1How the working-class vote does not depend solely on the predispositions which each individual develops and which are linked to sociodemographic characteristics. The institutional and political context in which an election takes place also influences voting behaviour ; this context can be more or less propitious to the expression of these predispositions at the polls. [1] The way in which the vote is organised in practical terms and the “electoral facilitation systems” on which this organisation depends, the nature of the party political offer, and the techniques used to mobilise voters also go some way to explaining voting practices. [2] Following an analysis of voting behaviour among the working-class conducted in France using a localised, multi-method approach, [3] we took the decision, in spring 2010, to establish a Franco/Brazilian research project. The ultimate aim of the project was to observe the election process in an area on the outskirts of Brasilia. [4] We sought to refine existing understanding of the role played by the political and institutional context in the voting behaviour among sectors of the population characterised, both in France and in Brazil, by their lack of education, low income and a limited interest in the political sphere. Optional versus compulsory voting, political choices which are structured around the notion of a left-right spectrum and around political symbols which don’t have the same resonance across both democracies ; “politics time” [5] which alters to a greater or lesser extent the usual ways in which neighbourhood communities operate, and different electoral transactions : all combine to shape contrasting political landscapes between the French and Brazilian suburbs which may help us to better understand the voting logics at work. The two rounds of voting in the majority of the main elections which punctuate Brazilian political life were all to be held on the 3^rd and 31^st October 2010 : first and foremost, the presidential election, but also elections for state governor, senatorial elections and elections to the federal parliament (Federal MPs), [6] the state parliaments (“state” MPs, or “district” MPs in the case of federal districts) [7]. In other words this set of elections provided a particularly favourable setting in which to conduct a study in electoral sociology.

A neighbourhood at the bottom of the socio-residential ladder

2We chose to focus our research on the neighbourhood of Estrutural. [8] The choice to use this very specific area as the focus of our research was made on the basis of several criteria. We selected the area first and foremost on account of the fact that the position which it occupies in the capital’s socio-residential sphere is relatively similar to that occupied by Cosmonautes, a district on the outskirts of Paris which we have been studying since 2002. Estrutural is actually located on the edges of Brasilia itself, 30 minutes by coach from the “centre” of the city. It is one of the poorest neighbourhoods in this urban area. The district developed at the end of the 1960s around the largest state-owned landfill site in the Federal District, known locally as lixao (literally, “large dump”), a pejorative nickname which has long been used to refer to the neighbourhood itself. Like the majority of the thrown-together settlements which form the suburbs of Brasilia, Estrural was first erected in “spontaneous” fashion, by migrants who, operating outside any legal framework, cobbled together informal dwellings, bestowing the appearance of a shantytown upon their urban surroundings. [9] Then, gradually – and in ways which we will later describe given that these continue to colour the dominant voting tendencies here – the area was legally recognised, the makeshift houses became permanent fixtures, shops and public services developed and the neighbourhood adopted the form which it retains to this day. Today, Estrutural is home to at least 25,732 people. [10] The population of Estrutural displays demographic, socioeconomic and cultural characteristics typical of the Brazilian working class. These characteristics can be depicted as statistics using the exit polls which we carried out on 3 October 2010, during the first round of the presidential elections.

Table 1

Sociodemographic characteristics of exit poll respondents (3 October 2010) – Estrutural/ Federal District [11] [12]

Average age 33 Under 26 34.3% Over 56 5.5% Educational qualifications < Primary school 58% Unemployment rate2 19% Income < minimum wage 44.4% Recipients of the bolsa familia3 21.5%

Sociodemographic characteristics of exit poll respondents (3 October 2010) – Estrutural/ Federal District [11] [12]

3As we will see, this data can be considered representative of the voters [13] and thus, given the high turnout, as reliable indicators of the characteristics of the adult population in the neighbourhood. The data reveals a very young population within which the elderly – those in retirement – constitute a tiny minority : 3.8% of the exit poll respondents. [14] In this environment, the level of education obtained remains very low indeed. Only a minority of voters had successfully completed secondary education : 24.6%. Furthermore, in this working-class neighbourhood, scarcely a single voter had graduated from an establishment of higher education : those who had accounted for 0.7% of all voters. Even among the youngest sector of the population, only a tiny minority enjoyed access to higher education : among the under 25s, only 9.1% attended or had attended university. Moreover, among this 9.1%, on the day of the election, not a single respondent was in possession of a full university degree. [15] In this neighbourhood where the population is very young, almost two-thirds of adult residents are either in employment or practise a professional activity (61.9%). Nevertheless, as far as unemployment is concerned, one of the defining characteristics of this working-class environment is the large number of people out of work : the unemployment rate stands at 19% compared to 7% in Brazilian society as a whole. This unemployment is one of the factors behind the low income on which the households in this neighbourhood are forced to survive : 44% of exit poll respondents stated that their “family income” was lower than or equal to the “minimum wage”, [16] or 510 reais a month (around 220 euros). Undoubtedly more significant still is the fact that only 21.8% of those surveyed earned more than twice the minimum wage. Furthermore, those that admitted to earning a salary five times higher than the minimum wage – the equivalent of 1,100 euros a month, accounted for only a miniscule minority of the voting population : 3.5% of respondents. Against this backdrop, a third of voters stated that they were in receipt of government welfare payments. In the majority of cases, the benefit awarded was the bolsa familia, a payment introduced under the Lula presidency, which ranges from 32 to 242 reais a month. It is means tested and awarded on the condition that the family’s children receive an education, and so is a benefit designed to assist Brazil’s poorest families. At the time our exit poll was taken, more than a fifth of voters (21.5%) were in receipt of this new government benefit.

A high level of participation leaning towards “the right”

4Whilst in socioeconomic terms, Estrutural occupies a structurally similar position in the socio-residential sphere as that of Cosmonautes, there are significant differences between the two neighbourhoods in terms of the electoral behaviour of residents. Firstly, the turnout among Estrutural’s residents is high and relatively stable : 83.4% on 3 October 2010, and consistently above 80% in previous elections.

