R&D – still changing after all these years

14th March 2023

  • R&D reliefs predicted to cost over £9bn by 2026-27 – by far the largest corporation tax relief
  • Fraud and error in schemes total over £1.1bn in last three years
  • R&D ‘claims farms’ continue to hard sell opportunities to claim refunds on expenditure that often does not qualify
  • Impact of changes to help tackle problem R&D claims firms not felt for another two years
  • HMRC not resourced to tackle historical incorrect claims


With costs running well into the billions, R&D tax reliefs are by far the largest business relief available in the UK1.They are consistently mentioned in budget and fiscal statements and it seems likely that they will appear again in the Spring Statement. Fraud and error in these schemes costs hundreds of millions of pounds a year yet, despite this, very little is being done to tackle historical abuse. New measures introduced in the Autumn 2022 budget won’t be felt for around another two years, and do nothing to claw back the billions already lost. Once again, the issue comes down to lack of resourcing for HMRC to tackle historic claims.

At the same time, serious questions need answering about whether the reliefs are acting as the incentive they were intended to be. Highly profitable finance companies are claiming millions in relief. Boundary pushing is rife – with unregulated advisers encouraging borderline fraudulent behaviour, while skimming off the top. Many companies are only applying for the relief after being contacted by tax advisers, suggesting the work would have been carried out regardless of the relief.

The scale of the reliefs

The original scheme was introduced in 2000 but has developed and extended to become one of the most generous schemes in the world. The relief schemes are part of government plans to spend 2.4% of GDP on R&D by 2026-272.

However, by 2017-18 the costs of the scheme were more than double the original forecast, coming in at £2.2bn. Since then costs have rocketed with reliefs costing the Treasury £6.6 billion in 2020-213, unusually a small decrease of 4% from 2019-20 at £6.9 billion (though almost certainly affected by Covid-19). Following changes to be introduced next year, the Office for Budget Responsibility (OBR) still predicts the cost of total reliefs will increase from £6.8 billion in 2021-22 to £9.2 billion in 2026-27, a very significant sum4.

The OBR has previously identified that the cost of ongoing tax reliefs receive little scrutiny compared with conventional public spending5 beyond annually produced cost data. With such a staggering amount of money spent on these schemes, it is vital to look at whether the money is going to the right place.

Does the funding have the desired outcome?

There are a number of different aspects which suggest the money is not achieving what it was intended to. As detailed below, a proportion has clearly been going to businesses who do not qualify, with some of that going to ‘rogue’ claims advisers. A further cut of the pot also appears to be going in interest and fees to finance companies offering upfront loans against future receipt of claim repayments. While it is understandable that businesses may wish to improve their cashflow when waiting for repayments, these are additional businesses making money out of the reliefs. It’s highly unlikely that when these reliefs were introduced the intent was to line the pockets of advisers and finance companies.

Conflicting data from the Office for National Statistics (ONS) and HMRC has caused confusion about how much R&D was taking place but recent adjustments to ONS methodology, resulting in very significant changes to their statistics, have brought the two data sets closer in line6. Some have suggested that there is now comfort from the ONS figures that error and fraud in the scheme are not as high as previously feared7. However, it seems likely that this comfort is misplaced as, if a business has claimed R&D tax relief, it is surely going to report in a government survey that it has conducted R&D, regardless of whether that expenditure genuinely qualifies. In addition, the ONS definition of R&D also includes capital expenditure and spending on the arts, humanities and social sciences so it would be expected that ONS figures for R&D would be significantly higher than HMRC data.

If the idea behind these reliefs is to increase innovation within the UK, then the fact that many companies do not even consider submitting claims until approached by advisers should be cause for concern8. If the spending has already taken place with no thought given to the reliefs, it can hardly be said that the R&D system is increasing spend in those circumstances.

When most people think of research and development, they might conjure up images of laboratory breakthroughs or advances in computing and technology. The data reveals a very different story. Given the broad nature of industry classification codes it is impossible to drill down exactly as to where the claims are being made, but there are some surprising results.

The ‘Financial and Insurance’ sector made 1,445 claims in 2021, on a spend of £2.59bn on R&D. ‘Admin and Support Services’ made 5,015 claims that year, on a spend of £1.33bn. That’s a whole lot of innovation in admin.

