INSIGHTS

Downloaded from http://science.sciencemag.org/ on July 16, 2020

P OLICY FORUM

AGRICULTURE AND ENVIRONMENT

The rotten apples of Brazil’s agribusiness

Brazil’s inability to tackle illegal deforestation puts the future of its agribusiness at risk

By Raoni Rajão,1 Britaldo Soares-Filho,1 American trade bloc. Among the concerns is EU may be contaminated with illegal defor-
Felipe Nunes,1 Jan Börner,2 Lilian Machado,1 that increasing greenhouse gas (GHG) emis- estation. Raising awareness is important to
Débora Assis,1 Amanda Oliveira,1 Luis Pinto,3 sions from deforestation and forest fires in press Brazil to conserve its environmental
Vivian Ribeiro,4 Lisa Rausch,5 Holly Gibbs,5 Brazil could cancel out EU climate change assets and to promote international politi-
Danilo Figueira1 mitigation efforts. The Brazilian govern- cal will for cutting telecoupled GHG emis-
ment and agribusiness contend that national sions. This could be achieved, for example,

I
n the increasingly polarized international laws ensure high conservation standards, through the environmental safeguards of
political arena, it has become difficult and hence trading bans should not include the Mercosur-EU trade agreement, which
to find common ground to solve Brazil’s legally authorized deforestation (1). Here, require EU imports to comply with the ex-
ongoing environmental crisis, which has we address the interlinkage between illegal port country’s legislation.
global as well as local implications. In- deforestation in the Amazon and Cerrado— Our study goes beyond previous assess-
ternational buyers of Brazil’s agricultural the largest Brazilian biomes with the highest ments of soy and beef supply chain trace-
commodities have raised concerns about rates of deforestation—and EU imports of ability and zero-deforestation commitments
products that are contaminated by defores- Brazil’s soy and beef, the country’s major ag-
PHOTO: ANDRE PENNER/AP PHOTO

tation (i.e., deforestation occurred during the ricultural commodities (table S9). Although
1
Universidade Federal de Minas Gerais, Belo Horizonte
process of producing the product) (text S12). most of Brazil’s agricultural output is defor-
31270-901, Brazil. 2Institute for Food and Resource
European Union (EU) criticism of the Brazil- estation-free, we find that 2% of properties in Economics and Center for Development Research,
ian government bolsters demands to boycott the Amazon and Cerrado are responsible for University of Bonn, Bonn, Germany. 3Escola Superior de
Brazilian products and to withhold ratifica- 62% of all potentially illegal deforestation and Conservação Ambiental e Sustentabilidade–IPÊ, Nazaré
Paulista 12960-000, Brazil. 4Stockholm Environmental
tion of the trade agreement reached in 2019 that roughly 20% of soy exports and at least Institute, Stockholm, Sweden. 5University of Wisconsin,
between the EU and Mercosur, the South 17% of beef exports from both biomes to the Madison, WI, USA. Email: rajao@ufmg.br

246 17 JULY 2020 • VOL 369 ISSUE 6501 sciencemag.org SCIENCE

Published by AAAS
 Cattle walk near an illegally burnt deforested area S7), even farmers complying with the soy
in the northern Brazilian state of Para. Beef exports moratorium are clearing the forest for pas-
contaminated by illegal deforestation are a ture or other crops within their holdings, and
key concern among some Brazilian trade partners. hence are still profiting from deforestation.
Despite uncertainties related to mapping
stand apart is its property-level Forest Code and geospatial data modeling (texts S5 to S7
(FC) and national CAR registry system, de- and S11), this represents an area of about 3.7
signed to monitor environmental compliance Mha of soy out of 17.2 Mha planted within
of its rural properties (6) (fig. S3). Brazil’s FC the CAR properties during the 2016–2017
regulates conservation on rural private prop- season (text S7 and figs. S14 and S15). This
erties (1), establishing areas of permanent figure, tantamount to a harvest of 11.3 ± 1.1
protection (APPs) along water streams and million metric tons (text S7 and table S6),
on hilltops as well as legal reserves (native represents a very high level of soy poten-
vegetation in a section of the property) (text tially contaminated with illegal deforesta-
S4). These legal reserves range from 20% of tion, including sizable volumes to the EU.
the property in most of the country (includ- Roughly 41% of EU’s soy imports come from
ing parts of the Cerrado) to 80% in the Ama- Brazil: 13.6 million metric tons per year, of
zon rainforest, the latter in recognition of its which 69% come from this region (table
environmental importance and economic po- S8). Although it is not possible to trace back
tential for forest products (7). soy imports to individual properties, we cal-
Of our CAR sample, roughly 162 thousand culate by using municipality export shares

