The National Fraud Initiative in Northern Ireland 2024

Foreword

Preventing and detecting fraud is a continuing challenge for all public sector organisations. Organisations must seek to minimise losses to fraud, in order to maximise the proper use of funding for public services. It is not possible to eradicate fraud but we must do all in our power to minimise its impact.

Since 2008, under statutory powers inserted in the Audit and Accountability (Northern Ireland) Order 2003 by the Serious Crime Act 2007, my Office has undertaken data matching exercises for the purpose of preventing and detecting fraud. The main data matching tool we use is the National Fraud Initiative (NFI), administered by colleagues in the Public Sector Fraud Authority.

This is the eighth NFI exercise to be undertaken in Northern Ireland. The cumulative outcomes from the NFI in Northern Ireland are over £48 million, representing current and past fraud and error stopped and potential future fraud and error averted. The low levels of outcomes in some datasets in this exercise, indicate how valuable the NFI exercise has been in making bodies more aware of fraud and encouraging them to put in place effective controls over fraud.

Nationally, outcomes from data matching through the NFI are now over £2.9 billion, demonstrating the continuing value of this cross-jurisdictional exercise. I will continue to work collaboratively with the Public Sector Fraud Authority, Audit Scotland, Audit Wales and participating organisations to ensure that the value and impact of the NFI is maximised.

These achievements would not be possible without the efforts of the 80+ participating organisations in Northern Ireland who review and investigate data matches in order to detect fraud and error. I thank all those involved for continuing to support this important work.

If your organisation does not already participate in the NFI, I would encourage you to join our collective effort. You can contact my Office at nficoordinator@niauditoffice.gov.uk for further information, or go to our website.

Dorinnia Carville - Comptroller and Auditor General

Key Facts 2022-24

£48 million - Cumulative value of NFI outcomes in Northern Ireland since 1 April 2008

£3.7 million - Total NFI outcomes in Northern Ireland in 2022-24 exercise

£3.1 million - The majority of savings continue to be identified from pensions matches

£0.5 million - There was a significant increase in rates outcomes in this exercise (compared to £57k in the previous exercise)

82 - Public sector bodies in Northern Ireland participated in this NFI exercise

Key Messages and Recommendations

1. Maximise the benefits

A number of organisations pay the NFI fee but are not effectively participating. In this exercise we identified three organisations that did not investigate any matches and five organisations that only completed minimal investigation work (see Audit overview). This represents poor value for money and minimal benefit to those organisations.

All organisations should maximise the benefits to be gained by actively participating and investigating NFI matches. This should include:

• Ensuring that an appropriate level of resource is available.
• Reviewing the guidance within the NFI secure web application, to help ensure the most effective use of limited resources when reviewing and investigating NFI matches.
• Prompt and timely review of matches.
• Prioritising high-risk matches.

We also encourage all organisations that have previously participated in pilot exercises to reflect on the benefits gained through successful pilots, for example through the GP registration data pilot exercise in 2018-20, and seek ways to staff and actively participate in these exercises on a regular basis.

2. Self-appraisal

Audit committees, senior management, and officers leading the NFI should review the NFI self-appraisal checklist. This two-part checklist is designed to assist audit committee members when reviewing, seeking assurance on, or challenging the effectiveness of their organisation’s participation in the NFI; and for officers involved in planning and managing the NFI exercise. Audit committees and management should ensure they are sighted on their organisation’s progress, and where necessary seek to understand reasons for low or nil outcomes.

3. Take improvement action

Where auditors recommend improving the timeliness and rigour with which NFI matches are reviewed, NFI participants should take appropriate action.

4. Identify new datasets

Organisations are encouraged to consider if there are any other datasets, not currently included, where there are clear indications that fraud may exist, which could be progressed into a pilot. Given the preparatory work that needs to be done for new untested datasets, it is essential that these are notified to us at an early stage to allow us to consider their inclusion in the data matching exercise.

5. Voluntary participation

We encourage all organisations to consider the benefits the NFI could bring to them in the detection and prevention of fraud. In particular, we continue to encourage housing associations to sign up to the NFI. Following our 2013 report on Tackling Social Housing Tenancy Fraud, the Public Accounts Committee commented that it expects all housing associations to participate in the NFI. However, to date, only three housing associations have participated in past NFI exercises, with none taking part in the current exercise due to other priorities. We continue to engage with the housing associations and the Department for Communities to encourage participation.

