Introduction
Until recently, Latin American electors have generally been treated as herds without any will of their own, easily manipulated by unscrupulous politicians. What distinguishes some of the current historiography is the questioning of such traditional assumptions. The growing interest in the electorate, therefore, has gone hand in hand with a concern for the study of its values and aspirations, with the aim of exploring the extent to which people were able to take advantage of the power conferred by the vote.Footnote 1
The symbiotic relationship between elections and fraud in pre-democratic regimes has been recently reassessed by historiography and political science. In an inspiring review of the literature on elections at the turn of the twentieth century, Posada-Carbó (2000) describes important developments in political historiographical perspectives. Scholars have re-examined common depreciative interpretations of Latin American experiences in the eighteenth and nineteenth centuries and granted them some of the thematic dignity that had been denied in the past. Inquiries about political practices, forms of political representation, clientelism, the role played by public opinion, the struggle for local and national power, as well as the dynamics of elections and voting behaviour are at the core of this change in approach.Footnote 2
Consider, for example, the debate on two interconnected themes that are analysed in this article: electoral fraud and political participation. Regarding the first, the novelty is the redefinition of the concept of gobierno elector, that is, the bureaucratic-administrative control over the electoral process through a network of agents acting at different levels, ‘from the president and state governors to local commissioners and intendants, councillors, justices of the peace, police delegates, superintendent registrars, and tax collectors’.Footnote 3 This concept, which is rooted in the classical analysis of the Argentine case by Natalio Botana, does not question the view that the government held control over the several stages of the electoral process from voters’ registration to the counting of votes, but reinterprets the dynamics and broader consequences of the whole process. The control over the electoral machine is the result of a complex operation where local elites constantly negotiate the organisation and management of the electoral stages. Substantively, it means that ‘the election results were not simply imposed. Political loyalties, clientelistic networks, the cohesion of the faction in power, the exchange of votes for favors, all these elements were put to the test as the election was approaching. Public and private sphere began to overlap and mix’.Footnote 4 Put this way, the issue of the governmental dominance over the electoral bureaucratic machine calls for a review of its most direct argument about electoral corruption. This is why Antonio Annino argued that the ‘intensity of the rhetoric on fraud was a symptom of the instability of the system of social influences’.Footnote 5
As one can see, the issue of fraud leads to the theme of political participation and, ultimately, of citizenship. According to some authors, electoral fraud was a way of doing politics at that time, because it promoted local gatherings of individuals who would have been alienated from the representative system otherwise.Footnote 6 , Footnote 7 Samuel Valenzuela acknowledges that, despite the fact that fraud was common in Chile in the nineteenth century, elections fostered the crystallisation of party identities and loyalties, as well as the participation of different segments of the civil society, ‘people who were politically active and militant, both men and women’.Footnote 8 It is revealing that, for some scholars, participation in elections was viewed as a collective and public act, never merely individual, determined by a complex network of politicians and their clientele.Footnote 9 The argument is not limited to Latin America. Richard Bensel stated provocatively that in the United States ‘the act of voting was profoundly democratic […], illiterate and ignorant men, steeped in poverty and lacking any other claim on social respectability, were not only permitted to cast their tickets, they were enticed, cajoled, treated and blessed as they did so’.Footnote 10 In Britain, Frank O'Gorman argues that voters were not dependent on employers before the first Reform Act of 1832, and that their relation was not based on servility and humiliation, but rather on deference and the exchange of mutual benefits.Footnote 11
However, none of these interpretations led to reviews of common depictions of the Brazilian republican regime between 1889 and 1930. The Constitution ratified in 1891 established direct elections for president, deputies and federal senators every four, three and nine years, respectively. From 1891 until the end of the regime in 1930, 13 elections for the federal legislature were held, uninterrupted by coups or arbitrary changes in election days. Despite the stability of electoral cycles, the literature frequently describes electoral corruption and physical violence as common practices at that time.Footnote 12 The electoral machine, comprising the stages of voters’ registration, the act of voting, the counting of votes, and the formal certification of winning candidates, has been labelled as a ‘fraudulent machine’.Footnote 13 A similar term, used in Argentina, defines this mechanism of control as ‘bureaucratic fraud’.Footnote 14 Even Eduardo Posada-Carbó, who questions the classic view of Latin American elections in this article's introductory epigraph, does not place Brazil among the cases requiring a reassessment. For the author, Brazil has followed a different path, as demonstrated by Richard Graham's studies, which ‘reinforce[s] the traditional stereotypes of Latin American politics’, and Murilo de Carvalho's study ‘on Brazilian citizenship, innovative from other angles, also reiterates traditional views on the electorate’.Footnote 15 It is symptomatic that Victor Nunes Leal's work, published in 1949, is still the classic reference for studies of elections during the republican period.Footnote 16
The main aim of this article is to review conventional interpretations of republican elections and to bring the Brazilian case closer to other Latin American experiences at the turn of the twentieth century.Footnote 17 In particular, the authors are interested in the study of electoral practices that reveal the way parties and individuals effectively participated in elections during the Brazilian First Republic. Moreover, this case study is used to draw insights for a better understanding of the general relation between electoral fraud and elections in pre-democratic political regimes. The Brazilian case indicates that fraud resulted from struggles among parties to control what is henceforth called the ‘electoral bureaucracy’. This interpretation is built upon analyses of formal complaints filed by defeated candidates in 11 elections for the Chamber of Deputies between 1899 and 1930 in 495 electoral districts.Footnote 18 These are official documents regulated by law, presented to the parliament in order to corroborate calls for election recounts from candidates who questioned its fairness.Footnote 19 The content of these formal accusations of electoral irregularities allows the understanding of some informal rules of the electoral game.Footnote 20 Figure 1 shows the proportion of districts where at least one fraud complaint was filed in each election year.
