Civil, structural, MEP, and product engineering. Some of your hours are research; most are not.
Design work that resolves a genuine technical uncertainty by analysis, modelling, or test can qualify under the four-part test. Drafting, permitting, and code compliance do not.
Photo by Marsumilae on Unsplash
Yes, for the part of the work that resolves real uncertainty. The credit does not reach engineering because it is difficult, or because it is billed. It reaches work on a business component where the information available to you did not establish whether the thing could be done, how to do it, or what the right design was - and where you resolved that by evaluating alternatives rather than by applying a method you already had.
That framing is why a transfer structure with no precedent in your details library, a drainage scheme for a site that defeats the standard approach, a strengthening scheme for a structure whose real capacity had to be established by investigation, or a product redesign proven by fatigue testing are all candidates, and why a permit set, a code check against published tables, and a bid package are not. Each candidate is tested against the IRC §41 four-part test on its own facts.
Then there is the second half of the problem, which is contractual rather than technical. Almost everything an engineering consultancy does is paid for by a client, so §41(d)(4)(H) has to be answered agreement by agreement before anything enters a base. The reported outcomes for design firms cut genuinely both ways, which is why we read the documents rather than the letterhead.
General and illustrative; eligibility depends on your facts.
Not activity categories - situations. If any of these read like a job you actually ran, there is very likely a qualified business component underneath it.
A long-span or transfer condition fell outside the framing you had built before. Alternative schemes were analysed and modelled, and one was carried only after the others were shown not to work.
Why it can qualify
Design uncertainty about the appropriate structural system, resolved by evaluating alternatives rather than by selecting a known detail.
Infiltration was poor and the outfall was constrained, so detention and treatment alternatives were modelled against measured site conditions, and the first layout failed the check before one was carried.
Why it can qualify
Hydraulic uncertainty set by site conditions, resolved by modelling alternatives rather than by applying a standard detail.
Settlement and capacity uncertainty forced two foundation systems to be developed and compared before either could be priced, and one was abandoned on the analysis.
Why it can qualify
Capability and method uncertainty in the engineering sciences. The abandoned scheme is part of the process, not wasted effort.
Capacity had to be established by investigation and analysis before a strengthening scheme could be selected, and the first scheme did not survive the second round of numbers.
Why it can qualify
The information available did not establish the method. Establishing it by test and analysis is what the statute asks for.
An unconventional conditioning or thermal-storage approach was modelled and evaluated because its behaviour in this occupancy and climate could not be predicted from the published data.
Why it can qualify
Technological in nature and genuinely uncertain. Routing a conventional system through a tight ceiling is coordination, not research.
A connection or facade assembly outside the tested range was analysed, mocked up, and instrumented because the manufacturer’s data did not cover the condition.
Why it can qualify
Design uncertainty resolved by physical testing. The mock-up materials consumed are supplies.
A new load case meant a lighter enclosure had to survive cycling it was never designed for. Three geometries were analysed, prototyped, and tested before one passed.
Why it can qualify
Performance and reliability improvement of a product, evaluated by analysis and physical test. Who owns the design decides whose credit it is - see the funding screen below.
Illustrative situations, not client work. Whether any of them qualifies for you depends on your facts, your agreements, and your evidence.
An engineering firm already produces the evidence a claim needs, in a form nobody files for tax: option studies with the schemes that were rejected, calculation sets with superseded revisions, model runs and their assumptions, prototype and test reports, mock-up photographs and instrumented results, and the meeting minute where the answer changed.
What is almost never in the record is the allocation - which hours went to the uncertain part of the project, and which went to documenting a settled design. A flat percentage applied across every project cannot answer that question, and it is the first thing an examiner asks about.
Representative activities that can meet the four-part test across structural, geotechnical, civil, MEP, and product engineering.
What spending counts toward the credit - tailored to a business whose costs are almost entirely people.
| Expense category | What goes into the base |
|---|---|
| Technical wages§41(b)(2)(A)-(B) | W-2 wages for engineers, designers, modellers, and analysts for time spent on qualified design work, plus direct supervision and direct support of it. |
| Contract research (65%)§41(b)(3) | 65% of amounts paid to U.S. specialist consultants and testing laboratories for qualified research performed on your behalf under an agreement entered into before the work. |
| Supplies§41(b)(2)(C) | Materials consumed in mock-ups, test assemblies, and prototypes built during design. Usually a small line for a consulting practice, larger for product engineering, and never depreciable equipment. |
| Computer rental§41(b)(2)(A)(iii) | Amounts paid for the right to use computers in qualified research, such as rented compute for a large simulation run. Per-seat software licences are not a §41 expense category. |
A directional shape for an engineering consultancy, not a benchmark. Wages carry essentially the whole claim, which is precisely why the hour-level allocation has to be defensible rather than assumed.
