Ordered by The Public Accounts Committee to be printed 10 April 2008

Report: 26/07/08R Public Accounts Committee

This document is available in a range of alternative formats.
For more information please contact the
Northern Ireland Assembly, Printed Paper Office,
Parliament Buildings, Stormont, Belfast, BT4 3XX
Tel: 028 9052 1078

Public Accounts Committee
Membership and Powers

The Public Accounts Committee is a Standing Committee established in accordance with Standing Orders under Section 60(3) of the Northern Ireland Act 1998. It is the statutory function of the Public Accounts Committee to consider the accounts and reports of the Comptroller and Auditor General laid before the Assembly.

The Public Accounts Committee is appointed under Assembly Standing Order No. 51 of the Standing Orders for the Northern Ireland Assembly. It has the power to send for persons, papers and records and to report from time to time. Neither the Chairperson nor Deputy Chairperson of the Committee shall be a member of the same political party as the Minister of Finance and Personnel or of any junior minister appointed to the Department of Finance and Personnel.

The Committee has 11 members including a Chairperson and Deputy Chairperson and a quorum of 5.

The membership of the Committee since 9 May 2007 has been as follows:

Mr John O’Dowd (Chairperson)
Mr Roy Beggs (Deputy Chairperson)

Mr Thomas Burns** Mr Trevor Lunn
Mr Jonathan Craig Mr Ian McCrea*
Mr John Dallat Mr Mitchel McLaughlin
Mr Simon Hamilton Ms Dawn Purvis
Mr David Hilditch

* Mr Mickey Brady replaced Mr Willie Clarke on 1 October 2007
* Mr Ian McCrea replaced Mr Mickey Brady on 21 January 2008
** Mr Thomas Burns replaced Mr Patsy McGlone on 4 March 2008

Table of Contents

List of abbreviations used in the Report

Report

Executive Summary

Summary of Recommendations

Introduction

Meaningful Information on Levels of Fraud and Error

Maximising the Uptake of Benefit Entitlements

Need to Improve Staff Skills and Training

Weaknesses in Management Performance and Accountability

Fraud Identification and Investigation

Interventions to Maximise the Use of Scarce Resources

Appendix 1:

Minutes of Proceedings

Appendix 2:

Minutes of Evidence

Appendix 3:

Chairperson’s letter of 26 February 2008 to Mr Alan Shannon, Accounting Officer,
Department for Social Development

Correspondence of 7 March 2008 from Mr Alan Shannon, Accounting Officer,
Department for Social Development

Appendix 4:

List of Witnesses

List of Abbreviations used in the Report

The Agency Northern Ireland Social Security Agency
The Department Department for Social Development
DLA Disability Living Allowance
IT Information Technology
DFP Department of Finance and Personnel
C&AG Comptroller and Auditor General
HM Treasury Her Majesty’s Treasury

Executive Summary

Introduction

1. The Northern Ireland Social Security Agency (the Agency), on behalf of the Department for Social Development (the Department), is responsible for the accurate disbursement of substantial sums of public money, some £3.5 billion in 2006-07, through the social security benefit system. The benefit system is inherently complex and the Committee recognises the good work being done by Agency staff in delivering services to customers, often under difficult and pressurised circumstances.

2. While the vast majority of benefits are correctly paid to customers, the Agency estimates that some £60 million was overpaid and £22 million underpaid to customers in 2006-07. The Committee considers that, despite improvements, this level of inaccuracy remains disappointingly high. It is particularly concerned that errors by Agency staff resulted in £28 million of overpayments and £19 million of underpayments. The Committee is also concerned that fraud and error in some benefits may have a detrimental knock-on effect on other entitlements, for example, housing benefits, health service charges and school meals.

3. The Department made the point that the Agency will struggle to maintain, let alone improve on, the fraud and error figures for 2006-07 given the scale and complexity of the benefit system, changes in benefits and reductions in resources. This is unacceptable. While the Committee acknowledges the challenges being faced by this Department and others, this report shows that there is considerable scope to make progress in a number of areas. These include, improved staff training and support; increased focus on getting the amount of benefit in payment right from the start; improved targeting of programme protection activity and greater accuracy in this work; effectiveness of fraud investigation; and a focus on cost-effective interventions to maximise the use and impact of scarce resources.

4. The Committee expects tackling benefit fraud and error to remain a key priority for the Department and Agency and that together they address the problem with vigour, so maintaining public confidence in the benefit system.

Meaningful Information on Levels of Fraud and Error

5. Public accountability requires full and meaningful reporting of the total levels of incorrect benefit payments and the actions being taken to address them. Given the complexities in measuring and interpreting the Agency’s estimates, it is essential that published data is consistent and comparable over time. The Committee considers that the total level of estimated inaccuracies in the benefits system should include all estimated overpayments, underpayments and incorrect payments resulting from ‘change of circumstances’ in DLA cases.

Maximising the Uptake of Benefit Entitlements

6. Although the Committee strongly supports well thought through advertising campaigns and other initiatives to deter fraud, it is vitally important that these are designed in a way which does not deter people in need, particularly the elderly, from making genuine benefit applications. The Committee welcomes the Agency’s work with the voluntary sector to actively promote benefit uptake.

Need to Improve Staff Skills and Training

7. While the Committee recognises that the benefit system is complex, it is astonished that almost half of overpayments are caused by Agency staff error and that the Agency’s programme protection activity, aimed at checking and validating benefits in payment, allowed errors to continue and, in some cases, incorrectly changed benefit entitlements.

8. The Committee welcomes the Agency’s initiatives, in response to the Audit Office findings, to strengthen its staff training programmes and introduce compulsory fraud awareness training for all staff. The Agency must also ensure that programme protection staff are better trained, experienced in the benefits under review and work to a much higher degree of accuracy. The Committee looks forward to seeing significant measurable improvements through substantial reductions in staff errors in the years to come.

Weaknesses in Management Performance and Accountability

9. The Committee is concerned at the degree of autonomy that was given to district managers for programme protection activity, which resulted in staff being diverted from this activity. Sound advice and guidance were openly ignored, and targets were not achieved. The Committee welcomes the recent improvements in accountability arrangements and performance against programme protection targets. However it is important that these improvements are sustained.

Fraud Identification and Investigation

10. The Committee believes that fraud investigation is a key weapon in the Agency’s strategy and expects that, where suspicions of fraud are raised, these will be appropriately investigated. However, the Committee is concerned that, prior to 2006-07, the Agency had difficulties recruiting sufficient numbers of fraud investigators which led to substantial backlogs in referrals for investigation and some referrals not being investigated at all.

11. The Committee considers that the business justification for increasing the numbers of investigators must be seriously examined. This is particularly important in the context of the introduction of the National Fraud Initiative which may generate even greater numbers of fraud referrals.

12. The Committee acknowledges the cross border co-operation on benefit fraud and welcomes the development of data matching protocols with the Department of Social and Family Affairs. This will allow the Agency to better identify incidents and the potential scale of this aspect of fraud. The Agency should explore similar types of co-operation with other public and private sector bodies, for example financial institutions, to better identify and tackle benefit fraud, including organised and complex fraud.

13. Given the substantial scale and widening nature of benefit fraud, the Agency should develop an Action Plan identifying the various types of fraud, its assessment of the associated risks and the action being taken to generate fraud referrals and tackle the various risks. Progress should be reported in the Agency’s annual report.

Interventions to Maximise the Use of Scarce Resources

14. Given that resources are limited, and the Agency indicates that it is likely to face further resource constraints, it is important that it has appropriate information to assess the risks of fraud and error and determine the cost-effectiveness of the various elements of its counter fraud and error strategy. It must also establish appropriate systems to evaluate the cost-effectiveness of different investigation methods and the costs of investigating cases from the range of referral sources.

15. The Committee welcomes the Agency’s increased focus on achieving sanctions and its commitment to developing a broader range of targets. The targets must contribute to the effective management of fraud investigations to ensure these are completed on a timely basis and conducted to the required quality standards. The Committee encourages the Agency to continue to work closely with the Public Prosecution Service to ensure it has a better understanding of the requirements for successful prosecution.

Looking to the Future

There is no doubt that the complexity of the benefit system contributes to the high level of fraud and error. It is the Committee’s view that the outmoded IT systems, which should be there to support Agency staff, exacerbate the problem. Going forward, the Committee expects both the Department and the Agency to take every opportunity to press for simplification of the benefit system and the introduction of improved and integrated IT systems.

Summary of Recommendations

Meaningful Information on Levels of Fraud and Error

1. The Committee recommends that the Agency, when reporting the levels of fraud and error, also reports any changes in methodology and adjusts historical data accordingly to allow for a meaningful interpretation of information and comparison of the levels of fraud and error over time.

2. The Committee recommends that the Agency reports the margins of error associated with its estimates of the levels of fraud and error. The Agency also should provide explanations of the margins of error and any other factors necessary for the proper interpretation and use of the published estimates.

3. The Committee recommends that the Agency reports the full extent of incorrectness in the benefit system, including all overpayments and underpayments of benefit and incorrect payments as a result of ‘change of circumstances’ in DLA cases. It is also important that the Agency reports the actions being taken to address the causes.

4. The Committee recommends that the Agency regularly quality assures the reasonableness and consistency of decisions underpinning its classification of fraud and error.

Maximising the Uptake of Benefit Entitlements

5. The Agency must ensure that its strategy for tackling benefit fraud and error, in particular its efforts at deterring fraud through, for example, its fraud publicity campaign, does not undermine initiatives aimed at maximising the uptake of benefits. The Committee recommends that any future media campaign should be designed in a way which does not deter people in need, particularly the elderly, from making genuine benefit applications.

Need to Improve Staff Skills and Training

6. The Committee welcomes the Agency’s initiatives to strengthen its staff training programme and recommends that staff receive the training and support commensurate with their individual roles and responsibilities. Where errors are being made, learning opportunities, at individual and organisational level, should be followed-up and lessons learned. The Committee looks forward to seeing significant measurable improvements through substantial reductions in staff error in the years to come.

7. The Committee expects programme protection work to be highly effective in detecting and correcting errors in the benefits system. The Committee recommends that the Agency must ensure that staff allocated to the work are highly trained and experienced in the benefits under review and work to a high degree of accuracy. District managers must not have the discretion to divert staff from this important area of work.

8. The Committee recommends that the Agency monitors the levels and quality of fraud referrals made by staff so that it can assess the overall effectiveness of its efforts to improve fraud awareness within the Agency.

9. The Committee recognises that the Department and Agency have important roles to play in addressing benefit fraud and error at the policy and strategic level as well as at an operational level. The Committee recommends that the Department and the Agency proactively participate in national initiatives, in particular those promoting the principle of benefit simplification.

Weaknesses in Management Performance and Accountability

10. The Committee welcomes the improved accountability arrangements that have been put in place for district managers and recommends that the payment of performance bonuses is clearly linked to the achievement of all targets, including programme protection activity and those related to reducing the excessive level of staff absence. It is also important that improvements in performance are sustained. The Committee looks to the Audit Office to check that improvements continue to be made in these areas.

11. While there is a clear need to reduce the high levels of error in the benefit system, it is important that the measures put in place to address them are cost-effective. The Committee considers that this requires a sound basis for assessing the financial impact of programme protection activity. The Committee recommends that the Agency develops a more rational and empirical basis for measuring this.

Fraud Identification and Investigation

12. The Committee recommends that more effort should be made to deter and prevent fraud and error arising in the first place, for example, improved and integrated IT systems, simplification of the benefit system, including the structure of application forms, improvement in the rates of ‘getting it right first time’; more extensive validation of claimant information prior to payment; and the use of risk assessment to focus additional checks on claims.

13. The Committee recommends that the Agency and Department of Finance and Personnel (DFP), along with HM Treasury, consider carefully, on the basis of spend to save, the business justification for increasing the number of investigators. This is particularly important in the context of the introduction of the National Fraud Initiative which may generate greater numbers of fraud referrals.

14. The Committee welcomes the work being done to develop data matching protocols and expects the Agency to keep it informed of developments. The Committee recommends that the Agency explore similar types of co-operation with other public and private sector bodies, for example financial institutions, where a multi-agency approach and intelligence sharing would be crucial to better identify and tackle benefit fraud, for example, organised and complex fraud.

15. The Committee recommends that the Agency develops an Action Plan identifying the various types of fraud affecting the benefit system, including, for example, ‘house drops’, cross border fraud and organised fraud. The Action Plan should provide an assessment of the risks associated with each type of fraud and the action being taken to identify frauds and tackle the various risks, including new or emerging threats. Progress should be reported in the Agency’s annual report.

Interventions to Maximise the Use of Scarce Resources

16. The Committee notes that the Agency has provided information which shows that spending on fraud investigation and programme protection produces a positive financial impact. The Committee recommends that this costing information should be further refined to enable the Agency to evaluate the cost-effectiveness of different investigation methods and the costs of investigating cases from the range of referral sources.

17. The Committee welcomes the Agency’s move to better target programme protection resources to achieve more effective outcomes. The Committee recommends, however, that the Agency approach to resource allocation should be improved through an assessment and publication of the relevant levels of fraud within each district.

18. The Committee recommends that the Agency reviews its current targets for benefit investigation and the underlying performance measurement processes. These should provide for effective management of fraud investigations to ensure they are completed on a timely basis, are conducted to required quality standards and deliver the Agency’s objective to maximise the number of sanctions.

19. The Committee acknowledges the Agency’s use of probability of success ratings to prioritise cases for investigation. However, the Committee recommends that this approach is regularly reviewed by Agency management to ensure that the thresholds for investigation continue to be set at appropriate levels.

20. The Committee recommends that the Agency monitors the reasons for frauds not proceeding to sanction and continues to develop action plans to address these.

21. It is crucial that cases sent for prosecution meet the evidential threshold set for taking them to court. The Committee recommends that the Agency works closely with the Public Prosecution Service to ensure that each body has a more complete understanding of their respective roles and responsibilities and the requirements for successful prosecution.

22. The Committee expects the performance of the Financial Investigation Unit to be kept under review. The Committee recommends that the Agency reports the costs and results of the work of the Financial Investigation Unit in its annual report. This together with other publications and reports, for example, DFP’s Annual Fraud report, should also be used to highlight the positive outcomes of its investigations which, the Committee considers, will further contribute to deterring fraud.

Introduction

1. The Public Accounts Committee met on 21 February 2008 for an evidence session with the Department for Social Development (the Department) and the Northern Ireland Social Security Agency (the Agency). The session related to the Comptroller and Auditor General’s (C&AG’s) report: “Social Security Benefit Fraud and Error”.