Table 2

Election turnout (1^st round) – Constituency n^o 9 (combines the neighbourhoods of Guará and Estrutural – Federal District 2002-2010

Year Turnout 2002 85.39% 2006 82.31% 2010 81.43%

Election turnout (1^st round) – Constituency n^o 9 (combines the neighbourhoods of Guará and Estrutural – Federal District 2002-2010

5Moreover, these high turnouts were calculated based on voters recorded on the electoral register, a number which includes almost all of the adults in the neighbourhood, since inclusion on the electoral register, like voting itself, is optional from 16 to 18, but compulsory between the ages of 18 and 65. All citizens in this age group are required to go to the regional Electoral Court in order to request a “título” – failure to do so constitutes a breach of the law. Moreover, the abstention rates include individuals who, although they may have failed to vote, nevertheless travelled to the polling station closest to them on election day to “justify” their behaviour. Just like their fellow citizens who had chosen to vote, they had queued in line in order to avoid the penalty which abstentionists risk if they have not been “excused”. In Estrutural on 3 October, 3% of those we interviewed were, in fact, “excused” voters (see Figure 2) who had just paid to travel to the polling station without casting a vote. One such voter even explained to us that, as he lived in the Federal District, he had not in fact been able to obtain an abstention permit, as such documents were reserved for those registered at a polling station outside the constituency. This voter, a 30 year-old man, explained that he had not been able to afford the bus ticket to travel to the polling station where he was registered, some tens of kilometres outside Estrutural.

6For those familiar with the French context and with neighbourhoods in which non-voting tends to become the electoral norm, [17] understanding the mechanisms behind this mass mobilisation of a population which is nevertheless very young, poor, and lacking in both qualifications and political awareness offers a fascinating opportunity for research. [18]

7It is also interesting to attempt to understand why this neighbourhood, among the poorest in Brasilia, votes overwhelmingly in favour of candidates at the “centre” or “right” of the local and national political spectrum. Such behaviour is in stark contrast with that observed on large urban housing estates in France, where the population votes largely for left-wing candidates. Conversely, in Estrutural, votes for the candidates of the Workers Party (PT) are always clearly in the minority. This is particularly true of local elections. When electing a candidate for governor of the Federal District of Brasilia, residents systematically award a clear majority to representatives of the Roriz family : to Joaquim Roriz until the most recent elections, and then in the 2010 elections – following his decision not to run as legal proceedings against him would have most likely seen him declared unable to stand by the Electoral Court, [19] – to his wife, Weslian Roriz (see Table 4). Yet Joaquim Roriz ran for election under the banner of the social-Christian party (PSC), a small organisation to the right of the political spectrum. [20]

8The same voting behaviour was observed during the elections to the district parliament. It was the candidates from parties classed as “centre” or “right-wing” that won the largest number of votes, leaving the PT trailing in their wake. Once again, the 2010 election saw a power struggle which worked severely to the detriment of the PT. The presidential election produced similar results. As with the local elections, the PT candidate once again received only a tiny minority of the votes. [21] Consequently, the aim of the research project developed in Estrutural was to enable us to identify the forces at work behind this mass mobilisation of the working class in favour of candidates whose parties are classed on the “centre” or “right” of Brazil’s political arena.

A receptive environment for measurement by questionnaire

9As has often been observed, when researchers in France attempt to use a tool such as a closed-question questionnaire to gauge opinion, they often come up against the refusal of those surveyed to cooperate. [22] This is particularly the case when the aim of the questionnaire is to measure political opinions and behaviour. Thus, the percentage of those willing to participate in exit polls rarely exceeded 50% of the total number of voters in the ten polling stations studied between 2007 and 2010 as part of the PAECE project. [23] To take a single example – the Cosmonautes estate – the exit poll response rate here stood at 51.4% during the first round of the 2002 presidential election and at 49.1% during the first round of the 2007 election. The response rate impacts heavily upon the quality of the data produced : indeed, the nature of this data is determined to a large extent by the social and political characteristics of the voters. Consequently, we were unable to reach entire sectors of the electorate through an exit poll during our research here. As a result, in Cosmonautes, older members of the electorate and right-wing and Front National voters are thus almost entirely absent from our database, having refused to participate in our survey.

10Within the context of a place such as Estrutural, these problems do not appear quite so acute. For researchers accustomed to surveying voters at French polling stations, several differences emerge almost immediately, from the moment initial contact is made with voters on the ground.

11Firstly, in Estrutural we noted that very few voters refused to respond to our exit poll : only 20% of those approached as they left the two sections of the polling station where the poll was conducted refused to do so. Furthermore, only in very few cases indeed did such refusals seem to be as a result of the sociodemographic characteristics of the respondents, as the table below, which gives a breakdown by age of all voters who entered the polling station, demonstrates particularly clearly. This table was compiled by a researcher who stood at the exit from the moment it opened until the moment it closed.

Table 3

Breakdown of voters who refused to respond to the exit polls by estimated age – Estrutural/ Federal District [24]

Age group Refusal to respond Exit poll respondents 16-20 19.4% 25.3% 21-30 29.3% 22.5% 31-40 23.5% 24.2% 41-50 13.8% 17.6% Over 50 2.6% 10.4% Unidentified 11.4% 0

Breakdown of voters who refused to respond to the exit polls by estimated age – Estrutural/ Federal District [24]

12Young women were more likely to sidestep our questions than men. They often came to the polling station with babies or young children, whom they claimed they wished to attend to once they had fulfilled their civic duty. Several voters, employed at the market, on the “large dump” [25] or as housemaids, also told us that they wished to return to work after voting. However, one voter, when asked why he refused to participate, stated that he wished his vote to remain secret. Broadly speaking, voters readily agreed to respond to the poll, without us needing to insist. In contrast, responses to the exit polls conducted in France were largely obtained by exerting a certain degree of pressure on voters. Furthermore, in Estrutural, we noted that very few of those surveyed refused to respond to the majority of the intensely personal questions asked, namely those regarding political tendencies or electoral choices. Conversely, in France, we often found people unwilling to answer these questions, including those who had agreed in principle to participate in the exit poll. [26] Admittedly, in Estrutural, questions relating to professional activity appeared to produce a certain degree of embarrassment and, on several occasions, we were given cause to doubt the responses given. It was abundantly clear that a number of those interviewed were looking to escape the stigma attached to those whose job was linked to the state-owned dump. However, what was striking in this neighbourhood was that voters were happy to tell us about their voting tendencies : for example, only 6.3% of respondents answered “don’t know” or refused to respond when asked to state how they had voted in the elections for governor, and only 4.9% of respondents did the same when it came to their choice for president. This very small percentage of people refusing to respond is all the more remarkable given that in Estrutural, the questionnaires were issued and completed by the researchers themselves. This went against recommended practice in France, where it is advised that voters complete the exit polls themselves using a system which allows voters to respond without revealing their voting choices or tendencies. In Estrutural, it was decided that researchers would complete the surveys, given the low average level of education among voters which, as we have seen, is a defining feature of neighbourhoods such as Estrutural. Indeed, 41.9% of the Estrutural residents who responded to our exit poll stated that they had failed to complete compulsory secondary education. A further 4.5% considered themselves to be (almost) illiterate.

13This particularly marked readiness among voters to participate in the survey ensured that our results were representative of all of the voters who entered the polling station studied and even, we may assume, of all the voters in the neighbourhood as a whole. Comparing these results with the election results recorded at the ballot box confirmed this hypothesis (see Table 4).

