This policy memo proposes and simulates the effects of a two-tier, nonrefundable tax credit for charitable contributions. A two-rate credit would expand access to tax incentives for charitable contributions to most Americans and increase charitable giving significantly, with substantial cost savings compared to alternative policy changes.
American charitable giving is at a critical juncture. Following the 2017 tax reform, the share of US households who take the itemized deduction for charitable contributions is expected to decrease sharply. Furthermore, the COVID-19 pandemic which, as of this writing, is precipitating an unprecedented global economic contraction, is further likely to reduce giving. So far, US policy attempts to bolster giving during the crisis are unlikely to mitigate the shock in a significant or cost-effective manner. If policy remains on its present course, less will be given in coming years to our charities and foundations, and by fewer people.
This policy memo proposes and simulates the effects of a two-tier, nonrefundable tax credit for charitable contributions. Relative to existing policy, a two-rate credit would expand access to tax incentives for charitable contributions to most Americans and increase charitable giving significantly, with significant cost savings compared to alternative policy changes.
Charitable giving is an important part of America’s economy and civil society. In 2017, giving exceeded two percent of GDP. The itemized charitable contribution deduction supports this giving by lowering the “price” of a charitable contribution for qualifying taxpayers. It does so by subtracting the value of donations from the donor’s taxable income, thereby reducing their tax. The value of the contribution deduction is higher when tax rates are greater (and thus the net cost of giving lower). The donor’s after-tax cost of giving has a large effect on charitable contributions.
However, the itemized deduction is not useful to many donors. Most tax filers are better off taking the standard deduction (a fixed reduction in taxable income) than deducting the sum of their charitable contributions and other itemized deductions because the standard deduction is greater than that sum and therefore reduces their tax burden more. Historically, about one-third of all Americans have itemized their returns in most years. The rest were either unable to benefit from the deduction on their tax returns or were not required to file a return at all.
The 2017 Tax Cut and Jobs Act (TCJA) is expected to reduce the share of taxpayers who itemize sharply to just over 10%. The TCJA increased the standard deduction and restricted many itemized deductions, making the standard deduction the better option for millions of Americans who had previously itemized their returns. Such households are likely to reduce their charitable contributions. Because millions of taxpayers will no longer receive a subsidy, charitable contributions are expected to decline overall, and giving will increasingly be done by a minority of mostly high-income households who will continue to claim itemized deductions.
This change has been partially offset by the Coronavirus Aid, Relief, and Economic Security (CARES) Act of 2020, which introduced a non-itemizers’ deduction for charitable contributions, limited to $300 of giving per return each year (P.L. 116–136 §2204). As I will explain, the structure of this deduction will make it a weak and expensive incentive, with most of the tax expenditure going to existing giving rather than encouraging new giving.
I propose a credit for charitable giving with two rates, where the generosity of the subsidy depends on the extent of the claimed giving. Specifically, I propose a credit rate of 10% (the lowest federal income tax rate) on gifts up to two percent of adjusted gross income (AGI), followed by a 37% credit (the highest income tax rate) on any further giving, though alternative subsidy rates are discussed below. The credit is nonrefundable and limited to gifts under half of AGI; filers can choose to take the credit or the itemized contribution deduction but not both. As I will explain, a two-tier tax credit would offer a tax subsidy to all taxpayers while also containing the cost of subsidies for charitable giving.
The arithmetic of this credit is summarized in Table 1. Everybody who donates (and owes income tax) can take the credit. Small gifts receive a small subsidy; proportionally larger gifts receive a proportionally larger subsidy.
|Marginal subsidy||Total subsidy||Minimum gift||Maximum gift|
|10%||0.1 × Contributions||$0||0.02 × AGI|
|37%||0.1 × 0.02 × AGI + 0.37 × (Contributions − 0.02 × AGI)||0.02 × AGI||0.5 × AGI or credit cancels all tax owed|
This simple structure offers many advantages over the existing contribution deduction. First, the two-tier credit is more equitable because its subsidy rate is the same for all returns. More subtly, the credit is also more effective than the itemized deduction at encouraging greater giving because it becomes more generous as taxpayers give more. By contrast, the itemized deduction rewards high income brackets, not large gifts. The itemized deduction’s value as a percentage of the gift increases in a household’s taxable income (because those households pay higher marginal rates) and decreases in value in the size of the gift (because giving decreases taxable income). For these reasons, Canada already implements its tax subsidies as a two-tier credit, and many other wealthy countries subsidize giving with tax credits rather than deductions.
