Fraud and error in housing benefit

The Department has failed to tackle problems with Housing Benefit fraud and error quickly or convincingly. As long ago as 1998, our predecessor Committee concluded that “the value of benefits wrongly paid is enormous”.1 Despite our predecessors’ concerns about the lack of information on trends in Housing Benefit fraud and on the types of fraud committed, by 2003 the Department was still unable to assure the Committee that the number of people committing fraud, its overall value or the level of fraud were reducing.2 The Department estimates it has spent £12.6 billion on Housing Benefit overpayments since 2000/01—billions of pounds that could instead have been used to improve the system. Overpayments increased from £980 million (4.6%) to £1.4 billion (5.8%) of Housing Benefit expenditure between 2010–11 and 2013–14; the highest rate of overpayments for any of the Department’s benefits. The Department claims it spotted that Housing Benefit overpayments were increasing in May 2013 but 18 months later was still unable to explain to us how it would be improving incentives for local authorities to tackle the problem. This was despite being asked by the Cabinet Office’s Fraud Error and Debt Taskforce in April 2014 to develop plans for reducing Housing Benefit losses. Housing Benefit overpayments accounted for 42% of overpayments across all welfare benefits, but the Department spent only 8% of its fraud and error funding on Housing Benefit. It argues that its initiatives to tackle fraud and error in other benefits would also reduce overpayments on related Housing Benefit claims. These initiatives are not, however, targeted at the biggest source of Housing Benefit overpayments which come from claimants who do not receive other benefits and who account for two-thirds of fraud and error. The rise in Housing Benefit 1 http://www.publications.parliament.uk/pa/cm199798/cmselect/cmpubacc/570/57002.htm 2 http://www.publications.parliament.uk/pa/cm200203/cmselect/cmpubacc/488/48802.htm 4 overpayments shows that the Department has not effectively targeted the major sources of fraud and error. Recommendation: The Department should review how it allocates money and resources to tackling Housing Benefit fraud and error. For each of the main sources of fraud and error, it should set out how—and by how much—its fraud and error initiatives aim to reduce Housing Benefit overpayments.

The Department is not doing anything to specifically target underpayments or the take-up of Housing Benefit, despite their importance to those most in need. The Department has reduced its funding to local authorities, who have in turn cut back on wider work to identify undetected fraud, claimant error and underpayment. Local authorities now focus on processing claims. Communicating directly with claimants is an effective way of finding fraud and error, but local authority interactions with claimants through interventions and reviews have declined. The Department argued that it is encouraging take-up through media campaigns. But its campaigns are targeted largely at overpayments and the need for claimants to report changes to their circumstances. The Department also argued that Universal Credit would make underpayments less likely but the timetable for implementation remains very uncertain. The Department did not highlight any other work to specifically tackle underpayments and improve the take-up of benefits. Recommendation: The Department must report back to us within 6 months of this report on what measures it has introduced specifically to target underpayments and encourage legitimate take-up.

The Department has provided weak incentives for local authorities to tackle fraud and claimant error. The Department now relies mainly on incentives in the subsidy regime to encourage local authorities to reduce fraud and error. The subsidy regime is used to reimburse local authorities for paying the correct amount of money to claimants and seeks to encourage the effective administration of Housing Benefit. Local authorities told us that it creates disincentives for finding overpayments in their caseloads, including claimant errors and fraud which account for 90% of Housing Benefit overpayments. If a local authority identifies a case of fraud or claimant error it loses 60% of the money paid out in housing benefit to the claimant. The weak incentives for local authorities to detect claimant error and fraud are exacerbated by constraints on local authority funding, including a 17% fall since 2010–11 in the funding available for administering Housing Benefit. This has resulted in local authorities reducing the amount of work they undertake to detect overpayments after the initial assessment has been made. The Department has previously provided additional incentives and funding to tackle fraud and error, for example through the Security Against Fraud and Error (‘SAFE’) framework. The Department accepts that it needs to review the subsidy regime to strengthen the incentives for local authorities in the period before Universal Credit is implemented fully. Recommendation: The Department should produce a proposal for how to strengthen incentives so that local authorities tackle Housing Benefit fraud and error more effectively. It should work with local authorities and gain approval from 5 the Cabinet Office’s Fraud Error and Debt Taskforce before sharing the proposal with us within 6 months.

