R&D tax relief •11 min read

Stage One Creative Services Ltd v HMRC

MJ

Matthew Jones ACA CTA

Last reviewed September 2026 · Editorial standards

Tribunal
First-tier Tribunal (Tax Chamber)
Decided
25 November 2024
Judge
Judge Anne Scott and Christopher Jenkins

The company won every issue. A month after Collins Construction, a second tribunal rejected the same two HMRC arguments about subsidy and contracting out. It also struck down two discovery assessments, because HMRC’s earlier guidance had established a practice the company had followed. That finding matters most in an enquiry: it is the only R&D decision in the published record in which paragraph 45 of Schedule 18 has been applied. Like Collins, it decided the old SME scheme.

What was at issue

Stage One Creative Services provides “engineering, construction and automation solutions for live events and installations” (paragraph 46) — the contracting pattern on our construction page. Its SME claims for the years ended 31 December 2017, 2018 and 2019 were challenged on four fronts.

Two were the arguments Collins had just defeated: whether the expenditure was subsidised within section 1138 CTA 2009, and whether it was incurred “in carrying on activities which are contracted out to the company by any person” within sections 1052(5) and 1053(4) (paragraph 10). For 2019 HMRC agreed that £3,600,829 was R&D expenditure (paragraph 89(a)); its objection was to the scheme claimed, not the science.

The other two concerned how HMRC reopened the older years. Discovery assessments were raised, not enquiries: £248,100.67 of tax plus interest for 2017 under paragraph 41 of Schedule 18 FA 1998, and £198,971.32 for 2018 under paragraph 52 (paragraph 90). That engaged two protections (paragraph 12): paragraph 44, on what an officer could reasonably have been aware of, and paragraph 45, the practice generally prevailing.

How the dispute got to a hearing

In 2016 HMRC enquired into earlier claims, met the directors on 18 March and closed by “authorising a substantial part repayment”; neither subcontracting nor subsidy was raised (paragraphs 65, 66). Later returns used the same format.

The 2017 return was filed on 10 August 2018, its enquiry window closing 9 August 2019; the 2018 return on 18 September 2019, closing 17 September 2020 (paragraph 211). The 2019 return arrived on 29 May 2020 and HMRC opened a compliance check on 28 July (paragraphs 70, 258). On 30 October the officer concluded that Conditions D and E of section 1052 were not met (paragraph 73). At a meeting on 24 November 2020 the officers “did not consider that the R&D had been sub-contracted” (paragraph 75) — a concession withdrawn on 18 December 2020 (paragraphs 79, 264).

Alternative dispute resolution produced an exit document on 10 November 2021 (paragraph 85). On 17 December 2021 HMRC’s note to the communication forum was read out: “HMRC not appealing Quinn but think Judge was wrong and HMRC not changing their view” (paragraph 87). The decision letter and 2018 assessment followed on 24 December 2021, and the 2017 assessment on 30 December (paragraph 90). The closure notice issued on 11 January 2022. The assessments had been appealed to HMRC by letters dated 10 January 2022, and the notices of appeal went to the tribunal on 21 January (paragraphs 91, 94).

It was heard by remote video on 20 to 22 November 2023 before Judge Anne Scott and Member Christopher Jenkins, and released on 25 November 2024 (paragraph 8 explains the delay). Charles Bradley of counsel appeared for the company, Francis Fitzpatrick KC and Sarah Black for HMRC.

What each side argued

HMRC’s substantive case was the one it had run in Collins: to provide “the deliverables provided for in the contracts, SOCS had to incur expenditure some of which fell within the definition of R&D expenditure. Had the contracts not been in place SOCS would not have incurred the expenditure” (paragraph 190). R&D being inherent, it need not be named (paragraph 191).

On discovery, HMRC’s position was that its interpretation had never changed. The November 2021 manual changes “simply improved CIRD and did not change HMRC’s interpretation” (paragraph 271). It had told the Chartered Institute of Taxation in November 2020 that any update would be “not a change in any way of HMRC’s long held interpretation” (paragraph 262). On paragraph 44, it argued the claim reports gave no contractual detail and did not address the tax issues (paragraph 217).

