Department For Work and Pensions: Progress in tackling benefit fraud

Executive Summary

National Audit Office Value for Money Report

1. Benefit fraud is a crime and undermines public confidence in the benefits system. Tackling benefit fraud is inherently difficult. It is in the nature of all welfare systems that some people seek to defraud them and that some will succeed. Good progress has been made in reducing the estimated value of benefit fraud, which represents a substantial achievement by the Department and its staff. Estimates by the Department for Work and Pensions [Footnote 1] (the Department) suggest overpayments due to fraud have fallen from 2 billion (to the nearest 0.5 billion) in 2000-01 to 800 million (to the nearest 0.1 billion) in 2006-07, although definitional changes, described in Figure 3, have helped to some extent. This is a reduction of 60 per cent. Reducing levels of fraud and error in the benefit system continues to be a key objective of the Department, which has had its main accounts qualified for the last 18 years due to the levels of fraud and error (Box 1).
   
   

   Box 1: What is the difference between fraud and error?

   The Department defines fraud as those cases where customers deliberately claim money to which they are not entitled. Customer error occurs when customers provide information to the Department which is inaccurate, incomplete or untimely, but without dishonest intent, and as a result the benefit paid is inaccurate. Official error occurs when officials fail to apply specific rules or do not take into account all the notified circumstances. We have not looked explicitly in this report at internal fraud. The Departments segregation of fraud and customer error estimates will always be subjective, being dependent on the conclusion on intent for individual cases investigated. Whilst official error is relatively isolated and attributable to administrative failures, fraud and customer error are harder to separate as the fundamental factor is proof of fraudulent intent.

   
   
   
2. Although fraud has fallen in recent years (Figure 2), we have chosen to examine progress in tackling external fraud because of the amount of money involved and the considerable attention the Department has given to it in recent years. To see what progress has been made since our last report in 2003, we considered:
   • What is benefit fraud and what progress has been made in tackling it? n How effective are the Departments interventions to tackle benefit fraud?
   • Do the Departments counter-fraud interventions represent value for money?
3. Protecting the integrity of the benefits system is at the heart of many of the Departments policy and operational initiatives. Since our last report, the Department has carried out major organisational change, including, for example, the roll out of Jobcentre Plus and the introduction of contact centres, designed to improve the services they provide at the same time as ensuring benefits are paid correctly to those eligible. These and other changes have taken place at the same time as the Department has reduced its headcount, expected to come down by 30,000 by March 2008. In addition, the development, for example, of regular interviews with jobseekers, medical examinations with people claiming disability and incapacity benefits and more checks at the outset of a claim for benefit are all designed to assist in getting the benefit right, encouraging compliance with regulations, and increasing the opportunity for the Department to identify those who are likely to be deliberately defrauding the system. Such measures have all been the subject of reports by the Comptroller and Auditor General. However, the Department has a range of interventions involving around 4,000 staff and costing more than 150 million deliberately designed to prevent fraud and identify it where it appears to be occurring. This report focuses on these anti-fraud measures.
   
   
   

   Figure 2: Comparison of benefit expenditure and losses due to fraud, 2001-02 to 2006-07

   	2001-02	2002-03	2003-04	2004-05	2005-06	2006-07
   Total benefit expenditure, billion	107.0	112.0	105.8	110.9	115.8	119.8
   Total number of benefit recipients, million	17.0	17.0	17.0	17.0	18.0	18.0
   Total estimated fraud and error, billion	3.0	3.0	2.6	2.6	2.7	2.7
   Total estimated fraud[1], billion	2.0	2.0	1.0	0.9	0.8	0.8
   Estimated fraud as a percentage of expenditure, per cent	1.9	1.8	1.0	0.8	0.7	0.6
   Source: Department for Work and Pensions Resource Accounts, 2001-02 to 2006-07
   1.[back] Estimated levels of fraud and error for 2003-04 were restated in 2004-05, using a more accurate methodology. The figures shown for 2003-04 are the restated ones.

   
   
   

   Overall, good progress has been made in reducing fraud.

