R&D tax relief: the fraud wave, and the crackdown now hitting the innocent

R&D tax relief is a case study in how a good policy, badly guarded, ends up punishing the people it was written for. The relief was generous and the controls were thin, so the abuse arrived on schedule. The response - a mass compliance operation - drove the fraud rate down and the legitimate small claimant out at the same time.

The relief and its purpose

Research and Development (R&D) tax relief was introduced in 2000 to encourage innovation by small and medium-sized businesses, alongside a separate Research and Development Expenditure Credit (RDEC) scheme for large companies. Under the original SME scheme, qualifying companies could deduct an enhanced proportion of their R&D spend - at its most generous, 230% of qualifying costs - from taxable profits, or claim a payable tax credit of up to 14.5% of losses if loss-making.

A profitable company spending GBP 100,000 on qualifying R&D could reduce its taxable profits by GBP 230,000. A loss-making startup could receive a cash payment from HMRC whether or not it was paying any corporation tax at all. For many early-stage technology and engineering companies, that cash-flow benefit was structurally important to survival.

The scheme broadened sharply over two decades. Claims grew from 15,585 in 2013-14 to a peak of approximately 79,205 in 2021-22, with total relief reaching GBP 4.76 billion in 2021-22 - a 575% cash increase from 2013-14.

The wave of fraud and rogue advisers

Generous rates, complex qualifying rules, cash-payable credits, and an HMRC approach that relied heavily on self-assessment with no pre-approval created the conditions for systematic abuse. It took two broad forms.

Organised criminal fraud. The House of Lords Economic Affairs Committee and TaxWatch UK documented large-scale, coordinated attacks on the scheme - claims filed for work that did not occur or did not qualify.

Rogue R&D advisory firms. A cottage industry of "no win, no fee" R&D consultancies emerged, cold-calling small businesses and offering to file claims on commission - typically taking 15 to 30% of any credit. Many of these advisers, unregulated and accountable to no professional body, prepared spurious or substantially inflated claims, sometimes without the business owner fully understanding what was being submitted in their name. The adviser collected the commission and moved on; any later HMRC challenge fell entirely on the company.

At its peak, HMRC estimated error and fraud in the SME scheme at 25.8% of all claims - approximately GBP 1.20 billion - in 2021-22. The National Audit Office qualified HMRC's accounts in consecutive years over error and fraud levels it described as among the highest reported across all government spending, including the programmes run in response to the COVID-19 pandemic.

The crackdown, and its collateral effect

From 2022, HMRC mounted a substantial compliance operation: mass enquiries into existing claims, a mandatory Additional Information Form filed with every R&D claim (from August 2023), and tighter scrutiny of whether activities genuinely met the R&D definition under section 1006 of the Income Tax Act 2007. It also introduced pre-claim notification and tightened the rules on subcontracted and overseas expenditure.

The figures show the measures working on fraud. Overall scheme error and fraud is estimated to have fallen from 17.6% (GBP 1.34 billion) in 2021-22 - a figure HMRC revised upward retrospectively as its random-enquiry data matured - to 7.8% in 2023-24 (an initial estimate later revised to about 6.5%) and 5.9% (GBP 481 million) in 2024-25. For the SME scheme specifically, the rate fell from 25.8% in 2021-22 to 14.6% in 2023-24 and 10.6% (GBP 339 million) in 2024-25.

But the crackdown has inflicted heavy collateral damage on legitimate claimants. SME R&D claims fell by 31% in 2023-24 against the prior year, with total claims dropping from around 86,000 at the 2021 peak to approximately 44,065 by 2024 - roughly half. First-time applicants fell by 45% in a single year. TaxWatch, no apologist for the fraudsters, noted that HMRC's "bulk compliance effort has been required, threatening the integrity of the scheme from the other end of the spectrum, with many valid claims being denied." Blick Rothenberg warned that HMRC's approach "must not risk penalising genuine small and medium-sized" claimants.

The chilling effect is real and documented. Many SMEs - particularly those without an in-house tax function - have concluded that the cost and risk of filing now outweighs the benefit, given the heightened prospect of a formal compliance check, the time and professional cost of responding, and the risk that the process triggers a wider corporation tax enquiry. The people who fabricated claims could afford to lose them. The small company that built something real, and feared it could not survive the enquiry, simply stopped claiming.

The merged scheme

From 1 April 2024, the separate SME and RDEC schemes were merged into a single scheme for accounting periods beginning on or after that date. The merged scheme applies an expenditure credit of 20% of qualifying R&D spend, broadly on the RDEC model. Loss-making SMEs that are "R&D intensive" - spending at least 30% of total expenditure on qualifying R&D - may claim under Enhanced R&D Intensive Support, which gives a higher effective benefit.

For many SMEs that relied on the old scheme's higher rates, the merger is a cut in the value of the relief. Combined with the compliance burden, the data points to a sustained contraction in R&D tax-credit participation by smaller businesses - precisely the population the relief was created to support.

International parallels

R&D-relief abuse is not unique to the UK. In Canada, the SR&ED scheme has faced persistent problems with aggressive contingency-fee claim filing, and CRA crackdowns that critics argue similarly chill legitimate small-business participation. In Australia, the R&D Tax Incentive - administered jointly by the ATO and AusIndustry - has been subject to similar scrutiny and to reviews of claims that fail the "core R&D activities" test. The pattern recurs across jurisdictions: generous relief, complex eligibility, unregulated advisers, mass compliance action, and legitimate-claimant collateral.