1. Introduction
Social policies are created in order to compensate for the distortions arising from the process of capitalist development, which discriminates and creates an increasing gap between the rich and poor. The role of the state is to provide funds in order to ensure the welfare and to comply with the established constitutional right that is the guarantee to health, education, work and food for every individual.
The state comes as a provider of these needs by creating social policies that ensure the survival for those living in extreme poverty. Based on this, income transfer programs were created, with the purpose of, in the short term, alleviate the problems arising from poverty and in the long term, investing in human capital, breaking the poverty cycle.
A conditional cash transfer (CCT) is a type of income transfer program with the aim to reduce poverty by creating welfare programs conditional to the beneficiary’s actions. The government only transfers the money to people who meet certain conditions, which may include the enrolment of children in school, going to the doctor and receiving vaccinations, among others.
Bolsa Familia is the largest conditional cash transfer in the world and it was implemented in Brazil (World Bank, 2005). and it consists in the financial aid to poor families as long as they keep their children often at school and vaccinated. The program aims to reduce poverty in the short and long term through conditional transfers of capital, which, in turn, aims to break the generational cycle of poverty from generation to generation.
The elements that determine the outcome of an election change considerably depending on the country, its historical moment, social, economic and political conditions and on the candidates who intend to occupy the position. In Brazil, this dispute has been polarized in the last two decades, around two political parties: the Workers’ Party (PT) and the Social Democrat Party(PSDB).As mentioned before, the Bolsa Familia Program was the unification and expansion of social programs created in the government of Fernando Henrique Cardoso (Social Democrat Party) in a single social program during the transition to the government of Luiz Inacio Lula da Silva (Workers Party) in 2003, thus both governments were in favor of the program.

In the last three presidential elections, there were significant changes in the electorate which determined the selection of the president, especially between 2002 and 2006. In the elections that led to the re-election of Luiz Inacio Lula da Silva, the largest share of votes was obtained in less developed regions of Brazil, unlike in 2002.

In 2010, the candidate Dilma Rousseff, who early in her campaign presented little expressive intention to being voted in the polls, gained strength and won the elections. This result was strongly attributed by the media to the unconditional support of then President Lula (who ended his second term with a high approval rating), the reasonable economic growth presented in its eight years of government and the spread of the Bolsa Familia program during the last government .

2. Bolsa Familia
Since the 1980s there was a debate about providing assistance to extremely poor families in Brazil. The granting of benefits and help was then done indirectly, usually by distributing food baskets to areas in need(especially the North and Northeast), sometimes followed by allegations of corruption due to centralized purchasing in Brasilia, and the deviation of goods due to the lack of logistical control. During the government of Fernando Henrique Cardoso (Social Democrat Party) finally the so called income distribution programs were effectively implemented in the country, some in partnership with Non-Governmental Organizations.

The creation of BolsaFamília was inspired by the Scholarship Program (BolsaEscola), a program created in the city of Campinas-SP in 1994, being implemented by the federal government of Fernando Henrique Cardoso (Social Democrat Party) in 2001.In 2002 Brazil already had a multitude of social programs that have benefited some five million families by, among other programs such as "BolsaEscola", under the Ministry of Education, "Gas Allowance", under the Ministry of Mines and Energy and "Food Card", under the Ministry of Health, each managed by different bureaucratic administrations.

The Bolsa Familia Program was the unification and expansion of social programs in a single social program, prepared during the transition from the government of Fernando Henrique Cardoso (Social Democrat Party) to the government of Luiz Inacio Lula da Silva (Workers Party) in 2003, with registration and administration centralized in the Ministry of Social Development and Hunger Alleviation, which facilitates their administrative efficiency and control. Bolsa Familia is the most important social policies of the Brazilian government and is now the largest program of conditional cash transfers in the world.