Figure 1. Percentage of Contested Districts, by Electoral year (1899–1930)
One piece of information is particularly important: on average, 56.8 per cent of districts were subject to some allegation of fraud, and all of them were subject to at least one complaint between 1899 and 1930. Such a high frequency attests to the relevance of formal complaints, an important instrument of protest in the hands of politicians. Figure 2 breaks down the data by state, distinguishing them in terms of occurrence or non-occurrence of fraud complaints.
Figure 2. Percentage of Contested Districts, by State (1899–1930)
Two features in the graph need to be emphasised. First, formal complaints were filed in all of the states which shows that the authors’ analysis is not biased for reasons of geographical or temporal concentration of formal complaints.Footnote 21 Notice, however, the contrast between states with a high frequency of fraud complaints (Bahia, Pernambuco, Federal District, and Rio de Janeiro) and those with a low frequency (São Paulo, Minas Gerais and Rio Grande do Sul). This variation is due to differences in the political situation observed in each state. In Bahia, Pernambuco, the Federal District and Rio de Janeiro, the process of institutionalisation among dominant parties was full of difficulties, affected by fierce disputes between factions and political forces.Footnote 22 In all of these cases, the frequent use of formal complaints, and, thus, of electoral fraud, is not surprising. Weak parties acting in these states did not dominate the political arena the way they desired, and were not equipped to do so. Diversely, the strength of republican parties in São Paulo (PRP), Rio Grande do Sul (PRR) and Minas Gerais (PRM) is well-documented, and they are commonly described as the best-organised parties in the electoral arena due to their effective capacity to control the political market.Footnote 23 Their monopoly of the electoral process allowed them to constrain the opposition and restrict political competition to party members, which obviously reduced their incentives to file complaints against electoral results. This explains the high number of formal complaints in some states, even if their chances of success were minimum.Footnote 24 That is, filing complaints was not an isolated act of defeated candidates, but a component of power struggles between the government and opposition groups.
Below, information is presented from formal complaints about the stages of the electoral process, from voters’ registration to the counting of votes and the declaration of results.
Parties, Factions and the Formation of Electoral Lists
The examination of formal complaints reveals that the nomination of official candidates was the first politically relevant moment in the electoral process. Compared to other Latin American countries, the Brazilian First Republic was characterised by the non-existence of national parties. Each of the 20 Brazilian states and the Federal District was represented in the National Congress by a distinct state party. Elections were carried out under the supervision of state governors, who favoured the victory of loyal candidates, to the detriment of the opposition.Footnote 25 In academia, this unique political configuration diminished the interest in analyses of political conflicts within states, and many authors say that the period was characterised by representative stagnation, ‘the almost absolute lack of party competition’, and ‘the freezing of competition in the states’.Footnote 26 , Footnote 27
However, political collusion during the process of candidates’ nominations are frequently reported in formal complaints and reveal the occurrence of counterintuitive events, keeping in mind the common pessimistic views of that political experience. In a report on the list of candidates presented by the Republican Party of Minas Gerais in the election of 1912, the newspaper Gazeta de Notícias stated that ‘the challenge of building a list is that it is a difficult task – because everybody claims his place’.Footnote 28 A few years later, a candidate, Helvécio Coelho Rodrigues, acknowledged that the republican peace in the state of Piauí had been disrupted by the mere organisation of the official list. In his words,
Some names have been indicated by party leaders after numerous manoeuvres to prevent the split that eventually ensued at the heart of the party; I do not desire to discuss the merit of the issue, but it is my intention to emphasise that the divergences did not have a political or economic character related to the life of the state; no, absolutely not! They had solely a personal character, and the split occurred only because it was impossible, humanly impossible, to allocate four seats to five deputies.Footnote 29
According to several accounts, political factions usually clashed during the nomination of party lists at various deliberative instances and with the involvement of several actors. These clashes were not limited to the executive committees of state parties, where governors interfered more directly. There is strong evidence suggesting that the process of nomination took place also at the local level. In some cases, there are explicit ‘nominations by municipalities’Footnote 30 or decisions based on local recommendations, as Freire (2000) observed occurring in the federal capital.Footnote 31 Hence, the final decision was not always consensual and frequently led to party splits, as reported by Salvador Felicio in 1903, when he justified his decision to run for the next legislature:
I did not have, indeed, the good fortune of being included in the list presented by the executive committee of the Republican Party of Minas Gerais, which recommended names of great honour, and whose merit I am pleased to endorse. My exclusion, however, could not lead me to hesitation or frustration. Being a son of the district, residing there, knowing its traditions, and living together with its people, I know how exceptional it is in terms of morality and independence. Official candidatures have always been received with suspicion and repugnance there. […] The city of Diamantina, the district seat, […] would rather feel disappointed with the refusal of its nominee by the party committee. And its most conspicuous citizens from all political leanings wanted to bring this [my] name to the list. It was not legitimate for me to decline the mission, even if it overwhelmed my forces.Footnote 32
