Natural Rate of Interest (r-Star): The Neutral Rate Explained

The natural rate of interest, or r-star, is a theoretical benchmark for the real short-term interest rate that would prevail when the economy is operating near its potential and inflation is stable. In that setting, the rate is often described as neither pushing demand above sustainable capacity nor holding it below that path. The New York Fed describes the Laubach-Williams and Holston-Laubach-Williams model estimates in similar terms: a real short rate expected when the economy is at full strength and inflation is stable. {source:nyFedRstarMeasurement}

“Natural” does not mean that one rate is ordained, observable, or identical across every country and model. It labels an equilibrium concept inferred from economic relationships. Some economists use “natural rate,” “equilibrium real rate,” and “neutral real rate” almost interchangeably; others reserve them for different model definitions or horizons. When a source gives an estimate, read its definition and horizon before comparing it with another number.

The benchmark is real, while the policy rate most readers see is nominal. The federal funds target range, for example, is an administered nominal rate. Economists adjust a nominal rate for expected inflation to discuss its real purchasing-power counterpart. The neutral rate itself is not the same as a mortgage quote, a Treasury yield, a deposit rate, or a central bank's announced target.

The useful intuition is conditional: if the real policy rate were around the relevant neutral rate, and other forces were unchanged, monetary policy would not add a large extra push or drag to demand over that horizon. The word “around” matters. The benchmark is uncertain, other shocks continue, and the same policy setting can have different effects as households, firms, credit markets, and expectations change.

A short-run neutral rate can move with cyclical forces that strengthen or weaken demand. Fiscal changes that boost near-term spending, greater investor risk appetite, or other demand tailwinds can raise the rate consistent with stable inflation and sustainable output for a time. Business uncertainty or weaker foreign demand can push it lower. This near-term benchmark is not a fixed value and need not equal the long-run rate. {source:fedBrainardNeutral}

The longer-run neutral rate refers to a setting after temporary cyclical forces have faded and output is growing near its longer-run trend while inflation is at its objective. It is a structural reference for where policy rates might settle over a longer horizon, not a weekly reading of current economic slack. Productivity, demographics, desired saving, investment demand, and the supply of safe assets can all influence it.

The two estimates can differ without either one being an error. A temporary demand boom might put the short-run neutral rate above its longer-run trend; a severe contraction or elevated precautionary saving might put it below. For near-term policy assessment, Brainard's speech emphasizes the short-run neutral rate when cyclical headwinds or tailwinds are present. A longer-run figure is a poor substitute for that question simply because it is easier to find in a table. {source:fedBrainardNeutral}

The longer-run federal funds rate shown in the Federal Reserve's Summary of Economic Projections is related, but it is not a direct measurement of r-star. Participants submit their individual views of longer-run normal values under their own outlooks and assumptions. The SEP is a collection of projections, not an FOMC target or a real-time measurement of the unobservable neutral rate. {source:fedPowellChangingEconomy}

R-star is a real-rate concept. To compare it with a nominal policy setting, inflation has to be brought onto the same horizon. A common long-run approximation is i* ≈ r* + π*, where i* is the nominal neutral rate, r* is the real neutral rate, and π* is the inflation objective or an assumed long-run expected inflation rate. That approximation relies on a long-run setting in which expectations are aligned with the objective and the rates use comparable horizons.

The exact Fisher relation is 1 + i* = (1 + r*)(1 + π*) when rates are written as decimals and use the same compounding convention and period. In a more general ex-ante comparison, expected inflation over the matching horizon belongs in the formula; an inflation target is a suitable substitute only when the long-run expectation is assumed to be anchored there. The Federal Reserve's conventional explanation similarly derives the neutral real rate by subtracting long-run inflation from the longer-run nominal federal funds rate. {source:fedPowellChangingEconomy}

For a clearly hypothetical calculation, suppose a fictional central bank has a long-run inflation objective of 2.0% and an estimated long-run real neutral rate of 1.0%. The additive approximation gives 1.0% + 2.0% = 3.0% nominal. Applying the exact form gives (1.01 × 1.02) − 1 = 0.0302, or 3.02%. The exact result is 0.02 percentage point above the approximation because the product includes the small interaction term. These inputs illustrate the arithmetic only; they are not a current estimate, an official target, or a forecast for any country.

Do not mix the relation above with a realized inflation number from a different period. The ex-ante real rate uses inflation expected over the life of the nominal rate; an ex-post real rate uses inflation that actually occurred over a matched period. A calculation can be numerically correct and still answer the wrong question if the inflation measure, horizon, or compounding basis does not match.

There is no market screen showing the economy's true r-star. Analysts infer it from observable data because equilibrium rates, potential output, and stable-inflation conditions are not directly measured. One influential approach, the Laubach-Williams model, uses real GDP, inflation, and the federal funds rate to estimate trends in economic growth and other factors related to r-star. The Holston-Laubach-Williams extension applies the framework to additional economies and later adapts it to changes in volatility and pandemic-era supply shocks. These are model estimates, not readings from a traded instrument. {source:nyFedRstarMeasurement}

A model has to separate persistent forces from temporary shocks using assumptions about how output, inflation, and interest rates relate. Different models choose different data, equations, trend structures, sample windows, and treatments of unusual periods. Even when models use the same data, their estimated latent rate can differ because the inference depends on the structure imposed by each model. There is no single estimation method that turns r-star into an observed fact.

