SUMMARYThis article seeks to help resolve the serious theoretical controversy about the contribution of international trade to economic development by constructing a simultaneous‐equations model that captures the most important quantitative aspects of the relationship between trade and development, and testing it by pooling data for 52 nations (representing all of the developing nations for which the required data are available) from 1961 to the present. The model is estimated by Full Information Maximum Likelihood, validated by dynamic simulation and utilized for policy and other counterfactual simulations. By the use of dummy variables and Chow tests, shifts of the structural relationships and slope coefficients over time are also tested for. The study finds strong empirical support for trade being very important for development, but to be more in the nature of a handmaiden than of an engine of growt