Every English council must publish every payment it makes over £500 — the raw chequebook. This variable counts payments that look suspiciously alike: same supplier, same amount, dates within a few days of each other. Genuinely paying an invoice twice is one of the most common and most avoidable ways councils lose money, so a high count of these look-alike pairs is a direct flag for weak financial controls.
| name | duplicate_payment_candidates |
|---|---|
| dataset | 11-transparency-code-spend |
| kind | constructed |
| type | numeric |
| unit | count and GBP value |
| grain | council-month |
| role | feature |
| source | count/value of transaction pairs with same supplier, same amount_gbp, payment_date within N days (e.g. 7), and — where available — different transaction_id |
| period | per council-month (or pooled months) |
| missingness | Requires prior schema normalisation; three-council sample only. Sefton has no transaction_id, so its candidates cannot be de-duplicated against the ledger and will over-count. |
| score_note | Direct avoidable-cost screen unique to transaction-level spend data; Sefton over-counts. |
| peer_group | council class (district / county / unitary / metropolitan / London) — not yet scored on the leaderboard |
The count and value of near-identical payment pairs (same supplier, same amount, dates within a window) that could indicate duplicate payments. These are candidates only: recurring identical invoices — rent, care packages, like the Sefton £511.00/£511.09 same-day pairs — are legitimate. Where a transaction ID exists it screens out same-ledger-entry matches; Sefton lacks one, so its figure is an upper bound.