Much of the current discussion about reducing the effects of climate change entails decreasing the amount of CO₂ released into the atmosphere. Other efforts revolve around technologies that capture CO₂ in the atmosphere and safely store it, often underground. Related to this carbon capture option is a natural alternative—the world’s existing trees—which absorb carbon from the air and store it within a tree’s structure in what scientists call a “carbon sink.”

This natural carbon sink raises an intuitive question: Can we plant our way out of climate change? The short answer is “No,” as it would require too many trees and too much time. It takes dozens of years for trees to mature and to absorb maximum amounts of carbon; and mature forests will not absorb net carbon. In other words, reliance on natural carbon sinks via the world’s forests will not resolve climate change.

However, that does not mean that trees, including existing forests, are ineffective in ameliorating climate change. Indeed, the Amazon forest contains over 100 billion tons of captured carbon, roughly equivalent to the total US emissions of carbon dioxide from 1990 to 2010.

The Brazilian Amazon, the subject of this innovative analysis, occupies 60 percent of the 2.7 million square miles that comprise the Amazon forest. And that section is shrinking: Roughly 17 percent of the Brazilian Amazon has disappeared over the last 50 years due to deforestation. From 1985 to 2021 alone, the area cleared for agriculture increased from 68.6 to 240.5 thousand square miles, an area about the size of Texas. The three maps below show this expansion across the biome at three snapshots — 1995, 2008, and 2017.

The reason deforestation is occurring in the Amazon forest is that land is potentially more valuable to a developer as a source of agriculture—most commonly, cattle ranching—than it is as a carbon sink. In other words, a developer has the option of cutting down part of the Amazon forest and raising cattle on the land, or that developer can receive a subsidy to preserve the forest. If no subsidy exists, or if the subsidy is too low, the developer has less incentive to preserve the forest. The effects of that developer’s decision extend beyond the local market and accrue to the rest of the world in the form of higher CO₂ levels in the atmosphere. Economists call these phenomena externalities, and in this case, the release of CO₂ through deforestation results in a negative externality. This negative externality, which is not priced into the local land market, is an example of market failure.

The contribution of this work is that it offers a framework for policymakers to assess how externally set carbon prices can affect the nexus between agriculture and forest. Importantly, the authors offer a model that captures the tradeoff between agricultural production and forest preservation (or regeneration) across space and time, while also incorporating uncertainty based on fluctuating beef prices and forest carbon measures.

To do this, the authors compose a unique dataset that incorporates cross-sectional variability in cattle farming productivity and in the potential absorption of carbon in the Brazilian Amazon. They divide the Amazon region into various subregions, from 1,059 sites measuring 67.5km × 67.5km, to 81 sites measuring 270km × 270km. The two maps below show the two key parameters the authors estimate for each site: cattle farming productivity (θ), which measures how economically valuable a site is for agriculture, and forest carbon density (γ), which measures how much carbon a site stores as standing forest. Together, θ and γ are the inputs the authors’ model uses to weigh agricultural production against forest preservation at each site across the biome.

The authors first attempt to derive the price of carbon implicit in the deforestation that occurred from 1995, the first year with reliable cattle price data, through 2008, which marks the beginning of the Amazon Fund, a pay-for-performance scheme financed primarily by the Norwegian and German governments. They show that this implicit price was only around $6 per ton of CO₂e, and use this price as a “shadow price” to produce simulations when no transfers to the Brazilian government prevail.

The authors then study the impact of adding outside payments for net capture of CO₂ in the Amazon to determine whether and by how much Brazil would gain if the country signed an agreement for a set of hypothetical dollar transfers per net unit of CO₂. The two maps below project the Amazon in 2050 under two policy regimes: (a) the baseline, where only the $6 shadow price prevails, and (b) a scenario where Brazil receives an additional $20 per ton of CO₂ captured. Each cell is colored by the share of land allocated to agriculture in 2050. Darker red sites show more deforestation, lighter sites show preserved forest.

The authors do not prescribe a single optimal carbon price, as that is the purview of policymakers and governments. However, they offer suggestive guidelines. According to existing research, deforestation in the Amazon will likely cross a tipping point of 20–25% if the carbon price stays at the shadow emission price of $6 per ton of CO₂e with no additional payments. The authors’ model reveals that additional payments of at least $15/ton would not only safeguard against reaching the tipping point, but would also trigger reforestation on a large scale.

If transfers per ton exceeded $20, optimal management of the Brazilian Amazon forest alone could deliver in excess of 39 billion tons reduction in atmospheric CO₂e over the next 30 years from restoration, avoided land conversion, and other forestry practices, a substantial contribution toward the goal of having a 50% chance of not exceeding 1.5 degrees of warming. The first chart below traces Z, the model’s measure of the share of arable land allocated to agriculture across the Amazon. The second traces X, the total carbon stored in the Amazon forest in billions of tons of CO₂e. Both are shown over 50 years (2017 to 2067) for each of the five payment scenarios: no transfer (b = $0) plus four levels of carbon payment ($10, $15, $20, $25 per ton).

By offering a dynamic approach that accounts for expected future prices of agricultural goods, as well as carbon emissions from deforestation and carbon capture from forest regeneration, and which also incorporates uncertainties surrounding measures of agricultural goods and forest carbon, the authors offer a timely framework for better decision-making on an issue of global importance.