This statement illustrates how important it was for state parties to control municipalities. It also indicates the nature of the selection mechanisms used by parties. Candidates’ social and family ties were not sufficient to guarantee their inclusion in the official lists. A similar impression stems from Eduardo Ramos’ response to a formal complaint filed against him at the sixth district of Bahia:
[…] when someone ventures in electoral struggles with some chances of success, it is unavoidable that he previously articulates a plan, an ideal, a programme; and when this programme perishes, or when the aspirant does not cultivate any ambition that justifies his admission among the ruling classes of his country, it is, at least, imperative that he joins a party.Footnote 33
With this argument, he questioned the real chances of his opponent: ‘[…] how did he want to get elected? Did he trust perhaps his family bonds, or on the simple condition of being born in one of the parishes of the vast 6th district?’Footnote 34
Essentially, the politician's response entails two conclusions. The first questions the chances of success of independent candidates lacking a party's support. The second discredits the power of family ties, which do seem to not be adequate to guarantee access to party lists. The autonomous actions and local strengths of individual candidates are, therefore, less relevant during the electoral process. It cannot be claimed that candidates’ connections to the territory are irrelevant, since these connections are frequently exalted in formal complaints. Closer attention needs to be paid to intra-party negotiations in order to understand the way the electoral market and the elaboration of party lists worked.
Independently of the political environment observed in each state, the data reveal that parties held the monopoly of representation and coordinated interactions between central and local instances. Although depreciative views of national representative institutions were widespread, and that parties ‘as the expression of electoral preferences were not valued [anymore]’,Footnote 35 state parties, even if not well organised, were the most effective instrument for allowing candidates access to parliament.Footnote 36
Voters’ Registration
Two aspects of voters’ registration, a crucial pre-electoral stage, are commonly emphasised in the literature. On the one hand, it can be said that the First Republic was characterised by limited suffrage.Footnote 37 On the other hand, it is also commonly claimed that the process of voters’ registration was prone to manipulation, as people who were already dead or disqualified from voting took part in the electoral process more than once, to the detriment of those who were legitimately qualified.Footnote 38 Indeed, there have been many reports of forged registrations, conducted secretly in favour of individuals loyal to the government, while excluding all others, everything under the supervision of committees in charge of reviewing registers. Reports mention the registration of people who had emigrated,Footnote 39 or even of the deceased who, ‘rose from their coffins to stuff electoral rolls’.Footnote 40 It is interesting to notice that legislators did attempt to reduce the occurrence of fraud. After the electoral reform of 1916, local committees were dissolved and the prerogative of registering voters was transferred to state court judges. An obvious consequence was that parties concentrated on enhancing their influence on the judicial system. Judicial misconduct included the irregular registration of minors and illiterate rural workers. Such practices were also manifest in appeals presented to court judges to protect citizens’ political rights.
Besides the numerous tactics of control and types of fraud during this pre-electoral stage, two other relevant facts are revealed by analysis of formal complaints. First, the process was conducted by ruling parties, through the actions of their local bosses,Footnote 41 men capable of controlling the electoral bureaucracy, particularly the local boards responsible for registering voters. Governors intervened indirectly, by transmitting ‘recommendations to the Juntas de Alistamento (voters’ registration boards) to obstruct voters’ eligibility’.Footnote 42 As stated by Telarolli, ‘in municipalities, where social relations are determined by the economic structure, everyone – and there are not many – with a modicum of competence for the public service is committed to politics’.Footnote 43 In this regard, ruling parties did not find it difficult to control this stage of the electoral process, virtually eliminating any chances of minorities attaining representation. Therefore, political competition was suppressed before elections actually took place.
Establishing who was eligible to vote was a good opportunity to influence candidates’ performance beforehand. When pursuing this strategy, parties granted voter ID cards only to their electoral corrals, while denying them to the opposition.Footnote 44 Substantively, the right to vote was a less relevant issue than the ability to obtain a voter ID card. After all, only those who possessed these cards could actually vote. This explains why parties strove to ensure that their voters brought their ID cards to polling stations, even allocating their own agents for this purpose, and, if possible, also aiming to restrict the participation of the opposition. This interference is illustrated by the fact that the names and addresses of party agents assigned to take care of voters’ registration were usually published in sympathetic newspapers.
