Contents

• Introduction
• Comparisons to Past Tax Changes
• Estimated Effects on Aggregate Charitable Giving
• Share of Contributions Affected
• Percentage Change in Price
• Elasticities: Price and Income
• Current Effects of a Future Change
• Income Effects
• Effects by Types of Charitable Objectives
• Estate Tax Issues
• Historical Changes in the Estate Tax
• Estimated Effects on Aggregate Charitable Giving
• Share of Contributions Affected
• Percentage Change in Price
• Price Elasticities
• Current Effects of a Temporary Change in the Estate Tax
• Effect of Change in Estate Taxes on Lifetime Charitable Contributions
• Wealth Effects
• Effects of Changes in the Estate Tax by Types of Charitable Objectives
• Policy Options
• Conclusion

Summary

The Administration's 2010 and 2011 budget outlines contain a proposal to cap the value of itemized deductions at 28%, for high-income taxpayers. In the 2010 proposal, the expected revenue was dedicated to addressing health care issues; as other sources are expected to finance health care, the proposal is now part of the increased taxes on upper income taxpayers. This proposal has generated considerable concern about its potential negative effect on charitable contributions. This concern has been heightened because charities are having difficulties in the current economic climate. The proposed tax change, however, would not go into effect until 2011 and thus the change could actually increase near-term contributions. Thus, it is the longer-term, or permanent, effect on giving that is the effect considered in this analysis. The analysis also considers the effects of other income tax changes and of the estate tax.

The estimated effects of the cap and other elements of the budget package depend on whether the proposals are compared with the current tax rates of 33% and 35% or the rates scheduled for 2011, 36% and 39.6%. Compared with current rules, estimated effects are between one-half a percent and 1% decline in charitable giving, depending on whether the effects of capital gains tax rates on gifts of appreciated property are included. When compared with tax rate provisions in 2011, charitable deductions are estimated to fall by about 1.5% if only the cap is considered, but if income effects from the entire budget package are included contributions actually rise 2.5%. The relatively modest effects of the proposal arise because (1) the effect of caps on the subsidy value is limited, (2) only a fraction (about 16%) of charitable giving is affected, and (3) because evidence suggests that behavioral responses to changes in subsidies are relatively small.

Different charities will be affected differently because the giving patterns of higher-income individuals differ from the average. Estimates show smaller reductions or larger increases for religious or combined charities, or charities directed at meeting basic needs, whereas the proposal is more likely to have negative effects for charities serving the health sector, and to a lesser extent art and education charities. Overall, contributions that benefit the poor will be less likely to fall or more likely to rise than the average contribution because the charitable purposes more favored by higher-income contributors are less likely to direct benefits to low-income recipients.

Estate tax changes would also affect charitable giving. The budget outlines hold the current 2009 estate tax rules constant. Allowing the estate tax to lapse in 2010, as would the current rules, could lead to reductions in charitable giving of around 4%. Returning to the higher estate tax rates currently scheduled for 2011 could increase charitable giving, by about 1%, while adopting the FY2010 Senate Budget Resolution provision could reduce charitable contributions by about 1%. Although a smaller share of charitable contributions are affected by the estate tax, the changes in subsidy value are much larger if the estate tax is repealed, and the estimated behavioral response is greater. The immediate effects on contributions and the distributional effects of changes are, however, uncertain. Over half of bequests involve gifts to foundations, which finance a variety of charitable objectives and provide benefits with a considerable delay.

Revenue from the cap on itemized deductions is currently directed at increasing revenues to finance other programs. If the cap is rejected either overall, or for charitable contributions, other revenue sources found, or the debt increased. Alternative revenue options include, among others, implementing a floor under charitable deductions and increases in tax rates on high-income taxpayers.

---

Charitable Contributions: The Itemized Deduction Cap and Other FY2011 Budget Options

Introduction

The Administration's 2010 and 2011 budget outlines contain a proposal to cap the value of itemized deductions at 28%, for high-income taxpayers. In the 2010 plan, the expected revenue was dedicated to addressing health care issues. Other revenue sources have been proposed for this purpose and the current proposal is part of the Administrations tax provisions for upper income taxpayers.

The itemized deduction cap has generated considerable concern about its potential negative effect on charitable contributions, especially in light of the difficulties charities are having during current economic conditions. The proposed tax change, however, would not go into effect until 2011 and could actually increase current contributions in the short term. Thus, it is the longer-term, or permanent, effect on giving that is considered in this analysis. The analysis also considers the effects of other income tax changes and of the estate tax.

Deductions normally save taxes at the marginal tax rate. If a taxpayer's top bracket is 35%, a dollar of deduction would lower taxes by 35 cents. The proposal would limit the reduction to 28 cents. Taxpayers claim deductions for charitable gifts when they itemize. Most taxpayers (70%) do not deduct their contributions because they take the standard deduction or in some cases do not file tax returns. In addition, the provision would affect only the top two marginal tax rates, which affect only about 1.4% of returns.

Limiting the value of itemized deductions or converting deductions into credits is not a radical idea. The Congressional Budget Office, for example, discussed converting deductions into a credit in its 2009 Budget Options study.¹ President Bush's Advisory Panel proposed to apply a credit to mortgage interest deductions and eliminate tax deductions although it would have retained and extended the charitable deduction.² The reason for proposed credits and caps on itemized deductions is that itemized deductions create subsidies that differ across income classes. For the 70% of households that do not file returns or file returns but do not itemize, there is no subsidy. For returns that do, the subsidy rates vary; the lowest income taxpayers receive a subsidy of 10% to 15%, while the highest income taxpayers have a subsidy rate of 35%.

In evaluating the proposed contribution deduction and its implications for charitable giving, this report first compares the magnitude of the proposal with past tax changes. The lack of any clear indication in historical experience of powerful effects of tax changes suggests a small response. The following section provides calculations of the consequences for charitable giving, which are estimated based on the share of giving that is affected by the cap, the magnitude of the price change, and the elasticity (behavioral response to tax changes). It also considers the effects of other income tax provisions in the budget and the effects on different types of charitable organizations. The study also discusses the role of the estate tax. It concludes with discussions of policy trade-offs and alternative policy options.

Comparisons to Past Tax Changes

Some insight into the expected impact of the deduction cap might be found by comparing tax changes to past tax revisions. The price of charitable contributions for itemizers is (1-t), where t is the tax rate at which contributions are deducted. For example, if the individual is in a 25% tax bracket, the tax price is 0.75, indicating that a taxpayer has to give up 75 cents for each dollar of contributions. That is, if the taxpayer in that bracket contributes a dollar, he or she saves 25 cents and only loses 75 cents that could have been used for other purposes.

The tax price of giving is affected by a cap on the rate at which deductions occur but is also affected by the marginal statutory rate. Consider the top tax rate. It has fluctuated substantially since the income tax was introduced in 1913, beginning at rates as low as 7% and rising as high as 92%. Starting in the mid-sixties, the top rate was 70% for many years (although it rose slightly with the Vietnam War surcharge). Beginning with legislation in 1981, the top tax rate has been reduced substantially. Effective in 1982, it was reduced from 70% to 50%. In 1986, it was further reduced to 28%. Rate increases occurred in 1990 and 1991, and decreases in 2001. Table 1 compares the magnitude of those past changes in tax price to the effects of the proposed itemized deduction cap. Two effects are considered for the cap: the effect of imposing a cap on the current (2009) top tax rate of 35% and the effect of imposing a cap on the top rate of 39.6% which is scheduled for 2011.

The percentage changes in tax price in the 1981 and the 1986 legislation were very large compared with the current proposal. If their effects are compared with the proposal's effect with current tax rates, the 1981 legislation is over six times as large and the 1986 change is over four times as large. The combined changes are 12 times the size of the proposed tax increase at the top rate (where the main effects of the proposal will be directed).

