Comparing the FOMC’s Estimate of R-Star with Alternative Estimates

The neutral interest rate adjusted for inflation is an important interest rate. From a monetary policy standpoint, the neutral real interest rate is often defined as the rate that would prevail if the employment and inflation gaps were zero—that is, if employment were at its “maximum” level and inflation were at the 2% target. In monetary policy circles, the neutral real interest rate is termed r*, or r-star.Another definition of the neutral real interest rate is the rate that equilibrates the supply of loanable funds (saving) with the demand for loanable funds (investment). The different measures of the neutral real interest rate have recently been discussed by Ricardo Reis, a professor of economics at the London School of Economics; see his April 2025 paper, “The Four R-Stars: From Interest Rates to Inflation and Back” (PDF).

In many models used for policy analysis, if the actual real federal funds rate were below r-star, monetary policy would then be classified as “easy” (or accommodative), which would tend to put upward pressure on inflation and economic activity. Conversely, if the real federal funds rate were above r-star, policy would be classified as “tight” (or restrictive), which would tend to put downward pressure on inflation and economic activity. The assumption built into the models is that monetary policymakers can change the actual real federal funds rate to achieve their objectives by changing the nominal policy rate, the federal funds rate target range.

The Difficulty in Determining R-Star

One challenge is that r-star is an unobservable variable—unlike, for example, the number of jobs or the number of people unemployed. This problem has been known for decades. In his 1967 presidential address to the American Economic Association, Milton Friedman articulated the challenge monetary policymakers face in setting policy based on the neutral real interest rate, also known as the natural rate of interest:

One problem is that it (the monetary authority) cannot know what the “natural” rate is. ... And the “natural rate” will itself change from time to time.Milton Friedman, “The Role of Monetary Policy” (PDF), American Economic Review, Vol. 58, No. 1, March 1968, pp. 1-17.

A similar conclusion was noted decades earlier by Harvard professor John H. Williams in 1931:

The natural rate is an abstraction; like faith, it is seen by its works. One can only say that if the bank policy succeeds in stabilizing prices, the bank rate must have been brought in line with the natural rate, but if it does not, it must not have been.Quoted from Athanasios Orphanides and John C. Williams, “Robust Monetary Policy Rules with Unknown Natural Rates,” Brookings Papers on Economic Activity, 2002, pp. 68-118.

Different Ways to Measure R-Star

Economists and policymakers have nonetheless used various methods to estimate r-star. The first figure below plots six different measures of r-star. One estimate is derived from a financial market instrument that some economists believe is a good proxy for r-star, the 10-year, 10-year (10Y10Y) forward interest rate that is based on Treasury inflation-protected securities (TIPS) yields.See Narayana Kocherlakota’s Sept. 8, 2015, speech, “Public Debt and the Long-Run Neutral Real Interest Rate.” The other five r-star measures are based on model-based estimatesModel-based estimates of r-star have been criticized on numerous grounds. For example, they tend to have wide confidence intervals and, importantly, they can be influenced by actions of the central bank. Regarding the latter, see Daniel Buncic’s 2024 article. Financial market-based estimates have also been criticized, because there are various risk premiums embedded within the yields. Sophia Cho and John C. Williams argue that these premiums are a plausible reason why financial markets are not good predictors of r-star. Also see Thomas A. Lubik and Paul Ho, “Examining the Differences in r* Estimates,” Federal Reserve Bank of Richmond Economic Brief, November 2024, No. 24-36. from the following sources:

• Holston-Laubach-Williams and Laubach-Williams
• Lubik-Matthes
• D’Amico-Kim-Wei
• Zaman

Several observations are evident from the first figure. First, the r-star estimates usually vary widely. For example, in the fourth quarter of 2025, the highest estimate—based on the 10Y10Y TIPS—was a little more than 3%. The lowest estimate—based on the Holston-Laubach-Williams model—was a little less than 1%. Second, they all generally tend to move in the same direction. Third, the model-based estimates are often lower than the financial market-based estimate (10Y10Y TIPS). Fourth, most of the estimates exhibit a relatively high degree of volatility; the exception is the estimate from Zaman. Finally, one estimate (D’Amico-Kim-Wei) usually comprises the lower bound of the range of estimates.

