The Allocative Efficiency of Global Deforestation
Agriculture, plantation forestry, and mining compete with forests for land. This paper presents land parcels as an abatement technology whose abatement cost is given by foregone economic activity. Thus, I derive a forest abatement cost curve across the globe which captures rich substitution patterns across land across a host of characteristics: land productivities, transportation costs, and persistence in land use. This abatement cost curve is a key input in a model of the global pollution taxes and standards. I demonstrate that status quo forest loss is neither cost-effective nor efficient: lost carbon in removed forests is more expensive than expected profits off of the land use replacing it. Under status quo policies, a spatial general equilibrium model predicts half of the carbon stock in global forest is lost to deforestation in equilibrium. Reallocating land under (privately) cost-effective forest loss would have prevented 31% of carbon biomass lost from 2010 onwards: this result is independent of any social cost of carbon estimate. Socially efficient deforestation under a Pigouvian tax removes 45% less forest than the status quo. However, efficient reallocation implies a major shift of agriculture out of the tropics, inducing a structural change that dramatically alters the geography of manufacturing and services as well. Marginal abatement costs of forest are lowest in areas with large amounts of inefficient agriculture. I show that, despite this inefficiency, countries in the tropics may individually rationally not choose to tax their own carbon stocks, highlighting the need for international transfers to support vital forest cover.
Trade openness and deforestation in agricultural markets
I study the staggered implementation of US standards on Mexican avocado production, aimed to curb local pests, on local deforestation. I find that improved market access for Mexican orchards dominates increased costs of doing business in effect size. The access effect leads to sustained deforestation up to 10 years out from when individual municipalities began exporting avocados under standards. Further, by using an index-based measure of deforestation (the normalized difference vegetation index, or NDVI) rather than a classifier, I convert this loss of tree canopy greenness into lost carbon capture. Using a $51 social cost of carbon, the policy created an environmental externality of at least $31 million across the affected region. Traditional classifier measures fail to capture the gradual deforestation carried out by small orchard owners, understating the true effect relative to my NDVI-based method by $9 million.
Agricultural carbon taxes and industrial farming
Carbon taxes on agriculture are a theoretical first-best. However, rich evidence suggests agriculture enjoys benefits from scale, and that large-scale industrial farms are also the heaviest polluters. This research interrogates the tradeoff between farm scale and environmental damages. There are generically two margins of adaptation to a carbon tax within the agricultural industry: first, farms can exit if cost burdens are too high; second, farms can scale to take advantage of scale efficiencies. Both channels have implications when damages and scale are correlated. If industrial farming is eliminated by carbon taxation, this suggests that there will be greater land costs to agriculture as farms turn to lower-yield agro-ecological farming technologies. If, on the other hand, industrial farming enjoys a major cost advantage, carbon taxation may lead to greater exit among farms which are technically cleaner. I estimate damage functions for a host of polluting agricultural inputs for farms in the European Union across various levels of farm scale. I then use shocks derived from the EU ETS carbon price scheme to identify firm exit and scaling as a consequence of input cost shocks. These estimates are used to calibrate a model of an agricultural carbon tax’s impact on the structure of the food supply in the EU.