Reaching the Hard to Reach: A Case Study of Brazil’s Bolsa Família Program

Reaching the Hard to Reach: Foreword

It has been said that development is about delivery: the will and ability to deliver interventions to very poor people in order to improve their lives. The development “space” is filled with great ideas and innovative solutions, from technological interventions to new policy initiatives. But the effects of these potentially game-changing ideas are severely mitigated if they do not actually get to the people they are intended to benefit. We think of this challenge in terms of “reach.” Solutions can only solve problems if they reach those who need them most.

Those who live at the base of the pyramid – the poorest of the poor – are also the hardest to reach. Many are homeless or live far in the countryside; they might lack formal identification; or they are socially marginalized because of their ethnicity or gender. As Anthony Lake of UNICEF puts it: “Disaggregate the data and we find that our statistical national successes are masking moral and practical failures. People are left behind simply because they live in rural communities or urban slums, in conflict zones, as part of indigenous groups, with disabilities or because they are girls.”

The Bolsa Familia program in Brazil is an excellent example of an innovative social policy program that successfully reaches the hardest to reach. Bolsa Familia is a conditional cash transfer scheme that targets Brazil’s poorest families. In addition to reducing poverty and improving health and education outcomes among those in the bottom income quintile, the Bolsa Familia program is noted for its precision, in not only identifying and finding the poorest of the poor, the socially excluded and the geographically distant, but also in delivering the cash benefit to these families. This report explains how the Bolsa Familia has achieved this. Our findings point to innovative policy design, the appropriate use of technologies, sheer legwork in the villages and urban favelas, and political attention paid to the challenges and imperatives of reach.

The following report on the Bolsa Família program is based on a year-long research project, carried out by Professor Joseph Wong of the Munk School of Global Affairs and Department of Political Science at the University of Toronto, and five research assistants: Nina Da Nóbrega Garcia, Nicoli Dos Santos, Sarah Ray, Ariel Sim and Alexa Waud. Funding for the project was provided by the Ralph and Roz Halbert Professorship
in Innovation at the Munk School of Global Affairs; the Canada Research Chairs program; and the MasterCard Center for Inclusive Growth.

The authors of this report would like to express our gratitude and appreciation to those we met and interviewed in Brazil, specifically those in the Ministry of Social Development and the Institute for Applied Economic Research in Brasília, and the inspiring people who are making Bolsa Familia reach in Salvador, Dias d’Ávila, Sapeaçu and Belo Horizonte.

This project is the first of several more focusing on the concept of development reach, and specifically innovative programs that are reaching the hardest to reach populations.

Introduction

Reach

Brazil is a highly unequal society. Despite having experienced years of robust economic growth, levels of social and economic inequality remain very high. Brazil’s Gini Coefficient ranks it among the most unequal societies in the world. In 2011, for instance, still more than 3% of the population lived under conditions of “extreme poverty,” earning a monthly income of less than R$70 (approximately US$18 in 2015). Meanwhile, almost half of the population – 49% – lived in “vulnerable situations,” earning a monthly per capita income of between R$149 to R$560 (US$39 to US$147). Poverty in Brazil is chronic, described by policymakers as “intergenerational poverty” – cyclical poverty in which the very poor tend to stay very poor.

The Bolsa Família program (BFP) was created in 2003 by President Lula to end persistent poverty. The BFP resulted from the consolidation of four pre-existing conditional cash transfer programs, such as Bolsa Escola, an education conditional cash transfer scheme. From 2003 onwards, the BFP has continually expanded its coverage, reaching nearly 13.8 million families by 2015. As of July 2015, the BFP covered 48 million people, or about one-quarter of Brazil’s population.

The program is managed by the central government in Brasília, though it is implemented and operated by Brazil’s 5570 municipalities. The average monthly benefit paid to a BFP family is R$163 (US$43 in 2015 dollars). The cost to operate the BFP equals just 0.5% of GDP, making it an inexpensive program to run, especially considering its coverage and reach.

What stands out about the BFP is that it is able to reach the poorest of the poor in Brazil. According to a 2007 World Bank report, in 2004, nearly three-quarters of cash transfers from the BFP reached those families whose income was in the bottom quintile; Brazilian officials estimate that upgrades to the online registration processes, regulations and more accurate data collection have improved the BFP’s targeting record. Eighty percent of the remaining 25% (or four-fifths) were transferred to the next lowest income quintile, or the “near poor.” This means that “fiscal leakage,” a common problem where corruption and rentextraction are prevalent, such as in Brazil, is not a severe problem in the delivery of Bolsa Família. Compared with other Latin American countries’ cash transfer programs, the BFP’s performance in actually reaching the poor stands out by a considerable margin.

The BFP’s reach is impressive. When it comes to social policy, reaching the very poor is very difficult to do. Reaching the hardest to reach involves identifying the very poor, finding them and, finally, delivering social services to them. But consider the challenges: those living under conditions of poverty tend to be informally employed; they likely live in informal housing without a permanent or fixed address; and they are often socially excluded if, for instance, they are from one of Brazil’s minority “traditional groups.” The very poor in rural parts of the country live far from town centers. People living in slums, especially precariously employed workers such as migrant workers, are constantly on the move. River dwellers living along the Amazon are effectively “off the grid.” And many poor people are not even aware of government-run social programs. In other words, identifying, finding and delivering social services pose daunting technical, administrative and political challenges. And yet the BFP has successfully reached the hardest to reach in Brazil.

What is Bolsa Família?

Eligible families that receive BFP benefits are those whose monthly per capita incomes are equal to or less than R$77 (US$20). Families that earn monthly per capita incomes between R$77 and R$154 are also eligible if they have children younger than 18 years old.