Other approaches provide cross-checks, not a perfect substitute. Statistical trend extraction may be distorted by large shocks. Surveys record what respondents believe about a future neutral rate, not an independent observation of the underlying equilibrium. Long-horizon market rates can contain expected future real short rates, term premia, liquidity effects, and risk compensation. Williams discusses why those market and survey measures can be informative while still requiring careful interpretation; model-based estimates also retain substantial uncertainty. {source:nyFedWilliamsGlobalRstar}

When reading an estimate, ask what the model calls neutral, whether the number is real or nominal, whether it is short-run or long-run, what country and data vintage it covers, and how uncertainty is reported. A point estimate without its method, date, and confidence range can look more precise than the underlying evidence allows.

A policy stance comparison needs rates that match in type and horizon. If the comparison is in real terms, start with the nominal policy rate and subtract expected inflation over a matching horizon to form an approximate ex-ante real rate. Compare that with an estimated short-run real neutral rate. If both figures are nominal, compare the nominal policy rate with a nominal short-run neutral estimate that uses the same inflation assumptions. Avoid subtracting a real r-star from a nominal policy rate without this adjustment.

A policy rate above the estimated short-run neutral rate may indicate a more restrictive setting relative to that model benchmark, all else equal; a rate below it may indicate a more accommodative setting. Neither sign guarantees a particular change in GDP, jobs, or prices. Brainard describes the gap between the policy rate and the nominal short-run neutral rate as a useful near-term reference when cyclical headwinds or tailwinds matter, but the estimate is only one input into policy assessment. {source:fedBrainardNeutral}

Market borrowing rates are not interchangeable with the central bank's policy rate. A mortgage or corporate bond rate also reflects its term, borrower risk, liquidity, and the expected path of future short rates. Nor is the long-run neutral rate a precise “correct” setting for today: cyclical conditions can move the short-run benchmark away from it.

An estimate can change because the economy may have changed, because measured data were revised, or because the model or its parameters changed. Those are different explanations. Productivity, saving, investment, and demographic shifts may alter the underlying equilibrium over time. Separately, revisions to GDP or inflation data can change the model's view of past and current trends even if the underlying economy did not suddenly move on the release date.

Real-time and revised estimates can therefore tell different stories. The New York Fed provides real-time series based on historical data vintages alongside current model estimates. Comparing them shows what an analyst could infer at the time versus what the same framework infers after later data arrive. The page also publishes replication code and dates its releases; its August 27, 2026 data release is a dated vintage, not a timeless value. {source:nyFedRstarMeasurement}

Model histories may be revised when a statistical agency benchmarks GDP, when new inflation data arrive, when an estimation window expands, or when researchers change the treatment of a major disruption. The Federal Reserve's real-time global note updates estimates for 11 advanced economies and warns that the series can be delayed, revised, or changed methodologically. It also notes that productivity inputs themselves are updated infrequently and can later be revised. {source:fedRealtimeGlobalNeutralRates}

A larger or smaller new estimate is therefore not, by itself, proof that the economy's sustainable real rate just shifted by that amount. Check which vintage is being compared, whether both estimates use the same model and definition, and how much of the difference comes from data or method changes. Use dated estimates rather than an undated headline.

Long-run neutral rates reflect forces that shape desired saving and investment as well as the return to productive capital. Productivity growth can affect investment opportunities; demographics can change the share of people saving or drawing down assets; the supply of government safe assets and demand for them can alter equilibrium yields. Global capital flows also transmit productivity and demographic developments across borders. These drivers can push in different directions and need not move together. {source:fedLongerRunNeutralRates} {source:fedRealtimeGlobalNeutralRates}

Some forces are global, while local institutions and conditions still matter. Williams describes common saving and investment forces across countries but also notes country-specific factors. Federal Reserve staff research likewise shows that central banks communicate neutral rates differently: some publish a range or a regular estimate, while others rely on speeches, models, or do not use the concept as a live policy guide.

Country comparisons require consistent definitions. One figure may be a real model estimate, another a nominal range communicated by a central bank, and another a survey median. They can differ in horizon, publication date, inflation assumption, and method. A 2022 comparison of major economies describes that period's practices and model results; it is useful evidence that practices vary, not a source of today's cross-country readings. {source:fedLongerRunNeutralRates}

An r-star estimate can organize a question: is the policy rate high or low relative to an inferred neutral benchmark, and what assumptions drive that comparison? It can help analysts compare scenarios, explain how persistent productivity or saving changes might affect long-run rates, and make a Taylor-rule input explicit. It does not reveal a single observable “true” rate or tell a reader which loan, deposit, or investment return is fair.

It also cannot mechanically label policy as tight or loose without matching the horizon, real-versus-nominal basis, and inflation assumptions. It cannot guarantee what inflation or output will do, nor identify the next central bank decision. Williams cautions policymakers against relying too heavily on precise r-star estimates; policy judgments also draw on other data, risks, and assessments. The neutral estimate is a reference, not a forecast or official target. {source:nyFedWilliamsGlobalRstar} {source:fedBrainardNeutral}

For the difference between nominal and real rates and the Fisher relation, see Real vs. Nominal Interest Rates: The Fisher Equation Explained. To see how an assumed neutral real rate enters a conditional benchmark, read The Taylor Rule: How Inflation and Output Shape Rates. The What Is the Output Gap? Potential GDP Explained guide covers another unobserved benchmark used in monetary-policy analysis.