BBSI's Profit Was Fattened by Released Insurance Reserves. That Line Is Now Shrinking.

On August 5, Barrett Business Services did something unremarkable and quietly damning. The small PEO, which processes paychecks and workers' compensation for thousands of small construction and blue-collar employers, grew second-quarter revenue 4% to $319.3 million. Its profit fell anyway, to $0.52 a share from $0.70 a year earlier. Analysts had penciled in $0.56. The stock has not recovered, trading near $32 after peaking around $47 a little over a year ago.

Revenue up, profit down. That kind of divergence is usually where the story gets interesting, and BBSI's is no exception. But the interesting part is not in the revenue line, and it is not in a hidden charge. It is in an accounting line most investors skim past, labeled something like "favorable prior-year development." Follow that line and the whole recent history of this stock rearranges itself.

What BBSI actually is

To see why, you have to see how a PEO makes money. A professional employer organization like BBSI co-employs its clients' workers: it runs their payroll, handles their HR, and — critically for this story — carries their workers' compensation risk. When a client signs up, the "gross billings" BBSI shows include the full amount it pays out to that client's employees and insurers, about $9 billion a year. Almost all of that flows straight back out again. What BBSI keeps — its actual revenue, about $1.24 billion last year — is the narrow spread between what it charges clients and what it pays out, plus fees.

BBSI does not buy a commercial workers' comp policy for its clients. It self-insures the exposure, setting aside reserves to cover future claims and only discovering in later years whether those reserves were too big or too small. And for several years, they were too big. Every quarter, the company updated its actuarial estimates and released the excess back into income. In 2024 that "favorable prior-year development" added $18.5 million to net income. In 2025 it added $18.7 million.

Put that in context. BBSI's full-year 2025 net income was $54.4 million on $2.08 of diluted EPS. So releasing old reserves supplied roughly a third of the entire year's profit — about $0.71 of that $2.08. The company was not only earning on the premiums clients paid this year; it was banking under-reserving that had quietly accumulated in prior years and now, because claims had come in light, was being handed back to the income statement.

That is not fraud. When claims really do come in below the actuarial estimate, releasing the reserve is the truthful accounting. It is real money, and the people who call every reserve release "earnings manipulation" are overreaching. But it is a distinction the market forgot. A reserve release is a one-time settling-up of how well the company estimated claims several years earlier. It is not recurring earning power, and treating it as such inflates how durable the business looks.

The line is turning

Here is the current quarter. In Q2 2026, BBSI recognized only $2.0 million of favorable prior-year development, down from $8.8 million in the same quarter of 2025. That $6.8 million swing is most of the arithmetic behind the earnings "miss." And it is not a quirk of one quarter: the workers' compensation market has firmed. Claims costs are climbing — the company points to rising litigated post-term cumulative-trauma claims, concentrated in California, where a claims-heavy and expense-heavy regulatory environment has pushed costs for years. BBSI's workers' comp expense rose to 2.5% of gross billings in Q2 2026 from 2.1% a year earlier, and gross margin as a share of billings fell to 2.8% from 3.3%.

The company's answer is that it is finally doing something about it. In what management describes as the first sustained run of workers' comp price increases in nearly a decade, BBSI has been raising what it charges clients for months, helped by approved California rate increases of 8.7% in 2025 and a further 6.6% effective this September. The argument is that 2026 is the "low watermark" for margins, and that cumulatively higher pricing will compound into 2027 as clients renew at the new rates.

That answer is plausible, and it deserves a fair test. The counterweight is timing. Clients renew on a rolling cycle, so higher prices take months to bleed through the entire book while the higher claims costs hit immediately. In other words, BBSI is selling insurance price increases into a market that is simultaneously raising its costs — and the question is whether the pricing power can outrun the claims inflation before the reserves the company has been releasing are exhausted. In a firming market, the sweet spot runs the other way: reserves that were released for years now have to be built up, which would push reported earnings down even while the underlying business prices itself healthier.

Growth is the second leg of the squeeze, and it is soft. Full-year 2026 guidance was trimmed to gross-billings growth of 3% to 4% and worksite-employee growth of just 2% to 3%, as clients — concentrated in construction and other blue-collar sectors — trim headcount. BBSI's net client retention was a strong 97% at the start of the year and it keeps adding new logos faster than expected. But the employers who stay are employing fewer workers, which caps the volume metric that actually drives revenue.

What the stock really costs

Now the valuation. The stock trades around $32.50, roughly 22 times trailing earnings. That multiple looks like the price of a steady, debt-free compounder, and in many ways BBSI is one: no debt, about $68 million in unrestricted cash, a dividend, and an aggressive buyback that spent $15 million in Q2 alone and has roughly $40 million left on its authorization.

The catch is the denominator. Trailing earnings still carry the benefit of reserve releases that are now shrinking, so the 22 times overstates how much you are paying for the profit the business can plausibly sustain. The sell-side has pushed its 2026 EPS estimate down to about $1.22 from $1.53 as the reality has set in, and the back half of the year starts from a first half that lost $1.9 million net, dragged into the red by a one-time tax charge in Q1. On the street's own forward number, the stock is closer to 27 times expected 2026 earnings than 22 times trailing — and that is before deciding how much of even that reduced estimate still depends on favorable development continuing.

There are three ways this plays out, and they cost the shareholder different amounts.

The benign case — call it resolution — is that management is right: 2026 is the floor, pricing compounds through 2027, favorable development settles at a small, steady level, and margins recover. Here the stock is roughly fairly valued, and the buyback quietly supports per-share results while you wait. This is the case the company is selling, and it is coherent.

The persistent-but-lawful case is that claims inflation keeps running ahead of what clients can be charged, that the reserve tailwind the business leaned on for several years does not just fade but reverses into reserve building, and that soft hiring caps growth for another year or two. In that world, earnings stay depressed longer than the "low watermark" framing promises, and a 20-plus-times multiple on a disliked earnings trough is not as cheap as the seller's sign suggests.

The materially-misstated case — that the reserve releases were manufactured — has no support in the evidence before us, and I am not alleging it. BBSI's disclosures are detailed, its accounting is transparent, it is audited by a major firm, and the favorable development was disclosed quarterly at a time when it was flattering results. This is not a hiding place in the sense of something concealed. It is a hiding place in the sense of something seen and then forgotten: the reserve-release line was there all along, and the market priced the profit as if it would recur.

That is the real invoice. The shareholder who bought BBSI's story paid for a durable compounder at a multiple that only made sense if the releases kept coming. The company's whistle did not blow on fraud; it blew on a change in the insurance cycle, which is a normal thing for an insurer to experience. The lesson is narrower but useful: when a self-insuring company shows you years of profit aided by "favorable prior-year development," that is a specific, reversible line, and you should price it as an asset that can turn. The next signposts — quarterly favorable development, the workers' comp expense ratio, and whether 2027 guidance finally shows gross margin turning back up — will say which of the three cases is actually paying out. BBSI's financial detective work has never been the hard part. The hard part is remembering that a tailwind, no matter how steady it has looked, is still a wind.