Without a good understanding of local levels of fraud and error, the Department is not able to target efforts effectively. The Department cannot tell which local authorities are doing well or badly in controlling fraud and error. It uses measures on the speed of processing claims as a proxy to assess local authorities’ performance, but these do not give any indication of the strength of fraud and error controls. With poor information on local authority working practices, it is unsurprising that there has not been much sharing of good practice in recent years. Despite the Committee’s repeated calls since 1998 for better information about local fraud and error, and repeated assurances from the Department that this would improve, the Department still works with national estimates which are inadequate to give local level insight into fraud and error. Given the scale of Housing Benefit overpayments—£12.6 billion since 2000/01—better information would be essential in strengthening the Department’s oversight role and targeting fraud and error initiatives. The Department told us that it is not considering expanding its measurement methodology to provide local estimates of fraud and error on the basis that it is too expensive. We are concerned that the Department has not properly assessed different ways in which it could improve information about local fraud and error, and has provided inflated cost estimates to justify continued inaction. Recommendation: Within the next 6 months, the Department must provide us with a full analysis of options to identify whether there is a more cost-effective way of producing local estimates of the level of fraud and error, and how it plans to assess the relative performance of local authorities in reducing Housing Benefit overpayments.

The Department is expecting significant benefits from HMRC’s real-time information, but is doing little else to tackle Housing Benefit fraud and error in the still unclear and very long transition before Universal Credit. The Department has begun to use data on claimants’ earnings from HM Revenue and Customs real- time information system to strengthen its response to income-related overpayments, which accounted for £637 million of overpayments in 2013–14. This offers a promising longer term solution for automating the use of earnings-related data under Universal Credit; but the timetable for implementing Universal Credit remains unclear and uncertain. The Department has not yet worked out how it will reduce Housing Benefit fraud and error in the transition period to Universal Credit. Neither has it produced a plan to tackle overpayments arising from sources other than income. Housing Benefit accounts for 42% of overpayments across all benefits and yet there are few initiatives to tackle other types of Housing Benefit overpayments. Given the lack of evidence that the Department is getting to grips with fraud and error we view with scepticism the Department’s confidence that it will meet its target to reduce total fraud and error overpayments to 1.7% of benefit expenditure by March 2015. Recommendation: The Department must demonstrate it has a convincing response to tackle Housing Benefit fraud and error before Universal Credit is implemented and the use of real-time information is automated. It should report to us within 6 6 months with a clear plan to tackle the major sources of loss on Housing Benefit. It should also set out what savings it has achieved across benefits against its 1.7% target, and which initiatives have realised these savings.

The Department’s introduction of the single fraud investigation service creates risks to other local services through reducing local knowledge. The single fraud investigation service (SFIS) brings together fraud investigators from the Department, local authorities and HMRC to investigate fraud across the whole welfare system. Initially the Department delayed the introduction of SFIS from 2011 to 2015 to be more in line with the introduction of Universal Credit. Now it tells us “I do not see what the connection is” between the two programmes and is not planning to delay SFIS to match Universal Credit’s new roll-out plans. The Work & Pensions Committee has recommended the Department aligns introduction of SFIS with the national implementation of Universal Credit. Now that Universal Credit has been delayed, local authorities have at least two more years and possibly many more when they will be expected to administer Housing Benefit. However local authorities have already lost much of their investigative capability and they are therefore losing their local knowledge and their ability to be effective in tackling fraud. Recommendation: The Department should provide a more complete assessment of the wider costs to local authorities of the SFIS programme, and consider how the benefits of local knowledge and data sharing can be maintained in the longer term. 7 1 The level of fraud and error in Housing Benefit

In 2013–14, the estimated level of Housing Benefit underpayments was £370 million, of which some £290 million was due to mistakes by claimants. In addition, not all potential claimants take up their entitlement, probably because of a lack of awareness. We were concerned about the level of unclaimed Housing Benefit and asked the Department what work it was doing to tackle underpayments. It argued that its awareness campaigns were encouraging take-up by those who are eligible for Housing Benefit. But those campaigns— are targeted at reducing overpayments by encouraging people to report fraud and raising awareness of the need for claimants to report changes to their circumstances. The Department acknowledged that its focus has been on reducing the level of overpayments rather than increasing uptake and identifying underpayments. We would expect the Department to be undertaking wider activities to ensure that Housing Benefit is received by the claimants most in need.15

The local authority witnesses acknowledged that they have cut back on the wider work that they used to undertake to identify fraud and error in claims that are already in- payment. They explained that their while their caseload had increased the funding received from the Department had gone down. As a result, local authorities have focused on the core requirement to process claims quickly and accurately, reducing their investment in wider initiatives to tackle fraud and error.16 12 C&AG’s Report, para 19 13 Q 134 14 C&AG’s Report, para 3.7 15 Qq 147–153; Department for Work and Pensions, Fraud and Error in the Benefit System: 2013/14 Preliminary Estimates (Great Britain), 15 May 2014 16 Q 1 10 2 Incentives for local authorities to tackle fraud and error

The Department uses a subsidy process to reimburse local authorities for paying the correct amount of Housing Benefit to claimants. The subsidy regime seeks to encourage the effective administration of Housing Benefit. Local authorities record the details of claims on a subsidy form, including where they have identified any overpayments, and whether these relate to claimant or official error. In 2013–14, the Department paid local authorities £23.5 billion in subsidy compared with total Housing Benefit payments to claimants of £23.9 billion. The overall rate of reimbursement from the Department to local authorities was 98%.17

