Fraud and error in the benefits system - Work and Pensions Committee Contents


3  Housing Benefit, Universal Credit and the role of local authorities

The administration of Housing Benefit

30. Housing Benefit (HB) helps claimants on low incomes, both in and out of work, to meet the cost of their rent. It is currently administered by local authorities on behalf of DWP but will become part of Universal Credit (see below). The main DWP-administered unemployment benefits—ESA, JSA and Income Support—can act as a "passport" to HB. Claimants of these benefits apply for HB via Jobcentre Plus (part of DWP). Claimants who are not receiving qualifying benefits apply directly to their local authority for HB and their entitlement is based on their income and the level of local rents.[25]

31. Excluding the State Pension, more is spent on HB than any other social security benefit—nearly £24 billion in 2012/13. The monetary value of HB overpayments due to fraud and error (£1.2 billion) is more than twice that of any of the other core benefits.[26]

THE MAIN HOUSING BENEFIT FRAUD AND ERROR RISKS

32. By far the largest risk of fraud and error in HB arises from the requirement for claimants accurately to report their income. Incorrect reporting of claimant income accounted for some 47% (£574 million) of HB fraud and error in 2012/13. Other notable categories of risk identified by the NAO include "passporting" issues i.e. where entitlement to HB is linked to entitlements to other benefits or services (part of a miscellaneous group of risks which made up 11% of overpayments, £134 million) and people claiming as a single person when they live with a partner (9%, £110 million).[27]

The Automated Transfer to Local Authority Systems

33. Since February 2012 local authorities have received daily updates of changes in benefit and Tax Credit entitlements via the Automated Transfer to Local Authority Systems (ATLAS) IT system, which DWP "hopes will lead to a significant reduction in fraud and error within locally administered benefits."[28]

34. The Local Government Association (LGA) told us that some local authorities had experienced considerable difficulties in processing the volume of data received via ATLAS. Councillor Sharon Taylor, Chair of the LGA Finance Panel, said that local authorities had to "wade through" a lot of information to find what they needed. The costs of "refining" the ATLAS data were currently borne by local authorities. The LGA argued that local authorities should be financially incentivised by DWP to carry out this work.[29]

35. Mike Driver, DWP's Director General of Finance, told us that this problem had not been brought to his attention. He accepted that ATLAS provided a lot of data but he highlighted that the system had been designed to notify local authorities automatically of all relevant changes in benefit and Tax Credit entitlements. He said that ATLAS was a relatively new system and as such was likely to develop further. DWP would consider reducing the volume of data local authorities receive through ATLAS, "if that was the solution."

36. In relation to the costs to local authorities of analysing ATLAS data, DWP noted that there was already a financial incentive for local authorities to use ATLAS data to reduce HB fraud. Mike Driver explained that, if a HB claim is found to be fraudulent, DWP only reimburses the local authority 40% of the money paid out to the claimant. If local authorities subsequently pursue and recover money paid out in relation to a fraudulent claim, they can claim 100% of the recovered money from DWP, in addition to the initial subsidy paid to them. However, he accepted that whether or not this system was sufficient to incentivise local authorities to allocate the necessary resources to analysing ATLAS data, in addition to committing resources to other activities to address fraud and error, was "another matter".[30]

37. We welcome the introduction of the Automated Transfer to Local Authority Systems (ATLAS). We recommend that DWP and appropriate representatives from local government conduct a joint review of ATLAS, including the level of resources needed to analyse ATLAS data, with a view to enabling local authorities to access the DWP data they need to verify Housing Benefit claims more easily. This review should be conducted and improvements implemented before the end of 2014.

