Government departments have 28 data-sharing agreements in place to detect fraud, but a report from the National Audit Office (NAO), has found there is insufficient focus on preventive measures.

The authors of the NAO’s Using data analytics to tackle fraud and error report noted that integrating preventative controls into existing business processes often requires real-time data sharing and cleaning of data for the controls to flag or stop potentially incorrect payments. The NAO said most of the data analytics tools in use by the government are to detect potentially incorrect payments that have already been made, and are not part of front-line preventative controls.

The NAO estimates that fraud and error cost the taxpayer between £55bn and £81bn in 2023-24. According to the report, data analytics can be deployed to make sure the right amount of money goes to the right recipient and to find potentially incorrect transactions.

The report covers basic tools that check if payment to a supplier is not duplicated, to the use of emerging technology like artificial intelligence (AI) to identify risky transactions. “With good-quality linked data, these technologies can deliver more immediate returns on investment, tackling fraud and error without requiring the wider system or organisational reform that fuller digital transformation would require,” the NAO said. 

The report highlights poor data sharing and poor adoption of data sources, which can be used to help tackle fraud. The NAO noted that there has not been widespread take-up of central initiatives, such as the National Fraud Initiative, which compiles data to identify potentially fraudulent activity. The NAO was told that among the reasons for low adoption were a lack of understanding of such initiatives among officials and the cost. 

In addition, the NAO found that counter-fraud teams are not always aware of datasets that might help them tackle fraud and error. While the government is trying to improve the quality of some key datasets that have the potential to unlock better fraud and error analytics in future, the NAO reported that inconsistent data formats and systems make it harder to use data to tackle fraud and error.

It recommended that HM Treasury make use of the National Fraud Initiative mandatory and agree with the Public Sector Fraud Authority (PSFA) the criteria for where public bodies should use other centrally provided tools. The NAO said HM Treasury and the PSFA should also review the charging model for PSFA central services to ensure they do not dissuade public bodies from making savings.

The report’s authors noted that public bodies continue to find it difficult and bureaucratic to share data to help tackle fraud, even though it is permitted under legislation. But they warned that greater data sharing could lead to fraud being exploited across departments.

“As more data is shared and systems linked, the risk increases that fraudsters penetrate one system to take advantage of another,” they warned in the report. 

The NAO said the PSFA and the Government Digital Service (GDS) should build on the Digital Economy Act 2017 data-sharing process to introduce a managed process to support public bodies to share data. They should also advise public bodies on how to ensure clear ongoing governance arrangements over shared data to manage the risk of fraudsters gaining access to multiple systems through access to one. It also urged the PSFA to consider using non-government and private sector databases, such as the fraud databases produced by Cifas.

Although the GDS Blueprint for modern digital government sets out a vision for government digital transformation, which covers the use of data analytics to tackle fraud and error, the NAO said it has not provided an implementation plan or considered that plan from the perspective of fraud and error data analytics.

Read more on Artificial intelligence, automation and robotics

• 
  

  Accounting watchdog ‘disclaims’ College of Policing financial accounts after serious IT failures

  By: Bill Goodwin

• 
  

  UKRI must do more to drive innovation agenda and avoid fraud

  By: Cliff Saran

• 
  

  AI and data experts to be brought into government through secondment programme

  By: Karl Flinders

• 
  

  Delayed Making Tax Digital programme to cost taxpayers extra £1.5bn

  By: Lis Evenstad