The company argued that its obligation was “to deliver the specified end product but not to solve any identified technical challenge or conduct any particular R&D” (paragraph 192). On discovery, “the centrepiece of our case” was “a radical change in HMRC’s published guidance” in November 2021 (paragraph 225). It called a specialist firm’s director of tax on what advisers and officers were doing, and the partner from the 2016 enquiry.

The provision the case turns on

Paragraph 45 of Schedule 18 to the Finance Act 1998 provides:

“No discovery assessment … may be made if—

(a) the situation mentioned in paragraph 41(1) or (2) is attributable to a mistake in the return as to the basis on which the company’s liability ought to have been computed, and

(b) the return was in fact made on the basis or in accordance with the practice generally prevailing at the time when it was made.”

Paragraph 44(1) is the separate protection. An assessment may be made only where the officer “could not have been reasonably expected, on the basis of the information made available to [him] before that time, to be aware of the situation mentioned in paragraph 41(1) or (2)”. That information includes the return and its accompanying documents.

The substantive conditions were section 1052(5) and (6) CTA 2009, and section 1138, which treated expenditure as subsidised “to the extent that it is otherwise met directly or indirectly by a person other than the company”.

What the tribunal decided

The appeal was allowed on all four issues, with nothing decided against the company. On subsidy it treated the Upper Tribunal’s remarks in Perenco about Quinn as “highly persuasive” (paragraph 162). It adopted Quinn’s reasoning with one minor exception (paragraph 181), and added a limit. On HMRC’s approach “no expenditure on R&D that is incurred in the course of delivering on a contract of any sort would be eligible for SME relief” (paragraph 178). The finding: “the R&D expenditure was not subsidised within the meaning of section 1138 CTA 2009” (paragraph 182).

On contracting out the key finding of fact was that “SOCS do not agree to carry out R&D or seek reimbursement of those costs. Any such expenditure is simply an incidental part of the delivery of a project and the clients did not undertake to reimburse such expenditure” (paragraph 166). Then: “the obligations in term of the contracts did not encompass R&D. The R&D was a means to an end, enured to the benefit of SOCS … and was an incidental part of delivery of the projects” (paragraph 208). It “was not contracted out within the meaning of section 1052 CTA 2009” (paragraph 209).

Then the discovery findings, taken in reverse order because the two were “inextricably interlinked” (paragraph 223). On the practice generally prevailing it applied Hargreaves v HMRC [2022] UKUT 34 (TCC) at [61]: a practice must be adopted by taxpayers and HMRC alike (paragraph 224). Comparing the manual before and after 30 November 2021: under the old version “there could be SME relief where there was a commercial contract and R&D was needed in order to deliver under the contract. The new version removed those factors and made it clear that only freestanding R&D could qualify” (paragraph 273). “HMRC’s approach changed in the latter part of 2019” (paragraph 276). The conclusion: “not only was there a PGP” — a practice generally prevailing — “but that SOCS’ tax returns were submitted in accordance therewith” (paragraphs 282, 283).

On paragraph 44, HMRC carried the burden (paragraph 285) and did not discharge it. Its own pleaded case turned against it. HMRC had pleaded, both before and during the hearing, “that only freestanding R&D was eligible for relief, and that had always been their stance”. On that footing “it would have been obvious from the R&D Reports that there was a potential loss of tax” (paragraph 295). The reports “would have sufficed to enable the hypothetical officer to conclude that the expenditure was both contracted out and subsidised” (paragraphs 296, 297).

The decision flags its own qualification: had paragraph 44 been decided by reference to the practice findings, “there was a discovery after the November 2020 meeting when the modus operandi was discussed” (paragraphs 298 to 301).

What HMRC did afterwards

HMRC has not appealed on the published record, and the forum minutes record that it accepts the decision. It amended CIRD81650 and CIRD84250 on 27 February 2025, the change log giving the reasons as “Updated to reflect FtT decisions” and “Redraft following FtT decisions”. Both carry the line “The guidance on this page has been amended to reflect HMRC’s view following recent First-tier Tribunal decisions”, and neither names the case.

