U-6 vs the Unemployment Rate, a thread:

1/ On Friday, I saw several folks lean on an expanded measure of underemployment (U-6) to argue that the jobs report was "worse than it looked".

I'd characterize this as "true, but not in the business cycle sense".

More downthread. 2/ Some definitions first.

U-3 is the standard unemployment rate. It defines unemployed as "doesn't have a job, but is actively looking in the past month" (or temporarily furloughed and expect to be recalled to work)

The final "scrapped together" claims data before the DOL begins publishing weekly reports again...

1/ Regular initial claims were "well-behaved" again in the week ended 11/08, no sign of a layoff surge, just modestly above year-ago levels (and nearly identical to 2 years ago) 2/ Maybe something bad is around the corner but it hasn't happened yet.

Next week's IC data will be the last to include part of the shutdown period.

Claims:

1/ Continuing claims running about 15K above my benchmark.

There's residual seasonality in the published series that paints an overly rosy picture of what is going on with this data... 2/ There's been a mild reacceleration in the CC data recently. We'll see if it goes anywhere; the May/June mild acceleration petered out.

1/ I've been curious how to reconcile the relatively upbeat "flows" from the CPS (hiring & quits rising modestly; layoffs falling modestly) with the downbeat stocks (esp upward-creeping unemployment rate)

https://twitter.com/EconBerger/status/1968709253654573229

2/ Out of the @mander_michaud / @KathrinPhD data, we see that job search after layoff has gone up non-trivially over the past year, but job search after quits has increased by less.

Still digesting, but my instinct is to think this is (on net) a positive development?

Some surprisingly upbeat news out of the CPS flows data for August (even as the CPS "stock" data was downbeat)...

1/ Hiring out of non-employment has stopped falling in recent months. May even be rising a little. 2/ We also have, courtesy of @mander_michaud and @KathrinPhD , measures of quits & layoffs into non-employment.

Neither of these are getting worse at present. As with hiring, may actually be improving!

Claims:

1/ An extremely concerning jump in initial claims to 263K (highest since the fall of 2021). This data includes Labor Day weekend, so maybe it's holiday noise, but I'm worried. 2/ It's well above last year's level.

BLS charts:

1/ Another bad jobs report. Weak job gains, and most concerningly the unemployment rate inches up further (now just a hair above 4.3%). 2/ On the plus side... prime working age employment population ratio, the flagship indicator of this report rises to 80.7%. No sign of deterioration in this metric. Steady since last fall.

July JOLTS:

1/ Some mild softening in hiring over the past 2 months. Too early to tell whether it is noise or renewed labor market cooling.

Quits pretty steady. 3/ The hiring rate, at 3.3%, is around where it was for much of 2011-13 (or if you like an average unemployment rate for that period, just over 8%). It's a really hard time to find a job.

1/ Interesting tidbits on the labor market from @uscensusbureau 's latest Business Trends & Outlook Survey - which indicates the US labor market may be warming up.

Thread... 2a/ First off are recent employer actions on headcount.

For the first time in the history of this survey (which is less than 2 years), we've had 4 consecutive weeks where actions have been more expansionary (as measured by a diffusion index) than a year earlier.

Claims:

1/ Continuing claims modestly above my benchmark projection for the 2nd week in a row.

The published series should be flattish for the rest of the year (a misleadingly upbeat portrayal of the underlying trend). 2/ The underlying trend is mid-single-digits Y/Y growth - a tiny bit better than 2 months ago, a tiny bit worse than earlier this year.

There's no sign of a sharp deceleration in the economy, nor is there any sign of acceleration.

Claims:

1/ Good news. The continuing claims scare of May/June faded further... continuing claims fell below my benchmark (which assumes a ~4% Y/Y increase).

They're likely to be (misleadingly) flattish for the remainder of the year, due to residual seasonality. 2/ Another way of seeing this: continuing claims were up about 4% Y/Y - comparable to before the scare.

That's still not *great*, reflecting a very mild ongoing deterioration. But that's not what got us worried two months ago.

