While we wait for the FOMC… a couple of our literature reviews. Enjoy!
In this note, we review two recent papers: “Inflation Perceptions during the Covid Pandemic and Recovery” by Fed Board economists David Lebow and Ekaterina Peneva, and “The Transmission of Supply Shocks in Different Inflation Regimes” by Sarah Arndt and Zeno Enders (Heidelberg University). Please, note that Ekaterina Peneva is the Chief of the inflation desk at the Fed Board, and David Lebow is the line officer of the same desk (translated: what they say reflects the Fed staff view).
The first paper shows that inflation perceptions (as opposed to inflation expectations) remain elevated. The paper discusses the possible reasons and the risks for policymakers (essentially that the “last mile” can be harder than anticipated).
The second paper estimates the response of consumers prices to supply shocks in two states of the world (high vs low-volatility regime) and offers a model to rationalize the results. We are a bit skeptical of the methodology, and we discuss why.
Paper #1: Lebow and Peneva (2024)
A new article (Inflation Perceptions during the Covid Pandemic and Recovery) by Fed economists David Lebow and Ekaterina Peneva studies the changes in US consumers’ inflation perceptions since the Covid pandemic. Inflation perceptions (as opposed to inflation expectations) closely tracked realized inflation in 2021-2022 but declined more sluggishly 2023. These patterns are stable across various demographic groups and suggest that the ‘last mile’ towards the 2% inflation target may be the hardest.
What the paper does
The paper looks at the evolution of US consumers inflation perceptions vs inflation expectations in the Michigan Consumer Survey (MCS). The paper correlates the consumers perceptions to their age, race, gender, income groups, and political affiliations, and compare them with realized and expected inflation.
Main results
Households’ inflation perceptions have tracked the rise in inflation closely but overstated the peak and persistence of inflation rates. Since 2016 the Michigan Consumer Survey (MCS) asks households about their perceptions of realized inflation over the previous 12 months, similar to the standard question on inflation expectations. Figure 1 shows the 12-month change in the CPI (solid black line) against inflation perceptions from the MCS. During the low-inflation period of the last decade, inflation perceptions were stable around the 2% target. Inflation perceptions tracked realized inflation accurately during 2021 and into 2022 but then peaked at 10% compared to a 9% peak for realized inflation. While realized inflation declined swiftly, inflation perceptions only decreased with a lag. At the end of 2023, realized inflation hovers around 3% while inflation perceptions remain above 6%. This pattern is robust across age, race, gender, and income groups. The biggest differences occur across political affiliations, with Republicans perceiving higher inflation rates than Democrats during the Biden presidency.
Figure 1. Realized inflation vs perceived inflation from Lebow and Peneva (2024).
News coverage of price increases may play an important role in shaping these patterns. First, the accuracy of inflation perceptions has been very high during the period of the inflation increase, suggesting households were well informed about the level of price increases. Second, since inflation was widely covered in the media, all demographic groups had access to similar information, though outlets with different political leanings may have reported differently on the topic. Third, inflation perceptions may have overshot realized inflation due to well-documented features of media coverage and household behavior: News coverage of unfavorable events is stronger than of favorable ones, and households remember price increases more than decreases. Price increases for frequently purchased goods such as groceries may also shape households’ perceptions lastingly.
While inflation perceptions fluctuated strongly, inflation expectations remained well anchored. 12-month forward-looking inflation expectations increased only moderately throughout the 2020-2023 period and long-term expectations remained virtually unchanged throughout. Interestingly, mentions of price increases in the MSC have decreased in line with realized inflation, but inflation perceptions remain elevated regardless. Consumer sentiment is negatively associated with inflation perceptions during times of high inflation but unrelated otherwise. High inflation perceptions may therefore explain the weakness in US consumer sentiment despite strong real fundamentals.
A critique of the paper
The authors provide a detailed and careful description of patterns in inflation perceptions over the last few years. Inflation perceptions are generally considered less than inflation expectations but may reveal interesting patterns about the state of the economy. From this perspective, it is interesting to see that consumer sentiment is negatively associated with inflation perceptions and that perceptions lag the decline of realized inflation.
The paper provides a sensible interpretation of the role of media coverage in determining inflation perceptions. News outlets provide the key source of information for households on economic issues and are the dominant channel through which households learn about central bank decisions. The fact that Republicans vs Democrats show the largest disagreement in inflation perceptions across all demographic groups is anecdotal support for the importance of the media in shaping household perceptions.
While not the focus of the paper, more evidence is required to establish the causal effects of news coverage on inflation perceptions and on the relation between inflation perceptions and consumer sentiment. Inflation perceptions may have risen irrespective of the type of media consumed and large disparities in media consumption within demographic groups may be hidden by the average perceptions. The determining factors of inflation perceptions therefore remain unclear and would need to be studied using detailed micro data at the household level.
Policy implications
The last mile on the way to the 2% inflation target may be the hardest because inflation perceptions remain high. While inflation expectations remain well anchored, households viewing inflation as remaining high can add persistence to the inflation process and force the Fed to adopt a higher-for-longer policy stance. High inflation perceptions increase the risk of wage-price spirals even when inflation expectations remain anchored as workers may negotiate for wages to catch up with real income losses.