2. The witnesses were:

• Mr Alan Shannon, Accounting Officer, Department for Social Development;
• Mr Bryan Davis, Chief Executive, Northern Ireland Social Security Agency;
• Mr John Nevin, Assistant Director of Disability, Incapacity and Security, Northern Ireland Social Security Agency;
• Mr Kieran Donnelly, Deputy Comptroller and Auditor General; and
• Mr Ciaran Doran, Deputy Treasury Officer of Accounts.

3. The Northern Ireland Social Security Agency is responsible for the accurate disbursement of substantial sums of public money, some £3.5 billion in 2006-07, through the social security benefit system. Inevitably, processing such sums through a complex benefit system for a large number of claimants will be prone to losses through fraud and error.

4. The Agency estimates £60 million of benefits was overpaid and £22 million underpaid to customers in 2006-07. In addition to being a drain on public funds, losses, through fraud and error impact on some of the most vulnerable people in society, affect public confidence in the benefit system and public administration generally.

5. The Committee expects the Agency to remain committed to getting the correct amount of benefit into payment first time – neither overpaid nor underpaid. Too much is still being lost through customer fraud and, of particular concern, are the large amounts over and underpaid to claimants, mostly as a result of staff error. Therefore, tackling overpayments arising from fraud and error should continue to be one of the Agency’s key priorities, as must also tackling underpayments caused by both Agency and customer error.

6. The Committee acknowledges that the Agency faces many challenges and recognises the good work by its staff in delivering services to customers, sometimes under difficult and pressurised circumstances. The Committee also considers that the Agency’s strategy ‘Tackling Fraud and Error in Social Security’ and the establishment of a Benefit Security Directorate, provide a good framework for a co-ordinated approach to tackling fraud and error.

7. It is clear that good work has been done but there is much more to do. The Committee believes the overall success of this strategy will largely depend on the Agency actively demonstrating a commitment to getting the right balance between deterrence, prevention, detection and prosecution of benefit fraudsters. It is clearly better to prevent fraud and error entering the system than trying to detect it at a later date and incurring additional costs to attempt recovery of the overpayment.

8. In taking evidence on the C&AG’s report, the Committee focused on six main issues. These were:

• the reporting of levels of fraud and error;
• maximising the uptake of benefit entitlements;
• ensuring staff have appropriate skills and training;
• management performance and accountability;
• how suspected fraud is identified and investigated; and
• maximising the use of resources through cost-effective fraud and error interventions.

Meaningful Information on
Levels of Fraud and Error

9. The loss of any public funds through fraud and error is of particular concern to this Committee. It is important that losses are minimised, not only because they are a drain on the funds available for other public services, but also that public confidence in the disbursement of public funds, and public administration generally, is maintained. Therefore, the Committee expects all public money to be safeguarded against loss of any kind; that any losses are fully and accurately reported; and that the recovery of any loss, particularly as a result of fraud, is robustly pursued.

10. The Northern Ireland Social Security Agency, on behalf of the Department for Social Development, administers the social security benefit system and the payment of substantial sums to claimants – some £3.5 billion was paid out on social security benefits in 2006-07. Of this the Agency estimated that some £60 million (1.7%) was overpaid due to fraud and error, with a further £22 million (0.6%) underpaid to customers, mainly as a result of errors by the Agency.

11. While the Agency is keen to promote the recent downward trend in the levels of benefit fraud and error, as a result of its strategy, the Committee considers that the level remains disappointingly high - £60 million is a significant amount of money.

12. The Committee acknowledges the Agency’s commitment to tackling the problem and recognises the good work done by its staff, sometimes under difficult and pressurised circumstances. However, the levels of benefit fraud and error remain too high. The Committee expects the Agency to continue to tackle benefit fraud and error with vigour and that this will remain a key priority for both the Department and the Agency.

13. The Agency assured the Committee that it has a robust methodology for measuring the levels of benefit fraud and error and recognised the need for consistency in its approach to facilitate comparison of trends over time. It also acknowledged that there were a number of definitional and technical changes, as well as the introduction of new benefits, for example Pension Credit in 2004-05, which impacted on its measurement methodology.

Recommendation 1

14. The Committee recommends that the Agency, when reporting the levels of fraud and error, also reports any changes in methodology and adjusts historical data accordingly to allow for a meaningful interpretation of information and comparison of the levels of fraud and error over time.

15. The Committee considers that reporting adjustments for changes in methodology can only go part of the way to providing the detail necessary for the proper interpretation of the levels of fraud and error. The Agency’s estimates of fraud and error are based on statistical samples, which by their nature are subject to margins of error. The quantification and reporting of these margins of error are necessary for the proper interpretation of data. This is not currently done.

Recommendation 2

16. The Committee recommends that the Agency reports the margins of error associated with its estimates of the levels of fraud and error. The Agency also should provide explanations of the margins of error and any other factors necessary for the proper interpretation and use of the published estimates.

17. The Agency’s business is to provide financial assistance to those who are most in need and overpayments and underpayments due to fraud and error impact on vulnerable people. The Committee considers the exclusion of all underpayments and the amounts paid in excess, and amounts underpaid, due to ‘changes in customer circumstances’ [1] in Disability Living Allowance (DLA), are material to providing a true picture of what is happening in the benefits system. Their exclusion has resulted in significant decreases in the reported levels of fraud and error. While the Committee notes the reasons given by the Agency for their exclusion, the Committee considers that accountability for public funds requires that the totality of incorrect payments (overpayments and underpayments, and DLA ‘change of circumstances’) is disclosed.

Recommendation 3

18. The Committee recommends that the Agency should report the full extent of incorrectness in the benefit system, including all overpayments and underpayments of benefit and incorrect payments as a result of ‘change of circumstances’ in DLA cases. It is also important that the Agency reports the actions being taken to address the causes.

19. While the Agency assured the Committee that the figures for fraud and error were credible, the Committee considers that there is an element of subjectivity in determining whether an overpayment is classified as fraud or error. The Committee remains to be convinced that the decreases in customer fraud in recent years are not, in part, related to reclassifications which have seen increases in customer error over the same period.

Recommendation 4

20. The Committee recommends that the Agency regularly quality assures the reasonableness and consistency of decisions underpinning its classification of fraud and error.

Maximising the
Uptake of Benefit Entitlements

21. Alongside the reduction in benefit fraud and error, the Agency still has much work to do to ensure that those in need take up the benefits for which they are entitled. The Committee welcomes the Agency’s inclusion of both these objectives as key priorities.

22. The Agency informed the Committee that it is actively promoting benefit uptake and is currently targeting 200,000 people, in co-operation with the voluntary sector, in a bid to minimise the amount that goes unclaimed. This is a welcome development. However, while a balance has to be struck between benefit uptake and fraud publicity campaigns, the Committee would be concerned if the anti-fraud campaign inadvertently put off people from claiming benefits to which they are properly entitled.

Recommendation 5

23. The Agency must ensure that its strategy for tackling benefit fraud and error, in particular its efforts at deterring fraud through, for example, its fraud publicity campaign, does not undermine initiatives aimed at maximising the uptake of benefits. The Committee recommends that any future media campaign should be designed in a way which does not deter people in need, particularly the elderly, from making genuine benefit applications.

Need to Improve
Staff Skills and Training

24. The Committee is astonished that of the estimated £60 million overpaid in 2006-07, almost a half, £28 million, is due to errors made by Agency staff. In addition, a further £19 million of benefit underpayments were the result of staff error. The levels of staff error are too high and must be reduced. The Committee recognises that the benefit system is complex. However the quality of benefit assessments and administration is directly under the Agency’s control and the Committee expects that benefit operations staff receive the training and support commensurate to the challenges they face.

25. The Agency recognises that training is a very important element of its approach to tackling benefit fraud and error and has recently reviewed its training processes, setting out a more structured learning programme. Current developments include the piloting of a new assessment tool to measure, more effectively, how well staff assimilate training and put this into practice. Benchmark targets have also been established to assess the accuracy of processing claims by district office staff so that individual training needs can be identified. However, the Agency acknowledged that more needs to be done in this area.

Recommendation 6

26. The Committee welcomes the Agency’s initiatives to strengthen its staff training programme and recommends that staff receive training and support commensurate with their individual roles and responsibilities. Where errors are being made, learning opportunities, at the individual and organisational level, should be followed-up and lessons learned. The Committee looks forward to seeing significant measurable improvements through substantial reductions in staff error in the years to come.

27. Internal controls are a widely used management procedure to ensure that processes are working as intended and, if not, to put things right. The Committee is therefore concerned at the level of error found in the Agency’s own sample check on programme protection activity. This identified errors which had not been detected and in some cases resulted in benefit entitlements being incorrectly changed. This calls into question the validity of financial savings claimed for programme protection activity.

Recommendation 7

28. The Committee expects programme protection work to be highly effective in detecting and correcting errors in the benefits system. The Committee recommends that the Agency must ensure that staff allocated to the work are highly trained and experienced in the benefits under review and work to a high degree of accuracy. District managers must not have the discretion to divert staff from this important area of work.

29. The Committee is concerned with the Audit Office survey findings that 25% of staff surveyed had not received any fraud awareness training or briefings, and this figure increased to 50% within the Disability and Carers group of staff. The Agency informed the Committee that it was surprised and concerned at the findings and had introduced compulsory fraud awareness training for all staff recruited in 2007, as well as including fraud awareness in the compulsory annual awareness training. The Committee welcomes the Agency’s positive response to the Audit Office findings.

Recommendation 8

30. The Committee recommends that the Agency monitors the levels and quality of fraud referrals made by staff so that it can assess the overall effectiveness of its efforts to improve fraud awareness within the Agency.

31. The Agency informed the Committee that training alone would not be sufficient to tackle benefit fraud and error and that a range of measures, including modern IT systems and benefit simplification, are required. The Committee was told that there were thousands of pages of advice and guidance on social security benefits and creating a benefits system to match individual circumstances necessitated this level of complexity. The Department informed the Committee that it contributed to a network of committees in London involved in developing social security policy and benefit simplification.

Recommendation 9

32. The Committee recognises that the Department and Agency have important roles to play in addressing benefit fraud and error at the policy and strategic level as well as at an operational level. The Committee recommends that the Department and Agency proactively participate in national initiatives, in particular those promoting the principle of benefit simplification.

Weaknesses in Management
Performance and Accountability

33. The Committee was alarmed at the wide variation in the districts’ performance against programme protection targets and that managers were, on occasions, ignoring the sound advice and guidance given to them by the programme protection unit. While the Agency acknowledged that ignoring the advice was unacceptable, it pointed out that managers had to balance a range of priorities at any one time. The Committee is, however, concerned that staff were taken off programme protection activity which, based on the reported figures, is highly cost-effective and district managers had been, to all intents and purposes, a law unto themselves.

34. The Committee considers that underlying management issues have contributed to the weaknesses in districts’ performance. The Agency informed the Committee that accountability arrangements have been tightened up; the programme protection unit’s interaction with district managers was now more effective and areas of concern about failure to meet targets are raised at senior management meetings and passed up to the Agency Management Board.

Recommendation 10

35. The Committee welcomes the improved accountability arrangements that have been put in place for district managers and recommends that the payment of performance bonuses is clearly linked to the achievement of all targets, including programme protection activity and those related to reducing the excessive levels of staff absence. It is also important that improvements in performance are sustained. The Committee looks to the Audit Office to check that improvements continue to be made in these areas.

Recommendation 11

36. While there is a clear need to reduce the high levels of error in the benefit system, it is important that the measures put in place to address them are cost-effective. The Committee considers that this requires a sound basis for assessing the financial impact of programme protection activity. The Committee recommends that the Agency needs to develop a more rational and empirical basis for measuring this.

Fraud Identification and Investigation

37. The Agency informed the Committee that its strategy for tackling fraud and error centred on preventing it entering the system; detecting if it is in the system; and correcting it when this occurs. It is clear to the Committee that the overall success of the strategy will largely depend on the Agency getting the right balance between these strands. However, the Committee considers that stopping fraud and error entering the system through, for example, checking for undeclared assets on receipt of a claim, must be preferable to trying to detect it later when benefit is in payment and having to incur additional costs to attempt recovery of overpayments.

Recommendation 12

38. The Committee recommends that more effort should be made to deter and prevent fraud and error arising in the first place, for example, improved and integrated IT systems, simplification of the benefit system, including the structure of application forms; improvement in the rates of ‘getting it right first time’; more extensive validation of claimant information prior to payment; and the use of risk assessment to focus additional checks on claims.

39. The Committee is concerned that a key strand of tackling benefit fraud, the investigation of fraud referrals, was constrained by a lack of sufficient investigators over so many years. The Committee notes the Agency’s difficulties in recruiting sufficient fraud investigators prior to 2006-07 which led to substantial backlogs in referrals for fraud investigation and indeed to some referrals not being investigated at all. The Agency told the Committee that, prior to the Fraud Act 2006, investigators were required to undertake surveillance in often challenging environments which made it unattractive work and presented recruitment difficulties. Since the introduction of the Fraud Act working practices have become more desk based and focused on data interrogation, and more attractive to staff; this has enabled the Agency to reach, what it considers to be, its full complement of investigators.

40. The Committee welcomes the Agency’s confirmation that the volume of referrals received each year is now being effectively managed. The Agency also informed the Committee that in 2006-07 there was more that a 4 to 1 return from investigations, with £10 million overpayments identified compared with £2.3 million spent on employing investigators.

41. While the Committee notes that the Agency has reached its complement of investigators in recent years, it is not clear how the current ceiling on numbers has been determined. Of further concern, is the potential impact of resource constraints on the Agency. Based on the evidence, the Committee believes that fraud investigation is a key weapon in the Agency’s armoury in combating fraud. The Committee recognises that public funds are not unlimited. But there is a danger in a broad brush approach to efficiency savings that becomes ‘penny wise and pound foolish’.

Recommendation 13

42. The Committee recommends that the Agency and DFP, along with HM Treasury, consider carefully, on the basis of spend to save, the business justification for increasing the number of investigators. This is particularly important in the context of the introduction of the National Fraud Initiative which may generate greater numbers of fraud referrals.

43. The Committee acknowledges the cross border co-operation on benefit fraud between the Agency, the Department for Work and Pensions in Great Britain and the Department of Social and Family Affairs in the Republic of Ireland. While the Agency believes that cross border fraud is not currently a major problem, the Committee considers that, in the absence of data matching between the relevant bodies, it is difficult to estimate the scale of the problem.

