The Bureau of Labor Statistics (BLS) reported on November 6 that October nonfarm payrolls rose 165,000, well above the 48,000 average monthly gain of the prior 12 months. August was revised up 53,000 to +186,000 and September up 31,000 to +60,000, together 84,000 higher than first reported. The unemployment rate was 4.1%.
Earlier coverage here treated the summer reports as a denominator story: the August revision to July showed how a trimmed payroll estimate understates frequency. This report changes a different input. The level matters less than where the jobs landed.
Where the payroll was added
Professional and business services added 70,000 jobs, against a 12-month average of 15,000. Health care added 42,000 (12-month average 31,000) and transportation and warehousing 25,000. Construction was down 3,000 and government up 6,000, both described by BLS as little changed.
For a workers compensation program, those are different exposures. Office-heavy professional services adds payroll with comparatively low injury frequency. Health care and warehousing are typically higher-frequency classes. A program that scales its 2026 payroll up 1% across the board is treating a 70,000-job office gain the same as a 42,000-job patient-care gain. The expected claim ratio (ECR, the share of payroll the actuary expects to emerge as claims) should move by class, not in total. The NCCI finding that payroll overstates exposure for high-wage employers points the same way: added dollars of payroll in high-wage office roles do not carry proportional claims.
Who it affects
Self-insured hospital systems and health care employers, self-insured logistics and distribution operators, and captives insuring those groups face payroll growth in higher-frequency classes. Professional-services parents with a captive may see payroll grow with little change in expected claims. Public-entity pools should note that government payrolls moved only 6,000, so the funding base is not growing with private payroll even if wages are.
The reserve mechanism
The lever is the exposure base in the expected claim ratio and in frequency-per-payroll checks. Three inputs move:
- Denominator. The 84,000 revision equals about 1.75 months of the trailing average gain. Programs that froze 2026 estimated payroll after the weaker summer prints are carrying a base that BLS now says was too low.
- Mix. Class weights determine the blended prior. Health care and warehousing gains argue for a higher blended expected frequency; professional services gains argue for a lower one. The net is a program-specific number.
- Wage-linked severity. Average hourly earnings rose 14 cents (0.4%) in October to $38.00, up 3.1% over 12 months. The Employment Cost Index showed private wages and salaries up 3.1% over 12 months as of September. Two independent BLS series agreeing at 3.1% is a sensible ceiling for indemnity trend where benefits track state average weekly wage. It does not touch the development pattern or tail factor.
Hours add a wrinkle. Involuntary part-time work (people working part-time who want full-time) rose 339,000 to 4.8 million, while average weekly hours held at 34.4. Headcount grew but hours per head did not, so for hourly workforces a payroll-based and an hours-based exposure measure can tell different stories. If your TPA reports claims per 100 full-time equivalents, check that the denominator is hours, not heads.
Where this shows up in your reserves
Open the exposure schedule behind the 2026 accident-year ECR in your last actuarial report. Look for whether payroll is audited, estimated or budgeted, and the date of the last refresh. Then compare the payroll by class to the class weights in the a priori loss rate. The BLS injury data due November 18 will give the frequency side of the same check.
What this means for your next review
Put one item on the agenda: re-estimate 2026 payroll by class code using the latest monthly figures, and ask the actuary to show the blended ECR before and after. If the movement is within a point or two, document it and move on. The Workers Compensation IBNR guide and what is driving your IBNR higher cover the underlying mechanics.
Our call
Do not read 165,000 as a reason to raise the aggregate expected claim ratio. The jobs that were added skew to office work, the higher-frequency gains are smaller, and hours per worker did not rise. Expect the net effect on a typical self-insured blend to be modest, with health care and warehousing programs the exceptions. Treat the 84,000 revision as a reminder to refresh the base, not as a trend. The November Employment Situation on December 4 will show whether the revisions hold; revisions have reversed before, so keep the update reversible.