Can a government contractor claim the R&D tax credit?

Often yes, but not on everything. The credit excludes research funded by another person, including a governmental entity, so research performed under cost-reimbursement contracts, where the contractor is paid whether or not the research succeeds, generally does not qualify. Research under fixed-price development contracts, where the contractor bears the cost of nonconforming work, can qualify, and so can independent research and development the company funds itself. Because a contractor's portfolio can span several contract types, eligibility is a contract-by-contract determination, not a yes-or-no answer.

Does IR&D qualify for the R&D tax credit even though the government reimburses part of it?

It can. Independent research and development is, by definition under the federal cost principles, effort not sponsored by a grant and not required in the performance of any contract. It is the contractor's own money at the contractor's own risk, which is what the credit was designed to reward. The fact that a share of IR&D is later recovered through overhead rates on government contracts does not, by itself, make the research funded, because no research agreement obligated another party to pay for it. The characterization still has to be confirmed from the company's actual agreements.

What happens if a contractor claims the R&D credit on cost-reimbursement contract research?

The funded-research exclusion was written for exactly that fact pattern, and it is a point the government actively examines and litigates. A credit claimed on research the government paid for, win or lose, is exposed to disallowance under examination, typically years later and with interest. A government contractor also has a second federal reader: its contract cost submissions describe the same research, and inconsistent stories in the two filings invite comparison. The protective move is to sort the claim against the contract terms before an examiner does.

R&D Tax Credit for Government Contractors: Funded Research

Q: What makes research 'funded' under the R&D credit rules?

Courts ask two questions of each research agreement. First, was payment contingent on the success of the research, so the researcher bore the financial risk of failure? Second, did the researcher retain substantial rights in the results, such as the right to use what it learned in its own business? If either answer is no, the research is treated as funded to that extent and does not qualify for the credit. The answers come from the contract's actual terms, which is why the determination belongs to the contract file rather than the general ledger.

In July 1982, Fairchild Republic won the contract to build the United States Air Force its next trainer jet, the T-46A. The deal carried the terms fixed-price development always carries: the Air Force was obligated to pay only for work that met the contract's specifications. If the airplane fell short, the government could reject the work, make Fairchild correct it at Fairchild's own expense, or accept it at a reduced price. The T-46A first flew in October 1985, was canceled in 1986 with three aircraft built, and took the storied Republic plant in Farmingdale, New York down with it. It was the company's last airplane.

What the program left behind was not a trainer fleet. It was a tax case. When the government later argued that Fairchild could not claim the federal research credit on the T-46A work because the Air Force had paid for the research, the Federal Circuit read the contract and disagreed. The Air Force had promised to pay only for research that succeeded. Fairchild had carried the cost of every failure. That made the research Fairchild's own, and the credit followed the risk. Three decades later, that decision still draws the line through every government contractor's research budget. The line runs through the contract file, not the tax return, and a credit study built without the contract file has never looked for it.

One dollar, two ledgers

The research credit has been in the tax law since 1981, built to reward companies that put their own money at risk on work that might not succeed. Congress wrote one sentence into it that matters more to government contractors than the rest of the statute combined: research does not count, to the extent it is funded by any grant, contract, or otherwise, by another person, including a governmental entity.

Whether research is funded is not decided in the tax department. It is decided by the contract file: who bears the cost when the research fails, and who keeps the rights to what it produces. Those are contract terms. They were negotiated by the contracts office, often years ago, for reasons that had nothing to do with taxes, and they now control the tax answer.

That is the seam. A government contractor's research dollar is recorded twice. On the contract side it is either a direct cost of a specific contract or it is independent research and development (IR&D), the contractor's own self-funded work, recovered as an indirect expense under the cost principles that govern federal contracts. On the tax side the same dollar is a deduction, and possibly a credit. The contract-side classification and the tax-side eligibility run on the same underlying facts. Nothing makes them meet: the credit is computed in the tax file and the cost classification lives in the contract file, kept by different people, reviewed by different agencies, on different calendars.

Funded, or not

The test the courts apply asks two questions of each research agreement. First, is payment contingent on the research succeeding? A contractor that gets paid win or lose bears no financial risk, and research it is paid to perform regardless of outcome belongs, for credit purposes, to the customer who paid for it. Second, does the contractor keep substantial rights in the results, the right to use what it learned in its own business? Fail either question and the research is funded, and the credit is gone to that extent.