That a deduction is worth more to that in higher tax rate brackets is obvious; the second point, that the deduction inefficiently penalizes large gifts, is subtler. To illustrate the efficiency gain of a two-tier credit relative to the deduction, consider Figure 1. The vertical axis marks the ratio of the reduction in tax received by a taxpayer to the dollars donated. The horizontal axis marks the ratio of the donation to income. A black, dashed line marks the two-tier credit. Up to 2% of income, the credit is worth 10% of the gift. For larger gifts, the value of the credit gradually approaches 37% of the gift because of the higher-rate credit tier.
The opposite dynamic is observed for a deduction, which is less generous for larger donations. A solid grey line plots the value of the itemized deduction for a specific middle-class household: a married couple, making $110,000 per year and filing jointly, with $30,000 worth other itemized deductions. Because $30,000 exceeds the standard deduction, they will always itemize. And because their taxable income after the other deductions is $80,000, they are in the 22% marginal rate bracket. Therefore, their first dollars given receive a 22% subsidy rate. Once this household’s giving reaches $2600, however, their tax bracket falls to the 12% rate on incomes between $19,051 and $77,400. Their tax reduction rate declines as their giving increases.
Say that, for some reason, this household decided to claim the $300 CARES deduction instead of the itemized deduction. This policy is plotted with a grey dotted line. Now the marginal benefit of giving more is zero once the household’s contributions exceed $300, so the value of the tax savings as a share of the amount given falls sharply beyond that point.
Of course, this is a visualization of a specific case. For most Americans, who do not itemize, the value of the itemized deduction is zero until their giving and other deductions exceed the standard deduction. The dynamics of the CARES deduction, however, are similar for everybody: giving in excess of $300 is not incentivized, and donors who already planned to give at least that much will pay less in tax with essentially no spur to further giving. The two-tier credit, however, encourages all households to give more, regardless of their itemizer status. Understanding the effects of the credit for charitable giving and other outcomes requires a policy simulation on a representative population of tax returns capturing the itemization behavior and economic status of all Americans.
I simulate the effect of the two-tier credit on taxpayer behavior and administrative outcomes. First, I construct a simulated dataset of 2018 tax returns based on the most recent IRS Public Use File dataset of individual tax returns. For itemized returns, all variables of interest are observed directly. For non-itemizers, who do not report their charitable giving, I impute the amount of charitable contributions using survey data from the Panel Study of Income Dynamics (PSID). To do so, I match non-itemizing tax returns to similar non-itemizing households in the PSID sample. Specifically, I create matches of tax returns to all non-itemizer PSID households with the same marital status and similar total income and wage/salary income. Matched pairs are weighted by the PSID cross-sectional sampling weights. I compute simulated federal income tax rates and giving subsidies, and the effects of those subsidies on giving and other behaviors, for each simulated tax return, using the TAXSIM calculator hosted by the National Bureau of Economic Research and assuming a tax-price elasticity of giving of −1.0. I adjust returns for inflation to 2018 dollars and scale aggregate outcomes for population and economic growth. This process accurately predicts pre-2017 levels of giving, and closely matches the simulations of the Tax Policy Center of the TCJA.
Before presenting these simulations, I would like to emphasize what these estimates do and do not show. First, these estimates extrapolate giving from a previous year’s tax data by applying a proportional change calculation to the long-run change in the value of tax subsidies for giving. Effects of short-run changes in incentives, such as shifting giving across 2017 and 2018 to maximize tax benefit, are not captured. Second, non-tax changes, such as macroeconomic factors, are only captured to the extent that I have rescaled all quantities appropriately for inflation and growth to 2018 levels; the currently unfolding macroeconomic disaster of COVID-19 is not simulated. Lastly, because the simulation assumes a change that is proportional to existing giving and changes in tax incentives, it does not model changes in the giving participation rate; people who give zero in the source data are still modeled as zero-givers.
Table 2 presents estimates for 2018 outcomes under five scenarios: (1) pre-TCJA law (that is, if the TCJA had never been passed); (2) actual 2018 law (post-TCJA); (3) actual 2018 law with the addition of a non-itemizer’s deduction limited to $300 of giving per return, as in the CARES Act; (4) actual 2018 law with the addition of a two-tiered credit, with taxpayers choosing between the credit or the itemized deduction (but not allowed to claim both); and (5) actual 2018 law with the addition of the two-tiered credit and the repeal of the itemized contribution deduction (so donors can only take the credit). I report dollar estimates for total charitable contributions and for the tax expenditure on those contributions under each scenario, as well as estimates of the number of tax returns claiming itemized deductions.
