Model Selection and Forecasting Ability of Theory‐Constrained Food Demand Systems

Abstract

Out-of-sample forecasting of annual U.S. per capita food consumption, applying data from 1923 to 1992, is used as a basis for model selection among the absolute price Rotterdam model, a first-differenced linear approximate almost ideal demand system (FDLA/ALIDS) model, and a first-differenced double-log demand system. Conditional-on-price consumption forecasts derived from elasticities are determined to be superior to direct statistical model forecasts. Models with consumer theory imposed through parametric restrictions provide better forecasts than models with little theory-imposition. For these data, a double-log demand system is a superior forecaster to the Rotterdam model, which is superior to the FDLA/ALIDS model. Copyright 1996, Oxford University Press.

Do U.S. Cotton Subsidies Affect Competing Exporters? An Analysis of Import Demand in China

“…A number of studies have compared model and elasticity-based forecasts using demand systems. These include Gustavsen and Rickertsen (2003), Kastens and Brester (1996), and Muhammad (2007). These studies concluded that demand forecasts derived from elasticities are superior to model-based forecasts.…”

“…For instance, Pan, Hudson, and Ethridge (2010) find that the elimination of U.S. cotton programs (direct payments, target price, and loan rate) would raise the world price index for cotton by slightly less than 5% per year over a 5-year period. Following Kastens and Brester (1996), import demand projections are derived using an elasticity-based forecasting equation. The unconditional elasticities are used because they encompass the complete effect of a price change making them more suitable for policy projections.…”

Procurement Risk Management in Beef Supply Chains

“…However, health concerns and a number of other factors contributed 3 The industry standard for demand modeling is the Almost Ideal Demand System of Deaton and Muellbauer (1980). It is used in almost all the work referenced and has been found to produce elasticities with desirable forecasting properties (Kastens and Brester, 1996).…”

Model Selection of Meat Demand System Using the Rotterdam Model and the Almost Ideal Demand System (AIDS)

“…The two most widely adopted especially in food demand studies are the Rotterdam model introduced by Theil (1965) and Barten (1969) and Deaton and Muelbauer's (1980) almost ideal demand system (AIDS). Both models are derived from consumer theory, and are used to impose or test behavioral restrictions that are deduced from that theory (Kastens & Brester, 1996). However, neither economic theory nor statistical analysis provides clear a priori criteria for choosing between these two models (Lee, Brown, & Seale, 1994).…”