Abstract

This article analyzes the sociodemographic composition of the Indigenous population in Brazil in the 1991, 2000, and 2010 demographic censuses, in addition to investigating inequalities in access to basic sanitation and electricity based on the 2010 Census. A methodology is proposed that classifies households with Indigenous residents as “homogeneous”, “mixed” with an Indigenous in the category “household head”, and mixed with a non-Indigenous in the category “household head”. Regional and situational differences overlapped with differences by type of household and location, with better conditions in urban than rural areas and in the Southeast and South. The North was the region with the least access in general. Regional differences and differences between urban/rural areas were observed, highlighting better urban conditions in the Southeast and South. These inequalities reflect fragmented public policies, economic pressures, and deterritorialization, among other factors. The development of new techniques and critical debate is essential to understanding and addressing racial inequalities in the country and promoting public policies appropriate to Indigenous peoples.

Key words:
Indigenous peoples; Census; Demography; Public policy

Introduction

The demographic censuses conducted by the Brazilian Institute of Geography and Statistics (IBGE) are crucial sources of data on the sociodemographic characteristics of the Brazilian Indigenous population for their coverage and the possibility of investigating a wide range of topics¹1 Campos MB, Estanislau BR. Demografia dos povos indígenas: os censos demográficos como ponto de vista. Rev Bras Estud Popul 2016; 33(2):441-449.

2 Silva LO. Domicílios com indígenas nos censos demográficos de 1991, 2000 e 2010 no Brasil: composição e análises de inter-relações entre composição segundo cor ou raça e condições socioeconômicas e sanitárias [tese]. Rio de Janeiro: Fiocruz; 2021.

3 Silva LO, Antunes MO, Damasco F. Povos indígenas nas estatísticas oficiais: identificação étnica, recomendações internacionais e a experiência brasileira [Internet]. 2018. [acessado 2023 ago 3]. Disponível em: https://api.saudeindigena.icict.fiocruz.br/api/core/bitstreams/1382140a-839b-47b5-a364-cb2764eba37f/content
https://api.saudeindigena.icict.fiocruz....

4 Raupp L, Fávaro TR, Cunha GM, Santos RV. Condições de saneamento e desigualdades de cor/raça no Brasil urbano: uma análise com foco na população indígena com base no Censo Demográfico de 2010. Rev Bras Epidemiol 2017; 20(1):1-15.

5 Santos RV, Guimarães BN, Simoni ATS, LO Silva, Antunes MO, Damasco FS, Colman RS, Azevedo MMA. The identification of the Indigenous population in Brazil's official statistics, with an emphasis on demographic censuses. Statistical J IAOS 2023; 35(1)29-46.^-66 Santos RV, Guimarães BN, Simoni AT. Cor ou raça: indígena? Contextos e recepções da inclusão de uma categoria no Censo Demográfico 1991. Rev Fr-Bras Geogr 2023; 59. DOI: 10.4000/confins.51565
https://doi.org/10.4000/confins.51565... . It is a “mosaic of micro-societies with small or medium-sized populations, with distinct demographic dynamics, revealing not only cultural autonomies but also very different epidemiological states or quality of life”⁷7 Azevedo MM. Os povos indígenas e os censos demográficos no Brasil. In: Berquó E, organizadore. Demografia na Unicamp: um olhar sobre a produção do Nepo. Campinas: Unicamp; 2017. p. 551-564.. Although censuses cannot portray the specificities of each population, their relevance to public policy purposes is undeniable⁶6 Santos RV, Guimarães BN, Simoni AT. Cor ou raça: indígena? Contextos e recepções da inclusão de uma categoria no Censo Demográfico 1991. Rev Fr-Bras Geogr 2023; 59. DOI: 10.4000/confins.51565
https://doi.org/10.4000/confins.51565...

7 Azevedo MM. Os povos indígenas e os censos demográficos no Brasil. In: Berquó E, organizadore. Demografia na Unicamp: um olhar sobre a produção do Nepo. Campinas: Unicamp; 2017. p. 551-564.

8 Anderson I, Robson B, Connolly M, Al-Yaman F, Bjertness E, King A, Tynan M, Madden R, Bang A, Coimbra CE Jr, Pesantes MA, Amigo H, Andronov S, Armien B, Obando DA, Axelsson P, Bhatti ZS, Bhutta ZA, Bjerregaard P, Bjertness MB, Briceno-Leon R, Broderstad AR, Bustos P, Chongsuvivatwong V, Chu J, Deji, Gouda J, Harikumar R, Htay TT, Htet AS, Izugbara C, Kamaka M, King M, Kodavanti MR, Lara M, Laxmaiah A, Lema C, Taborda AM, Liabsuetrakul T, Lobanov A, Melhus M, Meshram I, Miranda JJ, Mu TT, Nagalla B, Nimmathota A, Popov AI, Poveda AM, Ram F, Reich H, Santos RV, Sein AA, Shekhar C, Sherpa LY, Skold P, Tano S, Tanywe A, Ugwu C, Ugwu F, Vapattanawong P, Wan X, Welch JR, Yang G, Yang Z, Yap L. Indigenous and tribal peoples' health (The Lancet-Lowitja Institute Global Collaboration): a population study. Lancet 2016; 388(10040):131-157.^-99 Pagliaro H, Azevedo MM, Santos RV. Demografia dos povos indígenas no Brasil. Rio de Janeiro: Editora Fiocruz; 2005.. The population contingents enumerated by national censuses are commonly used in calculating numerous health indicators (generally as denominators)¹⁰10 Araújo EM, Costa MCN, Hogan VK, Araújo TM, Dias AB, Oliveira LOA. A utilização da variável raça/cor em Saúde Pública: possibilidades e limites. Interface (Botucatu) 2009; 13(31):383-394.

