EVANGELICALS, PENTECOSTALS AND POLITICAL REPRESENTATION IN BRAZILIAN LEGISLATIVE ELECTIONS (1998-2010)^*

EVANGÉLICOS, PENTECOSTAIS E REPRESENTAÇÃO POLÍTICA NAS ELEIÇÕES LEGISLATIVAS NO BRASIL (1998-2010)

ÉVANGÉLIQUES, PENTECÔTISTES ET REPRÉSENTATION POLITIQUE LORS DES ÉLECTIONS LÉGISLATIVES AU BRÉSIL (1998-2010)

Abstracts

What is the performance of Evangelical candidates in Brazilian legislative elections? Would Pentecostal candidates be responsible for the political representation of Evangelicals? Do Pentecostal churches have a high degree of electoral success? The Evangelicals influence in Brazilian politics has become increasingly prominent both in the media and in the academia. It is a consequence of the rapid expansion of Pentecostalism in Brazil. Previous literature assumes that Pentecostal candidates would use their congregations as “herd votes”. From a conceptual discussion of the Evangelical candidate and his church linkage, as well as a new database of Evangelical candidates, I present evidence that the share of Evangelical candidates remained stable in the last decade. Pentecostals politicians represent the vast majority of Evangelical politicians in Brazilian legislatures; however, in spite of that, the electoral success of Pentecostal churches is not as strong as it is asserted.

Evangelicals; Pentecostals; Candidates; Elections

Qual é o desempenho dos candidatos evangélicos nas eleições para o legislativo no Brasil? Seriam as candidaturas pentecostais responsáveis pela representação política dos evangélicos? Teriam as igrejas pentecostais um alto grau de sucesso eleitoral? A influência de evangélicos na política brasileira vem ganhando crescente destaque na mídia e na academia. Ela é consequência da rápida expansão do pentecostalismo no Brasil. A literatura prévia assumiu que candidatos pentecostais controlariam seus fiéis como um “rebanho eleitoral”. A partir de uma discussão conceitual sobre o candidato evangélico e sua relação com a igreja e de um novo banco de dados de candidaturas evangélicas, apresento evidências de que a proporção de candidaturas evangélicas se manteve estável na última década; os políticos pentecostais representam a grande maioria dos políticos evangélicos nos legislativos brasileiros; mas, a despeito disso, o sucesso eleitoral das igrejas pentecostais não é forte como se assevera.

Evangélicos; Pentecostais; Candidatos; Eleições

Quelle est la performance des candidats évangéliques aux élections législatives au Brésil? Est-ce que les candidatures pentecôtistes seraient responsables de la représentation politique des évangéliques? Les églises pentecôtistes auraient-elles un degré élevé de succès électoral? L’influence des évangéliques dans la politique brésilienne a pris une importance croissante dans les médias et les milieux académiques. C’est une conséquence de l’expansion rapide du pentecôtisme au Brésil. La littérature supposait, jusqu’alors, que les candidats pentecôtistes contrôlent leurs fidèles comme un «troupeau électoral». À partir d’un débat conceptuel sur le candidat évangélique et sa relation avec l’église et grâce à une nouvelle base de données sur les candidatures évangéliques, nous présentons des évidences selon lesquelles la proportion des candidatures évangéliques s’est maintenue stable au cours de la dernière décennie ; les pentecôtistes représentent la grande majorité des hommes politiques évangéliques dans les assemblées législatives brésiliennes mais, malgré cela, le succès électoral des églises pentecôtistes n’est pas aussi robuste que l’on ne l’affirme.

Évangéliques; Pentecôtistes; Candidats; Élections

Introduction

The 1986 general elections represented a turning point for Evangelical electoral mobilization in Brazil. Until the early 1980s, no more than a dozen Evangelicals had been elected at a time to the Câmara dos Deputados. In the 1986 elections, the number of elected Evangelical candidates grew to 32, and has grown further since then. However, despite the growth, the Evangelical presence in the Brazilian federal legislature is still far from reflecting its share of the Brazilian population. Despite being the case of most Evangelical electoral success in Latin America (Boas 2013), Evangelicals are still underrepresented in the parliament.¹

The underrepresentation of minorities in contemporary democracies is a major subject in the social sciences. Recent research sought to investigate why certain social groups do not enter into electoral politics, or, when they do, why they are not successful (e.g., Dancygier et al. 2015; Bueno and Dunning 2014; Wängnerud 2009). One of the underlying normative concerns of these studies is the concept of descriptive representation (Pitkin 1967; Mansbridge 1999). A representation in parliament that reflects a given constituency could contribute to better defend its interests and to increase its feeling of political inclusion, signaling that the political system is open to it as well as to other minority social groups. On the other hand, there are arguments that put descriptive representation into question. The idea of representation as a “mirror” of the social body might not produce, by itself, desirable results for minorities (Young 2000).

The underrepresentation of Evangelicals in Brazilian politics does not seem to produce in public opinion the same concern caused by other minority social groups. In fact, both the media and the academia have in general a negative assessment of the Evangelical electoral activism. Media coverage during elections suggests that Brazilian Pentecostal churches are politically successful, while possessing “currais eleitorais” [electoral corrals] and handling religious voters as their “votos de cabresto” [herd votes].² The political power of Evangelicals is frequently assessed as a threat, whether to the secular state, or to democracy itself, since the very vote of Evangelicals for Evangelical candidates would represent a distortion of the “true” preferences of the formers.³

Scholarly attention to the political consequences of Pentecostal growth in Latin America began to increase since the late 1960s. Studies on the subject stressed how Pentecostal churches would enter in politics by adopting clientelistic practices potentially harmful to democracy (D’Epinay 1970; Bastian 1994; Chesnut 1997; Gaskill 1997). Electoral support of the faithful would be an effective way to elect Pentecostal leaders, who, once in power, would benefit their churches with pork barrel and clientelism. The alleged argument is based on the implicit assumption that Evangelical voters would offer a high degree of electoral support to their leaders.

