Science1 publisher3 min readPublished
Bonn-led study ties 190 to 1,700 square kilometers of 2019 Amazon clearing to Bolsonaro government tweets
Bonn-led researchers estimate that anti-conservation tweets from Jair Bolsonaro's government added 190 to 1,700 square kilometers of Amazon clearing in 2019. Local fines and bans held steady, so the authors trace the extra clearing to lowered expectations of prosecution.
The Scientist · Science desk
Drafted by a language model from the sources cited here and checked against its claim ledger before publication. How we use AISend a correction
What happened
- From more than 3,600 posts by Bolsonaro and his agriculture, environment and foreign ministers in 2019, the team flagged 38 that signalled a lower risk of enforcement.
- Exposure in 578 Amazon municipalities was estimated from how often residents had tweeted about forests between 2015 and 2018.
- In the regions at greatest risk, deforestation rose significantly in months when the government published anti-conservation posts, according to the study.
- The effect was no stronger where Bolsonaro's support was high, but it was stronger where farm returns were high and past conservation policy had restricted clearing.
Compiled by The ScientistSomething wrong?How this is made
Why it matters
- constraint Fines and ban counts stayed flat while clearing rose, so an agency judging its deterrence by enforcement statistics alone would have missed this loss.
- exposure The places where farming pays well and past rules had held clearing back responded most, so the areas conservation policy already works hardest to protect are the most sensitive to a leader's signals.
- decision Anyone costing the damage has to pick a point in a range whose top is almost nine times its bottom, and only the 190 square kilometer floor is safe to cite as a minimum.
- precedent If expectations of prosecution can move while enforcement stays put, studies of compliance with other laws will need to measure what officials say alongside budgets and penalties.
The hard part of a study like this is the comparison. Brazil's environmental authorities were greatly weakened across Bolsonaro's presidency [14]. Setting 2019 against earlier years would therefore mix the effect of words with the effect of a weaker agency. As phys.org describes it, the Bonn team's comparison works on timing and reach inside 2019: clearing in months with a flagged post against other months, in places more or less likely to hear about those posts [6][5]. The clearing data came from MapBiomas Alerta, which reviews and validates deforestation alerts from several sources covering all of Brazil [15].
Reach is the weakest part of the design. It was measured from residents' own forest tweets in 2015 to 2018 [5], a window that closes before the 2019 government posts [3]. The step from there to a decision on the ground is an assumption, and the first-named author, Gustavo Magalhães de Oliveira of the University of Bonn, stated it as one [18][17]. "We assume that such messages spread through word of mouth," he said. "Supporters share them, as do critics, and they are discussed locally until they reach the people who actually make decisions about deforestation." [13] The thing this doesn't tell you is whether any given person clearing a plot saw any given post. A second exposure measure narrows that gap. "We see the same pattern when we analyze other forms of exposure, such as television," Oliveira said [7].
The strongest part of the causal case is the enforcement record. The statements had no influence on municipal environmental fines or deforestation bans [8]. "The effect is thus not attributable to any changes in the local enforcement of forest conservation laws," Oliveira said [12]. The political and economic splits point the same way: partisanship did not amplify the response, and profitability did [9][10]. That pattern fits people recalculating the cost of getting caught. "Our analysis indicates that, as a result of the messages, people change their expectations of the extent to which they will be prosecuted for breaking the law," Oliveira said [11].
The effect size is wide. For 2019 the estimate runs from 190 to 1,700 square kilometers [2], and the top is almost nine times the bottom [1]. The "up to 1,700" figure in phys.org's account is the ceiling of that range [2]. The article does not report total Amazon clearing for 2019, so the share of that year's loss tied to the posts cannot be worked out from it. The trigger was small: 38 flagged posts out of more than 3,600 examined, about 1 percent [4][3][2].
I think the design supports the causal reading for Brazil's Amazon municipalities in 2019. I would carry the 190 figure, not the 1,700, into any costing of the damage. The article suggests the effect may extend beyond forest conservation and beyond the Brazilian Amazon [16]. The evidence it describes comes from the first year of one presidency [19].
What to watch
- The paper's own central estimate and uncertainty intervals, which would show where in the 190 to 1,700 square kilometer range the evidence sits.
- Replications in other countries or for other laws that test the authors' suggestion that the effect reaches beyond forest conservation and the Amazon.
- Whether the same timing-and-exposure design finds a response to official statements in years after 2019.