Structural changes in consumer demand and implications for the world food market

Wusheng Yu, Purdue University

Abstract

The world food market has experienced significant structural changes in the past two decades, featured by a shift away from trade in bulk commodities. Much of the shift has been fueled by income growth and the subsequent upgrading of diets by consumers. Projecting future changes requires a directly estimated, globally regular demand system with flexible Engel effects. The objective of this work is to identify and estimate a system with these desirable properties, and to fully integrate the estimated system into a global applied general equilibrium model (AGE) to conduct simulation exercises aimed at predicting future food demand and trade patterns. An implicit directly additive demand system, AIDADS, is selected because of its superior flexibility. An AGE model and database, GTAP (Global Trade Analysis Project), is chosen as the basis for conducting the projection exercises. The AIDADS is estimated, calibrated to the GTAP database, and integrated into the GTAP model. Using the modified GTAP model, a demand-side simulation is conducted, in which regional income and population are shocked to their projected levels in 2020. This simulation shows that income growth would cause very disparate demand patterns, with the highest demand growth in livestock and other food and the lowest in fish and grain. However, these demand projections imply much more balanced production and import patterns across products. A decomposition analysis shows the importance of intermediate demand in balancing these patterns. At the global level, developing regions contribute more to growth in output requirements in most food products, while developed regions still cause more growth in imports of livestock, horticulture and fish. The same demand-side experiment is repeated under three alternative demand systems (Linear Expenditure System, Cobb-Douglas and Constant Difference Elasticity) in order to assess the differential impact on projections to 2020. The CD model is used as only a reference for its unitary income elasticites. These simulation results are compared to the AIDADS results to show where AIDADS does and does not make a difference. For high-income regions with moderate income growth, LES and CDE produce similar results to AIDADS. For developing regions with rapid income growth, LES and CDE under-predict non-food demand and over-predict food demand, compared to the AIDADS results.

Degree

Ph.D.

Advisors

Hertel, Purdue University.

Subject Area

Agricultural economics|Agriculture|Horticulture

COinS