6. NICS Fraud Forum

The NICS Fraud Forum acts as a Best Practice Liaison Forum. It is advisory in nature and aims to assist departments fulfil their responsibilities in respect of anti-fraud activity. We encourage the NICS fraud forum to meet regularly (at least twice a year in accordance with its terms of reference) and support and promote the NFI within the public sector.

7. Digital Economy Act (2017)

England, Scotland and Wales match data from His Majesty’s Revenue and Customs (HMRC) using powers under the Digital Economy Act (2017). There are significant aspects of this legislation that have not been adopted in Northern Ireland therefore we are currently unable to match to HMRC data. The Department of Finance are leading discussions regarding the implementation of The Digital Economy Act in Northern Ireland, however to date progress has been slow. We encourage the Department of Finance to work with the Cabinet Office to expedite discussions and implementation in Northern Ireland.

NFI Outcomes in Northern Ireland 2022-24

The NFI is a major data matching exercise undertaken every two years which aims to prevent and detect fraud. Appendix 1 provides an overview of the NFI process.

In Northern Ireland, 82 public sector bodies participated in the NFI. This includes all Councils, Health and Social Care sector bodies including Health Trusts, Further Education Colleges and all central government departments, and their arm’s length bodies.

Between 1 April 2022 and 31 March 2024, outcomes for the NFI in Northern Ireland were £3.7 million, compared with £4.4 million in the previous reporting period (a decrease of 16 per cent). The total figure comprises actual outcomes of £0.55 million and estimated outcomes of £3.16 million. Total outcomes for the eight NFI exercises to date in Northern Ireland are over £48 million (see Appendix 2). An NFI outcome describes the overall amounts for fraud, overpayments and error that are detected by the NFI exercise and an estimate of future losses that it prevents (Appendix 3). Figure 1 shows NFI outcomes for 2022-24 and the previous three exercises.

Insert Figure 1

In this exercise, outcomes in all areas decreased when compared to the prior period, except for rates and creditors which had increased outcomes. Rates outcomes have increased significantly (777%) compared to the previous reporting period, due to renewed focus on data matches now that staff are no longer occupied on high priority COVID-19 related work.

Figure 2: NFI outcomes in Northern Ireland compared to the previous exercise

Source: NFI web application

Note: COVID-19 grants data matching was a one-off exercise in the previous reporting period.

Occupational Pensions

How is fraud and error identified?

• Occupational pension information is matched to deceased records (known as mortality screening) to detect where a pension paying authority has not been notified of the death of a pensioner and so a pension continues to be paid after the date of death.
• Eight Northern Ireland public sector pension paying authorities submitted pensions data for mortality screening as part of the main NFI exercise. In addition, one of these authorities later submitted data for interim mortality screening offered by the NFI, and outcomes from that are included in this report.
• Occupational pension records are also matched to payroll records to identify cases of pensioners returning to work without notifying the pension paying authority, thereby possibly avoiding a reduction (abatement) in pension.

Outcomes

• 46 cases where pension remained in payment after the date of death of the pensioner, compared with 78 cases in the previous reporting period.
• £3.1 million* of actual and estimated savings, compared with £4.2 million in the previous reporting period.
• One pension abatement case identified of £2,728 compared with no cases in the previous reporting period.

*Note: The vast majority of pension outcomes are from a small number of cases where the pensioner died relatively young and was in receipt of a significant pension, leading to a large, estimated outcomes figure (see Appendix 3 for calculation methodology).

Case Examples

A pensioner died in September 2021 but the pension paying authority only became aware of the death via NFI data matches in February 2023. Overpayment of pension amounted to £14,108. The pension paying authority is pursuing recovery of the funds through contact with the deceased’s personal representatives however, to date the full amount is still outstanding.

A pensioner died in September 2022 but the pension paying authority only became aware of the death via NFI data matches in February 2023. Overpayment of pension amounted to £14,990. The pension paying authority recovered the funds in full in October 2023.