A second aspect revealed by formal complaints refers to the way the registration process worked. Reports indicate that it was perceived as a collective process, operated by parties at the local level. They registered whole groups of voters, going from door to door and inside their own places of residence.Footnote 45 Complaints did not question the exclusion of enfranchised citizens, since individuals, on their own, had little value, but rather the exclusion of entire groups of people, frequently called ‘our friends’ and distinguished between ‘opposition voters’ and ‘supporters of the government’.Footnote 46 , Footnote 47 During the 1912 election in Pernambuco, one of the defeated candidates, having obtained only two votes in the city of Villa de S. José da Coroa Grande, attached a document to his complaint where he alleged that ‘in the estates of my friends and mine we have three hundred and thirty-three voters’.Footnote 48
For all of these reasons, there does not seem to be much support for the classic idea of passive and politically apathetic citizens. On the contrary, two factors explain their pattern of behaviour, in contrast to arguments presented by the literature. On the one hand, onerous bureaucratic requirements discouraged a significant number of eligible voters to spontaneously apply for voter ID cards.Footnote 49 In order to register, applicants needed to show proof of literacy and of age. After 1905, applicants were also required to show proof of residence through a certificate issued by a judicial or police authority in their municipality and, if denied, through a declaration from three merchants or landowners living in the same locality. Still, there was no guarantee of success, as other political obstacles could eventually disqualify the applicant. The reason is that the municipality, being the locus of registration, managed by boards of local government members until 1904 and, later, by court judges and the most honourable tax-payer residents, permitted the direct interference of local politics on this important pre-electoral stage. In other words, the legal norms that established the criteria for the effective manifestation of the right to vote were insufficient, by themselves, to ensure the registration of eligible voters; they rather imposed high individual costs on the few who were determined to move forward in the process. On the other hand, it must be understood that the registration conducted by parties was an opportunity for individuals to obtain benefits and personal favours. Several accounts suggest that registration occurred in exchange for financial resources or was aimed at benefiting some individuals.Footnote 50
It is evident that this pre-electoral stage was when the ruling party first clashed against the opposition. Granting voter ID cards was the result of a process structured and supported by parties in collusion with civil servants in charge of registering people in accordance with the legislation in force. A candidate from Rio Grande do Sul was adamant when defending the legitimacy of the certificates delivered to his allies: ‘if the candidates of our list, as well as the independent candidates belonging to our party, won, this was because the minority […] did not take care of the registration of their voters’.Footnote 51 Thus, as in other Latin American countries, this was a key moment for rigging electoral results. Insofar as the ruling forces excluded opposition voters and ensured 100 per cent of registration of aligned voters, the subsequent stages of the electoral process occurred only to give visibility to the act of voting and to forge legitimacy for the republican representative system.
The Organisation of Polling-Station Committees
Polling-station committees played a central role in the electoral process, because, functioning at polling-station levelFootnote 52 they were in charge of collecting ballots, counting the votes, and issuing a report on voting to the electoral scrutiny board (junta apuradora). The literature repeatedly emphasises formal vices, fraud and manipulation that pervaded all of those activities.Footnote 53 Most of the denunciations in formal complaints confirm theses aspects. They point out problems in the counting of ballots, declaration of final results, issuing of voting reports, voters’ signatures and committee members’ initials, blurred, scribbled, or written by the same person, closing of the poll, and issuing final minutes, to cite only a few types of complaint.Footnote 54
The emphasis on so many irregularities after votes were cast has inhibited a deeper examination of the political disputes for the control of the composition of polling-station committees. The analysis of formal complaints reveals struggles for the formation of unanimous committees, which enabled control over events on the election day. A candidate responded to accusations of fraud in the city of São Gonçalo thus: ‘how could one rig the election while not controlling, as I did not, at least one committee member?!’Footnote 55 In the same vein, another admitted with sarcasm that ‘our generosity towards the challengers was so extreme that, in the municipality of Manaus, we allowed the board in charge of organising polling-station committees to grant the opposition with two effective members and two substitutes in each polling station’.Footnote 56 The strategies devised for controlling the committees ensured the exclusion of opposition voters (by not accepting their voter ID cards) and the admission of supporters whose registration had been previously denied.Footnote 57 Likewise, the opposition's right to denounce frauds was generally suppressed, promoting the fabrication of minutes, as well as reducing of some of the costs incurred by parties, such as those related to vote buying.Footnote 58 , Footnote 59 A unanimous committee also facilitated the counting of votes within the polling station immediately after close of poll. Without the presence of oppositionists, committees could ‘count votes their way’, often adjusting the results to favour some candidates.
