Economics Seminar - The Unintended Consequences of Managed Trade: Evidence from U.S.-Mexico Sugar Suspension Agreement

Managed trade policies, which refer to government interventions that use quantitative trade restrictions to achieve specific, measurable outcomes, have garnered significant recent attention. While such policies have long been used in agricultural and food markets to support commodity prices and enhance competitiveness, empirical evidence on their environmental consequences is notably limited, especially within the context of agricultural commodities.

 

To address this gap, this paper empirically examines the causal impact of managed agricultural trade on environmental outcomes in the U.S. sugar industry, which is shaped by a longstanding tariff-rate quota system. Our analysis focuses on the 2014 Suspension Agreement with Mexico, a unique quasi-experiment that shifted the industry from a free trade regime to a managed trade one. We examine agricultural fires and associated air pollution as our primary environmental outcomes, chosen for their direct link to farming practices and significant health impacts on nearby rural communities.

 

We employ a spatial difference-in-differences (DID) approach, leveraging comprehensive, high-resolution remote sensing data on land use and fires. The key to our identification strategy is a transportation-driven threshold that delineates our treatment and control groups. Sugarcane fields located closer to mills are considered highly exposed to the trade shock, as they can more easily adapt to market changes. This approach allows us to map a national-level trade shock to specific locations while explicitly accounting for spatial heterogeneity in exposure.

 

Our analysis yields two main findings. First, we establish a causal link between managed trade policy and an environmental externality: fields most exposed to the policy experienced a significant short-term increase in agricultural fires relative to less exposed fields. Specifically, the average treated field saw an additional 0.043 fire events compared to a baseline mean of 0.294, an effect equivalent to a 15% increase. Second, we document significant heterogeneity across regions and seasons. We find that the policy shock significantly increased fire activity only in Louisiana, with no significant changes in other states. Furthermore, the timing of agricultural fires shifted from late to early in the harvest season following the 2014 agreement. Mechanism analysis suggests this increase was primarily driven by a land-use shift from other crops, wetlands, and grasslands to sugarcane production.