VoxEU Column COVID-19 Environment Poverty and Income Inequality

Deforestation in Brazil is spreading COVID-19 to the Indigenous peoples

The existing literature on deforestation focuses on the environmental impacts. Using a novel panel dataset from Brazil, this column finds that deforestation is also playing a significant role in the transmission of COVID-19 to Indigenous populations, with one km2 deforested today estimated to result in 9.5% more new COVID-19 cases among Indigenous people in two weeks. In addition to being an environmental problem, deforestation is also a key health and economic issue, given the importance of curbing the spread of the COVID-19 to save lives and prevent an increase in inequality.

In 2020, Brazil became a worldwide epicentre of COVID-19 as well as deforestation. It represented 15.1% of confirmed COVID-19 cases and 14.3% of total deaths reported globally until the end of August. But the disease has hit ethnic groups differently in Brazil (Baqui et al. 2020, Bruce et al. 2020) and in other countries (Bertocchi and Dimico 2020, McLaren 2020, Vahidy et al. 2020), widening income and racial inequality gaps.  

At the same time, Brazil ranks first in deforestation worldwide. In 2019, the country accounted for over a third of all computed loss in humid tropical primary forests (Global Forest Watch 2020). In the first semester of 2020, deforestation has increased by 25% compared with the same period in the previous year . This year, 55% of the deforested lands have also been burned (Moutinho et al. 2020). On top of that, 72% of the deforested lands in 2020 are taking place in conservation areas and Indigenous lands, the forest's de facto main guardian (Laudares 2016, Baragwanath and Bayi 2020), with the consent of the central government. 

The over 300 Indigenous peoples living within Brazil’s borders are at the crossroad between COVID-19 and deforestation, mostly in the Amazon region.  NGOs  have been vocal in arguing that deforestation, along with illegal mining, land grabbing, timber logging, cattle ranching, and even health workers and missionaries, are transmitting COVID-19 to native populations. In the first semester of 2020, coronavirus infected more than 20,000 and victimised more than 800 indigenous people.  

The association between deforestation and the transmission of COVID-19 is not automatic because it requires human proximity. The reasons deforestation is related to the pathogen's spread are the following. First, the high rate of deforestation observed in Indigenous lands in 2020 entails some level of – peaceful or violent – social interaction. The Articulation of Indigenous Peoples of Brazil (APIB 2020) reports compelling cases of how deforestation can disentangle in conflicts. Second, independent of the motive that triggered the deforestation, Indigenous communities are already exposed to the virus through improper contact with infected people, especially through illegal mining (Sonter 2017, ISA 2020). Finally, the increasing deforestation, combined with droughts and wildfires, puts pressure on Indigenous people to displace to regions where the virus may already be present and worsens respiratory health risks, including COVID-19 cases, increasing the demand for health services and the locomotion to cities (Oliveira et al. 2020, Rocha and Sant'Anna 2020).

Figure 1 COVID-19 cases confirmed in indigenous people, deforestation, and illegal mining in Brazil

What exactly are the effects of deforestation on the spread of COVID-19 to Indigenous peoples? And what are the main transmission mechanisms? In a recent paper (Laudares 2020), I find that deforestation is a powerful and consistent variable in explaining the transmission of COVID-19 to Indigenous populations. Figure 1 exhibits a visual correlation between the variables, and Figure 2 scatters the data. With a new panel dataset covering 5,417 municipalities from 1 March 2020 to 31 August 2020, daily COVID-19 confirmed cases,   and deforestation alerts,  I use a fixed-effect panel estimation model. The estimates show that a one unit increase in deforestation per 100 km2 is associated, on average, with the confirmation of 2.4 to 5.5 new daily cases of COVID-19 in Indigenous people 14 days after the deforestation warnings. In the estimations, I use the lagged values of deforestation to account for the coronavirus' asymptomatic period (Lauer et al. 2020). On a weekly basis, it means that one km2  deforested today results in 9.5% more new COVID-19 cases among Indigenous people in two weeks.

Figure 2 COVID-19 cases and deforestation (km2)

I then consider a cross-section model by accumulating both variables for the entire period and adding other relevant covariates. I find that deforestation explains at least 22% of all COVID-19 cases confirmed in Indigenous people up to 31 August 2020. Population density and economic inequality are control variables that systematically explain the rising COVID-19 morbidity in this case and in developing countries in general (Ahmed et al. 2020). These results are even more concerning in the context of a pandemic that will certainly widen income and opportunity gaps between the rich and the poor (Campelo et al. 2020). 

The evidence suggests that the main mechanisms through which deforestation intensifies human contact between Indigenous and infected people are illegal mining and conflicts.  Municipalities that feature illegal mining or conflict are associated with a increase in defotestation of 179% or 63%, respectively. Using the 'bad controls' framework (Angrist and Pischke 2008),  I estimate that deforestation contributes to the transmission of COVID-19 to originary peoples through illegal mining  (≈84% to 91%) and conflicts (≈81% to 97%). Both transmission mechanisms also elevate the indicence of COVID-19 in the native population through other potential reasons aside from deforestation.

I use an alternative dataset from the Ministry of Health  that reports hospitalisations by race in the robustness check. I am aware that the effect of deforestation on COVID-19 hospitalisations is not direct, especially because a patient's clinical development requires the consideration of additional individual characteristics. However, it is worth using these data as a proxy for COVID-19 incidence to compare the coefficients across races. Relying on a panel data estimation, I find that deforestation is only positively correlated – and statistically significant – with COVID-19 hospitalisation in Indigenous people, but not in black, white, ‘pardo' (mixed), or East Asian people. 

