Analyzing the Jobs Report with online BLS tools and AI

On January 9, 2026, the Bureau of Labor Statistics (BLS) released the Employment Situation Report for December of 2025. Following is the first paragraph of that report:

Both total nonfarm payroll employment (+50,000) and the unemployment rate (4.4 percent) changed little in December, the U.S. Bureau of Labor Statistics reported today. Employment continued to trend up in food services and drinking places, health care, and social assistance. Retail trade lost jobs.

Shortly later, Stephen Moore posted a tweet on X which stated the following:

Don't believe the headlines calling this a 'weak jobs report.'

The household survey tells the real story:

A record number of Americans working and 200,000 more people joining the workforce.

Moore's claim that "200,000 more people joining the workforce" in December can be checked on the BLS web site as follows:

  1. Go to https://www.bls.gov/
  2. Click on "DATA TOOLS" and then "BLS Popular Series" on the top menu.
  3. In the Employment section, check the checkbox for "Civilian Employment (Seasonally Adjusted) - LNS12000000".
  4. Click the blue "Retrieve Data" button.
  5. In the "Change Output Options", change the From year to 2021, check the "include graphs" checkbox, and click the blue GO button.
The above steps should result in the following output:

This comes from the Household Survey. As can be seen, the table shows 163,760 thousand jobs in November of 2025 and 163,992 thousand jobs in December of 2025. That is an increase of 232 thousand jobs. Now repeat the same steps but for Series CES0000000001, total nonfarm employment:

  1. Go to https://www.bls.gov/
  2. Click on "DATA TOOLS" and then "BLS Popular Series" on the top menu.
  3. In the Employment section, check the checkbox for "Total Nonfarm Employment - Seasonally Adjusted - CES0000000001".
  4. Click the blue "Retrieve Data" button.
  5. In the "Change Output Options", change the From year to 2021, check the "include graphs" checkbox, and click the blue GO button.
The above steps should result in the following output:

This comes from the Payroll Survey. As can be seen, the table shows 159,476 thousand jobs in November of 2025 and 159,526 thousand jobs in December of 2025. That is an increase of just 50 thousand jobs, the number in the first paragraph of the jobs report. Which survey is best for measuring month-to-month job gains and losses? When asked the question, "Which survey is better for measuring month-to-month job gains and losses, the payroll survey or the household survey?, ChatGPT replied:

The payroll survey (Current Employment Statistics) is better for measuring month-to-month job gains and losses.

It has a much larger sample size, lower statistical noise, and is specifically designed to track monthly changes in nonfarm jobs, which is why the BLS uses it for the headline jobs number.

The household survey (Current Population Survey) is more volatile month to month and is better suited for identifying broader trends, turning points, and demographic patterns, not precise monthly job changes.

Answer: the payroll survey.

Asking ChatGPT "What is the sample sizes of the household and payroll surveys?" gives the following reply:

Here are the standard sample sizes used by the U.S. Bureau of Labor Statistics:

Household survey (CPS - Current Population Survey)

  • About 60,000 households
  • Represents roughly 110,000 individuals
  • Conducted monthly by the Census Bureau for the BLS
  • Measures employment status of people (employed, unemployed, not in labor force)
Payroll survey (CES - Current Employment Statistics)
  • About 122,000 businesses and government agencies
  • Covering roughly 666,000 individual worksites
  • Measures jobs, not people
  • Produces the headline monthly nonfarm payroll jobs number
Why this matters
  • The payroll survey's much larger sample is why it is far more reliable for month-to-month job gains and losses.
  • The household survey's smaller sample makes its monthly changes much noisier, even though it is essential for unemployment and labor force measures.
In short: ~60,000 households vs. ~122,000 employers (666,000 worksites).

This smaller sample size likely has much to do with why the first graph above, from the Household Survey, is much noisier than the second graph, from the Payroll Survey, which is much smoother. Also, note that the graph for the Household Survey often jumps from December of one year to January of the next year. This is due to adjustments to Household Survey population estimates. The most recent adjustment is described at this link. In any case, the noisiness of the Household Survey data can also work against Moore's argument of strong job growth. The Household Survey shows 163,831 thousand jobs in January of 2025 and 163,992 thousand jobs in December of 2025. That's an increase of just 161 thousand jobs in 11 months! The Payroll Survey shows an increase of 473 thousand over those 11 months. That's a relatively weak 43 thousand per month but far better than the 14.6 thousand per month implied by the Household Survey.

Comments

Post a Comment

Popular posts from this blog

From Record Monthly Surplus to the Largest Annual Deficit in 6 Years

Image
On May 25, 2018, Judy Woodruff interviewed Marc Short, White House Director Of Legislative Affairs, on the PBS Newshour. The video and transcript can be found at this link. At 2:43 in the video, Marc Short states the following: And, lastly, in April, we had the largest surplus on record in American history. That is a result of the growth and the revenues coming into the federal government because of the tax plan. READ FULL REPORT
Read more

Do the Trump Tax Cuts Mostly Pay for Themselves?

On April 9, 2018, the Congressional Budget Office released The Budget and Economic Outlook: 2018 to 2028 . The next day, the Investor's Business Daily ran an editorial titled It's Official: Trump Tax Cuts Are Boosting Growth And Mostly Paying For Themselves . That editorial begins: When the Congressional Budget Office released its updated budget forecast, everyone focused on the deficit number. But buried in the report was the CBO's tacit admission that it vastly overestimated the cost of the Trump tax cuts, because it didn't account for the strong economic growth they would generate. Further on, the editorial states: The CBO report also makes clear that this faster-growing economy will offset most of the costs of the Trump tax cuts. In a table buried in the appendix of the CBO report, it shows that, before accounting for economic growth, the tax cuts Trump signed into law late last year would cut federal revenues by $1.69 trillion from 2018-2027. But it goes on ...
Read more

How Big of a Benefit is the Doubling of the Standard Deduction and Child Credit?

Image
Note: This post has been updated to use the final GOP tax bill. On December 15th, the GOP released its final tax plan . The Committee on Rules has posted a Summary of the bill , a Joint Explanatory Statement , and the full bill . A fact sheet on the White House web site lists the following points, among others: The Tax Cuts Act nearly doubles the standard deduction. For individuals, the standard deduction will be increased from $6,500 to $12,000. For single parents, the standard deduction will be increased from $9,550 to $18,000. For married couples, the standard deduction will be increased from $13,000 to $24,000. The vast majority of American families will now be able to file their taxes on a single page because they lower taxes by claiming the standard deduction. American families will see their Child Tax Credit doubled to $2,000 per child under age 17. Families will receive a new $500 tax credit for dependents age 17 or older. The full Child Tax Credit will now be availa...
Read more