Infrastructure and Aggregate Agricultural Productivity: International Evidence

A relationship has long been recognized between a country's agricultural development and its investment in infrastructure capital, such as transportation and communication facilities, that is not directly related to the agricultural production process. The relevant literature spans Adam Smith's proposition "That the Division of Labour is Limited by the Extent of the Market"; the work of von Thtinen; numerous studies by historians and economic historians; and the work of economists such as Schultz, Kuznets, Wharton, Owens, and Johnston and Kilby, to mention a few.' Despite the apparently widespread recognition and acceptance that a relationship exists between agricultural development and infrastructural development, few, if any, empirical studies test this as a hypothesis and provide estimates of the magnitude of such a relationship.2 The aim of this paper is to utilize 1965 aggregate agricultural production data and the concept of the aggregate production function to study the effects of transportation and communication infrastructure on aggregate agricultural productivity. The aggregate data come from 47 less developed countries (LDCs) and 19 developed countries (DCs); the sample countries are listed in the Appendix. The use of intercountry comparisons for development studies dates to Clark and the pioneering work of Kuznets, and it has been exploited econometrically by Chenery.3 An important contribution to the study of international agricultural productivity was made by Hayami and Ruttan, who showed that an aggregate agricultural production function that accounts for intercountry differences in resource endowments, technical (modern) inputs, and human capital can explain a large proportion of the variation in aggregate agricultural output across more and less developed agricultures.4 Evenson and Kislev modified Hayami and Ruttan's model by including a variable measuring each country's investment in agricultural research.5 Essentially the same