It's Not About the Money: Why Natural Experiments Don't Work on the Rich

Abstract

An influential literature on the effects of marginal tax rates on the behavior of the rich has claimed that the elasticity of taxable income with respect to the net of tax share is very high possibly exceeding one. These high estimated elasticities imply that cutting taxes on the rich does not lose much revenue possibly increases it and that progressivity generates a large amount of deadweight loss. To identify this elasticity, these studies have conducted natural experiments' comparing the rich to other income groups and assuming that they are the same except for changes in their tax rates. This paper tests the natural experiment assumption using alternative data on the compensation of a panel of several thousand corporate executives and finds it to be false. Relatively, the very rich have incomes which trend upward at a faster rate are more sensitive to economic conditions, and are more likely to be in a form whose timing can be shifted in the short run. Interpreted broadly, these facts might reduce existing elasticity estimates by as much as 75%. The paper also suggests ways of improving existing methods.