Downloaded from http://science.sciencemag.org/ on July 16, 2020

of 362 thousand properties (45%) in the that a total of 1.9 ± 0.2 million metric tons
Amazon, and 217 thousand of 452 thousand of soy grown on properties with illegal de-
properties (48%) in the Cerrado, are noncom- forestation may have reached EU markets
pliant with the FC for deforesting APPs or fail- annually during the period of analysis
ing to conserve their minimum legal-reserve (table S1), of which 0.5 million metric tons
areas up until 2008—the deadline year for came from the Amazon (text S7, table S11,
granting amnesty to eligible past deforesters and fig. S21). In sum, 18 to 22% of all soy
(text S4 and figs. S8 to S13). Although these exported from the region to the EU is po-
noncompliance figures do not yet equate to tentially contaminated. Yet the level of con-
illegality, they do entail the obligation to start tamination may exceed the upper bound of
a program of environmental regularization 22%, given that our CAR sample covers only
by 2020, whereby landowners must submit 80% of soy planted in the region (text S11).
and follow a self-designed plan to attain FC With respect to beef, Brazil provides be-
compliance over a period of 20 years (8). tween 25% and 40% of EU beef imports
(2–5), because we explicitly link illegal de- A more pressing issue is illegal deforesta- (table S15). By matching GTAs issued in the
forestation on individual rural properties to tion. Roughly 120 ± 26 thousand properties states of Pará and Mato Grosso in 2017 with
their agricultural production and exports to (15% of our sample) in both biomes were de- CAR data, we identify the origin of 4.1 mil-
EU countries (text S12). To do so, we com- forested after 2008 (1). About 36 thousand lion heads traded to slaughterhouses. Of this
piled a comprehensive set of land-use and of these properties in the Amazon (84%) total, we estimate that 12 ± 2% (0.5 ± 0.1 mil-
deforestation maps for Brazil; information and 27 thousand in the Cerrado (35%) car- lion heads) come directly from properties
on 815 thousand rural properties’ boundaries ried out this deforestation, in all likelihood with potentially illegal deforestation (table
from the Cadastro Ambiental Rural (CAR), illegally (figs. S12 and S13), because these S13). In addition, 48 ± 10% of all slaughtered
the country’s online environmental registry properties had no forest surplus (i.e., veg- heads may be contaminated with potentially
(6); TRASE (Transparency for Sustainable etation above legal-reserve conservation illegal deforestation from indirect suppliers,
Economies) data; and GTA documents (cat- requirements) to be eligible for a deforesta- as the cattle pass from one property to an-
tle transport permits) that are issued when tion permit (see the figure) (text S5). other before being slaughtered (text S7, fig.
animals are traded between properties and A substantial share of this potentially il- S25, and table S13). Although beef exports
to slaughterhouses (table S1). We also de- legal deforestation is linked to agricultural from Pará are negligible, Mato Grosso state
veloped software to deal with the geospatial export commodities. Of 53 thousand proper- is the third largest Brazilian source of EU
data challenge of calculating the level of law ties growing soy in both biomes, 20% were imports (fig. S30). By tracing cattle between
compliance for each individual property, so deforested after 2008, about half of them in properties and slaughterhouses, and tracing
as to differentiate between its potentially a potentially illegal way (text S7, figs. S15 to beef exports from the latter to EU countries,
legal and illegal deforestation alongside its S17, and table S6). In the Cerrado, we find 9.3 we estimate that of 17.7 ± 1.2 thousand metric
production of cattle and soy (texts S3 to S7 ± 1.2 thousand properties with deforestation tons of beef exported from Mato Grosso and
and figs. S4, S6, and S7). after 2008 (43% with potentially illegal defor- Pará in 2017, about 46 ± 7% may have been
estation). In the Amazon, 1.5 ± 0.3 thousand contaminated with potentially illegal defor-
ENVIRONMENTAL COMPLIANCE properties were deforested since 2008, 91% of estation, including both direct and indirect
AND ILLEGALITY which were potentially illegal, despite the soy suppliers (text S7 and figs. S24 to S26).
Many countries have national or regional moratorium that prevents the trading of soy
environmental regulation to protect riparian grown on deforested lands in this biome (5). GHG EMISSIONS, AND A WAY FORWARD
forests, in addition to local zoning laws that Although only 1% of newly deforested areas China and the EU, Brazil’s major agricultural
limit deforestation and the expansion of agri- are being cropped with soy in the Amazon bi- product trade partners, acquired 29% and
cultural and urban areas. What makes Brazil ome, in contrast to 5% in the Cerrado (table 19% of the country’s agricultural exports,