44. The Agency informed the Committee that it was working with the Department for Work and Pensions and the Department of Social and Family Affairs to develop data matching protocols that will enable regular matching of records, and provide a means of identifying cross border fraud. It also confirmed that it was continuing to develop the general matching service [2], which produced high quality fraud referrals, and that it was working closely with the Northern Ireland Audit Office on the National Fraud Initiative.

Recommendation 14

45. The Committee welcomes the work being done to develop data matching protocols and expects the Agency to keep it informed of developments. The Committee recommends that the Agency explore similar types of co-operation with other public and private sector bodies where a multi-agency approach and intelligence sharing would be crucial to better identify and tackle benefit fraud, for example, organised and complex fraud.

46. The Agency informed the Committee of a range of initiatives aimed at tackling specific types of fraud. Given the substantial scale and wide-ranging nature of benefit fraud, there is a need for strategic management of the challenges being faced and the types of responses required.

Recommendation 15

47. The Committee recommends that the Agency develops an Action Plan identifying the various types of fraud affecting the benefit system, including, for example, ‘house drops’, cross border fraud, organised fraud. The Action Plan should provide an assessment of the risks associated with each type of fraud and the action being taken to identify frauds and tackle the various risks, including new or emerging threats. Progress should be reported in the Agency’s annual report.

Interventions to
Maximise the Use of Scarce Resources

48. The Agency told the Committee that the annual running cost of its counter fraud and error strategy is some £12.5 million, comprising programme protection £5 million, Benefit Investigation Services £5.3 million and policy and assurance group £2.2 million. However, it does not appear that the Agency analyses the cost-effectiveness of different investigation methods or cost of investigating cases from the various sources and types of referral.

49. The Agency also indicated that it is likely to face further constraints on its resources. The Committee acknowledges that there are challenges in reducing the levels of fraud and error. However, these should not diminish expectations of further improvement. The figures provided by the Agency show clearly that every pound spent tackling fraud and error does make a difference and results in a positive return on this investment. In the circumstances where resources may be further constrained, it is particularly important that the Agency has the appropriate information to demonstrate that it is making the most effective use of the resources available to it.

Recommendation 16

50. The Committee notes that the Agency has provided information which shows that spending on fraud investigation and programme protection produces a positive financial impact. The Committee recommends that this costing information should be further refined to enable the Agency to evaluate the cost-effectiveness of different investigation methods and the costs of investigating cases from the range of referral sources.

51. The Committee welcomes the Agency’s assurance that it is moving to a new and more developed risk based assessment that gives a better approach to allocating programme protection resources and targeting risks. While the Agency has stated that it is not unexpected that there will be little variation in risks between districts as each district is administering the same benefit, the Committee considers it is not good enough that it cannot provide a breakdown of fraud by district.

52. Targeting and allocation of resources needs to be based on an appropriate assessment of risk and to those districts that will achieve the most effective outcomes. It is also important that this rationale is applied to the allocation of programme protection resources across the range of benefit branches within the Agency.

Recommendation 17

53. The Committee welcomes the Agency’s move to better target programme protection resources to achieve more effective outcomes. The Committee recommends, however, that the Agency’s approach to resource allocation should be improved through assessment and publication of the relevant levels of fraud within each district.

54. The Agency told the Committee that it has, initially, concentrated on a single sanctions target and that this has been effective. The Committee welcomes the Agency’s increased focus on achieving sanctions in cases where fraud has been committed and its intention to now examine a broader range of targets to support the sanctions target.

Recommendation 18

55. The Committee recommends that the Agency reviews its current targets for benefit investigation and the underlying performance measurement processes. These should provide for effective management of fraud investigations to ensure they are completed on a timely basis, are conducted to required quality standards and deliver the Agency’s objective to maximise the number of sanctions.

56. The Committee was very concerned that the Agency closed approximately 3,000 older suspected fraud cases without adequate investigation or follow-up. There is a danger that this approach may send out the wrong signal to fraudsters, the wider public and the Agency’s own staff, almost half of whom indicated that they were not confident that suspected fraud is being investigated fully.

Recommendation 19

57. The Committee acknowledges the Agency’s use of probability of success ratings to prioritise cases for investigation. However, the Committee recommends that this approach is regularly reviewed by Agency management to ensure that the thresholds for investigation continue to be set at appropriate levels.

58. Between 2003 and 2007, a further 3,000 cases, which were deemed to involve fraud of some kind, were considered non-sanctionable by the Agency. This is a matter of deep concern. In the Committee’s view the problem of public attitude towards benefit fraud is not helped by the perception that those who perpetrate fraud are not sanctioned in any way.

59. The Agency told the Committee that 80% of these cases were closed in line with its sanctions policy. However, in almost 600 cases, the Agency was at fault in that half of those cases were due to delay in the system, mostly relating to the period when there was a shortage of investigators. The other element related to a failure in the investigation which reflected a lack of experience within the Agency.

60. The Committee welcomes the Agency’s candour in acknowledging its fault and failure in these areas and also notes the steps being taken to address these problems. The Committee considers that, given the time, energy and money that are put into investigation, it is important that investigations are sound and can be progressed to prosecution or other sanction, in the maximum number of cases. While there has been an increase in the number of sanction cases in recent years, evidence of improvement will be seen in a sustained increase in the number of sanctions achieved and in the proportion of fraud cases proceeding to sanction.

Recommendation 20

61. The Committee recommends that the Agency monitors the reasons for frauds not proceeding to sanction and continues to develop action plans to address these.

62. The Committee recognises and welcomes the fact that 97% of cases taken to court are successful. However, between 2003 and 2007, only some 800 cases out of 1,400 cases that the Agency passed to the Public Prosecution Service have actually gone to court. In approximately 250 cases the Public Prosecution Service withdrew the case due to the insufficient strength of available evidence.

63. The Agency stressed that a judgement call must be made on whether the evidence is strong enough to sustain a prosecution and that the Public Prosecution Service sets a high evidential threshold for taking cases to court. The Committee welcomes the steps being taken by the Agency to familiarise itself more with the work of the Public Prosecution Service and to implement a service level agreement.

Recommendation 21

64. It is crucial that cases sent for prosecution meet the evidential threshold set for taking them to court. The Committee recommends that the Agency works closely with the Public Prosecution Service to ensure that each body has a more complete understanding of their respective roles and responsibilities and the requirements for successful prosecution.

65. The Committee is concerned that recovery of overpayments resulting from fraud can take many years to recover. The Committee notes that the Agency is limited, by legislation, in the amount of money that can be recovered at any one time. However, this can send out a message that it is worth the risk to commit the fraud.

66. Prosecution of offenders is also crucial to deterring fraudsters and showing that those who commit fraud will be penalised. The Committee recognises that the Agency has no remit for sentencing policy. The Committee welcomes, however, the Agency’s engagement with officials from the Northern Ireland Court Service to ensure that they are aware of the Agency’s approach to tackling fraud and the impact of sentencing on its strategy.

67. The Financial Investigation Unit was set up by the Agency in 2005 to undertake financial investigation and confiscate assets as a step to civil recovery. The Unit costs £200,000 each year to run and to date has recovered £500,000. The Committee notes that it has taken until October 2007 for the financial investigation staff to be accredited. The Unit has however, finalised 14 cases, 21 are in the judicial process and nine are with the Assets Recovery Agency [3]. The Agency acknowledged that it is possible to measure the cost of activity and some of the outcomes, but it is more difficult to measure the Unit’s deterrent effect. However, it is important that the work of the Financial Investigation Unit is monitored and evaluated.

Recommendation 22

68. The Committee expects the performance of the Financial Investigation Unit to be kept under review. The Committee recommends that the Agency reports the costs and results of the work of the Financial Investigation Unit in its annual report. This, together with other publications and reports, for example, DFP’s Annual Fraud report, should also be used to highlight the positive outcomes of its investigations which, the Committee considers, will further contribute to deterring fraud.

Appendix 1

Minutes of Proceedings
of the Committee Relating
to the Report

Thursday, 21 February 2008
Senate Chamber, Parliament Buildings

Present: Mr John O’Dowd (Chairperson)
Mr Roy Beggs (Deputy Chairperson)
Mr Jonathan Craig
Mr John Dallat
Mr Simon Hamilton
Mr David Hilditch
Mr Trevor Lunn
Mr Ian McCrea
Mr Mitchel McLaughlin
Ms Dawn Purvis

In Attendance: Mr Jim Beatty (Assembly Clerk)
Mrs Gillian Lewis (Assistant Assembly Clerk)
Mrs Nicola Shephard (Clerical Supervisor)
Mr Ricky Shek (Clerical Officer)

Apologies: Mr Patsy McGlone

The meeting opened at 2.00pm in public session.

2.01pm Ms Purvis joined the meeting.
2.02pm Mr Craig joined the meeting.

3. The Chairperson welcomed Mr Kieran Donnelly, Deputy Comptroller & Auditor General (Deputy C&AG) and Mr Ciaran Doran, Deputy Treasury Officer of Accounts (Deputy TOA), to the meeting.
4. Evidence on the NIAO Report ‘Social Security Benefit Fraud and Error’.

The Committee took oral evidence on the NIAO report ‘Social Security Benefit Fraud and Error’ from Mr Alan Shannon, Accounting Officer, Department for Social Development, Mr Bryan Davis, Chief Executive, Social Security Agency (SSA), and Mr John Nevin, Assistant Director of Disability, Incapacity and Security, SSA.

The witnesses answered a number of questions put by the Committee.

Members requested that the witnesses should provide additional information to the Clerk on some issues raised during the evidence session.

4.10pm The evidence session finished.

[EXTRACT]

Thursday, 10 April 2008
Room 144, Parliament Buildings

Present: Mr John O’Dowd (Chairperson)
Mr Roy Beggs (Deputy Chairperson)
Mr Jonathan Craig
Mr John Dallat
Mr Simon Hamilton
Mr David Hilditch
Mr Trevor Lunn
Mr Ian McCrea

In Attendance: Mr Jim Beatty (Assembly Clerk)
Mrs Gillian Lewis (Assistant Assembly Clerk)
Mr John Lunny (Clerical Supervisor)

Apologies: Mr Thomas Burns
Mr Mitchel McLaughlin
Ms Dawn Purvis

The meeting opened at 2.00pm in public session.

2.12pm The meeting went into closed session.

6. Consideration of the Committee’s Draft Report on Social Security Benefit Fraud and Error.

Members considered the draft report paragraph by paragraph. The witnesses attending were Mr Kieran Donnelly, Deputy C&AG, Mr Brandon McMaster, Director of Value for Money, Mr David Murdie, Audit Manager, and Mr Joe Campbell, Audit Manager, NIAO.

The Committee considered the main body of the report.

Paragraphs 1 – 16 read and agreed.

Paragraph 17 read, amended and agreed.

Paragraphs 18 - 23 read and agreed.

Paragraph 24 read, amended and agreed.

Paragraphs 25 and 26 read and agreed.

Paragraph 27 read, amended and agreed.

Paragraph 28 read and agreed.

Paragraph 29 read, amended and agreed.

Paragraphs 30 – 32 read and agreed.

Paragraph 33 read, amended and agreed.

Paragraphs 34 – 36 read and agreed.

Paragraphs 37 – 39 read, amended and agreed.

Paragraphs 40 and 41 read and agreed.

Paragraph 42 read, amended and agreed.

3.39pm Mr Hilditch left the meeting.

Paragraphs 43 and 44 read and agreed.

Paragraph 45 read, amended and agreed.

Paragraphs 46 – 52 read and agreed.

Paragraph 53 read, amended and agreed.

Paragraphs 54 and 55 read and agreed.

Paragraphs 56 – 60 read, amended and agreed.

Paragraph 61 read and agreed.

Paragraph 62 read, amended and agreed.

Paragraph 63 read and agreed.

Paragraph 64 read, amended and agreed.

Paragraphs 65 – 68 read and agreed.

The Committee considered the Executive Summary of the report.

Paragraph 1 read, amended and agreed.

Paragraphs 2 - 6 read and agreed.

Paragraph 7 read, amended and agreed.

Paragraphs 8 – 10 read and agreed.

Paragraph 11 read, amended and agreed.

Paragraphs 12 – 14 read and agreed.

Paragraph 15 read, amended and agreed.

Paragraph 16 read and agreed.

Agreed: Members ordered the report to be printed.

Agreed: Members agreed that the Chairperson’s letter requesting additional information and the response from the Accounting Officer, Department for Social Development, would be included in the Committee’s report.

Agreed: Members agreed to embargo the report until 00.01am on 29 April 2008, when the report would be officially released.

Agreed: Members agreed that they would not hold a press conference.

[EXTRACT]

Appendix 2

Minutes of Evidence

21 February 2008

Members present for all or part of the proceedings:
Mr John O’Dowd (Chairperson)
Mr Roy Beggs (Deputy Chairperson)
Mr Jonathan Craig
Mr John Dallat
Mr Simon Hamilton
Mr David Hilditch
Mr Trevor Lunn
Mr Ian McCrea
Mr Mitchel McLaughlin
Ms Dawn Purvis

Witnesses:

Mr Bryan Davis Mr John Nevin Mr Alan Shannon

Department for Social Development

Also in attendance:

Mr Kieran Donnelly		Deputy Comptroller and Auditor General
Mr Ciaran Doran		Deputy Treasury Officer of Accounts

1. The Chairperson (Mr O’Dowd): You are all very welcome. We have received apologies from Patsy McGlone. I remind everyone that given that mobile phones interfere with the recording equipment, they must be switched off, especially as we require a record of this meeting for our report.

2. The substantive item of business relates to the Audit Office report ‘Social Security Benefit Fraud and Error’. I welcome Alan Shannon, accounting officer of the Department for Social Development (DSD), Bryan Davis, chief executive of the Social Security Agency, and John Nevin, assistant director of disability, incapacity and security in the agency. Our evidence sessions usually last for about two hours, but sometimes they go over that time. I do not think that you have been forced into attending one of these sessions before, Mr Shannon, have you?

3. Mr Alan Shannon (Department for Social Development): Yes, I have.

4. The Chairperson: Then you are familiar with our methods. I will begin proceedings, and then members will ask questions.

5. We are examining the Audit Office’s report into benefit fraud. To put it into context, in 2006-07, the Social Security Agency paid out £3·5 billion on social security benefits. Of that sum, the Audit Office estimates that £60 million, which represents 1·7%, was overpaid through fraud and error. Internal staff error cost £28 million, customer error cost £14 million, and customer fraud cost £18 million. However, that does not include another area of expenditure — underpayments. An estimated £21·8 million was underpaid to customers. Some £2·6 million of that underpayment was due to customer error, and £19·2 million was due to internal staff error. We are therefore dealing with a substantial amount of public money.