More interestingly, expenditure per claim is £1.8m for in the Finance sector – only ‘Mining and Quarrying’ has higher levels of expenditure per claim. Such high expenditure may be a result of the large salaries paid in finance.

Case study – Quadrature Capital Ltd

Quadrature is a hedge fund, recently identified as paying some of the highest salaries (average £3.6m) in the industry to its staff9. It shows profit after tax of £525m in the accounts to 31 January 2021. These accounts also show that they received an R&D tax credit of £5.67m (which [at 13% from April 2020] suggests total claimed R&D expenditure of £43.62m), and total credits of £12.6m over the last six years. As £409m of its £569m administrative expenses are salaries it seems highly likely that the high level of tax credits claimed relate to the very high salaries paid. While there is no suggestion that they have incorrectly claimed, there must be an argument that a business like this does not need support from the government to encourage it to innovate, and that support was not intended to fund salaries of that level. This is an example of the unintended consequences of untargeted support provided by tax reliefs.

While industry professionals and sector bodies argue that reliefs provide incentives for investment, which ultimately accelerates R&D as companies are willing to take on risk, there is also a recognition of the problems. They suggest that these arise, at least in part, from poor targeting and a lack of clarity and education around the definitions alongside the problem advisers and targeted attacks10.

The Autumn Statement announced a reduction in the SME scheme benefits impacting those businesses genuinely carrying out R&D that the government claim to want to support, raising concerns across the sector. This came with a promise to ‘work with industry to understand whether further support is necessary for R&D intensive SMEs, without significant change to the overall cost envelope for supporting R&D’11. This finally suggests a recognition that support needs to be better targeted but will require further consultation, and time, to arrive at something suitable, while in the short term impacting on genuine SME claimants.

‘Spurious’ claims and ‘rogue’ advisers

In 2021-22 HMRC estimated the level of fraud and error in the schemes at 4.9% of the estimated cost of the reliefs, equating to £469m, but this is biased towards the small and medium enterprises (SME) scheme, with 7.3% of claims estimated as incorrect, but only 1.1% of R&D Expenditure Credit (RDEC – mainly for large businesses) claims12.

Any kind of relief where refunds are potentially available is open to abuse. There has been much recent press and political coverage of the level of spurious claims, particularly in the SME scheme13. Many of these appear to arise out of what are often described as ‘rogue’ agents or advisers – usually unregulated firms, working on a contingent fee basis, that are even cold calling businesses persuading them to put in claims for relief. These, at the very least, push the boundaries of the definition of R&D and often go way beyond it.

Advertising of 100% success rates, which are meaningless in the context of the self-assessment system, and simplified methods for producing claims, are obviously attractive to businesses. Most businesses assume that the tax adviser industry is regulated14 resulting in them failing to apply due diligence in discriminating between the deals on offer, leaving them vulnerable to the less reliable claims agents. In effect, anyone can give tax advice. It seems relatively simple to set up a website advertising guaranteed success, cold call a business and talk them into submitting a claim, then skim a percentage off the top.

Despite the greater focus on the R&D claims market in recent months there are still plenty of firms out there trying to get a share of the spoils, persuading businesses to make claims that quite clearly do not fall within the definition of R&D. Businesses are still stating that only 10% of eligible companies are claiming their appropriate reliefs leaving 90% missing out, and often fail to be specific about the qualifying requirements, suggesting any new product could qualify.

We have seen advisers suggesting a variety of activities that could get relief when it is hard to see how they would fall within the definition of science and technology, such as staff training by local experts in domiciliary care. Firms have also made reference to claims made on behalf of Premiership football clubs and small roofing companies. While not impossible that the businesses and activities described are involved in advances in science and technology, they will be rare, and these examples appear to be misleading as to the nature of the requirements for claims.

Most strikingly, a simple search produces large numbers of adverts for sales roles within R&D tax credit claims businesses, which indicates how much of the relief is being spent on promoting the possibility of claims. One recent poll by an R&D adviser identified that 15% of businesses were getting cold called by such advisers at least once a day, with a further 28% at least once a week. While this will not be statistically significant it is evidence of the scale of the problem15.

One company advertising sales jobs claims to have dealt with £100m of claims in the last three years, and expects to claim £20m per month in the next year, suggesting a quarter of a billion pounds in claims by that one firm alone.