11 Campos MB, Borges GM, Queiroz BL, Santos RV. Diferenciais de mortalidade entre indígenas e não indígenas no Brasil com base no Censo Demográfico de 2010. Cad Saude Publica 2017; 33(5):e00015017.^-1212 Coimbra EA. Saúde e povos indígenas no Brasil: reflexões a partir do I Inquérito Nacional de Saúde e Nutrição Indígena. Cad Saude Publica 2014; 30(4):855-859..

Since the change in the skin color question to skin color or ethnicity in the 1991 Census, when the category “Indigenous” was introduced, a historical series started to serve as a reference for several studies on the demographics of Indigenous peoples and analyses on inequality patterns focusing on this segment of Brazilian society²2 Silva LO. Domicílios com indígenas nos censos demográficos de 1991, 2000 e 2010 no Brasil: composição e análises de inter-relações entre composição segundo cor ou raça e condições socioeconômicas e sanitárias [tese]. Rio de Janeiro: Fiocruz; 2021.^,33 Silva LO, Antunes MO, Damasco F. Povos indígenas nas estatísticas oficiais: identificação étnica, recomendações internacionais e a experiência brasileira [Internet]. 2018. [acessado 2023 ago 3]. Disponível em: https://api.saudeindigena.icict.fiocruz.br/api/core/bitstreams/1382140a-839b-47b5-a364-cb2764eba37f/content
https://api.saudeindigena.icict.fiocruz.... ^,55 Santos RV, Guimarães BN, Simoni ATS, LO Silva, Antunes MO, Damasco FS, Colman RS, Azevedo MMA. The identification of the Indigenous population in Brazil's official statistics, with an emphasis on demographic censuses. Statistical J IAOS 2023; 35(1)29-46.^,99 Pagliaro H, Azevedo MM, Santos RV. Demografia dos povos indígenas no Brasil. Rio de Janeiro: Editora Fiocruz; 2005.^,1313 Instituto Brasileiro de Geografia e Estatistica (IBGE). Tábua completa de mortalidade para o Brasil - 2015: breve análise da evolução da mortalidade no Brasil. Rio de Janeiro: IBGE; 2015.^,1414 Wong LLR. Tendências da fecundidade dos povos indígenas nos Censos Demográficos brasileiros de 1991 a 2010. Rev Bras Estud Popul 2016; 33(2):399-421.. The skin color issue, applied between the 1940 and 1980 Censuses, except for 1970, offered the options “White”, “Black”, “Yellow”, and “Brown”, and the IBGE²2 Silva LO. Domicílios com indígenas nos censos demográficos de 1991, 2000 e 2010 no Brasil: composição e análises de inter-relações entre composição segundo cor ou raça e condições socioeconômicas e sanitárias [tese]. Rio de Janeiro: Fiocruz; 2021. classified as “brown” Indigenous declarations. The language spoken (including Indigenous language) was investigated in 1950 and 1960. The other two occasions on which the Indigenous population was censused occurred in 1872 and 1890¹⁵15 Petruccelli JL, Sabóia AL. Características étnicos-raciais da população: classificações e identidades. v. 2. Rio de Janeiro: IBGE; 2013.^,1616 Dias Júnior CS, Verona AP. Os indígenas nos Censos demográficos brasileiros pré-1991. Rev Bras Estud Popul 2018; 35(3):e0058.. The collection of information on Indigenous language spoken at home and specific ethnic affiliation (i.e., Indigenous ethnic groups) was introduced in the 2010 Census.

The IBGE collected data on the Indigenous population in the last four demographic censuses carried out in the country in 1991, 2000, 2010, and, most recently, 2022. This population segment is distributed across all federation units and in most Brazilian municipalities, with spatial distribution patterns distinct from the Brazilian population in general. According to the 2022 Census, 4,832 (87.8%) of the 5,568 municipalities in the country had at least one Indigenous declaration¹⁷17 Instituto Brasileiro de Geografia e Estatística (IBGE). Indígenas: primeiros resultados do universo [Internet]. 2022. [acessado 2023 ago 7]. Disponível em: https://www.ibge.gov.br/estatisticas/sociais/populacao/22827-censo-demografico-2022.html
https://www.ibge.gov.br/estatisticas/soc... . These peoples’ sociodiversity was evidenced by the results of the 2010 Census, which identified 305 ethnic groups and 275 Indigenous languages ​​spoken in the country¹⁸18 Instituto Brasileiro de Geografia e Estatística (IBGE). Censo Demográfico 2010: características gerais dos indígenas (resultados do universo). Rio de Janeiro: IBGE; 2012.. The number of Indigenous peoples showed a progressive increase throughout the censuses, with significant differences between the 1991 and 2000 Censuses, up from 294,131 to 734,128 individuals, and between the 2010 and 2022 Censuses, up from 896,917 to 1,693,535 declared Indigenous peoples. The 2000 Census, in turn, highlighted the presence of Indigenous people in urban contexts, identifying a number 5.4 times greater in these areas and, in rural areas, 1.5 times greater than in the 1991 Census. Most of the population resided in rural areas in the 1991 and 2010 Censuses.

Oliveira highlights, regarding the possible reasons for the increase in the Indigenous population in official statistics since the 1990s, the interrelations between the Indigenous issue and the environmental agenda; the international guidelines regarding the participation of civil society in planning and decisions related to public policies; escalating the international agenda for discussing and establishing agreements and conventions related to the rights of Indigenous peoples; and the growing appreciation of identities and cultural heritage of these populations by the Brazilian government and perceived globalized world trends¹⁹19 Oliveira JP. Mensurando alteridades, estabelecendo direitos: práticas e saberes governamentais na criação de fronteiras étnicas. Dados 2012; 55(4):1055-1088.. Other factors that somehow dialogue with the reasons presented by Oliveira include adopting the skin color or ethnicity issue in other information systems and inequality studies; the “volatile” classification of skin color or ethnicity issue; and methodological innovations for the investigating Indigenous peoples in demographic censuses, among other factors⁶6 Santos RV, Guimarães BN, Simoni AT. Cor ou raça: indígena? Contextos e recepções da inclusão de uma categoria no Censo Demográfico 1991. Rev Fr-Bras Geogr 2023; 59. DOI: 10.4000/confins.51565
https://doi.org/10.4000/confins.51565...