Table 4

Representativeness of the first round exit polls – Estrutural/ Federal District

Results recorded in the two sections of the polling station studied (nos 281 and 285) Choices declared in the exit polls (non-responses excluded) Vote for the governor Weslian Roriz (PSC) 59.5% 56.1% Agnelo Queiroz (PT) 23.1% 24% Toninho (PSOL) 5% 4.5% Eduardo Brandão (PV) 2.6% 3.8% Blank/ null and void ballot papers 9% 5.2% Presidential vote José Serra (PSDB) 36.9% 33.8% Marina Silva (PV) 31.9% 31.7% Dilma Rousseff (PT) 23.4% 24% Plínio Sampaio (PSOL) 0.69% 1.3% Blank/ null and void ballot papers 6.5% 4.2%

Representativeness of the first round exit polls – Estrutural/ Federal District

14In addition to these issues relating to the representativeness of our data, it is also possible to view the administration of questionnaires at the exit of a polling station as a research experiment which reveals a great deal about the relationship between the sectors of society concerned and the political sphere. As such, the exit poll response rate in itself constitutes a social phenomenon triggering a variety of hypotheses and interpretations. In Estrutural, the warm reception received by the questionnaire as a measurement tool and the willingness of voters to interact with researchers can be explained, in part, by the sociodemographic characteristics of these voters. If we transfer knowledge acquired in the French context to that of Estrutural – in the form of hypotheses – it becomes possible to surmise that the young age of voters worked largely in favour of the successful execution of questionnaires. The fact that those conducting the surveys were also very young – the majority were students of political science at the University of Brasilia – undoubtedly also served to facilitate interaction with voters.

15The nature of the Brazilian political landscape is undoubtedly also among the factors which make voting choices “admissible” in public. Unlike in France, the presidential elections of 2010 were characterised by the relative “ideological” proximity of candidates and the absence of political forces which, like France’s Front National, would carry a particular political stigma. The three major candidates for the presidency, who together received 98.8% of the vote, [27] belonged to “government” parties on the centre or the centre-right of the electoral sphere : Dilma Rousseff was the candidate for Workers’ Party (PT), which, at the time of the election, had already been in power for eight years ; Marina Silva was the Green Party (PV) candidate ; and José Serra from the Brazilian Social Democracy Party (PSDB), a “moderate” party at the “centre” of the political spectrum. The personalities of these candidates also contributed to the similarities in the presidential electoral offer : both Dilma Rousseff and José Serra were active opponents of the military dictatorship ; the PT candidate going as far as to wage armed struggle against the regime ; José Serra was well known for having developed a policy to safeguard the interests of the working-class during his time as Minister (of Planning and Health) under the Cardoso government. Marina Silva was also a Minister in the Lula government between 2003 and 2008, responsible for the environment. [28] However, as far as values were concerned, throughout the 2010 campaign Silva – former leader of the PT, and the woman who under the dictatorship had defended the rights of workers in Amazonia – adopted positions traditionally associated with the conservatives. For example, she spoke out in favour of traditional family values and expressed a certain degree of hostility towards abortion. Moreover, it is highly likely that such positions explain at least part of her success in working-class areas during these elections, where abortion is a crucial topic of debate.

16Similarly, there were no such “divisions” in the choice of governor, or there were at least no factors which made voting choices impossible to admit to in public. In these elections, the two key candidates were Agnelo Queiroz for the PT and Weslian Roriz, for the Social Christian Party (see Table 4 above). The latter is a party to the right of Brazil’s political spectrum, but on the national level endorsed Dilma Rousseff. This position clearly illustrates the complexity of the partisan system and the incredibly fluid nature of the left/right divide. [29] Indeed, this divide proved to be of little practical use in explaining why politicians adopted certain positions, or, as we will see, to explain the pattern of voting. [30] The PT candidate Queiroz and his opponent Roriz together received 79.9% of the votes at Federal District level and 82.7% in the two sections of the neighbourhood which our survey examines. The “admissibility” of an individual’s choice for the post of governor was undoubtedly further bolstered in these elections by an entirely contextual factor : the overwhelming popularity of Weslian Roriz in the district of Estrutural. In terms of the “admissibility” of voting preferences, if we translate the dominant rationale in France to Estrutural, it should have been the votes for this right-wing candidate who was standing for election in a somewhat delicate situation – following her husband’s decision to withdraw from the race – which were more difficult to declare. However the enormous popularity of the Roriz family in this “makeshift” town which it had actively supported meant that it was easy for respondents to admit to having voted for her in our poll. As regards the negative reputation blighting the PT among vast swathes of the neighbourhood’s population, this did not appear to jeopardise the “admissibility” of votes for these candidates, even if one voter confessed when leaving the polling station that the neighbourhood “did not take kindly” to voting for the Workers’ Party.

17Finally, the fact that it was possible to obtain voting choices through the use of a questionnaire can also undoubtedly be explained by an even more decisive cultural factor : the relationship between voters and voting secrecy. In France it is widely accepted that secrecy has traditionally been a key element of legitimacy where voting and the expression of voting choices is concerned. [31] This particularly applies to the confines of the polling station where it would be entirely incongruous and even transgressive to publicly reveal voting preference. In this context, the practice of conducting exit polls as voters leave polling stations inevitably constitutes a departure (considered overly radical by some voters) from the school of thought which sees voting as a ritual in which the notion of secrecy takes centre stage. The Brazilian context appeared markedly different, and, in terms of our focus here, better suited to the gauging of electoral choices through a questionnaire. The works of Moacir Palmeira and Marcio Goldman on rural Brazil have demonstrated how, in electoral campaign situations, all of the voters from a given village in the 1980s and 1990s might become activists during “politics time”. [32] Within both the neighbourhoods and the families that the researchers observed, voters took ownership of the new cleavages which emerged as a result of supporting certain candidates, and publicly declared them. Indeed, as we will see later in more detail, during the election period politics occupies a more central and physically visible position in Brazil than it does in France, both up to and inside the polling station, where it is in no way unusual to come across voters proudly adorned in T-shirts or sporting stickers bearing a candidate’s name. A number of these voters thus resemble “sandwich-boards”, actively publicising a local candidate in the very place where votes are cast (see Figure 1).

Figure 1

Voters in polling stations sporting T-shirts and stickers bearing candidates’ names (3 October 2010, Estrutural polling station)

Voters in polling stations sporting T-shirts and stickers bearing candidates’ names (3 October 2010, Estrutural polling station)

18There is no doubt that this context serves to make the public admission of voting choices a banal occurrence, including within the polling station itself, and that it therefore acts, among other things, to further increase the “admissibility” of voting choices. [33]

Compulsory voting and the mobilisation of a neighbourhood

19Undoubtedly, the most striking observation where electoral behaviour is concerned is the high turnout recorded both in Brazil as a whole and in an overwhelmingly working-class neighbourhood such as Estrutural, where 83.4% of those on the electoral register went to the polls during the first round of the elections held in October 2010.

