Abstract
When the US Bureau of Economic Analysis adopted chain-type quantity indexes in 1996, real gross domestic product (GDP) and its components, when measured in units of the currency of a particular reference year, were no longer additive. We discuss some of the problems with the additive, fixed-weighted aggregates these estimates replaced, as well as problems that arise when users treat the new "chained dollar" aggregates as if they were additive. Using properties of the Fisher index formula, we describe new tools for economic analysis that traditionally relied on the additivity of fixed- weighted GDP aggregates. We also describe some tools that have been developed for the compilation and review of chain-type estimates. Finally, we discuss Claude Hillinger's proposed methods for creating an additive set of real aggregates and present empirical results that lead us to be skeptical about their usefulness.