The Bureau of Labor Statistics published its Employment Cost Index for the second quarter of 2026 on July 31, and one pairing in it undercuts a lot of pricing decisions made earlier this year. Total compensation costs in professional, scientific, and technical services, the industry that covers most dev shops and studios, rose 3.0% in the twelve months through June. Wages and salaries for professional and related occupations, a broader group spanning every industry that employs that kind of work, rose 3.5% over the same window, faster than the 3.4% all-civilian-worker average BLS reported the same day. A fixed-bid quote signed against January's rate card already missed both numbers, and neither shows up on its own as a change order.
What the Employment Cost Index actually measures
The ECI comes from BLS's National Compensation Survey, not a self-reported salary site or a single employer's payroll data. It tracks total compensation, defined as wages and salaries plus employer costs for benefits, across a fixed basket of jobs sampled from private industry and state and local government, so the index isn't distorted by workers shifting into higher- or lower-paid roles the way a raw average wage figure can be. It publishes quarterly, about a month after each quarter closes. The release used here (reference period Q2 2026, published July 31, 2026) reports figures not seasonally adjusted, and every number below is a twelve-month percent change through June 2026 on that basis.
The two numbers that moved in different directions
"Professional, scientific, and technical services" is an industry classification: who employs the worker. "Professional and related occupations" is an occupational classification: what the work is, regardless of which industry hired them. BLS reports both, and this release, they tell slightly different stories.
| Series | 12-month change, through Jun 2026 | What it covers |
|---|---|---|
| All civilian workers, total compensation | 3.4% | Wages plus benefits, every industry, every occupation |
| Professional, scientific & technical services, total compensation | 3.0% | Wages plus benefits, this industry only |
| Professional and related occupations, wages and salaries | 3.5% | Cash wages only, this occupation group, every industry |
Total compensation in the industry that houses most dev shops grew slower than the economy-wide average. Cash wages for the occupation group that includes most of the people doing the work grew faster than the economy-wide average. Read together, the honest inference is that employers in this industry are holding the line on total comp mostly by not growing benefits as fast as other industries did, while the base pay a subcontracted engineer or PM actually sees on their check is running hotter than the headline 3.4% figure suggests. BLS doesn't publish the reason behind either number, so that's a read of the data, not a finding in it, and I'd stand behind it on a call.
What this costs a fixed-bid contract signed in January
None of this shows up as a single event. A studio prices a nine-month fixed-bid build off January's payroll, wins it, and staffs it. By month nine, the people delivering it cost the studio 2 to 3 percentage points more to employ than they did at signing, compounding through raises, backfills at market rate, and the odd counteroffer, while the client's invoice hasn't moved since the contract was signed. On a build priced at a healthy margin, that erosion is an annoyance. On a build quoted thin to win it, which is exactly the situation an agency owner is in when rates are under pressure, 2 to 3 points of unplanned cost growth can be most of what was left.
A worked example
Picture a 14-person dev studio that quotes a 40-week fixed-price build in January, priced to clear a 20% margin at signing. Nothing about the build changes: same scope, same team, same client. Apply this year's wage growth pro-rated across the 40 weeks, and by delivery the studio's actual labor cost has drifted up by roughly three-quarters of a point of the total contract value, not because anyone made a mistake, but because the rate card was frozen the day the SOW was signed and the labor market wasn't. A 20% margin becomes a margin in the low 19s. That's survivable once. Stacked across four or five fixed-bid builds running concurrently, all priced the same way, it's the gap between a profitable quarter and one that quietly wasn't.
What actually protects the margin here
- Index engagements longer than two quarters to a public series, the ECI or a comparable one, with a stated adjustment date, rather than eating the drift silently or guessing at a flat annual bump.
- Separate the piece of your delivery this actually hits. The ECI measures US employer costs. If your engineering bench sits offshore and is priced in a different currency against a different market, this specific number isn't your engineering cost line. It's very much your cost line for the US-based account leads, PMs, and any onshore engineers who round out a disclosed hybrid team, and those roles are usually the ones a fixed-bid quote underprices first.
- Quote time-and-materials for anything past a two-quarter horizon where the client will accept it. A fixed price that far out is a bet that this year's wage growth stays flat, and the data above says it hasn't.
- Revisit the rate card at each quarterly ECI release, not once a year. A number that moves every quarter deserves a pricing review on the same cadence, even if the adjustment most quarters is small enough to just note and hold.
Where this doesn't fit
If a studio staffs a build entirely through an offshore subcontracted bench priced and paid in another country's currency, the US Employment Cost Index isn't the number to price against at all. The right reference point is that market's own wage data, and using a US index there would misprice the actual cost being incurred. And a studio already billing everything time-and-materials has less to fix here: rates can move project to project without touching a signed contract. The exposure this data describes belongs specifically to fixed-price, US-payroll-exposed engagements running more than a quarter or two, which is a narrower slice of the market than "every dev shop," even if it's a common one.
Studios quoting multi-quarter builds off a rate card that hasn't been checked against this data are pricing against last year's labor market. Tell us the scope of what you're staffing and we'll help you figure out where a disclosed hybrid team changes that math, and where it doesn't. The markup arithmetic in the economics of white-label development and the vetting checklist in how to vet a global development partner are both worth rereading against this quarter's numbers, and the turnover risk in a rising quits rate compounds the same margin problem from a different direction. If a rate card needs more than a pricing tweak, our services page covers how we scope a build from the start with this kind of drift priced in rather than discovered later.
Sources
- U.S. Bureau of Labor Statistics: Employment Cost Index news release, June 2026 results (published July 31, 2026)
- U.S. Bureau of Labor Statistics: Table 5, Employment Cost Index for total compensation, private industry workers, by industry
- U.S. Bureau of Labor Statistics: Table 8, Employment Cost Index for wages and salaries, civilian workers, by occupational group