A post by Atrios raises a question I’ve toyed with now and again for a few years. Minarchist libertarians by and large hold that crime prevention and the enforcement of rights are legitimate government functions and, by extension, legitimately funded through taxes of some sort. Redistribution, on the other hand, is held not to be part of the proper scope of government.

But what if it turned out that the marginal tax dollar could prevent more crime through, say, job training programs, or even direct payouts? Sure, there are a host of familiar pragmatic reasons to question the efficacy of these sorts of programs, and to worry about the institutional momentum they create. But prisons and police are institutions too, and it’s not prima facie clear that they don’t exhibit at least some of the same tendencies. So let’s stipulate for a moment that we’ve found some sort of redistributive programs that do seem to work pretty well at lowering crime rates and that it’s politically possible to implement these, and only these, programs. In this hypothetical, is there any principled libertarian objection to going this route? It’s hard to see what it would be. Rights protection, after all, is an end, to which police and prisons are only one means. And wouldn’t it be irresponsible, itself a wrong, for the state to spend coercively gathered tax revenue in anything other than the most efficient way?