The electoral mobilization of Pentecostal churches in Latin America is particularly strong in Brazil and has gained increasing scholarly attention from the 1990s on (e.g., Mariano and Pierucci 1992; Freston 1993; Pierucci and Prandi 1995; Fernandes 1998; Oro 2003; Borges 2009; Mariano and Oro 2011; Machado and Burity 2014). The growth of the Evangelical population, coupled with the opening of the Brazilian “religious market”, and the particularities of its electoral system (open-list proportional representation with high magnitude districts) are some of the major causes for the increase in the number of Evangelical representatives in Brazilian politics. The literature on the subject also indicates that the post-1986 Evangelical political growth was due mainly to the electoral mobilization of Pentecostal churches. Starting in the 1980s, these churches have adopted a corporate representation model, engaging in electoral dispute with “official candidates”, and competing for positions in local, state and federal legislatures. The best known case is probably Igreja Universal do Reino de Deus [Universal Church of the Kingdom of God], which throughout the 1990s has elected dozens of bishops and pastors to the Chamber of Deputies, state assemblies and city councils.

Although there is evidence that Pentecostal corporate representation has been responsible for increasing the Evangelical representation in Brazilian federal legislature (Freston 1993), few efforts have been made to investigate the share of Pentecostal corporate representation over the total of Evangelical representation from the 1990s onwards. In addition, no systematic empirical effort was made to scrutinize the Evangelical electoral mobilization based not only on elected politicians, but also on the total of Evangelical candidates of a given set of elections. The exclusive focus on elected politicians, ignoring other Evangelical candidates, is problematic for obvious reasons. The literature on recruitment and political careers emphasizes the importance of analyzing representation taking into account also the unelected competitors (e.g., Perissinotto and Miríade 2009; Norris and Lovenduski 1995). Besides, statements about Evangelicals’ “electoral strength”,⁴ either as a social group, or as churches, face the risk of committing serious mistakes for disregarding the overall picture of Evangelical candidates.

This work seeks to contribute in this direction. By using a new database of Evangelical candidates, containing information about church affiliation and electoral performance of all Evangelical candidates to the Câmara dos Deputados and to the 27 Brazilian state legislatures from 1998 to 2014, I intend to contribute in order to fill in the gaps identified in the previous paragraph. In particular, this study aims to answer the following questions: what is the performance of Evangelical candidates in Brazilian legislative elections? Would Pentecostal corporate candidates be responsible for the Evangelicals representation in Brazilian legislatures? Do Pentecostal churches have a high degree of electoral success?

This article is organized according to the following order: in the following section, I present a conceptual and methodological discussion about the Pentecostal “official candidate”. My goal is to provide a more precise analytical definition of the phenomenon, as well as to present the churches that adopt the Pentecostal corporate representation model. Next, I describe the strategies used to identify Evangelical candidates and their churches. The next section presents data on the relative and the absolute variation of Evangelical candidates, showing the candidates’ distribution by state and party. Then, I focus on the performance of Pentecostal corporate representation and on the churches that adopt it. The last section summarizes the findings, presents the conclusions and suggests ways for future research.

Evangelical and Pentecostal candidates: a conceptual and methodological discussion

Until the 1986 elections, most elected Evangelical candidates for the Brazilian lower chamber came from mainline churches, mostly Baptist and Presbyterian. The presence of representatives of Pentecostal churches was almost inexpressive. In 1982, only 12 Evangelicals had been elected to the chamber, with seven affiliated to the Igreja Batista [Baptists] and one to the Assembleia de Deus [Assembly of God]. This scenario has greatly changed in the 1986 elections to the Constituent Congress, when 32 Evangelicals were elected. This time, despite the presence of ten Baptist deputies along with representatives of other mainline churches, no fewer than 13 elected candidates were from Assembleia de Deus; two were related to the Igreja do Evangelho Quadrangular [Church of the Foursquare Gospel]; and one to the Igreja Universal. Therefore, it was a change of parliamentary profile, as well as a significant numerical change (Freston 1993; Pierucci 1989).

The fundamental explanation for this change was the adoption by Pentecostal churches of a corporate model of political representation. In this model, the church would adopt “official candidates” and promote them to the faithful. Freston (1993) identified the phenomenon in the official support of Pentecostal churches to their candidates for state and federal legislative elections. The adoption of the corporate model would be mainly restricted to three churches: Assembleia de Deus, Igreja do Evangelho Quadrangular and Igreja Universal do Reino de Deus (Freston 1993, p. 197). The “official candidates” would be organic representatives of the churches themselves and exist only in Pentecostal churches.

My characterization of the Pentecostal “official candidate” is based on Freston’s, but is not limited to it. Unlike much of the literature on Protestant growth and Latin American politics, more concerned about the consequences of this relationship for the Protestant field (e.g., Willems 1967; Martin 1990) than for politics, my concern lies mainly in the political and electoral consequences of the phenomenon – which does not mean that the analysis with such an approach cannot contribute to a more general sociological understanding of Brazilian Pentecostalism. I understand the Pentecostal corporate representation as an institutional resource that some churches offer to their candidates, and the Pentecostal “official candidate” as a Weberian ideal type.⁵ Corporate representation is a practice of Pentecostal churches, and not of mainline Protestant churches.

