Health benefits of reduced deforestation in the Brazilian Amazon

Deforestation and forest fires in the Amazon increased again after many conservation efforts were abolished by the Bolsonaro government.

“In fact, deforestation and forest fires in the Amazon increased again after many of these conservation efforts were abolished by the Bolsonaro government 65,79 .”

• cites [65] Barlow, J., Berenguer, E., Carmenta, R. & França, F. Clarifying Amazonia’s burning crisis. Global Change Biol. 26, 319–321 (2020). sourcePartly supported

  The Main text (p. 2) reports a marked 2019 increase in deforestation and active fire counts. It also says the government was undermining forest monitoring and cutting law-enforcement resources (p. 3), but does not say that many conservation efforts were abolished. The paper cautions that active fire counts rarely detect forest fires, so those counts do not establish an Amazon-wide increase in forest fires specifically. (LLM, conf=0.86)

  Experimental — accuracy not yet measured
• cites [79] Abessa, D., Famá, A. & Buruaem, L. The systematic dismantling of Brazilian environmental laws risks losses on all fronts. Nat. Ecol. Evol. 3, 510–511 (2019). sourceNot checked — no full text of the source

The effectiveness of private sustainability governance might be compromised by leakage if they operate on small market shares or at limited regional scales.

“Their effectiveness might be compromised by leakage if they operate on small market shares or at limited regional scales 80,81 .”

• cites [80] Paim, M.-A. Zero deforestation in the Amazon: the soy moratorium and global forest governance. Rev. Eur. Compar. Int. Environ. Law 30, 220–232 (2021). sourceNot checked — no full text of the source
• cites [81] Villoria, N., Garrett, R., Gollnow, F. & Carlson, K. Leakage does not fully offset soy supply-chain efforts to reduce deforestation in Brazil. Nat. Commun. 13, 5476 (2022). sourceSupported

In a geographic context, position alongside a border matters; two locations with equal distance to a long border are not comparable if they have very different north-south, east-west, or altitude positions along the border.

“As Keele and Titiunik 90 pointed out, the position alongside a border matters in a geographic context, as two locations with equal distance to a long border are not comparable if they have very different north and south, east and west, or high and low altitude positions along the border.”

• cites [90] Keele, L. J. & Titiunik, R. Geographic boundaries as regression discontinuities. Polit. Anal. 23, 127–155 (2015). sourcePartly supported

  Section 4.2, “Naive Distance” (page unavailable), says that equal perpendicular distances can conceal substantial separation along a long border, supporting the claim about position along it. The paper describes location in two dimensions, however, and does not address differences in altitude. (LLM, conf=0.86)

  Experimental — accuracy not yet measured

Overall findings indicate that tropical forest conservation has a significant public health premium that has so far remained unaccounted for.

“Overall, our findings indicate that tropical forest conservation has a significant public health premium that has so far remained unaccounted for 75 .”

• cites [75] Tallis, H., Kareiva, P., Marvier, M. & Chang, A. An ecosystem services framework to support both practical conservation and economic development. Proc. Natl Acad. Sci. 105, 9457–9464 (2008). sourcePartly supported

  In “Scientific Understanding that Could Improve the Likelihood of Win-Win Outcomes” (page unavailable), the paper links landscape change to disease and says conservation projects have not yet made use of that health connection. It does not establish a significant public-health premium specifically for tropical forest conservation; the citing authors present that as their own finding. (LLM, conf=0.86)

  Experimental — accuracy not yet measured

Deaths associated with respiratory problems decreased significantly.

“We also find significant reductions in deaths associated with respiratory problems.”

The death rate due to the circulatory system decreased significantly and negatively, amounting to 284 (±162) less deaths per year.

“Additionally, the treatment effects of the death rate due to the circulatory system, pregnancy-related health problems, and the perinatal period are also significant and negative, amounting to 284 (±162), 15 (±17) and 83 (±61) less deaths per year, respectively.”

Direct cost reductions of reduced hospitalizations are estimated at almost 6 million US-Dollars per year.