7 Azevedo MM. Os povos indígenas e os censos demográficos no Brasil. In: Berquó E, organizadore. Demografia na Unicamp: um olhar sobre a produção do Nepo. Campinas: Unicamp; 2017. p. 551-564.

8 Anderson I, Robson B, Connolly M, Al-Yaman F, Bjertness E, King A, Tynan M, Madden R, Bang A, Coimbra CE Jr, Pesantes MA, Amigo H, Andronov S, Armien B, Obando DA, Axelsson P, Bhatti ZS, Bhutta ZA, Bjerregaard P, Bjertness MB, Briceno-Leon R, Broderstad AR, Bustos P, Chongsuvivatwong V, Chu J, Deji, Gouda J, Harikumar R, Htay TT, Htet AS, Izugbara C, Kamaka M, King M, Kodavanti MR, Lara M, Laxmaiah A, Lema C, Taborda AM, Liabsuetrakul T, Lobanov A, Melhus M, Meshram I, Miranda JJ, Mu TT, Nagalla B, Nimmathota A, Popov AI, Poveda AM, Ram F, Reich H, Santos RV, Sein AA, Shekhar C, Sherpa LY, Skold P, Tano S, Tanywe A, Ugwu C, Ugwu F, Vapattanawong P, Wan X, Welch JR, Yang G, Yang Z, Yap L. Indigenous and tribal peoples' health (The Lancet-Lowitja Institute Global Collaboration): a population study. Lancet 2016; 388(10040):131-157.

9 Pagliaro H, Azevedo MM, Santos RV. Demografia dos povos indígenas no Brasil. Rio de Janeiro: Editora Fiocruz; 2005.

10 Araújo EM, Costa MCN, Hogan VK, Araújo TM, Dias AB, Oliveira LOA. A utilização da variável raça/cor em Saúde Pública: possibilidades e limites. Interface (Botucatu) 2009; 13(31):383-394.

11 Campos MB, Borges GM, Queiroz BL, Santos RV. Diferenciais de mortalidade entre indígenas e não indígenas no Brasil com base no Censo Demográfico de 2010. Cad Saude Publica 2017; 33(5):e00015017.

12 Coimbra EA. Saúde e povos indígenas no Brasil: reflexões a partir do I Inquérito Nacional de Saúde e Nutrição Indígena. Cad Saude Publica 2014; 30(4):855-859.

13 Instituto Brasileiro de Geografia e Estatistica (IBGE). Tábua completa de mortalidade para o Brasil - 2015: breve análise da evolução da mortalidade no Brasil. Rio de Janeiro: IBGE; 2015.

14 Wong LLR. Tendências da fecundidade dos povos indígenas nos Censos Demográficos brasileiros de 1991 a 2010. Rev Bras Estud Popul 2016; 33(2):399-421.

15 Petruccelli JL, Sabóia AL. Características étnicos-raciais da população: classificações e identidades. v. 2. Rio de Janeiro: IBGE; 2013.

16 Dias Júnior CS, Verona AP. Os indígenas nos Censos demográficos brasileiros pré-1991. Rev Bras Estud Popul 2018; 35(3):e0058.

17 Instituto Brasileiro de Geografia e Estatística (IBGE). Indígenas: primeiros resultados do universo [Internet]. 2022. [acessado 2023 ago 7]. Disponível em: https://www.ibge.gov.br/estatisticas/sociais/populacao/22827-censo-demografico-2022.html
https://www.ibge.gov.br/estatisticas/soc...

18 Instituto Brasileiro de Geografia e Estatística (IBGE). Censo Demográfico 2010: características gerais dos indígenas (resultados do universo). Rio de Janeiro: IBGE; 2012.

19 Oliveira JP. Mensurando alteridades, estabelecendo direitos: práticas e saberes governamentais na criação de fronteiras étnicas. Dados 2012; 55(4):1055-1088.^-2020 Muniz JO, Bastos JL. Volatilidade classificatória e a (in)consistência da desigualdade racial. Cad Saude Publica 2017; 33(Supl. 1):e00082816..

This article aims to analyze the composition by skin color or ethnicity in households with at least one Indigenous declaration through a household typology and measure the inequality between the proposed types in access to primary services to investigate differences between the observed population subgroups. The sociodemographic analysis of declared Indigenous people uses data from the 1991, 2000, and 2010 Census samples to sociodemographically characterize and identify significant patterns and differences between types of households. The analysis of access to services by type of residence, in turn, uses data from the 2010 Census universe, as these allow for greater spatial disaggregation in estimating values, emphasizing inequalities not observed in the sample data, such as between Indigenous people inside and outside Indigenous lands. Using data from the universe also facilitated using a coverage question for Indigenous identification, which increased the total number of individuals considered by 8.8%.

This is an unprecedented approach in research using census data, as it encompasses three recent national censuses. It uses a new typology of households per composition of different arrangements by skin color or ethnicity of residents, exceeding the Indigenous and non-Indigenous categories⁴4 Raupp L, Fávaro TR, Cunha GM, Santos RV. Condições de saneamento e desigualdades de cor/raça no Brasil urbano: uma análise com foco na população indígena com base no Censo Demográfico de 2010. Rev Bras Epidemiol 2017; 20(1):1-15.^,1111 Campos MB, Borges GM, Queiroz BL, Santos RV. Diferenciais de mortalidade entre indígenas e não indígenas no Brasil com base no Censo Demográfico de 2010. Cad Saude Publica 2017; 33(5):e00015017.. In this sense, the study has a dimension of methodological proposition by indicating a typology for households with Indigenous presence to analyze inequalities by skin color or ethnicity, remembering that other authors also used arrangements by skin color or ethnicity to analyze census data geared to Indigenous populations²¹21 Marinho GL, Caldas ADR, Santos RV. Indígenas residentes em domicílios "improvisados" segundo o Censo Demográfico 2010. Physis 2017; 27(1):79-102.