Run a typical defense portfolio through those two questions and it splits three ways. Cost-reimbursement work, where the government pays the contractor's costs as they are incurred, is generally funded: the contractor is made whole whether the research succeeds or not. Fixed-price development, where the contractor eats the cost of nonconforming work, is generally not funded: that is Fairchild, and the credit can be available. And IR&D is, by definition, effort no contract required and no grant sponsored. It is the company's own money at the company's own risk, which is exactly what the credit was written to reward, and it can remain eligible even though a share of it comes back to the company through its overhead rates on government work.

A contractor's portfolio can hold all three at once: fixed-price development on one program, cost-reimbursement work on another, IR&D underneath both. A portfolio like that turns “does our research qualify” from a yes-or-no question into a sorting problem, contract by contract and project by project, and the sort runs on documents that live in the contracts office, not the tax file.

What's at stake

The money runs in both directions, and both directions are expensive.

In one direction sits the credit that was never claimed. Companies that live on government work tend to absorb a simple, wrong rule of thumb: we do research for the government, the government pays for it, so none of it counts. The blanket answer surrenders the fixed-price development where the company carried the risk, and it surrenders the IR&D the company funded itself, year after year. The research credit is permanent law. For a manufacturer running real engineering spend it is recurring annual cash, and unlike most of what changed in the recent depreciation rewrite, which moves deductions in time, the credit is a permanent benefit. Leaving it unclaimed is leaving money on the table every single year.

In the other direction sits the credit that should not have been claimed. A credit study that samples payroll and projects but never reads the contracts will sweep cost-reimbursement research into the claim, research the government already paid for, win or lose. That is the exact fact pattern the funded-research exclusion was written to stop, it is a question the government actively litigates, and it fails at the worst possible time: under examination, years later, with interest. A government contractor carries one more discomfort here that a commercial company does not. The contract rulebook has its own symmetry principle, the credits rule, under which give-backs on costs the government paid belong, in fair share, to the government. A contractor whose cost submissions and whose tax filings tell two different stories about the same research has given two different federal readers a reason to compare them.

How it works

The failure is structural, not careless. The classification happens twice, by different hands. The contracts office sorts research costs for government accounting: direct to this contract, or IR&D to overhead. The tax side, often an outside credit-study provider, sorts research activity for the credit: qualified or not. Each sort is done competently on its own terms. Nobody runs the third sort, the one that decides everything: laying the credit claim against the contract terms, agreement by agreement, and asking who actually bore the risk and who actually kept the rights.

The past two years widened the gap. The 2025 tax law rewrote how research spending is deducted, restoring full expensing of domestic research costs and sending companies back into their research accounting to capture it. Small businesses were given a one-time window to apply the new treatment retroactively to 2022 through 2024, and that window closed on July 6, 2026. The window is gone. What did not expire is the point: the expensing is now permanent, the credit is permanent, and every fresh look at research accounting the new law forced is a fresh chance for the contract question to be asked, or skipped, again.

Where the rules stand now

The deduction side is settled and generous: domestic research costs are expensed in the year incurred, permanently. The credit side is settled and old: the credit is permanent law, and the funded-research line that governs contractors has been where Fairchild left it for thirty years. Risk and rights, read from the agreements. None of the recent legislation moved that line. For a government contractor the durable question was never whether the credit exists. It is which research the contract file says is yours.

What to do now

Read the contract file before the credit study, not after. A contractor that knows, agreement by agreement, where the risk of failure sat and where the rights ended up knows which of its research counts, before an examiner decides instead. The same reading protects the other flank: it shows which claimed dollars stand on cost-type work and should come out before anyone else finds them. The companies that get this wrong are not careless. They are organized the way almost every contractor is organized, with the contract answer and the tax answer in different rooms. The fix is not more research. It is one reading that covers both.

FactorTax reads a contractor's research spend through both ledgers at once, the federal contract cost rules and the tax rules, so the research credit rests on what the contracts actually say. To have your research read through both, schedule a discovery call.

The Research the Government Already Paid For