Fixed Price Contract Risk Starts at the Quote

The sentence is in the risk factors section of Graham Corporation's annual report, the part every public company must write about what could go wrong: "the original cost estimates in these or other contracts prove to be inaccurate, or the contracts do not permit us to pass increased costs on to our customers, our profitability may decrease or losses may be incurred". Graham makes vacuum and heat transfer equipment to order. I went looking for that sentence in other companies' filings and found a version of it in almost every custom manufacturer that files one. Fixed price contract risk is the one risk in the annual report that the sales and estimating organisation owns entirely, and the filings describe it as if it were weather.

I went through thirty two of them for this piece, every annual report filed since June 2025 by an engineered product manufacturer or heavy contractor that uses the words fixed price and cost estimates together, pulled from the SEC's full text search and classified by industry code. What follows is what they say, what they blame, what almost none of them name, and the one thing the auditors do after the price is fixed that nobody does before it.

In brief

• Fixed price contract risk is the risk that the cost of doing the work exceeds the price you agreed before you did it. The buyer is protected; the seller carries the difference. That is what "fixed price contracts place more risk on the contractor" means, and every filing agrees.
• Thirty two public manufacturers and contractors carry a warning about it in their latest annual report. Several carry the same sentence word for word, because the lawyers copy each other.
• The filings blame inflation, tariffs, steel, subcontractors, labour productivity, weather and scope changes. Most treat the estimate as something that "may prove inaccurate", like rain. Five describe estimating as work that people do.
• Where the reports give numbers, they are large: fixed price is 90 percent of Moog's over-time sales; $263.6 million of Tecnoglass's 2025 revenue; Textron expects a $60 million to $110 million charge on one programme "from when the program was bid".
• After the price is fixed, the cost estimate becomes a Critical Audit Matter, tested by outside auditors. Before the price is fixed, the same estimate is a spreadsheet nobody has tested. That asymmetry is the whole problem, and it is fixable.

What fixed price contract risk is, and who bears it

A fixed price contract pays the seller an agreed sum for an agreed scope, whatever it ends up costing to deliver. Oshkosh, which builds fire apparatus and defence trucks on such contracts, explains the plainest form in its annual report: "Under firm, fixed-price contracts with the U.S. government, the price paid to the Company is generally not subject to adjustment to reflect the Company's actual costs, except costs incurred as a result of contract changes ordered by the U.S. government." The softer form has an escalator: "Under fixed-price contracts with an economic price adjustment, the price paid to the Company can be adjusted upward or downward from the stated contract price based on cost indexes of labor or material that are specifically identified in the contract."

Textron states the consequence in one line: "we receive a fixed price irrespective of the actual costs we incur, and, consequently, we absorb any costs in excess of the fixed price." The buyer has transferred the cost risk to you, and paid for the transfer with a price that is usually higher than a cost plus arrangement would produce. Matrix Service, which builds storage tanks and terminals, says exactly that: "Fixed-price contracts typically present opportunities for higher margins, but carry a greater risk in terms of profitability because cost overruns may not be recoverable."

So the risk is not that fixed price contracts are bad. Custom manufacturers live on them; there is no other way to sell a thing that does not exist yet to a buyer who wants one number. The risk is specific and it has a location: the gap between what you estimated and what it cost, created on the day you quoted.

Firm fixed price versus fixed price with adjustment

The filings use several names for what sounds like one thing, and the differences are exactly the levers a seller has. A firm fixed price contract is the pure form Oshkosh describes: the price does not move for anything except a change the customer orders. A fixed price contract with economic price adjustment ties part of the price to a published index for a named material or labour category, so a steel spike moves the price by a formula rather than by an argument. BWX Technologies describes a third form, "fixed-price incentive fee contracts that provide for reimbursement of allowable costs incurred plus a fee", which shares overruns and underruns with the customer up to a ceiling; Matrix Service describes the same idea from the contractor's side, "we may agree to share with a customer a portion of any savings we generate while the customer agrees to bear a portion of any increased costs we may incur up to a negotiated ceiling."

For a custom manufacturer quoting an engineered product, the practical point is that the form is negotiable and the negotiation happens in the proposal, not in the contract. An escalation index, a validity period with a stated rule for what happens after it, a materials basis date, and a contingency line for the things you genuinely cannot know are all ways of moving a firm fixed price towards an adjusted one. Whether any of them appear depends on who wrote the quote and whether the company's quoting system tells them to. In most of the filings that blame tariffs, the honest sentence would be that the proposal did not carry an adjustment clause, and nobody at the quoting desk had a rule that said it should.

