What is the Sahm rule?

The Sahm rule is a recession indicator developed by economist Claudia Sahm in 2019. It triggers when the three-month moving average of the national unemployment rate rises by 0.50 percentage points or more relative to its lowest reading over the prior 12 months. Every US recession since 1970 has been preceded by a Sahm rule trigger, making it one of the most reliable and timely recession indicators available.

The logic behind the Sahm rule is rooted in the nonlinear dynamics of the labor market. Small increases in unemployment tend to be temporary and self-correcting, but once the unemployment rate begins rising at a certain pace, it generates feedback loops (reduced consumer spending, business caution, layoffs begetting more layoffs) that are difficult to reverse without a recession. The 0.50 percentage point threshold captures this tipping point.

What makes the Sahm rule valuable compared to other recession indicators is its timeliness. The yield curve may invert 12-18 months before a recession begins, and GDP data is released with a long lag and subject to major revisions. The Sahm rule uses real-time unemployment data that is available monthly and is rarely revised significantly, providing an early warning within the first few months of a downturn.

The indicator is published in real time by the Federal Reserve Bank of St. Louis (FRED series SAHMREALTIME). When the reading is well below 0.50, recession risk is low. As it approaches 0.30-0.40, markets begin watching closely. A breach of 0.50 has historically meant a recession is already underway or imminent.

Claudia Sahm originally developed the rule not primarily as a forecasting tool but as a trigger for automatic fiscal stabilizers. Her proposal was that stimulus payments should be sent automatically to households when the Sahm rule triggers, providing countercyclical support without waiting for legislative action. This policy application highlights the indicator's reliability, as automatic mechanisms require a high bar of confidence.