Central government in the UK has introduced performance management regimes that apply rewards and sanctions to local service providers. These regimes assume that organizational performance is attributable to decisions made by local policy-makers rather than circumstances beyond their control. We test this assumption by developing a statistical model of external constraints on service standards and applying this model to the outcomes of comprehensive performance assessment (CPA) in English local government. The results show that CPA scores were significantly influenced by the characteristics - such as social diversity and economic prosperity - of local populations. Thus 'poor' performance is partly attributable to difficult circumstances rather than bad choices.