“Regarding cost savings, we only calculate direct cost reductions of reduced hospitalizations, which we estimate at almost 6 million US-Dollars per year.”

The death rate due to perinatal period health problems decreased significantly and negatively, amounting to 83 (±61) less deaths per year.

“Additionally, the treatment effects of the death rate due to the circulatory system, pregnancy-related health problems, and the perinatal period are also significant and negative, amounting to 284 (±162), 15 (±17) and 83 (±61) less deaths per year, respectively.”

Carbon monoxide concentrations decreased by 4.9% (±3.9).

“Concentrations for carbon monoxide decreased by 4.9% (±3.9), nitrogen dioxide by 7% (±5.6), and sulfur dioxide by 6% (±4.6).”

The direct cost reduction from reduced hospitalizations amounts to 5.74 million (±4.84) US-Dollars per year.

“As the average hospitalization in the sample cost around 324 US-Dollars, this amounts to a direct cost reduction of 5.74 million (±4.84) US-Dollars per year (in 2010 US-Dollar rate).”

Around 680 annual deaths were avoided in a sub-sample population of ~4.12 million.

“Hospitalizations linked to fire and smoke-related health issues dropped by almost 18,000 per year and around 680 annual deaths were avoided in a sub-sample population of ~4.12 million.”

Aggregating all effects leads to a reduction of around 678 deaths per year for the local sample.

“Aggregating all effects leads to a reduction of around 678 deaths per year for our local sample.”

The parallel trends test is rejected for hospitalizations due to diseases of the circulatory system and health problems originating during the perinatal period, and the air pollutants nitrogen dioxide and ozone.

“The test is rejected for hospitalizations due to diseases of the circulatory system and health problems originating during the perinatal period, and the air pollutants nitrogen dioxide and ozone.”

The average hospitalization in the sample cost around 324 US-Dollars.

“As the average hospitalization in the sample cost around 324 US-Dollars, this amounts to a direct cost reduction of 5.74 million (±4.84) US-Dollars per year (in 2010 US-Dollar rate).”

The change in deforestation pressure reduces forest fire incidence.

“We find that the change in deforestation pressure reduces forest fire incidence.”

The death rate due to pregnancy-related health problems decreased significantly and negatively, amounting to 15 (±17) less deaths per year.

“Additionally, the treatment effects of the death rate due to the circulatory system, pregnancy-related health problems, and the perinatal period are also significant and negative, amounting to 284 (±162), 15 (±17) and 83 (±61) less deaths per year, respectively.”

Sulfur dioxide concentrations decreased by 6% (±4.6).

“Concentrations for carbon monoxide decreased by 4.9% (±3.9), nitrogen dioxide by 7% (±5.6), and sulfur dioxide by 6% (±4.6).”

The treatment effects amount to around 6430 (±5539) fewer hospitalizations related to pregnancy per year.

“In absolute terms, the treatment effects amount to around 7061 (±6676), 4180 (±2659), 6430 (±5539) and 46 (±53) fewer hospitalizations related to respiratory health problems, the circulatory system, pregnancy and lung cancer per year, respectively.”

The effect size translates to 296 (±120) less deaths in total per year.

“The effect size translates to 7.2 (±2.9) less deaths per 100,000 municipality-inhabitants, or 296 (±120) in total per year.”

The treatment effects amount to around 7061 (±6676) fewer hospitalizations related to respiratory health problems per year.

“In absolute terms, the treatment effects amount to around 7061 (±6676), 4180 (±2659), 6430 (±5539) and 46 (±53) fewer hospitalizations related to respiratory health problems, the circulatory system, pregnancy and lung cancer per year, respectively.”

Reductions in respiratory health problems were the overall largest contributor to these health improvements.

“Reductions in respiratory health problems were the overall largest contributor to these health improvements.”

Fine particulate matter concentrations in the air decrease as a consequence of reduced deforestation pressure.

“As a consequence, fine particulate matter concentrations in the air-a main vector for adverse health effects of fire smoke-also decrease.”

