Poverty again at the Center of Debate

Atila P. Roque; Joao Sucupira; Sonia Correa; Jorge Eduardo Durao

With the election of Fernando Henrique Cardoso in 1994 (who set up a social–democrat administration supported by a conservative coalition), Brazil would seem to have definitely entered the "adjustment era". (1) This experience of "late adjustment" has implied that the debate on the impact of economic restructuring and state reform has not taken place in Brazilian society as large, particularly regarding the exacerbation of social exclusion.

On the other hand, the Brazilian adjustment is taking place at a time when criticism of adjustment is being heard worldwide and possible alternatives are being formulated. The Declaration and the Action Program of the World Summit on Social Development (WSSD), for example, contain various indications of this direction.

These new circumstances may favor an alternative agenda that has as its main focus the eradication of poverty and the prioritization of social policies. Thus, the political struggle over the meaning of development may become even clearer.

This has already happened in the parliamentary debates on the constitutional reforms, in the media, in academic circles and in Brazilian society itself. The "Citizens’ Action against Hunger" is, perhaps, the most significant expression of this dispute.

Led by Herbet de Souza –known as "Betinho"– the "Action" –which is better known as the Hunger Campaign– introduced and made visible an agenda for eradication of poverty in an extremely unfavorable context. The campaign, began in 1993, however has its origins in the mobilization for ethics in politics that came out of the impeachment of ex–president Fernando Collor de Mello, in 1992. The NGOs also played their part during their 2nd Assembly of the Brazilian Association of NGOs (October 1992) which stressed the urgency of channeling citizens’ actions in the area of ethics in politics over the battleground of social exclusion.

The characteristic feature of Citizens’ Action is essentially the political treatment of poverty and hunger. Its greatest contribution has been to call attention to the problems of poverty and make these issues visible on the national political agenda. Brazilian society had, somehow, avoided the subject throughout the 1980s, and at the level of day–to–day perceptions and practices poverty had become a way of lifeand accepted. However, by 1993 there has been a marked change in perceptions and public discourse on the povery issue in Brazil. All of the party programs began emphasizing the question of inequality in reparation for 1994 presidential campaign. The elected president’s government program –which used as a reference the five fingers of the hand– gave priority to: agriculture, employment, health, education and security. In civil society, thousands of solidarity and social promotion initiatives began during this period.

Two years later poverty remains high on the agenda of Brazilian citizens. In January 1996, the most common answer to the question "What do you think of first when you think of Brazil?" was: Hunger and Poverty (12%). Yet this visibility and legitimacy have not carried over, into each of the fields of national public policies.

This indefinition, obviously, impacts on the implementation of the resolutions of the World Summit on Social Development, approved in March 1995 in Copenhagen.

With UNCED in 1992, the Brazilian government politically invested in of what is called the UN Social Cycle of conferences– Vienna 1993, ICPD 1994, WSSD1995 and Beijing 1995.(2) As of preparation for ICPD (Cairo 1994) the Brazilian position in negotiations was built on dialogue with academic sectors and civil society, and NGOs have been included on official delegations. Similar procedures were adopted in the preparation for WSSD. Meanwhile, taking into account the UN Social Cycle as a whole, Brazil’s action profile was more discrete in Copenhagen and its action was weak on subjects of an economic nature that could have mobilized the energies of various developing countries. Domestically, effective application of the Copenhagen resolutions faces political obstacles, reflected in fights over macroeconomic definitions and the budget allocations of the federal government. Such impasses will be further analyzed in other sections of this document.

Social Indicators in Brazil in the 1990´s

An accurate analysis of the incidence of poverty in Brazil encounters two limitations. The first is the reliability of data and statistical information. The national system of social statistics is in a very serious crisis. For example, the 1990 Census was completed only in 1991, and as of March 1996, its results have not been fully tabulated.