A pensioner died in April 2022 but the pension paying authority only became aware of the death via NFI data matches in January 2023. Overpayment of pension amounted to £9,522. The pension paying authority recovered the funds in full in June 2023.

Rates

How is fraud and error identified?

• Rates records are matched to the electoral register to identify:
  • properties where someone is registered to vote but the property is not on the Valuation List; and
  • properties where someone is registered to vote but LPS have not yet identified the liable ratepayer on the rating system.
• Lone Pensioner Allowance (LPA) gives a 20 per cent rate rebate to people aged 70 or over who live alone. LPA records are matched to death records, electoral records and state pension records to determine whether the award of LPA is still valid.

Outcomes

• 103 cases of rates avoidance detected, compared with 8 cases in the previous reporting period.
• £481,000 of actual and estimated savings, compared with £30,000 in the previous reporting period.
• 34 cases of incorrect award of LPA detected, compared with 50 cases in the previous reporting period.
• £23,113 of actual and estimated LPA savings, compared with £28,000 in the previous reporting period.
• Four national insurance number corrections within LPA records, helping to prevent future fraud and error.

Case Examples

NFI data matching to the electoral register prompted investigation of property ownership on the Land Registry system. A property was identified which was recorded as ‘pending ratepayer’ on the rating system but at which someone was registered to vote. Outstanding rates dating back to 2020, amounting to £10,356, are now subject to normal billing and recovery processes.

In another case, a person was registered to vote at an address, but the property was not on the Valuation List for rates. When the property was added to the Valuation List, rate arrears dating back to April 2017 of £12,535 were billed and are now being paid through an agreed payment arrangement.

LPS was not informed of the death of a ratepayer, therefore LPA remained on a pensioner’s rates account after the date of death. Once identified as part of the NFI exercise, the LPA award was cancelled effective from the date of death and £665 relief was removed from the rate account. All rates for this property have been paid in full.

Trade creditors and procurement

How is fraud and error identified?

• Data matching suppliers information allows organisations to identify duplicate payments, and can also highlight cases where system improvements or “housekeeping” are required, for example the removal of duplicate creditor reference numbers.
• NFI data matching between payroll, Companies House data and trade creditor records also helps organisations to detect links between staff on their payroll and companies with which they trade. Matches may reveal undeclared conflicts of interest which have resulted in a financial advantage to a staff member or someone with whom they are closely connected.

Outcomes

• 28 duplicate payments identified, compared with 16 in the previous reporting period.
• £90,475 recovered from suppliers, compared with £40,000 in the previous reporting period.
• 19 cases where action has been taken to correct non-monetary errors (such as a duplicate creditor reference number), compared with 65 such cases in the previous exercise, thereby helping to prevent fraud and error occurring in the future.
• No cases of undeclared conflicts of interest detected, compared with two cases in the previous reporting period.

Case Examples

A duplicate payment of £6,456.26 was identified due to the same invoice being keyed in twice in error. The invoice was keyed with uppercase letters for one payment and lowercase for the second. This meant the duplicate payment was not picked up internally. The overpayment was discovered in October 2023 as a result of the NFI exercise. The supplier was contacted and the amount was recovered in December 2023.

In another case, a duplicate payment of £23,167 was discovered in September 2023. The duplicate occurred because the invoice was provided twice by the supplier in error and on the second invoice a digit on the invoice reference was incorrectly captured. The supplier was contacted and the duplicate payment was recovered in November 2023.

Housing benefit

How is fraud and error identified?

• People on low incomes may receive housing benefit. Fraud and error can occur when calculations are based on inaccurate information, for example where:
  • the claimant does not declare a source of income; or
  • the claimant does not declare a change of circumstances, e.g. additional residents at the address.
• The NFI matches housing benefit records to a range of datasets, including public sector payroll and pensions, student loans, deceased records and housing tenancies, in order to detect inaccuracies.
• The Northern Ireland Housing Executive (NIHE) administers housing benefit for those who rent and who own their homes.