6. Mr Shannon, the agency’s business is to provide financial assistance to those who are most in need, and it supports one of the Department for Social Development’s strategic objectives. However, fraud, error and underpayment of benefits have an impact on vulnerable people. Paragraph 1.2 of the Audit Office report tells us that the accounts of both the Department and the agency have been qualified for several years because of the levels of fraud and error. The results are disappointing. What is the Department doing to improve performance?

7. Mr A Shannon: Thank you for your welcome, Chairman. I will now take the opportunity to make a couple of introductory points.

8. I appeared before the Committee last month when we discussed the performance of the Child Support Agency (CSA). In my introductory remarks then, I said that the CSA’s performance was unacceptable and was not a source of satisfaction to anyone, including the staff who worked there. Wherever that is the case, it is right to come to the Committee and say so.

9. Today, however, I will take a different approach: I will not say the same of the Social Security Agency (SSA). The SSA’s performance in the past couple of years has been very good, probably the best that it has ever turned in. That is in spite of its having implemented a programme of change and despite losing 674 posts. I put on record that staff in the SSA have done an exceptionally good job. They sometimes feel unfairly treated by the media, so today I will attempt to redress that balance.

10. I have three points to make in response to your question. As far as the Department and the Minister are concerned, reducing fraud and error are priorities. A couple of years ago, the Minister revised the targets for the agency. It now has five ministerial targets, of which reducing fraud and error is one. The report details the activity that underpins that target. There is a robust benefit security strategy. Underpinning that is programme protection, better targeting, greater accountability, data matching, for the first time there is a major publicity campaign, and there are more prosecutions than ever before. I am pleased that the Northern Ireland Audit Office (NIAO) report acknowledges the effort that has gone into that strategy.

11. That strategy is producing real results. The level of fraud is half what it was four years ago: it is now down to 0∙5%. That compares favourably with the position in Great Britain and elsewhere. The UK agencies — the London agency and ourselves — have been compared internationally and have come out very well with respect to countering fraud and error.

12. Members may point out that, as far as error is concerned, we have not statistically made the same progress as we have with fraud. Nevertheless, taking account of the underpayments — as well as the overpayments — that the Chairman mentioned, the accuracy rate was 98∙7% last year, which, given all the difficulties and complexities, is not a bad performance.

13. My third point comes to the crux of your question. Can the qualification of our accounts be removed? As you say, it has been there for a very long time. The same qualification is applied to the Department for Work and Pensions. I find it difficult to answer the question positively, given the context: the case load is 1∙1 million; there are 36 benefits; and, as you said yourself, £3∙5 billion is paid out; there are 500,000 new cases each year; and 800,000 changes of circumstances. We have to deal with all those with an IT system that is up to 20 years old.

14. The different IT systems do not talk to each other. An elaborate range of clerical workarounds is in place to get around that sort of difficulty. The benefit system is highly complex — staff have to administer benefits with the help of advice and guidance documents that run to thousands of pages. The Audit Office in London has said that it would be difficult to expect further progress without some serious attempt to simplify the current benefit regime. That is one reason that the Secretary of State for Work and Pensions published at the end of last year the Green Paper, which contains some proposals along those lines.

15. That brings me to future challenges; I know that those are of interest to Committee members. In the next three years, we will face three challenges. Later this year, we will introduce an employment support allowance. The transformation of pensions will be introduced. We have to cope with all that with a declining workforce because of the Budget settlement, which will require removing 500 to 700 posts from the organisation on top of the 674 that have been taken out in the past three years. In the light of the turbulence that will inevitably be created by all those changes, it will be difficult to maintain, let alone improve, performance.

16. The combined figure for fraud and error is 1·7%. It will be difficult to improve on what is statistically a rather low figure. In its report of January 2008, the Audit Office in London said that diminishing returns inevitably set in when fraud and error figures are so low. Small incremental advances cannot even be set, because the sampling tolerances are not sufficiently precise to do that meaningfully. We will struggle, not only to improve, but even to maintain, last year’s performance.

17. My last point is about materiality. The reason that the Audit Office qualifies our accounts is because it takes the view that any level of fraud and error above 1% is material and therefore must be qualified. I do not know what the rationale for that is. Earlier this month, my opposite number in London raised that issue with the Public Accounts Committee in London. We have examined some other sectors, and we understand that the insurance sector works with a fraud and error figure of around 5%. The retail sector works with a figure of between 2% and 5%. Our concern is that, if the challenge is to work to 1%, we will probably never achieve it. I do not know whether that is a sensible measure to apply in these circumstances. The Committee might want to explore that with us in due course.

18. The Chairperson: Thank you for that interesting response. Of course you attended the Committee last month; it was I who was absent, and I apologise for that.

19. Without prejudging the Committee’s report, I suspect that the Committee will not endorse the view that the target for fraud and error be raised to 5%. The figures that you gave are perhaps an argument for the nationalisation of the insurance and retail industries. That might be more useful than arguing that because reducing the figure below 1·7% is unachievable, the target should be changed to 5%. Across the Assembly, all Departments and their agencies have a 1% target that they must reach. I understand your difficulties in such a complex system — you are dealing with a massive budget, with the public and with constantly changing circumstances.

20. After hearing about the Child Support Agency’s experiences with its IT system, the Committee may not wish to go down the road of hearing about IT issues at today’s meeting, but we may wish to explore the issue of financing IT with the Committee for Social Development. Our experience of IT systems has not been good.

21. Moving on, the previous Public Accounts Committee’s ‘Report on The Administration of Income Support Benefit’, which was published in 2000, highlighted the need to get a reasonable grip on the scale of fraud arising from passport benefits, and you have touched on some of those matters. However, that recommendation was rejected on the grounds that the Department had already measured and reported on housing benefit fraud, and that other bodies, such as the Central Services Agency and the education and library boards, had developed their own anti-fraud initiatives. Mr Davis, what are the main passport benefits, and what information is exchanged between those bodies to help to detect and prevent fraud?

22. Mr Bryan Davis (Department for Social Development): I will ask our benefit expert to answer the question about passport benefits.

23. Mr John Nevin (Department for Social Development): The main benefit that can be received via a passport benefit is housing benefit. We have worked very closely with the Northern Ireland Housing Executive on that. The Housing Executive receives information from us and has access to our computer system, so that it can check on the level of fraud and error. We carry out benefit reviews for the Housing Executive to establish the level of housing benefit fraud.

24. There are passport benefits for Health Service charges, and we have arrangements with the Central Services Agency, whereby it can compare and data match details that it has against claims. In the case of claimants of free school meals, we pass information to the education and library boards. There is a lot of contact between the Department and various agencies at an operational level. I do not know what the levels of fraud and error are on either Health Service charges or school meals.

25. The Chairperson: Mr Doran, what steps has the Department of Finance and Personnel taken to ensure that there are no impediments to information sharing?

26. Mr Ciaran Doran (Deputy Treasury Officer of Accounts): The main initiative that the Department has taken in the past six months is to work closely with the Home Office on the resolution of the legislation issue that surrounds the national fraud initiative, which has become law in Northern Ireland. We are working closely with the Northern Ireland Audit Office on the implementation of that legislation across all Departments, not just the specific agencies that were mentioned. We hope to see real progress on that in the next six months.

27. The Chairperson: Moving into another area that is not covered by the Audit Office report, there is concern about absenteeism. I personally do not like the technique that is used for measuring staff absence; however, the average staff absence at present is 18·5 days, which you hope to reduce to 12·4 days. I acknowledge the comments that you made earlier in praise of the Social Security Agency staff, because it is a difficult and pressurised area in which to work. In one sense I understand the high levels of sick leave in that agency. What measures are you introducing, first, to help to support the staff in their work and in the pressures that they face, and, secondly, to try to reduce absenteeism and increase efficiency, given that you mentioned that you will be losing 500 to 700 posts?

28. Mr A Shannon: One of the changes in recent years has been the proportion of staff taking sick leave due to stress. We have put several measures in place to try to pre-empt staff becoming so stressed that they have to take time off. We have several in-house staff who can offer support. We have a helpline that is run by an outside firm, which will take calls from, and offer advice to, staff who want independent help. We have a range of other measures for staff who are off sick, for example, they receive help and support from the occupational health service, and we can allow staff to return to work gradually when they start to feel better.

29. Much effort has gone into pre-empting absenteeism. However, on the efficiency and discipline side, we have a proactive, well-oiled machine to ensure that Civil Service procedures are applied rigorously throughout the Department. We have benchmarked Civil Service procedures against best practice everywhere else, and we are pretty satisfied that we are doing everything that we can within the present rules. We are always looking at opportunities to re-examine ways of improving the rules.

30. I am glad to say that we are experiencing an improvement in absence rates. This year’s absenteeism figures have been reduced by two days for the SSA and by four days for the CSA. The core is well in line with other Departments.

31. The reason that the statistics for DSD look worse than others — and I am passing no judgements — is that, statistically, female staff take more sick leave than male staff; junior staff take more sick leave than more senior staff; and organisations that are dispersed tend to have more sick leave than organisations that are based in headquarters. All those structural factors apply to DSD, and it tends to be struggling against factors over which it has no control.

32. Mr Lunn: My questions relate to the programme protection activity and are for Mr Davis. Paragraphs 2.5 and 2.6 of the report state that the programme protection activity focuses on checking the types of cases where the risk of error is greatest, and it assumes that the risk of fraud and error is the same in each district. What has the agency done to establish the level of fraud and error in each district? If that has not been done, how can you be sure that your resources are being allocated to those districts that are most in need and to those that would achieve the most effective outcomes?

33. Mr Davis: We have analysed the risks of fraud and error across the districts, and we have found little variation between them. On that basis, we feel that it is right to allocate resources based on the particular risks that are identified in each district. It is not unexpected that we have found so little variation between districts. Each district is administering the same benefit, with the same IT system, with broadly the same process and with staff that are skilled in broadly the same way. We would therefore not expect a huge degree of variation among districts, and we would not expect to see a huge variation between what is happening here in Northern Ireland and what is happening in Great Britain. As the report recognises, the nature and types of errors that are occurring here are broadly similar to those in Great Britain.

34. We are moving to a new and more developed risk-based assessment that gives us a better approach at targeting. It will identify those cases that are most at risk and will give us greater scope for outcomes. We are allocating the money to the districts where those particular cases happen to be.

35. Mr Lunn: At present, is it up to the district manager to decide the level of that activity in his own district?

36. Mr Davis: No. The report refers to the fact that, prior to 2005, there was an element of that. Since then, we have introduced a much greater range of accountability arrangements of targeting some structural changes. We have seen that the variability of performance between the districts has reduced between 2005 and 2006 and that they have all been hitting the targets since then and in the current period. There is much less scope for district managers to dictate the activities that they will undertake.

37. Mr Lunn: Appendix 2 of the report shows high levels of overpayments in some benefits, for example carers’ allowance and incapacity benefit, and, based on the agency’s annual report and accounts, significant underpayments. What is the agency doing to tackle that anomaly?

38. Mr Davis: That goes to the heart of the fraud and error strategy that we have developed to tackle those generally across the agency. That strategy is centred on three broad concepts. The first is to prevent error entering the system in the first place. We do that by risk assessing IT systems and the processes, by checking and trying to ensure that we get the right information from clients in order to reduce the scope for error to enter the system.

39. Secondly, a range of programmes is in place to detect error, should it enter the system. Obviously, programme protection is an important element of that, as is our work on fraud.

40. Thirdly, a range of work is designed to correct error when it occurs. Therefore, quite an integrated, broad-ranging programme of work is being undertaken for all benefits, and it is used to tackle both overpayments and underpayments in the system.

41. Mr Lunn: Paragraph 2.19 refers to the £7·3 million of public money that was allocated for training and other activities, but no targets or measures have been implemented to assess the impact of those activities. Can you explain why that is allowed to happen? How can the impact of the investment be assessed in the absence of those targets and measures?

42. Mr Davis: It must be recognised that programme protection activities are part of an integrated programme, which, as I explained, is part of the overall fraud and error strategy. Those activities are complementary, and specific case checks are carried out as part of the programme protection plans.

43. To supplement that work, indirect activities are carried out, and those are centred on some training activities. Currently, they are focused more on having programme protection managers, fraud liaison officers and additional validators in place. Those people act as a link between the centre and the districts — they feed information back to the staff in our districts, and they provide advice and guidance on how to tackle fraud and error in those districts. Therefore, although they are indirectly involved in the process, they are an integral part of the overall programme protection activities. To that extent, the Committee will appreciate that they contribute to the overall outcomes of our programme protection activity, rather than being divisible of themselves. Notwithstanding that, we recognise that it is important to at least have management measures in place as regards the activities of those particular groups of staff. Those measures have now been introduced.

44. Mr Lunn: My final question concerns staff training. I am astonished by the level of internal staff error, which is referred to in paragraph 1.1 of the report. The figure for the year 2006-07 is £47 million. That begs the question: how effective are your training programmes? What training do new staff receive, and how is staff performance assessed? How are remedial training needs identified and monitored? My question is nearly as long as some of your answers.

45. Mr Davis: I will try to be as brief as possible, Mr Lunn.

46. Training is a very necessary element of our approach to tackling fraud and error, but, on its own, it is not sufficient. To tackle the £47 million of error, we must develop and use the broader range of measures that are included in the fraud and error strategy, which I described earlier. We need to carry out more checks, get the IT systems right, provide more training, simplify application forms, and review our processes. All that will contribute towards targeting the levels of fraud and error in the system.

47. You asked about the training that staff receive. A learning pathway is in place for our staff, who receive all the training that they need to do the job effectively. Our processing staff are given induction training. There is also a programme of benefit-specific training, which gives staff the legislative background to the benefits and takes them through the IT processes and so on.

48. That benefit-specific training is followed up by a period of consolidation training in which the trainees work alongside an experienced mentor. They receive on-the-job consolidation, and, as they are only learning the job, 100% of their work is checked. The trainees’ mentor then provides feedback on how they can improve. That element of training lasts for about 13 weeks.

49. Staff receive refresher training, and we try to focus much of that on areas where error is most likely to occur. We have reviewed the whole process of our training: over the past year, we have introduced a further enhancement of the learning pathways, and we have developed a new learning framework. We have clarified some of the roles and responsibilities and set out a more structured learning programme. We are piloting a new assessment tool, which will allow us to measure more effectively how the training programme has been taken on board. In the past, we tended to assume that staff assimilated the training that they were given, but now we use an assessment tool to check the extent to which that has happened.

