With recent shifts in inflation expectations, we’re reviewing three research papers on the topic. Our previous discussions on inflation expectations can be found here and here, and we’ve also explored how expectations influence the Fed staff forecast here and here.
The first paper, Disagreement About the Term Structure of Inflation Expectations by Ahn and Farmer, examines how long-term beliefs, along with private and public information, shape inflation expectations and the level of disagreement around them. The second paper, Do Household Expectations Help Predict Inflation? by Brandao-Marques et al., finds that household inflation expectations can serve as useful predictors, with most of their explanatory power coming from the median of the distribution. The final paper, The Causal Effects of Inflation Uncertainty on Households’ Beliefs and Actions by Georgarakos et al., shows that when inflation expectations rise, consumers tend to make purchases sooner to avoid higher prices, allocate more assets to inflation-hedging investments, and reduce job search activity.
Paper #1: Ahn and Farmer (2024)
Disagreement About the Term Structure of Inflation Expectations by Hie Joo Ahn (Federal Reserve Board) and Leland E. Farmer (University of Virginia) examines why professional forecasters disagree on inflation expectations over different time horizons. Using data from the Survey of Professional Forecasters (SPF), the authors analyse the sources of this disagreement and its broader implications.
What the paper does and main results
The authors introduce a structural dynamic factor model to analyse how individual forecasters form their expectations over different time horizons. Inspired by the Nelson-Siegel framework, the model breaks down expectations into three components: long-term beliefs, reactions to public information, and private signals unique to each forecaster.
Using data from the Survey of Professional Forecasters (SPF), the study examines how disagreement varies across forecasting horizons and economic conditions. The findings show that in stable periods, long-term disagreement stems from differences in long-run beliefs, while short-term disagreement is driven by private information. However, during economic downturns or times of high inflation uncertainty, forecasters interpret public information in different ways, leading to greater disagreement across all horizons.
This divergence in interpreting public information can slow the transmission of monetary policy, sometimes resulting in a “price puzzle,” where inflation temporarily rises after a policy tightening. In contrast, when disagreement comes from private information, monetary policy adjustments tend to stabilize the economy more quickly (see Figure 1).
Figure 1: Effectiveness of monetary policy (on industrial production “IP” and consumer price inflation “CPI”) when public information is the source of disagreement (magenta lines) vs when private information is the source (blue lines).
A critique of the paper
The paper’s key takeaway is that public information plays a major role in disagreements among professional forecasters, which in turn weakens the effectiveness of monetary policy. The authors suggest that better communication can help reduce this disagreement and improve policy transmission. However, the empirical results should be interpreted with caution, as the median forecaster tenure in the dataset is only 14 quarters. The limited number of observations per forecaster reduces the precision of the model’s decomposition, given the large number of parameters involved.
Paper #2: Brandao-Marques et al. (2024)
Do Household Expectations Help Predict Inflation? by Luis Brandao-Marques (IMF), Gaston Gelos (BIS), David J Hofman (IMF), Julia Otten (IMF), Gurnain Kaur Pasricha (IMF), and Zoe Strauss (Morgan Stanley) examines how the distribution of household inflation expectations influences inflation forecasts. The study finds that household expectations can be valuable predictors, with most of the predictive power coming from the median of the distribution.
What the paper does and main results
The authors analyze how household inflation expectations, based on the Michigan Consumer Survey, can help predict U.S. inflation rates. They focus on the distribution of these expectations rather than just the average, highlighting that household views on future inflation are often widely dispersed and multimodal. While the median expectation generally aligns with the inflation target, opinions range from deflation to double-digit inflation, especially during high-inflation periods when the distribution becomes more skewed.
The study finds that household expectations have predictive power for future inflation. Figure 2 presents regression results for one-year-ahead inflation, covering the full sample as well as different subperiods. Three key findings emerge: (i) Median expectations are strong predictors of inflation, (ii) Variance and skewness contribute little additional explanatory power (as measured by R-squared), (iii) The relationship between expectations and inflation is highly unstable over time. While the median expectation’s coefficient averages 1 over the full sample, it fluctuates widely across subperiods, ranging from -0.83 to 2.98.
Figure 2. Main result of Brandao-Marques et al. (2024)
A critique of the paper
Given the wide dispersion in household inflation expectations, a key question is what drives this disagreement. While data limitations may constrain a full analysis, identifying which households tend to have higher inflation expectations is crucial.
Are these expectations based on actual personal inflation experiences, or do some households consistently over- or underestimate inflation? Factors such as political polarization or generational differences could play a role. Another important aspect is whether households move within the distribution over time or if certain groups consistently occupy the same percentiles. Additionally, how much do extreme expectations at the tails influence the risk of de-anchoring inflation expectations? Understanding these dynamics could provide deeper insights into household behavior and its impact on inflation forecasting.
Policy implications
The relationship between household expectations and actual inflation is highly unstable, making it challenging for policymakers to rely on expectations alone to forecast inflation trends. Households may either anticipate inflation changes accurately or lag behind, depending on how quickly their expectations adjust. Given current conditions, it remains unclear whether elevated household inflation expectations signal a potential inflation resurgence or simply reflect a slow adjustment to past inflation trends.
Paper #3: Georgarakos et al. (2024)
The Causal Effects of Inflation Uncertainty on Households’ Beliefs and Actions by Dimitris Georgarakos (ECB), Yuriy Gorodnichenko (University of California, Berkley), Olivier Coibion (University of Texas Austin), and Geoff Kenny (ECB) examines how inflation expectations and uncertainty influence household beliefs and economic behaviour. Using randomized information treatments from the ECB’s Consumer Expectations Survey, the authors separate the effects of inflation expectations (the average forecast) from inflation uncertainty (the degree of confidence in those forecasts). The study analyses how these factors impact key household decisions in the euro area, including durable goods purchases, portfolio allocation, labour supply, and other financial behaviours.
What the paper does and main results
The authors use a randomized information treatment within a survey to independently manipulate inflation expectations and uncertainty, allowing them to analyse their distinct effects on consumer behaviour. This approach helps address key empirical challenges like endogeneity and clarifies the separate impacts of inflation expectations (average forecast) and inflation uncertainty (degree of confidence in forecasts).
The experiment, conducted in 2023 and 2024 as a randomized controlled trial (RCT), involved 19,000 households across 11 euro-area countries, with participants answering monthly online surveys. Respondents received different information treatments—some received no additional information, while others were given details such as professional forecasters’ mean inflation projections. This setup enabled the researchers to isolate the effects of inflation expectations from uncertainty.
Key Findings include:
Durable Goods Purchases: Higher inflation uncertainty reduces durable goods purchases by up to 23% in the following months, while higher inflation expectations have the opposite effect, prompting consumers to buy sooner in anticipation of rising prices (see Figure 3).
Portfolio Allocation: When inflation uncertainty rises, households shift more assets into liquid and safe investments, like savings accounts, while reducing exposure to riskier assets such as stocks and retirement funds. In contrast, higher inflation expectations lead to increased investment in inflation-hedging assets.
Labor Market Behavior: Uncertainty about inflation encourages precautionary job searches, leading to employment shifts such as moving from part-time to full-time work. However, when inflation expectations rise, job search efforts decline.
Other Financial Decisions: Households facing greater inflation uncertainty become more active in price comparisons while shopping and show a stronger preference for fixed-rate mortgages.
Figure 3: Dynamic effect of inflation expectations and uncertainty on durable goods purchases (x-axis: months, y-axis: % change durable good spending)
A critique of the paper
As with any survey-based study, the findings depend on self-reported behavior, which raises questions about possible measurement issues.