APPG Evidence Session: R&D Tax Relief

Founders warn that two to three years of near-constant change to the UK’s R&D tax relief regime, combined with an aggressive HMRC compliance stance, is having a chilling effect on legitimate claims. Successive governments have deterred innovators, inflated costs and in some cases driven R&D-intensive businesses to relocate abroad.

The All-Party Parliamentary Group (APPG) for Entrepreneurship recently held a quick-fire evidence session to investigate how recent reforms to the R&D tax relief system are affecting founders and finance leaders in practice.

Chaired by Lord Marks of Hale CBE, a Member of the House of Lords, founder and investor, the session was framed by opening remarks from Iain Butler of Buzzacott and brought together founders, CFOs, sector associations and tax specialists from across life sciences, deep tech, software and consumer-facing R&D.

While the session was held under the Chatham House Rule, we have summarised the core themes and policy ideas discussed below.

A scheme reshaped beyond recognition

A broad consensus emerged that the R&D tax relief landscape has been transformed in a remarkably short period. The merged Research and Development Expenditure Credit (RDEC) scheme, the mandatory Additional Information Form, claim notification rules, reduced SME rates, new overseas restrictions, the rewritten contracted-out rules, and the introduction of a 30% R&D-intensity threshold have all landed in the same window — accompanied by a substantially expanded HMRC compliance regime.

The cumulative effect, participants argued, is a system that legitimate claimants now struggle to navigate with confidence.

A chilling effect on genuine claimants

Several participants described clients and peers who have simply given up on the scheme because the compliance burden and the risk of a retrospective investigation now outweigh the benefit. Others described being counselled by advisers not to claim.

A relief designed to encourage R&D is increasingly being treated as a source of risk to be managed rather than an incentive to be planned around. Founders described running monthly forecasts around the intensity threshold and changing real investment decisions to stay above arbitrary lines — the opposite of what a well-designed incentive should produce.

The Advance Assurance gap

A recurring theme was the lack of meaningful certainty before money is spent. Participants pointed out that Advance Assurance, in its current form, is too narrow to be useful: it is unavailable to companies that have claimed before, and is closed off to programmes above a certain threshold.

One founder, whose group has registered around 60 patents and raised over £150 million, described being refused Advance Assurance on multiple occasions, despite the scale and complexity of the programmes being precisely the reason certainty was needed. The frustration was widely shared. Where Advance Assurance is most valuable — for large, novel, multi-year programmes — it is least available. It should be noted that HMRC has announced an Advance Assurance Pilot, although its scope is limited at this stage. The assurance offers advice on up to two areas, but it is important to note that securing assurance on these points doesn’t protect the rest of a claim, and HMRC can still audit other areas. Applicants cannot appeal a refusal, and the assurance is strictly limited to the lifespan of the reviewed project.

Government support for SMEs has historically been generous compared to international competitors. Now, however, the UK’s support is more on par with international comparators — meaning that the UK is not providing the same draw it used to, and participants are increasingly looking abroad. Participants pointed enviously to jurisdictions such as Canada, where API-based filing, dedicated technical specialists on the assessor side, and clearer feedback loops give claimants something closer to certainty. One adviser noted a client actively redomiciling their R&D-heavy business to Canada for precisely this reason.

Retrospective investigations and the “brown envelope”

The most visceral concern raised in the session was retrospective investigation. Several founders described compliance checks lasting from eighteen months to two and a half years, ending in either full or near-full vindication — but at enormous cost in time, adviser fees and board distraction.

A founder of a smart materials business shared a particularly stark account: a retrospective enquiry into 2021 and 2022 claims that ultimately recovered around £200 out of roughly £650,000 in credits awarded, but which stretched on for years and disrupted a subsequent legitimate claim worth close to £450,000.

For founders who have claimed consistently over five or more years, the cumulative exposure to a retrospective challenge can run to hundreds of thousands of pounds — enough to bankrupt the business or force an emergency raise. Participants called for:

1. A fast-track Advance Assurance route that survives across multiple claim years;

2. Greater transparency about why a claim has been selected for review — and that these reviews should be risk-based, rather than purely process-driven; and

3. A clearer statute of limitations on retrospective enquiries for non-fraudulent cases (one to two years was suggested).

There was sympathy for HMRC’s anti-fraud mission, but a strong sense that the current posture treats every claimant as a suspect by default.

Sector expertise — or the lack of it

Several participants raised questions about HMRC’s enquiry process and the lack of subject-matter expertise within HMRC’s compliance teams. In life sciences, founders described HMRC caseworkers as being unfamiliar with standard industry business models and challenging claims that are unambiguously R&D. A quantum electronics founder recounted receiving a multi-page letter demanding evidence of payment for each of roughly 12,000 individual line items — a request her advisers regarded as outside the scope of the regime. Such requests feel like they are about ticking boxes to show a process has been followed, rather than about risk.

The contrast with jurisdictions that pair claimants with technical specialists was felt sharply. In other jurisdictions, even where an assessor disagreed, participants noted the conversation at least took place in technical terms — rather than across what one participant called a “massive disconnect.”

This is not an accident of administration but a feature of the scheme’s original design. When the UK regime was set up, a deliberate choice was made not to adopt a Canadian-style model in which claims are reviewed directly by technical specialists. Given how positively the Canadian system is now spoken of, it is fair to ask whether that decision should be revisited.

HMRC has more recently established a panel of subject-matter experts, but it is important to be clear that, while it offers inspectors broader context on where the boundary of qualifying R&D lies in emerging areas, it does not review individual claims or make eligibility decisions. Responsibility for assessing claims still sits with the individual inspector. The practical consequence, as advisers stressed, is that claims must be written for a well-informed lay reader rather than a domain specialist — a real constraint on how founders can communicate the novelty of their work.

