Framing and Taxation: Evaluation of Tax Policies Involving Household Composition

Abstract

Three studies of attitudes toward tax policies were conducted on the World Wide Web. The results show several effects. In penalty aversion, subjects preferred bonuses over penalties, when policies differ only in how they are formally described. In the Schelling effect, subjects prefer both higher bonuses (for children) for the poor than for the rich and higher penalties (for being childless) for the rich than for the poor. In the neutrality bias, subjects preferred separate filing for married couples more when it was presented in a format that emphasized the effect of marriage (where it is neutral) than in one that emphasized the effect of the number of earners in a couple (where one-earner couples pay more). In the status-quo effect, subjects preferred the specified starting point to any change. Finally, in the metric effect, subjects favored more progressiveness in tax burdens when taxes were expressed in percent than when they were expressed in dollars.

The research suggests a general framework. Subjects approach a given decision problem with strong independent norms or ideals, such as, here, "do no harm," "avoid penalties," "treat likes alike," "help children," and "expect the rich to pay more." They then evaluate the problem on the basis of the most salient norms. In a complex area such as tax, independently attractive ideals are often in conflict, and the result is shifting, inconsistent preferences.

Note: This replaces an earlier version of the paper titled "Framing the Family: Normative Evaluation of Tax Policies Involving Household Composition," which circulated in SSRN journals in 2000.