[A] large share of survey respondents resist full equalization of after-tax incomes even when conventional optimal tax analyses would strongly recommend it. In a hypothetical situation meant to mimic the tax policy problem, between 50% and 95% of respondents choose not to fully offset inequality due to brute luck even when there are neither efficiency costs of redistribution nor differences in desert across individuals. These choices suggest that the two reasons why conventional optimal tax analyses tolerate after-tax inequality—the importance of encouraging effort and the possibility that some people "choose" to have low incomes—are not the only reasons why survey respondents, and perhaps Americans in general, accept it.
The second finding offers an explanation of the first: a large share of survey respondents prefer an alternative logic for taxation than that which is typically used in optimal tax analyses. The conventional logic stems from the use of a social welfare function that exhibits diminishing marginal social welfare of income. When presented with two possible justifications for their choices in the tax problem, between 62% and 79% of respondents prefer, instead of this logic, one tied to a centuries-old idea that Richard Musgrave (1959) named classical benefit-based taxation (CBBT). Under CBBT, taxes are assigned based on the benefit a taxpayer obtains from the activities of the state, with benefit being measured by the state's role in increasing the taxpayer's economic opportunities. In addition to being Adam Smith's first maxim of taxation, CBBT has a long history in public debate over taxes in the United States, from its use as a justification for the new personal income tax in 1913 to its use by presidents Franklin Delano Roosevelt and Barack Obama to advocate for progressivity. In that context, finding support for CBBT among the American public is natural, despite its absence from modern optimal tax theory.