The treatment effects amount to around 4180 (±2659) fewer hospitalizations related to the circulatory system per year.

“In absolute terms, the treatment effects amount to around 7061 (±6676), 4180 (±2659), 6430 (±5539) and 46 (±53) fewer hospitalizations related to respiratory health problems, the circulatory system, pregnancy and lung cancer per year, respectively.”

PM2.5 air pollution decreased to less than 20 μgm-3 on both sides of the biome border after 2006.

“While the pre-2007 annual PM 2.5 concentrations were well above 20 μgm -3 on both sides of the biome border and thus more than four times above the recommended long-term exposure limit of 5 μgm -3 by the WHO 70 , PM 2.5 air pollution decreased to less than 20 μgm -3 on both sides of the biome border after 2006.”

• cites [70] WHO Global Air Quality Guidelines (WHO European Centre for Environment and Health, Geneva, 2021).Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Acute exposure to PM2.5 post-2006 just inside the Amazon biome decreased significantly by, on average, 1.5 (±0.98) percentage points or 5.5 days per year.

“Results in Supplementary Table SI4 show that acute exposure to PM 2.5 post-2006 just inside the Amazon biome indeed decreased significantly by, on average, 1.5 (±0.98) percentage points or 5.5 days per year.”

Ozone was the only unaffected air pollutant.

“The only unaffected air pollutant is ozone.”

Pre-2007 annual PM2.5 concentrations were well above 20 μgm-3 on both sides of the biome border and more than four times above the WHO recommended long-term exposure limit of 5 μgm-3.

“While the pre-2007 annual PM 2.5 concentrations were well above 20 μgm -3 on both sides of the biome border and thus more than four times above the recommended long-term exposure limit of 5 μgm -3 by the WHO 70 , PM 2.5 air pollution decreased to less than 20 μgm -3 on both sides of the biome border after 2006.”

• cites [70] WHO Global Air Quality Guidelines (WHO European Centre for Environment and Health, Geneva, 2021).Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

PM2.5 concentration in the air decreased by 6.6% (±4.8) on average.

“On average, the number of fires decreased by 19.2% (±8.4),PM 2.5 concentration in the air decreased by 6.6% (±4.8), and mortality due to respiratory health problems decreased by 7.2 (±2.9) cases per 100,000 inhabitants.”

The treatment effects amount to around 46 (±53) fewer hospitalizations related to lung cancer per year.

“In absolute terms, the treatment effects amount to around 7061 (±6676), 4180 (±2659), 6430 (±5539) and 46 (±53) fewer hospitalizations related to respiratory health problems, the circulatory system, pregnancy and lung cancer per year, respectively.”

The average reduction in PM2.5 concentration was 1.97 μgm-3 (±0.95) in absolute terms for the period after 2006.

“We find an average reduction of 1.97 μgm -3 (±0.95) in absolute terms for the period after 2006.”

The number of fires decreased by 19.2% (±8.4) on average.

“On average, the number of fires decreased by 19.2% (±8.4),PM 2.5 concentration in the air decreased by 6.6% (±4.8), and mortality due to respiratory health problems decreased by 7.2 (±2.9) cases per 100,000 inhabitants.”

Nitrogen dioxide concentrations decreased by 7% (±5.6).

“Concentrations for carbon monoxide decreased by 4.9% (±3.9), nitrogen dioxide by 7% (±5.6), and sulfur dioxide by 6% (±4.6).”

The overall treatment effect of the prevalence rate of sexually transmitted diseases is insignificant.

“The overall treatment effect of the prevalence rate of STDs is insignificant, as expected.”

Amazon biome focused conservation policies further reduced hospitalizations due to respiratory health problems significantly by 1.71 (±1.62) cases per 1000 inhabitants per municipality per year.

“Still, we estimate that Amazon biome focused conservation policies further reduced hospitalizations due to respiratory health problems significantly by 1.71 (±1.62) cases per 1000 inhabitants per municipality per year.”

Mortality due to respiratory health problems decreased by 7.2 (±2.9) cases per 100,000 inhabitants.