The changes occurred in the Protestant field in Brazil, added to the Brazilian political system post-1988 (in particular the open-list proportional representation and the high magnitude of Brazilian districts), the pluralization of civil society and the end of Catholic monopoly in the religious market,⁶ contributed to produce the phenomenon of legislative candidates officially supported by Pentecostal churches. The social profile of church members, the process of sectarian socialization, the trust in pastors and religious leaders, all these factors would supposedly contribute to a strong support of church members to their “official candidates”.

As an ideal type, the Pentecostal “official candidate” is the representative of a church. Their relationship implies that the church recognizes the candidate and promotes him to the congregation. This recognition need not be publicized for civil society. In fact, most of the churches that adopt the corporate representation model do not publicize their support out of their congregation. The support takes place only inside the church or congregation. Although some candidates use “religious cues” to attract the Evangelical constituency, there are few cases where these “cues” include an explicit reference to a church. Generally, questionings made by outside researchers to church leaders about which, if any, candidates they support are not well received.⁷ However, the main characteristic of the “official candidate” is the fact that he is supported by the church. This definition excludes from the concept all candidates who profess a given Evangelical faith, but are not supported by their churches. They would be Evangelical candidates, but not “official candidates”. The distinction below illustrates the idea.

Suppose two candidates, A and B, each linked to an Evangelical church, A’ and B’, respectively. Candidate A is only an affiliate of church A’, that is, publicly shares the doctrine professed by church A’. However, church A’ is not necessarily committed to the candidacy of A. Candidate B, in turn, not only shares the publicly professed doctrine of church B’, but also is the candidate chosen by B’ to represent it, and therefore will be supported by it. Consequently, it appears that, in this example, only candidate B would be an “official candidate”, and not the candidate A. Interest in the “official candidates” does not lie in their degree of religiosity, but in the support relationship established between him and the church, a relationship that supposes the promotion by the church of the candidate in the congregation.

A third possible case of relationship between church and candidate that deserves to be cited is that of a church C’ which publicly supports a candidate C without C being a faithful or member of church C’. This is the case of many churches in executive elections: they publicly promote a candidate who, however, does not come from the ranks of the church. Regarding legislative candidates, which is the case of interest here, the vast majority of Evangelical candidates fits in cases A and B. I give examples of the three cases below.

In 2014, federal representative Benedicta da Silva, from Partido dos Trabalhadores (PT) [Worker’s Party] and the state of Rio de Janeiro, was re-elected for her fourth term in the lower chamber. Prior to that, she had been elected councilor for the city of Rio de Janeiro, also being the first black woman elected to the Senate. Evangelical, Benedicta was once a member of Assembleia de Deus, but today she is a Presbyterian. Besides being currently linked to a church that has no “official candidates”, Benedicta’s political career has always depended on little of the institutional resources of churches, being rather based on her political militancy (PT) and participation in social movements. Thus, Ms. Silva would be an example of the case A.

Jefferson Campos, from Partido Social Democrático (PSD) [Social Democratic Party], state of São Paulo, is a pastor of the Igreja do Evangelho Quadrangular, and in 2014 was re-elected to his fourth term as congressman. Previously, he had been councilor for the city of Sorocaba (SP) for two terms. Campos has a formal relationship with the church, which publicly adopts a corporate representation model. Thus, Mr. Campos is an example of case B.

In 2014, challenger Alceu Bueno, candidate from the tiny Partido Social Liberal (PSL) [Social Liberal Party], disputed a place for federal deputy for the state of Mato Grosso do Sul. Mr. Bueno was city councilor of Campo Grande. He was supported by Igreja Mundial do Poder de Deus [World Chuch of the Power of God], having recorded a video with the church’s leader, Valdemiro Santiago, in which Santiago asked the faithful to support Bueno. However, in April 2015, the candidate received a complaint about alleged involvement in a child prostitution network. Church leaders recognized that they have supported the candidate, but made clear that Mr. Bueno was not a member of the church.⁸ Thus, Mr. Bueno would be an example of the case C.

The Pentecostal “official candidate” is, therefore, one whose candidacy is promoted to the faithful of a church and whose electoral performance depends to a considerable extent on the support of this church. This definition excludes from the concept hypothetical cases where a candidate is backed by a church without being a member of it and without strongly depending on that support for his electoral viability.

If, at a conceptual level, this definition is not problematic, things are not that simple at the empirical level. One can think of a continuum of degrees of church support to a candidate; it would be difficult to establish empirically what the necessary degree for a candidate to be framed in that concept would be. I give an example. A pastor of Igreja Batista Vida Plena [Full Life Baptist Church], located in the city of São Bernardo do Campo (SP), told me in an interview that his church supports candidates for the city council. When I asked what kind of support, he said that the church (i.e., the pastor) presents the candidate to the congregation and argues that he is a member of the church, but that no one is obliged to vote for him just because of it. “We do not give him the pulpit to do advertising and there is neither financial nor logistical support.”⁹ This situation, even if it could be characterized as that of an “official candidate”, would represent a very different position in the continuum than that of, say, Igreja Universal, whose support for a candidate may imply political campaign in the pulpit and some pressure on the church members.¹⁰

The problem of having to set a minimum level of support of churches candidates can be circumvented through a change in strategy. My interest lies in identifying and separating Pentecostal “official candidates” from other Evangelical candidates in order to be able to compare them. Previous literature (Freston 1993) has not identified all Pentecostal “official candidates”, but correctly assumed that the phenomenon could be restricted to certain churches. Thus, “official candidates” would be those representatives of Assembleia de Deus (AD); Igreja do Evangelho Quadrangular (IEQ); and Igreja Universal do Reino de Deus (Iurd). In addition, I have decided to include as well the representatives of three other churches whose official support to candidates is publicly known: Igreja Internacional da Graça de Deus (IIGD) [International Church of the Grace of God]; Igreja Mundial do Poder de Deus (IMPD) [World Church of the Power of God]; and Igreja Maranata [Maranata Church]. Despite their differences, all these churches adopt the corporate representation model.