Tax rate changes differ from caps on deductions because they also have an offsetting income effect. This effect tends to be smaller than the price change, however, because of graduated rates, and in some cases were offset by other provisions. For example, the Tax Policy Center reports that allowing the individual rates to rise to their planned 2012 levels in the top 1% would result in an 11.6% reduction in tax price but only a 4.4% reduction in income.³ The 1986 Tax Reform Act, with one of the deepest tax cuts in the top rates, maintained the pre-existing revenue yield and distribution across income classes. Thus, it contained no income effects.

Table 1 addresses rate changes that affected the top rate, but the 1981 and the 1986 tax changes also reduced marginal tax rates across a broad range of taxpayers.⁴ In addition, non-itemizers were allowed an itemized deduction in 1985 and 1986.

Figure 1 shows the pattern of giving as a percentage of GDP over this period. There is no indication in this pattern of significant shifts due to tax rate changes. Contributions after 1981, despite tax price increases, remained relatively stable as a percentage of price. The small peak around 1986 is generally attributed by most researchers to a temporary rise in deductions reflecting a timing shift as tax cuts for 1987 and 1988 were pre-announced in the 1986 tax cut, but by 1989 contributions had returned to their previous levels. Contributions following the 1993 tax increase fell rather than increased.

A more detailed discussion of the potential effects is presented in the next section, but this historical comparison suggests it is unlikely that a significant effect on charitable giving will occur.

Figure 1. Charitable Contributions as a Percentage of Output, 1967-2007

Source: CRS calculations based on the Center for Philanthropy Giving USA 2008 and National Income and Product Accounts. Other charitable contributions (the difference between the two lines) include corporate, foundations, and bequests.

Estimated Effects on Aggregate Charitable Giving

To estimate the effect on charitable giving arising from the itemized deduction cap, three elements are required: the share of donations affected, the percentage change in price for those donations, and the permanent price elasticity.

Share of Contributions Affected

As shown in Figure 2, a relatively small share, 15.9% of all charitable contributions, would be affected by the itemized deduction cap (which affects 1.4% of taxpayers). According to the Tax Policy Center, the share of individual contributions that are subject to the two highest tax brackets are 4.7% for the current 33% bracket (36% in 2011 absent tax law changes) and 23.2% in the 35% bracket (39.6% in 2011 absent tax law changes).⁵ Thus a total of 27.9% of individual charitable contributions on itemized returns falls within these marginal rates. Some higher-income individuals will not be affected by this cap because they are subject to the alternative minimum tax, with a maximum rate already at 28%.

Figure 2. Share of Contributions Affected by Itemized Deduction Cap (in $ billions)

Source: Center on Philanthropy, Giving USA 2008, and CRS calculations. All percentages are as a share of total contributions.

There is already a limit on charitable contributions, which is often exceeded by high-income donors, and for some donors, contributions are not affected by changes in the tax rate for itemized deductions because they already contribute in excess of the maximum. Charitable contribution deductions cannot exceed a certain percentage of income (50% for ordinary contributions, less for gifts of appreciated property and gifts to foundations). The excess contributions are carried over to future years to be deducted, but for very large donations or for contributors who have large donations year after year, these deductions may never be taken. If they are, their value declines because of the time value of money. This effect is potentially quite significant among wealthy donors. One study of contributions from income and estate data found that the wealthy who appeared in the estate tax data contributed over a long period of time about twice the amount of their deductions claimed.⁶ The average adjusted gross income in this panel (covering 1987 to 1996) was $1.8 million. The relationship between contributions and deductions was highly variable from year to year, likely reflecting the importance of large one-time gifts.

Another study reported the percentage of giving over the limits for a single year (1995) and for a panel covering 1991-1995.⁷ For the adjusted gross income class of $2.5 million or over, the percentage of giving over the limit was 27.6% in 1995, whereas for the panel, the percentage of giving over the limit was 34%. Adjustments for the fraction of contributions in the top two brackets not affected by the tax change are based on this data, using the 1991-1995 panel. The overall excess of contributions over deductions in this data, covering income classes of $200,000 and over, weighted by giving, is 18.7%. Based on the estate tax study, that the highest income class, which accounts for about half the excess, is assumed to never deduct the excess, while the remaining income classes deduct it the following year, which at a discount rate of 5% would cause a loss of 5% of the value. This is equivalent to assuming that about half the excess contribution's deduction value is lost

This adjustment reduces the share of deductions affected in the lower income bracket from 4.7% to 4.5% and the share in the higher income bracket from 23.2% to 20.8%. Overall the share falls from 27.9% to 25.3%.

Finally, the estimate is adjusted for the share contributed by individuals who itemize, estimated at 63.2% for the latest year available.⁸ Although individual inter-vivos contributions are the bulk of the source of charitable contributions, about a quarter of contributions are made by corporations, estates, and foundations. Contributions are also made by individual non-itemizers. Thus the 25.3% of individual contributions affected by the rate change represent 15.9% of total contributions.

Percentage Change in Price

The price of charitable contributions for itemizers is (1-t), where t is the tax rate at which contributions are deducted. For example, if the individual is in a 25% tax bracket, the tax price is 0.75, indicating that a taxpayer has to give up 75 cents for each dollar of contributions because the contribution reduces taxes by 25 cents. Tax prices can also be affected by matching-donation programs; for example, an offer to match each dollar of contribution with an additional dollar from a matching-program sponsor would lead, even for a non-itemizer, to a tax price of 0.5 (i.e., it costs only 50 cents of income to achieve a dollar in contributions).

Effects are calculated compared to two baselines: current tax rules with tax rates of 33% and 35% and tax rules in 2011 with rates of 36% and 39.6%. In addition to the basic price effects, price effects are also estimated incorporating the proposed increase in the tax rate on capital gains from 15% to 20%, which affects gifts of appreciated property. This effect is only related to a comparison of current tax rules. A significant portion of high-income individuals' contributions are in the form of appreciated property. The value of donating property differs from the value of cash donations. Currently, taxpayers are allowed to deduct the entire cost of appreciated property, without paying the capital gains tax. Since the cost of a dollar of consumption from sale of an appreciated asset is 1/(1-at_g) where t_g is the capital gains tax rate and a is the share of value that would be taxed as a gain, the price of charitable giving is (1-t)(1-at_g).⁹ In Table 2, the base case represents the case with no appreciation. Two cases with appreciation of 50% of the value and 100% of the value are included.

Elasticities: Price and Income

The third element needed to estimate the effect on aggregate contributions of changes in the proposed policy is a measure of the responsiveness of contributions to prices. The estimate is based on the elasticity. Elasticities can be either price elasticities or income elasticities. A price elasticity is the percentage change in giving divided by the percentage change in price (with price equal to (1-t)). This relationship is negative, and an elasticity greater than one in absolute value is generally referred to as relatively elastic, while an elasticity less than one in absolute value is relatively inelastic.¹⁰ An income elasticity is the percentage change in giving divided by the percentage change in income and is positive. The remainder of this subsection discusses price elasticities.

Price elasticities greater than one indicate that a price subsidy induces more giving than the revenue loss; if price elasticities are less than one, more charitable spending could be achieved by other means, such as direct grants.

The evidence on these elasticities is generally drawn from tax data because variation in the price of giving generally occurs through the tax system. Most earlier studies of responses were based on comparing individuals who face different tax rates because they have different incomes; more recently researchers have used panel studies that allow for variation over time, while still focusing on the individual. However, aggregate changes over time also provide some insight into effects. As noted above in the historical comparison, the aggregate data on giving are not suggestive of a significant response to tax rate changes whether general (as in 1981 and 1986) or directed at higher-income individuals (as in 1990, 1993, and 2001). The share of giving did not change between 1980 and 1983. Similarly, with the 1986 change individual giving as a share of output was 1.47% in 1983 and 1.45% in 1989. While such basic comparisons do not control for many potentially important factors, the lack of evidence of large effects suggests that responses to tax changes cannot be very large.

