The BLS released Business Employment Dynamics for 2024 yesterday. Let’s discuss.
The Business Employment Dynamics report covers job gains by new businesses and job losses by businesses going out of business.
The BLS Birth Death Model is based off stale BED data with annual benchmarks. A Birth-Death FAQ has some interesting details.
Are net birth-death forecasts seasonally adjusted?
No, they are calculated using population data that is not seasonally adjusted, and the forecasts are applied to the sample-based not seasonally adjusted estimates. Months with generally strong seasonal increases such as April, May, and June generally have a relatively large positive forecast. Conversely, months with overall strong seasonal decreases, such as January, generally have a relatively large negative forecast.Since the 2020 benchmark, CES estimates have been subject to persistent and relatively large birth-death forecast errors. To help address these forecasting issues, BLS modified the model-based component of birth-death by incorporating current sample information to inform the forecasts starting with the 2024 benchmark released with the January 2025 Employment Situation. This modification was only applied from April to October 2024, known as the post-benchmark period. November 2024, December 2024, and January 2025, as well as future monthly estimates, use birth-death components calculated without this modification. [So, years late, the BLS made a modification for 7 months, now tossed]
Currently, BLS will not use this method outside of the post-benchmark period as there is not sufficient time during the monthly estimation period to incorporate the additional processing needed to use this method in the current month’s estimates.
BLS releases the Quarterly Census of Employment and Wages (QCEW) monthly employment data on a quarterly basis 6 to 9 months after a given collection month. During current month birth-death processing, the forecasted portion of the birth-death model uses inputs derived from comparing the following:
- the QCEW-based employment data including business openings and closings
- the same QCEW data excluding the birth units and imputing monthly values to the death units
The difference between these two QCEW-based monthly values is the net of births and deaths. Finally, BLS creates an over-the-month change from this series, and those monthly values become the inputs to an ARIMA time series model used to create forecasts up to and including the current month.
These lagged inputs and the ARIMA forecasting approach can produce accurate forecasts when business openings and closings for current months follow predictable seasonal patterns. However, sharp trend changes or extreme events like hurricanes and pandemics can result in large forecast errors.
Seasonal Adjustments
Birth-Death data is “not seasonally adjusted, and the forecasts are applied to the sample-based not seasonally adjusted estimates.”
Birth Death is not seasonally adjusted but the BED data is. There is also unadjusted BED data but it is impossible to work with.
For example the unadjusted BED numbers for 2014 Q2-Q4 are +2.355 million, -648,000 and +503,000 respectively.
The S/A to N/A comparisons are: -163,000 vs 2.355 million, -1,000 vs -648,000 and 287,000 vs 321,000.
The BLS should give seasonally adjusted birth-death numbers but it doesn’t. One thing you should not do is subtract unadjusted birth-death numbers from the seasonally adjusted nonfarm payroll reports.
Private Payroll Issue
Importantly, BED and QCEW reports are private payrolls. Nonfarm payroll data includes government jobs.
Curiously, the monthly BLS job reports includes wage and hour worked for private employees, but the BLS does not report private payroll data. It does provide private “employment” numbers monthly but that’s from the household survey.
The BLS does have weekly private payrolls so how hard can it be to provide monthly numbers?
258,000 Negative Revisions in May and June Nonfarm Payrolls
Earlier today I noted Payroll Disaster, Jobs Rise 73,000 but Massive Negative Revisions
In May, the BLS reported 144,000 jobs. And in June the BLS reported 147,000 jobs.
I kept asking “Does anyone believe these reports?”
Today, the BLS says oops. Employment in May and June was a combined 258,000 lower than previously reported.
Monthly Revisions
- The change in total nonfarm payroll employment for May was revised down by 125,000, from +144,000 to +19,000
- The change for June was revised down by 133,000, from +147,000 to +14,000.
- With these revisions, employment in May and June combined is 258,000 lower than previously reported.
I was pretty sure big negative revisions were coming because of QCEW data.
QCEW stands for Quarterly Census of Employment and Wages. QCEW represents about 95 percent of the data with mandatory reporting.
QCEW vs Nonfarm Payrolls
The above image from QCEW Concepts.
CES stands for Current Establishment Survey, the BLS nonfarm payroll report.
The monthly jobs report is a survey of 629,000 worksites vs 12.2 million vs QCEW. The response rate of QCEW is ~90 percent. The response rate on CES is ~45 percent.
QCEW Report Shows Overstatement of Jobs
On June 16, I commented QCEW Report Shows Overstatement of Jobs by the BLS is Increasing
The discrepancy between QCEW and the BLS jobs report is rising.
It is inexcusable for the BLS to not incorporate QCEW data as soon as possible.
Instead, it relies on poor sampling of a small subset. On that poor sample the response rate is pathetic.
In addition, there is survival bias. In recognition of survival bias, the BLS concocted its absurd birth-death model.
And on top that that, struggling businesses have no incentive to respond. In contrast, large corporations likely have someone dedicated to filling out government surveys.