The 30% intensity threshold

The R&D-intensity threshold drew sustained criticism as a piece of policy design. Because it operates as a cliff rather than a slope, the difference between 29.9% and 30% can translate into hundreds of thousands of pounds of claim value. Founders described tracking the threshold monthly in their management accounts and altering investment behaviour to avoid dropping below it.

For loss-making companies, the cliff edge is very real — and, as flagged in the evidence session, is probably driving a visible clustering of claims around 31% R&D intensity. The legislation was rushed through ahead of the Budget, and participants drew an analogy with the old “slab” structure of Stamp Duty, which famously distorted the housing market until reformed to a marginal basis. There was a strong call for the intensity threshold to be reformed onto a marginal footing so that small fluctuations in spend do not produce disproportionate consequences.

DSIT guidelines and the pace of technological change

Builders of proprietary software and AI systems argued that the DSIT guidelines underpinning the definition of “scientific and technological uncertainty” are out of date — much of the foundational material dates from 2004, and even recent refreshes do not, in their view, adequately capture modern software development or AI systems.

A more fundamental point deserves to sit alongside that criticism: the DSIT guidelines are intentionally high-level and cross-sector. By design, they cannot reflect the current state of the art in any single field, and the underlying framework can remain sound even when individual illustrative examples appear dated. HMRC has attempted to issue more detailed sector-specific guidance — notably for software and pharmaceuticals — but any such document will inevitably age as technology advances, and HMRC does not have the resources to continually refresh guidance across every sector. The practical implication is that companies and advisers should treat the guidelines as guidance, not as a detailed technical manual or a template for drafting a claim.

That said, the underlying complaint from founders is legitimate. In fast-moving areas, particularly AI and modern software development, the gap between what the guidelines describe and what businesses are actually building has widened to the point where claimants struggle to anchor their claims in the published framework at all. A clearer definition of R&D or innovation that is easier to understand and can be applied to technology as it evolves would be useful for founders and prevent having to update guidance so frequently.

The rise of the claims-farming industry

Several participants raised concerns about the proliferation of low-quality “we’ll do your R&D claim” outfits cold-calling founders. One CEO described receiving two to three such calls a week. This results in honest claimants being swept up in HMRC’s response to fraud driven by these intermediaries. More established founders are concerned that inexperienced founders — particularly those outside traditional R&D sectors — are at risk of being misled into making ineligible claims and facing subsequent investigation.

What the scheme still does well

Despite the criticism, participants were clear that the relief itself is vital for many. In life sciences, for example, it was described as being essentially built into business plans, and as a structural reason the UK is able to over-index on R&D relative to international peers. For deep-tech founders, it helps retain UK-based R&D staff. The merger of the RDEC and SME schemes was welcomed by some as a genuine simplification, particularly for companies juggling Innovate UK grants alongside tax relief.

The underlying message was not that the scheme should be abolished, but that constant tinkering had eroded the confidence that makes it function as an incentive. As one participant put it, the relief is increasingly treated as welcome money when it arrives, rather than something a founder can confidently plan R&D investment around.

Assessing the value of the scheme

A point worth raising alongside the operational issues is how the scheme is being evaluated. HMRC’s assessment has been framed primarily in “value-add” terms — how much additional R&D expenditure is generated for every £1 of support provided. Under the most recent review, the SME scheme was found to deliver a value-add ratio of approximately one, which would not typically be considered acceptable for a business support intervention.

The evidence presented in this session illustrates why that headline assessment risks missing the broader picture. A value-add ratio of one does not capture uncertainty reduction, risk-sharing between the state and the founder, the retention of high-skilled R&D staff in the UK, or the longer-term productivity effects of keeping R&D-intensive businesses domiciled here. Several participants pointed to precisely these effects when explaining why they continue to base their business plans around the relief despite its frustrations.

This reinforces the case for a fresh evaluation of the scheme as it now operates, following the introduction of the merged regime, rather than continued reliance on pre-merger evaluations and a narrow value-add methodology.

Policy ideas raised

A clear set of reform priorities emerged from the discussion:

1. Reform the 30% intensity threshold for loss-making companies onto a marginal rather than slab basis, to remove the cliff-edge behaviour distortion.

2. Refresh the DSIT sector-specific guidance — particularly for software, AI and data-driven R&D — while recognising that the high-level framework is, by design, generic.

3. Make Advance Assurance usable — including for repeat claimants and larger programmes — to give founders real pre-investment certainty.

4. Introduce a statute of limitations on retrospective investigations for non-fraudulent cases.

5. Investigate the success of the well-regarded Canadian-style model — where technical specialists assess claims.

6. Provide more clarity over exemptions for overseas R&D where the work genuinely cannot be done in the UK — particularly around specific patient populations in clinical trials and specialist manufacturing or testing.

7. Persistence and feedback: enable founders with long claim histories to build on prior approvals and provide a scaled-back claim report rather than starting from zero each year, in order to reduce the administrative burden.

8. Commission a fresh evaluation of the merged scheme that goes beyond a narrow value-add ratio to capture uncertainty reduction, skills retention, risk-sharing and longer-term productivity effects.

Re-ordering priorities

The session concluded with a sharp warning: if the scheme cannot be made to incentivise R&D investment again, some participants argued that the £7.6 billion it costs the taxpayer would be better spent elsewhere. That was offered not as a policy recommendation but as a measure of how far confidence has slipped.

The more constructive reading — and the one most participants endorsed — is that the relief remains one of the most important levers the UK has to keep R&D-intensive businesses, jobs and intellectual property in the country. But it is being undermined by complexity, uncertainty and a compliance posture that treats legitimate claimants as suspects. As the government looks to boost the UK’s R&D intensity, restoring confidence in the scheme — and the founders who depend on it — will be essential.