March 10 NYU Tax Policy Colloquium (with Eric Zolt)

Very interesting session of the Tax Policy Colloquium yesterday concerning Eric Zolt’s paper, Tax Deductions for Charitable Contributions: Domestic Activities, Foreign Activities, or None of the Above.

Eric’s paper responds to scholarship that has argued for (a) more vigorously embracing the idea that U.S. charitable deductions are appropriate even when they fund charitable activities that occur abroad, (b) allowing suitably vetted foreign charitable organizations to get tax-deductible contributions from U.S. donors, and (c) allowing for-profit firms to receive deductible charitable contributions when their activities advance charitable purposes.

Proponents of each of these claims relied at least in part on good old legal analogy. As in: This stuff is just as good as, and/or is hard to tell apart from, things that unambiguously and/or uncontroversially get the charitable contributions deduction. So if we allow the deduction in other cases, we should here as well. To which Eric, in addition to assessing those arguments, says: But what if the charitable contribution deduction is itself questionable policy? Then, rather than accepting reasoning by analogy, we should rethink it across the board. He notes that the Obama Administration recently proposed converting the charitable deduction into a 28% credit, the Bowles-Simpson Fiscal Commission Report suggested a merely 12% credit, and in general there are other subsidy delivery techniques (e.g, direct spending).

Early in the paper, Eric rightly observes: “We lack any coherent theory that explains successfully why governments should allow [charitable deductions]. It is not from lack of trying.” In response, I’d say that the theoretical literature on this topic is fairly mature, and supports 2 clear conclusions. First, the deduction is not theoretically correct in the sense of, say, allowing business deductions under an income tax. Second, as a practical matter, using it involves a well-understood set of tradeoffs. It certainly isn’t an optimal subsidy design, but with limited information and political economy problems it’s plausible that one might choose deductibility, subject to a few bells & whistles. But I personally would opt for something a bit different.

More particularly:

1) The so-called “donor” theory, under which allowing charitable deductions is a part of measuring the taxpayer’s income correctly, is in my view clearly wrong. The great Harvard Law professor Bill Andrews made this argument almost 40 years ago, and in doing so advanced the literature, but I would say that this view has rightly been rejected. Andrews supported this result by defining the relevant consumption that an income or consumption tax would want to reach as “private preclusive appropriation” of resources. But if you think about the distributional rationale for a non-lump sum tax on terms that are based on welfare considerations, and thereby are drawn to the idea that people at higher budget lines should pay more tax (and/or get smaller transfers) than people at lower budget lines, it clearly follows that voluntary outlays that people make in pursuit of their own preferences are not like business outlays incurred in the course of earning gross income. Alf likes going to restaurants, Brenda likes giving money to charity, and we wouldn’t say that Brenda is worse off because she spent her $100 on something she valued, rather than on the sort of thing that Alf values. If one thinks (as I do) about the income tax as insurance against ability risk and under-diversified human capital risk, one doesn't really need insurance against voluntarily deciding to spend money on things. (One could try to build a case for behind-the-veil insurance against having particular "expensive tastes" such as for charitable giving that raise one's marginal utility of a dollar, but that would be a steep uphill climb.)

2) Absent the donor theory, one is left (as everyone recognizes these days) with subsidy theories for the charitable deduction. There are lots of ways to provide public support for what we define as "charitable" activity. E.g., direct government expenditure is an alternative, and often we use more than one method to increase supply of the same output. The key features of the charitable deduction are:

(a) It's a decentralized approach rather than one where Washington decides. Well-understood set of tradeoffs here. (The paper emphasizes minority vs. majority preferences, but I thought this a better framework given the complicated question of how majority coalitions arise, e.g., via logrolling.)

(b) It's a "skin in the game" approach. That is, the way to get the federal government to cut a check for a given charity (albeit indirectly via the tax savings to you from the deduction) is to cut the charity a larger check yourself. This has the virtue of forcing people to take seriously the question of where federal dollars should be directed federal dollars, along with the vice of creating radically unequal degrees of voice. Rich people effectively get way more votes regarding where to direct federal resources, whereas generally we make some show of saying one vote apiece (however slanted actual political influence turns out to be). Suppose instead that we eliminated charitable deductions and gave each adult U.S. citizen the right to allocate $X in federal money to the charity of his or her choice (with criminal penalties for selling the right). This would equalize voice, but presumably would undermine how seriously people took it.

(c) To the extent the deduction is allowed, it causes the marginal reimbursement rate or MRR (how much per dollar contributed you get back from the federal government) to your marginal tax rate or MTR. In other words, the MRR automatically equals the MTR. There is absolutely no reason to think that the MRR should generally equal the MTR, as different considerations apply to optimizing them. The optimal MRR depends on a number of different inputs, including the elasticity of donor response. Batchelder, Goldberg, and Orszag argue for a refundable flat-rate MRR, here and elsewhere in the Internal Revenue Code. Elasticity considerations might conceivably lead one to favor a rising MRR, if one isn't worried about the unequal voice aspect of it, but there's no reason to think that the MRR would rise in lockstep with, and generally equal, the MTR.

MRR-type thinking supports charitable deduction floors (e.g., rules providing that you can only deduct charitable contributions to the extent in excess of $X or Y% of your adjusted gross income), but it doesn't necessarily support the existing deduction ceiling (e.g., that limiting certain charitable contributions to 50% of one's adjusted gross income). The latter limits are hard to rationalize other than as a response to unequal voice, but in that case one might want the ceiling to have an absolute dollar rather than percentage of AGI structure, since we presumably don't need to limit the exercise of voice by low-earners.

3) The case for subsidizing charitable activity is often rationalized in terms of alms. But in practice, so far as the charitable deduction is concerned, it mostly means universities, arts institutions, and the like - things that rich donors prefer, and that typically are rationalized on public goods grounds. I agree with the paper that the public goods explanations for a lot of the non-alms stuff that generates big deductions today can be questioned. For example, even if we agree that institutions like Harvard generate public goods and should be subsidized, does their $30 billion endowment suggest that they don't really need the money right now, especially in light of other U.S. fiscal burdens?

If I had a fixed charitable budget, I'd want to direct much more of it to true alms and less to affluent donors' favored institutions. Note that the public goods explanation often emphasized in this area need not apply to alms, which responds to a public goods problem only insofar as people generally would like the poor to be helped but would prefer not to pay for it themselves. But helping people in bad straits is good in itself (i.e., a direct argument in the social welfare function), not just good because people who happen to like it face a collective action problem amongst themselves.