50. We have introduced benchmarking into the district offices, which sets targets in relation to errors that have been made. That provides us with an opportunity to identify the individual training needs of staff members, and then to correct for those needs. Over the past few years, we have also introduced team time, which is an opportunity for staff to get together to discuss new information and systems. Team time is crucial for picking up on problems that staff are struggling with and feeding that back.

51. We are looking further at additional training, particularly for our decision makers, and are considering a programme of accredited training with a university in England to take that forward.

52. In summary, we are doing a lot on training, but we recognise that we need to do more. However, training on its own will not tackle the £47 million of official error. The broader programme must be in place to allow us to do that.

53. Mr Lunn: Thank you for that. You seem to acknowledge that things had to improve, and you have taken action.

54. Mr Davis: There is always room for improvement, particularly with regard to training, as we learn more about areas where staff are struggling and where there are problems. As Mr Shannon said, we are dealing with a hugely complex range of benefits. For IS and JSA, we are dealing with 7,500 to 9,000 pages of staff guidance. It is a huge task for staff to try to assimilate that guidance, and it is inevitable that errors will creep in. We try to correct for that, and examine other ways of tackling the problem, but training alone will not address it.

55. Mr Dallat: I know that tradition is important in how Departments go about their work; however, from what I have heard, it is clear that fraud and error have been lumped together in the one report as being serious drawbacks for the people the money really should reach. The level of underpayments is staggering, especially when most underpayments are caused by agency mistakes. Is the agency’s failure to pay less than the correct amount of benefit not pushing claimants further into the poverty trap? What are you doing about that?

56. Mr Davis: We treat underpayments very seriously. The ministerial targets that have been set for the agency relate to overpayments and underpayments. We use the full gamut of the fraud and error strategy to tackle not only overpayments but underpayments. In particular, we use the work of the programme protection side to try to identify where underpayments are most likely to occur.

57. In the last three years, the work of the programme protection unit has been such that it has identified and corrected for some £39 million of underpayments. Therefore, we take it very seriously: it is an integral part of our fraud and error strategy. Programme protection is a particular means of identifying it and we correct for it.

58. Mr A Shannon: It also complements the benefit take-up campaign which is based on the same theme.

59. Mr Dallat: Last year, I was lobbied by a consultancy firm looking for my full support for the fraud campaign; however, I do not recall being lobbied for the other campaign.

60. Is there not a danger that the fraud campaign, if it is not balanced in some way, will put off the very people who need, and are entitled to, benefits to bring them above the poverty line but do not claim them?

61. Mr A Shannon: I will turn to Bryan in a moment. That was the first time that we ran a campaign of that kind. We are obviously keen to learn lessons from it. In our judgement it was broadly successful, but some of the statistics which came out of it are not entirely satisfactory. For example, there seems to have been an increase in the number of people who thought that benefit fraud was acceptable. Bryan and his team are giving careful thought to where we go from here, what a second campaign might look like and what the key themes might be.

62. Mr Davis: A balance must be struck between the benefit uptake and the fraud campaigns. This time round the fraud campaign was about tackling attitudes to fraud and raising awareness about fraud issues. With respect to the results, it is a curate’s egg: we have found some very useful results in raising awareness. People realise that working and claiming benefits is wrong. However, we had mixed results with respect to the acceptability of fraud: that has increased from 12% to 17 %. Clearly, that is worrying, and the report recognises that.

63. Overall, the fraud campaign was a useful start, but no more than that. We need to focus on the changes of circumstances that people face when dealing with benefit claims, and trying to help people to understand when they need to notify the agency of such changes.

64. Mr Dallat: My next question has been answered in part. However, can you be more specific about what you are doing to reach out to those who do not claim the benefits to which they are entitled?

65. Mr Davis: We have undertaken a range of benefit-uptake programmes over the last three years. Benefit uptake is one of the agency’s five priorities.

66. We started out in a small way in 2005-06, when we targeted 3,750 pensioners who we thought were not getting their full entitlement. That produced some useful results, and enough for us to build on in subsequent years. We did that in 2006-07, when we targeted 20,000 people who were likely to be claiming less than their entitlement. Those were older people, people entitled to disability living allowance (DLA) and some entitled to carer’s allowance. From those two exercises, we were able to increase annually the degree of benefit uptake by £7 million.

67. Based on that, we have expanded even further the current year’s programme. We reach out to 200,000 people at the moment: older people, the disabled, some with mental health problems, carers and families. In co-operation with the voluntary sector, we try to target those with specific requirements. There was a view that the voluntary sector was often closer to people and could intervene with them more easily than could a large Government agency.

68. That has been tested and seems to work quite well. Certainly, the voluntary sector is praiseworthy for the exercise and the results that have come through. There is more to be done, and we intend doing that in the near future.

69. Mr Dallat: In his introductory remarks, Mr Shannon mentioned computer systems that are 20 years old. That must be a serious obstacle to cross-departmental checks to identify people who are not getting their benefits.

70. Mr A Shannon: That is correct; checks do not interact with each other. Improvement of those systems is one of the agency’s ambitions for the future. However, as the Chairman mentioned, changing them is a high-risk enterprise. Our colleagues in London, who lead such operations, pulled back from a major change programme in 2006. Therefore, any major improvements have been postponed.

71. Mr Dallat: That is disappointing. The report refers to the difficulty of recruiting investigators. Are you confident that the required number of investigators is now in post and that that level can be maintained?

72. Mr Davis: Recruitment has been difficult. That has been largely a historical issue: in the early days, fraud investigation was a difficult job that required surveillance and a lot of outdoor work, often in challenging environments. Investigators were not always popular, and several of our staff faced severe threats. The Fraud Act 2006 has changed that environment considerably. Fraud investigation is now much more of a desk-based, interrogative type of job. It has become much more attractive. For example, people have been attracted to it by the professionally accredited training that it offers. In the last couple of years, the agency has reached its staff complement. We hope that that will continue.

73. Mr Dallat: Your only target is to increase the number of cases that are dealt with. Is there not a danger that only soft touches will be targeted and more difficult cases will fall by the wayside?

74. Mr Davis: The previous range of targets included what was known as the monetary value of adjustment. Our strong view was that the monetary-value-of-adjustment target could actually produce perverse results, such as those you suggested, Mr Dallat. We could end up with a situation in which trying to increase the monetary value of adjustment (MVA) almost encourages fraud to take place. We want the MVA to decrease over time as the efforts of fraud investigators deter fraud in the first place, or catch it much earlier in the process, so that adjustments to amounts of money are much smaller.

75. The agency wants to focus much more on sanctions. That is supported by the Criminal Justice Inspectorate’s report. When the perspective of an organisation is changing, a simple, easy-to-understand target is needed at that stage. Therefore, we have, initially, concentrated on a sanctions target. That has been effective in re-adjusting the agency’s focus. The level of sanctions is starting to increase. We are considering whether we now need to examine a broader range of targets to support the sanctions target. However, a sanctions target on its own was right for that stage in the organisation’s development.

76. Mr Dallat: I endorse the remarks that you made earlier, Chairman. My experience in Coleraine, Ballymoney and Limavady is that the agency’s staff do a good job, sometimes under difficult circumstances. It is important that that is recorded.

77. Mr I McCrea: Appendix 6 on page 51 of the report shows that, as late as May 2006, the agency did not have a fraud response plan in place to set out roles, responsibilities, commitments and timescales. Can the Committee be assured that a plan is now in place? If so, why was it not implemented many years ago?

78. Mr Davis: The agency’s fraud response plan was published in 2007 and has been implemented. Prior to 2007, the agency’s plan was incorporated in the Department’s overall fraud response plan, which also applied to other areas of work in the Department. Therefore, it is not that we did not have one; rather, it was part of the wider departmental response plan.

79. That was supplemented by a fraud and error strategy, which we developed in 1998. It was available to staff and was reviewed and updated regularly. It has also been revised a couple of times since 1998. That strategy outlined what we were expected to do about fraud and error more generally.

80. Mr I McCrea: Mr Doran, I know that you have recently given evidence to the Committee on fraud in general. Will you outline for the Committee what guidance the Department of Finance and Personnel has issued regarding the management of fraud and what the main thrust of that guidance is, particularly on the issue of fraud response plans? Will you also tell us when that was issued?

81. Mr Doran: The starting point for the guidance is ‘Government Accounting Northern Ireland’, which applies to every accounting officer. There is a specific chapter about fraud in that document.

82. The original guidance regarding the fraud response plan — ‘Managing the Risk of Fraud: a Guide for Managers’ — was issued in 1998 and updated in 2003. More guidance was issued by the Fraud Forum, which has been operating in Northern Ireland for a couple of years. Furthermore, the Department of Finance and Personnel has issued specific guidance about the fraud response plan in the past couple of years.

83. If the elements that were recommended for a fraud response plan are considered, my view is that the Social Security Agency already has most of those elements in place, and has had for several years.

84. Mr I McCrea: Mr Davis, paragraph 3.8 and figure 7 of the report show large increases in the number of fraud referrals. To what extent has that increase resulted from the enhanced investigative powers that were provided by the Fraud Act 2006?

85. In planning for the introduction of that Act, what assessment was made of the likely impact of those powers on the agency, and what steps were taken to ensure that the agency could respond quickly and positively to make full use of them?

86. Mr Davis: Figure 7 of the report shows a mixture of ordinary fraud referrals, referrals from the general matching service and from our prisoner-matching service. Therefore, that increase is largely due to the matching in prisoner cases, rather than to a large increase in ordinary fraud referrals per se.

87. You asked about the impact of the Fraud Act 2006. We did plan for the implementation of that Act, because, as you rightly suggested, it changed fundamentally how we went about our job. Therefore, we considered a range of protocols and codes of practice, such as in our relationship with the financial services sector.

88. Furthermore, we considered the impact of the Act on our staff training needs. There was also a requirement to get legal authorisations for a number of our staff to undertake certain functions, and we had to plan for that.

89. We reviewed the processes that were required in conjunction with our colleagues in the Department for Work and Pensions. We looked at the volumes of fraud referrals and we adjusted the number of staff upwards in 2006 from 93 to 102 to reflect an increase in the workload — not so much as a result of the increase in fraud referrals, but more in the complexity of the fraud investigation process, which would take more time.

90. We have worked closely with the Public Prosecution Service on evidential issues and on sorting out relationships between us. I suspect that, like any organisation with a fairly major change in process coming in just after 2003, we are probably still on the learning curve, but I see significant improvements.

91. Mr I McCrea: Paragraph 3·12 shows that the agency has closed approximately 3,000 older fraud referrals without investigation or follow-up. I find that an astonishing approach to cases of suspected fraud. Is that allowing fraudsters to get off lightly without a proper investigation of the case? Do you feel that it sends out the right signal or the wrong signal? Is it any wonder, therefore, that half of your staff are not confident that suspected fraud is being investigated fully?

92. Mr Davis: It would be useful for me to give some background information. Those cases had gone through our FRISC assessment system and were identified as relatively low-priority cases. They had come from relatively low-priority sources, and at the time those cases were coming through — and this is going back three or four years, before they were weeded out — we were running with a very low level of staff in the investigative grades. There was a need to prioritise the work we undertook. That prioritisation, and the need for it, is accepted by the Criminal Justice Inspectorate and by the Audit Office. Where we were at fault was in not weeding out those cases earlier, as investigations were likely to prove unfruitful. Even the best sources of referral still yield only one third of fraud; the other 60% come up with nothing. Those cases should have been weeded out earlier. They fit with the prioritisation approach that the agency has adopted. However, I must stress that the decision to weed those cases was approved at assistant director level in the organisation.

93. The Chairperson: Can you clarify one point? In response to Mr McCrea’s comments on paragraph 3·8, Mr Davis said that that included prisoner referrals. What is meant by that?

94. Mr Davis: When someone is sent to prison, their benefit entitlement falls. We need to be notified of those cases so that we can match them against our records and ensure that those benefit payments have been stopped. We get a weekly listing of those cases. It used to be on a monthly basis, but that had the potential for giving rise to a greater level of overpayment. However, now that we receive those listings weekly, we can tackle any potential overpayment very quickly.

95. Mr McLaughlin: My questions are for Mr Davis and are mainly focused on management, because I associate with the previous comments about the staff. Paragraph 2.27 and figure 6 of the Comptroller and Auditor General’s report show that only two districts out of six met their targets. Of the remaining four districts, West, Belfast North and East Antrim missed their respective targets by a country mile — 38% and 31% underperformance respectively.

96. Are you satisfied with the reasons given in the report, and that we can depend on those explanations? Is there a localised dimension to the underperformance, given that that two districts achieved their targets while the other four missed by significant amounts?

97. Mr Davis: To put the matter in context, district managers faced a broad range of priorities and targets at that time. To a certain degree, they moved people about to try and best meet those targets, which meant that staff were taken away from some of the programme protection activities. At one level, that is not unsurprising, because the people that we need to carry out programme protection and undertake case reviews are very often the people with the most experience. When managers face problems and constraints it is not unsurprising that they turn to those people to try to fill those particular gaps. That is what was happening — but only to some extent — in this case.

98. This happened up until 2004 and, to some extent, in 2005-06. As I said earlier, we have since tightened up the accountability regime and we have refocused attention on quality in the agency. Our focus is on how well we do things rather than how quickly we do them. If we get more things right first time, we believe that the speed issue will sort itself out over time. We made that change in focus in 2005, and part of it was to focus more and to ensure that districts used their programme protection allocations to the full extent.

99. With regard to results, the variation in performance between districts was very much less in 2005-06. In subsequent years, districts have beaten the targets set for them. That information is reported all the way through the management line, right up to the agency management board, to ensure that that focus remains.

100. Mr McLaughlin: You have given me an explanation as to why the underperformance occurred, which is reflected in the report in front of me. Unless I have missed something, I do not understand why there is such a differential in performance between districts.

101. Mr Davis: You are looking at a snapshot of a couple of years. Those districts were facing particular problems in those years in relation to the availability of staff to undertake certain duties. We have corrected that by refocusing on the overall programme protection activities. We do not believe that that degree of variation will arise again.

102. Mr McLaughlin: If we drill down further, does that mean that there were management problems in those districts that fell so dramatically short of their agreed targets?

103. Mr Davis: Not so much management problems —

104. Mr McLaughlin: Taking your earlier remarks about the efforts and dedication of the staff into consideration, where was the problem?

105. Mr Davis: I fully accept that. They were trying to balance a range of priorities at any one time, and were moving people about to try to meet those priorities. They were not being assessed or measured purely against the programme protection activity, but against the broader range of targets that had been set for them. They were moving resources and using local discretion to try to ensure that the range and spectrum of targets were being met.

