In plain English: every two years the government's National Fraud Initiative cross-checks councils' records to flag payments that may never have been legitimate — pensions still paid to the dead, fraudulent discounts, duplicate invoices. This number asks how much money a council recovers for each flagged case it actually investigates — its "yield per investigation". A high number suggests investigators are chasing the right leads; a low one may mean effort spent on false alarms. No per-council data on either ingredient is published, so this can only be built from bespoke requests.
| name | nfi_hit_rate |
|---|---|
| dataset | 08-national-fraud-initiative |
| kind | prospective |
| type | numeric (ratio) |
| unit | £ outcomes per investigated match |
| grain | council, per biennial NFI exercise |
| role | derived yield metric |
| source | derived — nfi_outcomes_gbp / nfi_matches_investigated; both inputs require bespoke collection (committee papers or PSFA/audit-body web-application extract) |
| period | biennial exercises |
| missingness | inherits the very high missingness of both inputs; undefined where nfi_matches_investigated is zero or unknown |
| score_note | Best targeting-quality signal if inputs ever materialise; inherits both inputs' gaps. |
| peer_group | council class (district / county / unitary / metropolitan / London) — not yet scored on the leaderboard |
The average outcome value yielded per match a council investigated — a measure of how productive its NFI follow-up work was. It would help separate targeting quality from sheer effort. No per-council inputs are centrally published, so it is entirely prospective, and it inherits the actual-vs-estimated outcome mix plus format and period inconsistencies across councils' committee papers.