Allowing for Group Effects When Estimating Import Demand for Source and Product Differentiated Goods

In this study an import demand model (differential production model) is presented that is used in estimating the demand for source and product differentiated goods simultaneously. Unlike the traditional import demand models, this model can account for changes in relative group expenditures. Expenditure estimates differed when comparing the differential production model and Rotterdam model results. Results showed that if group revenue shares are relatively fixed, then the bias in expenditure estimates due to omitting group effects will be small when using traditional demand models such as the AIDS or Rotterdam models. As relative group shares significantly change and diverge the bias increases, particularly for imports representing a larger share of group expenditures.