Collaboration is a commonly prescribed method of public service improvement. If collaboration fails, blame is typically ascribed to transaction costs, organizational inertia, or premature evaluation. Drawing on a notable case of collaborative failure in England, however, we show that misdiagnosing public service problems as being of a type likely to be cured by joint working also generates poor results, and belongs conceptually prior to many “go-to” explanations of failure. Using stacked difference-in-difference estimators on 11 years of performance data relating to subnational tax collection, we show that inter- municipal cooperation produced no cost or quality improvements over independent service delivery. Supplementary testing attributes this less to governance problems, inertia or precipitate evaluation, than to a basic lack of interdependence – the specific “problem” to which collaboration is the “solution” – between large councils. Having exhausted scale economies internally, partners experienced no mutual reliance warranting their attempt to further economize through collaboration.