Ed Dolan's Econ Blog

The Charitable Deduction as a Tax Expenditure: What it Buys and What to Do About It (Part 2)

In the first part of this post, I argued that the charitable deduction is popular because people perceive it as a reduction in taxes that encourages charitable giving, but that perception rest on false premises. The charitable deduction is better viewed as a tax expenditure than as a tax reduction; surprisingly little of it goes to genuinely charitable purposes; and the degree to which it stimulates giving is often exaggerated. This second part of the post extends the critique and examines the pros and cons of possible reforms.

Efficiency and equity

Up to this point, I have emphasized the inefficiency of the charitable deduction, which violates the presumption that an efficient tax system should have a broad base and low marginal rates. The deduction could overcome that presumption if it were a highly effective way of promoting charity, but, as I argued in Part 1, the evidence does not support such a conclusion. Meanwhile, a commenter was kind enough to provide a link to a study from the Center on Philanthropy at Indiana University that further undermines the status of at the deduction as an effective way of helping the poor and needy. Using various surveys and data sources, the IU study concludes that

Only 8 percent of households’ donated dollars were reported as contributions to help meet basic needs—providing food, shelter, or other necessities. An estimated additional 23 percent of total private philanthropy (including donations from foundations, corporations, and estates) went to programs specifically intended to help people of low income–either providing other direct benefits (such as medical treatment and scholarships) or through initiatives creating opportunity and empowerment (such as literacy and job training programs).

The 31 percent of total giving that the IU study assigns to charity is very close to the one-third that I calculated earlier using less formal methods. The study differs in its assessment of some individual categories, however. It is less generous in allocating only 77 percent of the activities of combined-purpose charities to aid to the poor, and only 9.9 percent of giving to health-related organizations. It is slightly more generous than I was toward religious giving, allocating 20.1 percent, rather than my estimated 17 percent, to charitable causes.

While I have looked mainly at the efficiency of the charitable donation, many commentators have instead focused on its distributional effects. As Bruce Bartlett notes in his new book, the charitable deduction tends to benefit primarily the well-to-do, who face higher tax rates and are more likely to itemize deductions. As a result, the priorities of the wealthy dominate the pattern of giving.

Writing in The New York Times, Richard Thaler puts it this way:

Consider this scenario: Having decided that charitable giving is a worthy cause, the government subsidizes charitable gifts from certain households, and for those chosen to be part of the plan, every dollar donated to a charity is increased by a specified percentage. To qualify, taxpayers must have a substantial home mortgage; the subsidy rate increases with taxable income. Low-income taxpayers receive no subsidy, but donations from qualified high-income taxpayers are subsidized by as much as 40 percent — or more.

At this point, you may be wondering why I’d even mention something so preposterous. After all, why should a family’s eligibility for a donation subsidy depend on whether it has a large mortgage? And why should the government subsidize donations by the rich more than donations by the poor? The idea seems a nonstarter. And it would be, if not for one important detail: it is (approximately) the current law.

Some proposals for reform of the charitable deduction primarily address its distributional tilt toward the interests of upper income taxpayers; others focus on its inefficiencies. As we will see below, it should be possible to design reforms that address both, but first we need to address the concern that reform of any kind would be catastrophic to the cause of philanthropy.

How great is the risk of reform?

According to the World Giving Index from the Charities Aid Foundation, the United States was the most generous country in the world in 2011. What is the risk that reform could spoil that record? It is small, in my view.

First, it is worth noting that Americans do not earn their high marks for generosity solely, or even primarily, through giving money. The World Giving Index is a composite of three indicators. The first looks at peoples’ willingness to help strangers. The United States ranks second in that category. Of the seven highest-ranked countries, the other six are in sub-Saharan Africa. The second indicator measures peoples’ willingness to donate their time. The United States ranks fifth there; the nine others in the top 10 are all relatively poor countries in Africa and Asia. The third indicator is the percentage of the population who give money. The United States ranks tenth, behind Thailand, the UK, Ireland, and other mostly wealthy countries. It is only when the three indicators are averaged that the United States comes out on top for generosity.

Second, as I argued in the first part of this post, it is easy to exaggerate the degree to which the charitable deduction increases giving of money. This point is worth developing in more detail.

A 2008 NBER working paper by Jon Bakija and Bradley Heim provides a thorough analysis of the effects of the charitable deduction on giving. The authors begin with a review of the literature and then look for ways to improve on previous estimates, for example, by carefully distinguishing the short-term effects of tax changes on the timing of giving from long-run effects on the volume of giving. Earlier efforts, incuding a widely cited 1985 study by Charles Coltfelter, had found giving to be strongly responsive to changes in tax rates, with an estimated price elasticity of about -1.2. Using more recent data and adjusting for possible biases in earlier estimates, Bakija and Heim find a significantly smaller elasticity, about -0.7.

The price elasticity of giving means the percentage change in the quantity of giving that would result from a given percentage change in the “price” of giving. Some giving would take place even if there were no tax deduction, in which case the price of giving $100 to charity would be $100. If giving is tax deductible, any increase in the marginal tax rate represents a decrease in the price. For example, if the tax rate is 50 percent, the out-of-pocket cost of giving $100 to charity is just $50. If the tax rate increases to 51 percent, the cost of giving decreases to $49, that is, the price of giving falls by about 2 percent. With an elasticity of -1.2, such an increase in the tax rate would cause giving to increase by 2.4 percent; with an elasticity of -0.7, the increase in giving would be 1.4 percent.