“On average, the number of fires decreased by 19.2% (±8.4),PM 2.5 concentration in the air decreased by 6.6% (±4.8), and mortality due to respiratory health problems decreased by 7.2 (±2.9) cases per 100,000 inhabitants.”

• Also stated in: § Health · plain

  “The effect size translates to 7.2 (±2.9) less deaths per 100,000 municipality-inhabitants, or 296 (±120) in total per year.”

Hospitalizations linked to fire and smoke-related health issues dropped by almost 18,000 per year.

“Hospitalizations linked to fire and smoke-related health issues dropped by almost 18,000 per year and around 680 annual deaths were avoided in a sub-sample population of ~4.12 million.”

For the primary outcomes - number of fires, PM2.5 concentrations, and hospitalization/death rate due to respiratory health problems - the assumption of parallel pre-treatment trends cannot be rejected.

“However, for the primary outcomes of interest, namely the number of fires, PM 2.5 concentrations and the hospitalization/death rate due to respiratory health problems we cannot reject the assumption of parallel pre-treatment trends.”

• AssumptionThe authors state this as an assumption: In the absence of treatment, the treatment and control groups would have followed parallel trends (the parallel trends assumption).

The number of fires just inside the biome developed mostly in parallel to the number of fires just outside the biome before 2007.

“Visually, the number of fires just inside the biome seem to develop mostly in parallel to the number of fires just outside the biome before 2007.”

There is a strong correlation between acute elevated PM2.5 exposure and fire occurrences, even taking PM2.5 background levels into account.

“Considering Supplementary Fig. SI5, we see a strong correlation between acute elevated PM 2.5 exposure and fire occurrences, even taking PM 2.5 background levels into account.”

Placebo double-difference estimates on the pre-treatment period 2003-2006 show significant pre-treatment differences for nitrogen dioxide, ozone, the hospitalization rate for diseases of the circulatory system, and health problems originating during the perinatal period.

“We find significant pretreatment differences for the air pollutants nitrogen dioxide and ozone, as well as the hospitalization rate for diseases of the circulatory system and health problems originating during the perinatal period.”

The health improvements quantified cannot be replicated by private sector initiatives like the Amazon Soy Moratorium or the Cattle Agreement anywhere outside the specific policy context of the Brazilian Amazon.

“Finally, we do not claim that the health improvements we quantify here can be replicated by private sector initiatives like the Amazon Soy Moratorium or the Cattle Agreement anywhere outside the specific policy context of the Brazilian Amazon.”

Air pollution effects from local fires in the Amazonian region were measurable even up to 100 km away.

“Empirical evidence about the Amazonian region suggests that air pollution effects from local fires were measurable even up to 100 km away 33 .”

• cites [33] Morello, T. F. Hospitalization due to fire-induced pollution in the Brazilian Amazon: a causal inference analysis with an assessment of policy trade-offs. World Dev. 161, 106123 (2023). sourceNot checked — no full text of the source

Air quality improvements resulted in significant health benefits for the local population.

“We further demonstrate that air quality improvements resulted in significant health benefits for the local population.”

Smoke and its pollutants disperse aerially, which reduces the likelihood of a sharp discontinuous jump in air pollution and health outcomes at the border.

“Smoke and its pollutants disperse aerially, which reduces the likelihood of a sharp discontinuous jump in air pollution and health outcomes at the border.”

Several conditions must be in place for private sustainability governance to effectively result in tangible environmental and health benefits.

“This indicates that several conditions must be in place for private sustainability governance to effectively result in tangible environmental and health benefits.”

Due to the scale of the analysis, the study relies on remotely sensed measures of fire occurrences and air pollution.

“Second, due to the scale of our analysis, we have to rely on remotely sensed measures of fire occurrences and air pollution.”

State or federal laws concerning the Amazon apply to the defined administrative boundaries of the Brazilian Legal Amazon and thus basically to the entire sample, except for the Soy Moratorium, Central Bank credit restrictions and the Cattle Agreement.