Another way in which the price of giving changes is through matching grants. A small literature has developed which examines the response to matching grants through laboratory field experiments where solicitations vary randomly over individuals as to the existence and size of a matching grant. These studies have generally found modest effects of matching grants. In one study, the announcement of matching grants led to increases in donations but the magnitude of the match did not matter; the estimated overall price elasticity was low, at 0.3.¹¹ Another study found that the matching grant (up to a given amount) actually produced a smaller response than apprising prospective donors of a lead contribution that had already been made.¹²

In a study that surveyed philanthropy and high-net-worth households, one question asked was whether tax payers would reduce their charitable giving if there were no tax deductions for donations.¹³ Fifty-two percent of respondents said they would not reduce their donations at all, 37% said they would somewhat decrease donations, and only 10% would dramatically decrease donations. Since, at a 35% rate the tax price would increase by about 50%, this response suggests relative inelasticity. Note, however, that data on what individuals say they would do are not generally considered as reliable as evidence about what they actually do.

All of these types of findings suggest an inelastic response in charitable giving relative to price changes. For many years, however, tax economists estimating the response of taxpayers to changes in the tax price of charitable giving estimated a relatively elastic response of over one. In a study published in 1991, an estimate of typical values was set at a price elasticity of -1.27 and a typical income elasticity as 0.78.¹⁴ These studies generally used cross sections of individuals with different tax rates.

Researchers confronted many issues in estimating these relationships. One was separating income from price effects when the two moved together. Another (which had already become important in the discussion of capital gains realizations) was the possibility of timing of responses. For individuals with fluctuating incomes, there is an incentive to make charitable deductions in years with high tax rates; indeed, compared to permanent responses, this type of shifting over time is relatively costless. One could have no permanent effects on giving with a permanent change in taxes, but observe effects across individuals because of transitory effects.

The tax legislation of the 1980s provided a dramatic set of tax cuts and in 1992, researchers at the Treasury Department presented a paper to the National Tax Association which highlighted the shortcomings of research and its ability to predict behavior. This paper used panels that traced the same taxpayers to highlight two important points. When the same data were used to estimate responses in the traditional cross section approach (yielding a price elasticity of -1.1 and an income elasticity of 0.67, typical of values in the literature) and these data were then used to predict charitable giving during the 1980s, the results were significantly in error. For example, in the $1,000,000 and over income class giving in 1982 fell (after establishing a baseline using the income elasticity), due to price effects, by a third of the predicted amounts (12.6% rather than 36.8%). For the $200,000 to $1,000,000 class, the model predicted a fall of 21% in 1982 but contributions rose by 33.2%. In general, this pattern overall suggests elasticities that are too high. Moreover, all of the high-income groups had a rise in giving in 1986 suggesting an important timing effect. There was more-mixed evidence of a timing effect for 1981. The 1981 tax cut was not announced as far in advance as the 1986 tax cut.

As a result of these concerns, research turned to panel studies which could attempt to account for transitory price effects, which would be expected, at least among sophisticated high-income individuals, to be quite significant. At the same time, a temporary rise or fall in income should have a relatively small effect (consistent with the permanent income hypothesis). Thus, transitory price elasticities would be expected to be high, probably higher than permanent price elasticities, while temporary income elasticities would be small. For lower-income and less sophisticated taxpayers these results would not necessarily be expected.

There is no consensus from the panel studies and unfortunately, estimates of responsiveness show significant variation depending on time period, data set and methodology, suggesting that flaws remain. Table 3 shows the results of applying a low, a high, and more central range of price elasticities from the studies (which are discussed in greater detail in Appendix A). Compared with current tax rates, the central estimate suggests a reduction of somewhat more than one-half of 1% in charitable giving in response to the proposed change.¹⁵

Current Effects of a Future Change

One issue of concern raised in response to the Obama proposal is the effect on charities during the current economic downturn. The proposal is not scheduled to take place until 2011, and thus the price effects for the deduction cap (as shown in Table 3) would not apply until that time. Moreover, there should be an increase in current giving, as taxpayers shift their donations to the present in anticipation of higher costs in the future. Normally transitory price elasticities would likely be higher than permanent ones although the evidence presented in Appendix A is mixed. But even at the elasticity of 0.5, the reductions in Table 1 should be turned into increases if taxpayers make donations now (i.e., a 0.8% increase in giving).

Transitory price effects would also occur if the increase in top tax rates in 2011 were to be confirmed by legislation making all but those tax cuts permanent, without enacting the itemized deduction cap. Because taxpayers would be certain that rates would be higher in 2011, and thus the tax price lower, they would defer current giving. The magnitude of this effect would depend on the extent that taxpayers did not already expect those tax increases. Thus, from the perspective of a price effect, the price today for the top rate taxpayer is1.06 ((1-.35)/(1-.396)) as compared to the future or 6% higher; for the second-highest bracket the price would be 4.7%. At a maximum, at an elasticity of 0.5 contributions would fall by about 0.4%.

Income Effects

The results in Table 3 reflect only the changes in relative price, but there are other effects as well. There are income effects associated with the charitable giving deduction cap itself, which are small since on average charitable contributions are a small part of a taxpayer's budget, typically less than 3%.¹⁶ If the entire budget outline were considered, the income effects could be significant. Income elasticities, as noted above, are the percentage change in giving divided by the percentage change in income; they tend to be lower than one for necessities and higher than one for luxuries.

In calculating income effects, note that there is also uncertainty regarding the elasticity with respect to permanent income, and a range of estimates have been found. As with price elasticities, there are other kinds of evidence, and this evidence tends to suggest elasticities that are higher than the price elasticities used above. The line graph in Figure 1 indicates that over a very long period aggregate individual contributions were about the same relative to output, about 1.5%. Were the income elasticity less than one, a general downward trend would be expected, whereas with an elasticity greater than one an upward trend would be expected. For example, if income grew at 3% per year and the income elasticity was 0.5, the share of contributions should have fallen 1.61% in 1967 to 1.05% in 2007, whereas with a unitary elasticity it would have been the same; the actual ratio was approximately the same, 1.66%. If the only influence on charitable giving were income and price, this graph would be suggestive of a very low price elasticity and a unitary income elasticity.

Giving across individuals also appears to be relatively constant. Data presented for 1992 that included estimates for non-itemizers indicated that contributions were 4.4% for the lowest income class ($5,000 to $10,000), fell to 3.4% in the $10,000 to $15,000 class, and fell slowly ranging between 2.4% and 2.6% for all the classes between $30,000 and $1 million. Giving did rise at the very top income class as a percentage of income, to 3.1% from 2.6%. These data also tend to suggest a unitary elasticity of income; however, income may be correlated with other social factors that mean changes in income due to a tax cut could not be inferred from these data. (As an example, high-income individuals give a relatively small share of their income to religious organizations, while lower and moderate income individuals give more.)

Statistical studies are performed to control for other influences and for individual fixed effects (such as religiosity), but the income elasticities, (as can be seen in the Appendix A) vary considerably. They tend on average to be above the price elasticities. Estimates in this report use an elasticity of one.

For high-income individuals, these income effects are expected to be negligible compared with current law (where taxpayers could lose from the itemized deduction cap, but gain from the lower inframarginal tax rates, that is, tax rate on the first increments of income, and lower taxes on dividends). The Tax Policy Center estimates a 0.3% gain in income of the top percentile, which includes the estate tax.¹⁷ However, there are income gains for the entire population that average overall 3.5%; subtracting out a 0.2% overall gain from estate taxes, there is a gain due to all income taxes of 3.3%. This overall income effect would apply to all individuals, who constitute 74.8% of contributions. At a 1.0 elasticity, this would result in a 2.46% increase in giving.

For the comparison to current law, a 5.3% decline in income of the top 1% is projected because the lower marginal tax rates will not be extended, the cap on itemized deductions is imposed and the capital gains and dividends tax rates are raised.¹⁸ However, other tax cuts in the administration's package would lead to an overall effect of zero, and there should be overall no income effects.

The overall effects are reported in Table 4. Including all income tax effects, and taking into account gifts of appreciated assets, the effect on charitable giving compared with current tax rates would be a reduction of less than 1%. Compared with 2011 law, charitable giving should rise by about a 1% due to the proposals.

