Essays on Agricultural Greenhouse Gas Emission Footprints: Accounting, Bias, and Trade Policy

Abstract

Global agriculture and food systems account for roughly one-third of anthropogenic greenhouse gas emissions, yet substantial uncertainty persists in measuring emissions intensities of agricultural production and the carbon footprint of international trade at the commodity-country level. These uncertainties arise from fragmented emissions accounting frameworks. At the same time, a significant portion of agricultural GHG emissions arises from production within a country to meet external demand, and preferential trade agreements have proliferated over recent decades, reshaping global agricultural trade patterns.

The first chapter constructs a harmonized farm-gate emissions intensity database for major agricultural commodities using eight data sources. I document substantial cross-system variation in emissions intensity estimates at various levels and trace these differences to differing activity data, emission factors, allocation rules, and global warming potential assumptions embedded in emissions accounting. Mapping production-side heterogeneity in emissions intensity into bilateral trade patterns generates substantial differences in emissions embodied in agricultural trade. Decomposition analysis reveals that growth in agricultural trade volume is the primary driver of increasing embodied emissions, partially offset by efficiency gains at the exporter level. However, a persistent positive composition effect indicates a structural shift in global sourcing toward carbon-intensive exporters, particularly for beef and soybeans. These findings from variation and decomposition analyses highlight a critical challenge for climate and trade policy. Without harmonized accounting standards, emission benchmarking and carbon-based trade measures remain constrained by data uncertainty. Moreover, effective mitigation will require not only cleaner production technologies but also governance mechanisms that steer global sourcing toward cleaner exporters.

The second chapter examines how agricultural trade policy affects agricultural trade patterns and the emissions embodied in agricultural trade. First, I investigate the environmental bias embedded in existing tariff structures--defined as the extent to which tariff schedules systematically favor cleaner or more emissions-intensive exporters--and find that it varies sharply across commodities, ranging from an implicit carbon tax of 151 dollars per ton of CO2-equivalent to a carbon subsidy of 82 dollars per ton of CO2-eq. Counterfactual simulations using a general equilibrium Poisson Pseudo Maximum Likelihood framework reveal that agricultural tariffs, shaped by preferential margins, have historically tilted sourcing toward cleaner exporters among developed importers. Under full trade liberalization, shifts toward cleaner sourcing reduce embodied emissions, but this effect is offset by trade expansion. Introducing emissions intensity-linked tariff schedules demonstrates that the effectiveness of carbon border adjustment policy depends critically on institutional design: when layered onto existing tariff structures, environmental signals are diluted, but when applied independently and uniformly, such tariffs substantially reallocate trade toward cleaner exporters, reducing the carbon intensity of imports by up to 13%. The thesis highlights that effective climate mitigation in agricultural trade requires both improved emissions measurement and trade policy design that actively steers sourcing toward cleaner production systems. Without progress on harmonized accounting and coordinated policy frameworks, the climate effectiveness of trade-based instruments will remain constrained by data uncertainty and structural trade incentives.