AD was brought to Brazil in 1911 and is the second oldest Pentecostal church in the country, behind only the Congregação Cristã [Christian Congregation], founded in 1910. The Swedish missionaries Daniel Berg and Gunnar Vingren, coming from the USA, founded AD in the northern state of Pará. Its penetration was so rapid that, by 1940, it was already present in all Brazilian states (Rolim 1980). It is the largest Pentecostal church in Brazil. According to Borges Junior (2010), the General Convention (Convenção Geral or CGADB) is the AD’s committee responsible for the political organization. However, local ADs operate without an administrative connection to a national organization. The national link between the churches is made by pastors, affiliated to state conventions, which, in turn, are linked to a national convention. Nevertheless, the General Convention has no deliberative power over the churches (idem, pp. 60-61). Moreover, it is important to note that there are other national conventions independent of CGADB. The best known are probably the National Convention of the Assemblies of God in Brazil (Conamad for the Brazilian acronym), also called Ministry Madureira, and Assembleia de Deus Vitória em Cristo [Victory in Christ], from pastor Silas Malafaia. The name “Assembleia de Deus” was given in 1918 and refers to the American Assembly of God, but there is no institutional link between them.

IEQ was founded in the USA in 1923 by Aimee Semple McPherson and arrived in Brazil in 1951 through missionaries Harold Williams and Jesus Ramos. In 1953, Williams named it Cruzada, and in 1955, it was structured as Igreja do Evangelho Quadrangular (Santos 2002). Until 1987, the American headquarters retained the right to appoint the president of the Brazilian branch. However, IEQ growth in Brazil caused that, in 1988, the leadership was no longer to be indicated by the “mother church”. Unlike AD, which has a decentralized character, IEQ has the structure of an episcopal government. There is a National Council elected every four years and State and Local Councils as well.

Perhaps the most famous Pentecostal church in Brazil, Iurd was founded in 1977 in Rio de Janeiro by Edir Macedo. Since then, it has shown a remarkable growth, and, since 1990, reached all Brazilian states. Mariano (2004) and Oro (2003) relate its rapid expansion to the centralized structure of the Iurd, which would make its decision-making and administrative processes more dynamic and facilitate their investments. The Church has considerable media influence, controlling a national network of AM and FM radio, the Aleluia network and the Record TV station. Besides being the most famous, Igreja Universal is also the most politically and electorally successful Brazilian Pentecostal church.

Interestingly, two of the churches addressed here, IIGD and IMPD, were both founded by Universal dissidents. IIGD was founded in 1980 by Romildo Ribeiro Soares (a.k.a. R. R. Soares) in the city of Rio de Janeiro. It has considerable media structure and is based largely on televangelism. According to Mariano (1999), it has fewer professional administrative head offices than those of Iurd, from which Mr. Soares broke up. Despite the similarities between the two churches, the geographical scope of IIGD is much smaller than the Iurd, being concentrated above all in the southeast. IMPD, in its turn, was founded in Sorocaba (SP), in 1998 by Valdemiro Santiago. As Mr. Soares, Mr. Santiago was also part of Igreja Universal, but in the course of time broke up and founded his own church. In 2014, IMPD had more than 4,000 temples throughout Brazil and abroad¹¹.

Finally, Maranata Church was formed in Vila Velha (ES) in 1968. Although it appears to have a structure less centered in a leader than the others, its first president was the engineer Manuel de Barros Passos, who gives his name to a homonymous foundation linked to the church. Just as IIGD and IMPD, Maranata “exports” temples: according to its site, the church would have more than 50 temples outside Brazil.¹²

As shown in Figure 1, there is a big difference in size between the six churches. AD is by far the largest Pentecostal church, and it is also the largest Evangelical church in Brazil in number of believers, accounting for 29.1% of Brazilian Evangelicals. The second (of the six) is the Igreja Universal, with 4.4% of Evangelicals, closely followed by IEQ, with 4.3%. IIGD, IMPD, and Maranata come later, each having 0.8% of the country’s Evangelical population. However, the relationship between the number of believers and “electoral strength” is far from straightforward, as we shall see.

Figure 1
Share of Pentecostal Churches Over Total of Protestants (2010)

Data and measurement

Working with Evangelical candidates brings some difficulties, the most obvious being how to recognize or identify a candidate and his church. That would be a difficult challenge even if the analysis were restricted to elected Evangelical candidates for state and federal legislatures. The number of Evangelical churches in Brazil is huge, and even though there is no updated official survey, it certainly exceeds the thousands.¹³ However, as seen in Figure 1, the distribution of believers by churches is quite uneven: of the more than 26 million Brazilian Pentecostals Protestants declared in 2010, almost half (12.3 million) belonged to the Assembleia de Deus (2010 Census).