|Charitable contributions||Tax expenditure||Itemized returns|
|(1)||No 2017 tax reform (counterfactual)||$302 billion||$69 billion||48 million|
|(2)||Year 2018 policy (post-TCJA)||290||74||17|
|(3)||Post-TCJA policy + $300 non-itemizer deduction||304||91||17|
|(4)||Post-TCJA + credit or deduction||348||120||14|
|(5)||Post-TCJA + itemized deduction repeal + two-tier credit||349||77||13|
The share of tax returns that itemize will fall sharply for tax year 2018, from 48 to 17 million returns, because of the Tax Cut and Jobs Act. Charitable giving will decline as well: the simulation predicts a $9 billion decline in giving, or about 3%, because of the change in incentives. The decline in dollars given is not nearly as dramatic as the decline in the number of itemized returns for two reasons. First, a majority of households do not change itemization status, and many people give something to charity without a tax subsidy, so the decline represents the effects of lower subsidies for the minority of households affected. Second, the TCJA all but eliminated the Alternative Minimum Tax (AMT) for most filers, which actually increased giving incentives for a minority of returns who make very large donations. Many tax returns with tens of thousands of dollars in charitable giving and other deductions — enough that their total deductions had triggered the AMT — will now take itemized deductions against the ordinary 32–37% tax rate on high incomes instead of the 28% AMT tax rate, an increase in their subsidy rate of 14–32%. The greater incentive for these households, who already give large amounts, partially offsets the loss of others’ incentive to give.
Because of the increased value of the contribution deduction to households previously subject to AMT, the tax expenditure cost of the deduction will rise following the TCJA, even as the number of itemizers and the total dollars given decline. Under current law, the tax expenditure on the charitable contribution deduction rises 7% relative to previous policy. The greater cost of the deduction for the households who had previously been subject to AMT more than offsets the cost savings from rate reductions and reduced itemization rates for others.
Third, I add a $300 deduction for non-itemizers to the simulation. Given the small stakes, very few people choose to switch to non-itemizing because of this incentive (about 150,000 returns). But the non-itemizer deduction does have an effect: giving increases by $11 billion, more than undoing the forecasted decline of the TCJA. However, it does so at significant cost. The foregone tax revenue of the two deductions is now $91 billion, 23% greater than under actual 2018 law, or 32% greater than under pre-2018 law. Put another way, for every dollar in additional giving induced by this deduction, the treasury loses $1.55 in revenue. With a relatively low cap of $300, the full amount of the tax reduction flows to those who already gave more than that amount, but without any new incentive to increase their giving, blunting the policy’s effectiveness.
Next, I simulate the effect of offering the two-tier credit. Taxpayers may choose to take the itemized deduction or the credit, but not both. This policy significantly increases contributions: giving is over 20% higher than under actual 2018 law, and 15% higher than it would have been without the TCJA. This does come at a cost: tax expenditures on charitable contributions rise by about $41 billion, or about 79% of the increase in donations. Because this policy expands the credit to all filers with taxable income, and because it allows the taxpayer to pick the more generous subsidy from among two options, it is costly as well as beneficial.
In addition to these effects on giving and tax receipts, the credit reduces the number of households who claim itemized deductions, an additional benefit for taxpayers. Itemizing is a costly hassle: tax preparers charge about $100 more for an itemized return than a standard deduction return, and the indirect costs to filers are even higher — equivalent to 10–15 hours of work. While claiming a credit has its own compliance costs, the two-tier credit calculation is simple and need not require filling out an additional schedule. Allowing taxpayers to choose a credit instead of a deduction reduces the number of itemized returns by about three million, implying nationwide total savings of hundreds of millions of dollars in tax preparation fees and over 30 million taxpayer work hours.
The two-tier credit is so efficient, inducing substantially more new contributions than it forgoes in tax revenues, specifically because it is more valuable for large gifts than for small. To demonstrate this point explicitly, I include a final simulation: what would have happened in 2018 if the TCJA had enacted the two-tier credit and repealed the itemized contribution deduction? The result: simulated giving increases by roughly the same as if donors can choose between the two incentives, but the increase in tax expenditures of $3 billion is just 7% as large. Indeed, the simulation suggests that each $1 in foregone tax revenue induces over $19 in new giving.
This last simulation should not be taken literally. The estimated tax elasticity of giving of −1 is based on changes in tax rates for the existing system of itemized deductions beyond some standardized deduction cutoff. Substantially changing the set of itemized deductions would likely cause changes in other taxable behaviors over time, and actual policy effects of repealing the deduction would probably differ substantially from this simulation’s predictions. Rather, the point of this final exercise is to emphasize the cost-effectiveness of the two-tier credit. Many proposed policy reforms attempt to restrain costs by limiting or reducing the tax savings for the largest gifts using a cap or (in the case of the deduction) a decreasing subsidy rate. However, such policies increase the costs of a subsidy away from the decision margin while weakening incentives to give just a bit more. The two-tier credit is effective and efficient because it pairs a high marginal incentive for large gifts with a low tax expenditure on the first dollars given.
