R&D tax credits for engineering firms

Engineering firms sit awkwardly against the R&D definition, because the sector’s ordinary work is technical. Solving hard problems to a deadline is the job. Difficulty is not the test. The test is whether a competent professional could readily resolve the problem with knowledge already available in the field. Many engineering businesses have both kinds of work in the same year: a qualifying core, wrapped in delivery that is skilled, chargeable and outside the claim. A profitable company keeps 14.7p to 16.2p per £1 of qualifying spend under the merged scheme, so that boundary decides real money.

What engineering work qualifies as R&D?

The shapes differ by discipline.

• Design-and-build and product engineering. A constraint set established methods cannot satisfy together: weight against strength, temperature against cost, throughput against tolerance. Also substitute materials, where no performance data exists for your application and behaviour has to be established rather than looked up.
• Civil and structural. Ground that does not behave as the site investigation predicted, a form or span with no reliable precedent, temporary works whose behaviour has to be proved by instrumentation. More on our construction page.
• Mechanical. Fatigue, sealing, vibration and thermal behaviour published models do not predict for the duty in question; machinery pushed past the envelope the field has characterised.
• Electrical, control and instrumentation. Control of a system whose dynamics cannot be deduced in advance; power electronics and EMC behaviour at the edge of established practice; retrofits to legacy plant where the combined behaviour is unknown until it runs.
• Process engineering. Scaling from bench or workshop to production, where yield, tolerance and repeatability that held on the small version fail on the line. The Guidelines meet this directly: uncertainty will often arise from turning something already established as scientifically feasible into a cost-effective, reliable and reproducible process (paragraph 13); our manufacturing page develops it.

Attempts that failed still qualify: the relief follows the work to resolve the uncertainty, not the outcome. An iteration record of what was tried, what broke and what changed next is the best evidence a claim can have, the pattern set out in robotics prototyping and technological uncertainty.

Engineering also shades into sectors we cover separately: aerospace and defence, where certification standards drive the development, and cleantech and energy for scale-up from bench to pilot plant.

Where the qualifying line falls

Integration claims meet a standard objection: that each component came from a catalogue. The Guidelines answer it. There will be uncertainty if a competent professional cannot readily deduce how the separate components or sub-systems should be combined to have the intended function (paragraph 30), even where the principles for their integration are well known. The limit sits alongside it: assembling components to an established pattern, or following routine methods for doing so, involves little or no uncertainty (paragraph 29). A machine builder combining a servo drive, a vision system and a PLC as that firm has for a decade is doing skilled work with a predictable answer.

Routine application of established engineering practice does not qualify, and that covers a great deal of excellent work:

• Design to code. Published design codes, standard calculation methods and manufacturers’ data applied to a new duty, including simulation confirming a margin already understood.
• Made-to-order work on the firm’s established design approach: a one-off machine dimensioned for a new customer, where nothing about the engineering was in doubt. Work that only brings a company into line with what the field already knows is not an appreciable improvement (paragraph 24).
• Value engineering: the same specification at lower cost using known products. Commercial gain, not an advance in technology.
• Commissioning and tuning. Paragraph 14 is explicit: improvements, optimisations and fine-tuning which do not materially affect the underlying science or technology do not constitute work to resolve scientific or technological uncertainty.

The end of a project matters as much as its start: R&D ends when the knowledge is codified in a form a competent professional can use, or when a prototype with all the functional characteristics of the final product or process is produced (paragraph 34). We say no to work on this list regularly, including on projects a client was confident about.

Does precision machining or toolmaking qualify?

Subcontract machining and toolmaking sit at the hardest end of this boundary, because the work is exacting and the answer is usually already known. Holding a tighter tolerance by applying established fixturing, tool paths and metrology is skilled production: routine adaptation of an existing process does not advance overall knowledge or capability, even where it is new to the company or to its trade (paragraph 22).

The claim appears where the material, the geometry or the duty puts the process outside what the field has characterised — machining a substitute alloy whose behaviour under the required cut is not published, or holding a form whose distortion cannot be predicted from existing data, so the process window has to be established rather than looked up. Paragraph 13 puts the same test to production: uncertainty often arises in turning something already established as scientifically feasible into a reliable and reproducible process. The evidence takes the same shape as the rest of an engineering claim — first-off inspection records, the trials that did not hold and the reason they did not, and the settled parameters at the end.

Who claims when the engineering is done under contract?

For a contract engineering firm this is often the largest question in the claim. For accounting periods beginning on or after 1 April 2024, the customer claims where it intended or contemplated, when the contract was made, that R&D of that sort would be done. Where it did not, the contractor claims in its own right. A contractor whose customer carries on no trade within the charge to UK tax — an overseas customer with no UK trade, say — can also claim in its own right; a UK sole-trader customer is within the charge to income tax, so that route does not apply there.

Three shapes recur in engineering. On design-and-build for a client, where the contract, its scope and its testing regime describe the development itself, the client contemplated the R&D and claims it, taking 65% of what it pays an unconnected engineering firm. On subcontracted detail design — a package let against a performance specification, priced as a deliverable, with no technical unknown named in the tender — the firm that then has to resolve genuine uncertainty claims on its own costs. Own-product development sits outside the question entirely: nothing was contracted out, so the company claims. The same R&D cannot be claimed twice, so relief claimed on the wrong side of a contract is an incorrect claim. Read contracted-out R&D before you assume the claim is yours, or the customer’s.