Although many churches promote candidates for their congregations, this is usually not publicized outside the boundaries of the church. While media reports and research bodies periodically list the members of the Federal Evangelical caucus, these lists usually contain errors and outdated information. Examples of this situation are the lists of the Evangelical caucus made by the Departamento Intersindical de Assessoria Parlamentar [Union Department of Parliamentary Counsel] (Diap, in the Brazilian acronym). The Diap uses the criterion of Evangelical parliamentary faith, relating not only Evangelicals from mainline and Pentecostal churches, but also some not linked to any particular church.¹⁴ However, these lists contain errors and contain only elected Evangelical candidates. Restricting the analysis to elected candidates brings a serious selection bias problem. Although there is concern about the growing number of Evangelical representatives, the focus on the elected ignores the possible number of unelected Evangelical candidates. Thus, an investigation of Evangelical candidatures should take into account not only those who were successful, but also those who were not.

However, this could bring more difficulties. As said above, Evangelical churches do not publicize (except for their members) the names of the candidates for whom they offer electoral support.^15. How, then, to identify them? Firstly, it must be considered that it would hardly be feasible to identify all the candidates in a given election. The infeasibility is due not so much to the size of the effort, which, given the number of candidates in Brazilian legislative elections, would be considerable. The main problem lies in the fact that there are a large number of uncompetitive and inexpressive candidates, and it is virtually impossible to determine to which church they are linked with. If, for example, there were a high probability of finding out to which church belongs a candidate X who received ten thousand votes in a given election, the same probability would be very small for a candidate Z who has received only 50 votes. Next, I describe the methodological strategy used to identify candidates.

My empirical effort tries to take into account all Evangelical candidates for State Assemblies and Câmara dos Deputados between 1998 and 2014.¹⁶ For these five elections (1998, 2002, 2006, 2010, 2014), a total of 85,361 candidates competed for a place as federal or state representative. From candidates’ data in the Tribunal Superior Eleitoral (TSE) [Superior Electoral Court], I set up a unique database with all Evangelical deferred candidatures identified for the mentioned elections. The candidates’ identification effort was made in four main ways.

(a) The first was the collection of information about Evangelical candidates based on a review of the literature on Evangelicals and Brazilian politics. Given the mentioned difficulty to obtain this type of information, a series of case studies, whether focused on a district (state, city), or on an election, were made in which researchers sought to identify competing Evangelical candidates. This set of information allowed me to draw a first frame. However, most of these works pay attention only to elected candidates.

(b) The second was the use of two types of information contained in the TSE data: the candidates’ declaration of occupation as “priest or member of religious sect or order” and Evangelical titles used in candidates’ electoral names (see Table 2, below). In most cases, these two types overlap, i.e., the same candidate uses the title of “pastor” as his electoral name and declares his occupation as “priest”. I considered Evangelicals all candidates who used in their electoral names the titles of apostle, bishop, brother, minister, missionary, pastor, presbyter and reverend. As shown in Table 2, the title of “pastor” is responsible for the vast majority of Evangelical titles, followed by that of “brother”. Note also that, from 1998 to 2014, there was a small decrease on the titles of “pastors”, and an increase in that of “bishops” and “missionaries”. In absolute terms, there was a growth in the number of candidates with Evangelical titles. All candidates with titles were checked on the Internet, although not all have been confirmed.¹⁷

Table 1
Types of Link Between Candidate and Church.

Type	Description	Example
A → A’	Church candidate, but not official.	Benedita da Silva (PT-RJ)
B → B’	Church official candidate.	Jefferson Campos (PSD-SP)
C → C’	Official candidate, but not member of church.	Alceu Bueno (PSL-MS)

• Source: Elaborated by the author.

Table 2
Evangelical Titles of Candidates for Federal Chamber and State Assemblies (%)

	1998	2002	2006	2010	2014
Apostle	0.0	0.0	0.3	1.3	3.1
Bishop	2.2	9.4	6.7	5.9	5.1
Brother	19.0	19.8	25.6	20.5	23.0
Minister	0.0	0.7	0.0	0.0	0.0
Missionary	2.2	1.4	3.2	8.3	6.9
Pastor	75.2	67.6	62.9	62.0	61.2
Presbyter	0.0	0.7	0.6	1.0	0.3
Reverend	1.5	0.4	0.6	1.0	0.5
Total (N)	100 (137)	100 (278)	100 (313)	100 (303)	100 (392)

• Note: the percentage refers to the total of Evangelical titles in that election. Example: in 2002, from a total of 278 candidates using Evangelical titles, 19.8% used the “brother” title.
• Source: TSE.

(c) The third was the direct contact with Evangelical churches. This mode of identification turned out to be the least fruitful. Most mainline Protestant churches do not have “official candidates” nor a record of candidates linked to them – at least not one opened for researchers who are not church members. Thus, it seemed more efficient to get in contact with church lay members and ask them if and which candidates their church supported than seeking such information with the church leaders. Nevertheless, relatively few candidates were identified in this manner. With regard to Pentecostal churches, there was very little willingness to provide information to a university researcher. Politicians and advisers linked to AD and Iurd, for example, refused to provide any information at all (even if the refusal was not explicit or categorical). The exception was IEQ, the only one among the large Brazilian Pentecostal churches to collaborate with the research.

(d) Finally, for each Brazilian state, I looked for information about Evangelical candidates on the Internet sites of major newspapers of national or regional circulation. Stories detailing political and electoral disputes among Evangelical candidates often refer to their churches. This kind of information allows more thorough search not only on the relationship between that candidate and the church, but also on other candidates supported by that church.

Those were the four main ways adopted to identify Evangelical candidates in the analyzed period. It should be emphasized that, despite the research effort of over a year dedicated to the identification of Evangelical candidates, the data presented here may not be complete. Obviously, I would not be able to identify the candidates whose Evangelical faith is known only by themselves, and even for those publicly Evangelicals, it is possible that some candidates have not been identified. However, this possibility is above all related to non-competitive Evangelical candidates: the more competitive a candidate is, the more likely his faith (as well as his linkage to a church) is known. While recognizing this potential limitation, all the four strategies described above allow a reasonable degree of confidence in the validity of the data. Moreover, as discussed below, the data description reveals patterns that would be hard to explain if the data were too incomplete.

