Inflation expectations after the US tariff announcement

Following the Trump administration’s announcement of new tariffs on most trading partners, inflation expectations—as measured by the University of Michigan Survey of Consumers — rose markedly between January and April.

The survey provides estimates of households’ average expected inflation over the next year and the next five years. As shown in the figure below, the average one-year inflation expectation increased from 3.3% in January to 6.7% in April. Over the five-year horizon, expectations rose from 3.2% to 4.4%. These are substantial increases by historical standards.

Chart 1. Expected inflation (%)

Source: The Michigan Survey. Data as of 19 April 2025.

But how much significance should be attached to these developments? One interpretation is that consumers’ inflation expectations play an important role in shaping actual inflation. If households and firms expect rising prices, these expectations may be incorporated into wage demands and pricing decisions. From this perspective, the rise in expectations is concerning, as it implies a risk of near-term inflationary pressure that would gradually fade over time.

An alternative view is that the public overreacted to the tariff announcements and that these expectations play a limited role in wage-setting and inflation dynamics. If so, the rise in expected inflation may have little bearing on actual outcomes, and the risk of a sharp rise in inflation is correspondingly low.

To assess the predictive value of inflation expectations, monthly data from 1981 onward is used to examine how well they have historically forecast actual consumer price index inflation over the following one and five years. The first scatter plot below compares one-year-ahead expected inflation (horizontal axis) with realized inflation over the same period (vertical axis). If expectations were perfectly accurate, the observations would lie on a 45-degree line through the origin—implying a one-for-one relationship between expected and actual inflation. In such a case, expected inflation would explain 100% of the variation in subsequent inflation.

In practice, while the scatter plot shows a positive relationship, the data are widely dispersed. On average, a 1 percentage point increase in expected inflation is associated with a 0.93 percentage point increase in actual inflation—somewhat less than expected. The proportion of variation in actual inflation explained by expectations, as measured by the R-squared value, is modest at 34%.

Chart 2. Actual and expected inflation, one-year horizon (%)

Source: EFG calculations on data from the Michigan Survey and FRED. Data as of 19 April 2025.

Although the correlation is statistically significant, this is not a high bar. It is possible that both expected and actual future inflation are influenced by current inflation, raising the question of whether expected inflation adds any information once current inflation is considered. Unfortunately, this does not appear to be the case: expected inflation does not provide additional predictive power for one-year-ahead inflation beyond what is already captured by current inflation.¹

The next graph presents the same analysis for a five-year horizon. Here, the relationship is even weaker. A 1 percentage point increase in expected five-year inflation is associated with only a 0.44 percentage point increase in actual inflation. Although statistically significant, the R-squared is just 19%. Once again, expected inflation adds no forecasting power if current inflation is included in the analysis.

Chart 3. Actual and expected inflation, five-year horizon (%)

Source: EFG calculations on data from the Michigan Survey and FRED. Data as of 19 April 2025.

Taken together, these results suggest that the recent rise in inflation expectations following the tariff announcements should be interpreted with caution. Historically, expected inflation—at least as measured by the Michigan Survey of Consumers — has limited, and in some cases no, additional predictive value for future inflation once current inflation is accounted for. This suggests that the Federal Reserve is unlikely to place much weight on these data when considering the stance of monetary policy.

¹ Using data for the period March 1981 to March 2024 to regress inflation one year ahead on the current inflation rate and expected inflation rate, we can reject the hypothesis that the current inflation does not matter (p = 2.3%) but not the hypothesis that the expected inflation rate is irrelevant (p = 79.8%).