Table 4 does not report sensitivity analysis, although the only differential effect from Table 3 is for the comparison with 2011. Income elasticities reported in Appendix A, outside of those that were not statistically significant, range from 0.4 to 1.3, resulting in income effects of 1% and 3.2% and a net, using a 0.5 price elasticity of a decline of 0.44% or an increase of 1.76%.

Effects by Types of Charitable Objectives

Different types of charities may be affected differently by the change because the negative effects are more concentrated on higher-income donors. Higher-income donors contribute larger shares of their donations to contribute to health, education, art, environmental, and similar organizations, and less to religious organizations, those meeting basic needs, and combined purpose organizations.

The different types of recipients for all contributions are shown in Figure 3. If higher-income individuals contributed the same shares as overall contributions, the estimated effects in Table 4 would apply to all charities. However, evidence indicates that patterns for giving for high-income individuals differ from those of overall giving.

Figure 3. Donation by Type of Charitable Organization, 2007 (in $ billions)

Source: Data from Center on Philanthropy, Giving USA 2008.

Table 5 uses data in a study on high-net-worth philanthropy to estimate the differential effects across different types of organizations, combining both price and income effects (as shown in Table 4).¹⁹ (See Appendix B for the methodology used.) The study also estimated the shares of each type of charity that were directed to the needs of the poor; in general, higher-income donors give a small share of their donations to charities that benefit the poor.

As the table indicates, charities that are not relatively favored by high-income donors will have very small reductions or gains, depending on the comparison made. These include religious organizations, combined purpose charities and charities designed to meet basic needs. Organizations that are more likely to be recipients of donations from high-income individuals will be more likely to have reductions in gifts and those reductions will be larger. Health will experience the greatest declines, followed by arts, other, and education.

Estate Tax Issues

The modern U.S. estate tax was enacted as part of the Revenue Act of 1916 and almost from its onset a deduction for charitable bequests was allowed.²⁰ As noted above, charitable bequests make up approximately 8% of all charitable giving. As with the more familiar income tax deduction, the estate tax deduction for charitable bequests reduces estate tax liability by the dollar value of the charitable bequest times the estate tax rate. Since 1916, the top marginal estate tax rate trended upward (increasing the value of the deduction) for roughly 65 years, before reversing course and trending downward (reducing the value of the deduction) over the past 25 years. Under current law, the estate tax has been repealed in 2010, before reappearing at its 2001 level the following year, although there are legislative proposals to revise that. These fluctuations in estate taxation could have important implications for the level of charitable bequests in the coming years. In contrast, the President's FY2010 and FY2011 Budget Outlines maintain the estate tax in its 2009 form and the Senate Budget Resolution for FY2010 reduces the top marginal tax rate by 10 percentage points.

Unlike the personal income tax, the deduction for charitable bequests is not subject to a cap. That is not to say that similar distributional issues are not present. In fact, the benefit of the deduction for charitable bequests against the taxable estate is more heavily skewed towards higher wealth estates than under the income tax. This follows from the structure of the estate tax which has historically had a high exemption, resulting in a narrow base of roughly 2% of adult deaths each year since 1916.²¹ For this narrow base, however, the tax (subsidy) rate has been historically quite high, though trending downward from 70% to 45% over the part quarter century.

As in the case of the income tax discussion, this section examines historical changes in the estate tax, the estimated effect of different regimes on charitable giving through bequests, the effects of temporary changes, and the effects of the estate tax on gifts during the lifetime. It also discusses what types of charities might be most affected.

Historical Changes in the Estate Tax

Some insight into the expected impact of a change in the estate tax might be found by comparing tax changes to past tax revisions. The price of charitable bequests for estates subject to the estate tax is (1-t), where t is the tax rate at which contributions are deducted. For example, for an estate in the 45% tax bracket, the tax price is 0.55, indicating that the estate tax reduces bequeathable wealth by 55 cents for each dollar of charitable bequest.

As shown in Figure 4, the top marginal estate tax rate has fluctuated considerably since 1916, ranging from rates as low as 10% to rates as high as 77%. The estate tax's origin as a revenue source in times of crisis can be observed in its early years by spikes in the top rates which correspond, roughly, to World War I and World War II. The gradual shift away from this original purpose may be observed by the persistence of the World War II rate increase lasting through 1976. Beginning with the Tax Reform Act of 1976 (TRA76), the top rate has been consistently reduced.

Figure 4. Top Marginal Estate Tax Rate, 1916-2011

Source: IRS. Notes: Rates for 2010 and 2011 are based upon current law.

Table 6 examines how these reductions in the top marginal estate tax rate have affected the tax price of charitable bequests. Taken together, the tax price of charitable bequests has risen 139% since the enactment of TRA76. The result is that individuals engaged in estate planning face a much reduced incentive for charitable bequests than 40 years ago.

Figure 5 shows the pattern of charitable bequests as a percentage of GDP over the 1967-2007 time period. The pattern of bequests shows some evidence of shifting in response to tax changes. While little can be said concerning the 1976 tax change, it appears that estates may have increased charitable bequests prior to the tax price increases brought about by the 1981 and 2001 reductions in the top marginal estate tax rate. This evidence is harder to pin to the tax changes, compared to changes in the personal income tax rate, as the timing of bequests is tied to the timing of death, which is not generally thought to be strongly correlated with taxes. It is also worth noting that relative to GDP, bequests have generally stayed within a narrow range over this time period. For example, both the 1969 and 2000 levels of bequests were 0.2% of GDP. Given the inability to smooth charitable bequests over time and narrow affected populations year-to-year variation in bequests as a percentage of GDP is not unexpected.

Figure 5. Bequests as a Percentage of GDP Compared with the Top Marginal Estate Tax Rate

Source: Giving USA 2008 and IRS.

Estimated Effects on Aggregate Charitable Giving

To estimate the effect on charitable giving of changes in the estate tax, four elements are required: the share of bequests affected, the percentage change in price for those bequests, the price elasticity and the wealth elasticity. Although changes in either the estate tax rate or the estate tax exemption are expected to elicit behavioral responses, estimates in this study are confined to the effect of changes in the estate tax rate in this study.²²

Share of Contributions Affected

As shown in Figure 6, a relatively small share (4% to 6.4%) of total charitable contributions would be affected by the repeal and reinstatement of the estate tax (which affects approximately 2% of decedents). Using Giving USA 2008 and SOI data, approximately 53% of all charitable bequests in 2007 were made by estates subject to the estate tax.²³ In addition, another 32% of all charitable bequests were made by estates that filed estate tax returns, but were not liable for the estate tax.²⁴ Taken together, between 53% and 85% of all charitable bequests may be influenced by the estate tax. Thus between 4% and 6.4% of all charitable contributions may be affected by estate tax rate changes.

Figure 6. Share of Total Charitable Giving Affected by Changes in the Estate Tax Rate, 2007 (in $ billions)

Source: Center on Philanthropy, Giving USA 2008, and IRS and CRS calculations.

Percentage Change in Price

The price of charitable bequests for estates subject to the estate tax is (1-t), where t is the tax rate at which charitable bequests are deducted from the gross estate. For example, for an estate in the 45% tax bracket, the tax price is 0.55, indicating that the estate tax reduces bequeathable wealth by 55 cents for each dollar of charitable bequest. Table 7 illustrates how the tax price of charitable bequests under current law compares with those under the President's Budget Outline. Effects and the FY2010 Senate Budget Resolution are calculated compared to the President's Budget Outline or, equivalently, 2009 law.

Price Elasticities

The third element needed to estimate the effect on aggregate charitable contributions due to changes in the price of charitable bequests is a measure of the responsiveness of bequests to changes in the estate tax rate. As mentioned above, these measures are generally reported as elasticities: the percentage change in charitable bequests divided by the percentage change in price (with price equal to (1-t)).