As cited above, in spite of the identification effort, there is a subset of candidates for which it was not possible to identify the church (see Figure 2). The number of Evangelical candidates with an unidentified church varied according to the election, constituting less than 39% of 1998 Evangelical candidates and just over 53% of 2014’s. Overall, the number of Evangelical candidates increases on each election, and the tendency is that the number of those with an unidentified church increases as well.

Figure 2
Number of Evangelical Candidates without an Identified Church

Data analysis reveals that the existence of a contingent of Evangelical candidates with unidentified church is not a big problem. There is a clear and strong correlation between the probability of church identification and the number of votes. Histograms in Figure 3, displayed by election, present on the horizontal axis the voting of candidates with unidentified church. It can be noticed that the vast majority had very few votes, fitting into the category of those whose identification is practically impossible. In this article, my concern lies primarily on identifying (i) Evangelical candidates and (ii) the churches of Pentecostal candidates. Given that the “official” support of Pentecostal churches is an institutional resource that presumably offers Pentecostal candidates a contingent of votes, it is theoretically expected that the smaller the number of votes of any Evangelical candidate, the less likely he is to be a Pentecostal “official candidate”. Hence, it seems safe to say that there is a low probability that there are Pentecostal “official candidates” among those without an identified church.

Figure 3
Voting of Evangelical Candidates without an Identified Church

But there is one last difficulty to be explained. I argued that I understand the Pentecostal “official candidate” as an ideal type, and that, given the difficulty of establishing a dividing line between what would be considered “official” support of a church and what it would not, I chose to follow Freston (1993) and assume that certain Pentecostal churches constitute the phenomenon of “official candidatures” (the churches are AD, IEQ and Iurd, to which I added IIGD, IMPD and Maranata). It so happens that, to investigate the phenomenon of the support of Pentecostal churches to their candidates, it would be necessary to distinguish, of course, the candidates really supported by the church from those who, although linked in some way to the church, were not officially supported by it in the election in question.

This difficulty is absent in the case of Iurd because it does not allow other candidates (other than the “official” ones) to compete for congregation votes. As a rule, every candidate supported by (or linked to) Iurd is an “official candidate”. The situation is also non-problematic in the cases of IIGD, IMPD and Maranata, simply because these churches support a very small number of candidates and do not have a national structure of temples and churches. The problem begins to appear in the case of IEQ. Schoenfelder and Paz (2006), for example, report the case of IEQ in the state of Rio Grande do Sul, which, in the 2006 elections, had two candidates competing for the state legislature, but only one with “official” support of the Church. In the case of IEQ, the difficulty scale is still small, but it becomes greater with AD, the church that presents the largest number of Evangelical candidates in Brazilian elections since 2006.

There is no easy way out for this problem. If qualitative studies can investigate at length the type of support that a church gives to its candidate at a local level and distinguish between candidates officially supported and those who, although belonging to the church, did not receive official support, such investigation would not be feasible in a quantitative-oriented research as this work. My solution was thus to identify candidates linked to the churches, even if, in many cases, I do not know precisely whether the connection implied in formal support or not. Strictly speaking, I assume that candidates linked to Pentecostal churches can be a proxy of Pentecostals “official candidates”.

A general framework of Evangelical candidates for Brazilian legislative elections

The first question to be investigated concerns the variation in the number of Evangelical candidates for the Câmara and State Assemblies. Between 1998 and 2014, the absolute number of candidatures has increased. In 1998, 76 Evangelical candidates competed for seats in the federal, and 167 in states legislature. In 2014, these numbers have more than doubled, reaching 207 Evangelical candidates for the Chamber and 392 for the Assemblies. This increase in absolute terms would be expected, since the Evangelical population has also increased in the country. The question, then, is to know how was the relative variation, that is, the number of Evangelical candidatures on the total number of candidatures.

In this case, it is also possible to identify an increase. In 1998, 2.3% of all candidates for the Lower Chamber were Evangelicals. In 2014, this percentage reached 3.5%. However, it has to be noticed that after the increase between 1998 and 2002, Evangelical candidatures for the Chamber remained relatively stable. Between 2002 and 2014, the relative number of federal Evangelical candidates showed a tiny variation of approximately 0.2%. The situation is similar in the case of states. Between 1998 and 2002, there was an increase in the relative number of Evangelical candidates, which went from 1.6% to 2.6%. However, this number has remained fairly constant from 2002 to 2014. This means that, from the point of view of candidates’ supply, Brazilian Evangelical population is underrepresented. It would be necessary a sharp increase in the percentage of Evangelicals, and not stability, if the number were to approach the corresponding percentage of Evangelicals in the Brazilian population. Thus, with regard to candidatures, statements about the “strength” or “rise” of Evangelicals should at least be questioned.

It could be argued that these numbers may be underestimated, and that some candidates of Evangelical faith may not have been incorporated into the database. Even if this was the case, it would still be necessary to observe that the difficulty of identification tends to be greater the further in time the election is. It is easier to identify Evangelical candidates from 2014 than in 1998 elections, for example. This means that even if the numbers were underestimated, this bias would hardly alter the general trend presented here. Therefore, it is more likely that Evangelical candidates from older elections are ignored than candidates from recent ones. This does not change the pattern found; rather, it reinforces it. In addition, the evidence presented in Figure 4, showing that the share of Evangelical candidates is lower than the share of the Evangelical population, is consistent with studies of political careers that stress supply-side explanations for the underrepresentation of minority groups (cf. Norris and Lovenduski 1995). Constraints on resources, such as time and money, could severely limit the number of Evangelical applicants seeking a place as a candidate in Brazilian elections.

