Yesterday at the Colloquium on High-End Inequality, Alan Viard from the American Enterprise Institute discussed portions of his recent book (with Robert Carroll), Progressive Consumption Taxation: The X-TaxRevisited.
Here are summaries of my thoughts regarding the first two topics that I aimed to discuss. As it happened, we never got to Topic 2 (which shows that we had plenty to talk about).
1. For / against the progressive consumption tax
My two colleagues in leading the colloquium discussion both support shifting from the existing income tax to a progressive consumption tax – but for very different reasons! I had been thinking that their reasons for supporting it were wholly inconsistent – but upon reflection (and discussion) have concluded that they are instead merely entirely distinct, and potentially complementary.
Alan Viard: He starts from assuming that we need a general revenue system founded on ability to pay, and that this might be either an income tax or a consumption tax. Both burden work, but income taxation also burdens future consumption relative to current consumption. It thereby inefficiently deters saving. Plus, in practice enormous complexity (and undermining of objectives) results from the full playout of the realization requirement.
In evaluating a consumption tax, Viard emphasizes that, in principal, deferral of the tax (via saving instead of immediately consuming) does not reduce the expected liability in present value, so long as tax rates are constant.
To illustrate, say we have a flat 30% (tax-exclusive) consumption tax that will remain in place indefinitely. Adam and Barbara each consume $100K this year, so each pays $30K of current-year consumption tax, But Adam also has $100M of savings (Barbara has none).
A current-year-focused income tax advocate might ask: How can it be fair to charge both of them the same tax, when Barbara’s ability to pay is so much greater? But a consumption tax advocate might respond: Adam would have paid an additional $30M of tax if he had consumed all his savings this year. But by not doing so, he merely deferred it at the market interest rate, whatever that might be. Thus, the present value of the tax on this wealth is $30M whether he saves it for a year, a century, or a millennium. So the consumption tax does require him to pay more, as it should based on ability-to-pay – one simply needs a longer time frame to make the appropriate comparison. All we need to assume, to reach this conclusion, is a constant tax rate and realization at some point – or that indefinite deferral is effectively no better than realization at some point.
An income tax is likely to be more progressive than a consumption tax if they have the same rate structure, because higher-income individuals generally save a higher percentage of their resources than poorer individuals. But we can replace the lost progressivity from changing the tax base from income tax to consumption by making the new tax rates nominally higher and more progressive.
Again, all this is what I take to be the Viard view (although I agree with most of it); now on to the Robert Frank view.
Robert Frank: He questions the need for an ability-to-pay tax until we’ve exhausted all taxes on negative externalities. For example, we might meet at least a portion of our revenue needs by properly taxing pollution, carbon, congestion, etc.
But the taxes on negative externalities that Frank favors include a luxury tax for the high-end. The aim here is to address expenditure cascades, addressed in this paper, which we discussed at the High-End Inequality Colloquium on October 28.
In effect, he wants a high-end luxury tax, on the same grounds as pollution or congestion taxes. But since it’s a hopeless task to try to figure out all the items that ought to be on the luxury tax list, he proposes instead to start levying the tax at very high annual consumption levels.
This might even be an add-on to the current income tax, although it also could come out of a personal expenditure tax (PET) – i.e., an individual-level progressive consumption tax. The very high rates just at the top would presumably require thinking about multi-year inclusion of consumer durables such as a home, and one might also debate whether averaging should be allowed.
Could a general consumption tax at lower rate brackets be justified on externality grounds, rather than just under ability to pay principles? For Frank, the answer is potentially yes, but this would require a separate critique of keeping-up-with-the-Joneses style lateral positional wars via consumption.
Frank also posits that the luxury tax wouldn’t greatly reduce high-end labor supply, and does not posit adverse effects from unequal wealth-holding. His target is unequal current consumption, based on the analysis of expenditure cascades from the top on down.
My comments & concerns:
1) Bob Frank posits negative externalities from high-end inequality solely via the consumption pathway. But what if there are also negative externalities from high-end wealth-holding? Suppose that Wilkinson-Pickett adverse effects on health and on social gradient ills were found to be increased by wealth inequality alone, even controlling for current consumption levels.
2) As a matter of political economy, will progressive consumption tax rates (whether under the Bradford X-tax or the PET) be high enough? The political process sometimes over-focuses on marginal rates, without regard to the tax base or actual effective rates.
3) Is savings, reached by the income tax but not the consumption tax, a tag of ability? If so, then under standard optimal income tax theory, including it in the tax base can improve the available tradeoff between distributional gain and efficiency loss.
4) On more of a technical point, is the personal expenditure tax (PET) more administratively friendly to highish rates than the X-tax?
The conventional wisdom holds that, if you favor consumption taxation, then logically it’s at least against the grain to favor estate and gift and/or inheritance taxation. But I disagree.
The above logic is inescapable if one thinks of multi-generational family dynasties as if they were extremely long-lived individuals. But they aren’t – parent and child, bequestor and bequestee, are not in fact the same person.
Henry Simons famously argued for double-taxing gifts and bequests. No deduction to the donor, full inclusion for the done (leaving aside administrative arguments re. small items, and perhaps set of internal transactions within a household).
This was an argument about consumption, not income. And whether or not one buys his result, the logic of saying it triggered consumption by both was clearly correct. If I make a gift, including an altruistically minded bequest, then clearly I get utility from it, no less than if I had instead spent my $$ on vacation travel or a restaurant. (Accidental bequests are different, but there it’s efficient to tax them if we posit that they were precautionary saving, but the owner had no bequest motive and ended up not needing them precautionarily.)
So as a matter of judging individual welfare in ability-to-pay terms, if I give a $1M gift to my kids, this is $1M worth of consumption by me and by them. Double-taxing it would merely match the fact that there was double consumption.
Why wouldn’t we want to adopt the Simons solution? Not because it is logically wrong in defining ability to pay – but rather because we don’t like the policy outcome. As the literature discusses, there is an “altruistic externality” here. I valued my enjoyment of the kids’ consumption at $1M, but I didn’t value their enjoyment at a separate $1M. By contrast, if I bought a standard consumption item for $1M the total welfare gained would be just $1M.
This suggests tax-favoring gifts and bequests for policy reasons, relative to the Simons baseline. Taxing them “just once” is a salient, prominent, intuitive, and relatively simple way of getting to this result, but it doesn’t inherently get the incentive exactly right.
Now suppose we are worried about unequal wealth-holding when there is extreme high-end inequality. One might consider addressing this through a wealth tax, but one is in the ballpark if one does this periodically, such as by targeting gifts and bequests. So now, where there’s impact on high-end inequality, we might want to move towards taxing gifts and bequests less favorably than under the “tax it once” baseline.
True, in that scenario it’s really just a mechanism for a periodic wealth tax, with the tax on gifts needing to accompany that on bequests so it can’t be avoided via transfers to younger or healthier individuals. Basing it on gifts and bequests as such would merely be opportunistic. But now suppose that one’s high-end inequality concerns are triggered by dynastic wealth transmission, unequal opportunities in life, etc. Then the gratuitous transfer, not the wealth-holding, might indeed be worthy of direct focus.
My point here is not to sketch out the policy upshot, but just to suggest that (a) there is really no logical tension between favoring enactment of a progressive consumption tax to replace the income tax, and wanting to tax gifts and bequests, plus (b) if one is uneasy about the prospects for a progressive consumption tax to be progressive enough, a tax on gifts and bequests may help to address that problem.