The Commodity Structure of World Trade: Comment

Abstract

While making a statistical inquiry into commodity structure of world trade in an article which appeared in this Journal in August 1943, Albert Hirschman devised a method for subdivision of foreign trade of individual countries into various categories of interchange.' Having applied this procedure to world trade as a whole, he then showed that largest proportion of world trade consisted, each year between 1925 and 1937, of an exchange of primary products (foodstuffs and raw materials) against primary products (foodstuffs and raw materials).2 This led him to conclude that the traditional view which believes world trade to be based primarily upon exchange of manufactures against raw materials and foodstuffs is not even approximately correct.' I find this result rather surprising and am of opinion that, attempting to match corresponding export and import items for each country (conveniently subdivided into two broad categories of primary products and manufactured goods), Hirschman has not chosen best hypothesis or best procedure. Before I elaborate on two other alternatives, I believe that a brief outline of Hirschman's method is here in order. According to it, exports of manufactures are matched against imports of manufactures and exports of primary products are matched against imports of primary products on assumption that exports provide in each instance buying power for corresponding imports. The amounts left unmatched are then postulated to represent amounts of trade accounted for by exchange of manufactures against primary products. Thus, for example, conclusion reached by Hirschman on basis of his Scheme III4