The federal government created the Bolsa Familia program to support the poorest families and to ensure their right to food. To achieve this, there is a transfer of a monthly allowance directly to the families and these families can withdraw this allowance with magnetic card distributed by the Caixa Economica Federal bank.
There are four types of benefits within Bolsa Familia: * Basic Benefit: R$ 70, paid only for extremely poor families, with per capita income equal to or less than R$ 70; * Variable Benefit: R$ 32, paid by the existence in the family of children from zero to 15 years old - limited to five children per family; * Variable Benefit Linked to Youngsters: R$ 38, paid by the existence in the family of young people between 16 and 17 years - limited to two youngsters per family; and * Extraordinarily Variable Benefit: amount calculated depending on each case.
Families are free to use the money received however they want and they may remain in the program as long as the maintenance of the eligibility criteria and comply with the conditionalities indicated below. The cash transfer granted by the Bolsa Familia is associated with the development of other actions such as literacy programs, job training, support to family agriculture, employment, income generation and micro-credit. It also ensures access to those families who do not have children, as the case of the Indigenous and the homeless.
The BolsaFamilia seeks to address the problem of poverty in two stages. In the short term, the program aims to provide relief to the immediate and urgent problems of poverty such as hunger and the disintegration of the family environment. In the long run, BolsaFamilia is aimed at combating the transfer of poverty, leading to improved health and educational status of its beneficiaries through conditionalities, thus promoting better opportunities for the training and subsequent inclusion in future labor market.
The Bolsa Familia Program helped 11.1 million families in 2006, costing around R$ 8.2 billion, or 0.3% of GDP.
2.1 Conditionalities
Conditionalities are the commitments by both the families receiving the BolsaFamilia and the government to expand the access of families to their basic social rights. On the one hand, families should fulfill these commitments and to continue receiving the benefit. On the other, conditionalitiesmake the government responsible for the supply of public health services, education and social assistance.
In the health area, there are commitments to be met such as monitoring the health of pregnant women with routine screening, monitoring breastfeeding women and children under 7 years of age, keeping up to date with their vaccination card.
In the education area, all children and teenagers between 6 and 15 years must be properly enrolled in school and have a frequency of at least 85% of the classes. Students between 16 and 17 years must have frequency at least 75%of the classes, besides the participation of all families in food education actions when offered by the Government.
In the area of social assistance, children and adolescents under 15 years at risk or withdrawn from child labor by the Program for the Eradication of Child Labor (Peti), must participate of Familiarity Services and Strengthening of Family Bonds (SCFV) and obtain the minimum frequency of 85% from the monthly workload.
In this sense, the proper fulfillment of conditionalities is to make the beneficiary families have access to social policies, with the intention that in the long-term families are more likely to overcome their poverty.
The poorest part of the population has great difficulty in accessing services and benefits they need. In this situation, the main purpose would be to develop a strategy to facilitate access of the poorest families to services and benefits that is offered by the state within health, education, housing, etc. Thus, the Bolsa Família imposed conditionalities that should be met, giving the government the responsibility of providing access to health and education.
In a study on "Consequences and immediate causes of the recent fall in income inequality in Brazil" researchers at the Institute of Applied Economic Research (IPEA) say Brazil is part of the ten countries with the lowest concentration level of income the world. As reported by the study, the richest 10% of Brazil are 40% of income, and the poorest 40% share less than 10% of income. Given this framework, the main explanation for this disparity between rich and poor is the deficiency in education.
The low level of school attendance is the main obstacle for lowering the gap between rich and poor, and what is striking is that this poverty is transmitted from father to son, a cycle occurring from generation to generation. This means that the children of poor parents are more likely to be poor, and when they grow up and have children, they will have great chances of also being poor. There must be a break in this cycle, therefore the education and health come as conditions in the constitution of the Bolsa Familia Program, assuming that those who have more years of study have the best incomes and access to better health.
Surveys conductedbetween2005 and 2006 have shown improvement of living conditions of the poorest, with a reduction of social inequality and an increase in income of these families. The major responsible for this growth is the Bolsa Familia, which has caused great impact on the economy, being also responsible for the increase in consumption levels, mainly of food. According to research conducted by the Institute of Applied Economic Research (IPEA) Bolsa Familia was responsible for one third of the decline in inequality in Brazil between 2001 and2004.
3. Bolsa Familia and the Brazilian elections of 2006