The findings highlighted above show that the negative externalities resulting from deforestation are even larger than initially thought in pandemic times, especially for the most vulnerable segment of the population.  

I believe the reported results are also policy-relevant. The existing literature suggests that ending deforestation is an optimal environmental policy (Stabile et al. 2020, Souza-Rodrigues 2018). My findings show that deforestation is also a key health and economic issue, given the importance of curbing the spread of the COVID-19 to save lives and decreasing the intensity of the economic shocks the pandemic has been causing at the micro and macro level.   


Ahmed,  F  et  al.  (2020),  “Why  inequality  could  spread  COVID-19”, The  Lancet. Public  Health 5(5): 2468-2667.

Angrist, J and J-S Pischke (2008), Mostly harmless econometrics: An empiricist’s companion, Princeton University Press.

APIB (2020), Vidas  Indígenas  Importam!  COVID-19  e  Povos  Indígenas,  o  enfrentamento das violências durante a pandemia, Technical report, Articulação dos Povos Indígenas do Brasil (APIB).

Baqui, P et al. (2020), “Ethnic and regional variations in hospital mortality from COVID-19 in Brazil: a cross-sectional observational study”, The Lancet Global Health.

Baragwanath, K and E Bayi (2020), “Collective property rights reduce deforestation in the Brazilian Amazon”, PNAS  117: 20495–20502.

Bertocchi, G and A Dimico (2020), “COVID-19, Race, and Redlining”  COVID Economics 38. 

Bruce, R et al. (2020), “Racial Inequality in Health Care During a Pandemic”, Mimeo. 

Campello,  M,  G  Kankanhalli,  and  P  Muthukrishnan  (2020), "Corporate  Hiring  under  COVID-19:  Labor  Market  Concentration,  Downskilling, and  Income Inequality", NBER Working Paper 27208. 

Global Forest Watch (2020), 2019 Tree Cover Loss Data.  

Laudares, H (2016), From  Telegraphs to  Space:  Transport Infrastructure,  Development and Deforestation in the Amazon, The Graduate Institute of International and Development Studies.

Laudares, H (2020), “Is deforestation spreading COVID-19 to the indigenous peoples?”, COVID Economics 53. 

Lauer, S A et al. (2020), “The Incubation Period of Coronavirus Disease2019 (COVID-19) From Publicly Reported Confirmed Cases: Estimation and Application”, Annals of Internal Medicine 172 (9), 577–582.  

McLaren, John (2020), “Racial Disparity in COVID-19 Deaths: Seeking Economic Roots with Census data”, NBER Working Paper 27407.

Moutinho, P et al. (2020), The Amazon in flames: deforestations and fire during the COVID-19 pandemic, Amazon Environmental Research Institute

Oliveira, G d et al. (2020), “Smoke pollution’s impacts in Amazonia”, Science 369(6504): 634–635.  

Rocha, R and A Sant’Anna (2020), Winds of Fire and Smoke:  Air  Pollution and Health in the Brazilian Amazon, Instituto de Estudos para Políticas de Saúde (IEPS).

Socioambiental, Instituto (2020), The  impacts  of  the  pandemic  on  the  Yanomami  indigenous territory, . Instituto Socioambiental.

Sonter, L J et al. (2017), “Mining drives extensive deforestation in the BrazilianAmazon”, Science 8. 

Souza-Rodrigues, E (2019), “Deforestation in the Amazon: A unified frameworkfor  estimation  and  policy  analysis”, The  Review  of  Economic  Studies 86(6): 2713–2744.

Stabile, M CC et al. (2020), “Solving Brazil’s land use puzzle: Increasing production and slowing Amazon deforestation”, Land Use Policy 91:104362.

Vahidy, F S et al. (2020), “Racial and ethnic disparities in SARS-CoV-2pandemic: analysis of a COVID-19 observational registry for a diverse US metropolitan population”, BMJ  open 10(8).


1 Data from the Real Time Deforestation Detection System (Deter) of the National Institute for Space Research – INPE.

2 A concrete example of that is the proposed bill on rural land regularization (MP, n.º 910/2019), which incentivised deforestation and land grabbing. 

3 For instance, the SocioAmbiental Institute, the Articulation of Indigenous Peoples of Brazil (APIB), and the Missionary Indianist Council (CIMI). 

4 I use two datasets: one from Special Department of Indigenous Health, Ministry of Health, and another one from the Articulation of Indigenous Peoples of Brazil (APIB). The data of the indigenous healthcare services at the Ministry of Health does not integrate the universal health system, making the indigenous peoples highly underreported in the existing analysis focused on race (Baqui et al. 2020, Bruce et al. 2020).

5 The data on warning areas of deforestation in 100km2 within the Amazon Forest and the Cerrado (Brazilian Savannah) ecosystems was collected from Brazil's National Institute for Spatial Research (INPE).

6 Illegal mining and conflict are mechanisms that can also be the outcome of deforestation. Controlling for them would introduce bias in the estimation results and lead to the `bad control' problem. However, `bad controls' are useful to provide insights about the mechanisms through which the independent variable affects the dependent one, once compared the estimations that contemplate and do not include them (Maccini and Yang 2009).

7 SIVEP-Gripe database. This is the standard database used in the literature.

2,730 Reads