Despite its salience as a regulatory tool to ensure the delivery of unprofitable medical services, cross-subsidization of services within hospital systems has been notoriously difficult to detect and quantify. We use repeated shocks to a profitable service in the market for hospital-based medical care to test for cross-subsidization of unprofitable services. Using patient-level data from general short-term hospitals in Arizona and Colorado before and after entry by cardiac specialty hospitals, we study how incumbent hospitals adjusted their provision of three uncontested services that are widely considered to be unprofitable. We estimate that the hospitals most exposed to entry reduced their provision of psychiatric, substance-abuse, and trauma care services at a rate of about one uncontested-service admission for every four cardiac admissions they stood to lose. Although entry by single-specialty hospitals may adversely affect the provision of unprofitable uncontested services, these findings warrant further evaluation of service-line cross-subsidization as a means to finance them.