HMRC must apply lessons from innovative attempts to reduce Child Benefit fraud and error

• HMRC should continue to try new and innovative approaches to tackling fraud and error, but it should ensure those approaches have adequate governance and appropriate risk management.
• HMRC used Home Office travel data to identify people who were not living in the UK and who could be incorrectly claiming Child Benefit. While its first rollout of this initiative led to fraud and error savings of £60 million, over 8000 genuine, eligible claimants had their benefit payments suspended without warning. HMRC took steps to improve the process, within 2-4 weeks of problems emerging.
• When designing and planning future approaches HMRC must be clear about what level of risk it is willing to accept, ensure it can monitor claimant experience, and set up an efficient process for claimants to prove their eligibility.

HMRC used Home Office travel data in an innovative attempt to address a major driver of fraud and error in Child Benefit, which totals £270 million annually, but had insufficient governance and risk management, and did not consider the impact on claimants, a new National Audit Office (NAO) report finds. 

In 2021 HMRC identified an opportunity to match its Child Benefit data against the Home Office's airline data to identify claimants who were absent from the UK for more than 12 weeks for a possible error and fraud investigation. An early pilot in 2024 prevented an estimated £15 million in incorrect payments. 

In August 2025, HMRC began the first rollout of this new initiative, with changes to how the process operated compared with the pilot. It removed an early check of the Pay As You Earn (PAYE) system, leading to much higher levels of genuine, eligible claimants having their payments suspended. HMRC also sent out longer, more complex letters and questionnaires to most claimants to prove their eligibility. 

By October, HMRC had suspended payments for an initial 23,794 claimants. Some claimants reported financial and emotional impacts from the sudden loss of income, or difficulty and stress in trying to prove eligibility. During October 2025, HMRC reported an increase in calls from claimants in relation to the first rollout  in total, from August 2025 to February 2026, HMRC handled over 22,500 calls.² 

After problems began to emerge in mid-October 2025, HMRC took steps to improve claimants' experience of the process in October and November for the enquiries that remained open. This included reintroducing a PAYE check, automatically reinstating some payments, and allowing claimants to confirm eligibility over the telephone.³ HMRC also issued apologies to claimants who had been impacted by the rollout and as of April 2026, it had paid a total of £3,200 in compensation to 51 claimants.   

HMRC accepts there were weaknesses in its transition between the pilot and the first rollout, with shortcomings in its risk assessment and decision-making. For example, key decisions were made without sufficient scrutiny, and there was no formal risk assessment of the implications of decisions. HMRC also did not sufficiently consider the impact of the initiative on those claimants whose claims were incorrectly stopped. 

After commissioning an iternal audit to review the intervention, HMRC is now adopting a more controlled approach and learning lessons about strengthening governance, managing risk and the impact of its interventions on claimants. 

HMRC changed the process when it relaunched the initiative from March 2026. This includes giving claimants a month to prove eligibility before suspending payments and enabling claimants to submit evidence of eligibility online. 

HMRC calculated savings of £60 million from the first rollout and it projected that it would achieve savings of £366 million over the five-year period to 2029-30.⁴      

The NAO's report recommends that HMRC understands its risk appetite for future decisions and ensures that its performance data allow it to monitor customer experience alongside compliance effectiveness. HMRC should also disseminate lessons learned and seek to share these internally and with wider fraud and error teams in government.

Gareth Davies, head of the NAO, said:  

It's right that HMRC seeks new ways to tackle fraud and error. HMRC's initial work on using travel data to investigate potential Child Benefit overpayments suggested it could secure significant savings for the taxpayer.   

However, missteps in implementing the first rollout meant HMRC did not strike the right balance between detecting fraud and error and managing the impacts on Child Benefit claimants. While HMRC should not be discouraged from pursuing innovative ways to reduce fraud and error, it must learn and apply the lessons for future initiatives. 

ENDS
Notes to editors

Press notices and reports are available from the date of publication on the NAO website. Hard copies can be obtained by using the relevant links on our website.

1. In 2024-25, HMRC estimated the overall level of fraud and error resulting from overpayments of Child Benefit as £270 million.
2. By April 2026, HMRC information showed that, of the 23,794 claims initially suspended, it had confirmed 59% as eligible.
3. However, HMRC's performance metrics, which focused on compliance outcomes, provided limited data to track how effectively changes in the first rollout improved claimants' experience. For example, while HMRC told us that around 40% of eligible claimants in the first rollout had their payments automatically reinstated by mid-November 2025, it was unable to provide the NAO with average times for reinstating payments. HMRC has noted that, as regards claimant experience, the overall level of complaints submitted to HMRC and compensation paid in relation to the intervention for the first rollout have remained low.
4. Based on the pilot, HMRC projected gross savings for the whole initiative of £366 million over the 5-year period to 2029-30, with a return on investment of around 14:1. It plans to update its cost and benefit estimates for the Autumn Budget 2026.