The Taylor rule: formula, worked example and Fed policy since 1990

The Taylor rule is a simple formula that suggests where a central bank's policy rate should be, given inflation and the state of the economy. Stanford economist John Taylor proposed it in a 1993 paper, showing that a rule with just two inputs, inflation and the gap between actual and potential output, described Federal Reserve policy from 1987 to 1992 remarkably well. It has been the most widely used benchmark for judging whether the Fed is "too tight" or "too loose" ever since.

It is a benchmark, not the Fed's actual method. The FOMC looks at many rules alongside a great deal of other information, and it does not follow any of them mechanically. This guide gives the formula exactly as the chart on this page computes it, works through the calculation with real data for the third quarter of 2026, compares the rule with the actual fed funds rate from 1990 to 2026, and explains why the gaps are often large.

The formula

Each term has a job:

• r* + π keeps the real policy rate at its neutral level: if inflation is 3%, a 5% nominal rate is a 2% real rate.
• 0.5 × (π − π*) adds tightening when inflation is above target and easing when below. Combined with the π already in the formula, the rate moves 1.5 points for each point of inflation. Moving the nominal rate by more than inflation is what makes real rates rise as inflation rises.
• 0.5 × gap adds tightening when the economy runs above potential (a positive gap) and easing when it runs below.

Taylor's 1993 paper used the GDP deflator for inflation and the deviation of real GDP from a trend. The chart here substitutes core PCE inflation, closer to the Fed's 2% PCE goal, and the Congressional Budget Office's estimate of potential GDP. The Fed's own page on policy rules writes the formula the same way with a time-varying neutral rate, and describes a balanced-approach variant that doubles the weight on the output gap to 1.0.

Worked example: Q3 2026

Change one assumption and the answer moves a lot:

Every version sits well above the actual rate. That says the rule, given these inputs, would have called for tighter policy. It does not say why the FOMC chose otherwise; for the committee's reasoning, read its statements and minutes.

The rule against the actual rate, 1990–2026

Average gap between the rule and the actual rate (rule minus actual):

1990–2007: close on average, not quarter by quarter. Over 18 years the rule and the actual rate averaged out almost exactly, but individual years differed. In the late 1990s the actual rate ran above the rule (Q2 1998: rule 3.04%, actual 5.50%). In 2003–05 it ran below (Q1 2005: rule 4.82%, actual 2.47%). Q1 2007 was close: 5.09% versus 5.26%.

2009 and 2020: the zero lower bound. In Q1–Q3 2009 and Q2 2020 the rule prescribed a negative rate (−0.41% in Q2 2009, −1.96% in Q2 2020). The FOMC kept its range at 0% to 0.25% and turned to bond purchases instead; see Quantitative easing and tightening: the Fed’s balance sheet. A rule that ignores the lower bound cannot be followed there.

2015–2018: a slow climb under the rule. As the output gap closed, the rule rose to about 2.5% in 2015 and 4% in 2018, while the actual rate rose from 0.1% to 2.2%.

2021–2022: the widest gap. Core PCE inflation reached 5.56% for Q1 2022 with a positive output gap, and the rule called for 9.96%. The actual rate averaged 0.12%. Hikes began in March 2022.

2023–2026: closer, still below. As inflation eased and rates rose, the gap narrowed to 0.6–1.5 points in 2024, then widened to 2.6–2.9 points in 2026 as core PCE inflation rose back above 3% and the policy rate stayed in the mid-3s.

Over the full 147 quarters, the correlation between the rule and the actual rate is 0.52: the rule captures the direction of policy much better in some decades than in others.

Why the Fed does not follow it mechanically

The Fed's own explainer on policy rules lists the problems. In its words, the US economy is "complex and the understanding of it is incomplete" (Federal Reserve Board, policy rules page). In practice:

• Unobservable inputs. The neutral rate and potential GDP cannot be measured; they are estimated, with wide uncertainty, and they change over time.
• Revised data. GDP, potential GDP and inflation are revised for years. The chart uses today's data vintage, not what the FOMC saw at the time, so it can make past decisions look worse, or better, than they looked in real time.
• Choice of measures. Core or headline inflation? PCE or CPI? Output gap or unemployment gap? Each choice gives a different line.
• The lower bound. Rules can prescribe negative rates the Fed will not set.
• Gradualism and risk management. The Fed's page describes an inertial variant that puts 85% weight on last period's rate, reflecting a preference for moving in steps.
• Forward-looking policy. Rules react to current inflation; the FOMC acts on forecasts, financial conditions and risks the formula does not see.

How investors use it

• As a yardstick. When the actual rate is well below a range of rule prescriptions, analysts describe policy as accommodative relative to that benchmark, and vice versa.
• To frame debates. Disagreements about the "right" rate are often disagreements about r* or the output gap. The rule makes those assumptions explicit.
• With other guides. Compare it with the The Fed dot plot: how to read the FOMC’s rate projections, the market's expected path in 2-year yields (How Fed decisions move markets: expectations, surprises and yields) and the real policy rate.

Policy rules compared

How Finabase does it

Finabase does not calculate Taylor-rule prescriptions or display GDP, potential output or PCE data. The chart and every calculation on this page were made by Finabase Research from BEA, CBO and Federal Reserve data downloaded from FRED. In the app, the market's own estimate of where policy is heading shows up in short-term yields: the 2-year Treasury yield and the Treasury yields table. For news about Fed decisions, use Discover's Fed topic.

Frequently asked questions

What is the Taylor rule formula?

In its 1993 form: policy rate = 2 + inflation + 0.5 × (inflation − 2) + 0.5 × output gap, where 2 is both the assumed neutral real rate and the inflation target. The version charted here uses core PCE inflation and the CBO output gap.

Does the Fed follow the Taylor rule?

No. The FOMC consults several policy rules as benchmarks, but it does not follow any mechanically, because their inputs are uncertain and they ignore forecasts, risks and the zero lower bound.

What does the Taylor rule say now?

With Q3 2026 data, the 1993 rule as charted here prescribed 6.53%, against an actual quarterly average of 3.67%. With a 1.2% neutral rate the prescription is 5.73%.

What is the Taylor principle?

That the policy rate should rise by more than one-for-one with inflation, so the real rate rises when inflation rises. In the 1993 rule the coefficient on inflation is 1.5.

Summary

The Taylor rule turns two inputs, inflation and the output gap, into a suggested policy rate: 2 plus inflation plus half the inflation gap plus half the output gap. In 1990–2007 it matched the actual fed funds rate on average; since 2008 policy has mostly run below it, by almost 10 points in early 2022 and by 2.86 points in Q3 2026, when it prescribed 6.53% against an actual 3.67%. The prescription depends heavily on unobservable inputs, especially the neutral real rate and potential GDP, and on revised data. Treat it as a transparent benchmark for asking whether policy is tight or loose, not as a forecast of what the FOMC will do.