Every English council must publish every payment it makes over £500 — the raw chequebook. This variable condenses that chequebook into a single concentration score (the Herfindahl–Hirschman Index, borrowed from competition economics): near 0 means spend is spread across many suppliers, near 1 means a handful dominate. Heavy dependence on a few suppliers can mean weak competition, higher prices, and real trouble if a dominant contractor fails.
| name | supplier_hhi |
|---|---|
| dataset | 11-transparency-code-spend |
| kind | constructed |
| type | numeric |
| unit | index 0–1 (or 0–10,000) |
| grain | council-month |
| role | feature |
| source | sum_i (spend_i / total_spend)^2 over suppliers i, from normalised supplier and amount_gbp |
| period | per council-month (or pooled months) |
| missingness | Computable only after schema normalisation and supplier-name cleaning across the three-council sample; sensitive to supplier-name normalisation quality, and non-comparable across councils with different coverage exclusions (e.g. Sefton omits staff payments and redacted social-care items). |
| score_note | Strong supplier-dependence signal, though sensitive to name normalisation quality. |
| peer_group | council class (district / county / unitary / metropolitan / London) — not yet scored on the leaderboard |
Herfindahl–Hirschman concentration of a council-month's spend across suppliers. Values near 1 mean spend is concentrated in a few suppliers, a dependence/negotiating-position signal rather than evidence of waste. Compare only within like service areas — a county's social-care block contracts mechanically concentrate spend — and only after supplier-name normalisation, since split name variants deflate the index.