The sentence everyone uses

Read enough of these filings and the same sentences start to appear in different companies' reports. Babcock & Wilcox and BWX Technologies, which separated in 2015, still carry identical wording a decade later: "if our actual costs exceed the costs we estimate on our fixed-price contracts, our profitability will decline, and we may suffer losses." Both also say fixed price contracts "entail more risk to us because they require us to predetermine both the quantities of work to be performed and the costs associated with executing the work."

Orion Group, Great Lakes Dredge & Dock and Mirion Technologies, a marine contractor, a dredger and a nuclear instrumentation company, share "fixed-price contracts carry inherent risks, including risks of losses from underestimating costs". CECO Environmental and Fuel Tech share "the revenue, cost and gross profit realized on a fixed-price contract will often vary from the estimated amounts because of unforeseen conditions or changes in job conditions and variations in labor and equipment productivity over the term of the contract." ESCO Technologies and Moog share "we enter into fixed-price contracts, which could subject us to losses if we have cost overruns."

None of this is careless. Risk factor language is written by counsel, reviewed by counsel, and borrowed from filings that survived scrutiny. But it means the description of the risk has been optimised for the reader who might sue, not for the reader who might fix it. It tells you the risk exists. It does not tell you where it is made.

What the filings blame

I tallied the causes each report offers for a fixed price contract going wrong. The list is consistent, and it is a list of things that happen to the company after the price is signed.

Every row but the last is downstream of the quote. Several are real and uncontrollable: nobody at a switchgear maker sets tariff policy. But notice what the list does to the reader. It puts the risk outside the building. The estimate, where the price was actually decided, appears in most filings only as a passive thing that "may prove inaccurate", in the same grammatical mood as the weather.

The five that name the estimate

A handful of companies write about estimating as an activity. Powell Industries, which builds custom switchgear and is the cleanest example of a fixed price manufacturer I know, puts it in the active voice: "Failure to place competitive bids and adequately project future costs may result in losses on our fixed-price contracts with customers." Ducommun, an aerostructures supplier, is blunter: "We use estimates when bidding on fixed-price contracts," and "In many cases, we make multi-year firm, fixed-price commitments to our customers, without assurance that our anticipated production costs will be achieved."

Thermon, which makes heat tracing systems, locates the problem at the bid rather than the job: tariffs "may limit our ability to accurately estimate costs for our fixed-price contract bids." Astronics lists the failure modes in order and puts estimating among them: "Our failure to anticipate technical problems, estimate delivery reductions, estimate costs accurately or control costs during performance of a fixed-price contract may reduce our profitability or cause significant losses on programs."

Matrix Service is the only one of the thirty two that names the people. "Competent and experienced engineers, project estimators, project managers, and craft workers are especially critical to the profitable performance of our contracts, particularly on our fixed-price contracts where superior design and execution of the project can result in profits greater than originally estimated". And it says where the price comes from: "Fixed-price contract prices are established based largely upon estimates and assumptions relating to project scope and specifications, personnel and productivity, material needs, and site conditions."

That last sentence is, without meaning to be, a description of a quoting desk. Scope and specifications: the reading. Personnel and productivity: the rules. Material needs: the takeoff. Site conditions: the assumptions. Matrix's lawyers have written down the four things a quoting system has to get right, in a paragraph about why the company might lose money.

The numbers, where the reports give them

Most filings describe the risk without sizing it. A few do, and the sizes are worth reading slowly.

Moog: "In 2025, fixed-price contracts represented 90% of our over-time sales that we account for using the cost-to-cost method." Nine dollars in ten of that revenue rests on an estimate made before the work began. And Moog says where the losses concentrate: "Loss reserves are more common on firm fixed-price contracts that involve, to varying degrees, the design and development of new and unique controls or control systems to meet the customers' specifications." New, unique, to the customer's specification. That is the definition of custom work, and it is where Moog's own accountants expect to find the losses.

