July Was Another Jobs Recession Month

Behind The BLS' Garbage Data Is A Picture Of Job Loss, Not Job Growth

Another month, another Employment Situation Summary report from the Bureau of Lying Labor Statistics.

Another month, another report built around the phrase “little change” which obscures job loss and increased joblessness.

Another month, another dose of Lou Costello Labor Math.

Total nonfarm payroll employment changed little in July (+73,000) and has shown little change since April, the U.S. Bureau of Labor Statistics (BLS) reported today. The unemployment rate, at 4.2 percent, also changed little in July. Employment continued to trend up in health care and in social assistance. Federal government continued to lose jobs.

Update 2 (2245CDT 1 August 2025)

In firing Dr. McEntarfer, President Trump leveled a fairly drastic allegation against her: that the BLS was intentionally mis-reporting the jobs numbers.

If President Trump is right, Dr. McEntarfer and the entire Bureau of Labor Statistics have done irreparable harm to themselves and their work. Speaking as an analyst who uses BLS data frequently, to say this is a big deal is an understatement.

I have no inside information that allows me to conclude whether Donald Trump is right or wrong in this allegation. However, the BLS’ own numbers indicate that “something” is indeed wrong with the jobs numbers, and not just the outsized “revisions” that have become commonplace.

We do well to remember that the BLS employment data reporting provides a multidimensional look at what is fundamentally the same data—America’s population of working men and women. The Employment Level from the Household Survey and the All Employees metric from the Establishment Survey should broadly reflect the same trends.

Pre-COVID, this was what we saw in the jobs data.

While the Employment Level rose more slowly than the Establishment Survey metrics, its variance from the Establishment Survey metrics was fairly consistent.

Post-COVID, that changed.

The Employment Level flat-lined in 2022 and again between 2023 and 2024, while the Establishment Survey metrics did not. Indexed against January 2021 data, the Employment Level flat-lined again in 2025.

Remember, these are all measures of what should fundamentally be the same data. One metric showing no gain while the others show gain should not be possible. Yet that is what we have in the data.

That’s a problem. But it gets worse.

The JOLTS metrics are another perspective on the same employment information. We should expect that the net hiring captured by the JOLTS reports should align fairly closely with the change in the All Employees metric from the Establishment Survey. Pre-COVID, that is what happened.

Post-COVID, that happened also—except in 2025.

While simple measurement error means we should not expect net hiring from the JOLTS report to match the change in the All Employees metric exactly, since March the JOLTS net hiring figure has been considerably larger than the change in All Employees—from roughly 80,000 jobs to over 200,00 jobs larger.

That much variance between these two metrics does not happen, nor should it. One of these two metrics is flat-out wrong.

Such is the quality of data McEntarfer has been putting out the past few months.

Dr. Erika McEntarfer earned her firing, and then some.

Update 1 (1800CDT 1 August 2025)

I am clearly not the only one who believes that change is needed at the Bureau of Labor Statistics.

President Trump has fired the Commissioner of the Bureau, Dr. Erika McEntarfer, for sloppy statistics.

It is a firing that has been long overdue. The BLS data has been poorly handled for years.

Starting Point

Three charts confirm that the jobs report once again is fundamentally misleading.

The first one is the Employment Level from the Household Survey side of the jobs report, showing continued erosion in the number of working Americans.

The second and third are the indexed versions of the Employers’ Survey headcount of working employees, the ADP National Employment Report, and the Household Survey from report.

When we index from January of last year, we see that the Employment Level is once again diverging from the other trends.

When we index from January of this year, we see that the divergence has accelerated since the beginning of the year.

These three charts drive two very unpleasant realities home:

1. The “jobs created” number reported by the BLS and echoed by the corporate media is pure hoakum, and the ADP number is not much better.

2. The US economy is not adding jobs, but overall it is losing them.

Punctuating the latter reality especially, the employment-population ratio dropped in July.

So much for President Trump’s small jobs wins mid-month with the unemployment data.

I will have a much more detailed analysis later, but the top-level data tells enough realize that July was not a good jobs month at all.

The employment situation in the United States did change in July. Despite the obfuscations from the BLS, the data shows that it got measurably worse.

America’s jobs recession got measurably worse.

.