Most evidence on these elasticities comes from analysis of tax or probate records. Unlike studies of the personal income tax, all of these studies are essentially cross-sectional, since death is a one-time occurrence. In addition, these studies focus on the charitable bequests of the second-to-die spouse, since the incentives are difficult to specify for a first-to-die spouse given the existence of the marital deduction.²⁵ As noted above, a visual examination of historical charitable bequests and the top marginal estate tax rate provides qualified evidence of the impact of estate taxes on charitable bequests. This suggests that the behavioral response to a change in the estate tax is relatively elastic.

Another source of information on the responsiveness of bequests to changes in the estate tax rate is the expectations of high net worth households. In a recent study that surveyed high net worth households' philanthropy, one question asked was whether tax payers would reduce their charitable giving if the estate tax were repealed.²⁶ Fifty-three percent of respondents said they would not reduce their donations at all, 7.8% said they would somewhat decrease donations and only 2.1% would dramatically decrease donations. Given that a repeal of the estate tax would result in an 82% increase in the tax price of charitable bequests, this suggests that the behavioral response to a change in the estate tax is relatively inelastic, according to survey respondents.

Given the mixed message of these types of evidence, more sophisticated empirical evidence is considered. In this area, economists have generally found large behavioral responses to changes in the tax price of charitable bequests.²⁷ These studies use either probate records or estate tax returns to look at how individuals respond to estate taxes when planning their estates and the studies identify the tax price by way of either the rate schedule within a year or changes in the rate schedule over time. Note that unlike the studies of individual behavior, estate behavior cannot be examined by following an estate over time.

Table 8 shows the results of applying a range of price elasticities, representing the high, low and central estimates from the studies. Compared to the President's Budget Outline, the central estimate suggests a reduction of over 7.2% in total charitable contributions from a repeal of the estate tax, an increase of nearly 1.6% from reverting to 2001 estate tax law and a decrease of nearly 1.6% from adopting the Senate Budget Resolution. The low elasticity estimates, however, better fit the observed historical trend in Figure 5. Compared to the President's Budget Outline, the low elasticity estimate suggests a reduction of nearly 4% in total charitable contributions from a repeal of the estate tax, an increase of almost 1% from reverting to 2001 estate tax law and a decrease of nearly 1% from adopting the provision in the Senate Budget Resolution.

Current Effects of a Temporary Change in the Estate Tax

A note of caution is required to interpret the estimates presented above. Uncertainty surrounding the timing of death combined with the temporary nature of the estate repeal, under current law, could lead to decreased tax planning and subsequently changes in charitable bequests. Taken together with the President's outline to maintain the 2009 form of the estate tax the estimates are likely greater, in absolute value, than what expected results if current law were followed.

Effect of Change in Estate Taxes on Lifetime Charitable Contributions

Changes in the estate tax impact charitable giving through an effect on lifetime giving. A small literature has developed around this question and has generally found that lifetime giving to charities responds in a relatively inelastic manner to changes in the tax price of bequests.²⁸ In particular, a permanent repeal of the estate tax could reduce lifetime charitable giving by roughly 5.7%. That is, repealing the estate tax would raise the cost of making lifetime charitable contributions while alive, relative to the cost of giving gifts to heirs at death. In contrast, reverting to the 2001 estate tax law would increase lifetime charitable contributions by nearly 1.3% and following the Senate Budget Resolution would decrease lifetime charitable contributions by nearly 1.3%.

Wealth Effects

The results presented in Table 8, above, only reflect the change in relative price brought about by a change in the estate tax rate. A change in the estate tax will, however, also change after-tax bequeathable wealth. While some uncertainty surrounds estimates of the elasticity of charitable bequests with respect to wealth, the use of multiple information types allows a consensus to emerge.

One source of information is the examination of the level of charitable bequests over an extended time period. As observed in Figure 5, charitable bequests as a percentage of GDP tracked within a narrow range suggesting a wealth elasticity of near unity. By way of comparison, an upward/downward sloping line would have been indicative of a wealth elasticity greater/less than unity. If the only influence on charitable bequests were wealth and price, this figure would suggest a unitary wealth elasticity and price elasticity greater than unity. This observed wealth elasticity, however, may be influenced by other social factors.

Statistical studies which estimate the wealth elasticity, and attempt to control for other influences for charitable bequest, vary widely. Overall, the studies find wealth elasticities that are below, in absolute value, the price elasticities. Given that the estimates straddle unity, it is used in the calculations.

Some economists argue that the graduated rate structure of the estate tax virtually ensures that the price effect will dominate the wealth effect and that if the estate tax were assessed at a flat rate the wealth effect would dominate.²⁹ This assertion fits with the elasticities found in the economics literature and, thus, it is not surprising to find this same pattern in the overall effect reported in Table 9. Specifically, compared with the President's baseline repealing the estate tax would reduce total charitable contributions by nearly 4%, or a 50% reduction in charitable bequests. As way of comparison reverting to 2001 law would increase total charitable giving by close to 1%, while adopting the provision in the Senate Budget Resolution would decrease charitable giving by nearly 1%. Evaluating these alternatives at our lower price elasticity, the resulting total effects are significantly closer to 0, or -0.66%, 0.15% and -0.15% of total charitable contributions, respectively.

Effects of Changes in the Estate Tax by Types of Charitable Objectives

In addition to the overall effect, categories of charities are likely to be affected differentially. As mentioned above, higher-income donors contribute larger shares of their donations to health, education, art, environmental, and similar organizations, and less to religious organizations, those meeting basic needs, and combined purpose organizations during their lifetimes. Figure 7 breaks out charitable bequests by type of charity. Compared to lifetime giving, foundations receive a much greater share of charitable bequests. Of the remaining charities, only education also rises, in percentage terms relative to lifetime giving, whereas the arts and health organizations hold up better than religion or human services organizations.

Figure 7. Charitable Bequests by Type of Charitable Organization (in $ billions)

Source: Center on Philanthropy, Giving USA 2008.

Because of the large share of bequests that involve foundations, the effects on actual charitable spending changes are likely to be delayed as foundations tend to retain assets and spend in the future. (These effects can also occur with inter-vivos giving where gifts can be made to foundations, to endowments and supporting organizations, and through donor advised funds, all of which delay the spending of charitable funds; the importance of this effect is likely to be more limited, however, because of the smaller share of foundation giving in lifetime contributions.) The ultimate timing and distribution of contributions to foundations is difficult to determine, although they may likely be devoted to the same ultimate purposes as are generally favored by high-income individuals' lifetime giving.

Of course, when evaluating the effect of policy changes on different types of charities the initial share of bequests is only one determinant of the effect, with the relative elasticities for each charity being a second factor. According to one study, bequests to the arts and social welfare organizations are relatively unresponsive to changes in the tax price, while bequests to education and medical organizations, along with religious organizations, are relatively responsive.³⁰ Although these results generally conform to observed bequest patterns, the tax price effect on education appears to be an anomaly, likely due to aggregation bias.

Policy Options

Charitable contributions have a public goods aspect, which means they are undersupplied in a market economy, providing a justification for government subsidies to private contributions, provision of grants, or direct provision of goods (such as assistance for the poor, health research and care, education, and other goods).

The argument typically made for limiting itemized deductions, such as those for charitable giving, points to the differential subsidies for some individuals as compared to others due to the difference in marginal tax rates. In addition, if the price elasticity is less than one, the government loses more in revenue than the charity gains in donations, an attribute that is referred to as inefficiency.

Alternative policy options might be to find a different source of revenue. The type of revenue source could reflect efficiency goals (which might include limits on itemized deductions) or distributional objectives (such as raising revenue from higher-income taxpayers).

One approach if the concern about the itemized deduction cap is primarily directed at charitable contributions is to exclude charitable deductions from the cap. According to Internal Revenue Service Statistics of Income for 2006, charitable contributions accounted for 23% of itemized deductions for the $500,000-and-over adjusted gross income class, representative of the group of taxpayers affected by the itemized deduction cap. Thus about a quarter of revenue projected from the itemized deduction cap would be sacrificed by excluding charitable contributions.