Tecnoglass, an architectural glass and window maker, had "$263.6 million of the Company's total revenues for the year ended December 31, 2025" from fixed price contracts, according to its auditor. Textron, on a single programme: "Upon award of the LRIP option, which is largely fixed price, we expect to record an unfavorable cumulative catch-up program adjustment, reflecting higher costs than originally anticipated from when the program was bid, in the range of $60 million to $110 million." From when the program was bid. The filing does not say the costs rose; it says the bid did not anticipate them.

CPI Aerostructures went to its customer to reopen a price it had fixed six years earlier: "On May 7, 2025, the Company submitted to The Boeing Company a Request for Equitable Pricing Adjustment on the Boeing A-10 program addressing higher manufacturing costs on its 2019 firm fixed price contract." A request for equitable adjustment is the polite name for asking the buyer to share a loss the seller agreed to carry. Some buyers say yes. The point is that the number at stake was set in 2019, by an estimator, in a workbook.

Fixed price contracts, pros and cons, from the seller's side

The filings are honest that the arrangement cuts both ways. Astronics: "Depending on the fixed price negotiated, these contracts may provide us with an opportunity to achieve higher profits based on the relationship between our costs and the contract's fixed price." DBM Global, the structural steel fabricator inside INNOVATE Corp, says a fixed price contract "may benefit from cost savings but be unable to recover any cost overruns." ESCO Technologies puts the two sides in one sentence: fixed price contracts "enable us to benefit from performance improvements, cost reductions and efficiencies" while subjecting the company "to the risk of reduced margins or incurring losses if we are unable to achieve estimated costs and revenues."

So the pros are real: a fixed price rewards a shop that estimates well and executes better than it estimated, and the buyer pays a premium for certainty that a disciplined seller keeps. The cons are equally real and less often stated in full. Textron notes one that rarely makes the list: "fixed-price contracts generally require progress payments rather than performance-based payments which can delay our ability to recover a significant amount of costs incurred on a contract and thus affect the timing of our cash flows." An overrun on a fixed price job is not only a margin problem; it is money out of the door for months before the milestone that pays it back.

The asymmetry that matters most is not between upside and downside. It is between the two ways of getting the upside. A company can win on a fixed price contract by executing better than it planned, which is operations, or by planning accurately in the first place, which is quoting. Every filing in the sample invests in the first. Almost none describes any investment in the second, and the two are not the same skill or the same people.

Audited after, unaudited before

Here is the asymmetry that made me write this piece. Once a fixed price contract is signed, the cost estimate stops being a quoting artefact and becomes an accounting one. Revenue on long contracts is recognised as costs are incurred against the total estimated cost, so the estimate drives reported profit every quarter until the job ends. That makes it a Critical Audit Matter, the category auditors reserve for the judgements that most affect the financial statements.

Tecnoglass's auditor describes the procedures: testing "controls over the determination of estimated costs to complete fixed price contracts and controls over management's review and approval" of the costs allocated to each project, and "performing a comparison of the originally estimated and actual costs incurred". Babcock & Wilcox's auditor identified the same matter "due to the significant management judgment involved in estimating total costs and profit". Koil Energy's auditor cites "the complexity of these estimates and exercise of significant judgment by management". The phrase "estimated costs to complete" appears in 170 annual reports filed in the fifteen months I searched.

So after signature: documented controls, management review and approval, an outside firm comparing the estimate with the actual, every year. Before signature, at the moment the number is actually chosen and the risk actually created: an estimator, a workbook with forty tabs, rules that live in one head, a review that looks at the total for twenty minutes, and no comparison of estimate with actual because in most companies nobody has ever joined the two tables.

The auditor's procedure, "a comparison of the originally estimated and actual costs incurred", is the feedback loop I keep writing about, and it is the one control the quoting desk never has. The company is required to run it on the cost to complete, after the fact, for the accountants. It is not required to run it on the quote, before the fact, for itself. Most do not.

Why fixed price contracts place more risk on the contractor, precisely

The phrase people search for is true, but it is worth being exact about where the extra risk sits, because that is where the controls should go. A quote at a custom manufacturer passes through nine stations, and the fixed price risk is created at four of them.