22 Marinho GL, Bastos JL, Longo LAFDB, Tavares FG. Classificação de cor/raça de filhos em domicílios indígenas no Brasil. Cad Saude Publica 2019; 35(Supl. 3):e00006119.^-2323 Marinho GL, Santos RV, Simoni AT. "Você se considera indígena?": características da população residente em terras indígenas investigadas pelo Censo Demográfico de 2010. In: Santos RV, Guimarães BN, Campos MB, Azevedo MMA, organizadores. Entre demografia e antropologia: povos indígenas no Brasil. Rio de Janeiro: Editora Fiocruz; 2019. p. 179-200..

Methods

The household dimension is central to building and analyzing indicators in the health field, such as sanitation infrastructure and other public services. According to the IBGE methodology, households can be classified as private (permanent or improvised) or collective. Permanent private households serve exclusively as housing; improvised households are located in buildings without an area exclusively intended for housing, such as stores and factories. In both cases, relationships are governed by “kinship ties”, “domestic dependency ties”, or “socialization rules”, according to IBGE terminology²⁴24 Instituto Brasileiro de Geografia e Estatística (IBGE). Censo demográfico 2010: características da população e dos domicílios - resultados do universo [Internet]. 2010. [acessado 2023 jun 23]. Disponível em: https://biblioteca.ibge.gov.br/index.php/biblioteca-catalogo?view=detalhes&id=793
https://biblioteca.ibge.gov.br/index.php... . Hotels, motels, campsites, boarding houses, penitentiaries, prisons, and detention centers represent collective households. In these households, relationships are based on “administrative subordination rules”²⁴24 Instituto Brasileiro de Geografia e Estatística (IBGE). Censo demográfico 2010: características da população e dos domicílios - resultados do universo [Internet]. 2010. [acessado 2023 jun 23]. Disponível em: https://biblioteca.ibge.gov.br/index.php/biblioteca-catalogo?view=detalhes&id=793
https://biblioteca.ibge.gov.br/index.php... . This study only considered individuals living in permanent and improvised private homes.

Considering the variations observed in the three censuses regarding the Indigenous population regarding volume and urban-rural distribution, among other dimensions, this research compares, firstly, the ethnic-racial composition, structure by gender and age, the relationship between residents, number of residents, and spatial distribution of declared Indigenous people per the proposed household type, to highlight significant patterns and differences between the observed population subgroups. We used microdata from the 1991, 2000, and 2010 Census samples for this analysis since those for 2022 are not yet available for research.

Then, we proceeded to classify the data from the 2010 Census universe using the same criteria adopted for the sample data and adding the variables “Do you consider yourself Indigenous?” (yes or no) and location regarding Indigenous lands (inside or outside), not available in the sample data, and those related to basic sanitation and electricity infrastructures. Only permanent households were considered to characterize and analyze access to services since the IBGE does not collect this information for improvised households. Using this database, we characterized and measured access to services and conditions observed among Indigenous people declared in the 2010 Census using multivariate statistical modeling.

The choice of only one period for the analysis of sanitation and electricity conditions (2010) was due to the potential analytical possibilities arising from the disaggregation and the increase in population from the question “Do you consider yourself”. We should underscore that, in previous censuses, the question skin color or ethnicity was part of the sample questionnaire, and the variables “Do you consider yourself Indigenous?” and location concerning Indigenous lands were not part of the censuses.

Work variables and type of households with Indigenous people

The variables considered in this study were gender, age, housing conditions, skin color or ethnicity, total number of residents in the household, and access to electricity, water, garbage collection, and bathroom in the household. The spatial dimensions analyzed were Brazil, Major Regions (North, Northeast, Southeast, South, or Midwest), situation (urban or rural), and location (inside or outside Indigenous land).

The skin color or ethnicity is that declared by the informant to the census taker²⁵25 Instituto Brasileiro de Geografia e Estatística (IBGE). Censo demográfico 2010: manual do recenseador [Internet]. 2010. [acessado 2023 ago 22]. Disponível em: https://biblioteca.ibge.gov.br/index.php/biblioteca-catalogo?id=52601&view=detalhes
https://biblioteca.ibge.gov.br/index.php... for the question “What is your skin color or ethnicity?”, whose possible answers are “White”, “Black”, “Yellow”, “Brown” and “Indigenous”. The question “Do you consider yourself Indigenous?” was asked in the 2010 Census only within Indigenous lands when another skin color or ethnicity option was provided²⁵25 Instituto Brasileiro de Geografia e Estatística (IBGE). Censo demográfico 2010: manual do recenseador [Internet]. 2010. [acessado 2023 ago 22]. Disponível em: https://biblioteca.ibge.gov.br/index.php/biblioteca-catalogo?id=52601&view=detalhes
https://biblioteca.ibge.gov.br/index.php... .

The condition in the household characterizes the “relationship between the household head [...] and each of the other residents”¹⁸18 Instituto Brasileiro de Geografia e Estatística (IBGE). Censo Demográfico 2010: características gerais dos indígenas (resultados do universo). Rio de Janeiro: IBGE; 2012., in which the household head is “10 years of age or older, recognized by the residents as the household head”²⁴24 Instituto Brasileiro de Geografia e Estatística (IBGE). Censo demográfico 2010: características da população e dos domicílios - resultados do universo [Internet]. 2010. [acessado 2023 jun 23]. Disponível em: https://biblioteca.ibge.gov.br/index.php/biblioteca-catalogo?view=detalhes&id=793
https://biblioteca.ibge.gov.br/index.php... . This variable was also adopted to characterize the kinship or cohabitation relationships between the residents in the analysis of the household composition of the types created. The variables related to services focus on the type of access to electricity and water in the home, the disposal of sewage and residential waste, and the number of bathrooms in the home.

