The recession alarm goes off

US payrolls grew by just 114,000 in July and the unemployment rate jumped to 4.3 per cent, a nasty surprise for markets. The rise triggered the Sahm Rule, because the three-month moving average of unemployment is now 0.53 percentage points above its low of the past twelve months. Trading desks panicked.

The timing sharpened the reaction. Only two days earlier the Federal Reserve had left its policy rate at 5.25 to 5.5 per cent, where it has sat for a year, while signalling that a cut could come in September. The jobs report made that look late. Global markets were already fragile after the Bank of Japan's rate rise had set off a scramble out of yen-funded trades, and on Monday the selling spread from Tokyo to New York as investors priced in a sharp slowdown and much faster Fed cuts.

How the rule works

Claudia Sahm, then an economist at the Fed, devised the rule in 2019 as a trigger for automatic stimulus payments, not as a forecasting tool. Its logic is simple. Unemployment tends to rise gradually at first and then accelerate, because job losses reduce spending, which leads to more job losses. Once the three-month average has climbed half a point from its recent low, that feedback loop has historically been under way, and the rule has flagged every American recession since the 1970s.

Why the signal may mislead

The Sahm Rule has a perfect record of flagging recessions, but economists need to ask whether demographics are distorting it this time. Much of the rise in unemployment came from new entrants to the labour market, including immigrants, who have not yet found work. Large-scale layoffs played a smaller part. Even so, the speed of the slowdown suggests the Fed has waited too long to cut.

The immigration effect is large by historical standards. The Congressional Budget Office estimated that net immigration reached 3.3 million people last year, well above the pre-pandemic pace. A labour force that grows quickly can see unemployment rise even while employers keep hiring, because new job-seekers take time to be absorbed. That is a very different situation from the typical recession, in which the rise in unemployment comes from firms shedding existing workers. Ms Sahm herself has cautioned that her rule may be overstating the danger this time for exactly that reason.

Signs that it is real

Other evidence does not fully support the benign reading. Payroll growth has slowed from the pace of last year, and the rate at which firms hire has fallen below its pre-pandemic level. Job openings have come down steadily. Initial claims for unemployment benefits have risen, though they remain low by historical standards. Temporary layoffs also increased in July, which may partly reflect Hurricane Beryl's disruption in Texas, but they are often the first category to move when firms begin to cut.

Put together, the picture is of a labour market cooling faster than the Fed expected. Whether that slide becomes a recession depends on whether the cooling stops at lower hiring or moves on to widespread layoffs, and the Sahm Rule cannot distinguish between the two.

The bond market had been warning for some time. The gap between two-year and ten-year Treasury yields has been inverted since the middle of 2022, the longest such stretch on record, and inversions have preceded most recent recessions. After Friday's report the two-year yield fell faster than the ten-year, pushing the curve close to turning positive again, a move that in past cycles has tended to come shortly before or just after a downturn begins.

Precedents for a late Fed

History suggests the risk lies in moving too slowly. In 2001 and again in 2007 the Fed began cutting only a few months before recessions started, once the labour market had already turned, and the downturns arrived anyway. In 1995, by contrast, Alan Greenspan cut rates early after a year of tightening, and the expansion lasted another six years. In 2019 Jerome Powell himself made three so-called insurance cuts to guard against a slowdown driven by the trade war with China. Those episodes argue for moving before the data turn clearly bad.

The opposite precedent is 1998, when the Fed cut between meetings after markets seized up following the collapse of Long-Term Capital Management. Traders are now asking for something similar. Emergency cuts, however, are usually reserved for broken markets, not for a single weak jobs report, and a cut between meetings would risk signalling more alarm than the data justify.

Who gains, who loses

Short-term Treasuries have rallied hard, and two-year yields have fallen sharply as markets priced in bigger and faster cuts. Borrowers with floating-rate debt stand to gain once cuts arrive. Savers in money-market funds will see yields fall. Companies whose revenues depend on consumer spending face a weaker outlook, and cyclical shares have fallen further than defensive ones.

What to watch

The next pieces of evidence arrive quickly. Weekly jobless claims will show whether layoffs are spreading. The annual benchmark revision to payrolls, due later this month, may show that job growth over the past year was weaker than first reported. The Fed's annual Jackson Hole symposium in late August is the natural place for Mr Powell to set out how far and how fast the Fed intends to move.

The Sahm Rule has put the Fed on notice. Even if immigration blurs the signal, the labour market is starting to crack under high rates, and the Fed will probably have to start cutting at its next meeting. Expect a cut in September, possibly larger than the usual quarter-point, with floating-rate borrowers the first to benefit and new job-seekers bearing the cost of the Fed's delay in the meantime.