Updated 04 February 2013
Fraud and error in the benefit system as a percentage of benefit expenditure technical description
| Short title | Fraud and error in the benefit system, as a percentage of expenditure |
|---|---|
| Technical definition | This indicator measures the estimates of the levels of overpayment and underpayment, as a percentage of benefit expenditure, due to fraud and error across the benefit system in Great Britain |
| Rationale | This is the primary DWP indicator for levels of Fraud and Error in the benefit system. It is included in the DWP Business Plan and is a measure used in the DWP resource accounts and annual report. The indicator is important for DWP assurance on the impact of anti-Fraud and Error activity and has been used as the benchmark measure for the Fraud and Error Strategy. The indicator is used as a measure for Structural reform priority 1: Reform the welfare system, action 1.9: To reduce the level of benefit expenditure overpaid to a maximum of 1.7% by 2015. |
| Formula | The indicator is calculated from sample data, adjustments are made and data is grossed up to provide an estimate for the whole benefit population. The methodology is audited by the National Audit Office every year. Worked Example: Fraud and Error 2011/12 (final): The estimate of total overpayments due to fraud and error across all benefits is £3. 4 billion; this is 2. 1% of the total benefit expenditure, which was £159. 2 billion in 2011/12. The estimate of total underpayments due to fraud and error across all benefits is £1.3 billion; this is 0.8% of the total benefit expenditure in 2011/12. These estimates are subject to statistical sampling uncertainties as detailed in the “Robustness and data limitations” section below. |
| Start date | Published twice a year (May and October/November) |
| Good performance | A statistically significant decrease in the percentage of overpayments and underpayments would demonstrate improved performance but economic conditions and overall expenditure would also need to be taken into account.
The estimates are based on a random sample of the total benefit caseload and are therefore subject to statistical uncertainties. This uncertainty is quantified by 95% confidence intervals around the central estimate. These 95% confidence intervals show the range within which we would expect the true value of fraud and error to lie. The general reader can broadly estimate whether changes are significant by seeing how much the confidence intervals overlap. If they do not overlap then the differences between estimates are generally significantly different and indicate a real change in the estimates over time. If confidence intervals do overlap then the difference between estimates are generally not significantly different, indicating that any changes are more than likely to be due to sampling variation rather than real change. As detailed in the Latest data section above, the levels of overpayments and underpayments of Fraud and Error in the benefit system have both remained level over time since 2005/06. |
| Behavioural impact | This indicator should not lead to perverse incentives as it is based on a random sample of the live caseload, independently reviewed for fraud and error. |
| Comparability | The indicator is a recognised standard. It is recognised by the UK Statistics Authority, the National Audit Office and the National Fraud Authority.
Comparable statistics are available for Northern Ireland and HM Revenue and Customs Tax Credits. |
| Collection frequency | Published twice a year (May and October/November) |
| Time lag | 7-8 month time lag from the end of the measurement period to when the report is published (ie the period ending September 2012 will be published in May 2013). |
| Data source (which data collection it comes from) | The fraud and error measurement system consists of a random DWP sample survey of approximately 34,000 benefit cases covering Income Support, Jobseeker’s Allowance, Pension Credit and Housing Benefit. The Department also undertakes one off National Benefit Reviews (NBRs) for various benefits to estimate the level of fraud and error in a particular financial year using the same process. For those benefits not continuously reviewed or reviewed via NBRs, a proxy measured is used to estimate their rate of fraud and error. Proxy estimates for the unmeasured benefits are either linked to the results of a measured benefit that was likely to have a similar level of fraud and error, or are estimated by applying a rate based on the average value of incorrect payment for all measured benefits. The survey combines data collated from DWP administrative systems and Local Authority owned Housing Benefit systems with data collected from the claimant during an interview. The statistics are calculated from the results of a sample survey, which are recorded on an internal DWP database. For more information on the methods used to produce these estimates and how they are quality assured please refer to: |
| Type of data (Whether it is an official statistic, national statistic, survey, MI ) | National Statistics |
| Robustness and data limitations | Data is robust for statistical purposes at national level but the sample is not large enough to be able to produce statistically valid results at regional or local level. In any survey sampling exercise, the estimates derived from the sample may differ from what we would see if we examined the whole caseload. Where possible, uncertainties have been quantified to give an overall assessment in the form of 95% confidence limits. These confidence limits show the range within which we can be 95% sure that the true value lies. For more details on how confidence levels are calculated please see; More details on data limitations and the care required in interpreting results are included in Chapter 1 of the latest National Statistics report published at: |
| Collecting organisation | DWP (Statistical Services Division in the Information, Governance and Security directorate in Professional Services) produce these statistics. The Performance Measurement branch carry out the survey fieldwork and the Fraud and Error Measurement Analysis branch analyse the results to produce and publish the fraud and error estimates. |
| Return format | Data is shown as a percentage and is also available in monetary value terms (£) |
| Geographical coverage | National GB level – England, Scotland & Wales |
| How indicator can be broken down | The fraud and error estimates can be used for:
Obtaining an estimate for the amount over/under paid in total and by benefit, and broken down into fraud, customer error and official error, across the benefits administered by the DWP and Local Authorities. Obtaining estimates for the amount over/under paid by benefit, broken down into the types of fraud, customer error and official error, across Income Support, Jobseeker’s Allowance, Pension Credit and Housing Benefit. |
| Further guidance | Current and historical National Statistics publications for these estimates are available from: This page also contains links to further detailed documentation including guidance on Quality and Methodology; Uses and Users; Ad hocs and Pricing; and Variance and Confidence Intervals. The results data is published in reports and Excel spreadsheets on the above landing page. The source data is stored in a SQL database and analysis data is stored in SAS datasets. |
