Taxes Have a Potent Effect on Charitable Giving

Editor’s Note: President Obama has proposed a plan to reduce the value of the itemized deductions that wealthy people can claim for charitable gifts. To help explain the impact of Mr. Obama’s plan, The Chronicle asked four prominent economists to offer their analysis of how a change in the charitable deduction would affect giving. Following is an essay by Charles T. Clotfelter, professor of public policy, economics, and law at Duke University.

In light of the apparent incongruity between giving and the kind of self-interested behavior usually examined by economists, it is natural to wonder about what motivates people to make charitable gifts in the first place.

Yet economists, by and large, have paid little more attention to this ticklish question than they have to the reasons why households buy apples or oranges, pay to heat their homes, or take vacations. Instead, economists simply take preferences for these items as a given, leaving deeper explanations to the psychologists. Nor do they pay much attention to what people say about their motivations, preferring simply to observe behavior.

As they often do when analyzing consumption, economists tend to concentrate on the effects of two factors—income and price—on the quantity purchased. The approach taken in the case of charitable contributions is exactly analogous.

Income is defined as it often is in other applications, as disposable income—that is, after subtracting taxes. “Price,” however, takes on a rather special definition in the case of contributions. Owing to the deduction for contributions that is available in the income tax to those itemizing their deductions, a dollar donated is not exactly a dollar forgone, in terms of consumption.

An itemizer who is subject to a marginal tax rate of 20 percent, for example, sacrifices only 80 cents worth of consumption for each dollar contributed, owing to the reduction of 20 cents in tax for each dollar given away.

By virtue of the tax deduction, then, the net-of-tax price of giving is 80 cents per dollar. Thus taxpayers in higher tax brackets face lower net-of-tax prices of giving than taxpayers in lower brackets, if they itemize their deductions, while taxpayers who do not itemize (and therefore who receive no deduction for contributions) face a price of one dollar for each dollar they contribute.

Once defined, the income and price variables are then incorporated into statistical models to determine their quantitative effect on the amount given. As is the case with empirical studies of consumer demand for automobiles or food products, the structure of the models utilized to explain contributions does not suggest that price or income determines whether an individual makes contributions. Rather, the models embody the built-in assumption that these factors merely influence the amount contributed. Nor do the models exclude the possibility that contributions are influenced—even heavily influenced—by personal characteristics or other factors. Just because it implies that taxes can affect the amount of giving, the economic model is not, as one commentator has written, “based on the mistaken assumption that the primary driving force behind charitable donations is the income-tax deduction.”

Analysts have applied statistical methods to estimate behavioral parameters corresponding to these effects on giving. The models that have been estimated differ in specific mathematical form, but the primary focus in all of them has been to determine the quantitative importance of the effect of the two variables of primary interest: price and income.

Almost invariably, researchers have found it convenient to express these effects in terms of elasticity, defined as the percentage change in the amount given associated with a 1-percentage change in price or income. For example, an income elasticity of 0.80 implies that if Mr. Jones has an income percent higher than Mr. Smith, an otherwise similar individual, Mr. Jones will, on average, contribute 8 percent more than Mr. Smith.

As is the case in virtually every empirical application in economics, the precise estimates for the price and income elasticities differ from one study to the next. However, they have tended to cluster, which permits some degree of generalization.

Moreover, the clustering appears to have shifted over time. In studies published before 1990, most estimates of the price elasticity were in the range of -0.5 to -1.75, and most estimates of the income elasticity were in the range of 0.4 to 0.8.

More recent studies, however, have tended to produce larger income elasticities and smaller (in absolute value) price elasticities. To illustrate the application of these elasticities, consider -1.0 as a representative value for the price elasticity. This value implies that a 10-percent increase in the price of giving, for example from 0.60 to 0.66 (occasioned by a drop in the marginal tax rate from 0.40 to 0.34), would be associated in the long run with a 10-percent lower level of contributions. If 0.8 is taken as a representative value for the income elasticity, as in the example above, this would suggest that a 10-percent increase in after-tax income would imply a long-run level of giving 8 percent higher. The qualifying term “long run” is added here because the estimated models yielding these parameters are generally thought of as applying to long-run, equilibrium behavior, not necessarily to immediate responses to changes in these variables. All the same, parameters such as these imply that taxes have a potent effect on charitable giving.