The Goldfeld demand for money equation revisited

The purpose of this paper is to revisit the demand for money equation, originally suggested by Goldfeld and since then widely used, from various viewpoints by using the U.S. quarterly data over the period 1955.1-1978.4. Results from different estimation methods and specification tests suggest strongly that the Goldfeld demand for money equation suffers from misspecification. Moreover, misspecification does not seem so much to do with omitted variables, but rather with some more general specification errors like inappropriate dynamics and endogeneity of the RHS variables in the demand for money equation.