Figure 4
Evangelical Candidates for the Chamber of Deputies and State Assemblies

The second aspect to be investigated is the variation in the number of Evangelicals elected to federal and state legislatures.¹⁸ In the case of state level, these are 26 State Assemblies plus the Legislative Chamber of the Distrito Federal. The Brazilian Chamber of Deputies has 513 seats, and the state legislatures, 1,059. Figure 5, below, shows the data. Unlike the number of candidates, which, in absolute terms, increased for both federal and state legislatures, the number of elected Evangelical oscillated in the period. It is true that the total variation was positive. In 1998, 29 Evangelicals were elected for the Chamber, and 46 were for the As- semblies. In 2014, the numbers for Chamber and Assemblies reached 64 and 75, respectively. However, despite the considerable increase, it is necessary to point out, first, the drop occurred in 2006, and second, the growth recorded in 2010. Both trends can be seen at the federal as well as at the state level. On the 2010 federal growth, from 35 to 65 deputies, one can notice that it was even greater than that registered in 1986 elections. That did not change, however, the fact that the Evangelicals remained underrepresented in legislatures.

FIGURE 5
Evangelical Candidates Elected for Chamber of Deputies and State Assemblies

Some researchers have suggested that major corruption scandals occurred in the 2000s, allegedly involving a number of Evangelical parliamentarians, would have negatively impacted on the electoral success of the group (e.g., Machado and Burity 2014; Mariano and Oro 2011). Three of these scandals are noteworthy: the so-called “mensalão” [monthly bribes]; the scandal of “sanguessugas” [leeches or bloodsuckers]; and the “mensalão” of Democratas (DEM), a right-wing party.¹⁹ Having started with a story on Veja magazine in 2005 about a vote-buying scheme in Brazilian Congress, the “mensalão” scandal involved several politicians, among which Representative Carlos Rodrigues (Partido Liberal), Iurd bishop and one of the Church’s most important leaders.²⁰ In 2006, came to light the “sanguessugas” scandal, an overpricing scheme of ambulances purchase. Rodrigues was again accused of being one of the beneficiaries of the scheme. According to Souza (2009), the episode drew attention because of the fact that it allegedly involved ten parliamentarians linked to Assembleia de Deus and 14 linked to Igreja Universal. Nevertheless, in late 2009, the media highlighted the scandal of “mensalão” of DEM, held in the Distrito Federal and which involved Evangelical politicians as well. The episode was marked with a video showing state representative Junior Brunelli (Partido Social Cristão), linked to Igreja Catedral da Benção [Cathedral of Blessing Church], and others receiving bribes and then saying a prayer.²¹ Although there might be insufficient evidence to establish a causal relationship between corruption scandals and the electoral performance of Evangelical candidates, it is worth noting that the two high-profile cases (the “mensalão” and the “sanguessugas” scandals) occurred precisely in 2006, the year when the drop in the number of Evangelical elected occurred. The hypothesis of a connection between involvement in corruption and the decline in Evangelical representation will be addressed again later on, when I detail the electoral performance of the main Brazilian Pentecostal churches.

Figure 6 shows the number of Evangelical candidates per state, separating the number of total candidates (first graph) and the number of elected (second). There are a few things worth noting here. Rio de Janeiro and São Paulo are the Brazilian states with the highest number of Evangelical candidates. Next and way below come Minas Gerais, Distrito Federal, Espírito Santo, Bahia, and Goiás. Although it is the country’s most populous state, São Paulo only surpassed Rio in the number of Evangelical candidates in 2010. Until that year, Rio de Janeiro was in first place.

Figure 6
Evangelical Candidates per State (total and elected)

According to the 2010 Brazilian Census, in absolute terms, the states with the largest Evangelical populations are, from the highest to the lowest, São Paulo, Rio, Minas Gerais, Bahia, Paraná, and Pará. This shows that the Evangelical candidatures supply, even if it has some relation with the size of the state’s Evangelical population, is not determined by it. Take the example of Rio and São Paulo: in 2014, both had 69 and 82 candidates, respectively. These figures do not reflect the fact that the Evangelical population of São Paulo (~ 9.9 million) is more than twice as bigger as that of Rio (~ 4.6 million). The same applies to Distrito Federal (DF), for example, which presents a greater Evangelical candidates supply than the state of Paraná, even though the Paraná Evangelical population (~ 2.3 million) is more than three times higher than that of DF (~ 690,000). The explanation for variation in candidates’ supply must reside, at least in part, in the distribution of churches over the states as well as in organizational differences between the churches. The distribution of Evangelical candidates per election and state does not show clear evidence of a link between involvement in corruption scandals and candidates’ supply. In most states, there was no drop in the number of candidates in 2006. On the other hand, several states recorded a drop between 2006 and 2010, being Rio the most significant.

The number of elected candidates per state follows a similar trend as that of the overall number of candidates. Again, Rio de Janeiro stands out, being surpassed by Sao Paulo only in 2014. In addition to São Paulo and Rio, Minas Gerais also emerges with a significant number of elected candidates. However, it is something expected, given the fact that it is a populous state and has a considerable Evangelical population. On the other hand, the performance of Espírito Santo draws attention, since the state, in 2010, had approximately 1.2 million Evangelicals, but by 2010, it still had a number of Evangelical representatives higher than states with larger populations. Finally, when analyzing the elected candidates, it becomes clear that if the drop in the number of Evangelical representatives in 2006 occurred in several states, in none of them it was as strong as in Rio.

