Introduction: This article examines the incentive effects of delegating operational and financial decision making from central government to local healthcare providers. It addresses the economic consequences of a contemporary policy initiative in the English National Health Service (NHS) - earned autonomy. This policy entails awarding operational autonomy to 'front-line' organisations that are assessed to be meeting national performance targets. In doing so, it introduces new types of incentives into the healthcare system, changes the nature of established agency relationships and represents a novel approach to performance management. Methods: Theoretical elements of a principal-agent model are used to examine the impact of decentralisation in the context of the results of an empirical study that elicited the perceptions of senior hospital managers regarding the incentive effects of earned autonomy. A multi-method approach was adopted. In order to capture the breadth of policy impact, we conducted a national postal questionnaire survey of all Chief Executives in acute-care hospital Trusts in England (n = 173). To provide added depth and richness to our understanding of the impact and incentive effects of earned autonomy at an organisational level, we interviewed senior managers in a purposeful sample of eight acute-care hospital Trusts. Results: This theoretical framework and our empirical work suggest that some aspects of the earned autonomy as currently implemented in the NHS serve to weaken the potential incentive effect of decentralisation. In particular, the nature of the freedoms is such that many senior managers do not view autonomy as a particularly valuable prize. This suggests that incentives associated with the policy will be insufficiently powerful to motivate providers to deliver better performance. We also found that principal commitment may be a problem in the NHS. Some hospital managers reported that they already enjoyed a large degree of autonomy, regardless of their current performance ratings. We also found evidence that the objectives of providers may differ from those of both the central government and local purchasers. There is, therefore, a risk that granting greater autonomy will allow providers to pursue their own objectives which, whilst not self-serving, may still jeopardise the achievement of strategic goals. Conclusion: It is apparent that the design and implementation features of decentralising policies such as earned autonomy require careful attention if an optimal balance is to be struck between central oversight and local autonomy in the delivery of healthcare.