Qualifying families are identified in the Cadastro Único (CadÚnico), a unified registry of all low-income households in Brazil. The CadÚnico is managed by the central government, and currently has household data on more than 27 million families; of these, 13.8 million families are enrolled in the BFP.

Cash benefits are paid directly to each family through an electronic payment system. The amount paid each month varies, depending on several factors. The basic monthly benefit is R$77. Monthly child benefits of R$35 are added for each child 0 to 15 years old, as well as pregnant and lactating women (up to five beneficiaries). An additional benefit of R$42 per month is included for youths aged 15 to 17 (up to two youths for each family). And last, a superation top-up is calculated for families living in extreme poverty.

The BFP is a conditional cash transfer program. In order to withdraw their monthly cash transfers, beneficiaries must demonstrate that they (and their family members) have met the program’s health and education conditionalities. Health conditionalities require that children between 0 and 6 years old be vaccinated and attend regular health check-ups; and that pregnant and lactating women attend pre- and post-natal check-ups and participate in health promotion activities offered by local health units. Education conditionalities stipulate that children aged 6 to 15 must maintain an 85% attendance record at school, and youths aged 16 to 17 must achieve a 75% record. Municipal operators of the BFP monitor BFP families to record whether conditionalities have been met.

Impact

The BFP has had a positive impact on poverty reduction and human development in Brazil. Since the introduction of the BFP in October 2003, poverty incidence has declined continuously. Health outcomes, measured in terms of under-5 child mortality and nutrition, have improved among BFP families. Education outcomes are similarly impressive. For example, children of families that receive BFP benefits are less likely to drop-out of high school and thus more likely to graduate than children from non-BFP families. Empirical research on the effects of the BFP on employment show that BFP recipients are not incentivized to leave the labor market; rather than making poor people “lazy,” BFP recipients are likely to continue to work, to look for work, work as farmers or small-scale entrepreneurs.

Research

The research presented in this report was funded by the Canada Research Chairs program; the Ralph and Roz Halbert Professorship of Innovation at the Munk School of Global Affairs; and the MasterCard Center for Inclusive Growth. The research was conducted by Professor Joseph Wong; three undergraduate students, Nicoli Dos Santos, Sarah Ray and Alexa Waud; one graduate student, Ariel Sim; and a recent graduate of the Masters of Global Affairs program at the Munk School, Nina Da Nóbrega Garcia. Research was conducted throughout 2015, including primary fieldwork in Brazil during December 2015. Interviews and site visits were conducted in Brasília, Salvador, Dias d’Ávila, Sapeaçu and Belo Horizonte.

The research was vetted by the Ethics Review Board of the University of Toronto.

The report is organized into five sections. The key theme that runs through all five sections is reach: specifically about how the Bolsa Família program is so successful in delivering cash transfer benefits to Brazil’s poorest of the poor and most difficult to reach populations.

Section 2 describes the Cadastro Único and how the unified registry gathers information that provides both big data and precise data with which the BFP can effectively identify, find and deliver benefits to the most difficult to reach populations. Section 3 focuses on the cash transfer system, and specifically how BFP benefits are transferred from the central government directly to individual households. The fourth section describes how the BFP is implemented and operated by Brazil’s 5570 different municipalities. The program, though it is centrally managed by the Ministry of Social Development, is in fact decentralized in its operation, allowing considerable latitude for municipal innovation in implementation. Section 5 highlights the impact of health and education conditionalities on human development outcomes for those who receive BFP benefits. And last, Section 6 describes how the BFP’s success hinges in part on local, bottom-up participation by citizens themselves. Findings from our fieldwork confirm that participatory mechanisms, such as local or municipal councils, give poor people an active voice in how Bolsa Família operates, as well as a sense of stakeholder empowerment.

Cadastro Único

Identifying and Finding

To reach the poorest of the poor involves identifying, finding and delivering to them social services. How does the Bolsa Família program (BFP) find and identify the poorest of the poor?

The Cadastro Único Para Programas Sociais (Unified Registry for Social Programs) is the central government registry for all of Brazil’s low-income households. Created in 2001 by the Ministry of Social Development (MDS), the Cadastro Único (CadÚnico) houses information in a single database about poor families that potentially qualify for social assistance benefits. Currently the CadÚnico has data on more than 27 million households, or approximately 80 million people, accounting for 40% of all Brazilian households. Nearly 14 million households benefit from Bolsa Família.

Data recorded in the CadÚnico is used by the government to determine which families and individuals are eligible for 30 different governmentsponsored social service programs. In addition to Bolsa Família, other social programs that use CadÚnico data to determine eligibility include the Tarifa Social de Energia Elétrica (electricity subsidy), Programa Minha Casa Minha Vida (My Home, My Life), and the Bolsa Verde (environmental subsidy). In this regard, the CadÚnico actively identifies qualifying families; many poor families do not know about certain programs or are unaware that they are eligible to receive social program benefits. The CadÚnico ensures poor families do not fall through the cracks.

The CadÚnico is a tool for identifying and finding poor families. The detailed database provides a “map” of poverty in Brazil, showing the government where the most vulnerable and poor families live. By identifying and finding such households, the CadÚnico is also used to effectively deliver social services and programs such as Bolsa Família to eligible families.

“The Cadastro Único has become a tremendous resource. It has quantified poverty, identified geographical areas of socio-economic vulnerability, and it has revealed the face of poverty in Brazil is that of a young black female living in a rural area.”

MDS Official

Collecting Data

Information about low-income households is collected with the use of a lengthy questionnaire (the “green book”). The CadÚnico collects data on the family, including information about: household income; information about each family member such as education levels and employment status; the number of children; descriptions of the physical residence in which they live; and so on. Each family member is given a social identification number (NIS).