JOLTS: 1/ Hiring was a little on the soft side in June, but generally the main turnover indicators have been stable for some time. Cooling has been on pause since summer/fall 2024 and limited evidence that it has resumed. 2/ As stated earlier, hiring was on the soft side. A hiring rate of 3.3% is comparable to early 2012 levels, when the unemployment rate was a little over 8%. But there's a pretty good chance it's just noise around a stable, weak hiring trend.

Claims:

1/ Continuing claims a bit higher than I thought they would be, though still not that concerning.

We're probably going to be on a flattish trajectory in the published series for the remainder of the year due to residual seasonality. 2/ That flattish trajectory in the published series is an overly benign portrayal of what is actually happening in the underlying data, which is an ongoing slow creep upward.

1/ As with nonfarm payroll employment from the BLS, the breakeven monthly rate for ADP is much lower than it was a year ago.

Worth keeping in mind when seeing numbers like this…

https://twitter.com/nicktimiraos/status/1940387643390558393

2/ Without taking future revisions into account, we’ve had a similar trajectory to the unemployment rate in the first 5 months of this year with 124K/mo (vs 180K/mo a year earlier.)

Consider future revisions into account and we’re probably talking 154K vs 59K.

BLS charts:

1/ An increase of 139K in May, comparable to gains in the last two months (which were revised down).

My speculative hypothesis based on QCEW data through December 2024: after the eventual benchmark revision, these numbers will be closer to 70K/month. 2/ The unemployment rate is slowly creeping up, right along the track the Fed anticipated in March.

Was 4.244%, just shy of an unfavorable rounding.

Business Trends & Outlook Survey from the Census Bureau:

1/ Employers are becoming slightly less pessimistic, relative to a year earlier, about future headcount (probably reflecting less-high tariffs).

We haven't seen actual behavior (modestly contractionary) change much since 2/ The improvement in headcount plans (relative to a year ago) is coming from fewer firms planning to cut headcount, and more firms planning to expand it.

Again, they are more pessimistic than 5-6 months ago - but less pessimistic than 1-2 months ago.

Quick notes on the Q4 QCEW release this morning:

1/ TL;DR: We are very likely going to get large negative downward revisions to the growth of non-farm employment in the year from March 2024 to March 2025. 2/ To recap: the QCEW is the source data for the annual benchmark revision to nonfarm payroll employment.

The next benchmark date is March 2025. We'll get a preliminary estimate in late summer. We'll get the final benchmark revision in early February 2026.

Business hiring plans:

1/ With the reduction in US tariffs on China (from astronomical to very high levels), business hiring plans have become a little less downbeat. (Still more glum than a year ago, but the deficit is smaller.) 2/ This modest improvement in plans is due to more firms planning to increase headcount, and fewer firms planning to cut back.

But it's still true that fewer firms plan to expand than a year ago, and more firms plan to cut back than a year ago.

US employer sentiment update:

1/ US firms continue to be very pessimistic about their future headcount, but no change in that pessimism relative to 2 weeks ago.

We're not seeing any leakage, yet, from that pessimism into current action. (Soft data & hard data agree.) 2/ The deterioration in employer headcount expectations has come from both a falling share of firms planning to expand, and a rising share of firms planning to cut employment.

(But no additional worsening in the last 2 weeks.)

Thread on “hard data” vs “soft data”:

1/ Current conventional wisdom: the split between resilient “hard data” (e.g. government-published series on employment & activity) and weak “soft data” (sentiment surveys).

IMHO this CW is mostly wrong. 2/ Why wrong?

Survey/soft data that corresponds to concrete things people see at present are not that far from the hard data.

Questions about vague vibes, or about the future, look much worse.

Composite headline indices look bad because they mix "present" with vibes/future.

Jobs Report Chart Thread:

1/ Let's start with the BLS disclaimer on the report, regarding hurricane impact. 2/ Nonfarm payrolls were barely positive because of this hurricane impact (and also the Boeing strike).

Remember, an employed person who earns no money because they are absent from work counts as employed in the household survey, but non-employed in the establishment survey.