In the past few months, the Senate has made several attempts to extend the Medicaid bailout included in the stimulus package. States share the cost of Medicaid with the federal government, and because of the influx of new beneficiaries due to increased job loss, the federal government increased the portion of the cost that it would cover.

This was bad policy. Congress should avoid another bailout that would treat low-income Americans inequitably from state to state and further delay reforms to increase the fiscal sustainability of Medicaid.

Recent analysis by Heritage’s Brian Blase shows that when Washington increases Medicaid matching rates, aid is unfairly distributed. Richer states with more lavish Medicaid programs receive more assistance.

According to Blase’s analysis, of Americans living in poverty, 6.6 percent live in New York, yet New York received 15 percent of bailout money in the first stimulus. Conversely, Georgia is home to 3.6 percent of Americans living in poverty but received only 2 percent of federal assistance.

Blase writes, “Additional federal funds will have the effect of punishing states that essentially lived within their budgets while rewarding states that were more reckless or were counting on further federal dollars. Congress should stop subsidizing states that spend irresponsibly, because it rewards poorly managed state programs and further incentivizes mismanagement and fiscal disarray.”

At the same time that it encourages bad behavior, bailing out Medicaid delays the needed reform to put Medicaid on a fiscal path states can afford—even during a recession. Reform should be a policy priority at both the state and federal level, especially now. Blase explains that “without reform to the federal subsidization of Medicaid, Obamacare—which adds 20 million individuals to Medicaid—will exacerbate Medicaid’s budgetary pressures.”

Congress should encourage states to get serious about Medicaid reform, but bailing out the program does the exact opposite. Blase’s analysis and argument for this can be read in full here.