20However, we should clarify right from the outset that Brazil’s 2010 election campaign displayed none of the characteristics typical of “battle elections” which André Siegfried previously identified in an entirely different context as triggering the mobilisation of particularly high numbers of voters. [34] As previously noted, none of the three candidates in any way embodied a departure from the political status quo. All three had previously served under “moderate” governments. Furthermore, right from the very beginning, the election had been scripted as a contest which was to culminate in the ratification of the choice made by President Lula who, at the height of his popularity, had named Dilma Rousseff as his successor. Consequently, the degree of uncertainty surrounding the outcome of the election appeared minimal. Thus neither the political context nor the electoral framework were likely causes of the mass mobilisation of Brazil’s voters.

21It should also be emphasised that Estrutural is home to a population whose social and cultural characteristics largely work to the detriment of electoral participation. In the United States or France, voters in comparable neighbourhoods would shun the polls in record numbers. Indeed, as we have seen, two of the defining features of this environment are the young age of residents and the low level of education received (see Table 1). It therefore comes as no surprise that voters in this district mostly expressed very little interest in politics, including those who cast a vote on election day.

Figure 2

Voters in Estrutural wait for up to 1 hour 30 minutes to cast their vote (3 October 2010)

Voters in Estrutural wait for up to 1 hour 30 minutes to cast their vote (3 October 2010)

Table 5

Level of interest expressed in politics – Estrutural/ Federal District

None 44.3% A little 39.3% A lot 16.4% Total (n= 287) 100%

Level of interest expressed in politics – Estrutural/ Federal District

22Here we find ourselves faced with one of the classic paradoxes of electoral sociology and one which is even more marked in working-class areas : the need to understand the electoral participation of individuals who, even when emerging from the polling station itself, claim more often than not to have no interest whatsoever in politics. [35] And, moreover, the need to understand why those on the neighbourhood’s electoral register flock to the polls in their droves as soon as doors open at 8am, before queuing in line – sometimes, as we saw firsthand, for up to an hour and a half – to cast a vote which is almost always “indifferent” to the contest underway. [36] Indeed, there were many voters, who, once out of the polling station, were forced to rummage around for the “cola” (flashcard) distributed by candidates in the run-up to the election in order to be able to tell us who had received their vote just a few minutes earlier. [37] Voting day provided example after example of many voters’ detachment from the electoral process. For instance, a 43 year-old woman who worked as a home-help and who had recently moved to the area from the poverty-stricken state of Nordeste would tell us, with a sideways glance at her husband, that she had just voted for Agnelo Queiroz by mistake, having not been able to remember the name of the candidate her husband had advised her to support once she got to the voting machine. When asked about his own vote, the man confessed that he had not cast his vote out of conviction, but rather because “someone in his family knew the candidate from church”.

23Of course, the compulsory nature of the vote is pivotal in explaining the high turnout regularly recorded in the neighbourhood. [38] Effectively, voters are “obliged” to cast a vote at the polling station where they are registered. If their situation makes it physically impossible for them to do so – for example, they are travelling in a place far away from that in which they are registered – voters must travel to another polling station in order to “explain themselves”. Failure to do so puts voters at risk of incurring a whole range of sanctions, the most common of which are a four reais (1,70 Euro) fine, and a ban on holding either a passport or applying for a post in the civil service. These sanctions are meant to be imposed following three successive abstentions.

24A series of questions asked as part of the exit polls conducted during the first, and particularly the second, round of elections related to the obligation to vote. The aim of these questions was to better understand both voters’ perception of this requirement and how far it influenced their decision to go to the polls. The first thing to emerge from the responses to these questions was that only a small majority of voters wished the vote to remain compulsory : 52.2% versus 44.8%. The second striking result was the direct correlation between this response and that relating to the level of interest in politics : the lower a respondent’s interest in politics, the more vehemently opposed he or she was to the compulsory vote.

Table 6

Interest in politics and opinions on compulsory voting – Estrutural/ Federal District

Are you interested in politics ? Do you think the vote should remain compulsory ? Total Yes No Not at all 46.5% 53.5% 100% A little 58.9% 41.1% 100% A lot 70.2% 29.8% 100% Total (n = 286) 55.2% 44.8% 100%

Interest in politics and opinions on compulsory voting – Estrutural/ Federal District

Note : Chi-square 8.840 (sig. 0.012). [39]

25From these results we might surmise that the less politicised are more sensitive to the compulsory nature of the vote insofar as it forces them to cast an “indifferent” vote without any incentive or motivation to do so other than the wish to avoid punishment. If this is the case, the responses to the exit poll would reflect voters’ subjective perception that they had been transformed into “captive” voters by virtue of the legislation in force – a hypothesis which it would be useful to confirm at a later stage through in-depth interviews. Whatever the reality behind the figures, during the second round of the vote, we sought to gain a clearer picture of how voters perceived the sanctions linked to this obligation : it emerged that a large majority of respondents believed that the sanction incurred following abstention was fairly minimal. Indeed, 41.2% viewed it as “non-existent” compared to only 32.4% who believed it was “significant”.

26These differences in perception are due to the subjective dimension of this type of evaluation. However, they are also, to a large extent, the result of the fact that Brazilians themselves are far from conversant with the rules underpinning the obligation to vote. Furthermore, many voters, and certainly many abstentionists, are also unaware of the range of sanctions imposed and the exact sums of money they are liable to pay should they fail to vote. Our exit poll left no doubt that this was the case. Indeed, one closed question in the poll asked voters to estimate the size of the fine which abstentionists were at risk of incurring (Table 7).

Table 7

Estimated size of the fine – Estrutural/Federal District

Less than 5 reais 58% 6-50 reais 20.8% 51-100 reais 5.2% Over 100 reais 4.8% Don’t know/ left blank 11.2% Total 100%

Estimated size of the fine – Estrutural/Federal District

27Only a small majority (58%) guessed the correct amount : “less than 5 reais”. More than one in ten voters (11.2%) said they did not know the size of the fine. Consequently, almost a third of voters (30.8%) overestimated the financial penalty imposed for abstention, sometimes making guesses that were wildly off the mark : 5.2% believed the fine to be somewhere between 51 and 100 reais, and 4.8% even thought it exceeded 100 reais. In a neighbourhood where 44% of voters stated that their household income was below the minimum wage of 510 reais, it is likely that the economic cost which they wrongly associate with abstention acts as a deterrent powerful enough to bring to the polls those who, lacking any ethical or political incentive to vote, might otherwise have stayed away.