It should be noticed that the electoral law also stimulated struggles for the control of polling-station committees. Legal provisions established that the composition and addresses of polling stations should be announced a few weeks before the election day.Footnote 60 Some pre-electoral co-ordination was then required among local political forces. This is illustrated by a speech made by Marcelino Machado, a candidate from Maranhão in 1927:
Emissaries of the ruling party, who, two months before me, were already collecting signatures for the indication of committee members, soon realized that my action five days before 25 January would invalidate all of their calculations. They were strictly instructed to not allow me to assign even one member, and, facing a clear defeat, their boss ordered that they increased from ten to twenty the number of fake signatures in each list. […] Despite all this, out of the 13 polling stations in S. Luiz, I managed to nominate the chairs of five and three members in others, by appealing to the Board of Appeals.Footnote 61
This excerpt identifies the relevant players and the goals they pursued during the organisation of committees. On the one hand, it underscores the importance of party ‘emissaries’, the agents who collected voters’ signatures for the nomination of committee members. Months before the election, parties would publish in their newspapers the list of the committee members who they recruited and, as the election day approached, calls for the nominated members to turn up at the polls.Footnote 62 On the other hand, the previous excerpt directs attention to the location of polling stations. Until 1904, the law did not prohibit the organisation of committees outside the municipality. As senator Lauro Sodré pointed out, this facilitated the electoral success of some candidates, because the common practice was to set up polling stations in remote locations, including the estates of political bosses, ‘as a way of preventing voters from approaching’.Footnote 63
In summary, when the formation of polling-station committees is analysed, it is evident that not only was the contest for its control important, but also that the whole process, managed by local political organs, started long before the election day. After all, the Constitution of these committees followed the same logic of nominations for voters’ registration boards, discretionary choices made by members of the local government and their closest allies.Footnote 64
Voting: Voters’ Mobilisation Beyond Violence and Physical Coercion
In order to understand the dynamics of political representation in republican Brazil, parties’ effective capacity to mobilise and control the electorate cannot be neglected. Most authors depict voters as politically apathetic, commonly absent from the polls, and trivial participants in the electoral process. According to the literature, the role played by rural voters was even more irrelevant, due to coronéis’ capacity to carry out eleições de cabresto (‘halter elections’, in literal translation).Footnote 65 In general, physical intimidation and coercion at polling stations are claimed to be responsible for this pattern of passive behaviour, as voters were impelled to comply for fear of retaliation, a situation aggravated by the fact that the ballot was not secret. Indeed, formal complaints report all kinds of physical and moral intimidation, performed by thugs assigned to discourage electors from voting. Coercion attempts escalated a few weeks before election day. In Pernambuco, to quote candidate Annibal Fonseca, ‘arrests of voters, raids on rural properties, destruction of telegraph stations’, and even ‘police intervention to compel voters’ could be observed.Footnote 66
Despite the pervasiveness of intimidation, formal complaints suggest a political environment considerably more complex. Often, ballots were used as bargaining chips, and voters not necessarily found themselves in a position of subordination or direct intimidation. That is, voters’ dependency was not complete, because politicians commonly adopted other ruses to captivate the electorate. This indicates a relatively dynamic electoral market.
A first clue revealing that the electorate was not completely subject to coercive mechanisms lies in the characteristics of the practices adopted to persuade voters. Formal complaints uncover the existence of electoral campaigns that would have been unexpected during the republican period. As in other Latin American countries, a widely used campaign instrument was the party press, which channelled the flow of information to voters.Footnote 67 Partisan newspapers published calls for registration, lists of registered voters and members of polling-station committees, addresses of polling stations, as well as the lists of party members assigned to distribute party ballots in each polling-station.Footnote 68 Being official mouthpieces for political parties, newspapers published their main decisions, putting particular emphasis on the list of candidates a few weeks before the election. These pieces of information could, of course, work as instruments of voters’ mobilisation. Frequently, candidates published notes introducing themselves to readers along with a short political platform;Footnote 69 and voters, likewise, often signed manifests in favour of some candidacies. It was not uncommon, for instance, to find statements in formal complaints alleging that parties were using their press to discredit opponents. Daily, newspapers recommended that party supporters went to the polls, ‘without switching their vote to any other candidate’.Footnote 70 Moreover, telegrams from governors encouraging support for the ruling party were quite common in periodicals.Footnote 71
Besides the use of the printing press, some excerpts of formal complaints indicate that candidates engaged in typical campaign activities, sending letters and telegrams, visiting municipalities, organising rallies, and even door-to-door canvassing. Some politicians bemoaned the lack of individual guarantees for travelling through the district, as did Estacio Coimbra, a candidate from Pernambuco, who claimed to have overcome such difficulties after resorting to local emissaries.Footnote 72 Another candidate, acknowledging lack of means to deliver favours to local voters, declared he had organised ‘an intense campaign from municipality to municipality over the course of several months’.Footnote 73
Analogous tactics were pursued by Helvécio Coelho Rodrigues during the 1927 election in Piauí, when he approached citizens directly:
I travelled to 26 municipalities, always approaching the electorate publicly, presenting a plan of action, and denouncing the wrongdoings of the government. I have always been well received with general sympathies and did not obtain a larger vote share only because of official restraints and traditional habits among voters, who never met a candidate, but voted for the coronel without knowing why or for what; if the ballot is not secret, it is still cast in the dark.Footnote 74
Note that this quote distinguishes between the practices of the opposition and incumbent candidates. The latter are presented as political actors lacking any effective bond with their constituency, in contrast to aspirants of the minority, who travel to many municipalities across the district to campaign. Although it can hardly be said that the above example is representative of the general phenomenon, the formal complaint in which it was reported went further to demonstrate the self-confidence of incumbent candidates:
The secretary of justice travelled to the rural areas of the state delivering public speeches in favour of the ruling party and published telegrams in newspapers to advertise the support received by telegraph. He summoned working-class citizens several times to his private residence, where he made recommendations and exalted the incumbent party. The secretary of police chaired, on the eve of election day, a meeting at Cinema Olympia, in Therezina, and for about half an hour sought to persuade the audience about the need for party discipline, saying that it was not an infringement to receive ballots at the doors of polling stations.Footnote 75
Enjoying free access to the public administration and the prompt assistance of senior staff, campaigns of incumbent candidates were favoured, and their electoral success safeguarded. However, it should be acknowledged that elections had other costs, as candidates still needed to mobilise voters.