In summary, the two-tier tax credit can accomplish three desirable public policy objectives: (1) it would enable all Americans to receive recognition and subsidies for giving to the charitable organizations that support our civil society; (2) it would more than undo the projected decrease in charitable giving predicted to follow the TCJA; and (3) it would do so cost-effectively, and without making the tax system unduly complex.
The analysis above has focused on a very specific credit regime. Next, I consider the sensitivity of these projections to policy choices and the assumptions of the simulation. Specifically, I will simulate important effects of implementing the two-tier credit in addition to the itemized contribution deduction if the rate of subsidy above 2% of AGI differs from 37%, and if the share of AGI that must be donated before receiving the 37% subsidy rate differs from 2%, under a range of plausible tax-price elasticities.
Figure 2 presents simulations of outcomes of interest when the rate on giving in excess of 2% of AGI varies from 10% (that is, a flat 10% credit) to 50% (half the value of the additional giving). The top panel of the figure plots the additional charitable contributions the credit induces on the vertical axis against the higher subsidy rate on the horizontal axis. The bottom panel plots the ratio of additional contributions to foregone tax revenues—the “bang for buck” of the policy—against the top subsidy rate. A thick black line plots the simulations using a tax-price elasticity of giving of −1.0. Green bands display a range of simulations with taxpayer responsiveness ranging from −1.3 to −0.7, giving the reader a sense of the range of plausible policy outcomes.
Higher rates encourage more giving and do so more efficiently than lower rates. A flat rate of 10% encourages more giving than no credit, but almost certainly increases giving by less than it decrease tax revenues. Under the baseline simulation, the credit induces more giving than it costs in foregone tax for rates over 25%, and a credit with a higher-tier subsidy rate of 50% induces nearly 150% of its tax cost in additional giving. These efficiencies are possible because as rates increase, (1) the stronger incentive effect means more subsidies goes to new giving, rather than existing giving, and (2) more taxpayers choose to claim the credit instead of the deduction, improving efficiency.
Of course, efficient and cheap are not the same thing. A credit with a 50% top subsidy rate would increase giving by $120 billion, but at an additional cost to Treasury of $80 billion. While this would not make US giving subsidies excessively costly relative to peer countries, this change might be more than the current legislative environment can countenance.
We might also consider a 10%/37% two-tier credit where the share of AGI that triggers the higher rate differs from 2%. Figure 3 again plots the new contributions induced, and the “bang for buck” of the policy, when the AGI threshold varies from 0% (a flat 37% credit rate) to over 10% of AGI. Strikingly, new contributions are not highest at a zero cutoff; instead of being disincentivized by an initial subsidy tier of 10%, some households give more to reach the higher subsidy level, making a two-tier credit more effective in absolute terms than a flat rate. Beyond a 1% threshold point, new giving steadily declines, with each percentage-point increase in the tier cutoff decreasing new giving by about $3 billion. However, the cost-effectiveness of higher cutoff thresholds diminishes quickly and begins to decrease at AGI thresholds beyond 8%. This is because the higher revenue-efficiency of providing lower average subsidies to larger gifts is offset by the weakened incentive power of giving a flat 10% rate to most small gifts.
The best parameters for a two-tier credit therefore depend on the relative importance of scale, cost, and efficiency. Higher rates in the second credit tier encourage more giving and are more effective at encouraging giving, but they do so at a greater absolute cost in foregone revenue. Raising the threshold for the higher rate improves efficiency and reduces the tax cost of the credit, but with diminishing returns that peter out at around 8% of AGI. Higher thresholds encourage less giving in absolute terms.
The two-tier tax credit described in this memo would efficiently expand the share of Americans who receive a tax incentive for giving and increase the funds available to our charitable organizations. This credit would do so efficiently, at a cost in tax expenditures that compares favorably to the itemized deduction. A two-tier credit would a useful policy for building a fairer, more prosperous, and more civically engaged United States of America.
Funding source: Council of Michigan Foundations
I am grateful for financial support for this project form the Council of Michigan Foundations. I am further thankful for comments from Jim Ferris and audience members at ARNOVA and at the United Philanthropy Forum. All mistakes are the author’s.
© 2020 Nicolas Duquette, published by De Gruyter, Berlin/Boston
This work is licensed under the Creative Commons Attribution 4.0 International License.