106. Mr A Shannon: If one of those targets was to get the benefits paid first before doing anything else, it would be difficult to criticise a local manager for making that decision.

107. Mr McLaughlin: Let us move on to the question of the value-for-money ratio. We have funds that are voted for, and devoted to, programme protection. We see that there was a failure to review 76,000 cases.

108. Paragraph 2.29 shows that, by extrapolating the methodology that was the unit used, a detectable and recoverable £2·7 million of error, and a further £5·9 million of preventable error could have been identified. The figure of £1·7 million was devoted to that purpose, and the value-for-money ratio is quite significant. We are talking about identifying a total of £8·6 million, if that £1·7 million had been applied for the purpose for which it had been intended. Given that quite a significant amount of money could have been saved, does that not raise significant questions about the judgement of managers who took people off that cost-effective activity? To what extent are district managers’ salaries related to their performance?

109. Mr Davis: Obviously, district managers receive their basic salary, to which a performance element is attached. At the beginning of each year, all staff, including district managers, are set performance targets against which they will be assessed over that year. For a district manager, that will be a pretty broad range, but it will encompass programme protection. During the periods 2004-05 and 2005-06, a box 1 performance — the top box mark for someone who is doing extremely well — would have contributed about £400 to a district manager’s salary; in 2005-06, that figure had dropped to £150; and the current level for a top performer is about £100 a year. Therefore, performance bonuses are relatively small.

110. Mr McLaughlin: That is a surprisingly small sum. However, did every district manager receive a performance bonus, regardless of whether they achieved their targets?

111. Mr Davis: I cannot answer that at the moment. However, I can write to the Committee with that information.

112. Mr McLaughlin: I would appreciate that.

113. When reading paragraphs 2.35 to 2.37, I am left with the distinct impression that district managers were, to all intents and purposes, a law unto themselves. The protection unit was providing advice and guidance on two main areas: first, the need for increased general activity levels, given the identified shortfall in planned activity; and secondly, the need to focus activity on particular types of claims or case reviews, which emerging findings show would have the highest level of error. However, the report finds that in certain case studies, numerous recommendations were made and restated throughout the year, but that the districts failed to pick up on them because the unit had no authority to insist on their adoption.

114. A surprising example from 2004-05 shows that the unit recommended that one district would consider stopping the selection of a particular case type for intervention because of the very low rate of error detection. However, at the year end, that district had carried out almost 4,500 interventions of that type, which is 30% of its total activity. That would indicate that the district manager had neither regard for the advice, which appeared to be useful, nor any accountability when it came to ignoring that type of advice from the programme protection unit.

115. Mr Davis: I have to agree with you, Mr McLaughlin. That is unacceptable. We have improved accountability, and the report recognises those improvements. We have increased, and made more effective, the programme protection unit’s interaction and communication with district managers, and, if district managers are failing to meet any of their targets, we receive written reports from them. Areas of concern are highlighted to the assistant director and then discussed with the operations director at his senior management team meetings. In addition, those concerns are passed up to the agency’s management board. Therefore, the non-reaction to the programme protection unit’s advice that you witnessed will not happen again.

116. Mr McLaughlin: I welcome those measures. Does that mean that when specific recommendations are received, there is an on-the-record response — or memorandum — from the district manager in order that the process can be reviewed and individuals can be shown what they agreed to do in response to the advice?

117. Mr Davis: Absolutely, and, district managers issue us with what we call certificates of assurance, which highlight failings and problems. They assure the board and me that everything that is targeted and going on — not just in the area of programme protection — is being properly advanced by district managers.

118. Mr McLaughlin: Finally, paragraph 2.41 of the report discusses programme protection, essential in checking for and correcting errors in the benefit system. In order to do that, one would expect that individuals should require good knowledge of the benefit rules and processes. Can I presume that you would not dispute that?

119. Mr Davis: I absolutely would not; as I said before, generally, those people have good experience.

120. Mr McLaughlin: Nevertheless, when the report is examined, the results of the programme protection unit’s work indicate that there is a surprisingly high level of errors. How do you explain that? Is it because you do not have the right people who have the right knowledge and expertise?

121. Mr Davis: That brings us back to the general problem that the agency faces with benefits —they are complex and, as explained by Mr Shannon, the IT systems have limitations. Although an individual may be experienced, such problems persist and that person must still deal with the benefit system’s complexity and make judgement calls. Decisions are not all black or white; there are judgement calls to be made, and individuals interpret circumstances differently. Therefore, the nature of the benefit system and the supporting IT systems mean that a residual degree of error by officials will persist. Notwithstanding that, we are attempting to improve results, and subsequent validation exercises indicate an overall improvement in checking. That is heartening and demonstrates that we are moving in the right direction.

122. We have also introduced the specific post of district validator, whose job it is to carry out some of that checking and offer advice before decisions are checked by the independent standards assurance unit as part of the validation programme. We hope that that will provide an opportunity to generally improve checking at that level.

123. Mr McLaughlin: I have one further point to make —

124. The Chairperson: Other members are champing at the bit to ask questions.

125. Mr McLaughlin: You have no idea what he is like when he gets annoyed. [Laughter.]

126. Is there a case for road-testing new IT systems under the auspices of the programme protection unit?

127. Mr Davis: That is happening. As IT systems are developed, the programme protection unit carries out risk assessments of those systems and of the process itself. For example, the new IT system to administer employment support allowance, which will come online in October, is being developed by our DWP colleagues and, as an integral part of that work, its programme protection people are reviewing the process. When we implement that process in Northern Ireland, possible local variations will mean that it will be risk assessed by the programme protection unit.

128. Ms Purvis: Mr Lunn spoke about trying to remove errors, and you gave him a full and frank answer about training and explained that it was not the only way to deal with error. The NIAO survey of the Social Security Agency’s operational staff found that 25% of those who are involved in the administration of benefit claims said that they had not received any fraud-awareness training or briefing. That figure rose to 48% for specific areas such as disability benefit and carers benefit. Do those findings concern you? What measures are in place to ensure that all operational staff receive fraud-awareness training and briefing?

129. Mr Davis: Those findings certainly concerned us; they had to. Fraud-awareness training is important to us, so if any survey suggests that upwards of a quarter of staff feel that they had not received such training, we must take action. We reacted quickly and introduced compulsory fraud-awareness training for all staff who began working for us in 2007. That training was complemented by flyers, posters and desk aids, which reminded staff about fraud, the different types of fraud and about their role in fraud detection. We also published the fraud-response plan on the Internet so that staff could access it easily. We took the findings from the survey seriously and introduced measures to combat fraud. Although fraud awareness is part of the induction programme for all staff, we are also making it a compulsory annual cycle of awareness.

130. Ms Purvis: Earlier you mentioned a new assessment tool that checks the extent to which training and briefing is assimilated. Will you elaborate on that?

131. Mr Davis: I can elaborate only to a point; I will write to you about it in more detail. The general problem is that ticking boxes is easy where training is concerned. However, how much of their training do people actually understand? We are trying to assess the extent to which staff understand their training and how they have put it into practice. If problems persist, we must reassess the training with staff.

132. The assessment is of a practical nature, but I am happy to provide you with further details about it.

133. Ms Purvis: Thank you.

134. Appendix 7 shows that the various sources of referrals identified nearly £30 million of fraud and error in the past four years. However, the cost of identifying and pursuing those referrals is not shown. Do you measure the cost? If so, what is it?

135. Mr Davis: The follow-up work is done by the benefit investigation services, a unit that comprises 200 staff and 102 fraud investigators. The annual cost for that is just over £5 million, which includes the direct costs that are associated with the MVAs annually. Over those four years, the cost of that work was between £15 million and £20 million, because we were not fully up to complement.

136. Ms Purvis: Paragraph 4.11 states that the agency has had some success through large-scale exercises with selected employers to crossmatch wages records and employee details with benefit records. The last such exercise was carried out in 1999-2000. Have you carried out any similar exercises since then? Are there any plans to do so? What would be the costs of such an exercise, and do the costs outweigh the benefits?

137. Mr Davis: The exercise to crossmatch wages records and employee details with benefit records revealed some levels of fraud, but those tended to be relatively small amounts. Although the exercise was useful up to a point, it was not terribly so, particularly compared with the other mechanisms that have been developed, particularly, the general matching service. We are matching with organisations such as HM Revenue and Customs, and we are finding that that type of exercise is much more fruitful in identifying high-quality referrals. The table in appendix 7 of the report shows that around one third of those referrals led to a result for us.

138. Therefore, the exercise to which you referred is useful up to a point, but it is not the best compared with the other options that are available. The report refers to that when it mentions the success that we have had in that area.

139. Ms Purvis: Do you have any plans to carry out another such exercise?

140. Mr Davis: Not at the moment. We are continuing to develop the general matching exercises. Mr Doran referred to the national fraud initiative, and we will work closely with the Audit Office on that exercise, which will match a broader range of data systems to identify where potential fraud may exist.

141. Mr Beggs: Paragraph 1.7 and figure 2 of the NIAO report show that, although the level of customer fraud has more than halved between 2003 and 2004 and the end of December 2006, the level of customer error has virtually doubled over the same period. A sceptic might say that reclassifying fraud as customer error is an easy way of massaging the agency’s performance.

142. Mr Davis: That is definitely not the case. The figures, and the systems that are used to produce them, are important to us. Their credibility is crucial, so it would not be in our long-term interests to manipulate figures in the way that you have suggested. The figures show that, between 2005 and 2006, customer fraud fell by just over £10 million. In that same period, customer error fell by £200,000, which is a small amount. There is no reclassification in that statistic. The figures for 2003-04 and 2004-05 show that customer fraud fell by £10 million, but that customer error increased by £5 million.

143. The figures are the products of review and checking systems. They are validated by work that is done by our internal audit and by the Audit Office. Given the importance of the figures, we would not manipulate them.

144. Mr Beggs: How can you account for the significant increase in what is deemed to be customer error from 2003-04 onwards?

145. Mr Davis: Although the absolute amounts of customer error have increased significantly, the proportionate amounts have increased by between 0·2% and 0·4%. I suspect that that reflects a sampling issue. Some of the movements can be explained more by sampling variations than by a major increase in customer error. Figures will fluctuate on an annual basis, but it is important to examine them over a longer period. Over the four years that are shown in figure 2 of the report, there is a degree of constancy in the figures for customer error.

146. Mr Beggs: Paragraph 1.8 shows that fraud and error figures exclude underpayments and incorrect payments of disability living allowance. Is that exclusion aimed at making the figures look more respectable than they are, or does it mask the scale of the problem? Is the agency unconcerned about the underpayment of benefits to customers? John Dallat asked about that earlier. I am concerned that the exclusion of those figures masks a large amount of fraud and error. We are talking about £20∙2 million and £42∙8 million respectively. Why are underpayments and incorrect payments eliminated from those figures?

147. Mr Davis: You have asked a long question that raises several issues. Bear with me, and I will try to address each.

148. The figures for overpayment in the accounts reflect the approach that we have taken to focus on irregular payments. That is in line with what happens in DWP. However, as I said earlier, as an agency, our ministerial targets encompass both overpayments and underpayments. Our fraud and error strategy focuses on both, rather than dealing with just the overpayment situation. Therefore, we are not massaging the figures to the degree that you have suggested.

149. Dealing with changes in the circumstances of DLA claimants is a technical issue. It has been agreed that taking account of that is not legally an error; payments take proper account of the legislation. We are talking about a situation where the condition of someone who receives DLA changes gradually over the period of the award of the benefit. The claimant has a responsibility to notify the agency of any change in condition that will materially affect payment of benefit. It is difficult, with a complex benefit such as DLA, where there are 12 different potential outcomes, for a customer to readily identify a gradual change in condition that could then trigger a change in benefit payment. I understand why customers find difficulty with that.

150. We have undertaken a proactive review, which we call a periodic inquiry, into those cases on the DLA register that are most likely to change. Roughly 175,000 people are registered as being in receipt of DLA. With have taken the advice of statisticians and used a risk-based approach. We have identified a group of about 24,000 in which there is risk of a material change in condition. We are examining those with a view to reviewing fully individual cases, and we intend to send a mailshot to some of the others to remind them to notify us of any change. We have undertaken 7,500 full reviews, 3,200 of which resulted in changes, and we have made a £7 million adjustment to benefits as a consequence. Two thirds of the benefits were increased as a result of the review. That is not unsurprising, given that medical conditions are more likely to deteriorate than to improve.

151. We treat this matter seriously, but it is important to note that it is not legally regarded as error. That has been agreed with the Audit Office, DWP and the National Audit Office in GB.

152. Mr Beggs: Are the results of that adjustment in treatment responsible for the significant change in the figures in figure 4, where DLA fraud and error levels are significantly reduced?

153. Mr Davis: That is absolutely correct.

154. Mr Beggs: Paragraphs 1.3 and 1.4 record that the agency has responded positively to recommendations that the Westminster PAC has made. Does the agency monitor the costs of the various strands of its strategy? How much does all your counter-fraud and counter-error activity cost?

155. Mr Davis: There is a direct and an indirect answer to that question. The direct costs relate to the benefit security directorate, which has three main elements: the benefit investigation service; programme protection; and the policy and assurance group.

156. The work of the programme protection group costs approximately £5 million a year. In 2006-07, its work resulted in adjustments to benefit of some £39 million. The cost of investigating suspected fraud through the benefit investigation service cost £5·3 million. In 2006-07, that led to an adjustment in benefit of some £10 million. The third element, the policy and assurance group, is primarily a support and measuring group, and costs for it total £2·2 million. The total cost of all that activity is £12·5 million.

157. Mr Beggs: How do you determine in which group to invest resources to achieve the most beneficial outcome? Do you use any particular measurements?

158. Mr Davis: A large proportion of the agency’s budget is taken up by benefit delivery, and much of the remainder goes to changes to IT systems. We are therefore left with a relatively small amount to deal with fraud-related activities. The amount of money for the benefit investigation service matches roughly the level of referrals on the investigation actions that are needed to maintain that level of referral and stock.

159. However, over the past few years, as can be seen from the results, the overall level of benefit fraud has declined significantly and is now down to approximately 0·5%. As I said, the work of the benefit investigation service led to an adjustment to benefit of approximately £10 million. In the next couple of years, particularly when the agency faces other resource constraints — and Alan Shannon referred to the further 500 to 700 cuts that the agency must make — the question will arise as to whether it can afford to maintain that level of investigation service.