Households provide the following information in the Cadastro Único questionnaire:

• housing characteristics (number of rooms, type of construction, address or landmarks)
• family composition (number of family members, pregnant women, elderly)
• identification of each household member
• education level of each household member
• employment status of each household member
• family income
• family expense (rent, transportation, food)

It is important the CadÚnico records information about where the family lives. This is often not straightforward, however, as many poor people do not have a permanent residence. Some are homeless. Others live in slums without a formal address. Some live along the rivers. Hence, in addition to asking registered households in the CadÚnico to indicate their address, the questionnaire also asks them to physically describe their dwelling (i.e. apartment, home, shared room, block); to indicate the public space they live near (i.e. a street, avenue, river stream, an alleyway); and whether their residence is permanent, improvised or shared. Question 1.20 of the CadÚnico questionnaire asks specifically that the registered family provide additional descriptions of their residence to help social workers locate where they live. As an MDS official explains, the CadÚnico collects “clues” to help the MDS determine where the poorest of the poor can be found.

The household information collected for the CadÚnico is self-reported, though the responses are recorded by an interviewer who administers the green book questionnaire. Most often the head of household (for the purposes of registering for social programs) creates a file by visiting a Reference Center for Social Assistance (CRAS). The head of household is usually the mother (or female head of household). The interview data is digitally inputted. In larger cities, such as Salvador, the main CRAS in the city center receives approximately 500 visitors each day. The number of CRAS centers in each municipality varies. In the small town of Sapeaçu, for instance, there is one CRAS center to service the town’s 18,000 inhabitants, while in the large city of Belo Horizonte, there are 33 centers for nearly 2.4 million people.

The reality for many poor families is that the CRAS center is not accessible, either because it is too far away or because the registration process is too onerous or simply not a priority. To ensure that all lowincome households are registered in the CadÚnico, the CRAS centers will dispatch social workers to poorer or distant neighborhoods to directly register families for the registry. Social workers will often conduct home visits to assist families in the registration process.

Despite efforts to bring the registration process directly to homes, however, there remain many who are not included in the CadÚnico. Brazil’s poorest of the poor live in slums as well as the most remote and difficult to reach places, such as the countryside or jungles along the Amazon River. Many Brazilians also face what government officials refer to as “double exclusion”: both economic and social exclusion. Sixteen “traditional groups” are recognized by the MDS – including Gypsies, Quilombolas (descendants of slaves), Ribeirinhos (traditional fishing communities), indigenous groups, homeless people, among others – who are poor and socially marginalized, and often without any formal identification. Reaching these families is difficult.

To ensure the CadÚnico precisely identifies members of “traditional groups,” low-income households and those who qualify for social programs such as the BFP, the MDS employs a strategy of active search. Active search is implemented at the local level, addressing the specific challenges of municipalities. For example, in Dias d’Ávila, a town of 80,000, the CRAS dispatches a van – the Unidade Social Móvel (Mobile Society Unit) – to do weekly visits to families who live far from the town center. In Belo Horizonte, the government assigns a basic health unit to specific catchment areas within the city and family health teams go door-to-door to follow-up with families and to enlist new ones into the CadÚnico.

In addition to reaching those who are hard to reach, the CadÚnico must also ensure that the data it collects is up-to-date. Inaccurate data undermines the ability of the MDS to identify, find and deliver social services to poor households. During the initial implementation of the CadÚnico in 2003, the quality of household data in the registry was very poor. The CadÚnico was not updated regularly, there was a lot of missing data, and the MDS was unable to verify the accuracy of reported information. Thus in 2005, the MDS initiated a major push to improve the CadÚnico, to “clean” the existing database, and to put into place new mechanisms to ensure the continual updating of household information into the future. Municipalities were incentivized (with fiscal resources) to carry out this federal initiative. According to MDS officials, 85% of the current administrative work on the CadÚnico is spent on updating and verifying the database, while the remaining 15% is dedicated to enrolling new families.

Families are encouraged to update – self-report – their registry information whenever there are changes to their status, such as births or deaths; a change of address or residence; changes in employment status and income; and so on. The municipal CRAS centers are also pro-active in updating information, as well as in verifying household changes. To ensure the registry data are accurate, the CadÚnico is periodically checked, household data verified with local-level CRAS centers, and updated at least once every two years.

“On the families’ addresses: why are they so important? To send them their cards and letters. Finding them is the name of the game. It doesn’t matter whether the address is precisely right. People only need to be found. So how do we find them? There are many ways: the postmen who work in the favelas know where everybody lives; or people usually look for their mail at post office centers, and this is also true for people living in rural areas (including Amazonia); the zip code system is very good in Brazil (most people know their own), and people who live in big cities take advantage of it. These are the clues for the hardest-to-find situations.”

MDS Official

The CadÚnico and Bolsa Família

The CadÚnico is managed by the central government’s Ministry of Social Development (MDS), though the actual implementation of data collection is the responsibility of the 5570 municipalities in Brazil. The municipal coordinators of the CadÚnico and Bolsa Família are the main points of contact for beneficiaries, and municipal agencies, such as the CRAS centers, are in charge of data collection, entry, verification and updating. Household-level data collected by the municipalities are consolidated centrally. The federal bank Caixa Economica Federal (Caixa) operates the CadÚnico registry.

The MDS uses the CadÚnico registry to target – to identify and find – those households which are eligible for the Bolsa Família program. The targeting process involves three steps. First, municipalities collect household-level data about low-income households. Second, the data are consolidated centrally and managed by the Caixa Bank. And third, the MDS identifies the specific families who qualify for the BFP and enroll them into the program (see graphic below).

