The Bureau of Labor Statistics published the August 2026 Employment Situation on September 4, and the headline drew most of the attention: 162,000 nonfarm jobs added. The more consequential figure for self-insured workers compensation programs sits in the revision table. July, initially reported as a loss of 23,000 positions, was revised upward to a gain of 21,000. The swing is 44,000 jobs.
Any program that recalibrated its 2026 payroll exposure downward after the August 7 initial release now holds a frequency baseline that is too low for accident year 2026. The original July report and its reserve implications were covered here. The September 4 revision reverses those implications directly.
Who It Affects
Self-insured WC employers and group captives that run mid-year or quarterly exposure updates are the primary targets. If a risk manager or actuary reduced the expected payroll denominator for July after the initial negative print, that input is now 44,000 jobs short. The effect scales proportionally to each program’s share of exposure in affected sectors.
Public-entity self-insureds have additional reason to check. Local government employment added 42,000 positions in August, the largest sector gain in the report. Police, fire, and public works payrolls feeding civil service WC programs are growing, not contracting.
Reserve Mechanism: Frequency
Payroll is the denominator in workers compensation IBNR models. Expected claim count flows from payroll times a frequency rate per $100 of covered wages or per employee. An understated payroll produces an understated expected frequency, which pulls the Bornhuetter-Ferguson initial expected losses below where they belong and understates IBNR for accident months that have not yet fully developed.
The concrete exposure is the July 2026 accident month. Any study updated in August using the initial July payroll figure has an exposure unit that is 44,000 jobs short for that month alone. Combined with August’s 162,000 additions, cumulative 2026 payroll is materially higher than what an August-vintage IBNR study reflects.
Two supporting signals push in the same direction. Average hours worked edged up 0.1 to 34.4 in August, raising total payroll dollars. Average hourly earnings rose 0.3% for the month and 3.1% year over year to $37.75 per hour. Total payroll is hours times headcount times wages; all three components moved upward.
The sector mix reinforces the frequency exposure. Food services added 59,000 positions and construction added 22,000 in August. Both carry above-average WC claim frequency, particularly for musculoskeletal injuries. Healthcare added 13,000, relevant for hospital self-insured programs tracking professional-line exposure alongside WC.
What This Means for Your Next Review
Before your year-end 2026 accident year study is finalized, confirm with your actuary that the July payroll denominator has been corrected from the initial -23,000 to the revised +21,000. Ask whether the program has a minimum revision threshold that would trigger an interim correction outside the normal study schedule. For programs running quarterly updates, this correction belongs in the Q3 study, not deferred to year-end.
The September 2026 Employment Situation releases October 2. If August’s 162,000 figure is itself revised materially, the same adjustment logic applies.
Sources
- BLS Employment Situation Summary, August 2026, released September 4, 2026
- BLS Employment Situation, July 2026 initial release, released August 7, 2026