160. Programme protection costs £5 million and last year resulted in almost £40 million of monetary value of adjustment. That work is proving particularly effective, and, in many ways, we would like to invest even more in that activity. However, we are again constrained by the impending cuts, so we may not be able to afford any such increase.

161. Secondly, and probably more fundamentally, the type of people that we need to undertake programme protection work tend to be more highly skilled. We struggle to find such people in the system. Other members raised that point. I suspect that we will run with the current levels of investment for the next year or so, but we may be forced to review that in the light of future budgetary constraints.

162. Mr Beggs: Paragraph 1.9 of the Northern Ireland Audit Office’s report states:

“It is difficult to draw firm conclusions on trends”.

163. The reason it gives for that difficulty is that you keep changing the methodology, and, therefore, the figures for fraud are not comparable from one year to the next. In statistical analysis, one is taught to create consistency to facilitate comparison. Why do you keep changing your methodology?

164. Mr Davis: I agree fully with the thrust of your point. The agency is fully committed to adopting a robust approach to measurement. The report refers to an exercise that the National Audit Office undertook comparing our colleagues in DWP with other international organisations.

165. Their measurement system was found to be ahead of the game internationally. Our system was at least on a par with that used by the Department for Work and Pensions. We may even suggest that ours is slightly better.

166. Mr Beggs: My question was why does the methodology keep changing.

167. Mr Davis: We are faced with three main types of change. Pension credit was introduced in 2004-05, and the introduction of a new benefit adds new error, which has to be taken into account. That same discontinuity will probably occur next year with the impact of the introduction of the new employment support allowance.

168. Some definitional changes have also taken place. You spoke earlier about DLA changes of circumstances. There was a view that that should no longer be deemed as error; therefore, it needed to be stripped out of the figures. Overpayments were also stripped out to more adequately reflect the needs of the accounts. We were able to strip those figures out and produce a consistent run of figures. Therefore, there should be no great variation as a result of those two elements because, historically, they have been stripped out.

169. We also face technical changes. We now use what we call a common sample and increased stratification, which is a technical issue concerning our statistical approach. The aim is to give us a more accurate estimate of errors that are detected. Our statisticians tell us that that probably does not introduce great discontinuity to the system. It may have masked a slight improvement in the underlying error figures, because it found more errors as a result of the new stratification approach.

170. Mr Craig: Mr Davis, paragraph 4.12 of the report refers to the cross-border co-operation between the agency, the Department for Work and Pensions in Great Britain and the Department of Social and Family Affairs in the Republic of Ireland. Obviously, I hope that cross-border fraud is being investigated. What is the scale of that type of fraud, and has a value been put on it?

171. Mr Davis: We work closely with our DWP colleagues and with our colleagues in the Department of Social and Family Affairs in Dublin on the issue of cross-border fraud. We estimate that is not currently a major problem. Not many cases are being investigated and we are not being pressed by our colleagues in Dublin to follow-up a huge number of cases.

172. However, part of the difficulty has been that, until now, we have been unable to carry out the same data-matching exercises that we have been doing internally through the GMS system.

173. We are working with our colleagues in Dublin and in DWP to come up with data-matching protocols that will enable us to match records more regularly, which may actually be a better source for the identification of cross-border fraud. However, our systems are not currently identifying many instances of such fraud.

174. Mr Craig: In other words, any organised attempt to defraud the system would not have been identified because of the current lack of links between the different systems.

175. Mr Davis: Making use of the powers from the Fraud Act 2006 — which enables us to put the tentacles out much more — has not identified any particular cross-border fraud. However, the data matching will provide us with another way of investigating.

176. Mr Craig: I want to talk about another aspect of organised fraud — the notorious house drops. I have passed information about that to all Departments, but there have not been many prosecutions related to that type of fraud. Is the practice of using house drops targeted, and what results have been achieved in combating that type of crime? It is clearly carried out at an organised level.

177. Mr Davis: That is a difficult one for us. We have carried out exercises in relation to that type of crime. The last one was carried out two years ago, but it came up with very little, which confirmed an exercise that we had undertaken some time before. It has not been a very useful source of information about fraud. We tried those exercises a couple of times, but nothing came of them.

178. Mr Craig: Is the lack of results due to a dearth of information flowing to you, or is it due to a lack of support from other sources, such as the police?

179. Mr Davis: I do not think so. I will ask John to pick up on this matter.

180. Mr Nevin: One of the earlier exercises used a list of houses that were allegedly empty but were occupied. When those cases were investigated, there were legitimate tenants in those houses who were not claiming housing benefit, although there were housing benefit claimants in some of them. The information that we have received indicates that the practice of using house drops is not the major problem that it appears to be.

181. The Housing Executive, which is responsible for the administration of housing benefit, is taking action on a large scale to investigate the practice of house drops, and we will work with them to see whether there are cases that should be prosecuted. Historically, evidence has not shown that it is a major problem. The Housing Executive is looking at the matter, and we will be working with them.

182. Mr Hamilton: I turn to paragraphs 4.14 and 4.15 of the report and the growing complacency in attitudes towards benefit fraud. That exists despite the £500,000 advertising campaign to highlight the wider social impact and the cost of benefit fraud. According to the figures detailed in the report, the number of people surveyed who thought that benefit fraud was somehow acceptable rose from 12% to 17% in the space of a year, which is a worrying trend. What are you doing to ensure that there are strong deterrents against fraud, and that a loud and clear message about fraud is getting across to the public?

183. Mr Davis: I said earlier that the results of the campaign were like a curate’s egg. Some positive things can be taken from it, such as the level of awareness of the campaign, the extent to which people felt they were better informed about the issues, and the number of people that recognised working and claiming as a serious matter. You are right; the figures related to the number of people who said that benefit fraud was acceptable are of concern.

184. We have to venture out and use a similar approach to the anti-drinking-and-driving campaigns. We have to reiterate the message. A one-off shot will not be good enough. We are examining the learning points that came out of the campaign, and at a follow-up campaign that was running in GB at the same time. The initial campaign in Britain was similar to our own, and its results were broadly similar. The new campaign in Britain is more finely targeted. I talked to the authors of that campaign last week, and they are pleased that the level of results that are coming through now is much better than it was the first time out.

185. We propose to focus more on changes of circumstances and on reminding people when to let us know of changes: to tell us sooner so that that does not develop into a type of fraud. We are pleased that people see that working and claming is not right. We need to broaden the message that there are other types of fraud; in particular, people who have assets to declare, and people who are living together. We need a much more targeted campaign next time round, and learn lessons from the first campaign and the results in GB.

186. In respect of £500,000 costs, while that was not a prime reason for the campaign, we have followed up on some 4,000 referrals. People rang our hotline as a result of the campaign, and that resulted in a correction of £1 million of overpayments. The £500,000 of costs has resulted in the correction of £1 million of overpayments. However, that was not the purpose of the campaign.

187. Mr Hamilton: That was a positive outcome, and that is encouraging. Part of the problem of public attitude towards benefit fraud is not helped by the perception that those who perpetrate fraud are not sanctioned in any way. Between 2003 and 2007, according to the report, just over 5,500 cases were deemed to involve fraud of some kind. According to the agency barometer, nothing below £350 is taken to prosecution stage, and 1,600 cases of over £350 were considered non-sanctionable out of a remainder of 4,231. Therefore, around 40% of cases were not sanctioned in any way for a wide variety of reasons. Do you accept that all of those reasons are acceptable?

188. Mr Davis: I had a look at that. There were 2,960 cases, including the 1,056 of less than £350, that were not sanctioned. Our analysis is that 80% of those cases were closed in line with sanctions policy. However, that still leaves a residual number of 581 cases. In those cases, the agency was at fault. In mitigation, rather than as an excuse, half of those cases were due to delay in the system. Most of that delay related to the earlier period when we were short of investigators and did not have the capacity to take on a lot of that work. The report recognises a considerable improvement and progress on that.

189. The other element related to a failure in the investigation. Again, that reflected a lack of experience within the agency. We have put several improvements in place to counteract that. We have introduced a new fraud quality and disclosure policy, which will assess the quality of the evidence and the information that is available to us. We have introduced a new central decision-maker, who will ensure consistency in investigations. We are now up to complement, we have people trained, and we are building on experience. We have people who are more familiar with the outworkings of the Fraud Act 2006, we have new targets in place, and we are much more focused on sanctions. There is very close management monitoring of what is going on and, as a result, we see more sanctions flowing through the system. Even between 2005 and 2006, there was an increase of approximately 25% in sanction cases.

190. Although we do not have a full set of figures for 2007-08, it looks as if there will be a further rise of about 23% to 25% in sanction cases. Some mitigation is appropriate; and our actions to improve that percentage are starting to feed into the system.

191. Mr Hamilton: You are right: that does not excuse the number of non-sanctioned cases, but at least you are open about the agency’s fault and failure in that respect. Given the time, energy and money that are put into investigation, it is important that the investigation is solid and can be progressed to prosecution by the PPS. Enough money has been lost without more being wasted on shoddy investigations.

192. I appreciate that you may not be able to answer my next question at this stage and that you may have to come back to the Committee. How many individuals whose cases were not taken any further were later investigated for fraud?

193. Mr Davis: As you anticipated, I cannot answer that off the top of my head. However, we will investigate and seek to prosecute the cases that involved less than £350, if there is a repeat offence, and even if the repeat amount is less than £350. If we consider that someone has been particularly devious by, for example, altering a cheque, and even if the figure involved is less than £350, we are inclined to pursue that case. Some people have been detected, although we have not fully investigated their cases or maintained their records. However, if their names crop up again we will investigate fully and prosecute.

194. Mr Hamilton: Paragraph 4.27 of the report states that, since 2003, only 793 of the 1,412 cases that the agency passed to the PSNI went to court. Therefore, almost half of cases did not proceed to court. Why is that so? I suspect that your next response may overlap some of your previous answers.

195. Mr Davis: As you say, 793 of the 1,412 cases went to court, but 370 are still being considered by the Public Prosecution Service. You can see from the figures that the Public Prosecution Service withdrew 249 cases, due to the insufficient strength of available evidence.

196. I must stress that a judgement call must be made on whether the evidence is strong enough to sustain a prosecution. Our results show that when the Public Prosecution Service takes cases to court, there is a 97% success rate. On that basis, the PPS sets an extremely high level of evidential requirement before taking cases to court. We are learning about where that level is set and are about to implement a service-level agreement with the PPS. Following its implementation, the PPS will provide feedback on why it considers the evidence needs to be of a higher standard. That does not mean that there was a failure in the investigations; it simply means that the evidence that happened to be available was not strong enough to satisfy the evidential requirement of the PPS.

197. Mr Hamilton: It is crucial that the maximum number of cases that are sent for prosecution end up in court, and that relates back to the importance of public perception that has been discussed.

198. Mr Davis: That is correct. The agency is working closely with the Public Prosecution Service to ensure that each body has a more complete understanding of where the other is coming from.

199. Mr Hilditch: I have questions on deterrents and their effects, which may not be directly under your control, Mr Davis, but I value your opinion. Again the questions are to you, and I am sure that you must feel as though you are working overtime this afternoon. Paragraph 4.30 of the NIAO’s report states that the agency achieved convictions in 97% of cases that went to court. Simon has teased out why some cases do not reach court.

200. However, in some cases, successful prosecutions to recover the proceeds of fraud seem to take years. The report cites a case in which the recovery will take 100 years. Does that not send out a message that it is worth the risk to commit the fraud?

201. Mr Davis: The agency’s hands are tied on this issue. We must apply the legislation with regard to the amount of money that can be recovered at any one time. If someone has committed benefit fraud, the maximum amount of recovery is £12 a week. Even then, the amount of money that we recover is well down in the pecking order, so if the perpetrator has a social fund loan, they must pay that before they repay a fraud. Those precedents are set in legislation, so the agency can do nothing about them, except apply the rules as they stand.

202. Mr Hilditch: We will perhaps touch on that issue later. However, as seen in figure 15, in the majority of cases the penalty is a fine or conditional discharge. Given the nature or extent of the frauds described, do you consider that the penalties imposed will act as a sufficient deterrent to the fraudsters?

203. Mr Davis: You may not be surprised to hear that I am going to duck that question. The agency has no remit for sentencing policy and requirements. The judiciary and the court system hold that issue closely to their hearts. I am unable to comment on the nature of the sentences, but we have worked with, and met, the officials from the Court Service to ensure that they are aware of our approach to tackling fraud and of the efforts that we put into it. I would lose my head if I started to venture into the field of sentencing.

204. Mr Hilditch: Some of the issues contrast with the case in this morning’s newspaper of the Lottery fraudster. Given the variations in the penalties that you have outlined today, you have said that there has been some discussion. Will you elaborate on the intensity of that discussion? Has it been a one-off meeting, or is the matter being taken up with the Court Service?

205. Mr Davis: We have raised the matter a few times. John will pick up on that issue, as he was involved in those discussions.

206. Mr Nevin: We met officials from the Court Service to ensure that they appreciate the scale and volume of the problem, how we approach it and the degree of evidence that we can provide. A committee meets with magistrates, and they were provided with a DVD of programmes that were run by the BBC on benefit fraud in Northern Ireland. They said that there is no way that we can make direct representation, nor would we want to go to magistrates and make direct representations. However, the whole thing has been put in context, and there are ongoing liaison arrangements with them. As we develop different ways of carrying out our work, they will be kept in touch with what is going on. To be fair, there has been some sound evidence that the judiciary appreciates the seriousness and the extent of the crime. One magistrate said that social security fraud in the past had been regarded as being almost acceptable, but that that was no longer the case.

207. Quite a few suspended sentences, community service orders and jail terms have been handed down. We have been advised that the judiciary regards those three types of sentence as having equal weight. A community service order is imposed when the magistrate feels that a jail sentence is appropriate, but — for whatever reason — does not wish to impose that sentence.

208. Mr Hilditch: That appears to be in about 28% of cases, so it is good that there has been activity in that field.

209. Mr Nevin: It has been increasing in recent years.

210. Mr Hilditch: I see from paragraph 4.31 that the financial investigation unit has completed investigations in only 44 of the 134 cases referred to, and that it has closed 43 of those cases.

211. Why has it taken so long to complete the investigations and why are so many closed? I assume that was without recovery being affected. Why are more cases not referred to the Assets Recovery Agency (ARA)?

212. Mr Davis: I stress that the financial investigation unit was set up in 2005, so it is very new. Initially, it had a staff of two; during 2007-08, we increased that to eight. It has taken time for staff to get accreditation to be able to undertake that work and the training required to do it. Only well into 2006-07 did the staff become effective. Some 14 cases have been finalised; 21 are in the judicial process; and nine are with the ARA.