Formal complaints suggest that ruling parties targeted very specific segments of the electorate: their candidates participated in meetings of unions and associations, such as in 1927, when the governor of Amazonas, Ephigenio de Salles, was accused of ‘endorsing his candidates’ in some of those meetings.Footnote 76 João de Figueiredo, questioning the fairness of the 1918 election in the Federal District, claimed that he managed to achieve the support of about 400 working-class voters from the National Press after personally approaching ‘each of them’, which indicated how important it was to mobilise civil servants.Footnote 77 Some candidates targeted a ‘Catholic independent’ electorate.Footnote 78 Candidates’ endeavours to gain the support of some social categories reveal not only that they needed to engage with the electorate, but also some of these categories were particularly influential.Footnote 79
That party oligarchies did not fully control the vote through intimidation is evidenced by the common use of several other instruments of persuasion.Footnote 80 First of all, there was bribery, which ‘[competed] with fraud’.Footnote 81 There are several examples of its use, such as investment in local public works to the benefit of some groups, in this case, loyal businessmen. However, personal bribes are even more frequent in formal complaints. A denunciation published by a newspaper, for example, states that candidate Octavio Rocha Miranda delivered ‘to each one who voted for him a Reais$2 note at their place of residence’.Footnote 82 There was a market where votes were exchanged for ordinary handouts, where some voters asked for a ‘little money’ to pay for their ‘room’,Footnote 83 whereas others exchanged their vote for the relief of their fiscal debt.Footnote 84 Some candidates offered a simple meal on election day.Footnote 85 All of these examples indicate that incumbents’ ability to control the electorate should not be taken for granted, because each vote had its cost. This was the reason why no election was conducted in Itaberaba, Bahia, in 1927, as reported by candidate Arlindo Leoni: ‘The contender was there, and he heard from local bosses that rounding up voters would cost too much, and, therefore, they were incapable of summoning the electorate.’Footnote 86 A related reason in 1903 discouraged coronel Heredia de Sá in the Federal District who, after getting ‘tired of gathering votes, abandoned this activity to focus on rounding up thugs’.Footnote 87
In general, these narratives lead to a subsequent consideration of how voters were mobilised and persuaded. The examination of formal complaints reveals that voting was, at that time, a collective practice, rather than the exercise of an individual right and earlier in this article this was how voters’ registration took place. Things did not change, however, while voting was taking place. In 1927, a contender from Maranhão reported that 657 voters, ‘his friends, were prevented from voting’ and had to travel to the nearest town where ‘they publicly and notoriously voted in the notary public office (cartório) of his political adversary’.Footnote 88 Other candidates spoke of voters who ‘arrived to give testimony of their unsurpassable dedication’, and still another, acknowledging the limitless dedication of their voters, opted for ‘asking his friends, who were almost the entire electorate in that municipality, to cast only one third of their votes for him’.Footnote 89 , Footnote 90 Accounts from official documents filed by defeated candidates were also resolute in denouncing the imposition of constraints on the political rights of some voters. This how 330 voters justified their protest to the authorities in 1927, when they declared that ‘we were prevented from voting for our candidate – Dr. Marcellino – in the most absurd and arrogant, if not stupid and criminal, manner’.Footnote 91
The use of the expression ‘my friends’ is emblematic when procedures are effectively collective. A candidate in the 1899 election admitted: ‘I stood in the election of 31 December with the support of sincere friendships from the times when the electoral truth and the respect for the expression of the ballot box were valued; by friendships forged during 15 years of professional services dedicated to the progress of my home-state’.Footnote 92 These personal ties between the electorate and the candidate shed light on the nature of their relationship. The mobilisation for the support of a candidacy is collective and performed by ‘friends’ co-opted through means that invariably involve bribery and promises of future benefits. In any case, even when co-optation gives place to physical coercion, as frequently happened, this does not change the fact that voting was a collective activity.
We should now underscore that there was only one apparently clear cleavage during the Brazilian First Republic: the logic used to approach and control voters in urban and rural areas were different. Formal complaints suggest that in the countryside, where ‘arbitrary rule [was] the law’Footnote 93 , it was easier to control voting behaviour, as voters were more dependent and displayed a lower level of education.Footnote 94 Ironically, a deputy, whose victory had been challenged by a formal complaint supported by a petition of 102 signatures, said that it was weak evidence, as ‘this noble Commission knows very well how easy it is to obtain such a document from the electorate in the countryside’.Footnote 95 During a statement in the chamber in 1918, candidates Faria Souto and Pereira Nunes described the situation as follows:
[…] as the districts get farther from the cities, as vigilance becomes harder, as surveillance weakens, wrongdoings become more frequent, unanimity becomes an ordinary phenomenon, electoral quotients grow astoundingly, and the electoral rhythm acquires the cadence of numbers aligned in symptomatic uniformity.Footnote 96
An even more sarcastic statement was made in 1912 by Victorino de Paula Ramos, a candidate who believed that ‘the electoral activity becomes astonishingly more intense when one travels from the city to the countryside, that is, when distances are greater and the masses of the population decline’.Footnote 97 As can be seen, portrayals of Brazilian backwardness, an obstacle to its institutional development, a view firmly endorsed by Victor Nunes Leal, are recurrent in the voluminous pages of formal complaints.