Summary

The successful reach of the Bolsa Família program is due to the extraordinary data provided by the Cadastro Único that are used to find and identify the hardest to reach families. The CadÚnico provides the MDS with precise and updated household-level information to target and deliver conditional cash transfers to very poor, socially excluded as well as geographically distant families.

Direct Transfer

Delivering Directly

The Bolsa Família (BFP) cash transfer is delivered monthly to enrolled families through an electronic payment system operated by the Caixa bank. Beneficiaries receive a BFP bank card upon enrollment in the program. The card is used to withdraw funds, much like from a personal debit bank account, from Caixa ATMs, bank branches and ubiquitous lotteria houses throughout Brazil. The electronic payment system was adopted to ensure that cash transfers are reliable, safe and have a broad reach; for instance, sending a check through the postal system is less viable as many poor households do not have permanent housing or an address, and are unlikely to have a formal bank account. The BFP payment system does not require the beneficiary to have a bank account or a Caixa account. Simply put, being “unbanked” does not prohibit a poor family from claiming BFP benefits.

Because BFP funds are directly delivered to families through an electronic payment system, money is not transferred through either the state or municipal governments. This reduces fiscal leakage and mitigates the possibility of corruption and patronage.

Each beneficiary’s BFP card has a specific identity number. This number is not only unique to the individual beneficiary household, but also indicates what specific day of the month the funds are available to be withdrawn. According to current regulations, funds that are not withdrawn on the specified date are held for up to three months.

The BFP payment system is operated by the federal government-owned Caixa bank. Payments are paid directly to the beneficiary through Caixa’s existing networks of ATMs, bank branches and Caixa-owned lotteria houses. In addition to operating the payment system, the Caixa bank also provides administrative and technical support to the CadÚnico and the BFP, such as hardware, data analysis, information systems management and software, as well as local troubleshooting. As the operator of the payment system it is in Caixa’s interest to maintain a secure and efficient cash transfer mechanism. Continual optimization of Caixa’s payment system operations, the BFP database and the CadÚnico has resulted in significant reductions in administrative costs since the program’s implementation. In 2015, the MDS and Caixa introduced the BFP mobile app, a beneficiary-facing app which allows BFP beneficiaries access to their account, updates, conditionalities, and other important sources of information relating to their program status. Prior to the introduction of this app, BFP enrollees had to present themselves in-person at a government office to access their account.

Customized Messaging

BFP is a conditional cash transfer program, meaning beneficiary families must meet all education and health conditionalities to remain eligible for cash transfers. Benefits are suspended when families fail to fulfill the conditions. However, often times the beneficiary is unaware that their benefits are soon to be suspended or they have yet to update their school and health records. Before an account is suspended, the beneficiary will receive a warning slip directly from the ATM (or other cashpoint) when they go to withdraw the cash transfer. The slip explains why their account is or will soon be suspended. The warning is specific to that individual, and not a generic message, reflecting the precise data which the BFP uses to track and reach beneficiaries. The warning slip is not intended to be punitive. The messaging is positive, emphasizing to the beneficiary the importance of meeting the education and health conditions. The MDS finds that with customized and encouraging messaging, people are more likely to take action and fulfill the program’s conditionalities.

Reaching the Hard to Reach

The CadÚnico is the data backbone of the BFP. Unique household information about individual lowincome households is collected and updated by the CadÚnico in order to identify and find poor families eligible for BFP benefits. But identifying and finding poor families does not mean the challenge of delivering cash transfers to them is resolved; knowing where they live is not the same thing as getting cash to where they live. While the Caixa’s network of ATMs, bank branches and lotteria houses is extensive, it is also limited to the existing Caixa infrastructure. Banking regulations prohibit the diversification of possible cashpoints to grocery stories, post offices, retail stories and point-of-sale devices. Thus, those living outside the Caixa network – such as families in the remote countryside – are less likely to be reached. It is not uncommon for eligible families living in exceptionally remote areas to exceed the three month deadline to withdraw their BFP funds; not because they are negligent or have failed to meet their conditionalities but simply because they cannot reach a cashpoint (i.e. ATM) within the three month hold period. Government officials are currently contemplating changes to the hold period for families living in remote areas, extending it from three months to six months so that families are able to access a cashpoint.

Summary

Unlike other social services and cash transfer programs, the BFP directly transfers cash benefits to recipients through an automated electronic payment system. Though it is a challenge to get cash transfers to those who are unable to regularly access the Caixa banking network, the direct transfer system maximizes the BFP’s reach and minimizes fiscal leakage.

Local Implementation

When the Brazilian government was considering how to design and implement the Bolsa Família program (BFP), advisors from Mexico encouraged Brazilian policymakers to centralize the conditional cash transfer scheme. The Brazilians resisted, however. Social policymakers chose to maintain a centralized administration and cash payment system, but contrary to the advice of their Mexican counterparts, the BFP was implemented by municipal governments. Local implementation more accurately reflects the vast diversity of local contexts in Brazil; a one-size-fits-all approach would not work.

The Role of Municipal Governments

The central government, and specifically the Ministry of Social Development (MDS), outlines the goals and objectives of the BFP, though local municipal governments implement the program. Municipalities, and not the central government, are responsible for collecting and updating the data used in the Cadastro Único; actively search in order to find and identify families that are eligible to be enlisted in the unified registry; monitor and report that health and education conditionalities are met; and provide supplementary services to BFP recipients. Simply put, with 5570 municipalities in Brazil, there are also 5570 different implementation strategies.

