When Rules Change: The Economics of Retroactivity
New York University School of Law
New government decisions, in the tax law and elsewhere, often have retroactive effects, in that they alter the consequences of people's prior decisions. Retroactivity is typically considered acceptable in some circumstances (such as a change in interest rates by the Federal Reserve Board) but not others (such as ex post facto criminal legislation). Yet the voluminous literature on transition problems has made little headway in establishing a consistent framework for evaluating them.
This working paper, the first chapter of a book forthcoming from the University of Chicago Press, develops such a framework from three main insights. The first is the general applicability of rational expectations, suggesting a benchmark prediction that people will on average anticipate transition policy accurately given the repeated-play nature of the problem. (Thus, the Auerbach-Kotlikoff (1987) view of ex post capital levies as lump-sum is unpersuasive). Second, retroactivity is best decomposed into the problems of "transition risk" (raising familiar insurance issues) and "retroactive taxation" (raising the same issues as prospective taxation). Third, it is important to distinguish between rules' policy content (e.g., how tax burdens are distributed between taxpayers and allocated between economic activities) and their mere accounting content (e.g., exactly when a tax of fixed present value is collected). Major transition effects, such as from replacing the present income tax with a cash flow consumption tax, often result from accounting changes that are unrelated to the policy content of the change in rule, and this distinction is important in both a case-by-case and a "constitutional" analysis.
JEL Classification: H1, H2
Date posted: November 16, 1998