- 27 Jun 2014
- Working Paper
Positive and Normative Judgments Implicit in US Tax Policy and the Costs of Unequal Growth and Recessions
Executive Summary — What does United States tax policy reveal about Americans' values and beliefs, and about how those values and beliefs have changed over time? In this paper, the researchers use theory and data to back out the implicit priorities and judgments in US tax policy over the last several decades. They find a dramatic shift in the mid-1980s that persisted, and even continued, over the next 25 years and that cannot be reconciled with conventional assumptions about these values and beliefs. They explore evidence on a number of possible explanations for this shift, including a link between economic and political inequality. They also attempt to use their results to estimate the welfare costs of two key phenomena—rising inequality and recessions—and find that these estimates are highly sensitive to the explanation one adopts for the evolution of US policy. Overall, the researchers argue, uncovering the judgments implicit in policy provides a promising path toward both a better understanding of policy priorities and more objective comparisons for policy evaluation. Key concepts include:
- This paper uncovers the values and beliefs implicit in U.S. tax policy over the last three decades.
- Conventional assumptions about these values and beliefs are not consistent with the path of policy.
We use official data and standard optimal tax conditions to infer the positive and normative judgments implicit in U.S. tax policy since 1979. We find that explanations within this framework for the time path of U.S. policy require central parameters of the model, namely the elasticity of taxable income or the marginal social welfare weights on top earners, to take unconventional values. We use inferred social preferences to provide novel estimates of the welfare costs of unequal growth and recessions and find that they are sensitive to the assumed distortionary costs of taxation and the year from which preferences are derived. We explore several possible explanations for our findings with available data.