28It should also be noted that it is among the sector of the population least interested in politics that we find the largest number of voters likely to be discouraged from abstaining by the financial penalty which they would incur. Indeed, it is within this sector of society that the sum total of those who were unaware of the size of this penalty (9.5%) and those who overestimated it by a considerable margin (15.3%) is at its highest : approximately a quarter of the least politicised voters. There is thus no doubt that it is within this sector of the population that the deterrent provided by the financial penalty is most likely to be crucial in preventing abstention, by virtue of the fact that many believe it to be higher than it actually is.

29However, it would be overly simplistic to view the obligation to vote solely in terms of an individual and rational calculation which sets the costs of abstention (financial or administrative) against the costs of casting a vote, even if this vote is entirely “indifferent”. [40] Indeed, there is no doubt whatsoever that the fact that the vote is a legal obligation accompanied by sanctions has an impact on the normative dimension of voting practices. It is highly likely that the perception of the vote as a moral obligation is reinforced ; for it is undoubtedly the notion of electoral “duty” (rather than “right”) which is uppermost in the dominant social representation. However, this normative relationship with “good citizenship” and the intensity of the obligation as experienced subjectively by voters are difficult to establish via questionnaires. These are psychological matters, complex in nature and therefore difficult to measure.

30Furthermore, and perhaps most importantly, to consider the effects of the obligation to vote solely as an individual calculation of costs and benefits would be to overlook the fact that this obligation also has a significant impact on the intensity of mobilisation campaigns. [41] Indeed, to a certain extent, this obligation produces a climate ripe for the cultivation of a “captive electorate” or, in other words, an electoral market extended to all citizens of working age (the vote being optional for those over 65), whatever their social status, level of interest in politics, or place of residence. This in turn means that in order to mobilise voters, and ultimately win them over, candidates must target all strata of society – not forgetting the youngest, the poorest, the least educated and the least politicised, including those who live in makeshift “constructions”. Failure to do so sees their chances of emerging victorious from the polls considerably reduced. [42] Thus, in a typical cycle of circular causality, the obligation to vote creates the conditions for an intense, Brazilian-style campaign, which targets all sectors of society ; and this in turn fuels mass participation – the original purpose of the compulsory vote.

31What is immediately striking to an observer accustomed to French campaigns is the heavy physical presence of electoral messages : these appear on the countless small boards placed along the roadside, on the fronts of certain houses and even sometimes on the bodywork of cars, decorated from boot to bonnet in the colours of an election candidate.

32The production of promotional signs reached its peak on voting day when, as we have seen, they even appeared on voters’ bodies, on T-shirts or on stickers, sometimes stuck straight onto the skin. In Estrutural’s main square, home to the Sunday market, it was impossible not to know that elections were underway : the ground was littered with flyers (see Figure 9), and a number of stalls displayed electoral slogans out front. Such was the case, among others, of a stall re-selling pirate CDs, which particularly captured our attention because of the volume of the songs pumping out of its speakers. From a distance, we took the stall to be one backing Weslian Roriz, the main local candidate to the post of governor. However, as we drew closer, we noticed embarrassment and suspicion on the faces of the sellers who feared that we might be police officers, before confessing that they had been paid to fly the candidate’s electoral flag (see Figure 5).

Figure 3

Fronts of buildings displaying candidate numbers in the neighbourhood of Estrutural (3 October 2010)

Fronts of buildings displaying candidate numbers in the neighbourhood of Estrutural (3 October 2010)

Figure 4

Vehicles customised as part of the election campaign (Estrutural, 3 October 2010)

Vehicles customised as part of the election campaign (Estrutural, 3 October 2010)

Figure 5

A stand selling pirate CDs decked out in the colours of Weslian Roriz (Main square, Estrutural, 3 October 2010)

A stand selling pirate CDs decked out in the colours of Weslian Roriz (Main square, Estrutural, 3 October 2010)

33These election agents, paid and involved in the campaign to varying degrees, constituted one of the organised mechanisms of electoral mobilisation which candidates put in place. Agents were recruited from within all social milieux, and particularly from the poorest areas. We encountered such agents whilst administering our questionnaires, with a number of voters explaining to us that they had already “worked” for a party in exchange for a specific sum of money.

34On the eve of the election, we walked through a tiny makeshift “construction” known as Matadouro, home to around one hundred residents. The construction is made of sheet-metal dwellings and is located on the far edges of Brasilia. As we passed through Matadouro, we had the opportunity to observe such mobilisation in action (see Figure 6). In this environment, a sort of micro-shantytown where poverty is particularly rife, a parked car was decorated entirely in the colours of one of the candidates for parliament. Its driver, a local resident, had been hired as an election agent for the duration of the campaign.

Figure 6

A small rural “construction” mobilised in support of Jacqueline Roriz (2 October 2010)

A small rural “construction” mobilised in support of Jacqueline Roriz (2 October 2010)

35This small “construction” [43] provides an example of the way in which informal electoral mobilisation systems can be embedded into the social structures of a neighbourhood and how electoral transactions can be organised around exchanges far removed from the categories of specialised politics. Indeed, Renata, the (informal) representative of this “construction” – the leader of the residents’ association – described how residents had been mobilised and the way in which the majority had allegedly voted. It appeared that a meeting had been held with the majority of the families (80% of the 65 families resident according to the young woman) to discuss the upcoming elections. According to Renata, it had been decided that Matadouro would vote for Weslian Roriz, in light of the commitments that the candidate had made to the “construction”. She had, for example, guaranteed occupants that the construction would be legalised and that plots of land would be granted to those already in residence. The links between these residents and the Roriz family are thought to go back several years, more specifically to the time when, at the beginning of her first term as district MP, one of Weslian’s Roriz’s daughters, Jaqueline, worked personally to ensure that running water was installed in the district. During our discussion with her, Renata showed us her mobile phone several times, making clear to us that she was in direct contact with Weslian Roriz’s campaign team. She maintained particularly close contact with Jacqueline – who was once again standing for election to the Federal Parliament – and had been in conversation with her until just before our arrival. It is easy to imagine the advantages of the position for Renata, who was able to negotiate her influence in the elections with the candidate and, in return, establish her authority over a group thanks to the direct contact she enjoys with one of the most influential figures on the local political scene.

36There is no doubt that the mediation efforts of these “community” leaders given the role of “election agent” (cabos eleitorais) are crucial to electoral participation in working-class areas of Brasilia. [44] It is thanks to the daily involvement of these agents in the mobilisation of local residents that the chains of contact which link ordinary voters to candidates appear to take on tangible form. These vertical relationships built on the provision of “protection” and “services” in exchange for votes are consolidated through these collective microstructures of electoral mobilisation. [45] Our observation of the electoral campaign in Estrutural also made it clear that this type of group-based logic was present in transactions which form the basis for residents’ political choices. Whilst all evidence suggests that the bonds established between voters, local leaders and the region’s political elite almost resemble personality cults focusing on certain key figures, the more collective dimension of these processes – that which is conducted through the networks of cabos eleitorais – must not be overlooked.