There is currently an ongoing debate on whether, and how, do conditional cash transfer (CCT) programmes affect the electoral outcomes per country. Several studies have been made on this topic, particularly in South America, where CCTs are known to affect a large part of the electorate (up to 40% of the population in Ecuador). Such a broad scope must have at least indirect consequences on the voting behaviour of the electorate of these countries. In this part, we will consider one of the most famous examples of CCT-driven election results, the Brazilian presidential elections of 2006, and use the median voter theorem to discuss whether it had an impact in the presidential outcome.
3.1. Median voter theorem
The median voter theorem states that, if voters have single peaked preferences and they can place their options along a one-dimensional political spectrum, then the choice most preferred by the median voter is that which prevails. (Hindriks & Myles, 2006) In real life, this theory is difficult to apply because of its limitations: people usually vote for a package of measures promised by the candidate, some of which they agree with, some of which they do not. The single peakedness condition is very rarely met, because it is unlikely that one political candidate promises all the measures desired by a particular voter, while none of the other candidates offer (at least one of) the same measures. At the same time, unidimensionality of the political spectrum may be assumed to be straightforward in the form of the standard left-right spectrum. However, some voters also consider the dimension of libertarian-authoritarian of this spectrum, making it pluri-dimensional.
In our analysis, we rely on the current literature and assume, like in other papers ( (Zucco, 2009) and (Hunter & Power, 2007) ) that the CCT programme promoted by the 2006 presidential candidate Lula was the main reason why people voted for him. We will return to this limitation at the end of our analysis and present different reasons for the electoral success other than Bolsa Familia. Furthermore, we will assume a one-dimensional, left-right political spectrum when it comes to voting preferences.
3.2. The outcome of the 2006 Brazilian elections
In October 2006, Brazilians were asked to elect their president, which would stay in office until 2010. Since 2002, Brazil had been ruled by Luiz Ignácio “Lula” da Silva of the Workers’ Party (Partido dos Trabalhadores, or PT). Among other contenders, Lula’s main adversary was Geraldo Alckmin of the Social Democrat Party of Brazil. Whereas just one year before, opinion polls saw Lula lose at a significant difference to Alckmin, on election day Lula managed to stay in office, winning by 61% of the votes to ALckmin’s 39%. Most academics agree that this impressive comeback of Lula has its roots in the CCT programme that he promoted during his first mandate, the Bolsa Familia. This programme made the-then president lose most of the votes of his 2002 supporters and compensate them by winning votes from his 2002 critics. We will analyse this shift in political support, considered by some as “targeted social policy” (Hunter & Power, 2007) and by others a way of “maximizing electoral returns” (Bohn, 2011), linking it with the median voter theorem.
3.3. Lula’s first mandate
To better understand the shift in electorate which occurred between 2002 and 2006, we need to take a look at Lula’s first term as president, from 2003 to 2006. He was the presidential candidate on behalf of the PT in 1989, 1994, 1998 and 2002. His electoral package in these years corresponded to what his party stood for: transparency and social justice. His voters were mainly young, highly-educated and predominantly came from the South-east region and the urban areas. The poor did not support him in any of these elections. (Bohn, 2011)
In 2002, both the younger generation (16 to 24 years old), looking for a more progressive party, and the seniors (35 to 59 years old), fed up with the previous regime, supported Lula. Furthermore, the-then candidate won the support of 38% of the better-off Brazilians, offering him his first ever mandate as president. (Bohn, 2011)
The macroeconomic conditions left to him by former president Cardoso, who successfully battled hyperinflation and left the country running a large surplus, were very suitable for Lula to experiment with a CCT programme. Lula started by uniting four separate CCT programmes under the same umbrella, the Bolsa Familia, and continuously expanded its target group. While at the end of 2004, 59% of eligible families were reached, at the end of his first term, the proportion increased to 77%. In late 2006, Lula successfully targeted close to 100% of eligible families, representing 11.1 million families. (Hunter & Power, 2007) This achievement is deemed by many as the reason for his re-election in 2006.
3.4. Bolsa Familia as a targeted social policy
Through the Bolsa Familia, Lula tried to reduce the large level of income inequality existent in Brazil. It proved to be an effective and highly targeted measure which caused some shifts in his electorate. Unlike Cardoso’s structural reforms, Lula’s targeted policies produced more immediate electoral benefits because the results were so visible and credit for them could be taken immediately. (Hunter & Power, 2007)
In line with the median voter theorem, and considering the fact that Brazil features a high income inequality, Lula targeted a social group closer to the median voter, and much larger than his electorate: the eligible families of the Bolsa Familia account for roughly one fourth of the total population. (Johannsen, 2009) Since in Brazil, everyone is required to vote, the behaviour of the poor will have the greatest macro political effect. (Hunter & Power, 2007) This had beneficial effects for his supporters: especially in the poor North and North-eastern regions, where the public sector is the main driver of the economy, Lula’s supporters grew tremendously, both among the beneficiaries, and non-beneficiaries. In the more developed South and South-west regions, where the private sector is more relevant, these cash transfers did not create the same positive externalities as in the poorer regions, therefore support here (even only among the poor) was not as high as in the North. (Zucco, 2009)