We paired the databases of people and households from the microdata to “reconstitute” the households and residents surveyed to classify individuals. Only households with at least one Indigenous declaration were considered. Each declared Indigenous person was then classified as living in one of the following types of households: (a) homogeneous, those in which all residents were declared Indigenous; (b) mixed with Indigenous head, those with residents who declared themselves Indigenous and also of other skin color or ethnicity options, and the household head was declared Indigenous; and; (c) mixed with non-Indigenous head, those with residents who declared themselves Indigenous and of other skin color or ethnicity options, and the household head was not declared Indigenous. The item “Do you consider yourself Indigenous?” was also used in the classification of the 2010 universe data to identify Indigenous people.

Expanded sample

The weighting area variable, usually recommended as a stratum for data expansion, was unavailable in the 1991 Census microdata. For this reason, the municipality variable was adopted as the stratum for the 1991 Census. The weighting area was used as the stratum for sample expansion²⁶26 Dias AJR, Albieri S. As implicações do uso de calibração no procedimento de expansão da amostra do Censo Demográfico 2000. In: Anais do XIV Encontro Nacional de Estudos Populacionais. Caxambu: ABEP; 2004. for the 2000 and 2010 Censuses. As expected, the 1991 results showed more significant variability than in 2000 and 2010. The breakdowns by federative units showed high variability, so we limited the aggregations to Brazil and Major Regions. The frequencies and respective ranges of all the variables in this article can be found in the doctoral thesis that originated them².

Ranking of access to basic sanitation and electricity services and infrastructure

Principal Component Analysis (PCA) was used to select and reduce variables and construct a summary indicator to measure Indigenous people’s access to services and infrastructure. This resource is commonly used in studies on inequality that involve a large number of variables²⁷27 Caldas ADR, Nobre AA, Brickley E, Alexander N, Werneck GL, Farias YN, Garcia Barreto Ferrão CT, Tavares FG, Pantoja LN, Duarte MCDL, Cardoso AM. How, what, and why: housing, water & sanitation and wealth patterns in a cross-sectional study of the Guarani Birth Cohort, the first Indigenous birth cohort in Brazil. Lancet Reg Health Am 2023; 21:100496.

28 Hongyu K, Sandanielo VLM, Junior GJ de O. Análise de componentes principais: resumo teórico, aplicação e interpretação. ES Eng Sci 2016; 5(1):83-90.^-2929 Szwarcwald CL, Leal MC, Castilho EA, Andrade CLT. Mortalidade infantil no Brasil: Belíndia ou Bulgária? Cad Saude Publica 1997; 13(3):503-516..

The PCA assisted in selecting, adapting, and reducing variables in this research. The variables initially observed were the total number of residents, the number of existing bathrooms, the sources of electricity and access to water used, and the destination of sewage and garbage from the households. These variables were adapted as follows: for each variable, the classification “yes” was assigned when the individual declared to be a resident of a household with the following characteristics: (a) with up to three residents; (b) with access to electricity supplied by an energy distribution company; (c) whose sewage disposal is the general sewage (or rainwater) network or septic tank; (d) with at least one bathroom, and; (e) whose water consumption comes from the general distribution network. Cases that did not meet these conditions were classified as “no”. The “yes” cases for each variable were totaled for the application of the method. Thus, the model variables were the percentage of people (Indigenous and non-Indigenous) in households with up to three residents; the percentage of Indigenous people with access to electricity from an energy distribution company; the percentage of Indigenous people who live in households whose sewage disposal is the general network or septic tank; percentage of Indigenous people; percentage of Indigenous people whose household water supply is provided by a supply network; percentage of Indigenous people with garbage collected by a specialized service; and percentage of Indigenous people who live in households with at least one bathroom.

The explanatory power found for the first principal component of the PCA, that is, how much the model could explain total data variance, was 83.1% and aligns with other studies that also used this instrument to build inequality indicators²⁷27 Caldas ADR, Nobre AA, Brickley E, Alexander N, Werneck GL, Farias YN, Garcia Barreto Ferrão CT, Tavares FG, Pantoja LN, Duarte MCDL, Cardoso AM. How, what, and why: housing, water & sanitation and wealth patterns in a cross-sectional study of the Guarani Birth Cohort, the first Indigenous birth cohort in Brazil. Lancet Reg Health Am 2023; 21:100496.

28 Hongyu K, Sandanielo VLM, Junior GJ de O. Análise de componentes principais: resumo teórico, aplicação e interpretação. ES Eng Sci 2016; 5(1):83-90.^-2929 Szwarcwald CL, Leal MC, Castilho EA, Andrade CLT. Mortalidade infantil no Brasil: Belíndia ou Bulgária? Cad Saude Publica 1997; 13(3):503-516.. Sixty values ​​were generated (referring to 3x5x2x2 strata), given by the combination of the household type (three strata, namely, homogeneous, mixed with Indigenous head or mixed with non-Indigenous head), Greater Region (five strata, namely, North, Northeast, Southeast, South, or Midwest), situation (two strata, namely, urban or rural), and location (two strata, namely, inside or outside Indigenous land).

The values ​​of each subgroup were sorted in descending order, where higher scores express more satisfactory access levels. The importance of using this data modeling resource lies in the possibility of measuring the inequality between different population groups, that is, from their relational perspective, instead of the very values ​​obtained²⁷27 Caldas ADR, Nobre AA, Brickley E, Alexander N, Werneck GL, Farias YN, Garcia Barreto Ferrão CT, Tavares FG, Pantoja LN, Duarte MCDL, Cardoso AM. How, what, and why: housing, water & sanitation and wealth patterns in a cross-sectional study of the Guarani Birth Cohort, the first Indigenous birth cohort in Brazil. Lancet Reg Health Am 2023; 21:100496.

