Yesterday somebody contacted me and said the employed total for 2023 was wrong, so I dug into it, and this is what I found.
Updated 16 Feb 2025
Sent this email after downloading files again.
I thought I should alert you to this after downloading all data again on 16 Feb 2025.
Fred does not get revised.
Your own SA and NSA data on your website does not get revised.
There is no way to properly update the data without blindly following what you have stated here.
Two files I downloaded can be seen here.
On 22 Jan 2025 I contacted BLS with this question because I wanted to be accurate.
Hi, I am looking for the data that directly coincides with the monthly employment numbers, broken down by naics code. As an example, if we created 256,000 jobs last month, I want the data to show the 256,000 increase, broken down by naics code. I want to add this to my website at guestworkervisas.com Thank You Virgil
I received this response.
Good afternoon,
Thank you for contacting the Current Employment Statistics (CES) program, which produces monthly estimates of nonfarm employment, hours, and earnings by industry.
Please see this table, https://www.bls.gov/web/empsit/ceseeb1a.htm, which includes employees on nonfarm payrolls by industry sector and selected industry detail, seasonally adjusted for December 2024.
You can also access historical B-1 table data following the steps listed below:
- Open the Historical B-1 table data retrieval: https://www.bls.gov/webapps/legacy/cesbtab1.htm
- Check the box(es) next to the series and in the column of the type of data to view. Check as many boxes as desired.
- Click on Retrieve data at the bottom of the page, and a new window will open displaying up to 10 years of data for each time series selected. If this is not the desired format, more formatting options are available. See the More Formatting Options section of this page under Other Tips for further instructions.
If you have further questions, please feel free to contact us by email, or by calling 202-691-6555.
So I started programming and used that data after verifying that the download matched the 256,000 jobs for that month.
So we’re good, or so I thought.
Yesterday I received this.

It shows a increase of 3 million jobs, so I looked at my data that I downloaded in January and I see an increase of 1,996,000
So I dig into the data.
Their data shows 157,304,000 in Dec 2023 and 154,291,000 in Dec 2022 for a gain of 3,013,000 jobs.
The data I downloaded shows 158,926,000 and 156,930,000 for a gain of 1,996,000 jobs.
So I write them another email and this is what I received.
Thank you for contacting the Current Employment Statistics (CES) program, which produces monthly estimates of nonfarm employment, hours, and earnings by industry. CES data are subject to several revisions.
All data are revised for up to two months from their initial release due to the ongoing receipt of sample data. Information on these revisions can be found at https://www.bls.gov/ces/methods/benchmark-tables.htm. After this two-month interval has passed, CES data series are held constant until the annual benchmarking process, when they are revised again to incorporate new universe counts. Benchmark revisions are released with the January employment report and the 2024 benchmark revision article published on February 7th can be found here: https://www.bls.gov/web/empsit/cesbmart.htm.
The annual benchmarking process involves reconciling CES sample-based employment estimates with near-population counts from State unemployment insurance (UI) tax records. UI records are the sample frame for CES data, and they are used because most establishments with employees are required to pay these taxes.
In the benchmarking process, not seasonally adjusted sample-based all employee (AE) estimates for March of the most current year are set equal to an employment value derived from UI counts, plus a small adjustment for the few cases in which employers are not required to pay UI taxes. Only AE data are directly benchmarked. However, because women worker (WW) and production and nonsupervisory worker (PW) data are expressed as a proportion of AE, they too are subject to revision. Furthermore, PW revisions can carry through to hours and earnings data. Revisions to non-AE data, however, are typically quite small.
Resetting the not seasonally adjusted March level of AE causes a shift in the data for that month. Since it is unlikely that all of the difference between sample-based and UI-based employment occurred in March, that difference is proportionally distributed, or wedged, back to the previous April. Data subsequent to the benchmark month is re-estimated, using standard CES methodology, with the new March level as the base. Not seasonally adjusted data is subject to revision from the previous benchmark forward. For example, when data benchmarked to March 2024 levels was released in February 2025, not seasonally adjusted data was revised from April 2023 forward.
Simultaneous with the benchmark, the seasonal adjustment process is rerun for 5 years of data; seasonally adjusted estimates are therefore subject to five years of revision. So, when data benchmarked to March 2024 levels was released in February 2025, seasonally adjusted data were revised from January 2020 forward.
A comprehensive explanation of the benchmark process can be found at https://www.bls.gov/web/empsit/cesbmart.htm. The benchmark process for datatypes other than all employees is at https://www.bls.gov/web/empsit/cesbmart.htm#BMKRevOtherDT.
A final source of revision may be the switch from one iteration of the North American Industry Classification System (NAICS) to the next, which occurs every 5 years following the 2002 introduction. The most recent iteration was in February 2023 to the 2022 NAICS structure. These changes resulted in minor revisions to the industry structure and to CES estimates.
Information on the 2017-to-2022 NAICS conversion is available at https://www.bls.gov/ces/naics/naics-2022.htm. NAICS industry definitions are available at https://www.census.gov/naics/.
As CES data is revised, it is overwritten in our database. There are, however, sources for first preliminary published data; for example, employment data as originally published. This data is available on the CES Vintage Data page at https://www.bls.gov/web/empsit/cesvininfo.htm. Data is available at the CES supersector, 2-digit, and 3-digit levels of industry detail. Originally published data for women employees, production employees, hours, or earnings are not available on this page. Vintage data tables show all published values since the first preliminary estimates for May 2003.
If you have further questions, please feel free to contact us by email, or by calling 202-691-6555.
At this point I’m not sure where to go.
I’ll probably upload the new data.
But what gets me is this, we just had an increase of 1,017,000 in Employed.
Yet not a peep from the media anywhere.
Do they even question this stuff, or do they blindly spread it.