Addressing fraud and error in Universal Credit

38. As Universal Credit (UC) rolls out, HB and five out-of-work benefits and in-work Tax Credits will cease to be paid separately to claimants: they will be incorporated into a single household UC payment which, for the majority of claimants, will be paid directly to them each month. HB will no longer be administered by local authorities; support for housing costs will be centrally administered by DWP.[31]

39. The principal aim of UC is to create a single system of in-work and out-of-work benefits, with payments which taper off as claimants' earnings rise, in a way which more clearly encourages claimants to enter, and progress in, employment.[32] UC is intended to be a simpler system. As the NAO has noted, DWP hopes that:

    […] in the long term, this streamlining of benefit will remove or reduce some of the current complexities around benefit entitlement, verification of claimant circumstances and administrative requirements that can increase opportunities for fraud and error.[33]

IMPLEMENTATION TIMETABLE

40. We have been consistently supportive of the policy intentions of UC, but we have also noted delays in the development of the requisite IT and the consequent slow pace of implementation. While DWP still intends largely to complete national implementation of UC in 2017 as originally planned, the build-up of UC claims has proceeded at a far slower pace than originally anticipated; national implementation was originally scheduled for October 2013 but only 5,250 people were claiming UC by the end of January 2014.[34] When we took evidence for this inquiry no detailed published plans were available for extending implementation beyond the 10 "Pathfinder" areas, the first four of which began in the North West of England during 2013. This led us to conclude in our Report on Universal Credit in April 2014 that "it is difficult to envisage how the volumes required to meet the most recent timetable are to be achieved."[35] On 29 April DWP announced that from June 2014 UC will be implemented gradually in more areas of the North West, until all 90 Jobcentres in the region are accepting new UC claims. The announcement did not include any timetable for the completion of this phase of UC implementation.[36]

41. Our recent Report also highlighted the NAO's view that fundamental uncertainties around the IT solution persist, including: "how it will work; when it will be ready; how much it will cost; and who will do the work to develop and build it."[37] Rushing the implementation would have risked creating significant issues with fraud and error and so we agreed with the use of pilots and pathfinder areas. This inquiry could not therefore consider the detail of how UC will work in practice in relation to tackling fraud and error. Instead, we examined, at a more theoretical level, the opportunities for UC to reduce the main fraud and error risks and considered the actions DWP needs to take to protect the system.

REAL-TIME INFORMATION ON EARNINGS

42. As noted in chapter 2, and above in relation to HB, by far the largest risk of incorrect benefit payments is related to the requirement for working claimants accurately to report their earnings.

43. UC will utilise RTI, a new system which has been developed by HMRC to facilitate the collection of PAYE income tax. HMRC's RTI system receives data from employers about employees' PAYE income and tax, National Insurance and other deductions. Under UC, this information is automatically transferred to DWP, and used to calculate monthly benefit entitlements from this accurate and up-to-date earnings information.[38]

44. DWP stated that using RTI has the potential to:

·  Prevent fraud and error on new claims before payments are issued;

·  Prevent fraud and error from entering the system when there is a change of circumstances;

·  Detect existing fraud and error for undeclared or declared earnings and non-state pensions; and

·  Create efficiencies in the payment, referral and evidence gathering processes.[39]

45. Most witnesses acknowledged the considerable potential benefits of RTI, but some also pointed out that the system will not detect earnings incorrectly declared by self-employed people or those working in the cash economy. The Local Authority Investigation Officers Group (LAIOG) was concerned that RTI might even increase the number of people choosing to work for "[undeclared] cash in hand".[40]

46. Mike Driver confirmed that the Government's 2010 forecast, that RTI would produce £400 million of savings per annum through reduced benefit fraud and error, remained unchanged. He acknowledged that this figure was largely dependent on the full national implementation of UC; however, he also noted that DWP was exploring whether RTI could be applied to existing benefits before they are integrated into UC.[41] The day after giving oral evidence to our inquiry, DWP announced that it planned to use RTI to combat JSA fraud.[42]

47. Mr Gauke told us that RTI was already being used to verify Tax Credits renewals and had produced a "very quick return". HMRC expected to achieve savings of £820 million between 2014/15 and 2016/17 by using RTI in this way.[43]

48. By utilising real-time information (RTI) on PAYE income, Universal Credit has the potential to substantially reduce incorrect benefit payments due to inaccurate or late reporting of claimants' earnings; and it is projected to produce significant savings in Tax Credits. However, RTI cannot provide the complete solution, as it will not apply to a significant proportion of claimants who are paid outside the PAYE system, including the self-employed. Moreover, the full gains of RTI in relation to reducing benefit fraud and error are largely dependent on the successful national implementation of Universal Credit, which is at least three years away by the most optimistic schedule. We therefore welcome steps to apply RTI to existing benefits where possible and recommend that DWP and HMRC consider methods to automate this process.