Beyond the Bounds of Municipalities: The Role of Election Scrutiny Boards at the District
Once votes were counted at polling stations, each committee sent a report to the election scrutiny board of the district, which reviewed vote counts, issued certificates to the winners, and dispatched a load of documents to the National Congress. These boards were composed of the chair of the municipal government and other political authorities, and they were usually accused of forging final electoral results.Footnote 98 Their role changed over time. The electoral reform of 1916 centralised the election scrutiny process which took place in the state capital instead of each district where it was managed by only one election scrutiny board. Thus, the governor was endowed with stronger powers of influence. Moreover, the reform also altered the composition of those boards. They now consisted exclusively of a federal judge, his substitute and a delegate from the public prosecutor's office instead of the president of the district council.Footnote 99
One feature of this stage of the electoral process needs to be underscored. The formal complaints clearly reveal that election scrutiny boards followed a partisan logic. Accusations filed by defeated candidates targeted the collusion between the ruling party and the governor, who interfered directly on the counting of votes and issue of certificates. During the 1915 election in Amazonas, the governor, who was ‘aided by cronies acquainted with the instruments of fraud’, was accused of falsifying the minutes of polling stations where votes were registered.Footnote 100 These litigations often resulted in election scrutiny boards being split into two parts, each delivering certificates to its own candidates. Some situations were ridiculous. In 1909, different interpretations of minutes emitted by the third district of Rio de Janeiro led to the formation of two subgroups working in the same room, each of them occupying one end of the table and counting the votes in different ways.Footnote 101 In 1915, a split led to two subgroups working in the same building, one of them ‘surrounded by police forces’.Footnote 102 Again in 1915, a split led to the creation of a second board that met in a steamship, and, was therefore dubbed ‘aquatic’ by a contester.Footnote 103 In such situations, the National Congress was responsible for judging the legitimacy of the certificates issued by the boards.
Even after the reform of 1916, the influence of the executive on boards’ activities did not decline. Indeed, federal judges, although not dependent on state governors, were still subject to several forms of coercion and pressure. In 1927, Luiz Guaraná, a candidate for the second district of Rio de Janeiro, described some experiences incurred by the federal judge, who, under electoral law, chaired the activities of the election scrutiny board.
This magistrate was targeted and, after not yielding to compelling arguments aimed at preventing him from upholding the electoral truth, was subjected to even stronger arguments to nullify his action: – his son and his step-brother were imprisoned and those who searched them at the time of the imprisonment lamented that the federal judge was not present. […] The federal judge then decided, as a tribute to the authority of the position in which he was invested, to not be present during the works of scrutiny, where his decisions would not be respected and could, maybe, stimulate mockeries among the rulers of the state.Footnote 104
Once again, these facts attest to the relevance of the electoral bureaucracy and parties’ control over all stages of the electoral process, from voters’ registration to the counting of votes by polling-station committees, and, afterwards, by election scrutiny boards, a privilege aimed at safeguarding their electoral success.
Discussion
This analysis of the Brazilian experience supports some elements of other studies of pre-democratic elections. However, it is necessary to look beyond the classic framework of electoral studies, in which fraud is typically interpreted as synonymous with vote buying and violence against voters. A more comprehensive view of electoral practices at that time corroborates the notion of gobierno elector, which tackles the idea that the government ‘conducted elections and ensured the victory of candidates’ by monopolising the administrative apparatus of elections.Footnote 105 It is fair to say that several analysts acknowledge the importance of this idea during the Brazilian First Republic.Footnote 106 Reflecting on voters’ emancipation, a deputy of that time stated that ‘in modern times, the main enemy of such independence is administrative pressure’.Footnote 107 Nonetheless, control over several stages of the electoral process has been typically associated with fraud. Such an interpretation misses the point that elections were not a mechanism for expressing individual preferences, but rather an instrument to advance the interests of the ruling elites. This perspective forces a revaluation of the way the electoral process is understood. Denunciations of fraudulent practices were not substantiated by violations of political rights, but, rather by the inability of the local elites and parties to control the electoral bureaucracy. Ultimately, as stated by a contester, fraud was a ‘last resort’, or, as asserted by Victor Nunes Leal, ‘one resorts to violence only when other options are sluggish or less effective for reaching the target’.Footnote 108 , Footnote 109 This argument is innovative because it interprets electoral fraud as a mechanism used by parties in their struggle for the control of elections. It validates the view presented by Antonio Annino, according to which the instability of local social relations could explain the occurrence of fraud.
