Deforestation

A growing body of research employs quantitative spatial economic models to examine how trade influences deforestation. For broader reviews on the environmental implications of trade, see Copeland et al. (2022) and Cherniwchan and Taylor (2022). At the core of this literature is the idea that the spatial distribution of consumers and producers – and the trade linkages connecting them – matters for understanding the aggregate effects of policy. Trade policies, such as improved transportation or tariff changes, generate both direct effects in targeted regions and indirect effects that propagate across space through trade and migration. Capturing the full impact of such policies on forest loss, therefore, requires methods that can quantify both these direct and indirect effects.

Spatial Linkages

We start with a series of recent papers that examine the impact of various deforestation policies on Brazil’s forests. Asher et al. (2020) studied the deforestation impact of transportation infrastructure, finding that upgrading national highways has sizable effects on deforestation. Araujo et al. (2025a) adapt the market access approach designed by Donaldson and Hornbeck (2016) to assess the general equilibrium effects of transportation infrastructure with detailed data on transportation networks. They find that failing to account for the indirect effects of locally targeted infrastructure projects would underestimate their total deforestation impact by about a quarter.

Using the trade and land-use framework of Sotelo (2020) and Farrokhi et al. (2025), Gollin and Wolfersberger (2024) show that road expansion since the 1990s accounts for roughly one-tenth of total deforestation in Brazil. Leite-Mariante and Restrepo (2024) develop a dynamic spatial economy model and find that anti-deforestation policies trigger spatial leakages that unfold gradually over several years. Using longer time series data, Akerman (2025) examines how Brazil’s demographic transition influenced deforestation across regions and finds that the slowdown in population growth significantly contributed to curbing deforestation.

Beyond Brazil, Balboni et al. (2024) study forest fires in Indonesia and highlight the role of spatial externalities. Forest fires set for land-clearing purposes often spread to neighbouring plots of land, creating potential negative externalities. Madhok (2025) documents that forest encroachment by transport and other infrastructure accounts for 20% of species loss in India.

Supply Chains

Market incentives and policies can affect deforestation through supply chain linkages (for a review on deforestation supply-chain initiatives, see Lambin et al. 2018). Dominguez-Iino (2025) examines how tariffs influence deforestation using a spatial economy model in which farmers sell their output to multinational firms in imperfectly competitive markets in South America. The key insight is that these firms exert monopsony power, setting a wedge between producer and consumer prices. As a result, the impact of tariffs depends on the local degree of competition among multinational firms. Since international firms are often less present in remote areas, where carbon density is greatest, tariff increases have lower pass-through to local farmers. Consequently, the intended policy effect is weakened precisely where the potential carbon emissions from deforestation are highest.

Barrozo (2025) shows that carbon emissions in the Amazon beef supply chains are highest among small, informal domestic players and intermediaries – not exporters. Because exporters face external pressure to meet standards, they tend to be more efficient and cleaner. As a result, policies focused only on exporters risk ignoring major domestic contributors to emissions.

Tiew (2025) studies the EU’s 2017 palm oil import ban using Malaysian data. The author finds the ban to be mistargeted in that it penalizes downstream firms rather than directly addressing deforestation. Furthermore, the ban disproportionately harms smallholders. Coordinated supply restrictions among producers could raise total farmer payoffs while also reducing deforestation.

Global Perspectives

Another paper examining deforestation across a broader set of countries is Hsiao (2025), which focuses on Indonesia and Malaysia. These countries account for 84% of global palm oil production, which is a major driver of forest loss and carbon emissions. The study considers multiple destination countries for palm oil exports and highlights a key coordination problem: if only a subset of countries imposes tariffs on palm products from Indonesia and Malaysia, then global prices decline and prompt unregulated markets to increase their imports. Unilateral tariffs are therefore less effective than coordinated tariffs in mitigating deforestation. The study also quantifies the benefits of committing to long-run regulation, noting that deforestation responds only weakly to short-run regulation. Harstad (2012) offers a contrasting view for fossil fuels, arguing in a theoretical setting that a conservation coalition can avoid issues of leakage and commitment by directly purchasing foreign deposits and choosing to conserve them. In practice, this approach relies on enforcement by local authorities.

Complementing the papers above, which focus on particular regions or countries, Farrokhi et al. (2025) examine deforestation from a global perspective. Their analysis develops new analytical results showing how reductions in trade costs shape global deforestation. Two key insights emerge. First, global trade cost reductions can lead to an increase in total forest area if demand for agricultural goods is sufficiently inelastic. In this case, lower trade costs enable countries to source their food more efficiently across borders, producing the same amount of agricultural output with less agricultural land. Second, the correlation between comparative advantage and absolute advantage plays an important role in determining global forest area. While comparative advantage determines which countries specialise in agriculture, absolute advantage governs how much land is required in production. If countries with a comparative advantage in agriculture also have a higher absolute advantage, free trade can lead to lower global land use for agriculture. The authors develop and calibrate a dynamic spatial economy model that captures these insights and find that global reductions in trade costs can increase global forest area. However, they may simultaneously drive deforestation in specific regions with comparative advantage in agriculture, including Brazil and Canada.

Another recent paper that adopts a global perspective on deforestation is Mishra (2025). The paper applies the discrete choice framework used by Araujo et al. (2025b) and Souza Rodrigues (2019) to grid-level data on global deforestation, comparing private profits from deforestation since the 1980s with estimates of the social cost of carbon. The author then argues that a global Pigouvian tax on deforestation would be more efficient under free trade.

Regional Trade Agreements

Trade agreements often increase pressure on forests. A global panel study reveals that regional trade agreements (RTAs) have led to significant post-enactment deforestation in developing countries, primarily driven by agricultural expansion (Abman and Lundberg 2020). However, new evidence shows that environmental provisions in RTAs can offset these effects. When RTAs include clauses on forest protection or biodiversity, they mitigate the increase in deforestation observed in agreements without such provisions (Abman et al. 2024).

Future Work

Looking ahead, we see three directions for future research within this literature. First, incorporating more institutional dimensions of deforestation – such as enforcement capacity, land tenure, and governance – and studying how they interact with the spatial organisation of the economy. Second, improving the design of optimal trade agreements using quantitative models. While there has been theoretical progress in this area, including Harstad (2024), there remains a need for more quantitative work. Third, most existing studies assume representative agricultural producers. Future work could benefit from explicitly modelling farm-size heterogeneity and scale, which are important in shaping agricultural productivity.

Policy Takeaways

We emphasise several implications for trade and migration policy.

• Policy interactions. Trade and migration linkages have important interactions with trade policies, shaping their ultimate impact on forest outcomes.
• Unintended consequences. Infrastructure investments can have unintended deforestation consequences that are significantly larger when indirect spatial effects are considered.
• Market structure. Supply chain policies must account for the structure of markets. Targeting exporters may miss informal domestic actors responsible for the highest emissions.
• International coordination. International coordination is crucial for effective trade-based environmental policies, as unilateral actions can create leakage effects that undermine policy goals.

For full reference list see the end of the conclusion chapter.