Fortunately, needed change may finally be implemented in the national statistics system. The Brazilian Institute of Geography and Statistics (IBGE) is once again gaining support and visibility and there have been innumerable sectorial initiatives –at ministerial level – to improve relevant data for definition of social policies. Another sign of these changes is the initiative to publish a Brazilian Report on Human development. The UNDP–supported project is being executed by IPEA –Institute for applied Social Research– and has the collaboration of foreign consultants linked to academia and to the nongovernmental community. Publication is scheduled for the first half of 1996.

The second problem relates to the country’s enormous heterogeneity. Social indicators vary enormously according to region, and the variations must also be taken into account in terms of income, race and gender. Given the current precariousness of the data bases available, these disaggregation is not always available. There are very few indicators disaggregated by race, and those disaggregated by gender leave much to be desired. The same limitation applies to conventional indicators, such as infant mortality (since there are no data bases disaggregated by municipalities.) (3)

The statistical data presented here are organized in three blocks: 1) general demographic data (the section includes brief considerations on demographic transition); 2) sensitive indicators in terms of measuring inequality, including information relative to gender and race; 3) economic and social data specifically related to poverty. (4)

Poverty Indicators



In reporting on Brazilian poverty indicators, some preliminary factors need to be considered when seeking to report on poverty indicators in Brazil. The numbers vary considerably according to the sources and criteria used. Because of this, at present there is great debate as to the rate of poverty in Brazil, as well as to the methodologies used to measure it. There is also great heterogeneity in terms of the regional distribution and the nature of poverty throughout the country.

International Reports on Human development, for example, refer to the existence of 72 million poor in Brazil. The study done by IPEA in preparing the Hunger Map, in 1993, identified 32 million indigent. Given its close association with Citizens’ Action, this number focused the hearts and minds of the public. In a recent study, however, Rocha (1995) established criteria for differentiating poverty from indigence (absolute poverty) and came up with different results. The parameter used to identify the pattern of indigence takes into account the cost of food necessities. The criteria for measuring poverty takes into account the cost of basic necessities.

Based on these references the author concludes that in Brazil there are 42 million poor (30% of the population) and 16 million indigent (12% of the population). The data in the Brazilian National Report for the Social Development Summit, in turn, indicate that between 1980 and 1990 the total number of poor grew from 29.4 million to 39.2 million. In relative terms, however, the percentage of poor or extremely poor dropped from 34% to 27% of the population contingent.

However, over recent decades the proportion of poor has been falling in Brazil. On the other hand, there is no doubt that in Brazil inequality and the pattern of concentration of wealth have become more acute. While in 1960 the 10% richest held 34 times the income of the poorest 10%, in 1990 they held 78 times the income obtained by the most destitute sectors. Throughout the 1980s concentration affected above all a specific population group at the low–end of income distribution.

Considering the distribution of national income by strata: in 1990, 13.9% was in the hands of 1% of the population; the richest 10% held 45%; the poorest 50% controlled only 1.1% of the national income, and the poorest 10% accounted for barely 0.8%.

In terms of heterogeneity we should underscore the regional variations and the urban–rural differential. The incidence of poverty is higher in the North and Northeast, in fairly direct relation to the regional patterns of economic development. In 1992, the Southeast Region contributed 56.1% of national GDP; the South, 16.7%; the Northeast, 15.8%; and the Center–West and North, 5.6% and 5.5%, respectively. When compared to demographic distribution the disparity is even more evident. The Northeast, for example, holds 27% of the Brazilian population and contributes only 15.8% to the national GDP.

The incidence of poverty is higher in rural areas. Nevertheless there has been a rapidly growing "metropolization" of poverty. Concentration of the Brazilian population means in absolute terms the poor are concentrated in the more developed regions. The metropolitan poor accounted for 26% of the total in 1981 and 29% in 1990. More than 50% of metropolitan poor are concentrated in the cities of Rio de Janeiro and Sao Paulo, with the latter having the highest number of poor in the country (5.1 million). In the other states there is a relative equilibrium in the proportion of poor in metropolitan regions and rural areas. In proportional terms, the metropolitan regions of the Northeast have the highest number of poor (43% as compared to 22% in Sao Paulo).