Local authority witnesses told us that the subsidy system creates disincentives to identify fraud and claimant error. For example, if a fraud overpayment is identified by the local authority, the Department will reimburse 40% of the payment, with the local authority losing 60% of it. If the local authority is then unable to recover the fraudulent payment (as was the case when Islington identified a £35,000 fraud but the claimant moved abroad and the money was irrecoverable) then the local authority would have been better off not identifying the fraud in the first place.18

The Department could not recall when it last reviewed the subsidy regime and did not agree with our view that the current system provides conflicting incentives.19 The Department stated that the system encourages local authorities to process claims correctly and provides the Department with financial controls over the £23.9 billion of Housing Benefit expenditure.20 The Department maintained that the system does create incentives for local authorities as they are paid less if they make an incorrect payment and are allowed to keep any overpayments if they identify and recover them.21

The National Audit Office reported that the scheme is not designed specifically to target fraud and claimant error and does not create strong incentives to detect overpayments after the claim has been awarded. Fraud and claimant error accounts for 90% of all Housing Benefit overpayments.22 The Department confirmed that it does need to incentivise local authorities to deal with the large amounts of undetected fraud and claimant error in the caseload and is undertaking work with local authorities to review the subsidy regime.23 We asked witnesses from local authorities what incentives the Department could give to encourage them to put more resources into the administration of Housing Benefit.24 The local authority witnesses confirmed that there is an asymmetry 17 C&AG’s Report, para 2.11 18 Q 7 19 Qq 53–55 20 Q 67 21 Q 55 22 Q 63; C&AG’s Report, para 14 23 Qq 63–Q67 24 Q 4 11 between the costs incurred and the benefits that they are capable of achieving under the current funding arrangements.25 In an ideal world, any incentive scheme would not be labour intensive to apply for and would provide funding certainty to employ staff. The local authorities highlighted that the scheme in place up until 2006, the SAFE (Security Against Fraud and Error) initiative, provided more incentives such as rewarding prosecutions and sanctions.

The Department has limited information on local authorities’ performance in tackling Housing Benefit fraud and error. It does not have an accurate measure of levels of fraud and error at the local level to determine which local authorities are performing well. The Department uses information on the speed of processing claims as a proxy for local authority performance, but there is no correlation between this measure and the identification of fraud and error.26 Due to the lack of information on local performance, the Department had facilitated very little sharing of best practice between local authorities.27

Since 2000–01 the estimated total overpayments on Housing Benefit add up to £12.6 billion. We have made repeated calls for better information about local fraud and error.28 In 1998 we concluded “it is totally unacceptable that seven years after we last looked at this issue, Housing Benefit fraud should exceed £900 million and the Department still do not have information to show whether fraud is increasing, or all the information they need on the types of fraud, including variations at regional and local level”. Given the scale of Housing Benefit overpayments, better measurement of fraud and error at a local level would help identify the root causes of overpayments and provide valuable information to target the problem.29

The Department estimates the monetary value of fraud and error by testing a sample of over 12,000 Housing Benefit claims. It undertakes testing on benefit cases in payment to determine if there is any evidence of official error, claimant error or claimant fraud. This national measurement programme feeds into the Department’s published fraud and error statistics. The Department estimate that it currently costs £5 million to undertake this work.30 The Department stated that it is not considering expanding the national measurement exercise to provide local estimates of fraud and error as it would be too expensive to increase the sample size to produce statistically reliable results. 31

After the hearing the Department wrote to the Committee and claimed that it would cost £780 million to obtain local data at the same level of accuracy as the national methodology. The cost estimate is based on the Department’s view that it would need to sample 1.8 million cases—an average of 4,000 per local authority—to follow the methodology used to measure fraud and error nationally. It also estimated that it would 25 Q 6 26 Qq 137, Q139 27 Q 138 28 Committee of Public Accounts, Measures to Combat Housing Benefit Fraud, Session 1997–1998, Twenty Seventh Report, HC 366, 26 March 1998; Committee of Public Accounts, Tackling Benefit Fraud, Session 2002–2003, Thirty First Report, HC 488, 4 July 2003; Letter from Robert Devereux to Margaret Hodge 21 November 2014. 29 Q 154 30 Letter from Robert Devereux to Margaret Hodge 21 November 2014 31 Q 88 12 cost £180 million each year to estimate the volume rather than the value of overpayments.32 We question whether a sample of this size is necessary, and we would encourage the Department to explore options for producing estimates at a lower cost, for example focusing on the largest 190 local authorities which account for almost 80% of expenditure, or considering the level of accuracy needed. 3 Tackling fraud and error

