Inferences about the effects of money on economic activity may depend importantly on the choice of a monetary index because simple-sum aggregates cannot internalize pure substitution effects. This hypothesis is investigated by replicating five recent studies that have challenged an aspect of the "conventional wisdom" about the effects of money on aggregate activity. In four of the five cases, the qualitative inference in the original study is reversed when a simple-sum monetary aggregate is replaced by a Divisia index of the same asset collection. The results are mixed in the fifth case.

Inferences about the effects of money on economic activity may depend importantly on the choice of a monetary index because simple-sum aggregates cannot internalize pure substitution effects. This hypothesis is investigated by replicating five recent studies that have challenged an aspect of the "conventional wisdom" about the effects of money on aggregate activity. In four of the five cases, the qualitative inference in the original study is reversed when a simple-sum monetary aggregate is replaced by a Divisia index of the same asset collection. The results are mixed in the fifth case.

Interest-bearing checkable deposits are examined to test whether they should be included in measures of the U.S. money stock. Both Divisia and traditional simple-sum aggregates are constructed on the basis of tests for weak separability in a model of the demand for financial assets. Using nonparametric demand analysis, we find that several groups of assets are compatible with aggregation theory. We find empirical support for a narrow measure consisting of the components of current MIA. In tests based on a St. Louis equation and in terms of controllability, a Divisia aggregate performs better than the simple-sum MIA measure.

Interest-bearing checkable deposits are examined to test whether they should be included in measures of the U.S. money stock. Both Divisia and traditional simple-sum aggregates are constructed on the basis of tests for weak separability in a model of the demand for financial assets. Using nonparametric demand analysis, we find that several groups of assets are compatible with aggregation theory. We find empirical support for a narrow measure consisting of the components of current MIA. In tests based on a St. Louis equation and in terms of controllability, a Divisia aggregate performs better than the simple-sum MIA measure.