Seat licences for CAD, BIM, and analysis software are not a §41 expense category at all, however large the invoice. Neither is rent, insurance, or the overhead loaded into your billing rate: a fully-loaded cost model will overstate the credit every time.
Your project accounting is usually good enough to support an hour-level allocation, because it already separates phases and tasks. The work is mapping those tasks to business components and to the point at which the uncertainty was actually resolved.
Two of these are about the contract and three are about the hours. A claim that ignores either half is the kind that does not survive an examination.
§41(d)(4)(H)
Research is funded, and excluded, unless you both bear the financial risk of failure and retain substantial rights in the results. Payment that does not depend on success is funding, and so is an agreement that leaves you no substantial rights. For an engineering consultancy this is the whole ballgame, because almost everything is done under a client agreement. A fixed-fee agreement is a better starting position than a cost-reimbursement one, but it is only a starting position.
§1.41-4A(d)
The Eighth Circuit affirmed the Tax Court against a structural engineering firm in 2024 on contracts that were fixed-price, distinguishing proper performance from successful performance: an obligation to comply with applicable codes, or to work to a general standard of care, does not mandate success. What courts have looked for is express machinery - objective acceptance criteria, a right to reject non-conforming work, payment limited to accepted work, or a refund if benchmarks are missed.
§41(d)(3)(B)
Research relating to style, taste, cosmetic, or seasonal design factors is outside the permitted purposes. When an architect’s appearance decision hands you a real structural, envelope, or acoustic problem, it is your engineering that resolved it, not the look, that is the candidate. Architects face this exclusion far more often; it is covered in full on the architecture page.
§41(d)(4)(A), (B), (D)
Producing construction documents for a design that is already settled, permit sets, code checks, routine calculation to published tables, bid support, and construction administration are not qualified research. Adapting an existing design to a particular client’s requirement is separately excluded. On most projects this is the majority of the fee, which is why an honest claim here is a fraction of the hours rather than a percentage of the practice.
§41(d)(4)(C), (F)
Reproducing an existing design from its plans or specifications, such as rolling a proven design onto a new site, is duplication of an existing business component. And work performed outside the United States is excluded regardless of who employs the staff, which matters for practices using offshore drafting and modelling support.
Case outcomes for design firms have gone both ways on facts that look similar from the outside, which is the honest reason to read the agreements rather than to rely on a rule of thumb. Our note on funded research and contract clauses works through what the courts have actually looked at.
Most states with a corporate income or franchise tax run their own R&D credit off a similar research base, each with a different rate, cap, carryforward, and refundability - and a federal number multiplied by a state rate is not a state credit. For a practice with offices in several states, where the qualified work was performed matters to both numbers.
§174A matters here too, and independently of the credit. It restored immediate expensing of domestic research and experimental costs for tax years beginning after December 31, 2024, on a definition broader than the four-part test - so design costs that fall outside the credit can still be §174A expenditures. A study that only chases the credit leaves that on the table.
Summary only - the analysis is done per agreement and per business component, not per firm.
A hypothetical scenario to show how the pieces fit together. It is not a quote, projection, or promise of results.
Plus the §174A deduction on domestic research and experimental costs, on a definition broader than the credit’s.
Illustrative only. Figures are hypothetical and rounded; the federal credit commonly works out to roughly 6-10% of QRE depending on method, filing history, and the §280C election. Note how much the two screens above remove: that is what an honest engineering-firm claim looks like. This is not a quote or a promise of any result.
IRC §174A restores immediate, full expensing of domestic research and experimental costs for tax years beginning after December 31, 2024, on a definition broader than the §41 four-part test. For an engineering firm, that gap is the point: costs that do not clear the credit can still be §174A expenditures.
Tell us what your firm engineers and how the work is contracted, and we’ll separate the genuinely uncertain hours from the documentation - screening every agreement for funded research, and capturing §174A - reviewed and finalized by R&D experts and backed by Audit Protection. Contact us for pricing tailored to your study.
We typically reply within one business day.
