House of Commons - Public Accounts

Select Committee on Public Accounts Fifty-Eighth Report

APPROPRIATION ACCOUNTS 1996-97: CLASS XII, VOTE 1 (ADMINISTERED SOCIAL SECURITY BENEFITS AND OTHER PAYMENTS)

THE EFFECTIVENESS OF THE AGENCY'S MEASURES TO TACKLE BENEFIT FRAUD

20. Tackling fraud and abuse in the benefits system continues to be one of the Agency's top priorities. A key part of the Agency's fight against fraud is their Security Strategy, which over the five year period 1995-96 to 1999-00 aims to place more emphasis on the deterrence and prevention of fraud whilst continuing to develop methods for investigation and detection. This strategy is supported by a Security and Control Programme, which contains a number of projects, designed to achieve a demonstrable and sustained reduction in the level of fraud and error within benefit expenditure programmes. The Agency are investing £1.25 billion in their Security and Control Programme with the aim of securing weekly benefit savings of more than £6 billion over the five years.[17]

21. The latest estimates of fraud underline the need for an effective strategy for tackling fraud and the importance of shifting the emphasis from detection to prevention.[18] The Committee noted that taking error and fraud together the loss to the taxpayer was around £3 billion. We therefore asked the Agency why they were failing both taxpayers and customers so badly. The Agency agreed that it was totally unacceptable to have fraud and incorrectness that amounted to more than the Agency's running costs of £2.5 billion. The problems though were deep seated, and there was no simple or overnight solution. They had put in place a programme of action over a number of years to address the problem, including visits, reviews, data matching and co-operation with other parties.[19]

22. The Agency added that a key constraint to progress was the inadequacy of their information technology systems and that would take some time to put right. The income support computer system had been rolled out in 1989 and was based on mid-1980's technology; although the software had been updated 20 times since then there was only so much that could be achieved with such an old system. As an example, they told us that it remained possible for a couple to each claim for the same child. Although they could match the records of the two individuals and identify the duplicate claim after the event the current technology was not capable of detecting and preventing the second claim as soon as it was attempted. The Agency told us that they did not have specific funding to update computer systems—which contrasted with the specific "spend to save" funding allocated to the Security and Control Programme. Whilst they had embarked on a procurement path to replace all their computer systems this was a massive task that would take quite a number of years.[20]

23. We asked who was spearheading the attack on fraud. The Agency confirmed that leadership came from the Chief Executive and that their anti-fraud programme was headed by a Director. They added that the Director also had responsibility for benefit process design because they believed that there needed to be a more integrated approach.[21]

Income Support

24. The Committee asked the Agency what the increase in the estimate of fraud in income support, from £1.4 billion to £1.77 billion, told them about the success of their anti-fraud measures. They found it difficult to draw conclusions. In 1995 they had started new claims visits and targeted reviews to confirm claimants' circumstances and to identify unreported changes in circumstances. And they had increased the number of visits from 825,000 to 1.9 million between 1995-96 and 1996-97. But it was unlikely that this work would have shown through by the time of the second benefit review in September-November 1996.[22]

25. The Agency added though, that the benefit reviews had given them information of where fraud occurred, the circumstances under which it occurred, its characteristics and root causes. This had enabled them to focus their anti-fraud initiatives, such as visits to new claimants and targeted reviews, more effectively on the higher risk areas. Some of these reviews were also based on the local knowledge of district office staff.[23]

26. We asked whether there was a typical fraudulent claimant. The Agency said there were usually a combination of circumstances, and that there were higher levels of incorrectness and fraud around claimants living together, or working and earning and not declaring either working or the level of earnings.[24] They gave us two examples:

a case where a person was claiming as a single parent. However, review of Agency computer records by an investigator revealed that the person's husband was resident at her address. On a subsequent visit to this address a neighbour told the investigator that although he did not know when the claimant would be in, her husband would be home at around 7 o'clock that evening. When the claimant was interviewed she denied that her husband was living with her, and claimed that he came round to visit but she did not know where he lived. She subsequently returned her order book to the local Agency office stating that she and her husband were now living together.[25]

a case where, following an ineffective visit to the home of an unemployed single person, the claimant signed off stating that he would be starting work with his father. The Agency investigator made a follow-up visit to the claimant and was informed by the claimant's mother that the claimant had been working for his father for some time without pay.[26]