28 Hongyu K, Sandanielo VLM, Junior GJ de O. Análise de componentes principais: resumo teórico, aplicação e interpretação. ES Eng Sci 2016; 5(1):83-90.^-2929 Szwarcwald CL, Leal MC, Castilho EA, Andrade CLT. Mortalidade infantil no Brasil: Belíndia ou Bulgária? Cad Saude Publica 1997; 13(3):503-516.. For microdata processing, we used the SAS Enterprise Guide 8.1 statistical package and, for multivariate statistical analyses, the FactoMiner library in the R Studio environment (version 3.6.3). The IBGE authorized our access to the universe’s microdata, respecting all limitations imposed by the Institute to preserve the statistical confidentiality of respondents³⁰30 Instituto Brasileiro de Geografia e Estatística (IBGE). Confidencialidade no IBGE: procedimentos adotados na preservação do sigilo das informações individuais nas divulgações de resultados das operações estatísticas. Rio de Janeiro: IBGE; 2018.. The demographic census sample microdata are publicly accessible and can be consulted on the IBGE website.

Results

The total number of people who declared themselves as Indigenous, captured by the question on skin color or ethnicity in the 1991, 2000, and 2010 Censuses, in permanent and improvised private households, was 288,101, 718,310 and 818,632 individuals, respectively, encompassing 98% of the total number of Indigenous people listed in the 1991 and 2000 Censuses, and 91% of the total observed in the 2010 Census. These numbers were more significant in rural areas in the 1991 and 2010 Censuses, corresponding to 76% and 61% of the total Indigenous declarations. In the 2000 Census, more than half (53%) of the Indigenous population was in an urban context (Table 1).

Aggregation by type of households indicated distinct patterns for the total number of Indigenous people in homogeneous and mixed households. The number of Indigenous declarations in homogeneous households increased at the same rate of 1.7 in 1991/2000 and 2000/2010. Regarding Indigenous people in mixed households with Indigenous and non-Indigenous heads, there was a considerable increase in 1991/2000, followed by a decrease in the following period (2000/2010) at less pronounced levels.

The main difference observed between the 2000 and 2010 Censuses occurred in the urban context, including in absolute terms, with a declining number of Indigenous people from 380,877 to 320,334. The highest concentrations of Indigenous people were found in the following areas: in households in rural areas in which all residents were declared Indigenous (homogeneous), mainly in the North, Midwest, and Northeast; in mixed households in urban areas in the Southeast and Northeast; and homogeneous households located in urban areas in the North. The numbers of Indigenous people living in mixed households with an Indigenous and non-Indigenous head were not very pronounced in rural contexts in all regions and with similar relative frequencies.

Sociodemographic profiles in households with declared Indigenous people

Graph 1 shows the structure of Indigenous people by gender and age in 1991, 2000, and 2010 by household type. In homogeneous households, a young population can be observed compared to the country’s general population, with a high proportion of children and young people. The age pyramids of this population indicate a continuous decrease in these younger groups and a progressive increase in the proportion of older age groups. No significant differences were found in the proportions of Indigenous men and women in homogeneous households. The estimated gender ratios were 108 men for every 100 women in the 1991 Census and 105 men for every 100 women in the 2000 and 2010 Censuses.

Mixed households had a lower proportion of Indigenous declarations for children and young people, suggesting a tendency for younger residents to declare other skin color or ethnicity options, mainly “White” and “Brown” and, in particular, in households whose household head was declared Indigenous.

This possibility is reinforced by the analysis of the household status variable, for which we found that most declarations for this item were people declared as head, spouse, sons, daughters, stepchildren, or stepdaughters of the head. The percentages of Indigenous people without a kinship relationship with the responsible person were residual for the three household types in all the aggregations analyzed and with high coefficients of variation.

In mixed households with an Indigenous head, most Indigenous people were identified as the head. In mixed households with a non-indigenous head, the Indigenous residents were mostly spouses, sons, daughters, stepchildren, or stepdaughters of the declared head. A higher proportion of Indigenous men than women was observed in households where the head was declared Indigenous; and a higher proportion of Indigenous women than men in households where the head was declared non-Indigenous.

The estimated gender ratios for Indigenous people in mixed households with an Indigenous head were 206, 163, and 117 men for every 100 Indigenous women and, for Indigenous people in mixed households with a non-Indigenous head, 47, 48, and 60 men for every 100 Indigenous women in the 1991, 2000 and 2010 Censuses, respectively.

Mixed households: size and composition by skin color or ethnicity

Regarding mixed households, two points can be highlighted regarding size. The first concerns the concentration of Indigenous people in households with four or more residents, in which other statements of skin color or ethnicity are predominant. The second refers to the tendency for the number of residents per household to decrease, also observed in the national setting in recent decades³¹31 Cavenaghi SM, Alves JED. Domicilios y familias en la experiencia censal del Brasil: cambios y propuesta para identificar arreglos familiares [Internet]. 2011. [acessado 2023 jun 26]. Disponível em: https://repositorio.cepal.org/handle/11362/12879
https://repositorio.cepal.org/handle/113... .

According to the 1991 Census, 73% of mixed households with an Indigenous head and 82% of households with a non-Indigenous head had only one declared Indigenous resident, of which 45% lived in homes with four or more residents. In the 2000 Census, the percentages of mixed households with only one Indigenous declaration were 72% in households with an Indigenous head and 77% in households with a non-Indigenous head. As in the previous census, households with four or more residents were the most representative, as were those with two or three residents. In the 2010 Census, 72% of mixed households with an Indigenous head had a declared Indigenous person; in households with a non-Indigenous head, the percentage was 75%. A more balanced distribution was observed between household sizes, albeit with a concentration of 4 or more residents (Table 2).

