The paper attempts to revive the idea that the level and growth of output of countries in a balance of payments constrained framework are determined by the workings of the foreign trade multiplier; an idea first put forward by Harrod in 1933. Evidence is presented for a range of developed and developing countries, which suggests that the Harrod trade multiplier works in developed countries, but much more loosely in developing countries where capital flows and terms of trade changes are more important in determining growth performance. The relevance of the model for explaining regional growth, and the growth of industry in the early stages of development, is also discussed.
