Jobs: Firing the Head of the BLS Resulted in Positive Revisions for the First Time in 2025

SchiffGold US Debt Employment

In a rare move, the Household survey outperformed the Headline Report

Exploring Finance https://exploringfinance.github.io/
09-05-2025

This article first appeared on SchiffGold.

The analysis below covers the Employment picture released on the first Friday of every month. While most of the attention goes to the Headline Report, it can be helpful to look at the details, revisions, and other reports to get a better gauge of what is really going on.

The jobs report showed a very small increase of 22k jobs. This month actually saw a big deviation in the Household Survey to the upside which is quite rare. The Household Survey had 288k jobs. In almost all months (except in January), the Household Survey is constantly underperforming the headline number.

Figure 1: Primary Report vs Household Survey - Monthly

For the year, the Household Survey is significantly higher due to the data anomaly in January. However, if we removed the data correction in January, the Household Survey would actually show a loss of 790k.

Figure 2: Primary Report vs Household Survey - Annual

Switching to the headline report…

The BLS publishes the data behind their Birth/Death assumptions (formation of new business). In August, there were 90k jobs assumed into existence. Without these jobs assumed into existence, it is very likely that the jobs report this month would have been negative.

Figure 3: Primary Unadjusted Report With Birth Death Assumptions - Monthly

For the year, the birth death assumption is positive of 961k against a negative number for actuals (-1.47M). This means all positive growth YTD is directly tied to the birth/death assumptions.

Figure 4: Primary Unadjusted Report With Birth Death Assumptions - Monthly

Digging Into the Headline Report

The 22k jobs number was accompanied by a rise in the unemployment rate to 4.3%.

Figure 5: Change by sector

Jobs by Category

When looking at the last 12-month trend, most of the categories were actually above trend.

Figure 6: Current vs TTM

The table below shows a detailed breakdown of the numbers.

Monthly Average Change

Total % Change

Category

Total Employed

Current Month

3 Months

12 Months

3 years

Current Month

3 Months

12 Months

3 years

Private Sector

Construction

8,295

-7

-3.3

4.8

13.8

-0.1%

-0.1%

0.7%

6.0%

Education Health

27,452

46

58.0

71.8

82.3

0.2%

0.6%

3.1%

10.8%

Financial

9,250

-3

0.0

6.7

4.5

0.0%

0.0%

0.9%

1.8%

Information

2,926

-5

-5.0

-0.2

-4.6

-0.2%

-0.5%

-0.1%

-5.7%

Leisure Hospitality

17,050

28

9.7

19.3

29.8

0.2%

0.2%

1.4%

6.3%

Manufacturing

12,722

-12

-10.3

-6.5

-2.9

-0.1%

-0.2%

-0.6%

-0.8%

Mining and Logging

609

-6

-4.3

-1.1

-0.1

-1.0%

-2.1%

-2.1%

-0.5%

Other Services

6,050

12

3.7

6.2

9.5

0.2%

0.2%

1.2%

5.7%

Prof Business

22,536

-17

-17.0

-4.6

-4.2

-0.1%

-0.2%

-0.2%

-0.7%

Trade Trans Utils

29,082

2

-2.0

14.2

9.6

0.0%

0.0%

0.6%

1.2%

Government

Government Federal

2,918

-15

-11.3

-7.2

1.6

-0.5%

-1.2%

-2.9%

2.0%

Government Local

15,157

12

18.7

15.3

24.8

0.1%

0.4%

1.2%

5.9%

Government State

5,493

-13

-7.3

3.3

9.6

-0.2%

-0.4%

0.7%

6.3%

Total

All

159,540

22

29.3

122.2

173.7

0.0%

0.1%

0.9%

3.9%

Values in 1,000s of workers. Data as of: Aug 2025. Total Employed = Entire size of the labor market.

Revisions

The chart below shows how the jobs data has been revised. So far, every month this year has been revised lower by a significant margin, except for the most recent month (July). This shouldn’t come as too much of a surprise considering the head of the BLS was fired last week for posting too many negative revisions.

Figure 7: Revisions

Over the last twelve months, jobs have been revised down by about 26k per month and revised lower by 52k over the last three months!

3 Month Compare

12 Month Compare

3 Year Compare

Category

Current

As Of Published

Avg Month Diff

Current

As Of Published

Avg Month Diff

Current

As Of Published

Avg Month Diff

Private Sector

Construction

-1

10

-3.7

88

112

-2.0

513

566

-1.5

Education Health

198

176

7.3

881

841

3.3

2,999

2,847

4.2

Financial

10

15

-1.7

89

114

-2.1

181

249

-1.9

Information

-8

-4

-1.3

-5

20

-2.1

-164

-10

-4.3

Leisure Hospitality

28

53

-8.3

230

319

-7.4

1,097

1,490

-10.9

Manufacturing

-30

-30

0.0

-106

-69

-3.1

-66

71

-3.8

Mining and Logging

-11

-12

0.3

-8

-13

0.4

1

-1

0.1

Other Services

9

9

0.0

63

78

-1.2

331

314

0.5

Prof Business

-57

-38

-6.3

-70

-25

-3.8

-99

514

-17.0

Trade Trans Utils

-19

16

-11.7

161

197

-3.0

374

601

-6.3

Government

Government Federal

-45

-34

-3.7

-69

-63

-0.5

71

51

0.6

Government Local

31

42

-3.7

199

203

-0.3

865

814

1.4

Government State

-20

39

-19.7

62

115

-4.4

366

313

1.5

Total

All

85

242

-52.3

1,515

1,829

-26.2

6,469

7,819

-37.5

Values in 1,000s of workers. Because this data is focused on revisions, it is as of the month prior: Jul 2025. "Current" shows the change in employment after revisions. "As Of Published" shows change in employment at time of release. "Avg Month Diff" shows the average monthly change from the revisions.

More Detail in the Household Survey

Another level of detail in the Household report shows full-time vs part-time job holders. This month shows that full-time jobs were lost while part-time were gained.

Figure 8: Full Time vs Part Time

Historical Perspective

The chart below shows data going back to 1955.

Figure 9: Historical Labor Market

The labor force participation rate is still well below the highs before the Global Financial Crisis. This month it increased slightly to 62.3%.

Figure 10: Labor Market Distribution

Conclusion

Well, it’s hard to put a ton of stock in the data right now. After the amount of revisions and deviations in all the jobs reports recently, Trump fired the head of the BLS. Of course, the reason wasn’t inaccuracy or inconsistency… it was because he didn’t like the numbers.

The result this month was a revision higher for July, but an overall very weak headline report. With AI adoption and a relatively weak economy, it’s hard to see where the jobs numbers are going to get back on track and moving higher. At some point, Trump is going to have to come to terms with the fact that this economy is not very strong.