Given the level of urbanization and industrialization of the country it is also fundamental to assess the relationship between poverty and job market. The data available indicates that 52% of Brazilian workers receive less than two minimum salaries; in the last seven years the share of remuneration for work in national income did not exceed 30%; in the 1980s, real average wages dropped 14%.

In a paper investigating the correlation between poverty and employment in Brazil, Camargo et al. (1995) considered four variables: a) family size; b) rate of participation of family members in the work force; c) quality of jobs; d) worker skill levels. Significantly, the results indicate that worker skills, as well as job quality are the factors having the greatest weight in the production of poverty, as compared to the first two variables.

The open unemployment rate is not high in Brazil (4.1% in 1991), particularly if we take as a reference the European parameters. Nevertheless there is a high degree of informal employment (30 million workers in the informal segment). Worker turnover rates in the formal segment of the job market are also extremely high. Additionally, following implementation of the stabilization plan –particularly during 1995 – the number of jobs dropped precipitously, especially in the most industrialized areas. During the first week of January 1996 Sao Paulo recorded 8% open unemployment.

Finally we should consider the impact of inflation on the production of poverty and inequality in Brazil. The data available show that, in fact, the years with high inflation (1985–1994) brought a brutal wage erosion and deepening of the concentration of income in Brazil. Currently in official discourse and in the media there is a prevalence of arguments affirming that the reduction of inflation rates is in itself an instrument for reducing poverty. There is talk, for example, of a 30% increase in real income in the poorest segments.

Various authors, however, have also shown that it is not enough to consider stability and inflation reduction as exclusive measures for reduction of the persistent pattern of inequality. They emphasize the need to consider other variables and instruments, such as ensuring economic growth rates, whose indices were extremely unstable over the last 15 years and remain affected by the policy of high interest rates. But they especially underscore the need to create institutional mechanisms for reducing economic and social disparity via indirect income policies.

Sensitive Indicators

Life Expectancy

  • Evolution: 1980 – 61 years / 1990 – 65 years
  • Regional Variation: While in the Northeast life expectancy is 64 years, in the Southern region it is 68.
  • Variation by Income Level: If we consider income levels the variation is even more striking. Among those who earn up to 1 Minimum Salary (MS = US$ 100) life expectancy is 57 years. Those earning more than 5 MS live 73 years.

Infant Mortality

  • Evolution: 1980 – 65.8/1000 / 1990 – 51.6/1000
  • Regional Variation: South – 26.7/1000, Northeast – 88.2/1000.
  • Variation by Income Level: Among families earning up to 1 MS infant mortality is 75.2/1000. Among families with incomes of more than 1 MS the rate drops to 33.3/1000.

Child Labor

In Brazil 16.9% of children between 10 and 14 years of age and 50% of adolescents between 14 and 17 work. In absolute terms this corresponds to 3 million children and 4.6 million adolescents who work. Child labor is directly related to family income. In 1990 the activity rate for children 10 to 14 years of age living in extremely poor families (up to 1/2 MS) was 23%. Among children living in families with income over 2 MS the rate drops to 4.5%. The data also indicate a high percentage of working children in cases of poor households headed by women.


All the studies available refer to an improvement in formal education over the last 20 years. This evolution includes high enrollment rates for girls/women. Today, in some age segments and regions women receive more schooling than men. Brazil, however, still has 20.2 million illiterate persons 10 years of age or older. Data recently made public by the Ministry of Education report that barely 60% of children entering school reach the fourth grade. Those who finish fourth grade take an average of 6 yearsto complete their schooling. Only 43% finish eighth grade. Cohn W(1995) also indicates that there was a slowing of illiteracy reduction rates over the last 4 decades: the reduction in the number of illiterate in 1950 was 12.3%; in 1960 it was 7.3%; in 1970, 6.6%; and in 1980, barely 5%. This means that despite the expansion of the basic teaching network, the educational system has not be able to reduce illiteracy rates, while it continues to produce attrition, and continues to have serious problemsin terms of quality.