4. Overall, good progress has been made in reducing estimated levels of benefit fraud and in March 2006 the Department reported that it had met its Public Service Agreement (PSA) target to reduce Income Support and Jobseeker’s Allowance fraud and error by 50 per cent, by 2005-06. As part of the 2004 Spending Review, the Department was given a new target to reduce fraud and error in Income Support and Jobseeker’s Allowance by a further 15 per cent. In 2006-07, an estimated £800 million (to the nearest £0.1 billion) was lost due to fraud, around 0.6 per cent of total benefit expenditure, but down from an estimated £2 billion (to the nearest £0.5 billion) in 2001-02. In particular:
   • estimated fraud against Jobseeker’s Allowance, Income Support and Pension Credit has fallen from £900 million to £380 million since 1997-98 – a reduction of nearly 60 per cent;
   • since 2002, estimated fraud in Housing Benefit has reduced by around half to 1.0 per cent of expenditure in March 2006, although estimates for customer and official error have increased, and the Department reported that it has not met its PSA target for Housing Benefit, which covers both fraud and error.
   
   

   The Department has a range of targeted initiatives to tackle fraud

5. In our previous report in 2003 we recommended that the Department:
• gave greater focus to higher risk activity;
• ended the decline in fraud prevention and investigation activity;
• ensured it had reliable information on costs and results to allow resources to be allocated efficiently;
• did not make administrative savings at the expense of the reduction of losses; and
• increased the number of prosecutions and assessed the success of other deterrents and sanctions to prevent fraud.
6. The Department’s range of counter-fraud measures and their interconnections are summarised in Figure 1. Since our previous report, the Department has taken a number of initiatives to enhance its counter-fraud activity. In particular, it has:
• developed a more risk-based and intelligence-led approach to countering fraud, using a risk assessment tool called FRISC to prioritise cases;
• started to roll out in 2007 a new case management system called FRAIMS to speed up the management of cases;
• developed its advertising campaigns to target specific groups and those likely to commit fraud;
• restructured the Counter-Fraud Investigation Service into the Fraud Investigation Service (which investigates all fraud against all benefits administered by the Department’s agencies) and a newly developed Customer Compliance function (which deals with lower risk cases where full criminal investigation is judged unnecessary or where a sanction is unlikely to result);
• introduced more checks at the outset of a claim for benefit to prevent fraud entering the system in the first place, and begun to pilot voice risk analysis to evaluate its effectiveness in helping to identify high risk cases; and
• made extensive use of data matching and developed a more timely Data Matching Service, using private sector data to detect fraud.

The Department is ahead of equivalent bodies in other countries in understanding and tackling the problem of fraud

7. Our 2006 report International benchmark of fraud and error in social security systems[Footnote 2] considered the Department to be at the forefront of measuring welfare loss through fraud (Figure 3). We reported that whilst the United Kingdom has levels of social security fraud and error similar to those in comparable countries, which face many of the same problems, the Department for Work and Pensions has a better understanding than agencies in other nations of these problems, and is doing more to tackle them. We suggested that, although there was scope for the Department to learn from initiatives in some other countries, it was also very well placed to take the lead in encouraging the sharing of lessons abroad. Since that report, the Department has shared its experiences of tackling fraud with a number of interested countries.

Despite good progress, the Department’s efforts could be reinforced through improvements to current arrangements

