Mark Zandi runs the numbers at Moody's Analytics, and this week he put a probability-flavored verdict on them. The risks, he said, "are increasingly high that we're going into recession. We're not there yet - and maybe this thing gets turned around," though he added that turning it around is "increasingly becoming hard to do with each passing week" [1]. That is an opinion about the future, and this desk does not grade opinions about the future true or false. It is worth setting beside the record that already exists.
The steel-man is strong. Payrolls are a leading indicator, and the leading edge just bent down. Hiring does not usually stall in a vacuum; when it goes, confidence and spending can follow. Zandi is not guessing at a vibe, he is reading a real deterioration in the labor data and refusing to pretend it is noise.
Here is the record he is reading. US employers shed 23,000 jobs in July, following a downwardly revised 20,000 gain in June, and revisions to May and June left payrolls 103,000 lower than first reported [2]. That is the genuinely bad number. The unemployment rate, though, sat at 4.1% in July [2], and the last full read on output showed real GDP growing at a 1.5% annual rate in the second quarter, with consumer spending rising across both goods and services [3]. A shrinking payroll count next to still-positive growth and low unemployment is a mixed picture, not a contraction.
Two other economists in the same piece read the same data and declined to telegraph a downturn. Mark Blyth of Brown University: "It seems to me people are anchoring on the jobs report and using it to telegraph a recession. I'm not sure it does that" [1]. Harry Holzer of Georgetown put a timestamp on the humility: "It's too early to see whether this is a trend. The likelihood of a recession went up because of these job numbers, but it could be a one-time adjustment or a bump down that avoids negative growth" [1].
Claudia Sahm, whose name is on the recession indicator everyone cites, made the case for not pretending anyone can call this. "Recessions tend to be unforecastable," she said. "Often there's an event that causes people to lose confidence, change behavior and start a downward spiral" [1]. Note what that sentence is really describing: the recession as a thing people can talk themselves into.
One housekeeping conflict, because the numbers in circulation do not match. ABC's own account puts unemployment at 4.2% and three-month average hiring near 35,000 jobs [1]; the Bureau of Labor Statistics print for July shows 4.1% and the outright 23,000-job loss [2]. The likeliest reconciliation is timing, with ABC's 4.2% reflecting the June reading rather than July. We are not going to average them into a single tidy number, because they are measuring different months.
Which brings the stakes back to Sahm's spiral. A worker who reads the word "recession," decides the sky is falling, and cancels a car repair or a vacation is not insulating themselves from a downturn, they are supplying one of its ingredients. The July payroll loss is real and worth watching. The forecast built on top of it is a forecast, and treating the loudest version of it as already true is how a bump down becomes the thing it predicted.
