In addition to a list of newly posted papers, we’ll also include a write-up about at least one of the week’s additions. A quick administrative note: The best way to make sure we see your paper is to use the JEL code K34 (Tax Law) when you upload to SSRN. (If your reaction is “what’s a JEL code?”, check out Lea-Rachel Kosnik’s overview. And for a fascinating read on the tangled history of JEL codes, see Beatrice Cherrier’s article.)
(A point on terminology: By “accretion-based wealth tax,” Hasen does not mean a tax levied exclusively on accretions to wealth—i.e., changes in wealth. Rather, Hasen is referring to a wealth tax that “accretes” on wealth—i.e., is levied on the same base of wealth period after period.)
One of my favorite parts of Hasen’s paper is his section picking apart the argument that an income tax “double-taxes” savings—an argument often invoked by advocates of consumption taxation. Hasen uses the example of a taxpayer, A, who in period 1 earns $100 in wages that she can spend or invest at a 10% risk-free rate of return. Under a consumption tax at a 30% rate, A can purchase $70 of consumption today or $77 of consumption in the next period. With an income tax at a 30% rate, A can purchase $70 of consumption today or $74.90 of consumption in the next period. According to the familiar pro-consumption tax argument, $100 today is “equivalent” to $110 in the next period, and thus an income tax is non-neutral between these two values because it burdens the saver more than the immediate spender.
Hasen responds to this argument as follows:
The claim of equivalence derives . . . from the assumption of a precisely offsetting cost to putting off consumption until the future. That cost generally is understood to be the pain of deferral. In particular, the marginal saver is prepared to forgo current consumption at the price of 10 percent per year, which pain offsets the income earned on the saved amount. Note, however, that the pain of deferral is not particularly something to be accounted for under an income tax or, for that matter, under a consumption tax. Consider the pain of work (for those who do not enjoy working). No tax offset to realized personal services income is available for the pain of work under either an income or a consumption tax, nor does any claim of double taxation of work arise because the offset is not available. Analogously, the claim of double taxation of returns to savings seems inapt, inasmuch as income is received in exchange for the non-deductible pain of deferral.
A slightly different way to put the same point is: Individuals earn income by exchanging the disutility of labor for wages or by exchanging the disutility of deferral for a return on savings. A consumption tax (equivalent to a labor income tax) reaches only the former exchange; an income tax in its ideal form reaches both. In this light, it is a consumption tax rather than an income tax that appears to be “non-neutral”: it taxes one disutility-for-income exchange and not the other.
This is a clever argument for an income tax (at least in its ideal, mark-to-market form) over a consumption tax. How, though, does the argument for an income tax apply to a wealth tax? As Hasen explains:
An ideal accretion wealth tax is widely thought to differ from an ideal income tax only in that it applies to average rather than actual returns, because in other respects the taxes differ only in nominal rates. In particular, assuming in the simple case that returns to capital are uniform, a periodic wealth tax at a given rate is identical to an income tax at an appropriately higher rate. For example, if the risk-free rate of return is 10 percent, a capital income tax at 40 percent is the same as an accretion wealth tax at 3.64 percent. A taxpayer with wealth of 100 at the beginning of the period would have 10 of pre-tax income. Under the income tax, he would owe 4 on 10 of income, leaving 106 total; under the wealth tax, he would owe 3.64 percent of 110, which also is 4.
To be sure, as Hasen notes, an ideal income tax and an ideal wealth tax are not the same once one adds “real-world complexity” to the equation. Assuming that we do want to tax capital (or capital income), which of the two—a mark-to-market income tax or a periodic wealth tax—is preferable? Hasen argues in favor of a periodic wealth tax. I’ll focus here on a few of the distinctions Hasen mentions between a mark-to-market income tax and a periodic wealth tax, and why I think these distinctions weigh in favor of the former.
First, a mark-to-market income tax—but not a periodic wealth tax—reaches returns on capital that are consumed before the end of the tax period. If capital income is taxed, an individual who earns $100 in wages on January 1, saves at a 10% risk-free rate of return, and then withdraws her savings to consume on December 30 pays a tax on (almost) $10 of capital income (in addition to any tax on labor income). With a periodic wealth tax levied on assets as of December 31, however, the same individual would pay no wealth tax (though would still owe a tax on labor income). It’s hard to think of a reason why we would want our tax on capital to exempt individuals who consume their returns immediately.
Second, as Hasen notes, there is the issue of inframarginal returns. Imagine again a 10% risk-free rate of return and either a 40% capital income tax or a 3.64% wealth tax. As noted above, an individual who starts with $100 at the beginning of the tax period, invests at the risk-free rate, and does not consume the return pays a capital income tax of $4 or a wealth tax of $4. Now imagine instead that the individual identifies an investment with an inframarginal return of 20% rather than 10%. With a 40% tax on capital income, the individual pays a tax of $8. With a 3.64% tax on wealth, however, the individual pays a tax of only $4.37 (i.e., 3.64% times $120). Put differently, a 3.64% wealth tax is equivalent to a 40% income tax on risk-free returns and a 3.64% income tax on inframarginal returns. Why would we want to tax inframarginal returns (much) more lightly than risk-free returns?
Third, a wealth tax runs into constitutional complications, because the Sixteenth Amendment exempts income taxes but not wealth taxes from the requirement that “direct” taxes be apportioned among the states on the basis of population. Hasen suggests that we might address this problem by borrowing an idea from John Plecnik: the federal government could collect a wealth tax “without regard to apportionment, but return excess collections to the governments of the states from whose populations the excess is collected.” Yet apportionment would prevent a wealth tax from redistributing wealth across states (though it still could redistribute wealth among households within states). Note that the average income of the top 1% in Connecticut is five times the average income of the top 1% in West Virginia, and mean household income overall in Connecticut is nearly 1.7 times the mean household income in the Mountain State. Without some amount of Connecticut-to-West Virginia wealth-shifting (which is precluded by apportionment), the redistributive effect of a wealth tax is capped.
Why, then, might one want to tax capital through an accretion-type wealth tax instead of a mark-to-market income tax? One reason suggested by Hasen is that a wealth tax can exceed the rate of return while a capital income tax cannot. If the tax rate on capital income is greater than 100%, then presumably individuals subject to the tax will hold only cash (which generates no income and thus no tax). With a wealth tax, however, the government can drive the after-tax rate of return on capital below zero. “This feature of a wealth tax,” Hasen writes, “may be appropriate where the tax mitigates the . . . negative externalities deriving from wealth concentration.”
And yet a tax on capital need not exceed the risk-free rate of return in order for it to be effective in reducing wealth concentration. As readers familiar with the work of Thomas Piketty will recognize, capital’s share of income will rise if and only if r > g (where r is the return on capital and g is the growth rate of the economy). A tax on capital will reduce wealth concentration over the long run if it is set at a rate greater than r – g. Unless g is zero or negative, the tax need not be greater than r to have a wealth-deconcentrating effect. And a tax on capital at a rate as high as r is possible within an income tax framework without the undesirable side-effects of wealth taxation (i.e., exemption of immediately consumed returns, undertaxation of inframarginal returns, and the complications of apportionment).
In any event, whatever one thinks of a wealth tax, one will think more of it after reading Hasen’s article. And if the income-vs.-consumption-vs.-wealth tax debate is one that interests you, so too will Hasen’s contribution. As another law professor-cum-blogger might say, download it while it’s hot