“In our case, state or federal laws concerning the Amazon are applied to the defined administrative boundaries of the BLA and thus basically to our entire sample, with the exception of the three mentioned policies (Soy Moratorium, Central Bank credit restrictions and the Cattle Agreement).”

Smoke exposure from controlled burnings is mainly a function of distance to the fire and wind direction.

“Smoke exposure from controlled burnings is mainly a function of distance to the fire and wind direction 96 .”

• cites [96] Pearce, J. L., Rathbun, S., Achtemeier, G. & Naeher, L. P. Effect of distance, meteorology, and burn attributes on ground-level particulate matter emissions from prescribed fires. Atmos. Environ. 56, 203 – 211 (2012). sourceNot checked — no full text of the source

The study chose 2003 as the first year because the reference satellite for fire measurement changed in 2002, introducing an inconsistency in the time series.

“We chose 2003 as the first year in our analysis since the reference satellite for fire measurement changed in 2002, introducing an inconsistency in the time series.”

Because the biome border is not an administrative boundary, the results are unlikely to be driven by other policy effects.

“Furthermore, as the biome border is not an administrative boundary, our results are unlikely to be driven by other policy effects.”

The identification strategy explores a regulatory discontinuity at the Amazon biome border.

“Our identification strategy explores a regulatory discontinuity at the Amazon biome border, which has been used previously by different authors to estimate deforestation reduction impacts of various conservation policies 55,58,59 .”

• cites [55] Assunção, J., Gandour, C., Rocha, R. & Rocha, R. The effect of rural credit on deforestation: evidence from the Brazilian Amazon. Econ. J. 130, 290–330 (2019).Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

• cites [58] Heilmayr, R., Rausch, L. L., Munger, J. & Gibbs, H. K. Brazil’s Amazon soy moratorium reduced deforestation. Nature Food 1, 801–810 (2020). sourceNot checked — no full text of the source
• cites [59] Moffette, F. & Gibbs, H. K. Agricultural displacement and deforestation leakage in the Brazilian legal Amazon. Land Econ. 97, 155–179 (2021). sourceNot checked — no full text of the source

A difference-in-discontinuities comparison that only considers one-dimensional distance to the border is not adequate because it does not account for two- or three-dimensional aspects of space.

“Therefore, a difference-in-discontinuities comparison as in Supplementary Fig. SI7, which depicts the aggregate annual number of fires by distance to the biome border before and after 2007, is not adequate, as it only considers a onedimensional distance to the border running variable but does not account for the two or three-dimensional aspects of space.”

The combination of regression discontinuity and double-difference approaches relaxes the underlying assumptions of the individual approaches.

“The combination of approaches relaxes the underlying assumptions of the individual approaches.”

For nitrogen dioxide and the circulatory-system hospitalization rate, other factors cannot be ruled out as driving the estimated treatment effects.

“This is a problem for these outcome variables, as we cannot rule out that other factors are driving the effect results, especially for the significant treatment effects found for nitrogen dioxide and the hospitalization rate due to circulatory system health problems.”

The failure to reject parallel pre-treatment trends for the primary outcomes strengthens the robustness and credibility of the double-difference approach for those outcomes.

“This strengthens the robustness and credibility of the double-difference approach for these outcomes.”

Because effects are estimated with a double-difference approach, changes in urban emissions over time would affect the estimates only if they disproportionately affected one side of the border.

“In any way, our effects are estimated using a double-difference approach, so even if emissions from urban sources decreased or increased over time it would only affect our estimates if this change disproportionally affected only one side.”

Estimations using bandwidths other than the preferred 100 km produced results quite similar to the preferred bandwidth.

“The results from estimations with other bandwidths can be found in Supplementary Tables SI7-SI9 and are quite similar to our preferred bandwidth.”

The source of elevated air pollution is the fires.

“For this reason, we are confident that the source of elevated air pollution are the fires.”

Effects of the conservation policies on the bio-climatic variables cannot be ruled out.

“As conservation policies targeted at deforestation that can influence the local ecosystem, we cannot rule out effects on the bio-climatic variables.”