The Exact Excess Demand Specification

Abstract

The difficulty in estimating the standard disequilibrium model arises from the inability to observe demand and supply and hence the inability to separate the sample into periods of excess demand and excess supply. A solution is to add to the model information about excess demand that would aid in partitioning the sample. Goldfeld and Quandt (1975) provide a Monte Carlo study affirming the informational value of a priori sample separation observations. Bayesian estimation is an obvious approach to this problem, and for first attempts see Richard (1980) and Lubrano (1984, 1985). However, an extremely tractable model can be constructed by assuming exact sample split information. A basic model with exact information is presented next, and further modifications including dynamic links, an endogenous price, and a test for equilibrium are considered in later sections. Models with several markets in disequilibrium, and spillovers between these markets, can also be easily estimated as shown in Section 3.6.