Earnings and circumstance data matching
Matching each claim against authoritative records other bodies hold, such as employer-reported pay, tax records, and death and prison records, using same-period figures to adjust awards with a right to dispute, and sending every retrospective or uncertain mismatch to a caseworker who asks the claimant first.
Part of the recommended approachTechnique: Rules Engine
Data matching compares what a claimant declared with what an employer, the tax authority, or another agency has recorded for the same person and period.
In Universal Credit, earnings that employers report to the tax authority are used directly to calculate each monthly assessment, and a claimant who thinks the figure is wrong can dispute it with wage slips or bank statements, which is why the largest earnings overpayments come from the self-employed, whose income no employer reports.
DWP’s guidance also suspends a claim when expected earnings have not been reported for two assessment periods, until the claimant confirms what they received.
A match that looks backwards, such as a death record, a prison record, or a discrepancy with past tax data, goes to a caseworker, who checks the period it covers, contacts the claimant, and decides whether an overpayment arose.
The lesson of Australia’s Robodebt scheme is that a mismatch is a question, not an answer: the scheme raised debts from matches against tax office income, put the burden on people to disprove them, and was found unlawful in court.
This is the right first defence for overpayments caused by circumstances another body records.
It is the wrong tool for living arrangements and self-employed earnings, where no authoritative record exists and the random sample and case review have to carry the work.
Strengths for this problem
- Checks a circumstance directly against the body that knows it, so a match on employer-reported pay or a death record is evidence rather than a guess.
- Every match is explainable to the claimant, the caseworker, and a tribunal in terms of the record that triggered it.
- Catches unreported changes soon after they happen, before months of overpayment build up.
- Needs no labelled history, so it works from the first day a data sharing agreement is in place.
Limitations for this problem
- Covers only circumstances some other body records, so self-employed earnings, living together, and savings held outside the agency's reach are largely missed.
- Records cover different periods and definitions from the benefit rules, such as annual tax income against a monthly or fortnightly assessment period, so a mismatch is not an overpayment until a caseworker checks it.
- Depends on data sharing agreements, legal gateways, and common data standards that take years to put in place.
Data needed
- Employer-reported earnings per pay period, with employer, pay date, and the period the pay covers.
- Death, prison, and immigration status records, and benefit and pension records from other agencies, with source and effective date.
- Where the law allows it, information from banks on balances or income relevant to eligibility, limited to what the legal power specifies.
- The claim's declared circumstances for the same period, so the match compares like with like.
How it is measured
- Share of mismatches that a caseworker confirms changed the award, per match rule.
- Value of overpayments prevented by acting on a match before payment, against value recovered afterwards.
- Time from the change in circumstances to the corrected award.
- Share of match-driven decisions overturned on reconsideration or appeal.
Pitfalls
- Turning a mismatch directly into a debt and requiring the claimant to disprove it, which is what made Australia's Robodebt scheme unlawful.
- Comparing figures over different periods, such as an annual tax amount against fortnightly or monthly entitlement, without a caseworker establishing what was earned in each period.
- Letting match rules accumulate unreviewed, so caseworkers spend their time on rules that rarely change an award.
Human oversight
Where the rules use an authoritative same-period figure, such as employer-reported pay for the current assessment period, the award is calculated from it, and the claimant is told the figure used and can dispute it with a caseworker. A caseworker reviews every retrospective mismatch, every suspected error in the record, and every case that could lead to an overpayment debt or a fraud referral, contacts the claimant, and decides. The claimant is told what record the agency holds and can respond before any debt is raised. Policy and legal teams approve each match rule and the legal gateway behind it, and analysts review each rule's confirmation rate.