A revenue-raising alternative to the cap on itemized deductions is to impose a floor or ceiling, allowing deductions only in excess of a percentage of income or directly limiting the amount of deductions that can be taken.³¹ (There is a ceiling of 50% of adjusted gross income for charitable gifts, and smaller ceilings for other types of gifts, such as gifts of appreciated property. Ceilings already exist for mortgage interest deductions based on the amount of the mortgage, and percentage of income floors exist for medical expenses, casualty losses, and miscellaneous deductions.) A floor would likely preserve the incentive effects at the margin of charitable contributions. According to the 2004 Statistics of Income Public Use File, 96% of contributions are made by taxpayers who contribute at least 1% of income and 88% are made by those who contribute at least 2% of income. The overall cap on itemized deductions is projected to raise about $31 billion when fully effective in FY2012; thus the annual cost of excluding the charitable deductions would be about $7 billion. CBO estimated that a 2% floor on charitable deductions would raise about $25 billion. Thus, the revenue could be made up with a lower floor.

A more limited option would be to impose a floor under the higher tax rates for charitable contributions. That is, the general cap could be imposed, but itemized deductions in excess of some percentage of income would be excluded.

If the main objective is to raise revenue from higher-income taxpayers, an alternative is to increase the top rates or add a surcharge, a proposal that was made by then-Chairman Rangel in his 2007 tax reform proposal.³² Raising top income rates further would increase the subsidy to charitable giving, although a surcharge based on adjusted gross income (i.e., pre-deduction income) would not. Other provisions to raise revenue that affect high-income individuals would be to increase the tax rate further on dividends (which prior to 2003 were taxed at ordinary rates) and capital gains (taxed at various rates in the past). Another provision that would affect higher-income individuals and encourage charitable giving would be to return to the pre-2001 treatment under the estate tax.

Conclusion

Concerns about dire consequences for charitable giving in the wake of a proposed cap on the rate at which itemized deductions are valued appear to be overstated. Depending on which set of tax rates are used as the baseline, charitable deductions would decrease by less than 1% or less than 2%, and if income effects are incorporated, the budget outline taken as a whole would increase charitable giving. Moreover, since the change is not scheduled to take place until 2011 it would not reduce current giving and, indeed, could increase it.

The provisions on the estate tax also have consequences for charitable giving that could be more important per dollar of revenue. The retention of the current estate tax rates compared with the elimination of the tax now scheduled for 2010 would increase charitable giving, but would lower it compared with the pre-2001 rates scheduled to be in effect if the Bush tax cuts expired.

The effects of the cap on itemized deductions and any changes in bequests that might arise differentially affect certain types of charities.

Appendix A. Evidence on Elasticities for Inter-Vivos Giving

Table A-1 reports the results of six different studies (with a number of specifications which attempt to measure both permanent and transitory effects of changes in price and income on charitable giving.³³ Two of these studies (Bakija, and Bakija and McClelland) also provided some critiques of other studies and some sensitivity analysis that is useful in understanding the studies and their strengths and weaknesses. Results that are not statistically significant have an asterisk. Lack of statistical significance means that, although a relationship that most closely fits the data is estimated, there is such deviation from that relationship in the observations that there is not a clear causal effect. They are usually, although not always, associated with very small values that are close to zero.

While the studies differ in methodology, as discussed below, one difference is the type of data used. Tax return data are available for general use only to researchers in the Treasury Department and the Joint Committee on Taxation. (The Congressional Budget Office, CBO, has access to taxpayer data but must have uses approved by the Joint Committee on Taxation.) The data on giving and tax rates are probably superior in these studies and contain a larger sample of high-income taxpayers; however, such research cannot be replicated or subjected to any sensitivity analysis by others. Other researchers have to use public use data constructed from other sources. Of the six studies in Table 3, three (Randolph, Auten et al., and Bakija and Heim) used taxpayer data and all had as authors or coauthors a Treasury employee. The Bakija and McClelland study, with a CBO coauthor, included a sensitivity analysis for the Auten et al. study, but used a public use file, not the tax data. The other two studies also used a public use file.

Many of the studies listed below report multiple results using different specifications and, in general, an attempt is made to report the results that appear to be preferred by the author(s).

For comparison with this table and to illustrate the importance of dealing with transitory effects, Bakija and McClelland, who presented a range of strategies, also estimated a standard pooled cross section estimate, the type that had been done prior to the evidence shown by the 1980s tax cuts that did not deal with transitory effects. That estimate showed results that are typical of past cross section studies, a price elasticity of -1.22 and an income elasticity of 0.84.

In general, the theoretical expectation is that transitory price effects are large and transitory income effects are small (due to the permanent income hypothesis or consumption smoothing). Price elasticities and income elasticities in cross section studies are a combination of permanent and transitory effects. Thus, a lower permanent price elasticity and a higher permanent income elasticity would be expected than those observed in cross section studies. Only two studies, Randolph (1995)and Bakija and Heim (2008) find these results, and the Bakija and Heim income elasticity is only marginally higher.

Randolph (1995) was the first study to focus on the problem of transitory effects, and the technique used a 10-year panel that treated deviations from average income (and the resultant deviations from tax rates) as transitory. Permanent tax rates varied through changes in the tax law (and years around the 1981 and 1986 changes were excluded). This study allowed a long period of time to be transitory; therefore, it is possible that some of the permanent price and income effects are reflected in the transitory estimates, as the author acknowledges. Other studies tend to allow much shorter-term transitory effects, which might go too far in the other direction. Randolph's model allowed the price elasticity to vary by the share of giving, and he reports two measures: one unweighted with a price elasticity of-0.08 which is not statistically significant and one weighted more heavily toward large contributors, which Randolph appears to prefer. The results in the Randolph study are consistent in general magnitude with the expectations based on the aggregate data discussed in the text: a small permanent price elasticity, a large transitory price elasticity, an income elasticity of around one and a smaller transitory income elasticity.

Bakija (2000), who among other things, replicates the Randolph results with public use data, argues that the second weight, which yields an insignificant price elasticity, is more appropriate (although he criticizes other aspects of the model). In his own replications with public use files he finds effects similar to Randolph's unweighted results but suggests the appropriate measure of the aggregate elasticity evaluated over the full sample. These results are similar to Randolph's unweighted results: he also finds similar results for the elasticity when confined to incomes over $100,000.³⁴ Based on the specification he prefers and his replication, this approach basically finds no evidence of a permanent price response.

The Randolph study differs from the other studies in some important ways. By using average income over the panel as permanent income and estimating transitory effects based on deviations, he allows a broad scope for shifting over time, whereas other studies use shorter periods. This choice may be influenced by experience with capital gains realizations studies where using short periods to control for transitory effects were not successful in producing results that were reasonable.³⁵

Barrett et al. allow limited intertemporal shifting variation and also a lagged value of giving to deal with adjustment. They focus particularly on how quickly adjustment takes place, which they find to be very rapid. Their panel also does not include tax rate changes after 1986 which are an important exogenous source of variation. They find a lower price elasticity than a standard cross section, but also a small income elasticity. Like the other studies, this study includes individual fixed effects which are designed to control for heterogeneity among taxpayers (e.g. a taste for philanthropy, religiosity, etc). (Randolph could not employ individual fixed effects because he used an average over the entire panel for permanent income which was then indistinguishable from a fixed effect.) One drawback, however, of fixed effects, as Barrett et al. acknowledge, is that the fixed effect could also be picking up permanent income effects, and so suppressing the value of that elasticity. The Barrett et al. study also allowed a more limited scope for intertemporal substitution.

Auten et al. also use fixed effects and more limited intertemporal substitutions. As pointed out by Bakija and McClelland, they also had a problem in that they did not deal with known changes in the tax law (that is, 1986 was a higher-tax year than 1987 even though the high realizations in 1989 were associated with a pre-announced drop in tax rates) which would tend to bias their price elasticities up. This was a particular problem for panels that included 1986, and Bakija and McClelland reestimated their model using a public data file and found a much lower elasticity.

