UC Santa Barbara

Agriculture produces nearly all human food and supports the livelihoods of a large share of the world's population. It is also among the leading causes of habitat loss, greenhouse gas emissions, and air and water pollution. What agriculture costs the environment, and what constraining it costs producers and consumers, are therefore central policy questions. This dissertation takes up both in three chapters. Chapter 1 asks what a package of Brazilian anti-deforestation policies --- the Amazon Soy Moratorium, cattle agreements, and 2012 Forest Code --- costs the producers it restricts. A nested logit model of land use, estimated on a pixel-level panel of the Amazon and Cerrado, yields the compensating variation: those producers' welfare loss, or the money it would take to make them whole. That loss is 25 percent of their welfare, some 45 billion dollars in present value. The package retains 2.7 million hectares of Amazon forest, whose carbon justifies it at any social cost of carbon above 43 dollars per tonne. Chapter 2 asks whether yields are rising fast enough to meet projected 2050 crop demand without further cropland expansion. Fitting ten piecewise-linear trend models to 34 crop-country yield series for maize, rice, wheat, and soybean over 1961--2024 shows that yield plateaus are rare --- under 10 percent of global production for every crop. Where simulated potential yields are available, most systems, plateaued or not, remain well below their biophysical ceiling. Projected growth nonetheless falls far short of demand for maize and soybean, and closing that gap with land instead would take roughly 55 million hectares. Chapter 3 asks what drives the US agricultural nitrogen surplus, and how much of the apparent efficiency gain is real. An exact index decomposition of a new nitrogen budget spanning 18 crops and 2,829 counties separates five drivers. A near-flat total since 1990 conceals two large offsetting forces: rising yields added 1.99 Tg N of surplus, while falling surplus per unit of output removed 1.40 Tg N. Nearly a third of the apparent gain in conventional nitrogen-use efficiency reflects shifts in crop mix and geography rather than improved field practice.