Countries seeking to move away from a purely fee-for-service (FFS) system may consider a hybrid approach whereby only some procedures are paid by FFS while others are paid prospectively. Yet little evidence exists whether such a hybrid payment system contains overall costs without adverse influences on health outcomes. In 2003, Japan experienced a reform from FFS to a hybrid payment system in which only some inpatient procedures were paid prospectively. We exploit this reform to test how such a hybrid system affects overall costs and health outcomes. Briefly, we find that healthcare providers responded opportunistically to the reform, moving some procedures out of the bundled inpatient setting to FFS services, leading to no reduction in cost. There was some evidence of a moderate deterioration in health outcomes, in terms of a decline in the probability of symptoms being cured at discharge. In sum, our results suggest that in some cases, a hybrid payment system can be non-superior to either FFS or prospective payment system. K E Y W O R D S bundled payment, hybrid payment system, substitution behavior, price regulation 1 | INTRODUCTION Since 2000, global health spending has increased at an annual rate of 4.0%, about 1.5 times the 2.8% annual growth rate of the global economy (Xu et al., 2018). To contain costs, many countries have opted to replace traditional Fee-For-Service (FFS) payments with a prospective payment system (PPS), incentivizing providers to reduce costs by avoiding unnecessary services (Carroll, Chernew, Fendrick, Thompson, & Rose, 2018; Ho & Pakes, 2014). However, providers could also be incentivized to avoid unprofitable patients and discharge patients inappropriately early (Ellis, 1998; Gilman, 2000). Such behaviors may reduce costs at the expense of population health; for example, Cutler (1995) finds higher mortality and readmission rates at hospitals paid by PPS. Consequently, an increasing number of countries have experimented with hybridizing the two payment systems to preserve the advantages and mitigate the disadvantages of each (Ellis & McGuire, 1986; Ma, 1994; Robinson, 2001). The PPS nature of a hybrid payment system can blunt the excessive financial incentives of FFS and contain costs. The FFS nature, in turn, can mitigate the unintended incentives of PPS and prevent undertreatments. Empirical evidence is mixed regarding whether a hybrid payment system is better at containing costs than a pure FFS. A few experimental studies show that capitation-FFS hybrid systems have smaller deviations from the optimal level of treatments compared