1. Reading. The buyer's document, hundreds of pages written by an engineer who does not work for you, decides scope. A requirement missed here is a cost carried for free. On a fire apparatus request we studied, a foam system was plumbed and bracketed on three pages and never specified; on a crane section, one leftover sentence required nuclear certification. Neither is inflation. Both are fixed price losses waiting for a signature.
2. Scope and assumptions. Where the drawings and the specification disagree, someone decides, usually silently. Matrix's "estimates and assumptions relating to project scope and specifications" are made here, and they are the assumptions the change order argument will later be about.
3. The cost build. The workbook. Rates, hours, bought outs. This is the station the filings mean when they say "estimate", and it is the least risky of the four, because arithmetic is the thing companies are already good at.
4. The price. Margin against risk. This is where the escalation clause, the validity period and the tariff contingency either exist or do not. Powell's filing notes that "fixed-price contracts may prohibit our ability to charge the customer for the increase in raw material prices". Whether they prohibit it was decided by whoever wrote the quote.

Everything in the filings' list of causes lands on one of those four stations. Tariffs land on station four, where an economic price adjustment clause would have caught them. Subcontractor pricing lands on station three, where a quote older than its validity was used. Scope changes land on station two, where the assumption was never written down and so cannot be pointed to. Even weather lands on station four, as a contingency that was or was not priced.

A pre-bid control set for fixed price contract risk

If the estimate is going to be audited after signature anyway, the useful move is to put equivalent controls before it, where they can still change the number. These are the five we run when we audit a quoting system, recast as the controls a CFO would recognise.

Two more belong on the price itself rather than the process. Write the escalation rule into the quote, so that "the contracts do not permit us to pass increased costs on" stops being true by default; we showed a public case where a manufacturer held a price for 260 days against a 180 day validity because nothing in the proposal said what happened on day 181. And log the assumptions on the quote where the document was silent or contradicted itself, because a change order argument is won by whoever can point to the sentence.

None of this removes the risk. Steel will move, tariffs will change, a subcontractor will fail. What it does is separate the part of fixed price contract risk that is genuinely external from the part that was manufactured at the quoting desk, so that the annual report's list of causes stops being used as an excuse for both. That is the part a company can price, and the part this article is about.

What this means for a CFO reading the risk factor

The risk factor in your own filing was written to protect the company from the reader. Read it once as an operating document instead. If it says estimates "may prove inaccurate", ask who makes them, in what tool, checked by whom, and compared with what. If the honest answer is one senior estimator, a workbook, a twenty minute review and no comparison, then the sentence in the filing is not a disclosure. It is a description of a process nobody has instrumented.

The companies that name the estimate, Powell, Ducommun, Thermon, Astronics and Matrix, are not weaker for it. They are the ones whose counsel was told something true by the people who quote. That is the first sign of a company that treats its quoting as a system rather than as a person, and it is the kind of company we build engines for.

How the filings were read

The sample is every Form 10-K filed between June 1, 2025 and September 18, 2026 that the SEC's full text search returns for six phrasings that pair fixed price with cost estimates, overruns, estimating or profitability. That produced 193 filers. I classified each by its SIC code using the SEC's company records and kept the 68 manufacturers plus the heavy and specialty contractors whose work is quoted the same way, then read the fixed price sentences from the 32 whose products are engineered to a customer's specification. Every quotation above is verbatim from the filing, and the source note at the top of this article lists the company, form and filing date for each. The count of 170 reports containing "estimated costs to complete" is the search engine's own total for that exact phrase across all 10-K filers in the window, not only manufacturers.

Two limits are worth stating. Risk factor language is written to be complete rather than accurate, so a company that does not name tariffs is not one that is unexposed to them, and a company that names the estimate is not necessarily one that controls it. And the sample is public companies, which are the largest in each segment; a $200 million family owned precaster does not file, and its quoting desk is, if anything, less instrumented than the ones described here.

Where software fits

A quoting engine changes the first two stations. It reads the whole package and cites every requirement to its page, so the foam system and the nuclear sentence are on the list before the price is built. It puts the contradictions between drawings and specification in front of a person, so the assumption is made on purpose and recorded. It prices in the company's own rules, mined from its own archive, which is the same thing as having written them down. And because it is tested blind against filed quotes under a protocol we publish, it comes with the estimate versus actual comparison built in from the first day.

It does not decide the margin, and it does not know what tariffs will do. Those stay with you, as they should. What it removes is the part of the risk that was never external at all, the part that Matrix's lawyers described by accident: scope and specifications, personnel and productivity, material needs, and the assumptions nobody wrote down.