Sahm Rule: The Unemployment Threshold for Recession Identification

The Sahm Rule identifies recessions in real time using a simple threshold: when the three-month average unemployment rate rises 0.5 percentage points or more above its minimum level over the prior 12 months, a recession has begun. Developed by economist Claudia Sahm, the rule has caught every recession since 1990 with few false alarms.

The Rule in Plain Terms

The Sahm Rule answers a deceptively simple question: has unemployment spiked sharply and recently? If yes, a recession is underway.

The math is straightforward. Every month, the Bureau of Labor Statistics releases the unemployment rate for the prior month. The rule takes:

1. The unemployment rate for the most recent month and the two preceding months, then averages them (the three-month moving average).
2. The lowest unemployment rate recorded in the prior 12 months.
3. If the three-month average exceeds that minimum by 0.5 percentage points or more, the rule signals recession.

Example: Suppose in February 2024, the 12-month minimum unemployment rate was 3.8% (reached in March 2023). The three-month average for December, January, and February 2024 is 4.4%. The difference is 4.4% – 3.8% = 0.6 percentage points. Since 0.6 > 0.5, the rule signals. A recession has begun.

The threshold of 0.5 percentage points is intentionally small because unemployment doesn’t need to skyrocket to signal trouble. The rule is sensitive, not hair-trigger.

Why This Signal Works

Unemployment is a lagging indicator—it rises after an economy enters recession, not before. The National Bureau of Economic Research (the official arbiter of recession dates) often dates recession starts months before unemployment data reveals the damage.

But lagging indicators are still useful for real-time identification. By the time unemployment ticks up measurably, the recession is underway. The Sahm Rule, by measuring the speed of change, catches this transition faster than simply waiting for unemployment to reach some arbitrary level.

The 0.5 percentage point threshold is statistically robust. Claudia Sahm analyzed decades of labor market data and found that a 0.5 percentage point jump in the three-month average, relative to the recent minimum, has never occurred outside a recession (historically). The rule captured the 2001 recession, 2007–2009 financial crisis, and 2020 COVID recession with no false positives during the expansions of 1991–2000, 2003–2007, and 2010–2019.

The three-month moving average smooths monthly noise. A single month of weak hiring can jolt the unemployment rate, but that doesn’t mean recession is here. Averaging three months filters these blips and reveals underlying trend.

When the Rule Signals (and Doesn’t)

The Sahm Rule typically signals a recession a few months after it has technically begun. In the 2007–2009 recession, the National Bureau of Economic Research dates the start as December 2007. The Sahm Rule first signaled in June 2008, a six-month lag. In the 2020 recession (which lasted only two months), the rule signaled by April 2020, roughly one month after the February start.

This lag is acceptable because the rule is designed for certainty, not speed. Policymakers and economists care more about knowing they’re in recession with high confidence than being first.

The rule’s strength is in filtering false alarms. A single month of weak job creation (perhaps seasonal noise or a statistical revision) doesn’t trigger the rule. Only a sustained rise in unemployment crossing the threshold counts. This robustness made the rule popular among forecasters who’d been burned by leading indicators that fired on every hiccup.

In 2023–2024, the rule proved its value again: despite predictions of imminent recession, unemployment remained tight, and the rule never signaled. This non-signal was itself information—a signal that recession had not begun, even as tight credit conditions and higher interest rates raised fears.

However, in July 2023, the three-month average briefly approached the threshold due to normal monthly volatility, reminding observers that even robust rules have edge cases.

Relationship to the Business Cycle

A recession, by the National Bureau of Economic Research definition, is a significant decline in economic activity spread across the economy, lasting more than a few months. Unemployment captures one dimension: the labor market’s response. When firms stop hiring and begin laying off, unemployment rises sharply and measurably.

The Sahm Rule doesn’t measure inflation, investment, or industrial production—only labor market tightness. It’s thus a narrow signal, not a complete recession detector. But it’s reliable in a way broader, multi-factor indices are not.

The rule assumes that monetary policy and business cycles play out through the labor market with predictable timing. Interest-rate increases typically cool hiring within 6–12 months, then unemployment rises. The Sahm Rule captures the moment that cooling translates into measurable job losses.

Practical Use by Policymakers and Investors

The Federal Reserve and Treasury Department use the Sahm Rule as one input when assessing whether monetary or fiscal stimulus is warranted. Once the rule signals, policymakers know a recession is here and may consider quantitative easing or emergency spending programs.

Equity investors use the rule to confirm recession fears. When unemployment spikes and the Sahm Rule triggers, it’s typically a signal to reassess portfolio risk; earnings growth is likely to slow, and valuations may compress.

Bond traders view a Sahm signal as a floor under interest rates. A recession and rising unemployment typically prompt the Federal Reserve to cut rates, boosting bond prices. This dynamic has held reliably: most Sahm signals have coincided with rate-cutting cycles.

The rule is less useful for predicting how long a recession will last or how deep. A 0.5 percentage point spike might precede a shallow two-quarter dip, or it might be the start of a deeper, longer contraction. Other indicators (forward-guidance statements from the Federal Reserve, credit spreads, initial jobless claims) help investors refine that judgment.

Limitations and Edge Cases

The Sahm Rule is not perfect. It’s a univariate signal—it looks only at unemployment, ignoring wages, hours worked, and labor force participation. In some weak expansions, unemployment might tick up slightly due to demographic shifts or rising labor force participation, not recession.

The rule has also changed its track record slightly. In August 2022, with inflation running hot and the Federal Reserve aggressively raising rates, the rule briefly flashed during an expansion—though no recession materialized. Some analysts argue that in an era of low trend unemployment (around 3.5%), the rule’s 0.5 percentage point threshold might be too sensitive.

Additionally, the rule relies on accurate, timely unemployment data. Data revisions, which the Bureau of Labor Statistics makes each month for the prior months, can change the signal retroactively. A rule that once signaled might be revised away, or vice versa.

Finally, in fast, shallow recessions (like 2020), the lag can be long enough that policymakers have already begun stimulus before the rule confirms recession, making it less actionable for crisis response.

See also