For example, Belo Horizonte (BH) is a large, urban metropolis. The BFP and other social service programs are implemented by the Belo Horizonte Municipal Council of Social Assistance; the Social Assistance Secretariat; the coordinating office for the BFP; as well as neighborhood commissions and district councils. Given the size of BH, the city organizes nine main regional Social Assistance (CRAS) Centers. Additional CRAS centers and community centers are located in the most vulnerable and poor areas in BH. In Sapeaçu, a rural town of just 18,000, the implementation and organization of the BFP is much simpler than in BH, though similarly based at the level of the municipal government. Whereas in BH the CRAS centers and local health clinics are spread throughout different regions and catchment areas in the city, the entire BFP administration is located in one village block. The staff complement in Sapeaçu is significantly smaller as well.

How each municipality implements and manages the BFP differs. In BH, for example, social workers (CRAS social workers, community health agents) are assigned to specific catchment areas identified as “high vulnerability” neighborhoods within the city. The BH municipal government uses census data to identify which catchment areas have a higher concentration of poor families and allocate more social service resources to those vulnerable areas. CRAS social workers are responsible for their own catchment areas, registering families and monitoring them to ensure that conditionalities are met. Community health agents and workers will both deliver health services to BFP recipients and refer their patients to the CRAS center to register for or update their CadÚnico profiles. Because the catchment areas in BH are densely populated, distance is not an obstacle to reaching the poor. However, the population density, the spread of informal employment, and the lack of permanent housing make it difficult to find and identify especially vulnerable families.

In rural areas, on the other hand, the population is disparate, homes are widely spaced and can be far from the town center where the CRAS centers and health clinics are located. In Dias d’Ávila, a smaller city in the northeastern state of Bahia, social assistance vans are deployed to reach vulnerable households that live far from the city. BFP workers, CRAS social workers and community health agents travel up to an hour to find and reach those families who cannot or are unwilling to travel to the city center to register or update their CadÚnico profile. In the city of Salvador, Brazil’s third largest city but also one of the poorest, many poor families do not even know the CadÚnico and BFP exist. Mobile teams are thus deployed in very poor neighborhoods and favelas to register families in the CadÚnico and for the BFP. Many of these areas are without any internet and mobile connectivity, and word-of-mouth and door-to-door visits are the only way in which CRAS and BFP workers can reach very poor and marginalized households. Churches and neighborhood associations will also request the mobile teams to come to their specific areas in the city.

Monitoring BFP recipients to ensure that health and education conditionalities are being met is also a municipal responsibility. Prior to 2006, there was no comprehensive monitoring system in place. As a result, during the early days of the BFP, data reporting on health and education conditionalities was spare and inconsistent. Then, only 40% of BFP beneficiaries were monitored to ensure that health and education conditionalities were being met. This has improved since 2006, however, largely due to municipal efforts to increase their capacity to accurately monitor and report on whether individual households are meeting the BFP health and education conditionalities.

In BH, schools and basic health units receive a list of children and families living within a specific catchment area who are receiving BFP benefits. Teachers and family health teams are responsible for recording school attendance and medical check-ups for each beneficiary. This information is consolidated and sent to the municipal secretariats for education and health, and then collected centrally by the federal Ministry of Social Development (MDS). But just as it is very difficult to identify poor families for the CadÚnico and deliver cash benefits to many poor and hard to reach households, it is similarly challenging to accurately monitor and report on health and education conditionalities. Social workers assigned to specific catchment areas in BH often go door-to-door in poorer neighborhoods to check that BFP recipients are meeting the program’s conditionalities. In Dias d’Ávila mobile units are deployed to register families for the CadÚnico and to report that conditionalities are being met (or not) in the town’s schools and health centers. And in Sapeaçu, community health agents make daily home visits to remind families about meeting the BFP’s health and education conditionalities.

Decentralized Management Index

To ensure municipalities improve the implementation and administration of the BFP program (BFP), the Ministry of Social Development (MDS) has put into place the Decentralized Management Index (IGD), a performance-based incentive scheme for municipalities. The IGD scores municipal performance, rewarding municipalities that achieve high implementation and management scores with additional fiscal resources and subsidies from the central government.

The IGD evaluates each municipality on the following criteria:

1. The quality of the data collected for the Cadastro Único
2. The percentage of families that have updated their registry information
3. The percentage of children whose school attendance are monitored
4. The percentage of children with completed and accurate health records

The IGD incentive system, in one respect, was created to lessen the variation in implementation effectiveness of the BFP across municipalities. Municipal performance, while overall very good, nonetheless varies, depending on the resources each has, as well as the political will to effectively implement the BFP. The MDS finds that service provision tends to be less effective in poorer and smaller municipalities, and where, in some cases, municipal leaders do not have the political will to implement the CadÚnico and the BFP. The IGD scheme was created to incentivize municipalities to create political will and to provide fiscal resources to the most vulnerable municipalities to improve the implementation of the BFP. For example, vans were purchased with IGD subsidies in Dias D’Ávila and Sapeaçu for their mobile units.

Local Innovation

Local implementation of the BFP gives municipalities the opportunity to innovate in order to meet their specific needs and to address specific challenges. Reaching the urban poor in densely populated favelas where residents do not have an address or are constantly on the move presents different challenges than reaching the very poor in geographically isolated areas such as in the Northern region and along the Amazon River. Municipal autonomy to implement the BFP allows local governments to devise local solutions that best address local challenges.

BH has always been at the leading-edge of innovation in social services provision. It was one of the first municipalities to introduce a local conditional cash transfer program in 1996. Before the BFP was created, the BH government generated a vulnerability index of the city using georeferenced data from surveys and census data, identifying catchment areas in BH where the city’s poorest and most vulnerable live.

