Recommendation 3

Departments’ lack of urgency to robustly measure fraud and error hinders their ability to direct their counter fraud and error efforts. The robust measurement of fraud and error is a crucial aspect of any counter fraud response because it shows where controls need to be improved and effort directed. BEIS, DWP and HMRC have committed to producing an estimate of fraud and error for some of their COVID-19 schemes and publishing this in their annual report and accounts. But we are concerned that this work is not being given the priority it deserves and comes too late to direct efforts to minimise fraud and error in the schemes. HMRC does not expect to have a statistically valid estimate of fraud and error in the Coronavirus Job Retention Scheme until December 2021, around 22 months after it was first introduced. HMRC also does not intend to measure the rate of fraud and error in the Self-Employed Income Support Scheme and Eat Out To Help out schemes. DWP cannot say when it will set a target for reducing fraud, citing that COVID-19 has changed its traditional mix of cases within the benefits system. BEIS does not expect its fraud sampling exercise on the Bounce Back Loan Scheme to complete until the end of May 2021, 12 months after the scheme opened for applications. Other departments, that have not had as much scrutiny of their fraud and error risk, have not made public commitments to measure the extent of fraud and error in their COVID-19 schemes. Recommendation: HM Treasury should, within three months, strengthen current reporting requirements and ensure that all departments measure and report on the risks of fraud and error within each of their COVID-19 support schemes. This should include: • the estimated value of fraud and error within their COVID-19 response, • how identified risks of fraud and error are being addressed, and • any planned action to recover taxpayer money lost to fraud and error, including timescales. HMRC and BEIS should write to the Committee wit