In this note, we review two recent academic papers and we carry out some simulations using our models.
The first paper (“What drives core inflation? The role of supply shocks” by ECB economist Marta Banbura and coauthors) estimates the contributions of supply shocks to Euro area inflation. The results confirm previous findings by Ascari, Trezzi and coauthors (2022).
The second paper (“How do supply shocks to inflation generalize? Evidence from the pandemic era in Europe” by Viral Acharya (NYU St) and coauthors) provides a detailed study of the transmission of supply constraints to inflation through household inflation expectations and firms’ pricing power.
In part II of this note, we will carry out some simulations using our models to assess the possible amount of supply-driven disinflation ahead of us in the coming months.
Paper #1: What drives core inflation? The role of supply shocks
A new paper (“What drives core inflation? The role of supply shocks”) by ECB economists Marta Banbura, Elena Bobeica, and Catalina Martínez Hernández studies the drivers of euro area core inflation in a medium-scale vector autoregression model. The authors distinguish between several types of supply shocks and find that these shocks explain about half of the increase in core inflation in the Euro area. The results of Banbura et al. (2023) confirm previous findings by Ascari, Bonomolo, Hoeberichts, and Trezzi (2022).
What the paper does
The authors estimate a Bayesian VAR model of 17 variables for the euro area economy and identify 8 shocks driving these variables. In particular, the authors consider the following shocks: domestic demand, foreign demand, oil supply, oil-specific demand, gas price, global supply chain, domestic supply chain, labour market. The shocks are identified using a combination of sign and zero restrictions. For example, both oil supply and oil-specific demand shocks increase HICP inflation and core inflation, but the supply shock decreases oil production while the oil-specific demand shock raises oil production.
Main results
The paper provides a historical decomposition of core inflation into its structural drivers. Since the pandemic, core inflation has been strongly affected by supply-related factors (see Figure 1, Panel A). Shocks to global supply chain pressures and gas prices have had a much stronger influence than in the two decades before the pandemic.
Without supply shocks, core inflation would still have risen to record levels. In a counterfactual exercise, the authors show that core inflation would still have reached 4% even if supply shocks had been absent after the Covid period, suggesting that in any case demand factors did play a relevant role (see Figure 1 – Panel B). Overall, the results in Banbura et al. (2023) confirm the previous findings by Ascari, Bonomolo, Hoeberichts, and Trezzi (2022) (our review of Ascari et al. (2022) is here).
Figure 1. Main results of Banbura et al. (2023).
Panel A. Historical decomposition of core inflation.
Panel B. Core inflation and its counterfactual measure free of energy and global supply chain shock effects.
A critique of the paper
The authors provide a valuable approach to distinguish between different types of supply shocks in driving core inflation. Many papers focus on broad definitions of supply shocks, while the authors pay great attention to separating the contributions of different types of supply shocks. The identification scheme used to distinguish the different shocks appears reasonable overall and is transparent.
However, some types of supply shocks may not be cleanly identified. In particular, the authors assume that crude oil prices do not to react contemporaneously to gas price shocks. This assumption allows them to separate oil and gas price shocks. While it is based on previous findings in the literature, this literature is still in its infancy and further evidence is required to convincingly justify this point.
Estimation uncertainty surrounding the results is sizeable. The authors report the impulse responses of the variables in their model to the different types of shocks and generally find wide confidence intervals. This means the shocks are not sharply identified given the sample and the identification scheme. This estimation uncertainty also affects the historical decomposition shown in Figure 1 and weakens the interpretability of the results.
The model assumes the relationship between the variables stays constant over time. This is questionable. The Covid period constituted a structural break for many aspects of the economy, and substitution effects changed the structure of the economy. The importance of different shocks in driving core inflation are therefore affected by the likely misspecification of the model for the post-Covid period. This does not only mean that the idiosyncratic residuals in Figure 1 may be of a different size, but also that the relative sizes of all other shock contributions could be biased.
Policy implications
The recent inflation surge was partly out of control of central bankers but nonetheless warranted a strong policy intervention. Since supply shocks explain a significant share of the post-pandemic rise in core inflation and these shocks may be transitory, there is support for the idea that policymakers could look through these shocks. Simultaneously, the strong presence of supply-side factors suggests that the surge in inflation observed after the pandemic was out of the central bank’s control to a significant extent. At the same time, the authors show that core inflation would still have risen to record levels in the absence of supply shocks, justifying a decisive intervention by central bankers.