Households with an Indigenous household head had a higher number of Indigenous declarations from adulthood onwards (25 years in 1991 and 2000 and 35 years in 2010), with a predominance of other color or race options for children and young people. In households where the household head was not Indigenous, non-Indigenous declarations were more frequent in all age groups in the three censuses analyzed²2 Silva LO. Domicílios com indígenas nos censos demográficos de 1991, 2000 e 2010 no Brasil: composição e análises de inter-relações entre composição segundo cor ou raça e condições socioeconômicas e sanitárias [tese]. Rio de Janeiro: Fiocruz; 2021..

The primary skin color or ethnicity options used were “Indigenous”, “White”, and “brown”. In mixed households with an Indigenous head, the “Indigenous” option accounted for 36% of the total in the 1991 and 2000 Censuses and 39% in the 2010 Census. The second most used option was “Brown”, with 35% in the 1991 Census and 29% in the 2000 and 2010 Censuses. The “White” option, in turn, represented 25%, 28%, and 25%, respectively, as the third most frequently used skin color or ethnicity option in these households. In households with a non-Indigenous household head, the most frequently declared skin color or ethnicity options were: 27% “Indigenous”, 28% “White”, and 40% “Brown” in the 1991 Census; 30% “Indigenous” and “White” and 32% “Brown” in the 2000 Census; and 34% “Indigenous”, 31% “Brown”, and 26% “White” in the 2010 Census.

Access to sanitation and electricity services

The largest contingents of Indigenous people listed in 2010 were concentrated in rural areas within Indigenous lands (55%) and in urban areas outside Indigenous lands (33%). Indigenous people living in rural areas outside Indigenous lands represented 9% of the total, of which 79% were concentrated in the North and Northeast regions of the country. Individuals living in households located within Indigenous lands, in urban areas, accounted for 3% of the total Indigenous declarations. The housing conditions observed in these geographic areas revealed significant differences between the population groups analyzed.

Indigenous people living on Indigenous lands in rural areas had the lowest levels of overall access to services. These individuals lived mainly in homogeneous households in the North, Midwest, and Northeast. Almost all of this population did not report access to garbage collection (95%), did not have a bathroom in their homes (73%), consumed water from wells, springs, or lakes (70%), and access to the sewage system or septic tank was restricted to less than 12% of the total.

Indigenous people in rural areas outside Indigenous lands had housing conditions similar to those within Indigenous lands. This population was particularly significant in the North and Northeast, which accounted for 79% of the total Indigenous population in this spatial section.

Regarding Indigenous people living on Indigenous lands in urban areas, the numbers were significant for people in homogeneous households in the Northeast, 74% of the total. The conditions were more favorable in this section since 71% of the individuals declared access to the sewage system or the presence of a septic tank; 87% had a bathroom in the home; 79% were served by garbage collection, the same percentage as those with access to the water supply network; and practically all of these individuals had access to the energy distribution network (98%).

Indigenous people living in urban areas outside Indigenous lands were more numerous in the Northeast (32%), Southeast (26%) and North (20%). The South and Midwest each accounted for 11% of the total. In this context, 61% of Indigenous people lived in mixed households, 36% lived where the household head was Indigenous, and 25% lived where the household head was not Indigenous. The conditions observed indicated almost total access to the electricity grid, more than 80% of individuals had a bathroom in the home, and household garbage collection service was declared by 88% of these people.

Generally, the proportions of access to the services analyzed in homogeneous households were lower than in mixed households. The most significant difference was found in access to the sewage system or the presence of a septic tank, namely, 53% of access by the subgroup classified as homogeneous versus 66% and 67% of those identified as mixed with Indigenous and non-Indigenous heads, respectively.

Table 3 shows the values ​​generated for the summary indicator. The observed amplitude measures the inequality between the Indigenous subgroups in the spatial areas considered. The maximum value obtained was 6.2 times greater than the lowest estimated value relative to Indigenous people in homogeneous households in the North, rural areas, and within Indigenous lands.

In general, higher values ​​were observed for urban areas than rural areas. Regionally, the highest values ​​were for the Southeast, the South, Northeast, and Midwest, and the lowest for the North. In terms of location, Indigenous people living outside Indigenous lands presented higher values ​​than those living inside Indigenous lands.

The results of the types analyzed in each geographic section indicated the inequality by skin color or ethnicity between the groups. In this sense, Indigenous people in mixed households with a non-Indigenous household head generally had higher values ​​than those living in households with an Indigenous household head, and residents of homogeneous households had lower scores.

The analyses pointed to two distinct regional patterns, namely, the first, represented by the North and Midwest, is characterized by high Indigenous population contingents, mainly in rural situations, primarily within Indigenous lands, and lower scores; the second pattern predominates in the Southeast, South, and Northeast, with significant representation of individuals in urban situations, outside Indigenous lands, with higher scores, especially in the Southeast and South.

Discussion

The results of this research show the relevance of the proposed typology in characterizing the populations analyzed. We observed, among the types of households with Indigenous people, distinct profiles for each population subgroup, with differentiated and relatively consistent sociodemographic and spatial distribution patterns in the three censuses despite the fluctuations in the totals collected. The differences observed reflect, to a large extent, inequalities by situation and by Major Regions in agreement with those observed in the country²⁴24 Instituto Brasileiro de Geografia e Estatística (IBGE). Censo demográfico 2010: características da população e dos domicílios - resultados do universo [Internet]. 2010. [acessado 2023 jun 23]. Disponível em: https://biblioteca.ibge.gov.br/index.php/biblioteca-catalogo?view=detalhes&id=793
https://biblioteca.ibge.gov.br/index.php... , and regarding inequalities by skin color or ethnicity.

