Wildlife and Conflict: The Cost of Biodiversity Protection Policies
Scientific evidence shows that we are going through the first human-induced mass extinction of species, and as such, conservation-motivated trade restrictions are widely implemented among policymakers. However, there is little research on the harmful externalities of such policies. This paper combines global georeferenced data on wildlife habitats (including both animal and tree genera) with information on armed conflict, documenting the violent externalities generated by global bans on international trade in wildlife products aimed at conserving biodiversity. An event-study specification shows that the imposition of trade restrictions increases the likelihood of conflict in habitat areas. Two pieces of evidence suggest that the policy generates local windfalls that drive the conflict effect. First, two-stage least squares estimates for elephant ivory show that trade restrictions increase prices, which in turn increase the likelihood of conflict. Accounting for the spatial distribution of elephants, the implied effect size exceeds that of well-studied industrial conflict minerals. Second, for wild trees, satellite data suggests that, once restrictions are in place, production shifts from states with high institutional capacity to those with low capacity, generating local windfall rents that fuel additional conflicts. An analysis of the spatial distribution of battles before and after the policy shows that armed groups positioned to capture these rents expand into new areas and become more likely to gain territorial control. This pattern supports the “feasibility” mechanism, whereby windfall gains relax budget constraints and enable attacks. A quantitative social planner model shows that a targeted policy restricting trade in states with high institutional capacity and relatively small stocks of wildlife could preserve the resources while mitigating the conflict externality.
A Demand System Approach to Endangered Species Protection
This paper evaluates the effectiveness of the international treaty CITES, which seeks to protect endangered tree species by restricting their trade. I first construct a new species-level measure of harvesting activity using satellite data and validate its accuracy. Second, I estimate a staggered-adoption synthetic-control event-study specification to causally identify the treaty’s impact, revealing substantial heterogeneity in effectiveness across species and regions. Third, to inform policy design, I estimate own- and cross-price elasticities for precious tree species using an Almost Ideal Demand System (AIDS). Some protected species exhibit positive own-price elasticities—consistent with Veblen goods—complicating conservation efforts that rely on supply restrictions, as they may unintentionally increase demand. Finally, using the estimated elasticities, I quantify alternative counterfactual policies and propose targeted strategies that account for positive price responsiveness. In particular, counterfactual simulations suggest that restricting trade only in species that have negative own-price elasticities—obey the Law of Demand—or imposing simultaneous rather than sequential bans on endangered species and their close substitutes, leads to better environmental outcomes.
Wildlife Poaching and the Labour Market (with Etienne Le Rossignol)
Using data on elephant poaching and the harvesting of endangered precious tree species across Africa and Asia, we document substitution between employment in local agriculture---the labour-intensive sector---and wildlife poaching. Positive economic shocks to agriculture reduce poaching of both elephants and endangered trees. A national-level crop minimum price or crop disaster insurance policies reduce elephant poaching. A sector-biased, trade-liberalising reforms that lower agricultural employment increase elephant poaching. This is consistent with predictions from a standard 2 by 2 trade model, which we use as a conceptual framework. Finally, using household survey data on bushmeat consumption, we show that consumption patterns are highly correlated with poaching intensity and can serve as a proxy in regions where formal monitoring is limited. This approach expands the spatial coverage of poaching data and offers policymakers a practical tool for improving the monitoring and management of poaching crises.
Carbon Colonialism (with Etienne Le Rossignol and Orlando Roman)
This research investigates the socioeconomic and environmental consequences of newly established market-based forest preservation policy, implemented by private enterprises in partnership with local governments and firms in sub-Saharan Africa. In return, these initiatives invest in local education, health, and infrastructure. We hand-collect data on—nearly—the universe of privately-run carbon concessions across Africa and, using a staggered event-study synthetic control approach, estimate their impact on both forest preservation and economic activity using nighttime luminosity data. To explore the underlying economic mechanisms, we employ a spatial regression discontinuity design (RDD) with DHS data, comparing villages located just inside and just outside the concession boundaries.
Conservation Through Representation: Indigenous Politicians and Forest Protection in the Brazilian Amazon (with Gabriel S. Gaspar)
Indigenous populations in the Americas, while comprising a minority, exert significant control over extensive demarcated territories. In Brazil, indigenous communities govern approximately 15% of the national territory and nearly 25% of the Amazon rainforest. However, the influence of indigenous political representation on environmental outcomes remains poorly understood. This paper examines the effects of electing an indigenous politician to the city council, leveraging close elections as a natural experiment. Regression discontinuity estimates suggest a substantial reduction in deforestation rates in municipalities that narrowly elected an indigenous councillor compared to otherwise similar areas. These effects are particularly pronounced in locations with larger indigenous populations and greater shares of indigenous territory.