8. Having reduced estimated levels of fraud to 0.6 per cent of benefit expenditure, it is likely that the Department will find it increasingly difficult to secure significant further year on year reductions. However, a number of aspects of current arrangements could be improved. In particular:
• management information needs improving to enable the Department to fully understand the impact of counter-fraud activity and its cost effectiveness;
• the systems of targets used between, for example, Fraud Investigation Service and Prosecutions Division to drive performance in counter-fraud work currently appear to lack cohesion. When setting targets, the Department should take account of the desired outcome of its counter-fraud work as a whole from investigating particular types of cases through to securing sanctions, with an increased focus on the outcome rather than outputs;
• since the introduction of Customer Compliance in April 2006, cost effectiveness indicators have not yet shown an increase in the value of recoverable overpayments identified by counter-fraud activity. However, the value of identified overpayments does not take into account that earlier activity introduced through Customer Compliance prevents the fraud from continuing. Although still relatively new, from January to March 2007, Customer Compliance performance improved and early indications from 2007-08 suggest that improvements are continuing;
• compliance work is delivered at District level through local Jobcentre Plus offices, so that links can be made with those delivering benefits, but it remains difficult to assess the effectiveness of Customer Compliance work as, although there is a national champion, there is limited coordination between Districts. There is also little general sense of a counter-fraud community, with staff working on the different interventions often working independently with different objectives;
• there is a need to improve the efficiency of processes so that cases flow more easily from one stage to the next. Currently, the progress of cases can be delayed when input is required from parts of the Department not primarily concerned with fraud, which in some circumstances can result in a case not being taken to court; and
• recovering debt is an essential part of tackling fraud and needs to be integrated more into the Department’s counter-fraud work. Social Security Regulations limit the weekly amount recoverable and the level of deductions from benefit will depend on whether customers have other debt to repay and whether they have sufficient funds to live on. It may therefore take many years to recover fraudulent overpayments whilst customers are still in receipt of benefit, which may reduce the deterrent effect to commit further fraud.

The cost effectiveness of the different counter-fraud interventions vary considerably

9. We examined six major counter-fraud interventions in detail, using workshops with officials to examine the resources used and outcomes achieved for each intervention. The interventions are described in detail in Appendix 4. We estimate that in 2006-07 the Department spent around £154 million on these six interventions. However, information on the costs of countering fraud is not complete, making it difficult for the Department to compare the relative costs and benefits of its activities. In part, this is due to the complex interactions amongst its activities, but is exacerbated by avoidable weaknesses including:
• a lack of a consistent measure of effectiveness across its counter-fraud activities to enable quick comparisons or to determine the real effectiveness of its activities; and
• the difficulties in establishing accurate and complete figures for Customer Compliance. Whilst some basic information on budgets and staff numbers is available, better use could be made of the information in order to assess its cost effectiveness.
10. In the absence of the Department’s own assessment of cost effectiveness we have used all available data to establish a baseline for further work. The value of the deterrent effect of the Department’s activities, which could be substantial, is not captured in the analysis as no adequate measure for this exists. In order to compare interventions we used the value of identified overpayments, although this is not a reflection of how much money has actually been recovered by the Department or has been prevented from being paid incorrectly any longer. Our analysis found that:
• based on estimated overpayments at the outset of a referral, those referrals generated through the Data Matching Service are more cost effective than those generated by the National Benefit Fraud Hotline (£24 compared with £16 overpayments identified per £1 spent);
• data matching referrals to local authorities appear most cost effective (£27 per £1 spent);
• Customer Compliance identified during its first year £0.62 of overpayments for every £1 spent. Based on recoverable overpayments identified alone, the introduction of this new approach does not yet appear to be cost effective. To August 2007, the data suggests that for the financial year 2007-08 Customer Compliance is performing better;
• the overall cost of investigation and compliance activity has reduced by 18 per cent, due in part to the redeployment of 1,400 staff from early checking of benefit claims to benefit processing centres, with the aim of improving fraud prevention;
• overall, the value of overpayments identified has dropped by 26 per cent, from £144 million in 2005-06 to £106 million in 2006-07. A drop in recoverable overpayments, however, would be expected as the intention of Customer Compliance is to intervene in a case earlier and correct benefit payments, thereby preventing further overpayments being made;
• the value of identified overpayments does not account for the fact that earlier activity introduced through Customer Compliance prevents the fraud from continuing. Reducing the benefit in payment as a result of this activity stopped further incorrect benefit payments of £5.25 million per week, preventing a further £0.23 of overpayments for every £1 spent on Customer Compliance activity;
• cases progressed through fraud investigation cost six times more than compliance cases (£496 compared with £77), reflecting the complexity and staff intensive nature of investigation activity compared to compliance; and
• it is not possible to determine the deterrent effect of the advertising campaign, but increased calls to the Hotline as an indirect result of the campaigns identified an estimated £900,000 of overpaid benefits.
11. A major deterrent effect for potential fraudsters is the knowledge that the benefit will have to be paid back and the Department seeks to recover overpaid benefit through its 10 debt centres. For those people who owe money but are no longer on benefit, the Department seeks to agree instalments for recovery and where necessary may take court action or refer the debt to a private debt collection company. For those still getting means-tested benefit, rates of recovery are subject to Social Security Regulations. The recovery of fraud overpayments can vary between a maximum of £12 and minimum of £3 per week depending on whether other deductions are in force. Consequently, actual amounts recovered as a result of counter-fraud work may be a small proportion of the overpayment in that year and recovery may therefore be slow.
12. The Department reports that the number of overpayments due to fraud and error being referred to debt centres has increased from 580,000 between April and November 2006 to 825,000 in the same period for 2007. As a result, recoveries due to both fraud and error for this period increased by 30 per cent from £144 million to £184 million. In 2006-07 the Department recovered £22 million of fraud debt out of the total £339 million fraud debt stock, compared to new fraud debt of £45 million.

