Agri-food trade in GTAP-HET: Returns to scale in agriculture, and the Melitz model

Agricultural protection is almost always a highly sensitive issue in bi- and multi-lateral trade negotiations. The reasons for this are usually political, but it always means there is a high demand among policy makers for analytical tools which can assess the often complex impacts of liberalisation on farm viability, land use change, and consumer food prices. Since their inception, Computable General Equilibrium (CGE) models have made a rich contribution to this analysis. However, when standard assumptions of constant returns to scale and perfect competition are used, this can limit the ability of such models to answer the questions being asked by policy makers. Put simply, these questions often amount to “what are the opportunities, and what are the threats?” If returns to scale are not constant, or if other sources of productivity variation are present in a domestic farming system, the same liberalising policy may represent an opportunity to one farmer, and a threat to another, even if the two are producing the same commodity in the same country. Combining the macroeconomic width and rigour of a CGE model with the heterogeneity of domestic farm systems represents an exciting frontier in agri-food trade policy analysis. This paper will present the current evidence on returns to scale in agricultural sectors in selected countries – drawing on fixed and variable cost share data from the USDA, and European national Farm Business Surveys collated in the Farm Accountancy Data Network (FADN). Where the evidence for increasing returns to scale is clear, it follows that there is a clear case for the importance of making use of the Melitz model, or some alternative to the standard constant returns to scale assumption.