Two weeks ago, jobless claims in the US ostensibly spiked.

I say “ostensibly” as though the increase wasn’t real, and I suppose I should apologize already: It was surely real for somebody. Somebody — many several somebodies — filed in the week to October 5, which is to say folks were jobless. And every involuntarily jobless American is in some way a policy failure and a human tragedy.

But from a 30,000-foot, macro analysis perspective, it was farcical to suggest that one anomalous weekly claims print should take precedence at a time when the rest of the incoming data plainly suggested the Fed would be wise to take a breath and be patient about implementing additional rate cuts on the heels of September’s “recalibration.”

As I wrote at the time, the big jump in claims earlier this month was at least in part attributable to extreme weather, and it was all but guaranteed to reverse in the weeks ahead. Fast forward three weeks and guess what? That spike is now gone.

Initial US jobless claims were 227,000 in the week to October 19, Thursday’s release showed. That was down 15,000 from the prior week.

Consensus was — and try not to laugh — 242,000. Another day at the office for professional economists: Six figures a year to be dead-ass wrong habitually. It’s good work if you can get it! Like portfolio management: All the underperformance money can buy.

As the chart shows, the initial filers series is now virtually identical to where it stood prior to October 10’s “game-changing” “inflection” (note the scare quotes). The four-week moving average obviously drifted up because of “the maths,” but October 10’s anomaly will drop from that calculation early next month.

Continuing claims in the week to October 12 were 1.897 million. That was more than expected, and a new “since November of 2021” high, but you’re going to need more than that for a convincing recession narrative.

Barring a run of soft top-tier data next week (entirely possible!), the Fed will either have to pause rate cuts in November or else once again prove they’re only “data-dependent” when the data conforms to their preferred policy stance.