Abstract: Deforestation of the Amazon rainforest is a threat to global climate, biodiversity, and many other ecosystem services. In order to address this threat, an understanding of the drivers of deforestation processes is required. Spillover effects and factors that differ across locations and over time play important roles in these processes. They are largely disregarded in applied research and thus in the design of evidence-based policies. In this study, we model connectivity between regions and consider heterogeneous effects to gain more accurate quantitative insights into the inherent complexity of deforestation. We investigate the impacts of agriculture in Mato Grosso, Brazil, for the period 2006-2017 considering spatial spillovers and varying impacts over time and space. Spillovers between municipalities that emanate from croplands in the Amazon appear as the major driver of deforestation, with no direct effects from agriculture in recent years. This suggests a moderate success of the Soy Moratorium and Cattle Agreements, but highlights their inability to address indirect effects. We find that the neglect of the spatial dimension and the assumption of homogeneous impacts lead to distorted inference. Researchers need to be aware of the complex and dynamic processes behind deforestation, in order to facilitate effective policy design.
Introduction: The Spatial Blind Spot in Deforestation Policy
While policy interventions like the Soy Moratorium and Cattle Agreements have been hailed as successes for reducing direct deforestation by major commodities, forest loss in the Amazon remains alarmingly high. This study argues that current research and policy suffer from a "spatial blind spot." Most models assume that land-use changes in one municipality are isolated. In reality, land use is a zero-sum game: when a soy farm expands onto existing pasture, the displaced cattle are often pushed further into the forest frontier.
By using a spatial econometric model, this research reveals that traditional analyses severely underestimate the impact of soy and beef production by ignoring these indirect "leakage" effects. The growth of croplands in one area displaces other activities (like cattle ranching) into the forest elsewhere, making spatial spillovers the primary driver of modern deforestation in Mato Grosso.
Methods and Datasets
The researchers employed a Spatial Durbin Model (SDM), a sophisticated form of spatial panel regression that accounts for "spatial lags." Unlike standard linear models, the SDM accounts for how changes in one municipality affect its neighbors, allowing for the decomposition of impacts into "Direct Effects" (within a municipality) and "Indirect Effects" (spillovers to other municipalities).
The study covers 141 municipalities in Mato Grosso from 2006 to 2017 using high-resolution (250m) MODIS-based land cover maps, which track transitions between forest, pasture, and croplands. This was supplemented by agricultural statistics from the Brazilian Institute of Geography and Statistics (IBGE) on cattle headcounts and soy yields, as well as climate data to control for drought-induced changes.
Key Results: The "Leakage" Reality
The analysis demonstrates that croplands (primarily soy) are no longer the direct "chainsaw" of the forest. Instead, they act as a "shifter." They occupy already cleared land, which then forces cattle ranching to move deeper into the Amazon.
The Cost of Ignoring Space
If a standard model ignoring spatial connectivity is used, it would estimate that a 10% reduction in cropland impact avoids only 165 $km^2$ of deforestation. The spatial model shows the true figure is closer to 1,122 $km^2$. This discrepancy highlights how traditional models fail to capture the full environmental footprint of agricultural expansion.
Biome and Temporal Divergence
In the Amazon biome, deforestation is almost entirely driven by indirect cropland spillovers. In contrast, in the Cerrado and Pantanal, cattle ranching remains a significant direct driver of land clearing. Furthermore, direct pressure from agriculture was more pronounced between 2006 and 2011. In the later period (2012–2017), the pressure became almost exclusively indirect, suggesting that while policies are "cleaning up" supply chains, they aren't stopping the overall trend of forest loss.
Conclusion and Policy Implications
The Soy Moratorium has successfully deterred soy from being planted on freshly deforested land, but it lacks mechanisms to stop soy from displacing cattle into the forest. This highlights the failure of "command-and-control" policies that target single commodities in isolation.
To facilitate effective policy design, the authors suggest moving toward "root-cause" policies, such as land-use taxes or carbon taxes, which might be more effective than voluntary, commodity-specific agreements. Future research must expand this spatial analysis to national and international levels to capture even larger-scale leakage and include other vital vegetation types like the Cerrado's woody savannas, which are currently being converted at higher rates than the Amazon.
Summary updated