Value for money statement

Preventing and detecting benefit fraud is a high priority for the Department. Progress since 2003 has been good and represents a substantial achievement by the many dedicated staff within the Department. Estimated levels of fraud have gone down significantly since our 2003 report, from £2 billion to £800 million (approximately 0.6 per cent of benefit expenditure) although definitional changes have helped. However, the monetary loss is still a drain on public resources and the Department’s accounts have been qualified for 18 years as a result of the losses due to fraud and error.

It is inherently difficult to measure the deterrent effect of individual counter-fraud interventions because they form part of a whole strategy. The Department measures the full impact of its counter fraud work through the monetary value of fraud and error, but it is impossible to quantify the impact of deterring people from making illegitimate claims.

One relevant measure is to look at the level of overpayments identified as a result of the Department’s counter-fraud activities. Based on our examination of six of its largest counter-fraud activities, we estimate that the Department identified £106 million of overpayments in 2006-07 at a cost of £154 million. This suggests the Department is currently spending £1.50 to identify every £1 of overpayments. This assessment does not take account of the deterrent effect or the impact of stopping the fraud continuing, both of which are likely to be quite substantial. For example, the earlier intervention introduced through the Customer Compliance programme stops benefit fraud from continuing. Reducing the benefit in payment as a result of this activity has stopped further incorrect benefit payments of some £5.25 million per week.

Of the £106 million of fraud overpayment identified, £45 million was proven or admitted during 2006-07. £22 million of fraud debt was collected in 2006-07. However, this represents a relatively small amount of the £339 million of known fraud debt outstanding which the Department is continuing to pursue to the extent permitted by regulations.

The Department’s activities are proving effective in reducing fraud and should continue. Reaching a final judgement on their cost-effectiveness will always be difficult given some of the factors set out above. However if, as we recommend, the Department can improve its ability to compare the return from different elements of its current fraud strategy, it will be in a better position to target its limited resources at the most effective measures to counter-fraud.

Recommendations

In taking forward its counter-fraud work in the future, the Department should:

Improve its understanding of effectiveness

Issue: The Department’s management information on fraud could be improved, which would in turn enhance its ability to assess the cost effectiveness of the approaches it uses to detect and prevent fraud. To improve its understanding of the performance of its activities and to inform its strategy, the Department needs to:

1. Develop a stronger set of management information that would enable it to assess the ongoing effectiveness of fraud and error reduction strategies (paragraph 3.8)
2. Review the cost effectiveness of the Customer Compliance approach following completion of its first full year of operation. Thereafter, regularly review the costs, performance and effectiveness of Customer Compliance in order to ensure that this approach is contributing sufficiently to reducing fraud given its costs (paragraph 3.19)
3. Introduce a time-recording system in the Prosecution Division, as used in private law firms, in order to cost different types of cases (paragraph 3.9)
4. Record the outcomes of prosecution activities by case type to enable the Department to make informed decisions about the cases it should seek to prosecute (paragraph 3.12).