213. Overall, recoveries, which took place mainly in 2007-08, have amounted to £500,000. The unit costs about £200,000 to run. However, we are a part of the Home Office incentivisation scheme, whereby if we can secure confiscations, we are allowed to retain an element of that money to reinvest into further investigations. In 2007-08, that source provided almost £70,000, which was clearly a considerable abatement of the costs of the unit. Work is in progress: the newness of the unit explains why performance in the earlier years was relatively low.

214. Mr Hilditch: In the early stages, are you content with the £230,000 as the cost-effectiveness of the unit?

215. Mr Davis: We would like that figure to increase. John would like to contribute.

216. Mr Nevin: I am sorry; I did not catch the question.

217. Mr Hilditch: I was asking whether the agency was content with the cost-effectiveness of the unit: it costs £230,000, as against a recovery of £500,000.

218. Mr Nevin: The four new staff — the financial investigators — did not receive accreditation until October 2007. They have taken the cases through. Cases of an average of £30,000 are prosecuted, and it is part of the wider Government strategy to use the proceeds of crime to finance such units. The manager of the unit attends a meeting of the criminal financial section of the Organised Crime Task Force. We are told frequently that, with respect to bringing cases forward, we are the most active group in Northern Ireland. Cases require a lot of investigation. We have to prove that there are assets before we can take action. Many investigations reveal that there are no assets and they are dismissed as quickly as possible, so that time can be spent on complex cases.

219. Mr A Shannon: It is possible to measure the cost of the activity and some of the outcomes. It is more difficult to measure the deterrent effect of this activity, and the best we can do is fall back on the statistic that fraud, as a whole, seems to have diminished to 0∙5%. Deterrence is always difficult to measure.

220. Mr Lunn: I have a specific question for Mr Davis. When someone makes a claim for income support but forgets or fails to declare that they have substantial savings, how can the agency pick that up before payment is made?

221. Mr Davis: There are systems that will pick that up. John will address the detail.

222. Mr Nevin: Do you mean before the person is paid?

223. Mr Lunn: Yes. When a claim for income support is made, the agency will carry out normal checks, but how does it check for substantial savings?

224. Mr Nevin: At present, we are limited in what we can do at the time that the person makes the claim. At the end of the year, we data match with HM Revenue and Customs, and that exercise is likely to reveal that there will be interest payments, which will lead to investigation. We are developing a system, which, it is hoped, will allow access to up-to-date, online information at the time that a claim is made. If a previous claim revealed that a person had had capital in the past, we would ask them for their authority to go to their bank. If they refused to provide that information, we might be able to take steps under the Fraud Act 2006.

225. Mr Lunn: Can you carry out those checks at the end of the financial year, after you have started to pay the income support?

226. Mr Nevin: At present, the data matching occurs at the end of the tax year and for some time after that, when the tax returns from the banks become available.

227. Mr Lunn: Yes, but if someone had substantial savings when they made a claim, it is probable that they would have had those moneys at the end of the previous year. Could you not carry out a retrospective data match if you suspect that someone is not being entirely frank with you?

228. Mr Nevin: There are strict rules on data matching to which we must adhere. We get results from HM Revenue and Customs only when there is an indication that a particular match does not appear to be consistent with what we have been told. In other words, in the previous tax year, that person would not have had a claim to benefit, and his records could not be matched against the return from HM Revenue and Customs, because he was not receiving benefit. Are you with me?

229. Mr Lunn: I am with you.

230. Mr Nevin: Once his name appears on our benefit register, it can then be matched against returns from HM Revenue and Customs. That would probably happen a few months after the end of the tax year in which he made his claim.

231. Mr Lunn: Could I ask just one more question? If someone made a claim for income support halfway through the financial year, and you got information from a whistle-blower that that person had £250,000 in the bank, does that mean that you cannot do anything about it?

232. Mr Nevin: No. That is a different situation. If we had reasonable evidence, we would investigate that situation, probably very vigorously and quickly. Once we have sufficient evidence to suggest that a crime has been committed, under the Fraud Act 2006 we can write to financial institutions to obtain information. For example, if we knew that the money was in a particular bank, the case would be very easy to pursue. If it were being kept in several banks, we would have the power to write to them for that information.

233. Mr Lunn: What would happen if someone pulled up to the office in a Bentley? You said that you needed reasonable evidence.

234. Mr Nevin: If someone is sustaining a lifestyle that is obviously incompatible with the income that is being declared, I would like to think that they would be investigated rapidly, particularly if they turned up to the office in a Bentley.

235. Mr Beggs: I assure you that this is my final question, Chairperson.

236. Mr Shannon, both you and your colleagues have mentioned the complexity of the benefit system and the limitations of the IT system. Mention was made earlier of thousands of pages of advice and guidance. You also mentioned the new Green Paper that is out for consultation. How are you feeding into that system in order that a new system will emerge that will be simpler, but will meet the needs of vulnerable people in Northern Ireland and be simpler for your staff to administer?

237. Mr A Shannon: We have a policy unit that is not a part of the Social Security Agency. It is headed by John O’Neill, who has appeared frequently before the Committee for Social Development to discuss the Pensions Act 2004 and its associated regulations. John’s staff sit on a network of committees in London that are involved in evolving social security policy. They are involved in a benefit simplification committee — I forget its correct name — that examines possible alternatives to the current system.

238. I do not know how far that process will go. There is a trade-off between creating something that is simple and something that is matched to individual circumstances. When you simplify a system, there is always some pressure to include exceptions and to tweak procedures to make it suitable for individuals. That is when it becomes complicated. I cannot predict what the outcome of the exercise will be, but intensive work is going on.

239. Mr Beggs: I was just seeking reassurance about that, because neither I nor anyone else wants to be back here discussing the same issues in 10 years’ time.

240. Mr Craig: My experience of local benefit fraud is fairly poor. Most of the alleged benefit fraud turns out to be mistakes in the system. One might be tempted to say that deliberate mistakes might have been made in order to help people out at the time. Do you keep a record of errors and who made them? What action is taken if staff are found to have made mistakes? For example, are they retrained, or do you examine the reasons why such mistakes were made in the first place? Either I am dealing with very honest people, or I am the only person who seems to pick up on the fact that much of the so-called fraud actually results from mistakes in the system.

241. Mr Davis: If our staff have made mistakes, we give feedback to the individuals who are involved. As I said earlier, we have introduced a new benchmarking system whereby we examine the accuracy of a sample of the staff’s work. If we identify particular issues and problems with that work, we consider training or providing further information and advice for staff. That would pick up on the issues that you referred to, Mr Craig,

242. The Chairperson: We will conclude this section of the meeting. It has been a full and frank discussion. Mr Davis, especially, seems to have borne the brunt of the questions today. It is clear that a lot of good work has been done but that there is more to do. The Committee’s report will make several recommendations in that respect. I have no doubt that the media will, rightly, pick up on the broad headline that so many millions of pounds have been lost due to fraud.

243. Some individuals are involved in career fraud, and others who are involved in fraud to keep their heads above the poverty line. However, we should also concentrate on the fact that tens of millions of pounds are not claimed, especially by some of the most vulnerable people in society, such as elderly people or those who suffer from mental-health problems. I am glad to hear that the Department is working with the voluntary and community sector in a bid to minimise the amount that goes unclaimed.

244. Mr Shannon, the figure of 5% that you mentioned earlier has been rolling around my head during the whole session.

245. Mr A Shannon: I mentioned that 5% is the insurance industry’s target level of fraud. The proposition is being discussed in London as to whether 2% would be a more realistic target than 1%, because the level of fraud and error in Great Britain is down to 2·1% and our level is down to 1·7%. That is the area in which the target is possible. In our judgement, reducing the level to 1% is probably not possible.

246. The Chairperson: That is probably why the Chairman of the Finance Committee left; he went to lie down. [Laughter.]

247. I would hate to hear what Peter Robinson would have to say about that.

248. It has been a good evidence session. Thank you very much. The Committee will publish its report. There were two questions, to which you said that you would respond in writing, and we will forward those questions to you. Thank you.

Appendix 3

Correspondence

Chairperson’s letter of 26 February 2008

Public Accounts Committee
Parliament Buildings
Room 371
Stormont Estate
BELFAST
BT4 3XX
Tel: (028) 9052 1208
Fax: (028) 9052 0366
Email: cathie.white@niassembly.gov.uk

Mr Alan Shannon
Accounting Officer
Department for Social Development
Lighthouse Building
1 Cromac Place
Gasworks Business Park
Ormeau Road
Belfast
BT7 2JB

Date: 26 February 2008

Dear Alan

RE: PUBLIC ACCOUNTS COMMITTEE EVIDENCE SESSION 21 FEBRUARY 2008

Further to the evidence session at the Public Accounts Committee yesterday, please provide the following additional information which members requested at the meeting:

1. Please provide details of the performance bonuses, in addition to box marking bonuses, which were paid to District Managers over the years 2004-05 and 2005-06.
2. Please provide detailed information about the new assessment tool.
3. What is the full cost of an investigator and what is the average payback per investigator in terms of fraud detected? What additional staff is necessary to reduce the current backlog of referrals and what would it cost and save?
4. Paragraph 3.32 refers to the development of an IT system which will provide a basis for enhancing the management of fraud referrals and also benchmarking with GB. When is the system expected to be implemented? If it is delayed or does not deliver the anticipated benefits, what steps are you taking to ensure better management of fraud cases?

I would be obliged for a response by Monday, 10 March 2008.

Yours sincerely

John O’Dowd
Chairperson
Public Accounts Committee

Correspondence of 7 March 2008
from Mr Alan Shannon

Lighthouse Building
1 Cromac Place
Gasworks Business Park
Ormeau Road
BELFAST
BT7 2JB
Telephone: 028 90 829002
Facsimile: 028 90 829560
E-mail: perm.sec@dsdni.gov.uk

John O’Dowd
Chairperson
Public Accounts Committee
Room 371, Parliament Buildings
Stormont Estate
BELFAST
BT4 3XX

7th March 2008

Dear John

Public Accounts Committee Evidence Session - 21 February 2008

Thank you for your letter of 26 February 2008 seeking additional information following the evidence session of 21 February. The issues raised by Committee members and the Department’s response are as follows;

Performance Related Bonus

The Committee asked for details of performance bonuses and box marking bonuses awarded to District Managers in 2004/05 and 2005/06.

Box Marks

In the period 04/05 - 05/06 the total amount paid for Box Marking Bonuses equated to £1,850 (£1400 +£450), broken down as follows;

04/05

1 Box 1 @ £400 = £400.
5 Box 2s @ £200 = £1,000.

05/06 year

2 Box 1s @ £150 = £300.
3 Box 2s @ £50 = £150.

Performance Bonus

04/05 year

None were awarded for 2004/05 year.

05/06 year

A total of 5 bonuses were awarded at a cost of £3,750.

5 @ £750 = £3,750.

New Assessment Tool

Members asked for further details of the learning assessment tool mentioned during the evidence session.

The learning assessment tool forms part of a wider training package for staff with responsibility for processing and maintaining claims in District offices. In order to support and assist staff during their knowledge based benefit specific training, an IT based, Learning Assessment Tool to test trainee knowledge has been developed. More specifically the learning assessment tool will provide for:

• measurement of trainee knowledge at the beginning of the training event;
• measurement of the progress in trainee knowledge on completion of training event;
• identification of strengths and weaknesses in trainee learning/course content; and
• assurance that learning standards have been achieved.

Costs of Fraud Investigation

The Committee also asked for details of the outcomes from fraud investigation against investment and the resources needed to reduce the backlog of fraud referrals identified in the NIAO report.

The average cost of an investigator is £25,000 per year. Across the 2006/07 year there were on average 93 investigators in post, with overpayments identified in that same period of £10m as a result of counter fraud activity. Due to the continual flow of cases, outcomes in a particular year do not necessarily correspond directly to activity in that same year. However, in general terms of direct investigator costs against overpayments, this equates to a £2.3m investment for £10m return

The backlog referred to in the NIAO report related to a build up of cases due to reduced staffing complements in earlier years. That difficulty has since been overcome and the Benefit Investigation Service is now effectively managing the volume of referrals received each year. As we explained during the evidence session, many of the referrals received do not reveal criminality thus it is necessary to sift out low risk cases to ensure value for money and maximum impact from this investment. This remains a key element of our workflow management process and together with the increase in staff means that we are no longer dealing with a backlog situation. Those cases not selected for investigation are identified at an early stage and passed to Operations staff for consideration of follow-up action.

Fraims

Members also asked for details of when the new Fraud IT system is due to be implemented and what contingency plans are in place in the event it is delayed. A business case is currently being developed to assess the options around the introduction of a Fraud Referral and Interventions Management System (FRAIMS). Subject to the outcome of the business case evaluation, the new system should be implemented by April 2009 and will help workflow management of Fraud referrals and allow enhanced retrieval and collection of management information to assist analysis of fraud types/trends.

The Agency has introduced interim measures which can be developed further in the event the FRAIMS project is delayed or does not proceed. The risk scoring model referred to above was developed in 2007/08 and allows for better management of the workflow by identifying early and sifting out low risk cases. The risk model is regularly updated as further fraud outcomes are known thus allowing for the analysis of fraud types and trends to help target future activity. As a further contingency to the FRAIMS project, enhancements to the current Benefit Investigation IT systems are also under development.

I hope the Committee finds this additional information helpful.

Yours sincerely

Alan Shannon

Appendix 4

List of Witnesses
who Gave Oral Evidence
to the Committee

List of Witnesses who Gave
Oral Evidence to the Committee

1. Mr Alan Shannon, Accounting Officer, Department for Social Development.

2. Mr Bryan Davis, Chief Executive, Northern Ireland Social Security Agency.

3. Mr John Nevin, Assistant Director of Disability, Incapacity and Security, Northern Ireland Social Security Agency.

4. Mr Kieran Donnelly, Deputy Comptroller and Auditor General, Northern Ireland Audit Office.

5. Mr Ciaran Doran, Deputy Treasury Officer of Accounts, Department of Finance and Personnel.

[1] ‘Changes in circumstances’ refers to cases where the Agency considers that, due to the gradual nature of improvement or deterioration in the customer’s condition, the customer could not have been reasonably expected to have known to report those changes to the Agency.

[2] The general matching service provides data matching, the electronic comparison of data held on a computer system with data from the same or other computer system, to identify specific inconsistencies in that data. The Agency matches data on the benefit system with selected data from Her Majesty’s Revenue and Customs systems.

[3] From 1 April 2008, the Assets Recovery Agency has been subsumed within the Serious Organised Crime Agency.