The chief cause of general mortality in Brazil are circulatory system diseases (30%). Next are external causes (especially traffic accidents); followed by unidentified causes and cancer. The incidence of tuberculosis is 57.8/100,000. The HIV/AIDS epidemic affects 369 persons out of 100,000. Pockets of malaria continue to exist and affect 6.2 inhabitants/100,000. The three main causes for hospitalization in the country in 1994 were: obstetric (childbirth and abortion); psychiatric pathologies; and cardiac related disease.

At the end of the 1980s Brazil had 7,200 hospitals, most of them private, and 27,700 ambulatory services and clinics where the public sector presence is dominant. Recent studies show an expansion of the private sector, a trend negatively affecting access to services by the poorest of the population.

In the 1980s there was an expansion of the hospital network, but a decline in the number of beds per capita (from 4.3 to 3.7/1000). The doctor/inhabitant ratio is 1/641, while the proportion is much greater in the North and Northeast regions (1/1174 and 1/1070, respectively).


Poverty indicators disaggregated by gender are not available, since measurements are traditionally made on a household basis. It is possible, however, to measure the impact of gender inequalities in terms of income and political and social training through other indicators.

Wage inequality: Women on the average receive 54.1% of men’s salaries. In 1990, 61% of employed women received up to two minimum salaries.

Women heads of households: In 1990 in Brazil 7.3 million women heads of households were identified. With a conservative estimate of four persons per household, this implies the existence of approximately 30 million persons (1/5 of the population) living in households headed by women. 30% (approximately 2 million) of these women have no income. 60% have more than two children earning up to two minimum salaries. This information suggests that families headed by women are a major pocket of absolute poverty in Brazilian society. The Brazilian national Report for WSSD confirms the large number of women heads of household being poor. Whereas only 15% of couples under 34 years of age are poor, 35% of households headed by women in the same age segment are below the poverty line.

Maternal mortality: Estimates for the maternal mortality rates in Brazil vary between 100 and 200 deaths per 100,000 live births. 1994 hospitalization records indicate that in that year 1,041 women died at Brazilian hospital units during childbirth. According to IBGE data for 1992, 1,542 deaths due to abortion were recorded at public and private hospitals. In other words that adds up to a total number of some 2,500 hospital deaths. To this figure should be added the deaths outside the health network, as well as deaths where records do not clearly indicate an obstetric cause.

The causes for maternal mortality are associated with the same factors that explain the high infant mortality rates (poverty, malnutrition, lack of access to health services). They also constitute a fundamental reference for measuring gender inequality. Meanwhile, this data has not been given privileged status as sensitive indicators of social inequality.


While average gross income for the white population is 5.3 minimum salaries, income for the black or mestizo population is only 2.5 minimum salaries (a disparity of 50%). In analyzing the income situation of informal sector workers, Abreu et al. (1995) report that black and mestizo men workers receive an average of 3.7 minimum salaries, compared to 5.7 salaries received by white men workers. Black or mestizo women workers receive an average of 1.9 minimum salaries compared to 2.7 for white women workers.

Social Spending in the Federal Budget

Brazil’s 1996 proposed budget, now before the National Congress, is in the area of R$ 312.8 billion, slightly more than half the country’s GDP. At first, this may seem a relatively significant amount. However, after deducting from the overall budget the payments of debt principal (R$ 113.8 billion) (which are entirely financed by placement of new debt instruments), and interest and charges on debt (R$ 22.4 billion), what is left is R$ 176.6 billion to be allocated among the diverse State functions (education, health, justice ...). Of this last total, amounts allocated to federal government social programs account for almost 60%. Even considering the total amount of the budget, (including principal and interest, the share of social spending is still significant (33.5%).

If social spending is considered to include all expenditures covering costs and investments in social security and assistance, health, education and culture, labor, housing, sanitation and environmental protection, the data show a significant increase in social spending in the 1993–96 period (the provision for 1996 is 70% higher than for 1993). Consider the Table 1.