Bakija (2000) mainly contrasted his model with Randolph's by using legislated transitory changes in tax rates as the way to determine the transitory component of taxes. Bakija and McClelland base their analysis off Auten et al. and while they introduce a number of innovations, their main changes are to model expected tax changes and introduce adjustment lags.

Bakija and Heim use a panel approach with tax data, with fixed effects, with more limited substitution frameworks than Randolph, and with attention to expectations of tax changes. They characterize intertemporal substitution mainly through those pre-announced tax changes and allow shorter substitution periods. The main source of determining the price elasticity is the difference in response across taxpayers who had different changes in their tax rates. They also examine separate estimates for higher-income individuals. They obtain very different estimates depending on how they deal with fixed time effects (variables meant to control for changes that affect all observations in a given year), which cannot be introduced into the higher-income levels because they are so closely correlated across the sample with legislated changes in tax rates. The first one they reported, which did not use fixed time effects but incorporated a time trend, is included in the assessment.³⁶

Ultimately no study is perfect, and thus it is difficult to choose a central elasticity from among these. Excluding the high elasticities in Auten et al. for the panel that covers 1986 and that are likely overstated the elasticities range from essentially zero to 0.8. It seems likely that the unweighted Randolph estimate may be biased downwards, but some others may be biased upwards because of fixed effects or short periods for intertemporal substitution. Ultimately a center elasticity of 0.5 is used. Income elasticities may be biased in the other direction but should be above the cross section results for that era and typical of the past (0.84) so, for that reason, and for reasons stated in the body of the paper, a unitary elasticity is used.

Appendix B. Methodology for Estimating Effects by Type of Charity and Charitable Purpose

The first step in the estimation is to measure the ratio of the charitable giving share for each type of charity to the total share. For example, if affected donors give half a share to a given charity as their share overall, the price effects of the cap will have half the effect of the total. These shares were taken from Table 10 of the Center of Philanthropy's study of patterns of household giving.³⁷ Affected taxpayers fell into two categories in this study, $1 million or more and $200,000 to $1,000,0000. Shares were weighted based on the share of total giving in each tax-rate class, adjusted for limits on giving; the result was that 89% of the allocation was based on taxpayers with $1 million and over. Note that this methodology assumes that price elasticities are the same for different types of giving.

Once these ratios were obtained, they were multiplied by the overall price effects to obtain the reductions in demand due to price effects. To incorporate the income effects, data on the top 1% and the overall income effects, along with the share of income allocated to the top 1%, subtracting out the effects of the estate tax, were used. This income effect was used for their share of the contributions of each type, with the remaining 99% income effect used for the remaining contributions. For the comparison with current tax rules, income of the top 1% fell by 5.3% and income of the remaining 99% increased by 1.1%. For the comparison with law in 2011, there was a 0.1% increase in income in the top 1% and a 4.4% increase in income in the remaining 99%. Denoting "a" as the ratio of the share by high-income taxpayers of each category divided by the share of all contributions by these taxpayers. The weighted income effect was a times 0.1747 times the income effect of the top 1% + (0.75-a times 0.1747). The share of income in the high-income category is 17.47% and 75% of contributions are from individuals.

Appendix C. Evidence on Elasticities for Charitable Bequests

Table C-1 reports the results of seven different studies that attempt to estimate both the price and wealth elasticities of charitable bequests. Although these studies find a diverse set of estimated elasticities, they reach two common general conclusions: the price elasticity dominates the wealth elasticity and charitable bequests, generally, respond elastically to changes in the tax price of bequests. The exception to this second conclusion is provided by Greene and McClelland (2001) and is likely explained by their focus on the portion of the tax price related to the exemption level.