Several issues raised, such as the possible omission of declarations by Indigenous people, especially among children and young people, in mixed households located mainly in cities in the Northeast, Southeast, and North, may be further investigated as the complete results for Indigenous people from the 2022 Census are published. The first results for Indigenous people, published on August 7, 2023¹⁷17 Instituto Brasileiro de Geografia e Estatística (IBGE). Indígenas: primeiros resultados do universo [Internet]. 2022. [acessado 2023 ago 7]. Disponível em: https://www.ibge.gov.br/estatisticas/sociais/populacao/22827-censo-demografico-2022.html
https://www.ibge.gov.br/estatisticas/soc... , provide clues to some possibilities.

In this census, marked by a significant increase in the capture of Indigenous populations, the question of considering oneself Indigenous accounted for more than a quarter of all declarations registered outside Indigenous lands, highlighting that, in 2010, this question was only used within Indigenous lands. This fact, on the one hand, reiterates the relevance of the coverage question, together with skin color or ethnicity, for capturing these populations; on the other hand, it points to the need to investigate this “new” population contingent captured, visible in the official statistics in this most recent census, as is the example of Manaus, in Amazonas, with 4,040 Indigenous people declared in the 2010 Census, and 71,713 in the 2022 Census, of which 52,860 captured by the question “Do you consider yourself Indigenous?”¹⁷17 Instituto Brasileiro de Geografia e Estatística (IBGE). Indígenas: primeiros resultados do universo [Internet]. 2022. [acessado 2023 ago 7]. Disponível em: https://www.ibge.gov.br/estatisticas/sociais/populacao/22827-censo-demografico-2022.html
https://www.ibge.gov.br/estatisticas/soc... .

Also, from the perspective of the analytical-methodological possibilities opened by the 2022 Census, because the IBGE maintained the identification and ethnic-racial block questions in the universe questionnaire, keeping the location dimension, we will be able, for the first time, to compare the results of the universe of two censuses, expanding the possibilities of using the variables ethnicity and Indigenous language spoken at home.

The methodology used in this study allows the incorporation of these variables and others, such as those from the Indigenous approach questionnaire, another new feature of the 2022 Census, applied to political leaders of villages and communities before the start of the census, with questions about collective infrastructure, healthcare facilities, economy, and travel, among other topics, if this data is made public by the IBGE, obviously. The possibility of declaring more than one spouse within Indigenous lands and the existing kinship relationship for the composition of family units also represents another fertile ground for demographers of Indigenous themes and others interested in the topic.

However, the most significant novelty of the 2022 Census was the unprecedented investigation of quilombola populations. This new dataset also opens up possibilities for studies, including in its interface with Indigenous populations, since dual quilombola and Indigenous belonging was possible, a relevant fact in specific locations, especially in some states in the Northeast.

The prospect of exploring the ethnic-racial identification criteria in conjunction with other questionnaire variables is a viable option for constructing new data arrangements that allow the investigation of specific population groups. However, it is essential to remember that these new analytical perspectives need to consider the limits of census data, including the suitability of the selected variables to the reality of Indigenous populations¹1 Campos MB, Estanislau BR. Demografia dos povos indígenas: os censos demográficos como ponto de vista. Rev Bras Estud Popul 2016; 33(2):441-449., the disaggregation’s statistical robustness, and the impossibility of the demographic census to capture certain events with precision.

This last point was verified regarding the classification used in this study on the mortality block data from the 2010 Census, applied to households with an Indigenous declaration³²32 Okamoto da Silva L, Borges GM. Mortalidade de indígenas no Censo Demográfico 2010: análise segundo a composição por cor ou raça nos domicílios com indígenas declarados - produto para o projeto "Situação de saúde indígena no Brasil e seus determinantes sociais: abordagem das iniquidades étnico-raciais a partir de diferentes bases de dados secundários sobre demografia, políticas sociais e saúde de amplitude nacional" [relatório de pesquisa]. Rio de Janeiro: Fiotec; 2021.. Although the results obtained suggest worse indicators for Indigenous people, some analyses based on indirect calculations (the surviving children and maternal orphanhood method) were inconclusive, probably due to the high omission of deaths among Indigenous people, highlighting that the omission of deaths in demographic censuses as a whole is recognized by experts¹³13 Instituto Brasileiro de Geografia e Estatistica (IBGE). Tábua completa de mortalidade para o Brasil - 2015: breve análise da evolução da mortalidade no Brasil. Rio de Janeiro: IBGE; 2015.^,3333 Borges GM, Okamoto da Silva L. Situação de saúde indígena no Brasil e seus determinantes sociais: abordagem das iniquidades étnico-raciais a partir de diferentes bases de dados secundários sobre demografia, políticas sociais e saúde de amplitude nacional - produto para o projeto "Situação de saúde indígena no Brasil e seus determinantes sociais: abordagem das iniquidades étnico-raciais a partir de diferentes bases de dados secundários sobre demografia, políticas sociais e saúde de amplitude nacional" [relatório de pesquisa]. Rio de Janeiro: Fiotec; 2021. or is due to limitations of the data collected to estimate the indicators.

Classifying is, in essence, reducing and simplifying. However, despite this inevitable “flattening” of the realities captured in the research⁶6 Santos RV, Guimarães BN, Simoni AT. Cor ou raça: indígena? Contextos e recepções da inclusão de uma categoria no Censo Demográfico 1991. Rev Fr-Bras Geogr 2023; 59. DOI: 10.4000/confins.51565
https://doi.org/10.4000/confins.51565... , its historical-political implications, and the limitations of the capture criteria themselves, the relevance of quantitative analyses on minority groups for the production and implementation of appropriate public policies and compliance with existing legal precepts is undeniable. In this sense, searching for new ways of analyzing data to portray better the characteristics of ethnically differentiated populations, such as Indigenous populations, is necessary to deepen knowledge about social and health inequalities, a fundamental dimension for implementing and monitoring public policies.

References

Publication Dates

History