In 2002 the BH government introduced a supplementary social assistance program for high vulnerability areas: BH Cidadania. The BH Cidadania was rolled out first in nine pilot neighborhoods in BH, and has expanded to 34 catchment areas, each with its own BH Cidadania center. The city’s social assistance secretariat hopes to expand the program to 54 catchment areas in BH. The BH Cidadania centers are essentially a one-stop shop for all of the city’s social service programs, including supplementary services such as job training workshops. Social workers at the BH Cidadania centers keep track of low-income families, their eligibility for social programs and registration in the CadÚnico. In 2009, the BH municipal government created the Família Cidadã program, specifically targeting the most vulnerable citizens in BH. For instance, to qualify for the Família Cidadã program, families must earn, per capita, less than R$70 each month and score “extremely vulnerable” in the Belo Horizonte Family Development Index. Beneficiaries of the Família Cidadã program are given more attention with respect to their health and education needs.

Beyond the Municipality

State governments play an important role in the implementation of Bolsa Família as well. They are a bridge between the municipal and federal governments, and provide municipalities with financial and administrative support, including staff training. The Bahia state government, for example, keeps track of those municipalities that have scored relatively poorly on the IGD scheme, identifying specific municipalities for more technical and training assistance in data collection and monitoring conditionalities.

Like municipalities, state governments are also incentivized to improve their state-wide IGD scores. States are evaluated on their performance in registration and registry maintenance, monitoring, and overall administration of the BFP (and other social programs). Similar to municipal governments, states receive important fiscal resources for strong performance. Funds from the IGD system were used by the Bahia government to hire extra personnel to digitally input and update more than 40,000 entries into the CadÚnico.

Summary

Though the BFP is a federal social policy program, it is implemented by municipal governments. In Brazil there are 5570 different municipalities, each with its own unique conditions, needs and challenges. Municipalities have a fair degree of autonomy in how they implement the BFP, specifically in how each municipality finds, identifies, monitors and delivers to hard to reach populations. In this respect, municipal implementation encourages local innovation and the creation of new best practices.

Meeting Conditionalities

The objective of the Bolsa Família program (BFP) is to not only provide immediate poverty alleviation with the regular transfer of cash benefits, but to also break the intergenerational cycle of poverty in Brazil. As noted by social policymakers, the poor in Brazil tend to stay poor across generations. To break out of this so-called poverty trap, the very poor, the excluded and vulnerable need sources of human capital, notably health and education.

As the BFP is a conditional cash transfer program, beneficiaries must meet health and education conditionalities in order to receive cash transfers. The MDS stresses that the health and education conditionalities are not intended to be punitive, but rather to give access to otherwise excluded and poor families to Brazil’s health and education systems.

Health Conditionalities

The central government’s Ministry of Health and the municipalities share responsibility for delivering health care through Brazil’s Sistema Único de Saude (SUS, or Unified Health System) and for monitoring that health conditionalities are being met for the BFP. The specific health conditionalities of the BFP focus on health needs of children and mothers. Children ages 0 to 7 years are required to receive all vaccinations recommended by the SUS and to attend regular health check-ups to monitor growth and nutrition. Pregnant and lactating women are required by the BFP to attend pre-natal check-ups, post-natal follow-ups, and to take part in health promotion activities at local health units.

Monitoring that conditionalities are being met is difficult to do, given the challenges of finding, identifying and delivering the BFP to hard to reach populations. The MDS has set a target to monitor health conditionalities for about three-quarters (73%) of all BFP beneficiary households across the country. Given there are over 48 million individual beneficiaries of the BFP, who reside in 26 states and 5570 municipalities, the MDS has decentralized monitoring responsibilities. As with the implementation of the BFP, monitoring is by and large the responsibility of the municipalities.

At the municipal level in Belo Horizonte, for example, community health agents (CHAs) are assigned specific catchment areas within the municipality. The CHA is responsible for conducting regular home visits in their assigned areas to help monitor and encourage BFP families to meet the program’s conditions. Because community health agents work extensively within certain areas, the CHA becomes known to families and trust develops between the health agent and BFP families. During home visits, for example, the CHAs will request to see the children’s immunization cards or the mother’s checkup booklet. Monitoring reports are recorded by the CHAs and forwarded to the municipal secretariat and then the central Ministry of Health (MOH). This data is consolidated by the MOH and communicated regularly to the MDS.

Meeting the health conditionalities of the BFP has had a far-reaching impact on the health status of poor Brazilians. The Institute for Applied Economic Research (IPEA) reports that in 2012 the health care system delivered basic care services (i.e. pre-natal check-ups, infant check-ups, vaccinations) to approximately 8.6 million households, including to 5.1 million children, and 13.8 million women. Remarkably, of the monitored children, over 99% were up-to-date on their vaccines, and similarly 99% of pregnant women who had been monitored had met their pre-natal care conditionalities.

Education Conditionalities

 The BFP requires all children ages 6 to 15 of beneficiary families be enrolled in school, and that each child maintains a daily attendance record of a minimum of 85% school days per month. Additionally, parents must inform the school of absences and notify the municipal BFP coordinator if their child moves to a different school. Youths aged 16 to 17 years old are required to achieve a minimum attendance record of 75% school days per month.

The monitoring of education conditionalities, as with the health conditionalities, requires co-operation among various actors. Teachers record daily attendance. Attendance records are processed by each school’s administration and forwarded to the Ministry of Education. Absences are also reported to the Ministry of Education. Schools are required to establish a dialogue with families who fail to meet the BFP’s education conditionalities or who are at risk of failing to meet them in order to identify additional social assistance needs of the family and to connect them with the necessary support networks. The Institute for Applied Economic Research (IPEA) estimates that over 1 million individuals are involved in the process of collecting and tracking the information of more than 17 million students, from approximately 170, 000 schools.