37More than anything else, the 2010 election campaign resembled a “campaign of proximity”. Recent changes in legislation at national level placed more stringent conditions on the organisation of large meetings and banned the (musical) performances regularly funded or staged free-of-charge by the candidates. Such developments appeared to help reinforce the localised dimension of the campaign. Markets, small local associations, churches and, in particular, informal meeting places appear to have been the favoured sites of electoral mobilisation in Brasilia’s working-class neighbourhoods. Consequently, both in Estrutural and in the Matadouro “construction” discussed earlier, the campaign appears to have been conducted through intermediaries firmly rooted in the networks which structure voters’ everyday lives. These election agents, likely to occupy a position of leadership within their residential area, notably organised rallies which travelled to meet the voters, uniting neighbours behind a local candidate (particularly one who was standing for parliament). Thus, the electoral choices made in these neighbourhoods do not appear to be made solely as a result of the monolithic logic of classic clientelism underpinned by dyadic relationships ; such logic is not sufficient to take account of the complexity of the many chains of micro-relationships which become embedded in local networks, be these religious or partisan, friendship or community-based, and which support electoral mobilisation at a local level.

An institutional context which favours the cult of personality

38The intense focus on individual personalities which characterises the electoral transaction was immediately apparent from the posters, the boards put up along the roadside or the t-shirts to which we have already referred. In the majority of cases, these “props” contained only two or three pieces of information : the name and voting number of the candidate and possibly his or her photograph (see Figure 7). In some cases, three names, three numbers and three photographs appeared together if candidates had formed an alliance, with one standing for president, one for governor and one for parliament. Whether one or three names were present, what was interesting to a French observer was the almost total absence, right across the board, of “political” reference points, partisan labels, slogans and indications of belonging to either the “majority” or the “opposition”. [46]

Figure 7

A name, a number, a photograph : personality-centred campaign materials (outskirts of Brasilia, 3 October 2010)

A name, a number, a photograph : personality-centred campaign materials (outskirts of Brasilia, 3 October 2010)

39This communication strategy involving the use of the candidate’s name alone (along with the voting number which he or she had been assigned) is the result, to a large extent, of the rules and procedures in place for the casting of votes. Indeed, the ways in which votes are counted and candidates elected encourages candidates to promote themselves as individuals rather than as representatives of their political party. The fact is that the rules currently in force create a climate in which individual candidates are pitted one against the other, even when they appear on a single party list. Success in parliamentary elections depends on the fulfilment of two conditions : the candidate must appear on a list which has obtained a sufficiently large number of votes, and within this list, must be among the candidates elected with the highest number of votes. What is important to note here, is that the likelihood of a candidate being elected does not depend on the position which that candidate occupies on the party list. As a result, candidates have little incentive to promote their party – a campaign strategy which would also benefit their closest rivals. Instead, it becomes more effective for them to promote themselves as individuals.

40The never-ending stream of messages which focus exclusively on the name of the candidate and his or her voting number is also the result of a more directly cognitive phenomenon connected to the way in which votes are cast. In Brazil, individuals vote by typing the number of their chosen candidate on the keyboard of an electronic machine.

Figure 8

Vote box and voting machine (Estrutural, 3 October 2010)

Vote box and voting machine (Estrutural, 3 October 2010)

41The voting number comprises two digits for the presidential election and the elections for governor, three digits for the elections to senate, four digits for elections to the federal parliament (federal minister) and five digits for the elections to the state or district parliament. This system means that voters are obliged to “memorise” five different numbers. To give one example, on 3 October 2010, a voter who had voted Serra in the presidential elections, Weslian Roriz for governor, Cristovam in the elections to senate, Joaquim Roriz for the post of federal minister and José Edmar for the post of district minister was forced to enter the following four numbers : “45”, “20”, “123”, “3320” and finally, “45101”. After typing in a number, the candidate’s picture appears on the screen and the voter is asked to validate his or her vote. Even if, traditionally, the major parties have always been allocated the same numbers (PT candidates have a number starting with the digits “13” and PSDB candidates a number beginning with “45”, for example), the operation nevertheless remains complex and likely to create a great deal of confusion. It is therefore easy to understand why, in elections with five-digit voting numbers, the key challenge for candidates is to make their voting number and photograph as accessible as possible ; including and in particular around and on the premises of the polling station on election day. This undoubtedly explains why the ground of the dirt track to the school hosting the Estrutural polling station was covered in leaflets in the hours leading up to the vote (see Figure 9). And it also most certainly explains the abundance of voting messages on the bodies of certain voters over the course of the election day. Furthermore, from this perspective, the fact that 14% of respondents stated that they were entirely unable to tell us which way they had voted in the district assembly elections only a few minutes earlier is a perfect example of the difficulties which voters experience in getting to grips with the specifically cognitive dimension of the complex electoral framework. This inability was undoubtedly the result of having to remember – albeit for only a few minutes – the name and therefore the number of the person for whom they had just “chosen” to vote. And, as we have already pointed out, among those who did answer our question, many needed to find their “cola” in order to be able to do so.

Figure 9

The ground in front of the polling station littered with flyers (Estrutural, 3 October 2010)

The ground in front of the polling station littered with flyers (Estrutural, 3 October 2010)

42These intensely personality-focused forms of communication prove perfectly adapted to the characteristics of the electorate, and particularly appropriate to the sector in which political engagement is minimal – the majority of the population in this working-class neighbourhood. Indeed, in Estrutural, the cognitive tools belonging to the field of specialised politics are only relevant for a minority of voters. This is in no way surprising among a population in which 83.6% of respondents have “no” or “very little” interest in politics and in which 56% state that they “never discuss politics” or do so “rarely, and only during elections”. [47] This lack of political engagement can be gauged from the responses to questions on political orientation.

Table 8

Position of voters on the political spectrum – Estrutural/ Federal District [48]

Very left-wing 2.8% Quite left-wing 5.6% Centre 6.9% Quite right-wing 14.6% Very right-wing 6.6% Neither left nor right-wing 39.6% Don’t know/ No answer 24% Total (n=228) 100%

Position of voters on the political spectrum – Estrutural/ Federal District [48]

43As regards their position on the political spectrum, it is revealing to note that 24% of the voters interviewed left this question blank (which, as we have seen, was not a strategy to conceal their views) and the largest group in the poll (39.6%) comprises those who declared that they were “neither right nor left-wing”. Not only is it true that, in Brazil, these political categories do not have the same importance as they do in the French political system, but this first result leads us to emphasise that, for at least two-thirds of the voters in this workingclass district (63.6%), a sense of belonging to these political categories could not have been behind their votes. Furthermore, the result confirms the reactions elicited by the questionnaire, when certain voters claimed not to understand the meaning of “right” or “left”.

