Gifts to government

During the French Revolution members of the Assembly appealed to the people to make gifts (dons de patriotisme). MacFarlane (1844) relates that “nearly all the members of the Assembly took their silver and gold buckles from their shoes and offered them up on the altar of the country,” followed by women who “brought their spoons and their forks, wedding-rings and nick-nacks.” Even “the prostitutes of Paris and Versailles...offered a share of their earnings....” MacFarlane says that “The epidemic, like other novelties, raged for a time, and then was over.” Mussolini introduced the “ Gold for the Fatherland” initiative, encouraging the public to voluntarily donate gold jewelry such as necklaces and wedding rings to government in exchange for steel wristbands bearing the words “Gold for the Fatherland.” His wife Rachele is reported to have donated her own wedding ring.

See Bank et al. (2008) for a magisterial history of US wartime “sacrifice.”

Kang and Rockoff (2015) discuss the US World War I experience, and Jones (1989) discusses American fiscal propaganda during World War II.

Polenberg (1972) notes that the sale of war bonds in the USA during World War II illustrated the fine line between voluntarism and compulsion. In 1942, most of President Roosevelt’s advisors favored a compulsory savings plan, but the President decided instead to institute a voluntary plan. Those who wanted a compulsory plan argued that a formal and impartial compulsory plan would in fact be less oppressive than the haphazard and unequal community pressure that would be applied to a “voluntary” program.

Levi (1997) addresses military service as one way in which democratic governments assert their power to tax, and examines why at some times and in some places there is widespread draft evasion and at other times and places there is considerable patriotism and volunteering. In this context, it is fascinating to note that, during World War I, in the USA a system was devised to make conscription look as much like volunteering—even like voting—as possible. Local civilian volunteers would first register eligible young men in much the same way as persons registered to vote; in fact, registration was even held at each precinct’s voting location (Ellis and Noyes, 1990, p. 190).

While in what follows we usually refer only to tax evasion, we also have in mind tax avoidance. In this paper, it is the act of underpaying taxes that matters regardless of whether it is done legally or illegally.

Although in the empirical analysis of this paper we focus on gifts made to the federal government, we recognize that gifts to sub-federal governments are non-trivial, and there are reasons to think they are larger than gifts to the federal government. Brunner and Sonstelie (2003) show that, in the late 1990s, in financially strapped California annual contributions to local schools and school districts regularly exceeded $100 per pupil and, in a handful of cases, $1000 per pupil. It makes sense that, in the context of schools and school districts, a small community with close interactions among families has a group identity that could lead to an informal, collective decision about how much to contribute and might also create a social norm that, for some people, overcomes the incentive to free ride on the contributions of others. We leave for future research the question of whether property tax evaders give gifts to their local governments.

We use the assumption that there is only one kind of public good in order to focus on the question of the co-existence of gifts and evasion while keeping the analysis simple. This assumption, however, limits our ability to distinguish between unconditional and conditional (targeted) gifts. There is empirical evidence that giving increases when the donor believes she has some say in what the (non-government) gift goes to. It is suggested, for example, by a field experiment involving alumni of an Ivy League university described in Kessler et al. (2019). They find that this group, especially the rich and powerful among them, increase giving by 100–350% when they are given a sense of agency regarding what purpose their gift might be used for. To incorporate this aspect of giving, in Appendix D we extend our model by allowing for two kinds of public goods as well as conditional and unconditional gifts.

Strictly positive values of x and n can be guaranteed by assuming infinite marginal utility of x at \(x=0\) and of n at \(n=0\).

In this case, the Lagrange multiplier \(\lambda \) is equal to zero because the constraint \(Dg+E\le y\) cannot be binding. A binding \(Dg+E\le y\) in this case would lead either to a violation of \(x>0\) or of \(n>0\).

Li et al. (2015) provide support for this assertion. In a real-donation laboratory experiment, they compare giving to the US federal general revenue fund with directed giving to particular government organizations, and find that directed giving more than doubles both the likelihood of giving and the size of contributions.

Our model ignores some factors that might affect the aggregate amount of gifts to government. One is that people can get warm glow not only from donating (and paying taxes) to government, but also from donating to a variety of private charities, and they choose those that best fit their preferences. Another is that many people do not have complete information on the menu of charities available, or even about the possibility of giving gifts to government. Moreover, people are more likely to give when they are asked: Information dissemination and solicitation play an important role in charitable donation. These aspects could be incorporated by extending our model and adopting the logic introduced in Andreoni and Abigail Payne (2003). Such an extended model would suggest that, if the government wants to stimulate the gifts to government, it should raise the warm glow from gifts to government relative to warm glow from private donations, by for example raising people’s trust in the government or improving the government’s efficiency. Note that, in periods of (popular) wars, the warm glow from gifts to government could be naturally elevated because of rise in the patriotism and pride. Additionally, raising the awareness of the opportunity to donate to the government funds would likely raise the number of gift givers.

Note that the model does not address the point made by Yitzhaki (1974) that a high tax rate may increase evasion if the cost of (detected) evasion is proportional to the tax evaded, rather than—as here—the income understated.

The budget constraint can be expressed as \(x^{*}+pn^{*}=\alpha y\), where \(p=\frac{1-Dt}{{\widetilde{a}}-bDt}\) and \(\alpha =1-t+bt\frac{1-Dt}{{\widetilde{a}}-bDt},\) where \({\widetilde{a}}\equiv a(1-Dt)\).

The collected data are available from the openICPSR, see Slemrod and Kuchumova (2021) for the complete reference to the data.

Internal Revenue Service (2006).

In 1947, right after World War II, $12 million in gifts was donated. In 1991, during the Gulf War, $6 million was donated. In 2001, at the start of the Afghanistan war, $27 million was donated.

Philipps (2005). This is an example of a direct (although lagged) link between tax evasion and gifts to government.

See, for example, Tyler (2006).

This is highly negatively correlated with the share of total federal expenditure dedicated to welfare and social insurance spending, and so the estimated coefficients can also be interpreted as (the negative of) the effect of the latter.

The budget constraint can be expressed as \(x^{*}+pn^{*}t=\alpha y\) where \(p=\frac{1-Dt}{{\widetilde{a}}-bDt}\) and \(\alpha =1-t-bt\frac{1-Dt}{{\widetilde{a}}-bDt}\) and \({\widetilde{a}}\equiv a(1-Dt)\).