Is de-anchoring happening?
In this paper, Philippe Andrade and Michael Wicklein ask how much of the spring 2025 surge in U.S. household inflation expectations reflects the usual mechanical response to observable price changes (food, gasoline, core inflation), and how much is autonomous i.e. a potential signal of de-anchoring .
Data and methodology: The authors use one-year-ahead expectations from the University of Michigan Survey of Consumers, stretching back to the 1960s, and estimate a regression model in which households form expectations based on salient prices (gas and food CPIs) and broad-based core inflation.
Four historical episodes: The model identifies four episodes since 1960 in which expectations exceeded 8%: the early 1970s (1973–75), the late 1970s (1978–80), the pandemic surge (2021–22), and the spring 2025 surge .
The 2025 surge is uniquely unexplained: The early-1970s and pandemic surges are largely explained by salient-price increases. The late-1970s surge already showed a larger unexplained residual, consistent with genuine de-anchoring. But the spring 2025 episode is the most anomalous: less than 1 percentage point of the 8 percentage point surge is explained by food, gas, or core inflation, leaving roughly 7.5 percentage points unexplained, far above the 95% confidence interval . The 2025 surge structurally resembles the episode that eventually required the Volcker disinflation to reverse.
The role of tariffs: The authors consider tariff-related uncertainty as a contributing factor, noting that consumers’ expectations tracked tariff perceptions closely early in 2025. However, they emphasize that estimated tariff pass-through (0.8–2.2 pp according to Barbiero & Stein 2025) falls far short of the observed 8 pp spike, so tariffs alone cannot account for the magnitude .
Important caveat: The authors note that longer-term inflation expectations from professional forecasters and financial markets have remained broadly stable near the Fed’s 2% target, and that other household surveys show more muted increases than the Michigan survey. De-anchoring risk is therefore flagged as elevated but not yet confirmed.
Why Have Inflation Expectations Surged Recently? A Historical Perspective
Authors: Philippe Andrade, Michael Wicklein
From: Federal Reserve Bank of Boston, University of Chicago
Why does de-anchoring risk matter?
Luigi Bocola, Alessandro Dovis, Kasper Jørgensen, and Rishabh Kirpalani build a New Keynesian economy with Rotemberg price adjustment costs, populated by households and monopolistically competitive firms. The central bank minimizes a standard loss function over inflation and output, but crucially operates without commitment i.e. it sets policy under discretion each period . The central bank can be one of two types: a Hawk or a Dove but agents cannot directly observe which type they face. Instead, they update their beliefs using Bayes’ rule every time they observe an interest rate decision. The central bank’s reputation is defined as the private sector’s probability that it faces the Hawk .
The key theoretical result: When private agents are uncertain about the central bank’s type, the optimal policy response to an inflationary supply shock is more aggressive than under full information: the central bank deliberately accepts a deeper recession to signal its hawkishness, lower inflation expectations, and build credibility. The strength of this “reputation channel” depends on one sufficient statistic: the elasticity of long-run inflation expectations to monetary policy surprises .
Empirical identification: The elasticity is estimated via high-frequency identification, regressing changes in long-run market- and survey-based inflation expectations on surprises in short-term OIS rates in a two-day window around central bank meetings, for Brazil, Chile, Mexico, the euro area, the UK, and the US (2010–2024) .
Findings
- Large cross-country heterogeneity in the elasticity. For Brazil, a 100 basis-point surprise rate hike reduces five-year-five-year forward inflation compensation by 50 basis points and survey-based four-year-ahead expectations by 18 basis points, both statistically significant. For the US, euro area, and UK, the long-run elasticities are close to zero and not statistically distinguishable from zero .
- The “credibility dividend.” As reputation increases from 0 to 1, both the inflation bias (lower average inflation) and the stabilization bias (smaller inflation and output volatility in response to shocks) improve.
- A better measure of de-anchoring risk. The paper argues that the level of long-run expectations is a poor indicator of de-anchoring risk since a low-reputation central bank can keep inflation expectations at the same level as a high-reputation one by tightening more aggressively. The elasticity of long-run expectations to monetary surprises is a more informative indicator, because it reveals how fragile anchoring truly is .
Conclusion: Bocola et al. show that even if observed long-run expectations appear stable (as they broadly did in 2025), the underlying de-anchoring risk can still be high if the elasticity of those expectations to monetary surprises is large which is exactly the scenario that Andrade and Wicklein’s “unexplained residual” hints at for short-run expectations.
Monetary Policy without an Anchor
Authors: Luigi Bocola, Alessandro Dovis, Kasper Jørgensen, Rishabh Kirpalani
From: Stanford University, University of Pennsylvania, European Central Bank, University of Wisconsin