The media plays an important role in the transmission of monetary news to households. While monetary policy decisions affect the wider population, reaching households directly is a difficult business for central banks. This gives central banks only indirect control over the narrative surrounding its policy decisions and the resulting perceptions that consumers form about the state of the economy and the central bank’s actions. A fragmented media landscape may artificially result in heterogeneous inflation perceptions even within a monetary union. While not of first-order importance, the changing nature of news media and increasing prevalence of disinformation campaigns, especially on social media, may affect the transmission of monetary policy.
Paper #2: Arndt and Enders (2024)
A new paper (The Transmission of Supply Shocks in Different Inflation Regimes) by Sarah Arndt and Zeno Enders (both at Heidelberg University) argues that supply shocks have larger effects on inflation when inflation volatility is high. The authors explain this result with a model in which price flexibility is more beneficial for firms when inflation volatility is large. We are skeptical of the methodology and the interpretation of results, as we discuss below.
What the paper does
The authors estimate the responses of monthly CPI inflation to a measure of supply shocks for two different states: a high-volatility and a low-volatility regime. Supply shocks are identified as movements in crude PPI inflation that are 1) coinciding with outliers in the crude PPI inflation series, and 2) occur when crude PPI and crude industrial production (IP) move in opposite directions. This yields a monthly time series of 0-1 indicators (no shock/shock) that are used as instruments to identify the causal effects of supply shocks.
Main results
US inflation responds more strongly to supply shocks when inflation volatility is high. Figure 2 shows the results. The red dashed line shows that CPI inflation responds more strongly to the measure of supply shocks when inflation volatility is high (state 2 in red/pink). The response of crude IP is also stronger.
Figure 2: State-dependent impulse responses from Arndt and Enders (2024).
Shocks to upstream prices have strong effects on final goods prices. The authors apply their methodology to crude PPI (upstream), intermediate goods PPI, and finished goods PPI (downstream) to study the effect of supply shocks at different stages of the production chain on CPI inflation. The effects of 1% shocks to upstream prices are naturally weaker than those for downstream prices – which are much closer to CPI – but remain sizeable (peak effects of 0.25% compared to 0.5%).
A model of endogenous price flexibility can rationalize the results. The authors consider a framework in which firms face costs to changing their prices but can affect this cost by investing in price flexibility. Firms are willing to invest in price flexibility if the present value of discounted cost of this investment is smaller than the expected gain from being able to reset prices. The profit gain from resetting prices is larger when the overall price level fluctuates more strongly. Similarly, higher price flexibility means that more firms are willing to adjust their prices in response to shocks, such that inflation responds more strongly to shocks when inflation volatility is high.
A critique of the paper
The paper asks an important question: Which factors determine the passthrough of supply shocks to inflation? Unfortunately, we are not convinced by the approach the authors have chosen to address this question, nor by the interpretation of their results. Here’s why:
- The direction of causality is unclear: Do shocks have larger effects on inflation when inflation volatility is high simply because the shock increased inflation volatility? The authors use a Markov switching approach to determine the inflation regime in a data-dependent and agnostic way. This is good for avoiding a priori biases affecting the results, but also limits the interpretability of the results. Is there an underlying factor that drives both inflation volatility and the responsiveness of inflation to shocks?
- There is no clear distinction between shocks and the outcome variable. A key prediction of the authors’ model is that even large shocks will have a weak effect on inflation if they happen in a low-volatility regime. But in the empirical application, the shocks are constructed from the time series of inflation, so we do not observe if a large move in inflation is the result of a small or a big shock. This dramatically undermines the interpretation of the results.
- The model is based on the idea of investments in price flexibility. This makes the cost of price adjustment endogenous. We are skeptical how quickly firms can achieve meaningful differences in their price flexibility. Since most high-volatility regimes only last several months, many firms may not find it optimal to invest in more price flexibility if they cannot make use of the lower cost of price adjustment anytime soon. Instead, a standard model of menu costs predicts that firms adjust their prices more in response to large price changes, which could explain the findings. Similarly, models of rational inattention predict that firms adjust their prices more frequently when inflation is high, providing another alternative explanation.
Policy implications
The paper touches upon two relevant topics for policy makers. First, the pass-through of upstream price pressures to the overall price level can be strong. The role of supply chains and the production network for aggregate inflation movements is being increasingly studied. Recent work suggests that certain upstream sectors (such as energy) may disproportionately affect inflation dynamics and that optimal inflation stabilization policy should weight sectors not just by their size but also by their upstreamness.
Feedback effects from endogenous choice of price adjustment can contribute to inflation volatility. The source of these feedback effects can vary. Investment in price flexibility, menu costs, or rational inattention all predict that once inflation passes a certain threshold, the benefit of price adjustment increases for firms and contributes to the propagation of initial shocks. This provides a rationale for the persistence of inflation and suggests that policymakers should focus on early interventions to choke off inflationary pressures.