This notion has two strong implications. The first is the review of current explanations for electoral reforms that were supposedly aimed at ending electoral corruption. Recently, Susan Stokes and colleagues presented a sophisticated model that interprets the end of electoral corruption in the United Kingdom and the United States as a consequence of industrialisation and economic growth.Footnote 110 In particular, these two events supposedly affected the relationship between parties and brokers, reducing the centrality of the latter and favouring a movement towards a more programmatic electoral logic. In Britain, some authors emphasised the importance of institutional factors, such as the Corrupt and Illegal Practices Act of 1883, or the electoral reforms that increased the number of voters and the size of constituencies.Footnote 111 , Footnote 112 In Latin America, however, there is an element that deviates from this explanation. The adoption of independent electoral courts during the early twentieth century promoted the judicialisation of electoral disputes and ended parties’ intervention on the management of the electoral process.Footnote 113 In Brazil, for example, the literature is unanimous in considering that the adoption of electoral justice in 1932 was intended to reduce electoral fraud. It can be understood that explanations of the mechanisms created to assure clean and fair elections must take into account the fact that fraud may assume different formats and characteristics. Hence, in Latin America, differently from the British and American example, the restoration of electoral governance ‘does not emerge as a simple by-product of economic modernisation’, but as an attempt to solve disputes between parties over the control of elections.Footnote 114 That is, the study of efforts to end fraud should focus on the attempts to reduce or eliminate what Natalio Botana insightfully identified as the essence of the gobierno elector.
The second implication relates to the role played by voters, which can be divided into two aspects. First, elections had, predominantly, a local dynamic. It can be observed in the electoral legislation of the First Republic, which afforded great powers to the municipal bureaucracy, and in voters’ mobilisation, which occurred primarily within municipal boundaries. This agrees with the view of Antonio Annino, for whom ‘elections have always been experienced in all countries as a local affair, with local rules that could not be bent legitimately by any law from the “center”’.Footnote 115 Second, voting was a communitarian act, representative of a given locality and done collectively. It was not an individual act. As asserted by Raffaele Romanelli several years ago, the idea of the voter as an individual in the nineteenth century ‘is nothing more than an abstraction’.Footnote 116
These two aspects, the local and the community, are central for understanding the evolution of voting practices and the notion of political citizenship in that period. For decades, the legacy of Thomas Marshall led scholars to interpret enfranchisement as a mere concession of rights.Footnote 117 The literature tends to disqualify the low voter turnout between the late nineteenth and mid-twentieth centuries because of fraud and physical violence. It also became critical of the numerous constraints on the right to vote during that period.Footnote 118 Both of these attitudes are overly reductionist. On the one hand, it is assumed that voters’ deference was complete, and that they were entirely subject to the wishes of coronéis and local bosses. However, numerous studies attest to the need to review the idea that the coronel easily controlled a large amount of ‘halter votes’ (votos de cabresto) and was capable of guiding ‘a crowd of voters as someone who tends cattle’.Footnote 119 Basically, voter dependence may not be the only reason leading them to exchange political loyalty for future protection. As pointed out by other studies, clientelistic relations were imperfect and unstable, following the logic of formation and perpetuation of clientelistic networks, but not necessarily within the private sphere.Footnote 120
Additionally, and more troublesome, the history of the vote is generally presented as the history of universal suffrage, promoting the perception that the access to the ballot paper automatically implies the emergence of a voting process based on individual liberties. The main problem of this literature is the insufficient attention paid to the electoral practices of the nineteenth century, which were based on localism and communitarianism. Instead of insisting on the view that political rights were being gradually expanded, focus should be on the transition from the local to the national vote, and from the collective to the individual vote. In the first case, the issue is the formation of a party system that gradually evolves in terms of organisation and structure in order to compete at national level; the second case refers to the loss of parties’ control over voters.Footnote 121 Statements that these two correlated processes depend on the expansion of the suffrage are overly simplistic. We should remember that many countries had already adopted inclusive electoral systems in the nineteenth century, although this cannot be interpreted as evidence that voters were free.Footnote 122 In Spain, the adoption of universal suffrage in 1890 did not make the system less stable, because it ‘depended on local influences’.Footnote 123 A similar environment is found in Imperial Germany, where voters were subject to intimidation from priests and noblemen even after the adoption of universal suffrage in 1871.Footnote 124 As emphatically stated by Malcolm Crook about the paradigmatic case of the 1848 French elections, ‘voting at this point was far from constituting the individual act with which we are familiar today […] because electoral participation was a collective business, which had traditionally taken place in assemblies that affirmed the identity of the community’.Footnote 125 In Hispanic America, political participation, and the notion of citizenship by extension, was associated with the Iberian notion of vecindad (neighbour or resident), which entails a territorial conception of identity linked to local dynamics and values.Footnote 126 These examples indicate that the transition from the collective and local vote to the individual and national vote should be analysed, independently of the expansion of suffrage. Ultimately, this appears to be a crucial step for a better comprehension of the transition from oligarchic regimes to democracy. The main challenge is to understand how voters stopped being controlled as they were at the turn of the twentieth century. From an institutional perspective, the effectiveness of the secret ballot was more important than the mere expansion of political rights. After all, as illustrated by the Chilean case, the introduction of the secret ballot in 1958 reduced significantly ‘the controls on votes’.Footnote 127 It can be said that ensuring privacy during the act of voting definitely reduces, and even relieves, the social and community pressures on the voter.
