Alan Auerbach's Wall Street Journal op-ed (subscription required) concerning the consumption tax option in fundamental tax reform is, of course, excellent and well worth reading. But his analysis reflects some underlying assumptions that I have for a long time questioned.
Though not a consumption tax advocate (his aim is simply to explain the main dimensions of the choice), Alan cites studies from his academic work suggesting that a consumption tax could increase GDP, in the long run, by as much as 9 percent (though perhaps only half as much - still not a trivial gain - if current progressivity is maintained).
Although I am more of a consumption tax advocate (assuming retained progressivity) than he is, I question these studies due to an important underlying assumption that he is entirely straightforward about. A key component of the GDP (and efficiency) gain that he finds is the wealth levy on holders of existing assets that would arise upon enactment without transition relief. Since this is ostensibly a one-time-only wealth tax, he models it as lump-sum (i.e., as having no effect on future behavior because people do not expect their wealth to be taken again).
My objection, spelled out at length in my book When Rules Change, is twofold. First, an ostensibly one-time wealth levy will not necessarily be viewed as such, since it may show a political predilection or at least willingness to use wealth levies, hence potentially affecting people's expectations, given in particular that its logic is infinitely repeatable. (Take wealth once with the solemn promise never to do it again, then repeat complete with fresh promise.) This is not necessarily to condemn the wealth levy, but to suggest that it might have similar distortionary effects to a standing wealth tax. Who is right on this depends on how people's expectations actually end up being affected, a tricky empirical question that would likely depend on the context. Alan's view, presumably, is that the wealth levy's being a kind of byproduct of the dramatic change in tax base makes the claim that it is being levied once and once only more politically credible. Certainly not a silly view, but consider that rate increases in a consumption tax can have the same wealth levy effect as introducing the tax (which is itself merely a rate increase from 0% to whatever is the starting rate or set of rates). And consider that the plea for transition relief would presumably have been made and consciously rejected by Congress, adding to the likelihood of its affecting expectations.
Second, I think of the wealth levy as conceptually and practically distinct from the shift from an income tax to a consumption tax. As I discuss in When Rules Change, it results, not from the change in tax base itself, but from the change in what I call "accounting methods." That is, it results from shifting from a system with income tax style accounting, where certain expenses are capitalized and allowed only over time, to a system with cash flow accounting or expensing of all business-related outlays. Yet one could have a consumption tax with income tax-style accounting (deferred deductions but interest on basis to make them expensing-equivalent in present value). Indeed, David Bradford actually proposed this in his later X-tax refinements because he didn't like the transition hit that Alan relies on. In theory, moreover, though making it work would be tricky, one could have an income tax with expensing-style accounting (e.g., you are allowed up front the present value of the entire future stream of deductions for economic depreciation).
If you like the ostensibly one-time capital levy, you could have it without switching tax systems. As Bradford showed, wiping out income tax basis, once and once only, while otherwise retaining the current system would impose a comparable transition hit, although admittedly it might be trickier to make people believe the "once and once only" claim. Or we could enact, say, a one-time 25% wealth tax, plus a constitutional amendment to stop us from doing it again.
At the least, the merits of the wealth levy seem to me analytically separate from the income to consumption tax change, since it results from a conceptually distinct "accounting" change and could be done without the broader change (or not done while doing the broader change).
The 9% GDP growth finding must therefore be viewed in the context of its including the impact of a wealth levy that is distinct from the change of tax base as such (even if brought about thereby), and that is assumed to be credible to prospective investors as a one-time-only event.
Alan and I have discussed this issue in the past, and I think we agree about the basic analytics. Perhaps we will have a chance to discuss it further. For those who follow the NYU Tax Policy Colloquium, which I co-taught with David Bradford until the tragedy last February, I will be co-teaching it with Alan in winter/spring 2006, which he will spend at NYU as a visitor.