How to Decide Whether to Take Prevailing Wage Work

Why this matters

Almost every shop decides this one job at a time, and one job at a time the answer is always no. The wage is a pass-through only if you priced from the determination before you bid; the capability never is. A shop that estimates from its own rates and finds the determination after award has not passed anything through, it has eaten the difference. Learning to read a wage determination, reconfiguring payroll to carry classifications and fringe, changing how technicians record time, and producing a weekly sworn submission are costs you pay once and then keep paying at a lower rate. Spread across one project a year they are brutal. Spread across four they nearly disappear.

So the question is not whether this project pays. It is whether you intend to do enough of this work for the capability to amortize, and that is a decision about your next twelve months, not about the invitation on your desk. This walks through how one shop actually ran it.

Confirming a job is genuinely covered - by the federal Davis-Bacon Act on federally funded construction, by a Davis-Bacon Related Act on federally assisted work, or by a state prevailing wage statute whose coverage and thresholds differ from the federal ones and in some states reach maintenance - is a separate step and comes first. See related: A Prevailing Wage Job Comes Up, Should You Bid It. Check one thing before anything below: if your state's statute has a low threshold and reaches maintenance, a meaningful share of the institutional work you already do may already be covered, the volume analysis is moot, and the only question is how fast you can build the capability.

The shop

Six field technicians, roughly 9,000 field hours a year, mixed commercial and residential mechanical service, no covered work ever. A general contractor invites them to price a six-week mechanical scope on a public elementary school renovation, about 240 field hours. The owner's instinct was to decline because he had heard certified payroll was a nightmare. He ran it instead.

Question 1: what does the capability cost to stand up, once?

Count office hours, not feelings. This shop's list, built by calling a payroll provider and a peer who already did this work:

• Reading and mapping classifications from a determination onto their own people: 8 hours
• Reconfiguring payroll to carry per-classification rates, fringe, and split-classification days: 10 hours
• Changing timekeeping so technicians record classification changes as they happen: 6 hours
• Producing and correcting the first two weekly submissions: 6 hours

Thirty office hours. Amortize over three years and call it 10 office hours a year, which is the right treatment because the configuration persists and the learning does not evaporate.

Why this step is first: everything downstream is a per-project number, and per-project numbers are small enough to talk yourself into. The fixed number is the one that decides the answer, so it goes on the table before anything else does.

Question 2: what does each covered project cost on top of that?

Two recurring items. Classifying and pricing the scope off the applicable determination before bid: about 3 office hours per project. Weekly certified payroll production, review and submission during the project: about 2.25 office hours per active week in the first year, dropping toward 1.25 once the process is routine and the technicians' records stop needing correction.

The drop is not optimism. It is the specific hours spent chasing a technician who wrote "9.25 hours, two jobs" instead of splitting his day, and that stops happening after about the third project because the crew learns what the form needs.

Question 3: what is your ordinary office load per hundred field hours?

You need a baseline or the covered number means nothing. This shop measured its ordinary administrative load at about 4 office hours per 100 field hours across all work. Measure yours; do not borrow this one.

What breaks if you skip it: the covered project's admin load looks alarming in isolation and unremarkable against the baseline, and only the difference is a real cost. A shop comparing a covered project's load against zero will always decline.

Question 4: run both volumes

One project a year, 240 field hours. Attributable office hours: 10 amortized setup, plus 6 weeks at 2.25 equals 13.5, plus 3 for classification and pricing. Total 26.5 office hours against 240 field hours, which is 11.0 office hours per 100 field hours. Subtract the 4.0 baseline and the incremental load is 7.0 office hours per 100 field hours.

Convert to money without using money. In this shop an office hour costs about 60 percent of what a field hour costs, so 7.0 incremental office hours per 100 field hours equals 4.2 field-hour-equivalents per 100, or 4.2 percent added to labor cost. Labor runs about 45 percent of this shop's sell price, so that is 1.9 percent of the sell price. Against a 12 percent net margin, the capability consumes 1.9 divided by 12, which is about a sixth of the margin on that project.

Four projects a year, same size, 960 field hours. The setup amortization is unchanged at 10 office hours a year; it is now spread across four times the work. Weekly production settles at 1.25 hours across 24 active weeks, which is 30 hours. Classification and pricing, 3 hours each, is 12. Total 52 office hours against 960 field hours, or 5.4 office hours per 100. Subtract the 4.0 baseline and the incremental load is 1.4 office hours per 100 field hours.

