Generalized Habit Formation in an Inverse Almost Ideal Demand System: An Application to Meat Expenditures in the U.S.

The Inverse Almost Ideal Demand System (IAIDS) model of Moschini and Vissa (1992) and Eales and Unnevehr (1994) is extended to include: (1) general, nonlinear, nonadditive habit effects; and (2) a specification for habit stock terms that allow purchases from the distant past to influence current consumption (long memory). The resulting models are compared with a linear habit effects model and a static specification. The empirical estimation is on U.S. quarterly meat expenditures (1961-1993), with each model being subjected to a battery of misspecification tests. Results of these tests, along with tests of homogeneity and symmetry restrictions, indicate clearly that the most generalized dynamic specification--the one with nonlinear, nonadditive long-memory habit stock effects--is preferred. Furthermore, persistence effects are found to be qualitatively important in that flexibility, consumption scale, and habit flexibility estimates differ, in some instances substantially, between alternative specifications.