Suppose I'm a utilitarian convinced that human consumption of meat causes a huge amount of animal suffering. And suppose I love meat, and giving it up would leave me worse off. I would happily comply with a no-meat-eating rule if I thought others would likewise comply. But in the absence of a mechanism (whether internal/moral or external/political) to enforce compliance, I rationally believe that my compliance with the no-meating-eating rule will have zero effect on market demand for meat. And suppose I rationally believe my heeding the rule will only make me worse off while making no animals better off. In that case it is perfectly rational to continue to eat meat even if I believe that it would be immoral to eat meat under conditions of general compliance with utility-maximizing rules. I think Matt's voluntary taxpayer case is exactly analogous.
Matt seems to think there's something significant about the fact that Americans contribute lots of money to charity, but I can't quite see what it is.
Unlike Will, I have a hard time coming up with a model in which it makes sense for Warren Buffett to refuse to contribute more to the government unless the rest of us also are forced to contribute more. In order to come up with such a model, I would have to assume some extreme "lumpiness" of public goods. That is, at the current level of government revenues, the marginal dollar produces no public goods, but a huge "lump" of new dollars would create a threshold effect that suddenly would produce a lot of public goods. That model strikes me as totally unrealistic relative to where we are today.
Also, I think that the fact that Americans contribute to charity is relevant. It shows that coercion is not necessary in order to get people to pay for public goods. An interesting question is whether at the margin we get more public goods (including income redistribution) from charitable contributions than from government.
Mark Thoma links to a post from Ed Dolan, who complains that religious charities do not do as much redistribution as one would like. But neither does government, if what you want is redistribution from wealthy to poor (it redistributes a lot from young to old and from poorly-connected to well-connected).
My personal preference would be to err on the side of letting each individual decide how much to give to charity and which charity to support. The alternative is confiscating money through the democratic process.
I think it is difficult to evaluate what happens to a marginal dollar that is confiscated or donated. I tend to believe that a marginal dollar donated will do more social good.
(As you know, I am not so high on non-profits. It would not surprise me if it turns out that a marginal dollar of Warren Buffett's would do the most social good if given neither to a non-profit nor government. It might be best invested in a profit-seeking firm. But you don't have to agree with me on that.)
I think it is worth imagining a world in which government competes on a level playing field with other charities. That is, imagine a world in which government relied on voluntary donations. In such a world, government would be smaller and other providers of public goods would be larger. To me, that sounds like a win-win.
I remember the old bumper sticker that says "It will be a great day when we spend billions on education and the Pentagon has to hold a bake sale go buy a bomber." I think it will be a great day when the government has to hold a bake sale to buy anything.
It is easy to come up with a model in which in theory voluntary contributions do not produce enough public goods. But in practice, I am not convinced that taxation produces an attractive mix of public goods. At the margin, if we had less taxation and more charitable contributions, might we be better off?