44The question on identification with a political party also allowed us to confirm that a large section of the electorate does not cast votes based on what, in other electoral contexts, constitutes a key cognitive shortcut. [49] Over half of respondents (56.7%) claimed to have no affinity with any of Brazil’s political parties. This was in spite of the fact that the closed nature of the question itself encouraged voters to state an affinity, albeit with the possibility that they would do so in artefactual fashion. Moreover, among those who responded to the question, many asked the interviewer to indicate which party acronym corresponded to “Marina’s party” or the “Roriz party”, information of which they appeared to be entirely unaware.

45In conclusion, if the transactional forms of reasoning at work in the personality-centred approach to voting – deeply embedded in the social fabric and more clientelistic than “political” – prove so effective in this context, it is because they are adapted to both the cognitive resources of a significant proportion of the population in a neighbourhood such as Estrutural and to the institutional rules which favour this type of transaction.

The Roriz family : effective generation of votes in the least politicised sector of the neighbourhood

46In the context of Estrutural, the electoral strategies developed by the Roriz family have proven particularly effective. In broader terms, these strategies form part of a larger web of relationships based on “protection” and “exchange” which, like any sustainable clientelistic system, maintains a section of the population in thrall. [50]

47It was whilst residents built and settled permanently in this neighbourhood [51] that the bond between them and the Roriz family was established. Initially, in 1997, the local government, at the time controlled by the PT, opposed the presence of this “construction” on the edges of the rubbish dump and called in the armed forces to remove illegal occupants. The move sparked a bloody clash and resulted in a significant number of victims. Once elected governor, Joaquim Roriz earned the lasting support of the neighbourhood’s population when he took steps to legalise the “construction” – support which was bolstered further by his decision to transfer plots of land to the occupants. Thus he became a “protector” and “benefactor” for the residents. However if this protection-based relationship is to last it must be constantly maintained – not least because of generational renewal (57% of voters are under 35) and because turnover in the neighbourhood is high : as the exit poll conducted on 3 October revealed, almost half of voters arrived less than ten years ago.

48Maintaining this relationship involves a certain amount of memory work, reminding residents of the role which Joaquim Roriz played, along with other local mediators, [52] in ensuring Estrutural became a permanent fixture. However, conversely, it also involves reminding voters of the hostile attitude shown towards the neighbourhood by the PT. The “museum of blood” [53] which was opened in the neighbourhood, and which we would hear about from the local hairdresser who had worked for a time as an election agent for the Roriz family, constitutes the clearest illustration of efforts to cement the legacy of that family in the collective consciousness. The story of Estrutural’s origins is still repeatedly told here and the association of the PT with traumatic events serves to maintain a highly negative image of the party.

49However, the Roriz family’s attempts to retain their popularity also involve the (re)production of classic clientelist transactions. Such transactions were discussed in connection with the small neighbouring “construction” Matadouro, whose residents were grateful to Jaqueline Roriz for having installed running water and for her subsequent promise to formally transfer plots of land to families living in the district should she win the elections. In Estrutural itself on the day of the vote, families told us that they had “always” voted Roriz out of loyalty towards the community’s benefactor, whilst some found it difficult to contain their emotion at the mention of his name and spoke of the “love” which they felt for the family. [54] Finally, there is no doubt that Joaquim Roriz’s “personable” style also contributes to his popularity and his image as a man close to the people, and serves to endow him with all the attributes of a political “protector”.

50In order to better identify the social and cultural factors which increase the chances of a vote being cast in favour of the Roriz family in Estrutural, we conducted a regression analysis. The aim of such an analysis was to enable us to verify the validity of the observations, hypotheses and theories developed up to that point. It would allow us, in particular, to confirm that the electoral strategy which, on a local level, ensures the Roriz family’s victory at the polls, is most effective among the youngest sector of the neighbourhood’s population, and amongst those most lacking in educational capital or interest in the political sphere.

51In order to conduct this analysis, we deliberately limited the first round of the elections to a duel between the two key candidates for the post of governor : Weslian Roriz and Agnelo Queiroz of the PT. These candidates alone had received 80% of the votes cast by those who responded to our questionnaire. Only 4.5% of respondents voted for the candidate in third place. Reducing the contest to one between the two key candidates made the development of a statistical model easier, and in particular facilitated the use of a binary logistic regression analysis. Furthermore, it allowed us to more clearly identify the social and cultural factors at work during the vote. Given that we were first and foremost researching the factors behind a vote in favour of Weslian Roriz, we decided to use Queiroz as the reference mode in order to facilitate the interpretation of the results.

52We used the following as independent variables : age (divided into four groups) ; sex ; length of residence in the district ; level of education received (limited to two values : below or above primary education) ; religion (limited to two values : voters who declared themselves to be “of no religion” and those who claimed to have a faith) ; whether voters defined themselves as “left-wing” ; level of political engagement (limited to two values and defined on the basis of interest declared in politics and the frequency of political discussions) ; their assessment of the Lula government in terms of what he had done to improve living conditions in the neighbourhood ; and the moment at which they had chosen their preferred candidate for governor.

53The Chi-squared of the model and the pseudo R-squared demonstrate that these variables provide a significant contribution to explaining votes cast in favour of Weslian Roriz.

Chi-squared of the model

54,261 significant above a threshold of 0,000

Chi-squared of the model

Pseudo R-squared

Cox and Snell 0,221 Nagelkerke 0,318

Pseudo R-squared

54The regression analysis enabled us to confirm the hypotheses and analyses developed to date. The majority of the explanatory variables integrated into our regression model meet requirements of statistical significance. [55] This particularly applies to religion, the assessment of the Lula government and level of education, but also to certain age groups, levels of political engagement and the moment at which a voter selected his or her chosen candidate for governor. Others displayed a risk of error which was slightly too high to be classified as “significant” – duration of residence in the neighbourhood (significance : 7.4%), left-wing or not left-wing (significance : 10%) – but offered results interesting enough to be discussed subject to validation at a later stage using larger “samples”.

55It comes as no surprise that a voter’s negative assessment of the Lula government in terms of its contribution to improving the personal living conditions of residents (only 28% of voters held such an opinion) considerably increased the likelihood of voting in favour of Weslian Roriz : her popularity rating in this case increased to 9 against 1. The likelihood of votes going to this candidate also depended on the age of voters. The categories among which this likelihood was highest were the 16-25s and 26-35s. Level of education attained was also a highly significant variable. Effectively, a low level of education also significantly increased the chances of a voter favouring Weslian Roriz : her popularity increased to 3.3 among voters who had not progressed beyond primary education. Finally, religion also visibly influenced the vote. Indeed, religious affiliation considerably increased the chances of voting for Weslian Roriz : her popularity rating here climbed to 6.8 against 1.