Outcomes

• Four cases of housing benefit fraud, error and overpayment detected, compared with 13 cases in the previous reporting period.
• £10,600 of actual and estimated savings, compared with £51,000 in the previous reporting period.
• Housing benefit outcomes have reduced in recent exercises following the transfer of NIHE housing benefit fraud investigation to the Department for Communities (DfC) in 2017. The DfC uses Real Time Information from employers and pension providers as its main focus for housing benefit investigations. Therefore whilst there continue to be matches as part of the NFI exercise, many of these cases have already been identified by DfC and therefore do not count as NFI outcomes.

Case Examples

A housing benefit claimant had not declared that they were in receipt of a student loan. This only became apparent via an NFI data match. Overpayment of housing benefit amounted to £1,812.10 for the period September 2022 to February 2023.

In a similar case, the claimant was overpaid £751.41 for the period October 2022 to February 2023.

In both cases, the housing benefit claim has been cancelled and recovery of the amount is in progress.

Private supported care home residents

How is fraud and error identified?

• Health Trusts may contribute to the care home fees of older people. If care homes fail to notify Trusts, either fraudulently or erroneously, that a resident has died, payments may continue after the date of death of the resident.
• The NFI matches Trusts’ private supported care home payment records to death records, to help identify such cases.

Outcomes

• One case of an erroneous payment to a care home detected, compared with two cases in the previous reporting period.
• £543 of actual and estimated savings, compared with £35,000 in the previous reporting period.

Case Example

A wrong date of death was recorded on the Abacus system which resulted in an overpayment of £542.88 to a care home. The date was corrected and the overpayment was recouped in April 2024.

Blue badges and concessionary travel passes

How is fraud and error identified?

• Blue badges are administered by the Department for Infrastructure (DfI). A blue badge entitles the holder to concessions such as use of parking spaces designated for blue badge holders and free on-street parking in “pay and display” areas. Matching blue badge holder records to death records helps to identify potential fraudulent use of a badge after the death of the registered badge holder. Northern Ireland blue badge data is also matched to data from England, Wales and Scotland, to identify cases where a person may be holding more than one badge.
• Concessionary travel passes are administered by Translink on behalf of DfI and are issued to a number of eligible groups, such as people aged 60 and over. Details of travel pass holders are matched to death records to identify cases where a pass is still in circulation, and could therefore be used fraudulently, after the death of the pass holder.

Outcomes

• Data matching identified 2,537 blue badges still in circulation after the death of the badge holder, an 18 per cent increase on the previous exercise. DfI examined a 10 per cent sample of the matches and found that all 250 cases were already known to them and cancelled, based on General Register Office for Northern Ireland (GRONI) data. The Northern Ireland Blue Badge System is updated directly each month from a report from GRONI and marked as ‘not for automatic renewal’. No issues of fraud were found in the sample examined.
• Data matching identified 3,455 travel passes still active after the death of the pass holder (3,883 in the previous exercise). Only four of these had actual usage recorded, to the value of £58.77, after the date of death (16 cases and £349 in the previous exercise). Translink had already cancelled 1,655 of the passes, based on monthly information provided by the GRONI. It cancelled the remaining 1,800 passes by the end of March 2023.

Other outcomes - Payroll

How is fraud and error identified?

• The NFI matches payroll data across all participating organisations to identify cases of employment fraud, for example:
  • employees working for one body while on long-term sick leave from another; or
  • employees with two jobs where shift patterns overlap, so that it would not be possible to cover both jobs.

Outcomes

• No cases of payroll error, compared with two in the previous reporting period.
• £nil outcomes, compared with £2,270 in the previous reporting period.

Social Housing

How is fraud and error identified?

• In Northern Ireland, the majority of social housing (around 48,847 properties) is owned and managed by the Northern Ireland Housing Executive (NIHE). In addition, 19 registered housing associations own and manage around 43,000 properties. The NIHE is a mandatory participant in the NFI (see Appendix 1) but the housing associations can participate voluntarily. In this NFI reporting period, none of the housing associations opted to participate.
• The NFI matches local tenancy data to the tenancy data of other participating social housing providers across the UK, and also to housing benefit, to help detect tenancy fraud by identifying where a person appears to be resident at two properties and therefore may be subletting one property unlawfully. Housing tenants’ data is also matched to death records to ensure proper reassignment of a tenancy on the death of a tenant.
• In addition, the Northern Ireland social housing waiting list, which is maintained by the NIHE, is matched to housing tenants, housing benefits and deceased records to detect undisclosed tenancies and undisclosed changes of circumstances.

