We use the cost models to calculate capitation payment as the cost predicted for a patient given their age, gender, deprivation and morbidity but removing the effect of the patient’s practice and replacing it by the average of the practice effects. Capitation payments, at patient and at practice level, are sensitive to the choice of estimation method and morbidity measure. We also find that the difference between average cost and capitation for some types of patient is often substantial, so that there are incentives for patient selection, though less than when capitation is based only on age and gender.