What Is the Sahm Rule? Formula, Trigger, and Limits

The Sahm Rule tracks a change in the U.S. national unemployment rate, or U-3. It uses the seasonally adjusted rate from the Current Population Survey (CPS), a household survey. U-3 is the number of people classified as unemployed divided by the civilian labor force. Under the BLS CPS definitions, people generally must be available for work and have actively searched in the previous four weeks; people on temporary layoff expecting recall are an exception. The Sahm Rule does not use U-6, unemployment-insurance claims, payroll growth, or the number of job openings.

The indicator looks for a sustained increase relative to a recent unemployment low. It is expressed in percentage points, not as a percentage change in the unemployment rate. FRED's real-time Sahm series describes the formula and tracks monthly readings using data available at each point in time.

First calculate the average of the latest three monthly U-3 readings. Then subtract the lowest three-month average from the previous 12 months:

Sahm indicator = current three-month average − lowest three-month average in the previous 12 months

A value at or above 0.50 percentage points meets the trigger. For example, a change from 3.8% to 4.3% is 0.50 percentage points. It is not a 0.50% increase in the unemployment rate, and the comparison is not against the lowest single monthly reading. The FRED series definition specifies the moving-average comparison and threshold.

The rule uses the U.S. national, seasonally adjusted U-3 series. A calculation using a state rate, U-6, or a non-seasonally adjusted series would not be the same indicator. Keep the rate's unit, geography, and adjustment status attached to any calculation.

Suppose the three latest seasonally adjusted U-3 readings are 4.2%, 4.3%, and 4.4%. Their average is 4.30%: (4.2 + 4.3 + 4.4) ÷ 3. Now suppose the lowest three-month average in the preceding 12-month comparison period is 3.80%. The indicator is 4.30% − 3.80% = 0.50 percentage points, exactly at the trigger.

Both the monthly readings and the comparison low in this example are invented for teaching. They are not current BLS observations, a historical episode, or a forecast. In an actual reading, use the same release vintage and the documented U-3 series throughout the calculation.

The example also shows why a percentage-point difference matters. If the current average were 4.29% against 3.80%, the difference would be 0.49 points and below the threshold. Rounding a published value can make a near-threshold result look different, so use the underlying series precision when reproducing a calculation.

A monthly unemployment reading can move for reasons that do not describe a lasting change in the labor market. Averaging three months dampens some of that month-to-month noise. The reference low is also an average, so the rule compares like with like instead of contrasting a smoothed current measure with a single-month trough.

The 12-month window makes the benchmark recent. It lets the trigger respond to a rise from a low reached during the last year rather than from a fixed unemployment target. As the window rolls forward, the benchmark can change. The rule therefore answers a narrow question about how far unemployment has risen from a recent smoothed low; it does not estimate the economy's long-run sustainable unemployment rate.

FRED publishes a real-time series, SAHMREALTIME, and a current-data series, SAHMCURRENT. SAHMREALTIME uses the unemployment rate and recent history that were available in each month. SAHMCURRENT applies the same formula to the historical values available now. The series can differ when seasonal factors are revised, so a present-day reconstruction is not always the same as the reading that could have been observed at the time.

The BLS seasonal-adjustment methodology explains the annual review of CPS seasonal factors. FRED notes that the unemployment rate's seasonal-factor updates can change estimates in recent years. When describing what the trigger showed in a particular month, identify the real-time series and release vintage. When recalculating with today's data, label it as a current-data result and do not describe it as what a reader knew then.

The trigger signals that the smoothed U-3 rate has risen materially from a recent low. It is evidence of labor-market deterioration and can prompt closer attention to recession risk. It does not prove that every part of the economy is contracting, identify the cause of the increase, or guarantee that a recession will follow. It is based on an unemployment measure that can respond after labor conditions have already weakened, so it should not be presented as a long-lead forecast.

The Sahm Rule is also separate from the National Bureau of Economic Research's business-cycle dating. The NBER dates U.S. peaks and troughs retrospectively using a broad assessment of economic activity, not this single threshold. Its business-cycle dating FAQ describes that process. Crossing the Sahm threshold does not itself amount to an NBER announcement or an official legal declaration of recession.

Claudia Sahm proposed the trigger in a 2019 Hamilton Project policy proposal. The proposal used the unemployment condition as a possible trigger for direct payments during downturns, with design details intended to support a faster response. That policy context explains why the indicator is sometimes discussed alongside automatic stabilizers.

A threshold in an analytical proposal does not authorize a payment by itself. A government program needs legal authority, funding, eligibility rules, and an operating mechanism. The Sahm Rule can be studied as an indicator or considered in policy design, but do not infer that crossing it automatically sends money or changes benefit eligibility.

Start by naming the series, geography, period, and data vintage. State whether the value is SAHMREALTIME or SAHMCURRENT, identify the latest three months, and show the rolling comparison low. Keep the result in percentage points and distinguish exactly 0.50 from values above or below it.

Then compare the signal with other evidence: payroll employment, hours worked, output, income, and other labor-market measures. The Sahm Rule compresses one pattern in U-3 into a single number; it cannot replace a broader diagnosis or establish what caused a change. The guides to soft landing versus recession, the Beveridge curve, and U-3 versus U-6 cover related questions from different angles.

Finally, separate a signal from a forecast and a policy decision. Describe the series and release date before discussing whether the threshold was met. Do not treat one rounded observation as a guarantee, a formal recession date, or a promise about what policymakers will do.