The Department has a target to reduce overpayments to 1.7% of benefit expenditure by March 2015 but was unable to set out its strategy to tackle Housing Benefit overpayments, which comprise 42% of all the Department’s benefit overpayments.33 Instead it highlighted its proposed changes to the fraud and error measurement methodology, which are expected to reduce the net losses on Housing Benefit from 3.5% to 3.1%. As a result, the Department remains confident that it will meet its 1.7% target for overall benefit overpayments.34

We have repeatedly called for the Department to use information better to reduce fraud and error.35 The Department collects information on fraud and error by claimant group, risk type and the year it entered the system. The major cause of overpayments is changes in claimants’ income, which accounted for £637 million (46%) of fraud and error in 2013–14 (Figure 2). Overpayments due to claimants mis-declaring or not reporting their income has increased by 32% since 2011–12. 36 32 Letter from Robert Devereux to Margaret Hodge 21 November 2014 33 Q 48 34 Q 50 35 Q 154, Committee of Public Accounts, Measures to Combat Housing Benefit Fraud, Session 1997–1998, Twenty Seventh Report, HC 366, 26 March 1998; Committee of Public Accounts Tackling Benefit Fraud, Session 2002–2003, Thirty First Report, HC 488, 4 July 2003. 36 C&AG’s Report, para 3.10 13 Figure 2 Source: C&AG’s Report, Housing Benefit fraud and error, Session 2014–15, HC720, 17 October 2014, Figure 15

The Department is now focusing its efforts on reducing fraud and error due to earnings by using HMRC’s real-time information. In summer 2014, the Department started a project to check local authorities’ claimant data against the real-time data on earnings. It forecast that this would identify 223,000 incorrect Housing Benefit claims and reduce fraud and error by £30 million. Resource constraints have limited local authorities’ ability to process the changes and the expected reduction is now estimated at £15 million in the current year.37

Real-time information will provide information on claimants’ earnings only for people paid through the PAYE system. Other groups, such as the self employed or the cash economy, will not be captured and the potential to significantly reduce overpayments is not yet known.38 We challenged the Department on how they are going to target claimants not covered by real-time information.39 The Department stated these claimant groups will be covered under Universal Credit when claimants will interact with the Department on a monthly basis. The Department indicated that it needed to explore more regular contact with claimants who declare themselves as self-employed before Universal Credit is implemented.40 37 C&AG’s Report, para 3.19 38 C&AG’s Report, para 21 39 Q 46 40 Q 47 14

The Department is introducing the single fraud investigation service, which aims to bring together fraud investigators from the Department, local authorities and HMRC to investigate fraud across the welfare system.41 The Government announced that the new service would roll out in 2014–15.42 The service was originally due to be introduced in 2011 but was delayed to be more in line with the introduction of Universal Credit.43 The Department has now decided that the single fraud investigation service should go ahead despite the delayed roll-out of Universal Credit and told us that it did not see any connection between the two.44 Local authority witnesses stated that this could cause a significant gap in local fraud knowledge in the interim period, during which they remain responsible for administering Housing Benefit claims.45 The Work and Pensions Select Committee concluded that it made no sense to roll out the service nationally, ahead of the national implementation of Universal Credit, while local authorities retain responsibility for Housing Benefit. It recommended that the implementation of the service be aligned where practicable with the expansion of the Universal Credit Pathfinder areas and with national implementation of Universal Credit. 46

Local authority witnesses expressed concerns that the introduction of the single fraud investigation service will reduce their ability to identify fraud. They were concerned at the loss of local intelligence garnered by local fraud investigators, particularly around household occupancy. Local authorities share claimant information collected from a wide range of services—blue badges, single person discount for council tax, financial assessments for social care and parking permits—enabling them to help claimants but also useful in tackling fraud. 47

The Department estimated the single fraud investigation service will deliver an additional £559 million by 2021, approximately half of which will be Housing Benefit. We challenged the Department on the wider implications of the loss of local investigations teams.48 They assured us that centralising fraud investigators will enable the redeployment of resources into areas of higher levels of fraud. We also questioned the Department on how they will replicate the knowledge and capacity of fraud investigators once they have moved out of local authorities and whether the number of fraud investigators will be the same after implementation. The Department remains hopeful that this will be the case. 49 41 C&AG’s Report, para 1.15 42 Letter from Robert Devereux to Margaret Hodge 21 November 2014 43 House of Commons Work and Pensions Select Committee, Fraud and error in the benefits system, Sixth Report of Session 2013–14, HC 1082, para 64 44 Q 129 45 Q 15 46 House of Commons Work and Pensions Select Committee, Fraud and error in the benefits system, Sixth Report of Session 2013–14, HC 1082, conclusion para 9 47 Q 15 48 Q 107 49 Qq 109, 125 15