Last week there were two disturbing reports about Medicaid, a program of health insurance for the poor that is mostly managed by the states and jointly paid for by the federal and state governments. The Congressional Budget Office predicted that states with a large number of poor people would not expand their Medicaid programs as required by the health care reform law now that the Supreme Court had made expansion optional. And a Harvard study unrelated to the court decision made it clear that a failure to expand Medicaid would likely doom thousands of low-income people to death or poor health. Revising earlier estimates to take account of the decision, the budget office said that making expanded assistance optional could leave three million more people uninsured in 2022, saving the federal government $84 billion through 2022 because it would not have to subsidize their coverage. The analysts made no effort to predict which states would or would not expand Medicaid. Instead, they looked at various factors that might influence the states’ decisions and predicted that some would not respond to even the extremely generous matching money that the reform law provided. They forecast that some states that already had trouble paying for their existing Medicaid programs would put off expanding Medicaid until their economies improved, or would expand the program partially, not to the extent originally required. Other states may shun the program entirely, as some Republican governors have vowed to do. By 2022, the budget office projects, only two-thirds of those who would become newly eligible for Medicaid if all states expanded to the levels sought by the reform law will actually gain eligibility. One-third will not, either because their states refuse to expand Medicaid at all or expand it only partially. Leaving low-income people uninsured will almost certainly damage their health. A study by three researchers affiliated with the Harvard School of Public Health, published by The New England Journal of Medicine, compared three states (New York, Arizona and Maine) that had already expanded coverage of childless or disabled adults with four neighboring states (Pennsylvania, New Mexico, Nevada and New Hampshire) that had not. Deaths among people ages 20 to 64 dropped in the three expansion states by about 1,500 a year, adjusted for population growth, whereas death rates in the comparison states went up. Expansion also brought a 21 percent reduction in cost-related delays in getting care. Some critics, mostly conservative Republicans, contend that Medicaid does not improve the health of beneficiaries and may even harm them. The new study should lay that canard to rest. State officials who want to save money by not expanding Medicaid will be harming their most vulnerable residents, and will most likely shift the cost of any emergency care they need to safety net institutions, taxpayers and charities.