This paper develops a simple model of payment incentives and empirically evaluates provider payment reform in Hainan Province, China. We use a pre-post study design with a control group to analyse two years of claims data to assess the impact of a January 1997 change to prospective payment for a sub-sample of the hospitals. This difference-in-difference empirical strategy allows us to isolate the supply-side payment reform effects from demand-side changes, in contrast with previous studies of China's reforms. Our results validate the theory that Chinese providers' behavioural response to payment incentives is similar to that reported in the literature derived from the experience of industrialized countries. We find that prepayment is associated with a slower rate of growth of overall expenditures, programme spending and patient co-payments per inpatient admission, compared to fee-for-service (FFS). These findings suggest cautious optimism regarding the effectiveness of prospective payment for controlling costs and should be encouraging for policymakers in developing and transitional economies considering replacement of FFS with more aggregated forms of provider payment.