27. In our report on Measures to Combat Housing Benefit Fraud, we expressed concern about the significant level of fraud in housing benefit that arises because of the failure to properly control claims for income support. In particular, we were concerned about the failure of the Agency to co-operate fully with local authorities.[27] We asked the Agency what they were doing to put this right. They told us that they recognised the problem and that they had started a programme of work to improve co-operation with local authorities. This programme included provision of remote access terminals to enable selected local authorities to gain access to information held by the Agency, and setting clear objectives for operations directors and area directors within the Agency to improve communication and relationships with local authorities, including examining the feasibility of co-locating staff.[28]

28. We asked the Agency when they expected to bring the level of income support fraud under control. They told us that they were unable to provide a clear timetable or target date, given that on their own the Agency could not address the issue totally. In their view, all members of society, the public, everybody had a role to play in recognising and accepting that it was the Agency's job to ensure that those who should not get benefit do not get benefit.[29]

29. We pressed the Agency on why they were reluctant to set targets for reducing fraud on income support. They told us that although the area benefit reviews were intended to give them a better indication of levels of fraud on a continuous basis, they did not believe that they could get an accurate figure because they could not know about frauds until they found them.[30]

Disability Living Allowance

30. Prior to their benefit review, the Agency had already identified the need to strengthen claim procedures for disability living allowance. Their Safeguarding Project had made recommendations for improvement that had been implemented, and which were expected to tighten the gateway to the benefit.[31] The Agency told us that when the results of the benefits review became available, the Safeguarding Project had been merged with the Benefit Integrity Project into a single coherent plan.[32]

31. The Benefit Integrity Project was designed to re-examine entitlement in some 450,000 cases. The focus was on those cases where the claimant was being paid the higher rate mobility component with either the higher or middle rate care component; that is those claimants the Agency considered to be least likely to identify that they should be contacting the Agency about a change in their circumstances.[33]

32. The Benefit Integrity Project has come under severe criticism from disabled people and organisations of disabled people as being unfair and targeting the wrong people. The Committee therefore asked about the operation of the project. The Agency told us that disability living allowance was an extremely difficult benefit to administer. It was not about the medical condition of the person, but the impact of their condition on their mobility and their care requirements. A similar condition could affect claimants in different ways. They acknowledged that there had been weaknesses in procedure, and problems with the accuracy of old records. They told us that following their review of the early results of the Benefit Integrity Project, the Secretary of State had announced on 9 February that, in future, they would require additional information to be collected to validate cases where the outcome was a recommendation to lower or disallow benefit.[34]

33. We asked the Agency whether the blanket coverage of the project was appropriate, or whether a more targeted approach would be more efficient. They pointed out that the approach they had adopted was targeted because they were not addressing all 1.8 million people who were on the allowance. They emphasised that the project was looking at incorrectness, rather than fraud, and that in nearly 1000 cases claimants' entitlement had been increased.[35]

34. We asked the Agency when they expected to be able to quantify the results of their measures to strengthen procedures over disability living allowance claims to make it more difficult to get fraudulent claims into the system. The Agency thought this would take a number of years, as it would be difficult to see how effective any changes in forms and procedures were, and to estimate what the level of fraud might have been without these changes. They were looking at a range of techniques to measure the effectiveness of their anti-fraud initiatives, including comparisons of caseload levels with other health trends and a variety of other attributes.[36]

Specific Measures to Tackle Benefit Fraud: Order Book Control System and the Benefit Payment Card

35. When the Committee last took evidence on this account, they looked at two projects aimed at reducing the level of benefit fraud: the Order Book Control System and the benefit payment card.

36. The Order Book Control System had been introduced in the South East of England. Bar-coded order books were identified by the system, and this prevented the encashment of order books reported lost or stolen or that were otherwise invalid. The Committee recommended extending the system to other parts of the country.[37]

37. The Agency told us that since they last appeared before the Committee, the Order Book Control system had been extended to areas outside the M25, mainly large conurbations, through the introduction of mobile units. They added that full implementation of the Order Book Control System would follow the timetable planned for the benefit payment card as they intend to use the same equipment for both systems.[38]

38. The development of the benefit payment card was aimed at claimants who received payment of benefits through the Post Office network. The card would be swiped through machinery at the post office to give the teller access to details of benefit entitlement and other relevant information about the claimant. In 1996, the Committee was told that the Agency intended to trial the use of the benefit payment card for child benefit payments in a limited number of post offices at the end of 1996. The use of the card would then be rolled out across other benefits and all post offices over a two-year period.[39]

39. The Agency told us that trials of the benefit payment card had started at the end of 1996 and had now been extended to include 205 post offices. Although plans were still under discussion, the Agency expected that a live trial of all services available through the payment card technology would begin in January 1999 involving up to 305 post offices. Subject to the successful completion of this trial, a national rollout would start in April 1999 and be completed before the end of the year 2000.[40]