Outcomes

• No social housing outcomes have been recorded from data matching exercises completed in this NFI exercise.

Making the NFI work for your organisation

The following tips will help you to get the best out of the NFI exercise:

Roles and responsibilities

• The Senior Responsible Officer (SRO) should nominate an appropriate Key Contact who has the necessary time and influence to ensure NFI work is started and progressed effectively, in line with the suggested timetable.
• SRO and Key Contact should agree the approach and timeframe. The NFI recommends that key reports and high risk matches are prioritised. We would encourage sign-off of the proposed approach by the Audit Committee.
• Users nominated to investigate data matches must have a good knowledge of the business area they are investigating (payroll, creditors etc.).
• Users should familiarise themselves with the latest guidance to ensure effective working. Comprehensive guidance is available under the Help menu in the NFI web application. The Public Sector Fraud Authority publication NFI Matters (under help/documents) highlights relevant hints and tips to make the most effective use of the NFI web application.

Investigating matches

• Prioritise key reports and higher risk matches.
• Use the report comment facility to record your intended approach.
• Follow up matches promptly so that fraud and error can be stopped at the earliest opportunity.
• Work within the secure web application – this streamlines the process, allows information to be shared easily and ensures data security. Exporting of data should be kept to a minimum.
• Do NOT investigate every match. Use a risk based approach.
• Periodically review shared comments from other organisations and respond appropriately to any queries.

Recording and reporting

• Users should record short but informative comments on matches within the NFI web application; this allows the SRO, Key Contact, NFI Coordinator and auditors to determine progress.
• Use the report comment facility to record a comment for multiple matches where appropriate, rather than entering the same comment numerous times. This saves time and effort in processing matches.
• All outcomes, both quantitative and qualitative (e.g. national insurance number corrections), should be recorded in the comments and outcomes boxes. Only record outcomes that are a direct result of the NFI exercise.
• Use outcomes to make informed system improvements e.g. strengthening controls.
• SRO should report progress and outcomes (including nil outcomes) to senior management, the Board and the Audit Committee.
• Take positive assurance from having few matches and no monetary outcomes.
• Use the NFI self-assessment checklist to quality assure your approach to the NFI.
• Use the outcome of your NFI work to inform your Annual Governance Statement.

Audit overview

Audit teams in the NIAO monitor progress by participating organisations, to help ensure they are getting the best out of the data matches provided and to encourage progress where necessary. We reviewed progress at the end of December 2023, 11 months after matches were released, and rated organisations’ NFI approach as Green (satisfactory), Amber (mostly adequate) and Red (unsatisfactory) (see key below).

Overall performance had improved when compared to the previous exercise with 90 per cent of participating organisations rated as satisfactory (82 per cent in the previous exercise). Three organisations (4 per cent of total participants) had not looked at any data matches. Two of these organisations had been classified as unsatisfactory at the same point in the previous exercise. NIAO staff continued to monitor those organisations categorised as red or amber and by the end of March 2024, most had made further and sufficient progress. However, one organisation had not processed any of its data matches, attributing its lack of activity to ‘staff shortages’.

Unsatisfactory

1. No reports opened or data matches processed 11 months after match release
2. No clear timetable in place to review and investigate matches
3. Internal issues in these organisations diverted time and resources away from the NFI

Mostly adequate

1. Not all key reports were opened 11 months after match release
2. Low number of matches processed
3. High risk matches not prioritised
4. Comments/queries from other organisations not being responded to

Satisfactory

1. Work commenced promptly
2. Key reports and higher risk matches prioritised
3. Clear and informative comments recorded
4. Outcomes clearly recorded
5. Report comment facility used appropriately
6. Shared comments from other organisations reviewed and responded to as appropriate
7. Clear arrangements for internal reporting of the NFI exercise

Insert Figure 2

Looking to the future

The Public Sector Fraud Authority continues to develop the NFI and the NIAO continues to monitor how developments might be applied in Northern Ireland to extend the scope of the NFI through increased participation and additional data matches:

Private landlords: We are currently engaging with DfC and Public Sector Fraud Authority colleagues in relation to a pilot in Northern Ireland to match private landlord data with housing benefit and LPS landlord rates rebate data to identify any landlords who have not registered with NI Direct. It is hoped that this pilot will go ahead in Q4 2024/25.

