The Armington procedure (AP) has become increasingly popular in agricultural trade analyses. However, some arguments have arisen concerning the relevance of using the procedure for such analyses. This study examines the assumptions commonly made when using the Armington procedure and suggests modifications for agricultural trade analyses. Results from models utilizing rice-trade data suggest that the assumptions of the single constant elasticity, in particular, may not be appropriate for analyzing agricultural trade. These results also suggest that, with proper modifications, the AP can be applied to agricultural trade. Further, results of a modified Armington procedure indicate that trade in rice exports is highly competitive and that changes in market shares of individual exporters are not independent of changes in budget expenditure allocated to imports.