Some libertarians argue that the state should provide a minimum basic income, mainly because this approach would be preferable to the costly and bureaucratic amalgamation of redistribution programs that currently exist.

It’s hard to disagree with the notion that the current system is a failure. The Cato Institute’s Michael Tanner has produced a searing indictment of the modern welfare state, pointing out that more than $1 trillion is spent every year on redistribution programs for the ostensible purpose of alleviating economic hardship, yet (or more likely as a result) the poverty rate is at an all-time high. Perhaps one reason poverty remains high is that such programs make leisure more attractive than work, as painstakingly illustrated in a study produced by Tanner and Charles Hughes.

Moreover, welfare programs create very high implicit marginal tax rates, making it very difficult for poor people to improve their living standards by engaging in additional productive behavior. It’s almost as if the system was designed to create permanent dependency.

So it’s very understandable that well-meaning folks want to dismantle the current approach and try something new. The basic income certainly has the merit of being a big, bold idea. And advocates of the basic income can cite some very iconic libertarian figures who support at least some version of their approach, including Milton Friedman, Friedrich Hayek, and Charles Murray.

But if we’re going to launch a very difficult battle to radically reshape well-entrenched bureaucracies and long-standing policies in DC, it seems reasonable to ask whether there might be a better road to Rome.

There are several very practical and compelling reasons to be skeptical about the basic income proposal.

The first concern is that a guaranteed basic income may simply replace one form of government-financed dependency for another. An economy’s output is a function of the quality and quantity of labor and capital that are productively deployed. That’s an argument for reducing the size of welfare state in order to encourage more work. But it’s not an argument for creating a lump-sum form of redistribution.

That being said, a well-designed basic income presumably will not have the punitive implicit tax rates on those who decide to engage in productive behavior. Moreover, the army of bureaucrats who monitor and oversee the current system wouldn’t be necessary (though skeptics understandably wonder whether politicians actually would shrink the federal workforce even if redistribution programs were axed).

The second concern revolves around the question of durability. More specifically, there’s no guarantee that politicians in the future won’t re-create the current panoply of programs that the basic income is supposed to replace. The political incentives that existed when those programs were first created, after all, presumably would remain in a world where people get a pre-determined amount of cash from the government.

This is largely speculative, to be sure. Interest groups would have an incentive to lobby for the creation of new programs with targeted beneficiaries. But perhaps the general public would understand that even one new handout would be the first step on a slippery slope leading back to something akin to the current system.

A third concern is that the basic income may become overly generous, particularly in a fiscal system where a relatively small slice of the population pays an overwhelming share of the tax burden. In that kind of regime, there would be an incentive for a majority of voters to increase the amount of redistribution.

It’s unknown whether this would happen. Simply stated, the few experiments that might teach us about incentives with a system of basic income have been too limited to draw any firm conclusions. For those who want to review that literature, Jim Manzi has a good summary showing less-than-robust evidence for a basic income and Megan McArdle raises some very practical objections to the concept. Proponents of the basic income surely would respond by arguing that voters would understand that they were taxing themselves and thus be reluctant to make the basic income too large.

The final concern, at least for many libertarians, is that the federal government shouldn’t be in the business of redistributing income. The Constitution, for instance, does not list food stamps, welfare, housing subsidies, or Medicaid as enumerated powers of Congress in Article I, Section VIII.

Supporters of the basic income surely would respond by stating that it’s too late to make that argument. Limits on the role of government largely evaporated during the 1930s and it’s pointless – even though perhaps intellectually satisfying – to make such arguments today.

The bottom line for advocates is that anything would be better than the current system, so why not try something new?

They’re right, but there’s actually a better way of approaching the issue. Why not take all income-redistribution programs, put them into a single block grant, and then transfer the money – and responsibility – to state governments? In an ideal world, the block grant would gradually diminish so that states would be responsible for both the collection and disbursement of all monies related to welfare.

But that’s a secondary issue. The main benefit of this federalist approach is that you stop the Washington-driven expansion of the welfare state and you trigger the creation of 50 separate experiments on how best to provide a safety net. Some states might choose a basic income. Others might retain something very similar to the current system. Others might try a workfare-based approach, while some could dream up new ideas that wouldn’t stand a chance in a one-size-fits-all system run out of Washington, DC.

And as states adopted different systems, they could learn from each other about what works and what doesn’t work. And since it’s easier to influence decisions that are closer to home, taxpayers at the state level almost certainly would have more ability to impact what happens with their money.

Moreover, there’s actually some evidence that this approach is practical. The 1996 welfare reform legislation isn’t a perfect analogy, but the core feature of that law was the elimination of a national entitlement and the provision of a block grant so that states could decide (with some strings attached) how to deal with poverty.

By most measures, the 1996 reform was a success, with Ron Haskins of the Brookings Institution explaining that it resulted in lower levels of welfare dependency and reductions in child poverty.

This isn’t to say the 1996 law was ideal. The advantage of a comprehensive federalist approach is that policy experts can push states to experiment with different policies. And given the vast differences between various American states, it’s almost guaranteed that there will be lots of diversity.

This diversity not only will inform policy makers about what works and what doesn’t work. It also will satisfy the libertarian desire to get Washington out of the business of income distribution, while presumably producing a system that actually does a better job of helping the less fortunate escape government dependency.

In other words, all the advantages of the basic income plan without the potential system-wide downsides.