The U.S. Tax Reform Act of 1986 (TRA) included many provisions changing the taxes due on capital income received by both individuals and corporations. Measuring the distributional effects of these tax changes is not as straightforward as it might appear. What is easily reported is the change in tax payments on capital income for each individual resulting from the tax reform, holding constant this individual's portfolio and the rate of return earned on the portfolio. These calculations were done originally when projecting the distributional consequences of the legislation; the equivalent exercise could now be done using current portfolio choices.

What we would like to measure, however, is the change in welfare of individuals in different parts of the income distribution resulting from the 1986 tax reform, taking into account the option individuals have to change their behavior in response to the reforms as well as the impact of the resulting price changes. When behavior does change in response to a tax reform, the gain to each class of taxpayers will be underestimated when measured by changes in tax liabilities holding fixed their behavior at their ex ante values, because such a measure ignores the benefit these taxpayers derive from altering their behavior. In contrast, any measure of the gain based on ex post data will normally be overestimated, because this measure ignores the nontax costs that ultimately limit the extent to which people are willing to change their behavior to save on taxes.