Many R&D specialist firms will also pay a percentage of their own fee as a referral fee to accountants or other advisers who refer on their clients for specialist R&D assistance, creating another group of people incentivised to persuade businesses to claim.

In reality, establishing eligibility for relief is complex and takes time and expertise. Many people within the industry have been advocating for compulsory professional regulation for R&D advisers (and more widely in the tax adviser industry)16. Research has shown that 80% of advisers that are not members of professional bodies have no professional qualification17, which is surely unusual in the financial services industry and clearly gives rise to significant risks to both clients and HMRC. However, consultation last year on improving the tax advice market resulted in no changes and intentions for a further consultation that has not yet appeared. In the meantime, the largely unregulated claims adviser market is continuing to abuse the system and potentially cause financial damage to those unwittingly taken in by them if boundary-pushing claims are eventually refused. This mirrors the situation which has arisen in recent years with participants in disguised remuneration schemes subsequently liable for the loan charge18 and demonstrates the practical impact of such lack of regulation.

HMRC compliance activity

The long-running problem of ‘rogue’ R&D advisers has not been helped by the lack of targeted compliance activity by HMRC. While the reduction in SME scheme benefits will likely make it less attractive to potential fraudsters, there have already been large amounts incorrectly claimed according to HMRC estimates. In the last three years alone, estimates of error and fraud in the schemes total over £1.1bn (21/22 £469m, 20/21 £336m, 19/20 £311m19).

Incorrect claims will range in seriousness from genuine errors, through pushing (or exceeding) the boundaries of the R&D definition and eligible costs, to those that are undeniably fraudulent and relate to costs which were not occurred.

There is work that could be done by HMRC to improve definitions, guidance and knowledge that could help prevent businesses making errors in their claims. More education would also potentially enable businesses to be better able to identify when ‘rogue’ agents are proposing they make unacceptable claims20. However, the issue of problem R&D agents will not go away until HMRC take real action to challenge them, but at the moment HMRC appear to be waiting until new legislation requires advisers to be named in claims (see below).

HMRC took action in April of last year following identification of what they describe as organised crime group attacks on the scheme, pausing repayments for two weeks and doing more checks before paying money out. Letters issued to around 1,600 claimants appears to have been well-targeted, based on results so far21, but there are also reports of distress being caused to recipients who are genuine claimants and have interpreted the letters as allegations of fraud. However, these have been focused on claims thought to be part of the organised crime attacks rather than the ‘rogue’ agent type cases.

HMRC have also reported eight recent arrests in relation to R&D claims22, including of an adviser involved in the claims, but this seems likely to be just the tip of the iceberg and yet again demonstrates the limited deterrent effect of HMRC’s prosecution policy. With hundreds of millions of pounds of fraud each year, eight arrests is pretty insignificant.

Other changes to be introduced to legislation from April 2023 (applying to accounting periods starting after 1 April 202323) will require pre-notification that a claim is to be made and more details around the claim, including naming an officer of the business responsible and any adviser. This is likely to act as a deterrent for some of the more outrageous claims. This requirement will also provide HMRC with much more information, and time to carry out risk assessments, including the possibility of engaging with claimants before they submit their claims24. This is a move in the right direction but most first pre-notifications won’t happen until September 2024 and it will be getting on for two years before they are receiving information in a digital format to enable proper targeting of risk assessment. This is plenty of time for many more spurious claims which could lead to further losses in the region of £1bn based on most recent estimates of fraud and error.

The reduction in SME scheme benefits are intended to make it less attractive to potential fraudsters but this is only forecast to reduce the level of fraud and error in the scheme from 7.3% to around 6%,25 which raises questions about whether this justifies the loss of benefit to genuine claimants.

HMRC staff dealing with R&D reliefs have been moved around a bit in recent years with long-standing specialist units created in 2006 having been restructured26. Numbers working on R&D were said to have been increased by 100 in March 2021 but anecdotally a number of experienced staff have been lost and there is a call for the right resources to be in place to deal with cases27.