40. The Agency emphasised that once fully operational, there would be £50 billion a year flowing through the system, with some 20 million people getting benefit paid through post offices each week. This would require the automation of some 40,000 counter positions in about 19,300 post offices across the country. As a result, it was important to ensure that the system would be secure and reliable prior to full implementation, as any problems with the system once roll-out began at the rate of 300 post offices a week would put the project back many years.[41]

Achievement of Fraud Savings

41. The Agency are aiming to secure weekly benefit savings of more than £6 billion over the five years of their Security and Control Programme from an investment of £1.25 billion.[42] In 1995-96 the programme reported savings of £403 million and in 1996-97 of £786 million, but these figures reflected sharp reductions following validation by the Agency's internal audit..[43]

42. In view of this shortfall, and delays to key projects, we asked the Agency about the prospects of achieving their savings target. Although not fully confident, they believed they would achieve a significant proportion of the projected savings. The primary cause for reservation stemmed from the difficulty in knowing how quickly the results of preventative action would materialise, and how it could be measured and demonstrated. However, they pointed out that it was a progressive programme, and some of the early work was around increasing detection in order to understand where they needed to address prevention.[44]

43. The Agency added that they were also working with the National Audit Office, The Treasury and the Benefit Fraud Inspectorate on how they could identify and measure the benefits of fraud prevention.[45]

Conclusions

44. The Agency's computer system is a key constraint to detecting fraud. The Agency account for benefits totalling over £36 billion, and it cannot make sense to seek to control these payments with systems that use out of date technology and are not up to the task. While we recognise that the procurement of new systems is now underway, we are concerned that anti-fraud measures in such an important area are hampered in this way. We look to the Department and the Agency to let us know their plans for these new systems, and the timetable, as soon as possible.

45. We are disappointed that despite the Agency's anti-fraud initiatives and significant expenditure under their Security and Control Programme, they appear to be no closer to getting a grip on fraud on income support. Indeed, the Agency's estimate of fraud has risen to £1.77 billion.

46. In our report on Measures to Combat Housing Benefit Fraud, we urged the Agency to improve co-ordination of their anti-fraud measures with local authorities because fraud in awards of income support provided a gateway to fraud in housing benefit. We note the steps the Agency now plan to address this.

47. We share the concerns expressed by the disabled, and by the Social Security Select Committee ,[46] about the design and operation of the Benefit Integrity Project. We note the steps taken recently to validate proposed changes in benefit after review, but we look to the Agency to take urgent stock of the lessons emerging from their review of cases and appeals. This is necessary to ensure they get a firm grip on fraud whilst not penalising those who have merely made mistakes or errors due to their disability or faulty benefit design.

48. Given the high level of fraud estimated in disability living allowance, and concerns about the Benefit Integrity Project, we are concerned that the Agency does not expect to be able to measure accurately the effectiveness of their anti-fraud initiatives for some years. They need to develop methodologies and tools to enable them to do this as quickly as possible.

49. When we last looked at benefit fraud in 1996 we took some comfort from the Agency's major projects to tackle fraud, in particular the benefit payment card. While we recognise the need to make sure that the £50 billion of benefits that would be paid through these cards is secure and reliable, we are extremely disappointed at the significant delays that have arisen. And the Agency are still unable to provide a firm date by which the project will be fully operational. We expect the Agency to resolve the future of the benefit payment card project quickly.

50. We are concerned that the estimates of fraud savings claimed by the Agency tend to diminish when subjected to scrutiny by the Agency's internal auditors. And we are disappointed that the Agency are unable to provide any targets or timetable for reducing benefit fraud. Unless there is confidence in the level of savings claimed, based on the achievement of targets for fraud reduction, it will be difficult to evaluate the effectiveness of the substantial spending on the Security and Control Programme. We look to the Agency to improve the accuracy of their assessment of savings to measure the impact of their anti-fraud work on reducing benefit fraud and to set a timetable for containing and reducing benefit fraud.

17 C&AG's Report (HC 251-XII of Session 1997-98), paras 50-51, 77 Back

18 ibid, para 61 Back

19 Qs 3, 26-27, 57 Back

20 Qs 25-28, 75 Back

21 Qs 52-53 Back

22 Qs 3, 65 Back

23 Qs 11-13 Back

24 Q12 Back

25 Q117; Evidence, Appendix 1, p19 Back

26 Evidence, Appendix 1, p19 Back

²⁷ PAC 27th Report HC 366, Session 1997-98, paras 84-88 Back

28 Q6 Back