56If we stick with the three major significant sociocultural variables within our regression model, it becomes clear that the sector of the population on which the Roriz family’s electoral strategy is most effective is that comprising the district’s youngest residents (under 36), along with the least educated and those who define themselves as “catholic” or “protestant”. In this context, a lack of religious affiliation or a higher level of education are factors clearly detrimental to the hegemony of these electoral strategies.

Table 9

Binary logistic regression analysis conducted on the vote in favour of Weslian Roriz – Estrutural/Federal District

Variables in the equation Coefficient (Standard errors) Exp(B) Man 0.388 (0.373) 1.474 Ref. Woman 16-25 1.160* (0.576) 3.189 26-35 1.365* (0.642) 3.915 36-45 0.040 (0.528) 1.041 Ref. Over 45 Primary education or below 1.182** (0.409) 3.262 3.262 Ref. Above primary school Religion 1.922*** (0.561) 6.834 Ref. No religion Over 10 years’ residence in the district 0.644 (0.360) 1.904 Ref. Under 10 years’ residence in the district The Lula government has not improved my situation 2.201*** (0.560) 9.036 Ref. The Lula government has improved my situation Not very politicised 1.055* (0.468) 2.872 Ref. Politicised Not left-wing 0.977 (0.594) 2.658 Ref. Left-wing Choice of candidate before the campaign 0.862* (0.377) 2.368 Ref. Choice of candidate during the campaign Constant -5.207*** (1.233) 0.005

Binary logistic regression analysis conducted on the vote in favour of Weslian Roriz – Estrutural/Federal District

Note : * = Significant below 5% ** = Significant below 1% *** = Significant below 0.1%

57The analysis of the other variables included in the regression model allowed us to refine the explanatory model. First of all, it is interesting to note that length of residence in Estrutural could significantly influence voter choice. The longer residents had lived in the district, the higher the likelihood of them voting for Weslian Roriz appeared. Consequently, her popularity rating appears significantly higher among that sector of the population which has been living here for more than ten years : 1.9 versus 1 (with a significance level of 7.4%, or slightly above the thresholds used here). This result appears to indicate that living in the district for a long time and accumulating a significant quantity of “autochtony capital” [56] are factors likely to compel voters to support the Roriz family. Doubtless we should interpret this as one indicator of strong probability that voters will belong to networks on which the Roriz family’s power depends, as well as an indicator of the increased likelihood that these same voters have been involved in the history of this “construction” and in the political allegiances which go with it.

58Finally, we must analyse the issues related to politicisation. In line with the analytical framework previously outlined, the regression analysis confirms that it is among the least (or non-) politicised sector of the population (the numerically largest sector) that the Roriz family’s electoral strategy appears to be most effective. The popularity of Weslian Roriz is here 2.9, against 1 in the most politicised sector of the population. Logically and inevitably, this politicisation is particularly unfavourable to this candidate when the voter is left-leaning. Conversely, if voters are right-wing or politically neutral, the chances of them voting for Weslian Roriz appear to increase : her popularity rating here stands at 2.7 against 1 (with a significance level of 10%). This appears to further confirm the hypothesis according to which the section of the electorate on which the Roriz’s power depends comprises right-leaning voters and those with little or no interest in the political sphere. Finally, it is interesting to note that Weslian Roriz wins a large majority among voters whose voting choice had already been made when the election campaign got underway. This appears to provide further evidence of this lasting bond of loyalty which ties a section of the electorate to candidates who they deem authentic political “protectors”. This result can also undoubtedly be explained by the fact that, here, the PT candidate is an outsider who has to rely on the electoral campaign in order to compensate for his outsider status, and promote both his candidacy and manifesto among electors.

59In sum, the regression analysis appears to validate the analytical framework which we initially offered with a view to accounting for the electoral successes of the Roriz family in this overwhelmingly working-class area on the outskirts of Brasilia. If the Roriz family emerge from the polls with a clear majority, it is due to the fact that they manage to rally behind their banner the “right-leaning” section of the politicised minority, and in particular due to the fact that the electoral strategy which they have developed has proven particularly effective among the youngest, least educated, most religious and least politicised residents in the neighbourhood. Indeed, this strategy, which combines religious influence, clientelistic “protection” and a personality-centred leadership style has proven itself perfectly adapted to the cognitive frameworks and socially constructed expectations present among this section of the population. In a neighbourhood such as Estrutural, the positive effects of such a strategy are all the stronger given that youth, religious belief, and a lack of education and political involvement are characteristics typical of the vast majority of both residents and voters.

60***

61Despite the lack of interest in politics shown by the majority, voters in Estrutural are highly mobilised as a result of the electoral laws. However, this brief foray into a Brazilian neighbourhood also makes it clear that they are not the only ones affected by this obligation to vote ; it also impacts political professionals. By assigning the same electoral value to the vote of those least politicised as that attached to the vote of the most educated, the compulsory vote explains a number of the distinctive features of electoral campaigns in Brazil. In a neighbourhood such as that in which we conducted our research, at election time the candidates were everywhere and presented proposals particularly adapted to residents’ day-to-day concerns. They were physically present, organising face-to-face meetings with voters. They were also present through a whole host of electoral signs – posters, flags and flyers literally covering pavements, roads, walls and markets. Taking over urban public spaces, these signs target passers-by, with the aim of reactivating their political loyalty as much as making it easier for them to cast their vote. Finally, candidates ensured their presence through a stream of election agents. Indeed, on voting day, all those who wore stickers or signs on their front or back expressing loyalty to a particular candidate, were playing such a role.

62Whilst entirely informal in nature and specific to the particular context, this mechanism for the mobilisation of citizens nevertheless demonstrates that this is an electoral context which favours the participation of the least politicised. This in turn makes it clear that a legislative framework which declares the vote compulsory is entirely effective in Brazil. Working-class neighbourhoods in France such as Cosmonautes appear political deserts in comparison. What strikes a Brazilian observer a week away from the 2010 presidential elections in the district of Saint-Denis, is the extent to which in France nothing signals that elections are round the corner : aside from a few metal boards erected as part of the official campaign, it is impossible to guess that preparations for a vote are underway. And whilst certain residents did admit to having opened their door to the representatives of a left-wing party who had come to encourage them to vote, the bewilderment expressed at this intrusion of an activist into their homes highlights the extent to which, in France, we are still a long way from the situation observed in Brazil, [57] where the working-class make the electoral contest their own.