Has the U.S. Economy Transitioned to a Higher Long-run Real Interest Rate Regime?

In a September 2015 speech, then-Minneapolis Fed President Narayana Kocherlakota argued that the long-run financial market real interest rate had declined sharply relative to levels that prevailed from 2004 to 2011. He argued that the decline in the long-run real market interest rate likely reflected both a decline in the long-run neutral real interest rate (what some economists and policymakers refer to as r*, also called r-star) and the market’s expectation of less accommodative (“tighter”) monetary policy in the future. The latter was reflected in the market’s expectations of inflation below the Federal Open Market Committee’s 2% inflation target over the longer run.

A little less than a year later, then-St. Louis Fed President Jim Bullard announced the “St. Louis Fed’s New Characterization of the Outlook for the U.S. Economy” (PDF). In Bullard’s view, the U.S. economy’s output growth had downshifted to a “low growth regime” characterized by real GDP growth of 2%, inflation of 2%, and an unemployment rate of 4.7%. The low real growth regime reflected, in Bullard’s view, a shift to a low labor-productivity growth regime. Although the two analyses by the then Fed presidents differed, the conclusions were the same: The U.S. economy had likely transitioned to a regime of low real financial market interest rates.It is important to note, as Bullard did, the distinction between real returns on financial assets (bonds) and tangible capital. Regarding the latter, see Paul Gomme, B. Ravikumar and Peter Rupert’s 2017 paper.

This blog post examines these past findings in the context of the recent rise in real interest rates to explore 1) whether the period of persistently low real interest rates has ended and 2) whether the long-run neutral real interest rate has increased.

Changes in the Neutral Real Interest Rate

Monetary policymakers tend to closely monitor estimates of the neutral real interest rate. One reason is that the real interest rate is used in some monetary policy rules, such as a Taylor rule. According to one definition, the neutral rate of interest is the real short-term interest rate that would prevail if there were no transitory financial disturbances.See Frank J. Fabozzi and Frank J. Jones, Foundations of Global Financial Markets and Institutions. Fifth Edition. The MIT Press, 2019, pp. 369-70. This definition was, more or less, the framework used by Bullard, who measured the real interest rate as the yield on the one-year constant maturity Treasury security less the 12-month percent change in the Dallas Fed’s trimmed mean personal consumption expenditures (PCE) price index.

The long-run neutral real interest rate is an important indicator for monetary policymakers and investors. It is commonly 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. If the real federal funds rate were below the real neutral rate, monetary policy would be classified as “easy,” which would tend to put upward pressure on employment and inflation. Conversely, if the real federal funds rate were above the neutral rate, policy would be classified as “tight,” with downward pressure on employment and inflation.

One challenge for policymakers is that the neutral real interest rate is an unobservable variable—unlike the yield on the one-year Treasury security. The “latent” feature of the neutral real interest rate has spurred some economists to estimate the long-run neutral rate using a formal modeling approach. Estimates in this vein include the r* estimates by Holston-Laubach-Williams or Laubach-Williams (New York Fed) or Lubik-Matthes (Richmond Fed).

Model-based estimates of r* 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. In this vein, see Daniel Buncic’s 2024 article. By contrast, Kocherlakota used a long-term financial market-based interest rate to argue that the lower real interest rate regime—for lack of a better word—was probably due to a decline in the long-run neutral real interest rate or the expectation of a more restrictive (“tighter”) monetary policy in the future.

The two figures below are updates of the series referenced by Kocherlakota in 2015. The first figure plots the 10-year, 10-year (10Y10Y) forward interest rate that is based on Treasury Inflation Protected Securities (TIPS) yields. In this case, the 10Y10Y forward rate is the market’s expectation of the average annual real interest rate that will prevail over the 10-year period that begins 10 years from the current observation. The forward interest rate can be viewed as the “market’s consensus regarding future interest rates.”Fabozzi and Jones, pp. 393-99. In the figure, the latest observation is Feb. 23, 2024. The chart also plots the 100-day average of the daily observations.

10-Year, 10-Year Forward TIPS Yields

SOURCE: Haver Analytics.

The second figure plots the market’s inflation expectation over the same horizon. The market’s inflation expectation is the nominal 10Y10Y Treasury forward yield less the 10Y10Y TIPS yields.

10-Year, 10-Year Forward Inflation Compensation

SOURCE: Haver Analytics.

The vertical line in each figure occurs at the date of Kocherlakota’s speech, Sept. 8, 2015. On that date, the 10Y10Y forward TIPS yield measured 1.5% and the 10Y10Y inflation expectations measured 2%. Both the real rate and the market’s long-run inflation expectations were well below levels that prevailed over, roughly, most of the previous decade. Eventually, the long-run neutral real rate would plummet to nearly zero (on a 100-day average basis) in late 2020.