THE IDENTITY ASSURANCE PROGRAMME

49. The Government's original intention was that UC would be "digital by default"—the assumption was that claims would be made and managed online.[44] One of the key areas of uncertainty in the IT is the development of a sufficiently robust system for verifying claimants' identities online. The NAO reported in September 2013 that the system the Government intended to develop to ensure that "all digital public service users can assert their identities safely, securely and simply", the Identity Assurance programme (IDA), was "missing" from the UC Pathfinder in June 2013.[45]

50. In December 2013, before we launched this inquiry, DWP indicated that IDA was at a very early stage of development. Howard Shiplee, DWP's Senior Responsible Officer for UC, told us that it would take "some considerable time to get to a totally online system".[46]

POTENTIAL FOR BIOMETRIC SYSTEMS

51. Nuance Communications, which delivered a voice-based identity verification system used in the Australian public sector, believed that traditional systems were often both "time consuming and inconvenient" to claimants and "vulnerable to fraud". As far as Nuance Communications was aware, the IDA programme being developed for UC would rely on people providing personal information over the telephone and online. Its view was that systems such as these can also facilitate identity fraud.[47] CIFAS, a statutory anti-fraud body, reported that in 80% of confirmed identity theft fraud cases, personal data of this kind had been obtained from the internet.[48]

52. Nuance Communications argued that a voice-based system could both improve customer service and reduce the risk of fraud in the UK benefits system. It explained how the system worked in Australia:

    When a citizen first enrols in the system, they can opt for a voice sample to be collected, and a voice print to be produced and stored for future use. The next time the citizen contacts the organisation, perhaps wishing to chase up a payment, a second voice sample is collected and compared to the stored voice print to confirm the person is who they say they are. If needed, this comparison can be completed within a few seconds.

Nuance Communications claimed that the system was "highly accurate" and "impossible to impersonate or reverse engineer". [49]

53. Other private sector witnesses were supportive of biometric systems of this kind but stressed that ID verification systems were not primarily anti-fraud measures; fraud protection was merely a positive by-product.[50] Sean Duffield of Nuance Communications acknowledged that biometric systems were "not the be all and end all" but could be "part of the overall solution" to reducing fraud in the system.[51]

54. Lord Freud told us that DWP had piloted a biometric system "a few years ago" but that the Department "did not find it hugely helpful". He said that responsibility for the development of IDA had now passed to the Cabinet Office, which was exploring the possibility of an identity verification system which could be applied "across all of Government." DWP would "look to use that system for UC." We were not given any details about how the Cabinet Office system might work or the timescale for its implementation.[52]

55. We believe that in the longer term biometric identity systems could have an important role to play in identity verification processes across government. We recommend that the Government continue the Cabinet Office work to explore a government-wide system; and evaluate the benefits of biometric identity verification in the social security system and more widely across public services, including by examining the effectiveness of the voice-recognition system currently used in Australia.

THE INTEGRATED RISK AND INTELLIGENCE SERVICE

56. The LGA believed that, as support for housing costs is integrated into UC, DWP will need access to a range of information currently held by local authorities, on individual properties and the composition of households within those properties. It was concerned that DWP would be unable to replicate the cross-checking of data currently undertaken within and between local authorities in verifying HB claims.

57. Councillor Taylor explained that local authorities could cross-check claims "across the whole raft of council activities, so everything from Blue Badge car parking, Council Tax support or tenancy."[53] The LGA and LAIOG were concerned that, without automated access to these types of data, UC might be more vulnerable to fraud and error, in relation to housings costs, than HB.[54]

58. In April 2013 the Communities and Local Government Committee expressed similar concerns that, without the ability to cross-check local authority data, UC could be vulnerable to fraudulent multiple housing claims. In response, the Government stated that an IT system called Integrated Risk and Intelligence Service (IRIS) would act as a central data hub. IRIS would have access to a "huge number" of databases. It would be able to cross-check data and provide similar safeguards against fraudulent claims to those introduced by local authorities administering the HB system.[55]

