Project Liberty permitted rapid expansion of the total volume of services provided by community-based organizations without interfering with the provision of traditional services, although a modest effect was seen for smaller agencies. Although the results do not imply that "supply side" planning for disaster needs would not improve system response, they do imply that implementation of flexible "demand side" financing can call forth a large volume of new services rapidly and without interfering with other community services.

This paper tests whether capitated payments to Medicaid managed care plans induce to plans' strategic undercutting of treatment for specific diagnostic groups. I focus on treatment (measured by length of stay and cost) in acute care hospitals in Massachusetts. I use a "differences-in-differences-in-differences" approach, where the third differences compare treatment patterns between managed care plans that receive capitated payments with those that do not. I find that the first reduce treatment significantly more to mental health patients than to patients in other disease groups, whereas the latter reduce hospital resource use more uniformly across disease groups. These results highlight the importance of using payment mechanisms in public programs that reflect the variability in costs of beneficiaries.