|
The Almost Ideal Demand System (AIDS) and Rotterdam models were used to test for the existence of structural change in Australian per capita consumption data on meats (beef, lamb, pigmeat and chicken) for the period 1965:1 to 1992:4. The models were estimated in single equation as well as in nonlinear system of equation versions. Both models showed the presence of structural change with the AIDS model producing more consistent results. The changes occurred in the early 1970s for beef, lamb and pigmeat and in the early 1980s for chicken. However, the effects of the changes were very small with only a 2.44 percent drop in expenditure share of beef, a 2.96 percent drop in lamb, a 0.71 percent increase in pigmeat and a 2.45 percent increase in chicken over the period of time assuming prices and expenditure held constant. These facts support the results of previous studies where the main determinant in quantity of meat demanded are changes in prices and total expenditure. Because of high collinearity of the data, the LA/AIDS model generated similar estimated elasticities to the 'true' AIDS model and due to the similarity of their right hand sides the first-differenced LA/AIDS model estimated similar elasticities to the Rotterdam model. This study also reveals that based on a test developed by Alston and Chalfant (1993) Australian per capita consumption data on meat rejects the Rotterdam model in favour of the AIDS model. |
|