Same conversion: 1.4 times 0.6 equals 0.84 field-hour-equivalents per 100, so 0.84 percent added to labor cost, 0.38 percent of the sell price, about 3 percent of the net margin.

The finding. The identical capability costs about five times as much per unit of work at one project a year as at four. Not because anything got cheaper, and not because the crew got faster in the field. The setup spread over four times the volume, and the weekly production hours fell by nearly half once the record-keeping stopped needing correction. Both effects push the same way and they compound.

Note what this comparison does not do: it does not compare a corrected covered-project figure against an uncorrected ordinary-work figure. The 4.0 baseline is this shop's real administrative load on all its work, measured the same way, and it is subtracted from both scenarios.

Question 5: what does it do to the wages of the people who are not on it?

This is the cost that does not appear in any spreadsheet and it is the one that persists after the project ends.

The determination rate is billed through, so paying it does not cost margin. But a technician who works six weeks at a determination base plus fringe and then returns to your ordinary rate has now seen a number, and so has everybody he eats lunch with. If your ordinary rate sits close to the determination, this is a non-event. If it sits well below, you have just run a very effective advertisement for what your people could be earning, and the pressure lands three months later in a raise conversation you did not plan for.

Size it before you bid rather than after. Compare the determination's base plus fringe for your main classification against what you actually pay that classification, as a ratio. A shop paying 0.9 of the determination has a manageable conversation ahead. A shop paying 0.7 of it should decide deliberately whether it wants that conversation, because the project will start it either way.

Both directions are real. A shop already paying at or above the determination gets the reverse effect: covered work confirms its wage structure to its own crew, and it competes well against shops whose numbers have to move.

What the shop decided

He declined the single project as offered, then spent two weeks finding out whether four were available: he registered for the state's bid notification service, called the two general contractors who work that district, and asked the community college's purchasing office what went out to bid annually. The answer was that four to six covered projects of that size passed within reach each year and that most were not competitively pursued by service shops.

He took the school project as the first of four, priced it with the incremental load carried at the one-project rate rather than the four-project rate, and treated the difference as the cost of entry rather than as a loss. That last choice is the one worth copying: he did not price the first project optimistically on the assumption the other three would materialize.

What would change this. If you operate in a state whose prevailing wage statute has a low threshold and reaches maintenance work, a meaningful share of the institutional work you already do may be covered, and the capability is not optional at any volume - the volume analysis becomes irrelevant and the only question is how fast you can build it. If a payroll service will produce certified payroll for a per-payroll fee, most of the fixed cost converts to variable and the break-even project count falls sharply; get that quote before running Question 1, because it can change the shape of the answer entirely. And if you are a sub, the general contractor usually collects and forwards your submissions with its own, though some agencies take subcontractor payrolls directly. Either way it does not sign your statement of compliance and does not own your classifications, so the burden shifts less than subs expect.

The failure mode. A shop takes one covered project a year, prices it like ordinary work, absorbs about a sixth of its net margin in unbilled office time, and concludes that prevailing wage work does not pay. What did not pay was doing it once. The tell is an owner who can name the compliance burden in detail and has never measured his own baseline administrative load, so he has no number to subtract.

How to verify you got this right

After your first covered project closes, reconstruct the actual office hours it consumed from your own records rather than from the estimate above, and recompute the per-hundred-field-hour figure. If the actual sits far above the estimate, the gap is almost always in correction time on technician records, which is fixable and falls fast with practice, rather than in the submission itself.

Then re-decide. The volume answer is not permanent. If two of your four expected projects did not appear, the capability is running at close to the one-project cost, and the honest options are to go find volume or to stop.

References

• Davis-Bacon Act, 40 U.S.C. 3141 to 3148, and 29 CFR Part 5 for contract provisions and payroll requirements on covered federal and federally assisted construction
• State prevailing wage statutes, which differ from the federal scheme in threshold, covered work and submission requirements
• See related: What Prevailing Wage Changes About a Job You Already Know, Certified Payroll and What It Actually Requires of You, A Prevailing Wage Job Comes Up, Should You Bid It