The recession rule that's flashing amber

After the October jobs report the Sahm Rule indicator reached 0.33 percentage points, an amber warning. Devised by Claudia Sahm, the rule says a recession is under way once the three-month moving average of the national unemployment rate rises 0.5 percentage points above its low of the previous twelve months. Economic forecasting is full of unreliable rules of thumb, but monetary economists treat this one with unusual respect.

How the rule is built

Sahm, then an economist at the Federal Reserve, proposed the rule in 2019 as a trigger for automatic stimulus payments to households, so that support would arrive early in a downturn without waiting for Congress. Its appeal lies in simplicity. It uses one widely published number, smooths it over three months to remove noise, and compares it with its own recent low. Since 1970 it has signalled every American recession, usually within a few months of its start, with very few false alarms.

The current reading is easy to reproduce. Unemployment was 3.8 per cent in August and September and 3.9 per cent in October, for a three-month average of about 3.83 per cent. The lowest three-month average over the previous year was 3.5 per cent, set early this year when the jobless rate touched 3.4 per cent, a level last seen in the late 1960s. The gap between the two is the 0.33 points now on the gauge.

Unemployment starts to climb

Payroll growth slowed to 150,000 and the unemployment rate rose to 3.9 per cent. Once unemployment starts rising it tends to keep going, because companies cut back and households grow cautious, and each feeds the other. If the rule is triggered, it will suggest that the Federal Reserve's rapid tightening has finally tipped the economy over.

The details of the report pointed the same way. Payroll gains for the previous two months were revised down, and wage growth eased, with average hourly earnings rising only 0.2 per cent on the month. Part of the weakness in manufacturing reflected the car workers' strike against the Detroit carmakers, which has since been settled, so some of those jobs will return. The broader message is that hiring has slowed from the brisk pace of the summer.

Why rising unemployment snowballs

The mechanism behind the rule is a feedback loop. When firms stop hiring and start trimming staff, the newly unemployed cut spending, which lowers demand for other firms' products, which leads to further cuts. Households who still have jobs grow cautious and save more. Banks tighten lending as they anticipate higher defaults. That is why unemployment rarely rises by a small amount and stops; in post-war American history, increases of half a point in the jobless rate have almost always been followed by much larger ones.

Monetary policy works with a lag, which strengthens the case for concern. The Fed raised its policy rate from near zero in March 2022 to a range of 5.25 to 5.50 per cent by July this year, and held it there at its meeting on 1 November. Much of that tightening has yet to feed fully through to corporate borrowing costs, as fixed-rate debt taken on at low rates matures and is refinanced. Long-term Treasury yields came close to 5 per cent in October, adding to the squeeze.

The case for a false alarm

There is a reasonable argument that this time the rule may mislead. Unemployment can rise because more people are looking for work as well as because firms are shedding jobs. Labour-force participation has recovered strongly since the pandemic and immigration has picked up, so part of the increase in the jobless rate may reflect a larger supply of workers finding jobs more slowly, which is far less dangerous than layoffs. Weekly claims for unemployment benefits remain low by historical standards, and layoff announcements have not surged.

That argument has force, and it is why the reading is amber rather than red. Rules built on past recessions can fail when the structure of the economy shifts, and the pandemic scrambled many of the usual relationships. The rule was also not designed as a forecast; it tells observers that a recession has probably started, not that one is coming. Even so, a rising unemployment rate at a time when policy is at its tightest in two decades deserves more weight than the soft-landing consensus gives it.

Bond markets already lean this way. Futures prices imply that the Fed has finished raising rates and will start cutting next year, and Treasury yields fell after the jobs report as traders brought those bets forward. A trigger of the rule would speed up that repricing, pushing short-dated yields down fastest and steepening a yield curve that has been inverted since last year.

What would trip it

The arithmetic is not demanding. If unemployment rose to 4.1 or 4.2 per cent over the next few months and stayed there, the three-month average would climb above 4.0 per cent, enough to cross the threshold of half a point above this year's low. The indicators to watch are the November and December jobs reports, weekly jobless claims, which would signal rising layoffs before the monthly data do, and hiring of temporary workers, which tends to fall early in downturns.

The amber reading should worry the Fed. Labour markets tend to turn slowly and then all at once, and another few months of rising unemployment would probably be enough to trip the rule, putting rate cuts on the table by the middle of next year.