Uh oh: BLS preliminary recalculation sheds half-million jobs in 2018-19

Every year, the Bureau of Labor Statistics recalculates its benchmarks for employment calculations. It’s not every year, however, when those revisions chop a half-million jobs out of its previous estimates. In its preliminary calculations, the new benchmark does just that — although it remains to be seen whether those benchmarks will hold up for their final implementation next February:

Each year, the Current Employment Statistics (CES) survey employment estimates are benchmarked to comprehensive counts of employment for the month of March. These counts are derived from state unemployment insurance (UI) tax records that nearly all employers are required to file. For national CES employment series, the annual benchmark revisions over the last 10 years have averaged plus or minus two-tenths of one percent of total nonfarm employment. The preliminary estimate of the benchmark revision indicates a downward adjustment to March 2019 total nonfarm employment of -501,000 (-0.3 percent).

Preliminary benchmark revisions are calculated only for the month of March 2019 for the major industry sectors in table 1. The existing employment series are not updated with the release of the preliminary benchmark estimate. The data for all CES series will be updated when the final benchmark revision is issued.

The chart shows how broadly the retrenchment goes in the US economy. Nearly every industry appears to have had its employment levels overstated:

Keep in mind that this is only a proposed new benchmark. The BLS will continue working on it to tweak it for better accuracy, although this release shows that this is well advanced from the spitballing stage. The new preliminary calculations have some economists already assuming the figures are solid, which would be a mistake:

The economy had about 501,000 fewer jobs as of March 2019 than the Bureau of Labor Statistics initially calculated in its survey of business establishments. That’s the largest revision since the waning stages of the Great Recession in 2009.

The newly revised figures indicate the economy didn’t get a huge boost last year from President Trump’s tax cuts and higher federal spending. They also signal the economy is a bit weaker than previously believed and could give the Federal Reserve even greater reason to cut interest rates in September.

“This makes some sense, as the 223,000 average monthly increase in 2018 seemed too good to be true in light of how tight the labor market has become and how much trouble firms are said to be having finding qualified workers,” said chief economist Stephen Stanley of Amherst Pierpont Securities.

The average 223,000 monthly increase in employment in 2018 — the strongest in three years — could be trimmed to 180,000 to 185,000, economists estimate.

That’s certainly possible, but it’s not certain yet at all. There’s reason to think that this understates reality too, especially when it comes to wage growth. That has objectively accelerated over the last couple of years, which only makes sense if job growth was pitched high enough to put pressure on employers to increase compensation more rapidly.

That leads us to another point: the measures from BLS do not create reality but reflect it, as best as surveys can do, anyway. During this same period, US economic growth as measured by the Bureau of Economic Analysis ran higher than 3% annualized GDP growth in four of the last nine quarters. That itself would indicate a higher level of sustained job growth than the 180K level, although that would be an indirect measure of job growth, to be sure, about which more in a moment.

Still, the BLS release has begun to catch eyes, and one surprising set in particular:

It’s a legit headline, and a legit story — as long as the proper context is provided. This is an adjustment of measures rather than reality, and job growth is itself an indirect measure of economic growth. The GDP reports provide the direct measures of economic growth, and those have been reviewed repeatedly over the last two-plus years.

The bottom line: This doesn’t actually change anything about the economic reality of the moment. It’s not a massive job loss, but rather a recalculation of previous growth intended to better reflect that reality — and even that’s a preliminary change to a previous survey model, not the data itself.