I critically consider four purported economic-efficiency arguments for egalitarian redistribution of income or wealth. (1) Jeremy Bentham’s “greatest aggregate happiness” criterion has been used (by Bentham, John Stuart Mill, Alfred Marshall, A. C. Pigou, Abba Lerner, and more recently Richard Layard) to argue for wealth transfers toward the poor based on the supposition that they register higher happiness from a marginal dollar. Drawing from Vilfredo Pareto and Lionel Robbins, however, I argue that modern economic theory is not about individual happiness, let alone aggregate happiness, and therefore does not support (nor refute) any happiness-based case for wealth redistribution. (2) Theories based on a “social welfare function” misapply the economic way of thinking in a different way. (3) Other writers have framed redistribution as a public good, and public goods provision by the state as a voluntary collective means of satisfying individual preferences, thereby using modern economic theory to formulate a rationale for redistributive policies based on Pareto-efficiency. I criticize this rationale for resting on suppositions about actual preferences that are self-immunized against falsification. (4) James Buchanan made a related case for taxing inheritances based on the supposition that in constitutional deliberation behind a veil of ignorance we would agree to such a policy based on our preference for a certain kind of fairness. I find this argument non-economic, equally non-falsifiable, and no more plausible than alternative suppositions about our common preferences. The economic way of thinking does speak clearly, however, about how taxes on income or wealth discourage its production, the more so the higher the marginal tax rate.