SCIENCE sciencemag.org 17 JULY 2020 • VOL 369 ISSUE 6501 247

Published by AAAS
INS IGHTS | P O L I C Y F O RU M

respectively, over the past 5 years (fig. S2) are responsible for 62% of all potentially il- vegetation (7) that still cover 60% of the
(9). All economic partners of Brazil should legal deforestation (text S10 and table S18). Brazilian territory (1).
share the blame for indirectly promoting de- This small but very destructive portion of the In the EU, public and private initiatives are
forestation and GHG emissions by not bar- sector poses a threat to the economic pros- building up to ensure agricultural imports
ring imports and consuming agricultural pects of Brazil’s agribusiness, in addition to free of tropical deforestation (11), and soon
products contaminated with deforestation, causing regional and global environmental Chinese companies may follow suit (12). Yet
illegal or not. We calculate by superimpos- consequences. It is not enough to claim to so far there is a strong emphasis on private
ing a biomass map on deforestation maps be the world’s most sustainable agriculture certification schemes that are costly, lack
(text S8) that EU soy imports alone could be while a share of the sector fails to comply transparency, and encompass only specific
responsible for the indirect emission of 58.3 with the country’s own environmental laws farms and hence a small part of the sector.
± 11.7 million metric tons of CO2 equivalent and supports the government’s undoing of Here, we demonstrate that thanks to Brazil’s
(MtCO2e) from both legal and illegal defor- past environmental achievements (text S1). already existing CAR registry (6), mapping
estation in the major Brazilian biomes be- Instead, the government and agribusiness and monitoring programs (13), and animal
tween 2009 and 2017 (table S16), taking into should take concrete steps to achieve coun- tracking system (GTA), it is possible to imple-
account municipalities’ export shares. Yet the trywide environmental compliance. This is ment a national and public monitoring sys-
EU share is likely to increase as a result of economically viable, given that about 4.1 Mha tem that enforces environmental compliance
the Mercosur-EU and U.S.-China trade agree- of legal-reserve debts in soy farms could be at the property level to substantially reduce
ments. If implemented, these agreements will compensated by purchasing forest certifi- deforestation in the country’s major agricul-
increase EU demand for Brazilian products cates from landowners with FC surplus (10). tural supply chains. Brazil certainly has all
(text S2) because of lower tariffs and to fill Additionally, the required restoration of 0.6 the elements to feed the world with a respon-
in the gap as U.S. exports to the EU could be Mha of all riparian APPs together with 4.3 sible agricultural sector that tackles climate