Social Public Spending by the Federal Government

Brazil: 1993-1996
  In R$ million at average 1996 prices
Social Spending 1993 executed 1994 executed 1995 budget 1996* budget
Social Security 38,748 47,041 55,263 60,953
Health 9,107 16,396 17,955 19,388
Medical/Health Care 7,429 12,234 11,907 12,794
Food & Nutrition 426 745 1,271 1,124
Education & Culture 6,973 10,369 11,O6O 10,039
Higher Education 3,374 5,115 5,14O 4,76O
Basic Education 1,888 2,853 2,376 2,569
Labor 3,761 3,126 9,587 9,505
Housing 159 54 1,877 2,098
Social Assistance 2,212 1,024 1,379 1,919
Sanitation 713 326 739 851
Environmental Protection 25 45 247 96
TOTAL 61,716 78,381 98,107 104,849

Source: Technical Note N¦ 04/95 prepared by AOFF/CD and CONORC/SF.

* Bill N¦ 35 for 1995, National Congress.

Given that social security, health, and education and culture account for almost 90% of 1996 budget resources allocated to social programs of the federal government, this article will concentrate on these three macro sectors.

Social Security: Social Security allocations in the 1996 budget represent 58.2% of social spending. Although significant, compared with the others, the allocation is dropping as a share of total social spending (62.7% in 1993). This figure becomes more significant when we take into account that as of the 1988 Constitution there has been a trend towards structural increase of spending with benefits based on the increase in the universe of beneficiaries and new rules for maintaining and updating the value of thesebenefits. Therefore, if other social programs in education and health had not been given relatively more resources, the tendency would be precisely an increase in the share of social security spending.

The budget situation for social security is worrisome and must be resolved structurally. There are those who defend the idea that equilibrium in social security in Brazil depends less on spending cuts than on structural factors, such as increasing the per capita average salary and employment. Since discussions on the budget reform will continue over this year, the effects on governmental accounts will only be perceived as of 1997.

Health: Of all the sectors, health is perhaps the sector having the most financing problems. The 1988 Constitution significantly increased (more than 40% of the Brazilian population) the number of persons entitled to benefits from the health system. In Brazil it is common knowledge that "public health program spending is scant and misspending is rampant." For example, in 1992 health spending was a little over half 1989 spending. As of 1993 spending in this sector increased, but judging by the chronic problems it would seem that the theory of misspent resources is indeed a fact.

For 1996 the budget (R$ 19.4 billion) provides for an 8% increase over 1995, and 18% over 1994. However, for health to be given more resources the Minister in this area must stuggle for linked budgetary items (creation of a tax on financial transactions –CMF– still not approved by the National Congress), when, given the extremely poor conditions of hospitals and equipment in general, not to speak of personnel remuneration, resources should be allocated without conditioning them to approval of the CMF. In that way, since expected collections from the CMF are estimated at R$ 6.0 billion, if this new source is not approved the expected amount drops to R$13.4 billion, 25% less than that provided in 1995 (an 18% decrease compared to 1994).

Education: Most federal spending on education goes to universities, since primary and secondary school education fall within the competency of municipal and state spheres. Considered one of the five priorities from the time of the president’s election campaign, this sector was not provided with budgetary resources beyond the level already in place. In reality, the amount foreseen for 1996 for Education and Culture is below the amount for 1994 (a real decrease of 3%). The Higher Education program, for example, whose actual expenses in 1994 were R$ 5.1 billion, had its budget allocation cut to R$ 4.8 billion. The allocation for spending on Basic Education is slightly higher than for 1995, but even so it is less than for 1994 (a real drop of 10%).

The data indicated on the earlier table show that with the exception of the Environmental Protection and Labor items, the others show a trend toward increased spending. Considering that Federal revenues grew substantially over the last year, it should be noted that education was not privileged in the distribution of resources in the 1996 budget.

