From a recent and, yes, very dorky conversation on Medicaid policy:

Me: [Long rambling about how, in the absence of a drastic overhaul in health care, Medicaid needs to be expanded, not cut, and how one of the worst features of the program is the set of complex eligibility requirements that dissuade many otherwise eligible poor people from signing up.]

Person X: "So the best thing to do is to turn Medicaid into some kind of block grant program, and let states decide for themselves how they want to administer the program and what sorts of eligibility standards are best. That's the whole point of federalism!"

Me: "Argh, no, besides the fact that block grants almost always amount to cuts, basically, especially since most states aren't allowed to run deficits to pay for expanded welfare rolls during downturns, blah blah blah, states will also start pulling dirty tricks to lower their costs. Like offering only minimal insurance, or requiring high co-pays. And they'll do it because they're each competing with other states to offer crappier and crappier benefits so that no one gets stuck with a bunch of rent-seeking new immigrants. Federalism just means a race to the bottom."

X: "Oh, states don't do that."

Me: "Yes they do! Everyone knows..."

X: "No, I read a study on that back in college. Everyone thinks it's true but it's totally false."

And that argument sort of ended there, because who was I to argue with a study?" Luckily, back at home with the power of the internet, I can look this stuff up. The famous paper I was thinking of—but couldn't for the life of me recall—was "Welfare Magnets" by Paul Peterson and Mark Rom (1990) which suggested that between 1976 and 1989, neighboring states often lowered their benefits in competition with each other. Even more recently, though, Michael Bailey tweaked (pdf) that model—allowing for immigration between non-contiguous states and doing some other technical stuff—and concluded that the "race to the bottom" still held. States are pressured to discourage poor people from migrating by reducing benefits. Searching around some more, it's hard to find any countervailing studies, so I'd like to know what paper this fellow read (he's promised to try to find this fabled "paper").

Anyway: Medicaid block grants bad. And no, this isn't all I talk about in my spare time.