Improve its use of resources

Issue: A number of different directorates within the Department are responsible for counter-fraud work, each having their own process- focused targets. These do not always support each other. This means there is insufficient oversight of the entire end-to-end management of cases of fraud.

5. The Department should align its resources across different interventions to ensure the smooth flow of cases from one stage to another which should help prevent delays (paragraphs 2.22-2.24)
6. The Department should also review those targets which impact negatively on other interventions, for example, those encouraging investigation of cases that make it harder for another part of the Department to secure a positive prosecution and achieve its target (paragraph 2.29).

Issue: Although there are a number of process-related targets used to assess the effectiveness of the interventions, there is no central monitoring of these to ensure that they are realistic, challenging and align resources with strategic priorities. The Fraud and Error Strategy Division is not responsible for counter-fraud performance, but is best placed to have overall responsibility for the Department’s counter-fraud targets. The Department should:

7. Give overall responsibility for all counter-fraud targets to the Fraud and Error Strategy Division, with clear lines of progress reporting (paragraphs 2.17)
8. Appoint a national head of Customer Compliance to lead activity across Jobcentre Plus, with responsibility for oversight of performance to ensure that Customer Compliance activity remains a priority for Jobcentre Plus managers in the regions (paragraphs 2.9)
9. Provide dedicated decision makers for fraud cases, in order to prevent undue delays where processing new claims takes priority over assessing the value of overpayments as a result of fraud (paragraph 2.22).

Issue: The Department’s counter-fraud interventions are closely inter-connected and the outputs of one might provide inputs to another. As a result it is important to have strong links and systematic feedback throughout the process, especially where the success of one intervention may reduce or increase that of another. At present, there is insufficient feedback and learning between interventions.

10. The Department must take practical steps to create strong links and feedback loops between interventions that are directly or closely inter-connected to improve learning across and within interventions (paragraph 2.20).

Improve the deterrent effect of the Department’s work

Issue: The Department currently uses the value of overpayments identified as an indicator of the success of some of its counter-fraud work. This is not a reflection of how much money is recovered, which could be as low as 25 per cent. The overall deterrent effect of the Department’s counter-fraud work will ultimately be reduced unless money is recovered.

11. The Department must review recovery of overpayments in fraud cases and consider setting appropriate targets for recovery from customers who have committed fraud and have come off benefit (paragraphs 3.4-3.5)

Issue: Since the introduction of the customer compliance model, more cases of suspected fraud are being reviewed by the Department. However, there is a risk that as this model becomes established, and more lower level frauds are dealt with this way, the Department’s hard message of prosecuting fraudsters may be eroded.

12. The Department should communicate with claimants that cases dealt with through Customer Compliance are serious and may be subject to a fraud investigation should further evidence come to light or should the fraud be repeated (paragraph 2.11).

Assessment of implementation of recommendations

To enable prioritisation of these recommendations and allocation of the Department’s limited resources, we have set out below our assessment of the impact these recommendations might have. In the table we have assessed the likely costs of implementing the recommendations and the length of time we would expect it to take the Department to implement. We have also provided an indication of how soon improvements might be seen once the recommendation is implemented. This is the National Audit Office’s assessment, and does not take into account issues of capacity that the Department faces or the availability of resources.

1. [back from footnote 1] The Departments evidence supporting the fraud and error estimate is based on a range of reviews and exercises of the different benefits. As part of our audit of the Departments accounts, including the Departments estimate of fraud and error, we reviewed the Departments approach to sampling and the methodology applied in checking, including re-performance of a sample of items. The National Audit Office also reviewed the processes for interpreting the data and generating estimates of error and fraud. As a result, we are satisfied that the figures reported by the Department are the best estimates available.
2. [back from footnote 2] Comptroller and Auditor General, HC 1387, 2005-06.