What is secular decline in real interest rates?

The short answer

The real interest rate is the nominal interest rate minus expected inflation. The “natural” or “equilibrium” real interest rate, called r-star, is the rate that would prevail if the economy were operating at full potential. Wicksell introduced the concept in 1898; Laubach and Williams turned it into an estimable time series in 2003.

What makes r-star economically important is that it serves as the benchmark for monetary policy. When the actual policy rate is above r-star, monetary policy is restrictive; when below, it is accommodative. A falling r-star means central banks have less room to cut before hitting the zero lower bound — which is exactly the policy challenge that emerged after 2008.

The puzzle is that r-star fell persistently from the early 1980s through 2020 across nearly every advanced economy. Since 2022, however, the empirical evidence has fragmented — different models give very different answers, and what looked like a structural certainty is now an open question.

→ New to monetary economics? Financial education hub

What the data shows

The two most cited models are Holston-Laubach-Williams (HLW), updated quarterly by the New York Fed, and Lubik-Matthes (LM) at the Richmond Fed.

The empirical context (HLW, Lubik-Matthes, ECB, surveys, 1980-2024):

• US r-star (Lubik-Matthes): from ~3.5% in early 1980s to roughly 0.5-1% pre-COVID
• US r-star (HLW): broadly similar trajectory, around 0.5% pre-COVID
• Euro area r-star (HLW): -0.7% in Q2 2024
• Survey-based r-star (US, ECB Survey of Monetary Analysts): rose from 1.0% to 2.0% between 2021 and 2024
• Lubik-Matthes US estimate has risen post-COVID; HLW estimate has fallen — divergence of 100-200 bp by 2024
• FOMC long-run policy rate projection (Dec 2024): 2.5% nominal, implying ~0.5% real r-star

The exception to highlight: the average standard error on HLW r-star estimates is “very large,” per Holston-Laubach-Williams (2017) themselves — they acknowledge that “r* is barely identified.” The point estimate hides considerable uncertainty.

→ Dataset: US Real Interest Rates History

Why it happens — the macro mechanism

The persistent fall in r-star reflects a global excess of desired savings over desired investment.

Channel 1 — Demographics. Aging populations increase the supply of savings (people in their peak earning years save more) while reducing investment demand (slower workforce growth means less need for new capital). Carvalho-Ferrero-Nechio (2016) estimate that demographics alone explain roughly 1.5 percentage points of the decline in advanced-economy r-star since 1980. See our FAQ on aging and growth.

Channel 2 — Productivity slowdown. The HLW model explicitly defines r-star as trend productivity growth plus a residual. When TFP growth fell from ~2% to ~0.5% per year after 1973, this mechanically pulled r-star down — firms expecting lower returns to investment will pay lower interest rates. The angle that distinguishes this view: if you believe the productivity slowdown is partly intangibles mismeasurement, then the real fall in r-star may be smaller than HLW suggests. See our FAQ on TFP slowdown and our FAQ on intangibles.

A third channel runs through the global savings glut.

Channel 3 — Global savings imbalances. Bernanke’s 2005 “global savings glut” thesis emphasized excess savings from China, Germany and oil exporters flowing into US Treasuries, depressing global real rates. Rachel-Summers 2019 update this with corporate net savings from intangibles-heavy firms — see our FAQ on the savings glut and our FAQ on the capital investment slowdown.

Synthesis by regime. In the Volcker era 1980-1990, r-star was around 3.5% in the US, supported by strong productivity growth and demographic dividend. From 1990 to 2007, r-star drifted lower (to roughly 1.5-2%) as China integrated into the global trading system and demographic peak earnings boosted savings. From 2008 to 2019, r-star approached zero across most advanced economies, with HLW and Lubik-Matthes broadly agreeing. Since 2022, the divergence between models reflects genuine uncertainty about whether the post-COVID surge in inflation, fiscal expansion and AI-related capex marks a regime change — and what was once a confident secular narrative has become a real-time empirical contest.

R-star was supposed to be a slow-moving structural variable. The post-COVID divergence between models suggests it is, instead, a contested estimate of an unobservable quantity.

→ Reading framework: Monetary regimes

What it means for different economic actors

Bond investors. Lower r-star implies lower long-term equilibrium yields on government bonds, holding inflation expectations constant. The post-COVID rise in nominal yields reflects partly higher expected inflation, partly debate about whether r-star has actually risen. The 30-year Treasury at 4.5%+ is consistent with either narrative.

Equity investors. Discount rates for future cash flows depend on real rates. A persistently low r-star supports higher equity multiples; a rise in r-star compresses them mechanically. The CAPE ratio’s behavior across the post-2022 period is partly a real-time bet on r-star direction.

Central banks. A low r-star means policy hits the zero lower bound more often, which is why the Fed and ECB both moved to make balance-sheet tools (QE, LSAPs) part of their permanent toolkits after 2008.

A common error is to read any single year’s real rate as evidence about r-star. Real rates fluctuate enormously around the natural rate; only multi-year averages and structural models can isolate the trend.

Practical observation

Go deeper

Related questions

• What is the savings glut thesis?
• Why has capital investment slowed despite low rates?
• What is total factor productivity and why has it slowed?
• Real vs nominal interest rates: what matters?

Frequently asked questions

Why have HLW and Lubik-Matthes r-star estimates diverged so sharply since 2022?

The two models use different identifying assumptions. HLW imposes a tight structural link between r-star and trend productivity growth, which has remained slow — pulling its r-star estimate down. Lubik-Matthes uses a more flexible statistical approach with no such restriction, which lets r-star rise in response to recent high real rates. The divergence reflects genuine uncertainty about whether post-COVID conditions represent a temporary disturbance or a structural regime change. As Lubik himself notes, having multiple competing models is itself useful for sharpening monetary policy thinking.

Is r-star the same as the Fed Funds rate target?

No. The Fed Funds rate is the actual policy instrument the Fed sets. R-star is an unobserved equilibrium concept — the rate consistent with full employment and stable inflation. Monetary policy is restrictive when the real Fed Funds rate exceeds r-star and accommodative when below. The FOMC’s longer-run projection of the policy rate (2.5% nominal as of December 2024) implicitly assumes inflation at 2% and a real r-star around 0.5%.

Could the secular decline in r-star reverse permanently?

Possible, but uncertain. Three structural forces could push r-star higher: demographic shifts as global aging slows the savings glut, fiscal expansion as government deficits increase the supply of safe assets, and AI-driven capex if it materializes at scale. The Carvalho et al 2025 paper suggests US r-star has risen modestly post-COVID. But the empirical evidence is too recent and too noisy to call a definitive reversal of a 40-year trend.

Last updated — 12 July 2026