Paper #2: How do supply shocks to inflation generalize? Evidence from the pandemic era in Europe
A new paper “How do supply shocks to inflation generalize? Evidence from the pandemic era in Europe” by Viral Acharya (NYU Stern), Matteo Crosignani (NY Fed), Tim Eisert (Nova SBE) and Christian Eufinger (IESE) provides a detailed study of the transmission of supply constraints to inflation through household inflation expectations and firms’ pricing power. Using European micro data for the pandemic period, the authors show that supply-side shocks can generate broad-based inflation through an amplification mechanism via inflation expectations and markups.
What the paper does
The paper starts from explaining how production constraints transmits to inflation. The mechanism is split into several parts, and the paper aims to study each part empirically. The key theoretical steps in the transmission mechanism are:
- Firms in certain industries are hit by a supply shock, which leads to cost-push inflation in these industries.
- Households learn about the increase in inflation and adjust their inflation expectations. This happens because they experience higher prices when making purchases and because they learn about production constraints in the news.
- As household inflation expectations rise, firms with high pricing power can capitalize on these expectations and increase their markups. This is because households exhibit a lower elasticity of demand when inflation is high: price comparisons across competitors become more costly when prices are changing frequently, and the expected rise in price levels reduces the benefit of finding the best deal today compared to consuming today instead of tomorrow.
Empirical results
The authors use rich micro data on consumer and producer prices as well as survey data on consumer expectations and firm constraints to study these channels empirically. They document that:
- Industries with tighter supply chain constraints exhibited stronger PPI and CPI growth after the pandemic.
- Countries with stronger supply chain constraints across industries experienced a larger and more broad-based increase in household inflation expectations.
- Countries with stronger aggregate supply chain constraints experience a stronger increase in CPI growth even for goods produced in industries not initially exposed to supply chain constraints. This suggests spillover effects. The effect is driven by countries with high household inflation expectations, and not by high wage growth expectations.
- Firms with higher pricing power were more likely to maintain or increase their markups when inflation expectations were high. This is true even if these firms were not facing supply chain constraints.
A critique of the paper
The paper covers all steps of the transmission channel that it describes in great detail. It makes good use of micro data, allowing it to provide a more granular understanding of the drivers of inflation. The rich cross section in their sample allows the authors to draw tight conclusions even with few time series observations, which poses a common challenge for standard approaches in empirical macroeconomics (for example VARs).
While the authors show empirical evidence on the direction of each part of the channel, it is harder to interpret the magnitudes they are estimating. The paper measures inflation expectations and firm constraints using diffusion indices based on survey data. For example, the inflation index may indicate the share of respondents indicating a “rise” in expected inflation versus a “decline”. While often correlated with the average inflation expectation in the economy, the index does not give direct information about the level of expectations. This makes it hard to derive quantitative implications from the findings presented.
Which sectors are hit by supply shocks probably matters for aggregate inflation but is not discussed in this paper. Sectors of central importance to the economy such as the energy sector or transportation can easily lead to spillovers and broad-based inflation following localized shocks, while shocks to other sectors may rarely lead to a generalized increase in inflation.
It is unclear how much the channel described in this paper contributed to inflation observed in the post-Covid period. While the authors show that there is evidence in favour of the channel they describe, it is unclear what the actual contribution of this channel to inflation numbers has been. The channel may be precisely measured but of small importance. Similarly, it is unclear how much of the rise in inflation that the channel can explain is due to the initial impulse of supply side constraints and how much due to the amplification via inflation expectations and pricing power. This distinction has important implications for the ability of central banks to see through shocks.
Policy implications
Supply side shocks to certain sectors can lead to broad-based inflation via an interaction with firms’ pricing power and household expectations. This questions the ability of central banks to “see through shocks”. This strategy is only valid if inflation expectations stay anchored. Monitoring and managing inflation expectations becomes crucial. Central banks may want to act pre-emptively to avoid inflation spirals, even if initial disturbances are concentrated in a subset of sectors. The amplification mechanism through expectations also makes expectation management and transparent communication important policy tools.