SOME INVESTIGATORS HAVE CONCLUDED THAT the estate tax is important for maintaining public philanthropy (Auten and Joulfaian 1996; Boskin 1976; Clotfelter and Schmalbeck 1996; Bakija, Gale, and Slemrod 2003; Joulfaian 2000, 2001; Kopczuk and Slemrod 2003). (See Bakija and Gale (2003) for a review of these and other reports.) Most of these studies use multiple-regression estimates to show that the cost of donations (one minus the tax rate) is important in the donation decision. They find that the net effect of a rate reduction is to reduce total donations, even though after-tax wealth increases puts an upward pressure on donations. This partial-wealth effect is implicitly more than offset by the partial decline-in-donations effect induced by the increased cost of donations, i.e., a rate decrease, like an increase in the personal exemption, in their theory reduces tax-deductible benefits and hence reduces donations, while a rate increase enhances donations. We define this behavior as a negative wealth propensity. The declaration that most taxpayers, in terms of volume of donations, conform to this theory is labeled the negative-wealth-propensity hypothesis (NWPH). Below are the implications this view holds for marginal utility functions with respect to both donations and heir distributions. Similar predictions would also follow from eliminating the tax (Auten and Joulfaian 1996; Joulfaian 2000; Clotfelter and Schmalbeck 1996, who employed a simulation in their study). In fact, estimates of the total decline in donations, if the estate tax were eliminated, range from a low of 12 percent (Joulfaian 2000) to as high as 45 percent (Clotfelter and Schmalbeck 1996).
These studies identify tax deductibility as the principal driver of charitable donations; taxpayer wealth plays a secondary role. In an earlier study, Steuerle (1987) de-emphasizes this cost-of-donation effect. Barthold and Plotnick (1984) find practically no evidence of the cost of donations influencing charitable decisions.
The NWPH has a disturbing implication: if one assumes standard axioms of economic theory, and if the taxpayer has an optimum allocation between heirs and charities before a rate change, the change precludes achieving household equilibrium.
A more intuitively plausible hypothesis is that rate declines, or an increase in the personal exemption, encourage an overall net increase in charitable donations--the positive-wealth-propensity hypothesis (PWPH). This proposition is consistent with the well-known observation that donations rise with individual wealth and, of more importance, is compatible with axioms of accepted economic theory. The theories also differ in their emphasis. As implied above, the PWPH stresses wealth, while the NWPH emphasizes tax deductibility of donations.
To test these hypotheses, indifference curves are constructed that express individual preferences between donations and heir distributions, a set representing each hypothesis. These curves yield contradictory predictions: Do taxpayers generally behave as predicted by the PWPH, or as implied by the NWPH? Our empirical results support the former. Moreover, this approach also avoids the collinearity/ identification problem that multiple-regression investigators frequently face (as noted by Poterba 1998). Finally, it applies the long-neglected tool of indifference-curve analysis to questions of public policy.
In short, the main problem for multi-regression investigators was untangling the simple correlations between personal wealth, marginal estate-tax rates, and charitable donations. Their available tools were not powerful enough to compellingly identify the separate influences that wealth and tax rates have upon donations. Rather than study the general effects of estate taxes on donations, we address the important, but narrower question: Does a reduction in estate-tax rates increase overall donations? …