Legal Commentary

Does federalism promote policy experimentation? Thirty years ago, Susan Rose-Ackerman offered a now-famous riposte to the idea that states are laboratories of democracy. According to Rose-Ackerman, states suffer from a free-rider problem: A governor or mayor bears the political risk that their novel policy experiment will be a flop, but other politicians who stand on the sidelines and wait to see how the idea fares can reap the political rewards simply by copying the idea if the policy turns out to be a success. As I have noted in an earlier post, Brian Galle and Joseph Leahy have a smart refinement of Rose-Ackerman's basic insight, noting that subnational politicians' ambitions for higher office might not induce them to take the risks of being a policy pioneer because, citizen ignorance about politics being what it is, the average voter will not be able to tell who was the first mover when credit is being claimed. One might, therefore, expect subnational politicians to hang back rather than experiment, waiting for neighboring politicos to test new policies.

My casual empiricism suggests that this is not an accurate account of how subnational politics operates. But the question of whether states are decent labs o' democracy, in any case, is: "Compared to what?" The issue is not whether federalism is a perfect laboratory but rather whether federalism, imperfect as it is, outperforms the unitary state. Assuming that governors and mayors have incentives to hold back with the experiments, they might still outperform federal agencies and Congress if the federal officials are even more paralyzed by fear.

The recent fracas over whether the U.S. Department of Agriculture should allow New York City to launch a pilot program barring the use of Supplemental Nutrition Assistance Program (SNAP) benefits to purchase sugary soft drinks provides an excellent illustration of the following hypothesis: However cowardly and risk-averse subnational politicians might be, members of Congress and federal agencies might be even more craven.

The central uncertainty with SNAP and soft drinks is whether, by limiting recipients' entitlement to buy unhealthy beverages, one will reduce the high levels of obesity and diabetes among indigent households. There is no doubt that low-income households are disproportionately subject to both ailments, and there is no doubt that sugary soft drinks are a major culprit in these health hazards. But opponents of New York City's proposal argue that, by barring food stamps from being used to purchase soft drinks, one will simply stigmatize the recipients of food stamps without making much of a dent in their consumption of unhealthy drinks.

Maybe the opponents are right. But the claim that restricting food stamp purchases will produce substantial stigma strikes me as, at best, a hypothesis in need of confirmation. Food stamps already are ineligible for use in purchasing non-food items, alcoholic beverages, tobacco, and prepared foods. Therefore, food-stamp recipients already have to use cash along with their Electronic Benefit Transfer (EBT) card to purchase a normal mix of groceries and personal items. Why exactly will stigma of using an EBT card increase because users have to put their Coca Cola in the "cash" pile on the checkout counter along with beer, Stouffer's frozen dinner, and Kleenex? On the other hand, if the restriction reduces the consumption of sugary soft drinks, then it could save lives and avoid diabetes-induced amputations. Of course, food education might also change eating habits, but, given the difficulty of changing eating habits, it is not obvious that education will work better than simply restricting the use of federal benefits.

In short, the question is an empirical question for which we need data rather than rhetoric. But my bet is that, despite having significant discretion to authorize a waiver, the Department of Agriculture will not permit New York's policy experiment to go forward.

Why not? Here's a hypothesis. Perhaps both incumbent congresspersons and agency specialists are allergic to controversy, and perhaps controversy is endemic to important policy experiments launched at the national level. Consider, as an example, New York City's anti-soft drink policy: In Washington, it attracts empty libertarian or egalitarian slogans like a dead cat attracts bluebottles. As Jonathan Adler put the question, "is this an example of over-weening paternalism? Or is it a responsible limitation on government assistance?" Of course, framed in this manner, the policy debate is an immediate call to arms for social conservatives, libertarians, and egalitarians to battle it out on the high plains of abstraction. Highlight the "government assistance" aspect of the case, and egalitarians immediately smell an assault on the dignity of the indigent -- that is, an implicit claim that poor people make poor food choices. Libertarians, by contrast, are incensed by a claim of entitlement to what they regard as governmental largesse -- charity rather than private freedom. We are accordingly treated to tired rhetoric when we need fresh data.

Bloomberg can sidestep these controversies, because he rules a one-party town where libertarian slogans have little purchase. Everyone knows that Bloomberg (who makes war on everything from public smoking to trans-fats) is motivated by technical health concerns, not bash-the-poor rhetoric. A meddlesome nuisance he may be, but an indigent-bashing Tea Partier, he ain't. So he can go forward with a policy experiment free from the paralyzing cant that generally freezes Washington, D.C. into a perpetual battle of slogans worthy of bumper stickers and C-Span. Put more generally, political homogeneity increases rather than decreases the power of politicians to sidestep gridlock-inducing controversy. Increase the number of antagonistic factions, and Madison's Nightmare haunts one's efforts to launch any controlled experiment.

Of course, this theory about the incapacity of federal officials to authorize experimentation is just a hypothesis. Maybe I am overlooking a plethora of exciting policy experiments being launched by field offices of federal agencies throughout the nation. In any case, the Rose-Ackerman/Galle-Leahy thesis will only be persuasive when likes are compared to likes -- when subnational politicians who are allegedly free-riding off of each others' experiments are compared to shell-shocked federal bureaucrats huddled in their agencies just trying to avoid the latest salvo in the national Culture Wars. Assuming that the Department of Agriculture does what I (sadly) predict and declines Bloomberg's petition for a waiver, we will have one data point on the question of relative institutional competence to experiment: On an issue where the ratio of rhetoric to data favors the former overwhelmingly, the feds turned down a chance to get a little more of the latter.

In short, these waiver fights are one laboratory from which we can all learn a great deal about the relative merits of federalism and Leviathan.