Reducing socioeconomic health inequalities is a key goal of most health systems. When care providers are paid prospectively, e.g., by a fixed sum per patient, existing inequalities may be sustained by the incentives to undertreat relatively unhealthy patients. To counter these incentives, prospective payments are often risk-adjusted based on observable patient characteristics. Despite that risk adjustment (RA) is widely used, empirical evidence is lacking on how it affects the behavior of care providers. This paper provides such evidence using detailed administrative data from a Swedish region. We examine how a novel RA model applied to the prospective payment for primary care providers – capitation – affected socioeconomic differences in care utilization among individuals with a chronic condition. On average, the new RA model implied substantial increases of the capitation for patients with low socioeconomic status (SES). Yet, we do not find any robust evidence of greater access to primary care for individuals with low SES relative to individuals with high SES after the model was introduced. We find a small increase in hospital emergency department visits (a substitute to primary care), but no effects on hospitalizations. These results do not suggest that the new RA model reduced socioeconomic health inequalities. Our findings highlight that a risk-adjusted prospective payment may not by itself guide treatment decisions. We discuss other governance and management policies that may address undesired health inequalities.