After September 2015, longer-run market-based inflation expectations would fall to around 1.5% in mid-to-late 2016, briefly return to about 2% from mid-2017 to early 2019, but then drop back to around 1.5% in mid-2020, after the onset of the pandemic earlier that year.

Since then, the long-run (10Y10Y) real market interest rate and inflation expectations have rebounded close to levels that prevailed from 2004 to 2012, or thereabout. In other words, the economy in the aftermath of the pandemic may have transitioned out of the previous regime of low real interest rates and a long-run inflation rate under 2%. Using the theory and logic of Kocherlakota, then, the increase in market rates may be the result of an increase in the long-run real neutral interest rate or market expectations of more accommodative (“looser”) monetary policy in the future (conditioned on the assumption of an unchanged long-run neutral interest rate).Measuring the long-term real market interest rate and inflation expectations over a shorter horizon—a five-year, five-year forward calculation—does not alter the conclusion that there has been an upward shift in the real rate and inflation expectations over the past two years.

Productivity and the Neutral Real Rate

If the long-run real neutral rate has increased to more than 2% over the past two years, then the question is why. One possibility, as stressed by Bullard, is that labor productivity (output per hour in the nonfarm business sector) growth has increased. The third figure below plots the annual (four-quarter) percent change in labor productivity since 2005.

Growth of Labor Productivity in the Nonfarm Business Sector

SOURCE: Bureau of Labor Statistics.

Labor productivity growth has exceeded its longer-run average in three of the past five years. Thus, it’s possible that there has been a shift to a faster productivity growth regime. However, productivity growth is notoriously volatile and, moreover, model-based estimates by James Kahn and Robert Rich suggest that there remains a more than 90% probability that the U.S. economy is still in a low-productivity growth regime (1.3% productivity growth on average).

The first and second figures suggest that the U.S. economy may no longer be in a low long-run neutral real interest rate regime. Besides the possibility of a shift to a higher labor productivity growth regime, another possibility that could explain a higher long-run real market interest rate: a permanent increase in the level of public debt that is financed by higher future taxes or lower future transfer payments (e.g., Social Security or Medicare outlays). A key to this argument, raised by Kocherlakota, is whether households increase their savings today to pay the higher taxes in the future to service the debt. This is known as Ricardian equivalence, following Robert Barro’s seminal article (PDF). If Ricardian equivalence holds, then the increase in savings today results in an unchanged real interest rate. But if Ricardian equivalence does not hold (“non-Ricardian”), as Kocherlakota noted (referring to a different class of models), then an increase in the supply of debt leads to an increase in the long-run neutral real interest rate.

The final figure below plots the level of federal debt held by the public relative to the trend that prevailed from 2007 to 2019. The chart suggests a permanent increase in both the level of federal debt and the growth of federal debt in the future using current projections from the Congressional Budget Office. While visual evidence is far from definitive, an increase in the long-run neutral real rate under the expectation of a permanent increase in federal debt seems plausible under non-Ricardian assumptions noted above.

Federal Debt Held by the Public

SOURCE: Congressional Budget Office.

NOTE: Actual data through 2022; CBO projections through 2030.

Conclusion

The evidence presented in this blog post suggests that the long-run real market interest rate has increased over the past two years, nearly returning to the levels that prevailed from 2004 to 2012. The higher market real rate could reflect a higher long-run neutral real interest rate or the expectation of “looser” monetary policy in the future that results in an inflation rate above the 2% target set by the FOMC. That said, two years is probably not a sufficiently long enough time to declare with much confidence that the U.S. economy has shifted to a higher real interest rate regime. Indeed, as noted by Bullard, regime switches are generally not forecastable. Still, a higher long-run neutral real interest rate, if it has occurred, has important implications for monetary policy in the long run, as noted above.

Notes

1. It is important to note, as Bullard did, the distinction between real returns on financial assets (bonds) and tangible capital. Regarding the latter, see Paul Gomme, B. Ravikumar and Peter Rupert’s 2017 paper.
2. See Frank J. Fabozzi and Frank J. Jones, Foundations of Global Financial Markets and Institutions. Fifth Edition. The MIT Press, 2019, pp. 369-70.
3. Fabozzi and Jones, pp. 393-99.
4. Measuring the long-term real market interest rate and inflation expectations over a shorter horizon—a five-year, five-year forward calculation—does not alter the conclusion that there has been an upward shift in the real rate and inflation expectations over the past two years.