59. The NAO found that IRIS, like IDA, was "missing" from the UC Pathfinder in June 2013.[56] In oral evidence on 17 March 2014 the LGA and LAIOG said that they were not aware that it had yet been developed or whether it was in use in any of the UC Pathfinder areas.[57]

60. Lord Freud told us that DWP was using a "relatively straightforward data-matching process" in the UC Pathfinder. He insisted that, once UC is fully implemented, DWP will use an automated system to cross-check claims by a "very similar process" to that used by local authorities in relation to HB. He told us that IRIS was still in development and reported that it was now DWP's intention to employ "IRIS analysts" to "watch processes go through". However, it was not clear how or when DWP would achieve automated access to the range of property data currently available to local authorities.[58]

61. Local authorities expressed concerns about DWP's current capability to administer the housing element of Universal Credit with the same level of safeguards which local authorities are currently able to apply to Housing Benefit administration, and therefore without increased risks of fraud and error. The Integrated Risk and Intelligence Service (IRIS), or a similar system to allow DWP to cross-check claims against the range of property data held by local authorities, will be vital. Such a system will need to be fully developed and tested before national implementation of Universal Credit commences.


25   Gov.uk web page [accessed 9 April 2014] Back

26   NAO [FAE0020) para 6 Back

27   NAO [FAE0020) figure 5 Back

28   NAO, Report by the Comptroller and Auditor General: Department for Work and Pensions 2012-13 Accounts, December 2013, para 45 Back

29   Q43 Back

30   Qq175-80 Back

31   Work and Pensions Committee, Fourth Report of Session 2013-14, Support for housing costs in the reformed welfare system, HC 720, chapter 8 Back

32   Work and Pensions Committee, Third Report of Session 2012-13, Universal Credit implementation: meeting the needs of vulnerable claimants, HC 576, paras 5; 135 Back

33   NAO, Department for Work and Pensions: 2012-13, Report by the Comptroller and Auditor General, December 2013, para 5 Back

34   DWP, Universal Credit - Experimental official statistics to January 2014, April 2014 Back

35   Work and Pensions Committee, Fifth Report of Session 2013-14, Universal Credit implementation: monitoring DWP's performance in 2012-13, HC 1209, para 30 Back

36   DWP press release, 29 April 2014, "Universal Credit: First year of welfare transformation and North West next steps" Back

37   Work and Pensions Committee, Fifth Report of Session 2013-14, Universal Credit implementation: monitoring DWP's performance in 2012-13, HC 1209, para 45 Back

38   Work and Pensions Committee, Third Report of Session 2012-13, Universal Credit implementation: meeting the needs of vulnerable claimants, HC 576, paras 142-9 Back

39   DWP (FAE0013), paras 47-8 Back

40   LAIOG (FAE0008) Back

41   Q217 Back

42   "New measures to protect the integrity of the benefits system", DWP press release, 8 April 2014 Back

43   Q218 Back

44   Work and Pensions Committee, Third Report of Session 2012-13, Universal Credit implementation: meeting the needs of vulnerable claimants, HC 576, paras 17-20 Back

45   NAO, Universal Credit: Early progress, HC 621, September 2013, figure 10 Back

46   Oral evidence taken on 9 December 2013, Q31 Back

47   Nuance Communications (FAE0014) para 4 Back

48   Q127 Back

49   Nuance Communications (FAE0014) para 14 Back

50   Q125 [Simon Dukes and Jan Smith] Back

51   Q122 Back

52   Qq206-7 Back

53   Q38 Back

54   Qq59-64 [Debbie Gibbons and Cllr Taylor] Back

55   Communities and Local Government Committee, Ninth Report of Session 2012-13, Implementation of welfare reform by local authorities, HC 833, paras 123; 126; Q253 [Lord Freud] Back

56   NAO, Universal Credit: Early progress, HC 621, September 2013, figure 10 Back

57   Q57 [Debbie Gibbons and Mick Hopkins]; Q60 [Cllr Taylor] Back

58   Qq181-4 Back


 
previous page contents next page


© Parliamentary copyright 2014
Prepared 15 May 2014