Discipline by discipline: inside and outside the claim

Which costs go into an engineering claim?

Staff costs apportioned to development time, covering workshop, test and site engineers as well as the design office. Materials consumed in prototypes, rigs and trial builds qualify as consumables where they are not sold; materials absorbed into something you sell in the ordinary course of business fall outside, which catches firms whose prototype ships as the deliverable. Where the unit was always going to ship, the boundary reaches development time too, not only materials: can I claim R&D tax relief on a prototype that is later sold? Agency staff enter as externally provided workers at 65% of payments to unconnected providers, and only so far as their earnings are within UK PAYE and Class 1 NIC — where any part of a worker’s earnings is UK-payrolled, all of them qualify. Unconnected subcontracted development enters at 65% too; connected parties are restricted instead to the lower of the payment and the other party’s own relevant expenditure. Software and cloud costs used in the R&D count.

No capital expenditure qualifies, whatever the machine or test rig cost, though capital spending on R&D can attract R&D allowances instead; rent and patent costs qualify for neither. Two location restrictions apply: subcontracted R&D counts only where the work is done in the UK, and externally provided workers only so far as their earnings are within UK PAYE and Class 1 NIC. Overseas R&D costs covers the narrow exception.

What is an engineering claim worth?

Take £100,000 of qualifying spend. The merged scheme turns it into a £20,000 expenditure credit at 20%, and because that credit is itself taxable, what the company keeps depends on its rate: £15,000 at the 25% corporation tax rate, £14,700 where augmented profits fall between £50,000 and £250,000 and the 26.5% marginal rate applies, and £16,200 at the 19% rate or for a loss-making company, subject to the PAYE cap. Profitable firms are the more common case in engineering.

The alternative is narrower than it looks. A loss-making SME whose relevant R&D expenditure reaches 30% of its total relevant expenditure takes up to £26,970 on the same £100,000 through ERIS. The denominator is the catch: total relevant expenditure is broadly the trading costs in the accounts for the period, not the R&D ones alone, and connected companies enter both halves. An engineering business buying materials and letting subcontract packages rarely clears 30%.

Grant funding no longer reduces relief under the current schemes; see grant funding and R&D tax relief. Run your own figures through the claim value calculator, and the ratio through the ERIS intensity calculator.

Our Midtec Products case study works a real engineering claim: a £40,000 benefit across staff, agency and material costs, on a retrofit emissions device developed after new DEFRA eco criteria. A change in industry legislation is one of the most reliable triggers for qualifying work.

What evidence does an engineering claim need?

The records most firms already produce, kept with the claim in mind rather than filed against the job cost. Design reviews and drawing revisions show the constraint being fought. Test reports, failure investigations and non-conformance records show the uncertainty being worked through, with dates attached. A trade study that closed an option because the data would not support it evidences an unresolved uncertainty better than any later narrative.

Apportionment is the harder half, because engineers move between the qualifying core and ordinary delivery inside the same week. HMRC’s guidelines for compliance accept an estimated proportion of known expenditure where the estimate is arrived at using evidence and reason and the apportionment basis is recorded. Tie the records to a project boundary with a stated advance and uncertainty, then write the Additional Information Form from them rather than from memory; it has been mandatory for claims made on or after 1 August 2023, in practice 8 August 2023. HMRC checked around one in six R&D claims (17%) in 2023-24, the most recent year it has published. See what records an R&D claim needs and HMRC R&D enquiries.

What to bring to a first conversation

You do not need a prepared claim. Five things make the first hour productive:

• A shortlist of projects, not contracts. Two or three pieces of work where the answer was genuinely not known at the start.
• The engineer who led each one. HMRC anchors a claim in the judgement of a competent professional, and we interview them directly.
• The contracts behind that work, or the tender documents. Who claims turns on what they say.
• A rough cost shape: development headcount, agency use, subcontracted work, prototype materials.
• Your period dates and claim history. First-time claimants, and companies that have not claimed in the three years ending with the notification deadline, must notify HMRC within six months of the end of the period of account or the claim is invalid. See claim notification.

Talk it through with a chartered adviser

LimestoneGrey is a firm of Chartered Tax Advisers and Chartered Accountants specialising in R&D tax relief, regulated by ICAEW. We prepare engineering claims that are meant to be checked: costed carefully, evidenced from design and test records, signed off by a chartered adviser, with enquiry support included as standard and the fee agreed before work starts. Part of that work is telling you which projects do not qualify.

For a straight view on your development work, and on which side of your contracts the claim sits, call 0330 223 4 223 or send us a message.

Sources

• Guidelines on the meaning of R&D for tax purposes — paragraph 13 on turning feasibility into a reliable process, 14 on fine-tuning, 22 on routine adaptation of an existing process, 24 on appreciable improvement, 29 and 30 on combining standard technologies and the limits of system uncertainty, and 34 on where R&D ends.
• CIRD161000 — the intended-or-contemplated test and customers outside UK corporation tax; CIRD162000, HMRC’s worked examples.
• GfC3: Recommended approach to claims and record keeping (part 5) — estimates arrived at using evidence and reason, and recording the apportionment basis.
• Check what R&D costs you can claim — the cost categories, the 65% rule, and the exclusion of capital, rent and patent costs.
• R&D tax relief: the merged scheme and enhanced R&D intensive support — the 20% credit and the ERIS conditions.