Downloaded from http://science.sciencemag.org/ on July 16, 2020

redirected to China. Mha of legal reserves on low-yield pasture- change and protects some of the world’s most
Most of Brazil’s agricultural properties are lands in both biomes would remove 1.4 ± 0.3 biodiverse regions. But to achieve this goal,
deforestation-free. Of our CAR sample, 15% GtCO2e (text S9, fig. S28, and table S17). This the country and its international partners
of properties were deforested after 2008, half will greatly benefit agribusiness because its must acknowledge their shared environmen-
of them potentially illegally. However, only productivity depends on the rainfall regu- tal responsibilities as a main step to seek
2% (17,557) of all properties in both biomes lated by the vast forests and other native common solutions. j
RE FERENCES AND NOTES
1. B. Soares-Filho et al., Science 344, 363 (2014).
Linking rural property deforestation to agriculture and exports 2. J. Karstensen, G. Peters, R. Andrew, Environ. Res. Lett. 8,
024005 (2013).
Though most of Brazil’s agricultural output is deforestation-free, 2% of properties are responsible for 62% of 3. W. Carvalho et al., Perspect. Ecol. Conserv. 17, 122
all potentially illegal deforestation. Roughly 20% of soy and at least 17% of beef exports from both biomes to (2019).
the EU may be contaminated with illegal deforestation. 4. J. Alix-Garcia, H. Gibbs, Glob. Environ. Change 47, 201
(2017).
5. F. Gollnow, L. Hissa, P. Rufin, T. Lakes, Land Use Policy
78, 377 (2018).
6. Sistema Nacional de Cadastro Ambiental Rural,
www.car.gov.br/publico/imoveis/index.
7. J. Strand et al., Nat. Sustain. 1, 657 (2018).
8. D. Rother et al., Trop. Conserv. Sci. 11,
10.1177/1940082918785076 (2018).
9. Ministry of Agriculture, Livestock and Food Supply,
AgroStat—Estatisticas de Comércio Exterior do
Agronegócio Brasileiro [Foreign Trade Statistics of
Brazilian Agribusiness] (2020); http://indicadores.
agricultura.gov.br/agrostat/index.htm.
10. B. Soares-Filho et al., PLOS ONE 11, e0152311 (2016).
AMAZON 11. Sustainable Trade Initiative, National Committee of
the Netherlands, European Soy Monitor (IDH and
IUCN NL, 2019); www.idhsustainabletrade.com/
uploaded/2019/04/European-Soy-Monitor.pdf.
12. L. Jun, “We can feed the world in a sustainable way, but
we need to act now” (World Economic Forum, 2019);
www.weforum.org/agenda/2019/01/we-can-feed-the-
world-in-a-sustainable-way-but-we-need-to-act-now/.
13. Instituto Nacional de Pesquisas Espaciais, Prodes
Project—Deforestation Monitoring of the Brazilian
Amazon Rainforest and Cerrado Biome by Satellite
(2019); www.terrabrasilis.dpi.inpe.br/app/dashboard/
GRAPHIC: ADAPTED FROM RAJÃO ET AL. BY X. LIU/SCIENCE

deforestation.

ACKNOWL EDGMENTS
Supported by the Climate and Land Use Alliance, the
Gordon and Betty Moore Foundation, Conselho Nacional de
Desenvolvimento Científico e Tecnológico, and Fundação de
Conservation units Amparo à Pesquisa do Estado de Minas Gerais. Also suppor-
Indigenous lands CERRADO ted by the Alexander von Humboldt Foundation (B.S.-F. and
Military areas J.B.) and the German Federal Ministry of Education and
Research (J.B.). Dataset, model, and results are available at
Properties with potentially illegal deforestation
csr.ufmg.br/radiografia_do_car.
Properties with potentially legal deforestation
Properties without deforestation SUPPLEMENTARY MATERIA LS
Undesignated areas and other uses science.sciencemag.org/content/369/6501/246/suppl/DC1
Biome boundary
State boundary 10.1126/science.aba6646

248 17 JULY 2020 • VOL 369 ISSUE 6501 sciencemag.org SCIENCE

Published by AAAS
The rotten apples of Brazil's agribusiness
Raoni Rajão, Britaldo Soares-Filho, Felipe Nunes, Jan Börner, Lilian Machado, Débora Assis, Amanda Oliveira, Luis Pinto,
Vivian Ribeiro, Lisa Rausch, Holly Gibbs and Danilo Figueira

Science 369 (6501), 246-248.

DOI: 10.1126/science.aba6646

Downloaded from http://science.sciencemag.org/ on July 16, 2020

ARTICLE TOOLS http://science.sciencemag.org/content/369/6501/246

SUPPLEMENTARY http://science.sciencemag.org/content/suppl/2020/07/15/369.6501.246.DC1
MATERIALS

REFERENCES This article cites 7 articles, 1 of which you can access for free
http://science.sciencemag.org/content/369/6501/246#BIBL

PERMISSIONS http://www.sciencemag.org/help/reprints-and-permissions

Use of this article is subject to the Terms of Service

Science (print ISSN 0036-8075; online ISSN 1095-9203) is published by the American Association for the Advancement of
Science, 1200 New York Avenue NW, Washington, DC 20005. The title Science is a registered trademark of AAAS.
Copyright © 2020, American Association for the Advancement of Science