Josh Barro notes that Mitt Romney’s approach to Medicaid reform has changed considerably between 2002, when he was running for governor of Massachusetts, and his 2012 presidential campaign:
Mitt Romney’s 2012 campaign platform calls for converting those federal payments into block grants—a state would get a fixed amount of money, regardless of how much it actually spent on Medicaid—and then capping the growth of those block grants at CPI plus 1 percent. Because health care costs tend to rise faster than CPI + 1, this would mean that, over time, states would have to pay a larger share of Medicaid costs.
But in 2002, Romney’s approach was nearly the opposite. In a detailed PowerPoint presentation on health care reforms, Romney called on the federal government to flatten its Medicaid matching practices and pick up 77 percent of the tab for every state, including Massachusetts. At the time, he noted this would have saved Massachusetts taxpayers $1.7 billion a year.
But this savings to state taxpayers is a cost to federal taxpayers. Far from controlling federal Medicaid costs, Romney’s 2002 proposal would have meant increasing federal spending on the program by 35 percent. At today’s Medicaid spending levels, that would mean an annual increase in federal Medicaid costs of $96 billion a year.
Josh recommends an approach somewhere in between Romney’s 2002 and 2012 proposals; he believes that by requiring states to pick up 100 percent of costs at the margin, block grants encourage cost control. Yet he also believes that the federal contribution should be relatively generous, partly to reflect the fact that the federal government has more fiscal flexibility than states (i.e., it can borrow more easily during downturns).
My view is somewhat different. As we’ve discussed in this space on numerous occasions, I think that the federal government should operate and fund the Medicaid program exclusively, relieving state governments of a significant and growing burden. At the same time, the federal government should devolve other responsibilities, e.g., over most transportation functions, K-12 education, and social welfare programs apart from EITC, to state governments. The best case against this approach is that state governments have a significant advantage in operating Medicaid programs that are well-tailored to local conditions. My assumption is that a federalized Medicaid program would be somewhat less generous than the most generous state Medicaid programs and somewhat more generous than the least generous state Medicaid programs. A unified program might be somewhat less susceptible to beggar-thy-neighbor dynamics in state governments that run cost-effective programs are actually punished for their success. In a similar vein, focusing federal anti-poverty efforts on a unified cash transfer program like EITC limits the potential for mischief while giving state governments wide autonomy to pursue more specialized interventions that are tailored to meet local conditions.
Ultimately, the goal is to have states compete in domains where competition leads to a race to the top (quality of infrastructure, K-12 education, and the local environment, etc.) while having the federal government dampen subsidy-chasing competition by taking a larger role in other domains. The end result would in theory be a somewhat larger federal government in the short- to medium-term coupled with lower overall spending levels, factoring in state and local spending. By fixating on federal spending, we miss the extremely rapid expansion of state and local spending.
P.S. Josh Barro disagrees:
There’s a reason that social welfare programs are often funded in part by the federal government, but operated by state and local governments: states and localities have a social services apparatus in place, and the federal government does not.
Because Medicaid is means tested, you need to monitor participants for their continued eligibility. Medicaid beneficiaries also often have many simultaneous social services needs, and it’s helpful to have one caseworker handling them all together, especially when your goal is to get beneficiaries to a point where they don’t need welfare anymore.
In other words, Medicaid is about much more than making payments to doctors and hospitals. The federal government could build an apparatus to do this, but it would be duplicative, and you would lose the benefit of integration with other, state-run assistance programs.
It might make sense to federalize the aged and disabled components of Medicaid, which closely resemble Medicare and often serve the same beneficiaries. But for the portion of the program that serves low-income adults and children (a bit less than half its cost) we’re better off retaining the state-local structure and using a well-designed block grant to improve fiscal incentives.
I’d be open to federalizing the aged and disabled components of Medicaid; my sense is that the remainder of the program would function more effectively were we to do so.