UC Berkeley

In this dissertation, I investigate the determinants of regional production choices,primarily focusing on agricultural commodities, and the aggregate consequences of those decisions. The dissertation is guided by the connecting theme of understanding why regions choose to produce specific commodities and the implications of these choices on international trade and development. The study comprises three interconnected chapters, each addressing a different aspect of agricultural and resource economics and international trade. The first chapter examines the clustering of agricultural production in Mexico and demonstrates that agricultural value chains (AVCs) explain these spatial patterns more closely than land suitability alone. The chapter highlights the presence of economies of scale in AVCs that act as barriers to entry and contribute to the clustering of agricultural production. I use Mexico as a setting, where highly productive, export-oriented agriculture coexists with subsistence agriculture. The paper develops a general equilibrium model incorporating intermediaries with market power and large fixed costs of entry. This model allows me to explore several counterfactuals, such as the impact of climate change on agricultural production and the role of fixed costs in generating patterns of clustering. The second chapter, jointly authored with Thibault Fally, investigates the role of primary commodities in international trade, highlighting the potential understatement of aggregate gains from trade when key features of commodities are ignored. By developing a general-equilibrium model that accounts for these features, including low responsiveness of both demand and supply to changes in price, reliance on natural resources, and a high concentration of these natural resources among only a few countries, this chapter offers a more comprehensive understanding of the economic gains from trade. The model and data are used to simulate a number of counterfactuals which confirm that ignoring the specific features of commodities leads to a wide understatement of the gains from trade. The third chapter, jointly authored with Joel Ferguson and Kangogo Sogomo, presents a computationally feasible approach to remote sensing outcomes at large scale. We develop a novel dimensionality reduction technique to predict agricultural outcomes using only widely available, aggregated training data. This methodology is adaptable to various problems, including classification of crop types, measurement of forest loss, or urban growth, and offers a practical toolkit for researchers and policymakers who require timely and cost-effective data for decision-making purposes without the need for additional sources of information. We focus on measuring crop yields, particularly maize yields in Mexico, which is important for ensuring food security and informing policies such as food aid and transfer programs to assist smallholders affected by climatic shocks or better target responses to natural disasters in rural areas. Taken together, these chapters provide insights into the determinants of regional production choices, the development of innovative techniques for crop yield estimation through remote sensing, and the importance of primary commodities in international trade. By exploring the forces shaping regional agricultural and resource production and the distribution of scarce natural resources, this dissertation contributes to a more integrated understanding of how trade and agricultural and resource sector and play a role in shaping regional economic outcomes. The findings have implications for agricultural policy, international trade, and for regional development.