Contrary to the median voter theorem, the beneficiaries of the Bolsa Familia were not required to vote for Lula in 2002 or 2006 in order to receive the state aid. Although some studies show that beneficiaries generally voted with Lula in 2002 (Bohn, 2011), it is unlikely that this was a significant factor in the election, considering Lula’s target electorate. Due to this discrepancy with the median voter theorem, we will only focus on the 2006 election (when the shift in electorate actually took place), and assume that people cast their vote as a reward of the results in the past (retrospectively), not because of the promises for the future.
3.5. The shift in Lula’s electorate
While Lula’s promotion of Bolsa Familia can give a hint of a shift in his electorate, he was actually gaining the support of the poor without losing (a significant amount of) his support from middle and high-class voters. We can therefore say that Bolsa Familia did not contribute alone to the shift in Lula’s electorate. Together with the increasing support from the poor voters came the loss of support from Lula’s previous supporters, the better-off Brazilians. But this happened for a political, rather than social reason. This was the president’s distancing from his political party and its ideals, and eventually due to the party’s large corruption scandals.
In sketching a picture of his party, the PT, Bohr (2011) describes it as a very democratic left-wing party that promotes transparency, social justice, political discipline and which is open for all classes (unlike other smaller parties which targeted a specific class). The PT was a refreshing sight on the Brazilian political scene due to its members’ honesty, efficiency and discipline. (Bohn, 2011) This was the main reason why it was appealing to young, middle-class idealists with an extensive knowledge of politics. However, this started to change during Lula’s first term.
Due to the large fiscal surplus that Brazil was running, Lula decided to gain the poor’s support by implementing, along with the Bolsa Familia, other measures to appeal to this target group. This was a cost-effective measure since he did not have to raise taxes and thus lose support from the well-off voters. He promulgated the “Statute of the Elderly” in October 2003, giving large benefits to the elderly, who supported him in 2002. (Bohn, 2011) Until 2006, he increased the minimum wage from R$200 to R$350. He, and his party, ignored the reforms (most notably the agrarian reform) that the party advocated when it was in opposition. Now they were focusing instead on gaining as much support as possible. (Hunter & Power, 2007) This did not go unnoticed by his well-off electorate: many people felt that the fresh ideals that Lula was bringing to the political stage were already fading by 2004. His electorate was now composed of both well-off Brazilians, and poor voters, either of which had very different interests.
In June 2005, the worst corruption scandal ever to hit the PT started. Nicknamed “Mensalão” (Portuguese for monthly stipend), it revealed that PT members were so focused on their vote-maximizing goal that they would stop at no cost to reach it. The PT was shown to have bribed several Congress members with monthly payments in return for their support of the policies advanced by the PT. The scandal nearly broke down Lula’s government and determined the party leader, the secretary and a number of key members to resign. (Bohn, 2011) With this scandal, it was clear to the idealistic supporters of the PT that their favourite party abandoned its ideals and was just as corrupt and self-interested as the other parties. Lula lost the support of well-off Brazilians and now had only that of the poor voters.
3.6. Lula’s rebranding and his 2006 triumph
In October 2005, ratings were at an all-time low for Lula, just one year ahead of the presidential elections. His membership to the PT linked Lula to the scandal and tamished his reputation. Until September, he registered a 20 percentage point decline in popularity among highly-educated voters. For less-educated people, those with just high school education, the effect was smaller, Lula losing just 9%. On top of this, due to the large discrepancies in the Brazilian society, the press could not spread the story to all citizens. Therefore, particularly in the poor regions of the North and North-east, citizens were not so well informed, and due to their low level of education, they were not as punitive as their better-off counterparts in the South. (Hunter & Power, 2007)
After the Mensalão scandal, Lula sought to distance himself from his party. He used his Northern origins to appeal to poor voters and even started questioning the integrity of his fellow party members. This convinced some low-educated voters that he had nothing to do with the scandal. Furthermore, during his campaign in 2006, he went so far as to greatly reducing the size of the party’s symbol on his advertisements and barely mentioning the party in his free TV time. (Hunter & Power, 2007) In fact, some academics create a distinction in the voters between “lulistas” (supporters of Lula) and “petistas” (supporters of the party), with one being poor and largely interested in material benefits, and the other being a highly-educated idealist. (Samuels, 2006)
Since 2002, in Brazil the income rose for poor people, roughly stayed the same for the middle-class and slightly decreased for the rich. Poor voters, therefore, had every reason to reward Lula with their vote for another term, while well-off Brazilians intended to penalize him. Table 1 below shows the results of the 2006 election, by comparing them with those of 2002:

Table 1: Election results of 2006 (Corrêa)
We can see that overall, the percentage of the population that voted with Lula roughly stayed the same. However, the proportion of where the votes come from changed significantly. This was largely due to the support received from the poor voters and from the seniors. Some papers find a correlation between the penetration of the Bolsa Familia in 2006, and the shift of votes towards Lula as compared to 2002: the correlation coefficient was found to be at 0.621, showing that Bolsa Familia had a very big impact in Lula’s reelection. (Hunter & Power, 2007)
3.7. The reelection from a median voter theorem viewpoint
Considering all the facts and statistics illustrated above, it becomes much more simple to apply the median voter theorem to Lula’s reelection in 2006. First of all, the problems that we presented in section 3.1 regarding the incongruence of the two limitations are now solved. We can now say that voters had a unidimensional policy space, because of the large impact that Lula’s “rebranding” played on the Brazilian political scene. In effect, in 2006, voters had to choose between the then-president, which was in favour of protecting the poor, and the other adversaries who criticized his decisions while being in opposition. Also, since we assume that Bolsa Familia was Lula’s biggest “trademark” (and not his party, which was his main asset in 2002), we view voters as having single peaked preferences: either they would like more social policies like Bolsa Familia (but not limited to; in the end, Lula also increased minimum wage, offered benefits to seniors and raised public pensions), or not.
Although it has been showed that the median voter theorem fails under multidimensional voting (Hindriks & Myles, 2006), using the simplifications listed above and focusing on the Brazilian elections of 2006, we can say that the outcome of the election was in favour of the median Brazilian voter. For the most part, this was due to the targeting social policies implemented by Lula, which aimed at a much larger share of the population (one fourth), than his previous target group.
Critics of the Bolsa Familia impact on Lula’s re-election chances affirm that if Lula had not implemented Bolsa Familia and if he had kept everything else constant, he would have kept his electoral base of 2002 and get re-elected by those people. (Corrêa) In fact, judging by the electoral profits and losses displayed in table 1, such critics may have a point. However, our opinion is that the senior voters would not have voted Lula again, because their main reason for voting him in the first place was that they thought he would bring a fresh change from Cardoso’s rule. With no effective change that might have affected seniors, we believe that this electoral group would have sought change somewhere else, namely in Lula’s adversaries.
4. Voting Process for Dilma Roussef
Noting that the profile of the electorate PT in 2010 resembles that of 2006 and not 2002, Pereira, Shikida and Nakabashi (2011) sought to investigate which aspects are most relevant in determining the fraction of votes for Dilma in the municipalities. The evidence presented denote that programs of this nature cash transfer programs, regardless of the success in maintaining economic growth and reducing poverty and inequality, have been key to electoral success.

