This paper employs a nationally representative, household-based dataset in order to test how the compensation method of both the specialists and the primary care providers affects surgery rates. After controlling for adverse selection, I find that when specialists are paid through a...