September was another month of slow jobs growth, the government reports. Private payrolls increased 104,000 last month, according to the Labor Department’s establishment survey. (Total payrolls, which include government jobs, rose 114,000.) That’s up from August’s revised gain of 97,000 for the private sector, but no one will be impressed with these numbers. Nonetheless, there’s no smoking gun here for arguing that the labor market is collapsing. It may be suffering a slow and lingering death, in which case the last rites may be administered at some point down the road. In the here and now, however, private-sector job growth of 104,000 doesn’t signal that the economy’s in recession, even if it’s not hard to envision a change for the worse in the near-term future. Nor does the 1.7% year-over-year growth in private payrolls through last month tell us that the jig is up. It’d be another matter if September data so far from other sources screamed of free-fall. But that’s not the case either (see here and here, for instance).

Meantime, a curiosity worth mentioning: A separate employment survey from the Labor Department shows a much-brighter picture for September. The so-called household survey, which is calculated with a separate methodology, is used for estimating the monthly unemployment rate. The official jobless rate for the nation reportedly fell to 7.8% last month vs. 8.1% in August. Why? The household survey reports that employment rose an incredible 873,000 last month–the best month in nearly a decade. Huh?!?! That’s almost certainly overstating the case, and dramatically so, which is probably why most economists tend to focus on the establishment survey. Indeed, the household numbers are far more volatile on a monthly basis vs. the establishment data. (For some details on how the two series compare, see the Labor Department’s explanation here.)
On that note, let’s review the history of private payrolls via the establishment numbers in recent years for some context with today’s update. As the chart below shows, the labor market is still expanding. It’s not impressive (notwithstanding the household numbers), nor anywhere near sufficient to inspire much confidence about the longer-term economic outlook. But if we’re looking for signs that September will eventually be labeled the start of a new recession, as defined by NBER, the news in today’s payrolls report isn’t dark enough.

It’s important to distinguish between what’s increasingly looking like a chronic illness for growth vs. nowcasting the current state of the business cycle. The key points that everyone agrees on: 1) economic growth decelerated over the last six months or so; and 2) the all-important pace of job growth has suffered too. The obvious warning signs in this process can’t be ignored, namely, slow growth at some point, if it drags on long enough, will bring us to a new recession.
When the tipping point is reached, however, it’ll be obvious in the data. That includes payrolls, but even this key indicator can’t be relied on for issuing timely signals about the state of the business cycle. That requires a deeper dive into the data, which is the inspiration behind The Capital Spectator Economic Trend Index (CS-ETI). When we last checked in with CS-ETI for the read through August, the odds that a new downturn had started remained low. The numbers so far for September don’t change that view, although it’s still early and today’s jobs report reminds that we should be prepared for weaker-than-expected news in other September indicators as they arrive in the coming weeks.
Pessimists will point out that data revisions will darken the cyclical skies and so profiles a la CS-ETI are bound to be misleading. Nice try, but the methodology behind CS-ETI minimizes revision risk on several fronts. First, CS-ETI measures a broad array of leading and coincident indicators so that we’re not relying on the kindness of the revision fairy for any one data point. That’s crucial because revisions are both positive and negative through time across a broad data set. By looking at a mix of indicators, the revisions to some degree will cancel one another out.
Another defense against revisions in CS-ETI is tracking mostly year-over-year percentage changes. That’s important because the year-earlier figures are less likely to be revised dramatically, if at all. Year-over-year changes also provide us a clearer read on the cycle compared with trying to see through the seasonal distortion that routinely harasses in shorter time frames.
In addition, CS-ETI uses several indicators that are immune to revisions, i.e., the market data for stocks, oil prices, and interest rates. And the surveys for manufacturing and services via the Institute for Supply Management, along with the consumer sentiment survey from the University of Michigan, have a history of minimal revisions vs. conventional economic numbers.
Finally, to round out this point, I’ve analyzed the vintage data for the last 10 years and CS-ETI’s signals hold up quite well. In other words, revision risk isn’t likely to blind us for very long or very deeply, thanks to the design of CS-ETI.
All that aside, there’s no denying that economic growth remains weak and is perhaps getting weaker. In terms of a binary analysis—growth or contraction?—the overall summary is that recession risk is still low. A qualitative analysis may tell us otherwise, of course, such as looking at a nowcast of third-quarter GDP, based on the latest numbers (here’s last week’s update). I’ll have an update on that front soon, along with fresh numbers for CS-ETI.
Meantime, the sluggish expansion drags on. It may not be improving, but it’s not necessarily getting worse either. We are, it seems, still caught in economic purgatory, somewhere between heaven and hell.