As we can see in Figure 1, there is a high correlation between the percentage of votes obtained by Lula in 2006 and the vote share of Dilma in 2010 in the municipalities. There are two possibilities for this similarity in the voting pattern: the same social and economic factors that led to the less developed regions to vote for Lula were important in the election of Dilma (such as economic growth and great expansion of income transfer programs) and /or Lula’s electorate remained loyal, and full support of the president's nominee for the Workers’ Party favored the vote for Dilma, regardless of economic conditions and benefits received.
Figure 1 shows the scatter diagram between the proportion of valid votes obtained by Dilma in the second round of presidential elections in 2010 (vertical axis) and Lula's vote share in the second round of the elections of 2006 (horizontal axis), which indicates the importance of the support of then President Lula and the belief that their candidate would carry on their economic and social policies.
Figure 1: Proportion of valid votes obtained by Dilma in the second round of presidential elections in 2010 andLula's vote share in the second round of the elections of 2006

We can observe in graphs 2 and 3 a high correlation between the proportion of votes for Dilma Roussef and the level of development of municipalities (education and Human Development Index)
Graph 2: Municipal proportion of valid votes for Dilma in the second round (2010) and education

Graph 3: Municipal proportion of valid votes for Dilma in the second round (2010) and Human Development Index (HDI)

The three diagrams suggest that the 2010 election scenario strongly resembles that of 2006: less developed municipalities (with the lowest HDI in 2000) and lower average schooling of the population over 25 years yielded a higher proportion of votes for the candidate.
In those municipalities with higher levels of education and development, the desire for alternation of power prevailed , reflected in the increased proportion of votes for the other candidate, Jose Serra. As graphs 4 and 5 show, we can observe a clear correlation between the proportion of families benefiting from BolsaFamilia and the vote for Dilma.
Graph 4: Municipal proportion of valid votes for Dilma on the second round (2010) and the scope of Bolsa Familia

Graph 5:Municipal proportion of valid votes for Dilma on the second round (2010) and the resources of BolsaFamilia (per capita)

According to this study, the scope of the Bolsa Familia apparently played a leading role in the continuity of the Workers Party in power for the 2010 elections.
Bibliography

Bohn, S. R. (2011). Social Policy and Vote in Brazil: Bolsa Família and the Shifts in Lula’s Electoral Base. Latin American Research Review , 46 (1), 54-79.
Ministério de Desenvolvimento Social e Combate à Fome (MDS). Retrieved on 11/12/2011 from http://www.mds.gov.br/bolsafamilia
Brasilino, L. (2006). As perspectivas do BolsaFamília. Retrieved on 11/12/2011 from http://www.brasildefato.com.br

Corrêa, D. S. Conditional Cash Transfer Programs and their Electoral Effects. Illinois: University of Illinois.
Hindriks, J., & Myles, G. (2006). Voting. In Intermediate Public Economics.
Hunter, W., & Power, T. (2007). Rewarding Lula: Executive Power, Social Policy, and the Brklian Elections of 2006. Latin American Politics and Society , 1-30.
Johannsen, J. (2009). Conditional Cash Transfers in Latin America. Problems and Opportunities. Inter-American Development Bank, Social Protection and Health Division, Manila.
Pereira, A. E. G., Shikida, C., & Nakabashi, L. (2011). Análise introdutória dos determinantes da eleição de Dilma, 99-107. Retrieved on 18/12/2011 from http://www.economiaetecnologia.ufpr.br/boletim/Economia_%26_Tecnologia_Ano_07_Vol_026.pdf#page=105
Samuels, D. (2006). Sources of Mass Partisanship in Brazil. Latin American Politics and Society , 48: 2 (Summer): 1-27.
World Bank (2005). The Impact of Bolsa Família on Education and Health Outcomes in Brazil.
Zucco, C. (2009). Cash-transfers and voting behavior: An empirical assessment of the political impacts of the Bolsa Familia program. Princeton University.

http://www.ipc-undp.org/publications/mds/46P.pdf http://www.icsw.org/global-conferences/Brazil2006/papers/rosane_mendonca.pdf http://www.scielo.br.proxy.library.uu.nl/scielo.php?pid=S0104-62762011000200008&script=sci_arttext