In the Research and Development Communication Forum minutes of 6 May 2025 HMRC records that it “accepts the tribunal’s decisions and has updated its position and approach based on those”, while opening no correction route for companies already final. Members commented that this “could put companies that followed HMRC’s guidance in a worse position compared to those that did not”. HMRC acknowledged the point. It noted that the CIRD manual is internal guidance for its own teams that is also published, and that customers can take a contrary position and “appeal/challenge HMRC’s position”.

The same minutes set out HMRC’s own statement of the principle, as a point of clarification. It is “vital to the integrity of the self-assessment system that a tax position determined upon the practice generally prevailing at the time cannot be changed once it becomes final, subject to discovery and overpayment relief provisions”. HMRC added: “This principle applies even if a subsequent tribunal or court decision changes our interpretation of the legislation.”

What it changes for a claim being prepared now

  • A guidance change is a fact you can argue. Where HMRC rewrote its manual and then assessed an old year on the new reading, paragraph 45 is live. Four things need evidencing: the guidance in force when the return was filed, the wording that replaced it, that officers and the profession were both working to the old version, and that the return was filed on that basis. Manual text, trade-press commentary, forum minutes and a practitioner witness were all accepted (paragraphs 225 to 283).
  • Look at how the year was opened, not only the merits. Discovery assessments carry conditions, and on paragraph 44 HMRC carries the burden. A claim that is hard to defend technically can still survive if the year was reopened badly — see what happens if an R&D claim is rejected.
  • Keep the earlier enquiry file. The 2016 enquiry, in which HMRC examined the same work and paid without raising either point, mattered here (paragraphs 65 to 67, 278).
  • R&D as a means to an end still counts. The company solved problems in order to deliver, rather than researching for clients. Under the old scheme that sufficed.

For periods beginning on or after 1 April 2024 the question differs. Section 1133(2)(c) CTA 2009 asks whether “it is reasonable to assume, having regard to the terms of the contract and any surrounding circumstances, that the person intended or contemplated when entering into the contract that research and development of that sort would be undertaken”. Subsidised expenditure no longer reduces relief. See contracted-out R&D and, for a contractor in Stage One’s position, who claims when engineering is done under contract.

Where it sits against the other decisions

On the substantive limbs this follows Collins Construction and, behind it, Quinn (London) Ltd [2021] UKFTT 437 (TC). The strongest support on subsidy comes from outside the R&D code. In HMRC v Perenco UK Ltd [2023] UKUT 169 (TCC) the Upper Tribunal construed paragraph 8 of Schedule 3 to the Oil Taxation Act 1975, a petroleum revenue tax provision using the words “met directly or indirectly”. It held that “A’s payment is properly to be regarded as consideration for what A receives and not as a way of meeting B’s expenditure” (paragraph 74). At paragraph 75 it added that its approach “is very similar to that of the FTT in Quinn”. This tribunal treated those comments as “highly persuasive” (paragraph 162) rather than as binding.

Pointing the other way is Hadee Engineering, relied on by the officer in her conclusion letter (paragraph 74(d)) and distinguished on its facts. On discovery the companion decision is Realbuzz Group Ltd v HMRC [2025] UKFTT 493 (TC), where the taxpayer conceded the merits and still defeated the assessment under paragraph 44.

Has it been appealed?

No onward appeal has been reported, and no Upper Tribunal or Court of Appeal decision naming Stage One Creative Services appears on the published record as at the review date above. Applications for permission are not consistently published, so silence is not proof. HMRC’s position is better evidence: the forum minutes record that it accepts this decision and Collins Construction, and it amended the manual in February 2025.

This is a First-tier Tribunal decision on its own facts: it binds only the parties, sets no precedent and changes no legislation. See the register of R&D case law.

Sources

Every CIRD paragraph cited above (CIRD81650 and CIRD84250) is indexed on our CIRD reference index, with HMRC’s own title for each, a line on what it says and the date HMRC last revised it.

First-tier Tribunal decisions bind only the parties to them and set no precedent. This entry describes the decision as it stood at the review date above, as general information rather than advice on your circumstances — see our terms. For an answer on your own facts, talk to us.