The BFP’s educational requirements are supplemented by two other programs: Saúde na Escola (Health in Schools) and Mais Educação (More Education). The second program, More Education, has increased the national average number of school hours per day from four to seven. The implementation of the More Education program specifically prioritizes those schools in which at least half of the students are beneficiaries of the BFP.

The education conditionalities of the BFP have produced impressive outcomes, particularly in the poorest Northeast region of the country. Historically, education outcomes for students in the Northeast are considerably below the national average. For example, whereas the national average of the percentage of students who have completed elementary school before the age of 15 is 79.4%, in the Northeast this figure is only 64%. Within the Northeast, however, students who are beneficiaries of the BFP rank above their non-BFP peers in elementary school graduation rates by the age of 15 at 71.3%. According to data from the School Census and the MDS, students in the BFP also perform much better and are less likely to dropout from high school than their peers who are not recipients of the Bolsa Família conditional cash transfer. Again, the relatively stronger performance of BFP recipients is particularly pronounced in the poorest region of Northeast Brazil.

Strengthening Health and Education

The BFP’s success in monitoring and encouraging better health and education decisions by poor – and hence, better outcomes – is a reflection of not only the BFP conditionalities but also the social program infrastructure in place. In other words, meeting health and education conditionalities requires that adequate health and educations services are available. When it comes to health, the decentralized delivery of health services in the SUS has improved patient access, particularly in poorer areas and those further from city centers. In addition, in many municipalities, especially those in which there are a large number of poor families, the health service providers, the CRAS centers and the Bolsa Família programs are tightly integrated, meaning that CRAS social workers will encourage BFP recipients to meet the program’s health and education conditionalities while community health agents will encourage poor families to register in the Cadastro Único.

There is always room for improvement, however, when it comes to the delivery of health and education services. So long as poor families face challenges in accessing adequate health and education services, meeting the developmental goals of the Bolsa Família program will be mitigated. The MDS reports that health and education services access remains a challenge for families that live far from village and town centers and for those who are doubly excluded because they are poor and culturally marginalized.

Summary

The BFP is a conditional cash transfer scheme, intended to not only alleviate poverty through cash benefits but also to break the cycle of intergenerational poverty. BFP beneficiaries must therefore meet health and education conditionalities in order to receive the cash transfer. Monitoring and reporting on health and education conditionalities are municipal responsibilities. The BFP conditionalities have had a positive impact on both health and education outcomes for families enrolled in the BFP.

Citizen Participation

Participation Matters

The passage of the 1988 Constitution in Brazil was a watershed moment. In addition to putting Brazil on a democratic path, the Constitution also committed the government to protect all citizens’ social rights. Often the gap between the rhetoric of a constitutional commitment to social rights and the actual practice of inclusive democracy is large. Indeed, democracy can sometimes be double-edged, in that democracies can be committed to social provision for the poor but also rife with corruption, clientelism and political exclusion. As one observer of Brazilian politics puts it, “the groups at the bottom lack the means to influence the country’s economic and political development.”

To ensure social programs actually reach the poorest of the poor, bottom-up participation is required in the design, implementation, feedback and continual evaluation of the program. Participatory budgeting – a system used in many developing countries at the local level – is a good example of how the involvement of regular citizens in budgeting improves government effectiveness. Similarly, a recent study of social policy in Brazil shows that in those municipalities where local policy councils involving citizen participation are active, health and well-being outcomes are better than in those cities without such councils.

But how do these councils matter; what do these local councils do? Empirical surveys and our field research in Brazil confirm that there is significant variation among municipalities when it comes to local participation, especially with respect to the role of local policy councils. In some municipalities, local policy councils do not exist at all, or merely exist “on paper”; while in other municipalities, councils are very active. And among those that are active, what it is they do varies quite considerably as well.

Municipal Councils

In general, local municipal councils enjoy a veto voice over the functions of the various social services secretariats (i.e. health, education, social assistance). Each municipality, however, organizes and operates their municipal councils differently.

Local participation in Belo Horizonte (BH) occurs at three levels: local commissions, district councils, and municipal councils. BH is a very large city, tremendously diverse in terms of household income, neighborhood composition and traditional group minorities. The local commissions operate at the most micro-level to reflect specific neighborhood needs and initiatives. All citizens who live within a catchment area are invited to attend local commission meetings, which are held monthly. Citizens provide feedback, and voice opinions and suggestions for improvement to the local commission. Above the local commission are the district councils. A city the size of BH needs to organize municipal governance into regional districts; each district is represented by several functionally-specific district councils, such as health councils, education, and social assistance councils. The district council’s membership includes city policymakers, municipal officials, and representatives from the local commissions. The municipal council operates at the city-level and comprises city officials, representatives from the district councils, non-governmental organizations and local health, education and social assistance secretariats.

The city of Salvador in the northeastern state of Bahia, also a large city, similarly organizes municipal councils, the most active being in health, education and social assistance. Like BH, the health, education and social assistance secretariats, including the BFP manager, are accountable to the municipal councils, which have the authority to audit and approve initiatives taken by the secretariats. The social assistance secretariat in Salvador, for example, had its budget suspended for several months because the municipal council overseeing social assistance programs did not approve the budget. Municipal councils in Salvador comprise equal numbers of members from government and civic representatives.