Footnotes

1.	Congressional Budget Office, Budget Options, 2009, p. 192, http://www.cbo.gov/ftpdocs/102xx/doc10294/08-06-BudgetOptions.pdf.
2.	Simple, Fair and Pro-Growth: Proposals to Fix America's Tax System, The President's Advisory Panel on Federal Tax Reform, November 2005.
3.	Urban Brookings Tax Policy Center, Table T09-0162 http://www.taxpolicycenter.org/numbers/displayatab.cfm?DocID=2235; Table T09-0144 http://www.taxpolicycenter.org/numbers/displayatab.cfm?DocID=2194.
4.	See Gerald Auten, James M. Cilke, and William C. Randolph, "Effects of Tax Reform on Charitable Contributions," National Tax Journal, Vol. 65, September 1992, pp. 267-290.
5.	Urban Institute and Brookings Institution Tax Policy Center, Charitable Contributions by Statutory Tax Rate, Table T09-0175, http://www.taxpolicycenter.org/numbers/displayatab.cfm?DocID=2249.
6.	David Joulfain, "Charitable Giving in Life and at Death," in William G. Gale, James R. Hines, Jr. and Joel Slemrod, Rethinking Estate and Gift Taxation (Washington, DC: Brookings Institution Press, 2001).
7.	Gerald E. Auten, Charles T. Clotfelter, and Richard L. Schmalbeck, "Taxes and Philanthropy Among the Wealthy," in Joel Slemrod, Does Atlas Shrug?: The Economic Consequences of Taxing the Rich (Cambridge: Harvard University Press, 2000).
8.	Based on Internal Revenue Service Statistics showing $186.6 billion of itemized deductions in 2006 and $294.9 billion in deductions as reported by Giving USA 2008, The Annual Report on Philanthropy for the Year 2007, prepared by the Center on Philanthropy at Indiana University, Giving USA Foundation, 2008.
9.	One provision that is not considered in calculating tax price changes is the phaseout of itemized deductions. Despite the term used to describe it, the phaseout of itemized deductions does not reduce the value of itemized deductions at the margin. It is triggered by an increase in adjusted gross income, and, if itemized deductions grow with income, as is commonly the case, its effect is to increase the effective marginal tax rate by 3% in a way that does not affect the subsidy. (The itemized deduction is itself two-thirds phased out in FY2009 and is scheduled to be fully phased out in 2010, but will be restored in 2011.) A simple model with the phaseout, L = t(y+.003(Y-Yb)–D), where L is liability, t is the tax rate, Y is income Yb is the point at which the phaseout begins and D is deductions, illustrates that the change in taxes with a change in D is t. Generally, state income taxes are enough to cause deductions to grow by 3% of income, but there may be occasional circumstances where deductions do not grow fast enough. In that case the itemized deduction phaseout would reduce the value of charitable deductions. No data are available on the size of this effect but it is likely to be small.
10.	In general, necessities for which there is no close substitute (such as insulin, water, or food) or commodities that take up a very small fraction of the budget should have low elasticities while goods that have close substitutes should have higher ones. As with many commodities, charitable contributions have features that don't place them squarely in one type or another; they do not have close substitutes and tend to be a small part of the budget for most people but they are not necessities in the sense that food and water are. (However, a necessity is in the eye of the beholder; religious or ethical beliefs or the interaction of charitable giving with social status may make charitable contributions more of a necessity than many other expenditures).
11.	Dean Karlan and John A. List, "Does Price Matter in Charitable Giving? Evidence from a large-Scale Natural Field Experiment," The American Economic Review, vol. 97, no. 5, December 2007, pp. 1774-1793.
12.	Daniel Rondeau and John A. List, Matching and Challenge Gifts to Charity: Evidence from Laboratory and Natural Field Experiments, National Bureau of Economic Research Working Paper 13728, January 2008.
13.	The Center on Philanthropy, The 2008 Study of High Net Worth Philanthropy, Sponsored by Bank of America, Indiana University-Purdue University, Indianapolis, March 2009.
14.	See Charles T. Clotfelter, "The Impact of the Tax Reform Act of 1986 on Charitable Giving: A 1989 Perspective," in Henry J. Aaron and William G. Gale, eds., Economic Effects of Fundamental Tax Reform (Washington DC: Brookings Institution, 1996).
15.	This number is smaller than that estimated by the Center for Budget Policy and Priorities, which estimated 1.3% primarily because they used an elasticity of 1. See Paul N. Van de Water, Proposal to Cap Deductions for High Income Households Would Reduce Charitable Deductions by Only About 1%, Center on Budget Policy and Priorities, http://www.cbpp.org/cms/index.cfm?fa=view&id=2700. As they indicate, a larger number estimated by Len Burman, Urban Brookings Tax Policy Center, is about 2% and compares the effects to the higher 2011 rates. The Tax Policy Center has also increased the estimated share of affected taxpayers since those estimates were made.
16.	See Charles T. Clotfelter, "The Impact of the Tax Reform Act of 1986 on Charitable Giving: A 1989 Perspective," in Ed. Henry J. Aaron and William G. Gale, eds., Economic Effects of Fundamental Tax Reform (Washington, DC: Brookings Institution, 1996).
17.	Urban Brookings Tax Policy Center, Table T09-0136, http://www.taxpolicycenter.org/numbers/displayatab.cfm?DocID=2220.
18.	Urban Brookings Tax Policy Center, Table T09-0138, http://www.taxpolicycenter.org/taxtopics/2010_budget_tables.cfm.
19.	Patterns of Household Charitable Giving by Income Group, 2005, prepared for Google by the Center on Philanthropy at Indiana University, Summer 2007.
20.	Frank J. Doti, "Estate Tax Repeal: Historical Data Indicates the Philanthropy May Suffer," Tax Notes, April 14, 2003.
21.	Darien G Jacobson et al., The Estate Tax: Ninety Years and Counting, U.S. Department of the Treasury, Internal Revenue Service, Statistics of Income Division, SOI Bulletin, Washington, DC, Summer 2007.
22.	See Congressional Budget Office, The Estate Tax and Charitable Giving, July 2004, for a discussion of how changes in the estate tax exemption affect charitable giving and the magnitude of the behavioral responses. In addition, the treatment of capital gains is expected to have a minor impact on charitable bequests, due to the low tax rates on capital gains. The expected effect is further reduced, since capital gains are relevant only to the share of charitable bequests that arise from investment appreciation and the likelihood that heirs would not recognize all of the appreciation immediately.
23.	The Center on Philanthropy at Indiana University, Giving USA 2008: The Annual Report on Philanthropy for the Year 2007 (Giving USA Foundation, 2008) and U.S. Department of the Treasury, Internal Revenue Service, Statistics of Income Division, Estate Tax Returns: By Tax Status and Size of Gross Estate, October 2008, http://www.irs.gov/pub/irs-soi/07es01fy.xls.
24.	Estates are required to file an estate tax return if the gross estate is greater than the exemption level of $3,500,000 in 2009. Deductions for spousal bequests, charitable bequests, funeral expenses and discharge of indebtedness are subtracted from the gross estate to calculate the taxable estate, upon which the estate tax is calculated.
25.	Wojciech Kopczuk and Joel Slemrod, "Death and Dollars: The role of gifts and bequests in America," in Alicia Munnell and Annika Sunden, eds., Tax Impacts on Wealth Accumulation and Transfers (Washington, DC: Brookings Institution Press, 2003).
26.	The Center on Philanthropy, The 2008 Study of High Net Worth Philanthropy, Sponsored by Bank of America, Indiana University-Purdue University, Indianapolis, March 2009.
27.	Michael J. Brunetti, "The Estate Tax and Charitable Bequests: Elasticity Estimates Using Probate Records," National Tax Journal, vol. 58, no. 2 (June 2005); Jon M. Bakija, William G. Gale, and Joel B. Slemrod, "Charitable Bequests and Taxes on Inheritances and Estates: Aggregate Evidence from across States and Time," Papers and Proceedings of the Annual Meeting of the American Economics Association, vol. 93, no. 2 (May 2003); Michael J. Boskin, "Estate Taxation and Charitable Bequests," Journal of Public Economics, vol. 5, no. 1-2 (Jan./Feb. 1976); and David Joulfaian, "Estate Taxes and Charitable Bequests by the Wealthy," vol. 53, no. 2 (September 2000) are four recent examples of this literature. Pamela Greene and Robert McClelland, "Taxes and Charitable Giving," National Tax Journal, vol. 54, no. 3 (September 2001) in contrast, finds an inelastic behavioral response.
28.	Some prominent examples of this literature are Gerald Auten and David Joulfaian, "Charitable Contributions and Intergenerational Transfers," Journal of Public Economics, vol. 59, no. 1 (January 1996); Congressional Budget Office, The Estate Tax and Charitable Giving, July 2004, http://www.cbo.gov/ftpdocs/56xx/doc5650/07-15-CharitableGiving.pdf; and David Joulfaian, "Charitable Giving in Life and Death," in William Gale and Joel Slemrod, eds., Rethinking Estate and Gift Taxation, (Washington, DC: The Brookings Institution, 2001).
29.	Wojciech Kopczuk and Joel Slemrod, "Death and Dollars: The Role of Gifts and Bequests in America," in Alicia Munnell and Annika Sunden, eds., Tax Impacts on Wealth Accumulation and Transfers (Washington, DC: Brookings Institution Press, 2003).
30.	David Joulfaian, "Charitable Bequests and Estate Taxes," National Tax Journal, vol. 44, no. 2 (June 1991).
31.	Ceilings and floors for charitable deductions were discussed in proposals in past Congresses to expand the deduction to non-itemizers. See CRS Report RL31108, Economic Analysis of the Charitable Contribution Deduction for Non-Itemizers, by [author name scrubbed].
32.	See CRS Report RL34249, The Tax Reduction and Reform Act of 2007: An Overview, by [author name scrubbed].
33.	The studies are William Randolph, "Dynamic Income, Progressive Taxes, and the timing of Charitable Contributions," The Journal of Political Economy, vo. 103, August 1995, pp. 709-738; Kevin Stanton Barrett, Anya M.McGuirk and Richard Steinberg, "Further Evidence on the Dynamic Impact of Taxes on Charitable Giving, National Tax Journal, vo. 50, June 1997, pp. 321-334; Jon Bakija, "Distinguishing Transitory and Permanent Price Elasticities of Charitable Giving with Pre-Announced Changes in the Tax Law," October 2000,Mimeo; Gerald E. Auten, Sieg Holger, and Charles T. Clotfelter, "Charitable Giving, Income and Taxes: An Analysis of Panel Data, American Economic Review, Vol. 92, March 2002, pp. 371-382; Jon Bakija and RobMcClelland, "Timing vs. Long-Run Charitable Giving Behavior: Reconciling Divergent Approaches and Estimates," December 2004, Mimeo; Jon Bakija and Bradley Heim, "Does Charitable Giving Respond to Incentives and Income? Dynamic Panel Estimates Accounting for Predictable Changes in Taxation," National Bureau of Economic Research Working Paper 14237, August 2008.
34.	He finds a positive but insignificant permanent price elasticity of 0.322, a transitory price elasticity of -1.202, a permanent income elasticity of 1.188, and a transitory income elasticity of 0.195. For incomes over $100,000 he finds an insignificant permanent price elasticity of -0.155, a transitory price elasticity of -0.744, a permanent income elasticity of 0.611 and a transitory income elasticity of 0.145.
35.	For a discussion of some of these issues, see two reprinted CRS reports by [author name scrubbed], "Can A Capital Gains Tax Cut Pay for Itself," Tax Notes, vol. 48 (July 9, 1990), pp. 209-219; and "Limits to Capital Gains Feedback Effects," Tax Notes, vol. 51 (April 22, 1991), pp. 363-371. As these two reports taken together show, studies of capital gains realizations with short intertemporal effects continued to produce the high elasticities that appeared much larger than reasonable, given that realizations cannot exceed accruals.
36.	When they used time dummies for the higher-income sample they got results much like cross section results suggesting they were identifying effects in a similar way.
37.	Patterns of Household Charitable Giving by Income Group, 2005, Prepared for Google by the Center on Philanthropy at Indiana University, Summer 2007, p. 15.