It is clear that, in response to concerns about abuse of the scheme, HMRC has stepped up their compliance activity, but they appear to be taking a scattergun approach. We have been told that even in cases where lots of information has been provided they are asking large numbers of apparently standardised questions but then taking significant amounts of time to respond. In addition, there appears to be a shortage of experienced staff involved so treatment is inconsistent between HMRC officers, BEIS guidelines are being misinterpreted and progress of enquiries is very slow. One industry insider told us ‘in response to the issue of fraudulent claims, HMRC has instituted a backlash of compliance activity, and unfortunately the way this has been carried out puts genuine claims at risk and undermines the whole aim of the R&D tax credit reliefs to support innovation in the UK.’

One obvious issue is around the application of the definition of R&D, which is critical for establishing eligibility but is not a tax concept. The definition belonged to the Department for Business, Energy, and Industrial Strategy (BEIS) (possibly now Science Innovation and Technology or Department for Business and Trade?), who seem likely to have far greater experience in the field than HMRC staff. This raises a question as to whether there are non-tax professionals within government who would be better able to test claims against the eligibility definition, alongside the tax compliance staff dealing with other aspects. It seems obvious that, if they were involved in this aspect of compliance. they would enable better consistency of treatment and assurance that reliefs are being given where they should be. However, there is no suggestion that this is being considered with the current talk being of a new HMRC team being set up from April 2023 with few details of how this will be staffed or will operate.

The Financial Secretary to the Treasury recently stated28 that she believes HMRC is now putting enough compliance resource into the problem (or will be with the new team), but that further information to come may change her opinion. This largely refers to a random enquiry programme underway into claims, effectively a sampling exercise, which HMRC believe will identify how widespread the problems are. It’s hard to understand why it has taken so long to pursue this information bearing in mind the long-standing concerns about incorrect claims, and raises questions about the current error and fraud estimates. These are said to come from historic compliance activity but it is difficult to know how significant that has been as previously HMRC have said that providing details of the number of enquiries would potentially prejudice the assessment or collection of tax29. However, those enquiries that have taken place appear to have been particularly successful, with 84% of the cases settled since April 2022 having identified inaccuracies in claims, with average additional amounts due of £128k30. However, without any idea of the numbers of cases it is hard to evaluate the full compliance programme.

At the same hearing Nicole Newbury, HMRC Director for Wealthy and Mid-sized Business Compliance, referred to the fact that there is currently no automated risking scheme related to R&D claims as this requires electronic provision of data behind the claim which will not start to come in until probably around March 2025. This suggests that, until then, the only way of identifying risky claims is by manual risk assessment – something which appears impossible for the currently almost 90,000 claims a year with the numbers of staff reported. HMRC are trialling less resource intensive approaches involving ‘one to many’ letters where the same letter is issued to a number of claimants with similar characteristics nudging them to consider their claims on the basis of guidance provided in the letter31. While this is an efficient way of recovering some taxes where claimants have made errors it is less likely to provide results in more deliberate cases.

While the changes are welcome and should begin to reduce abuse of the reliefs in future, they are being introduced very late in the day and are moving the focus away from historical, current, and even the next two years claims. It seems that many hundreds of millions of pounds incorrectly claimed cannot possibly be pursued with the limited resources that are or will be available.

It is also of concern that, despite the apparent ease of identifying R&D claims advisers promising large refunds for potentially non-qualifying expenditure, HMRC don’t appear to be doing much to directly tackle those problem advisers.


This is now one of the most expensive tax reliefs. The historically low level of compliance activity appears to be a good example of how failing to take timely action results in increasing abuse as people become more confident in their ability to get away with it.

While the coming changes will almost certainly improve the situation, the focus appears to be on tackling claims that are to come rather than the large sums already incorrectly paid out. Bearing in mind HMRC’s reported return on compliance investment (estimated around 18:132 but varying dependent on customer group) it seems incomprehensible that HMRC are not provided with more general long-term compliance resources to deal with historical issues such as incorrect R&D claims. Are we just to accept that billions of pounds incorrectly claimed through fraudulent activity are to be forgotten about?

More generally, if the UK government seeks to promote innovation in science and technology, they need to be sure that the rules they have in place ensure that any funding is targeted to appropriate areas – improving the definitions, guidance and skills involved in applying them will be crucial for the changes to have this effect.


