Empirical research has shown that hospitals have higher than average l
everage, the extent of leverage is related to the extent of cost-based
reimbursement, and not-for-profit hospitals are not as highly levered
as their for-profit counterparts. Previous theoretical work does not
unambiguously predict all of these results. This article develops a mo
del of reimbursement policy and proves that, if reimbursement follows
the indicated pattern, the Modigliani-Miller capital structure irrelev
ance theorem fails to hold and, given reasonable restrictions upon the
utility functions of not-for-profit donor/investors, each of the empi
rically observed results would be expected ex ante.