Sapeaçu is a very small village in Bahia, with a population of 18,000. Similar to Salvador and BH, the three main local councils in Sapeaçu represent health, education and social assistance. In Sapeaçu the health council is the most active, however. The health council (at the time we visited) is composed of eight members: three from the Sapeaçu government, two representatives from the health clinic, and three members from civil society. The local health policy council members regularly visit and evaluate the village’s local health clinics. The council is responsible for approving municipal health budgets. The most recent head of the Sapeaçu health council mandated that one representative on the council be from the countryside.

The Dias d’Ávila municipality organizes six policy councils to address health, education, social assistance, youth affairs, the elderly, and the disabled. Each council is composed of equal representation from government and civil society. Council membership must change every two years. Unique to Dias d’Ávila, civil society representatives in the various councils must have previously been involved in a civil society organization (i.e. social movement, community organization), thus precluding regular citizens from participating. The rationale for this is that societal representatives on policy councils need to reflect prior experience in a specific policy area. Similar to the other policy councils visited, the functionally-specific policy councils in Dias d’Ávila have the authority to approve municipal budgets earmarked for social services.

In summary: despite variations in how local policy councils are organized, they all share important common characteristics, notably, to represent broad and diverse citizens’ interests and needs; to provide a forum for citizens’ voices to be heard; and to generate a sense of stakeholder empowerment among citizens.

Local Demands, Municipal Action

BH is large city, divided into nine administrative regions. Local matters are discussed and debated within these regions. Recently, the BH municipal government proposed to move the boundaries of the Central and South regions of the city so that the city’s social services could be more efficiently utilized. Moving the regional boundaries, however, meant that existing catchment areas for the community health centers would be re-aligned as well. Families that had used the social services in one particular health center and had developed deep bonds of trust with the service provider were now required to access a different center. Family health teams – the network of community health agents that worked with families in specific catchment areas – would also be re-organized as a result.

Citizens living in the Central and South regions of BH opposed moving the regional boundaries. The issue was raised during a monthly health commission meeting, and subsequently a counterproposal was drafted and presented to the district health council. At the municipal level, the citizens’ voices were heard and the regional boundaries were reverted as they had been before.

This example of local demand and municipal action is significant for two reasons. First, it demonstrates how voices from the bottom-up can be heard by municipal decision-makers. And second, BH’s threelevels of policy councils – commissions, district councils, municipal councils – are a mechanism through which community participation is achieved and can be effective.

Summary

Participation by citizens in the design, implementation and evaluation of the program is important in ensuring the BFP operates efficiently and effectively. Though municipalities vary in terms of how local participation is organized, common to all of the research sites visited was the presence of local policy councils that included citizen members (i.e. non-governmental representation); exercised some policy authority; and conveyed citizen opinions and feedback to municipal decision-makers.

Lessons Learned

The Bolsa Família program (BFP) is one of Brazil’s signature – and celebrated – social policy initiatives. Launched in 2003, the conditional cash transfer program currently reaches 13.8 million families or about 48 million people, and remains very inexpensive to operate. The BFP has helped to reduce poverty in Brazil and has improved social and economic outcomes for the poor. The BFP is well on its way to breaking the cycle of intergenerational poverty in Brazil. Most impressive is the fact that the BFP has successfully reached Brazil’s poorest of the poor families, society’s most vulnerable and socially excluded people. It is an example of how precise targeting can have a positive impact on not only improving the lives of the poor, but of Brazilian society as a whole. BFP is a model of “reaching the hard to reach,” one that ought to be emulated and adapted in other national settings.

What lessons can be learned from the Bolsa Família program?

Big and Precise Data

The backbone of the BFP is the CadÚnico, an impressive database that identifies and finds poor families in Brazil. The CadÚnico is not just a source of big data, however; rather, it is an extraordinary example of precise and accurate data. It is this precision that allows the government to effectively and efficiently target the poorest of the poor, identifying and finding those who are eligible for all of its social programs, including Bolsa Família.

Payment to the Poor

Cash benefits are directly transferred to BFP recipients, thus bypassing the state and municipal governments. This reduces the transaction costs – such as fiscal leakage, corruption, and administrative costs – typically associated with cash transfer programs. Most importantly, the direct and automated payment system, operated by the federal Caixa bank, ensures that benefits reach their intended beneficiaries.

High-Tech and Low-Tech

The effectiveness of the BFP depends, in part, on high-tech solutions. The management of the CadÚnico and the operation of the automated payment system, for example, rely on technological interventions. Yet, at the same time, the BFP’s effectiveness in reaching those families that are most difficult to reach – such as the homeless, slum dwellers, those who live in rural areas – reflects also low-tech solutions such as human effort. Social workers, community health agents and local BFP managers, among many others, are responsible for home visits, door-to-door canvassing, mobile units and other strategies of “active search.”

Local Implementation

Brazil is an extremely diverse country, and not surprisingly, the challenges of reducing poverty are equally diverse. Each municipality – and even neighborhoods within a municipality – has its own challenges when it comes to addressing the problem of poverty. Hence, even though the BFP is a federal social program, administered and managed by the central government, the implementation and operation of the program is decentralized, with each municipality in charge of registering and updating the CadÚnico, implementing and improving the BFP, and monitoring that conditionalities are met. Municipalities have a fair degree of latitude, and hence each municipality implements the BFP to best suit its specific needs and challenges.

Citizen Stakeholders

Citizen participation is critical to ensuring the BFP is effective in reaching the poorest of the poor. Empirical studies show that active local councils are associated with better BFP performance and improved social and economic outcomes. Local councils have a positive effect on the BFP when they include citizens as part of their membership; listen to citizens’ voices and feedback; and give citizens real authority in establishing local budgetary and policy priorities