both prices and employment - Transmission mechanisms: Within affected industries, uncertainty raises producer prices (precautionary pricing) while reducing employment. Uncertainty propagates heterogeneously: supplier industries experience falling prices and employment, while customer industries face rising prices and falling employment as input costs increase.
- Aggregate implications: Uncertainty shocks are contractionary, but inflation responses depend on sectoral origins. Upstream uncertainty is persistently inflationary, while downstream uncertainty is not.
Uncertainty through the Production Network: Sectoral Origins and Macroeconomic Implications
Authors: Matteo Cacciatore, Giacomo Candian
From: HEC Montréal, Bank of Canada
The amplification channel
Efrem Castelnuovo, Giovanni Pellegrino, and Laust L. Særkjær investigate the temporal and state-dependent dimension of uncertainty through the inflation cycle. The authors use U.S. monthly data from January 1962 to August 2024, including CPI inflation (yearly growth rate), industrial production (monthly growth rate), 3-month Treasury Bill rate, and 10-year Treasury Bill rate.
- Empirical framework: They employ a nonlinear stochastic volatility-in-mean threshold VAR proposed by Alessandri and Mumtaz (2019), where endogenous variables depend on their lags and are directly affected by common stochastic volatility, a proxy for macroeconomic uncertainty. The model switches between two regimes based on whether lagged CPI inflation exceeds a threshold of approximately 7% (estimated) or 6% (imposed robustness check). The high-inflation regime covers 12% of observations, primarily the mid-to-late 1970s/early 1980s and 2022.
- Theoretical framework: The authors develop a nonlinear New Keynesian DSGE model with Calvo price-setting, approximated around positive trend inflation rates (0%, 2%, 6%), and featuring time-varying second-moment TFP shocks. The model is solved using third-order pruned perturbations.
Findings
State-dependent effects:
Uncertainty shocks have dramatically asymmetric effects depending on the inflation regime:
- High inflation regime (>6%): Peak inflation response of 0.37% , much larger than in low inflation regime. In addition, there is a significant industrial production decline lasting 18 months, and a policy rate increase of 60 basis points.
- Low inflation regime (<6%): Moderate inflation increase of 0.16% at peak, while there is a muted real activity response, and essentially no policy rate reaction.
The inflation-uncertainty amplifier
The theoretical model reveals a novel mechanism: the interaction between firms’ upward pricing bias (due to nominal rigidities and convex marginal revenues) and trend inflation creates amplification. When trend inflation is high, re-optimizing firms set even higher prices because their relative prices erode faster under price stickiness, magnifying the precautionary pricing response to uncertainty.
Variance decomposition
In high-inflation states, uncertainty shocks explain approximately 20% of inflation volatility, 25% of long-term interest rate volatility, 60% of policy rate volatility, and 10-11.5% of industrial production volatility. Contributions are substantially smaller in low-inflation regimes.
The Inflation Uncertainty Amplifier
Authors: Efrem Castelnuovo, Giovanni Pellegrino, Laust L. Særkjær
From: University of Padova, Aarhus University