For any elasticity and any tax rate, we can calculate the total amount by which a tax deduction would increase giving and the cost of the deduction in terms lost tax revenue. It turns out that with an elasticity of -0.7, the loss in revenue is greater than the gain in total giving for all tax rates. With an elasticity of -1.2, the gain in giving is very close to the loss in revenue for rates in the vicinity of the top tax rate of 35 percent. The elasticity of giving would have to be substantially greater than -1.2 to produce gains in giving by top-bracket taxpayers that noticeably exceeded the loss in revenue.

The disparity between loss of revenue and gain in giving is, of course, much greater if we look not at total giving, but just at its charitable component, which we have estimated to be about a third of the total. The loss of revenue from the tax deduction is greater than the gain in charitable giving for all tax rates and for any plausible elasticity. For a 35 percent tax rate and an elasticity of -0.7, the tax deduction costs the Treasury about $4.50 in revenue for every $1 of added charitable giving. Even at an elasticity of -1.2, the revenue loss is three times the gain in charitable giving.

The examples we have given are purely numerical; they have no political or ideological content of their own. Progressives and conservatives would presumably interpret them differently.

Progressives are likely to think that helping the poor is an important goal of public policy and that government programs to help the poor operate at least as efficiently as private charities. Accordingly, they could reasonably support a package that eliminated the charitable deduction; left tax rates unchanged; and spent the resulting increase in revenue on expansion of public programs to aid the poor. Doing so would result in more aid to the poor with no increase in tax rates.

In contrast, conservatives are likely to think that reducing tax rates is itself an important policy goal and that private charities do a better job of helping the poor than do government programs. They would prefer a package that eliminated the charitable deduction; “spent” the resulting gain in revenue to reduce tax rates; and left charity to the private sector. (Alternatively, they might prefer to spend part of the increased tax revenue on reducing the budget deficit and national debt.)

Either way, progressives and conservatives should be able to agree on elimination of the charitable deduction even if they disagree about what to do with the extra revenue.

Abolish the charitable deduction or reform it?

Up to this point, I have framed the policy choice as one of retaining the charitable deduction in its current form or abolishing it altogether. Before leaving the subject, however, we should consider ways of reforming the deduction without eliminating it.

One popular proposal is to replace the deduction with a tax credit. The tax credit could be set at a rate, perhaps 10 to 15 percent, that would have the same net effect on revenue as the present deduction. If so, the switch to a credit would give lower-income households and middle-income households that do not itemize a relatively larger say in the pattern of charitable giving, while reducing the influence of high-income taxpayers. Probably such a change would have no dramatic effect on total giving. Its primary impact would be on the perceived equity of the charitable tax preference rather than on its efficiency.

A second reform proposal is to cap the deduction for high-income taxpayers. One variant of a cap is included in the Obama administration budget proposal for 2013. Any resulting increase in revenue could go either to deficit reduction or to expansion of government aid to the poor. A cap would be likely to find favor with those who would like to see a shift from private to public efforts to aid the poor, and who would also like to shift a larger share of the total tax burden to upper income households.

A third possible reform would be to leave the deduction in place, but target it more narrowly at true charity. For example, as part of the process of qualifying for tax-exemption, organizations could be required to submit budgets showing what share of their spending is devoted to truly charitable purposes. Only that portion of contributions would be deductible. For example, if I gave $100 to a church that spent 20 percent of its income on charitable missions, I would get a tax deduction of $20; if my alma mater spends 10 percent of its income on scholarships for needy students, I would get a $10 deduction for each $100 gift, and so on.

It is easy to think of ways to combine or modify these basic kinds of reforms. If asked to choose among them, my own preference would be to change the deduction to a tax credit and target it narrowly at truly charitable giving. I am less enthusiastic about the idea of capping the charitable deduction; even if one thinks high-earners ought to bear a larger share of the total tax burden, fiddling with the charitable deduction seems an odd way to go about it.

On the whole, however, I am inclined to favor eliminating the charitable deduction altogether. Any reform runs the risk of adding to the complexity of an already overly complex tax code while producing small gains, if any, in efficiency and equity. I am convinced that private charity would survive the change in good health, as would philanthropy in the broader sense that extends beyond aid to the poor and needy.

One Response to “The Charitable Deduction as a Tax Expenditure: What it Buys and What to Do About It (Part 2)”

Jon BakijaApril 27th, 2012 at 8:52 am

Thanks for citing our work. You should check out the final published version of our paper, which is substantially different and I think substantially improved relative to the early NBER working paper version. An un-gated version is available here:

The final version demonstrates that when you use a more convincing identification strategy relying on differences in changes over time in tax incentives for charitable giving across states for people with similar income levels, the estimated price elasticity of charitable giving is larger, generally in excess of one. Estimates are smaller when you instead effectively use the middle class as a control group for upper-income people. But the middle class is not really a plausible control group for upper-income people, because there are really dramatic changes to income and wealth for upper-income people over time, and available observable data is not really sufficient to control for those differences.