Housing tenancy fraud: This pilot, launched in Autumn 2022 with a number of pilot bodies in England, combines public and private sector data to help detect subletting of social housing and unlawful succession. The second phase was launched in spring 2024 including matches from new data. To date the Northern Ireland Housing Executive and Northern Ireland housing associations have not been involved in the pilot but are keen to participate in future phases.

Digital Economy Act (2017): We will keep this in view and continue to seek updates on progress (see associated recommendation in Key messages and recommendations).

Appendix 1 – Overview of the NFI

What is the NFI?

The National Fraud Initiative (NFI) is a major data matching exercise, run by the Public Sector Fraud Authority (PSFA) every two years, which detects fraud and error. The PSFA was launched on 3 August 2022 as an integrated partnership between the Cabinet Office (CO) and HM Treasury (HMT). The PSFA engages with government departments and public bodies to understand and reduce the impact of fraud. The NFI compares data from over 1,200 organisations across England, Scotland, Wales and Northern Ireland, making it a powerful tool. The NIAO co-ordinates the exercise in Northern Ireland, with over 80 local public sector organisations participating. These include government departments, local councils and health bodies.

Governing legislation states that organisations which are audited by the C&AG or a local government auditor may be required to participate in data matching exercises. These are known as mandatory participants. Other bodies may participate on a voluntary basis where the C&AG considers it appropriate. The purpose of data matching must be “to assist in the prevention and detection of fraud.”

Key steps in the process

• Participating organisations submit the required datasets, e.g. payroll, pensions, domestic rates, to the secure NFI website.
• The NFI matches the data and identifies inconsistencies, which are returned to participants as ‘matches’ via the secure website. Examples of matches include:
• Payroll to payroll - could indicate a person working elsewhere while off sick.
• Pensions to deceased records – could indicate that a pensioner has died but the pension-paying authority has not been informed.
• Rates to electoral register – could indicate a person is resident at a property which is not on the Valuation List for rates, or where a liable ratepayer has not yet been identified.
• Organisations have around 12 months to review and investigate their matches. Organisations are not expected to investigate all matches but should determine their approach based on their knowledge of key fraud risks and the risk scores applied to each match by the NFI system.
• Organisations record relevant comments and outcomes within the secure NFI website. Outcomes may be monetary (actual and estimated, see Appendix 3) or non-monetary (such as correction of national insurance numbers or duplicate creditor references).
• Organisations with few matches, or no fraud or error in the matches they investigate, can take positive assurance that their internal control environment is operating effectively, and use the NFI to inform their Annual Governance Statement.
• The Public Sector Fraud Authority collates the total outcomes and the Comptroller and Auditor General (C&AG) reports on the outcomes for Northern Ireland.

Data protection and data security

Data protection and data security are fundamental to the NFI. The C&AG’s Code of Data Matching Practice3 and the NFI privacy notice ensure compliance with data protection legislation. They let individuals know why their data is being matched, the standards that apply and where they can find further information. The C&AG has statutory authority to data match and does not require the consent of the individuals concerned.

The NFI uses a secure web-based application for the transmission of data and for access to matches by the participating organisations. The NFI’s systems are accredited to handle, store and process information up to the restricted classification level.

Appendix 2 – Total NFI outcomes in Northern Ireland

Notes:

The bulk of the outcomes for social housing (£1.6m) relate to a pilot exercise carried out in 2018-20, matching care home data with social housing data. This exercise was a one-off and not repeated.

GP registration data matching was a one-off pilot exercise carried out in 2018-20.

COVID-19 grants data matching was a one-off exercise in the previous reporting period.

Appendix 3 – Calculating outcomes and estimated forward savings

A summary of the outcomes methodology calculations applied to the NFI datasets included in the 2022-24 exercise in Northern Ireland is set out in the table below.