Since nobody knows what the real r-star estimate is in real time (or anytime, for that matter), using an average of competing estimates—rather than a specific model-based estimate—seems reasonable.This view seems consistent with Federal Reserve Gov. Christopher Waller’s suggestion that “we always need to be humble in citing a numerical value for r*.” See Waller’s May 24, 2024, speech, “Some Thoughts on r*: Why Did It Fall and Will It Rise?”

The figure also includes the geometric mean of the six estimates. This measure is termed the alternative r-star estimate. In the fourth quarter of 2025, the alternative r-star was 1.43%, which was slightly lower than the previous quarter when it was at its highest rate since the 2008-09 financial crisis. One might argue that the alternative r-star is perhaps biased upward, since four of the six estimates fall below the average. Nevertheless, there are earlier periods when the r-star average might have been biased downward. For example, in several periods, there were only one or two estimates below the geometric average (e.g., the 2001-07 expansion).

How Does the Alternative Measure Compare with the SEP’s Measure?

The second figure compares the alternative r-star measure with the median estimate published in the Federal Open Market Committee’s quarterly Summary of Economic Projections (SEP). At the conclusion of its March 17-18, 2026, meeting, the median FOMC participant projected that the appropriate nominal federal funds target rate over the longer run was 3.1%.The Summary of Economic Projections (SEP) in its current form began in 2012. Each participant determines his or her appropriate policy for the purposes of attaining the 2% inflation target and maximum employment over time. Since there are 19 participants, it is thus conceivable that there are 19 different views on the economic outlook and thus 19 different views on appropriate policy. The FOMC’s median longer-run appropriate target, minus the FOMC’s 2% target rate, is often viewed as its r-star estimate. Hence, I will call this the SEP r-star.

Several takeaways are worth noting in the second figure. First, the SEP r-star estimate changes relatively infrequently compared with the alternative r-star estimate. This is especially true for 2019 to 2023. Second, the median SEP r-star estimate tends to differ sharply from the alternative r-star estimate. An exception occurred briefly from 2016 to 2018, and periodically from 2020 to 2022. Third, the SEP r-star tends to adjust toward the alternative r-star estimate, albeit with a lag. Finally, the SEP r-star estimate remains below the alternative estimate; however, this gap has been closing over the past year.

Notes

1. Another definition of the neutral real interest rate is the rate that equilibrates the supply of loanable funds (saving) with the demand for loanable funds (investment). The different measures of the neutral real interest rate have recently been discussed by Ricardo Reis, a professor of economics at the London School of Economics; see his April 2025 paper, “The Four R-Stars: From Interest Rates to Inflation and Back” (PDF).
2. Milton Friedman, “The Role of Monetary Policy” (PDF), American Economic Review, Vol. 58, No. 1, March 1968, pp. 1-17.
3. Quoted from Athanasios Orphanides and John C. Williams, “Robust Monetary Policy Rules with Unknown Natural Rates,” Brookings Papers on Economic Activity, 2002, pp. 68-118.
4. See Narayana Kocherlakota’s Sept. 8, 2015, speech, “Public Debt and the Long-Run Neutral Real Interest Rate.”
5. Model-based estimates of r-star have been criticized on numerous grounds. For example, they tend to have wide confidence intervals and, importantly, they can be influenced by actions of the central bank. Regarding the latter, see Daniel Buncic’s 2024 article. Financial market-based estimates have also been criticized, because there are various risk premiums embedded within the yields. Sophia Cho and John C. Williams argue that these premiums are a plausible reason why financial markets are not good predictors of r-star. Also see Thomas A. Lubik and Paul Ho, “Examining the Differences in r* Estimates,” Federal Reserve Bank of Richmond Economic Brief, November 2024, No. 24-36.
6. This view seems consistent with Federal Reserve Gov. Christopher Waller’s suggestion that “we always need to be humble in citing a numerical value for r*.” See Waller’s May 24, 2024, speech, “Some Thoughts on r*: Why Did It Fall and Will It Rise?”
7. The Summary of Economic Projections (SEP) in its current form began in 2012. Each participant determines his or her appropriate policy for the purposes of attaining the 2% inflation target and maximum employment over time. Since there are 19 participants, it is thus conceivable that there are 19 different views on the economic outlook and thus 19 different views on appropriate policy.