4Letter to House of Lords Economic Affairs Finance Bill Sub-Committee, HM Treasury, 28 November 2022, https://committees.parliament.uk/publications/31759/documents/178686/default/

5Fiscal risks report, Office for Budget Responsibility, July 2019, https://obr.uk/docs/dlm_uploads/Fiscalrisksreport2019.pdf

6Comparison of ONS business enterprise research and development statistics with HMRC research and development tax credit statistics, Office for National Statistics, 29 September 2022, https://www.ons.gov.uk/economy/governmentpublicsectorandtaxes/researchanddevelopmentexpenditure/articles/comparisonofonsbusinessenterpriseresearchanddevelopmentstatisticswithhmrcresearchanddevelopmenttaxcreditstatistics/2022-09-29

7Letter to House of Lords Economic Affairs Finance Bill Sub-Committee, HM Treasury, 28 November 2022, https://committees.parliament.uk/publications/31759/documents/178686/default/

8Corrected oral evidence: Draft Finance Bill 2022-23, House of Lords Economic Affairs Committee Finance Bill Sub-Committee, 21 November 2022, https://committees.parliament.uk/oralevidence/11607/pdf/

9Quadrature Capital: The hedge fund that pays staff £3m each, The Times, 13 November 2022, https://www.thetimes.co.uk/article/quadrature-capital-the-hedge-fund-that-pays-staff-3m-each-p068vjqd2

10Corrected oral evidence: Draft Finance Bill 2022-23, House of Lords Economic Affairs Committee Finance Bill Sub-Committee, 14 November 2022, https://committees.parliament.uk/oralevidence/11558/pdf/ and Corrected oral evidence: Draft Finance Bill 2022-23, House of Lords Economic Affairs Committee Finance Bill Sub-Committee, 14 November 2022, https://committees.parliament.uk/oralevidence/11556/pdf/

13Free money from HMRC’: firms cash in on questionable tax credits, The Times, 28 October 2022, https://www.thetimes.co.uk/article/fraud-tax-credit-hmrc-government-scheme-hmrc-investigation-330bxbk8c

15How often are you targetted by random companies peddling tax credits?, Rufus Meakin, 28 February 2023, https://www.linkedin.com/feed/update/urn:li:share:7036323766594859008

18Disguised remuneration schemes and the Loan Charge briefing, Taxwatch UK, March 2020, Loan-Charge-Debate-Briefing_20200318_FINAL.pdf (taxwatchuk.org)

20Corrected oral evidence: Draft Finance Bill 2022-23, House of Lords Economic Affairs Committee Finance Bill Sub-Committee, 14 November 2022, https://committees.parliament.uk/oralevidence/11558/pdf/

21Letter to House of Lords Economic Affairs Finance Bill Sub-Committee, HM Treasury, 28 November 2022, https://committees.parliament.uk/publications/31759/documents/178686/default/

22Eight held over research and development tax credit fraud, The Times, 31 October 2022, https://www.thetimes.co.uk/article/eight-held-over-research-development-tax-credit-fraud-conspiracy-86nhgwl8q

25Research and development tax credit cut ‘risk to innovation’, The Times, 6 December 2022, https://www.thetimes.co.uk/article/research-and-development-tax-credits-cut-risk-to-innovation-82flfn7g9

26Letter to House of Lords Economic Affairs Finance Bill Sub-Committee, HM Treasury, 28 November 2022, https://committees.parliament.uk/publications/31759/documents/178686/default/

27Corrected oral evidence: Draft Finance Bill 2022-23, House of Lords Economic Affairs Committee Finance Bill Sub-Committee, 14 November 2022, https://committees.parliament.uk/oralevidence/11558/pdf/

28Corrected oral evidence: Draft Finance Bill 2022-23, House of Lords Economic Affairs Committee Finance Bill Sub-Committee, 21 November 2022, https://committees.parliament.uk/oralevidence/11607/pdf/

29HMRC Freedom of Information response 19 December 2022

30Letter to House of Lords Economic Affairs Finance Bill Sub-Committee, HM Treasury, 28 November 2022, https://committees.parliament.uk/publications/31759/documents/178686/default/

31Letter to House of Lords Economic Affairs Finance Bill Sub-Committee, HM Treasury, 28 November 2022, https://committees.parliament.uk/publications/31759/documents/178686/default/

32Treasury committee, Oral Evidence: The work of HMRC, HC 382, House of Commons, 30 November 2022, https://committees.parliament.uk/oralevidence/11972/pdf/