In summary, the data set forth above showing the high share of social spending in total budget resources, as well as the growth in certain sectors, are highly relevant, since they underscore two fundamental aspects. The first is that the recommendation to allocate a minimum of 20% of the federal budget to social spending, as suggested recently at the world conference on Social Development in Copenhagen, with a view to reducing poverty, lost its appeal in the case of Brazil. In fact, if the reduction of social inequalities were to depend only on that criterion, Brazil would have been improving its social indicators for quite some time now.

The second aspect relates to the idea that it is not enough merely to increase the quantity of resources, as shown in the data on the table, but chiefly to look after the quality of spending. As indicated by the economist André Cesar Medici ("Os Gastos Públicos Federais com as Políticas Sociais", published by ABONG in September 1995), "One of the big deficiencies of the Brazilian State’s management mechanisms is the lack of mechanisms to evaluate the results of public spending". It is, thus, fundamental for State reform to include an evaluation of public spending.

A Brief Note on the "Community Solidarity Program"

The emphasis given to the social dimension in presidential discourse is not matched in the sphere of government policies. The simple fact that the government is still discussing a general strategy for state action in the social area, after a year in office, reveals the marginal attention given thus far to social policies in the Cardoso government.

The Community Solidarity program, geared to playing a uniting and coordinating role for the various government initiatives in the social area, has been fulfilling that role in a limited way and with great difficulties. Based on a diagnosis indicating that the problem with social policies in Brazil is not so much the insufficiency of public spending in social areas as it is the poor quality of that spending, the program was established as a mechanism to increase the efficiency and impact of social spending geared to fighting poverty.

This objective, in practice, has been stalled by the political choices of the government which, in making economic stabilization its major or sole objective, adjusted its accounts in a way that basically sacrificed social policy spending. The budget difficulties faced by the health sector are a dramatic example of this fact.

While we do not fail to recognize the positive aspects of Community Solidarity, especially the emphasis on the partnership with civil society, its development is still heavily conditioned by the larger objectives of the stabilization policy. Getting beyond this contradiction is the greatest challenge now facedby the government of Fernando Henrique Cardoso.


COHN, Amelia. Relatorio Nacional Brasileiro Para a Cupula Mundial de Desenvolvimento Social. Ministerio das Relacoes Exteriores, 1995.

ROCHA , Sonia. Diferencas locacionais da pobreza e os seus rebatimentos para a politica social. Documento de subsidios para o Relatorio Brasileiro de Desenvolvimento Humano, 1995. (mimeo)

CAMARGO, J. M.; BARROS, Ricardo Paes; e MENDONCA, Rosane.

Determinantes da Pobreza no Brasil. Documento de subsidios para o Relatorio Brasileiro de Desenvolvimento Humano, 1995 (mimeo).

ABREU, Alicia de Paiva; JORGE, Angela Figueira; SORJ, Bila. Desigualdade de Genero e Raca: o informal no Brasil. Revista Estudos Feministas, Numero Especial, CIEC/ECO/UFRJ, 2o semestre de 1994.

MEDICI, Andre Cesar. Os gastos publicos federais com as politicas sociais. Cadernos ABONG, No10, Associacao Brasileira de ONGS, setembro de 1995.


1 At the level of society and of parliament the prevailing reaction to the measures proclaimed by the Washington Consensus reflected, over more than ten years, an impasse, or a standoff of political forces. The gamut of actors and interests involved in this anti-adjustment alliance is extremely complex and requires a more in-depth analysis, which goes beyond the scope of this article.

2 This strategy in international politics reflects a new logic in foreign policy –called stellar by the new Minister of Foreign Affairs – and a greater opening up of the Ministry of Foreign Affairs to civil society.

3 To get beyond this limitation, in 1994 UNICEF produced an analysis of infant mortality incidence (on a municipal basis) based on approximate indicators.

4 The two sources used predominantly were the aforecited Brazilian National Report for WSSD and the connected reports produced for preparation of the Brazilian Report on Human Development. Additional