The next step is to assess the distribution of Evangelical candidates per Brazilian political parties. Despite the presence of some Christian parties, it should be noted that there are no formally Evangelical parties in the Brazilian political system. Borges (2009), investigating this issue, notes that political identity is a central variable for the formation of political parties. However, this identity does not exist among Evangelicals; there would not be a political identity that would make Evangelical voters and politicians distinct from other social groups. To what extent Evangelical candidates prefer certain parties, or, conversely, are spread across different Brazilian parties is the question to be investigated below.

Brazil post-1988 has one of the largest party fragmentations of the world (Clark et al. 2006). Part of the literature attributes this fact to certain specific characteristics of Brazilian electoral system, such as open-list proportional representation, high magnitude districts, and the existence of coalitions for legislative elections (e.g., Lima Jr. 1999; Kinzo 1997). This relationship, however, is still unclear, and has been challenged by recent research (Calvo et al. 2015). In Brazil, the effective number of parties is similar both in low as in high magnitude districts, suggesting that there might be other explanations for the phenomenon of fragmentation.

Party distribution of Evangelicals in Brazil is highly dispersed. In 1998, Evangelical candidates were elected by 15 different parties (considering elections to federal and state legislatures). In 2002, the number increased to 18; in 2006, it fell to 17; in 2010, it rose to 19; finally reaching 23 in 2014. By itself, this picture reveals little, and could be in line with the upward trend of party system fragmentation in the last elections. However, as one can see below, the distribution of Evangelicals is far from balanced or random. Figure 7 shows the total number of Evangelical candidates (horizontal axis) and the number of elected Evangelicals (vertical axis) distributed among the different Brazilian parties.²² I included in the chart only the parties that have elected one or more Evangelical candidates.

Figure 7
Evangelical Candidates (Total and Elected) per Political Parties

Note that the vast majority of parties is concentrated in the lower left quadrant, indicating a comparatively small number of elected and presented candidates. A much smaller number of parties occupies the chart center, and an even smaller number is in the upper right quadrant, which represents a larger number of candidates (total and elected). It should be noted, in this quadrant, the presence of the Partido Republicano Brasileiro (PRB) and the Partido Social Cristão (PSC). Although not formally Evangelicals, both parties excelled in the last legislative elections as those with the largest numbers of presented and elected Evangelical candidates.

Created in 2005, the PRB is considered by many as the “political wing” of Igreja Universal, a fact denied by both the Party and the church. Despite the denial, the vast majority of Iurd candidates enter the electoral race via the PRB, and the success of the church’s candidates certainly explains the high success of the party’s Evangelical candidates. The PSC, in turn, was registered in 1990. According to its statute, it is defined as a party built in the “Christian social doctrine”, ²³ that, according to its “quick guide” for the 2014 elections, advocates for “life and human dignity, respect for traditional moral order” and other principles.²⁴ Although it has no formal ties to Evangelical churches, the Party, along with the PRB, has registered a considerable number of Evangelical candidates, who, especially in 2010 and 2014, achieved remarkable electoral success.

Finally, it is worth noting that the two parties with the largest Evangelical representation, PRB and PSC, even though having elected federal and state representatives, have little expression in executive elections. In 2014, the PSC presented its own candidate for the presidential elections, Pastor Everaldo Pereira, who finished the race in fifth place, getting only 0.75% of valid votes. The two largest Brazilian parties with concrete chances to win executive elections, the center-left Partido dos Trabalhadores (PT) and the center-right Partido da Social Democracia Brasileira (PSDB), although not highlighted in Figure 6, are distinguished from each other in terms of number of Evangelical candidates. Between 1998 and 2014, PT had an average of 5.2 Evangelical candidates per election (both for federal and state legislatures), electing 1.6. In the same period, PSDB had an average of 21.8, electing circa 8 candidates.

Comparing Evangelical candidates and Pentecostal “official candidates”

As discussed in previous sections, corporate representation is an institutional resource that certain Pentecostal churches offer their candidates. According to Freston (1993), it was this phenomenon that made possible the remarkable growth of Evangelical representation in the Brazilian Congress in 1986. Indeed, nearly half of the Evangelical congressmen elected that year were Pentecostal official candidates. Although the change in Evangelical representation has multiple outcomes (for example, the post-1988 Evangelical deputies have a different ideological and socio-economic profile), which deserve further investigation, the focus of this section lies in the elected Pentecostal official candidates and the electoral success of Pentecostal churches. Therefore, I will concentrate my analysis on the performance of the six mentioned Pentecostal churches (AD, IEQ, Iurd, IIGD, IMPD, and Maranata). While there may be other candidates from Pentecostal churches, I assume that the phenomenon of Pentecostal corporate representation can be reasonably measured in these six churches. I do not say that there is no other; only that these six are the most significant and best represent the phenomenon.

Figure 8 shows the number of Evangelical and Pentecostal elected candidates for the Chamber and Assemblies. The analyzed time frame (1998-2014) is short, which makes difficult any attempt of long-term trend identification. However, some conclusions can be drawn. In the case of the federal legislature, it is clear that the share of Pentecostal official candidates is high and has been growing over the total of Evangelicals. It reached its highest level in 2002. In that year, from a total of 42 elected Evangelicals, 37 were Pentecostal official candidates, making up almost 90%. In 2006, the Pentecostal participation decreased, but despite the fall, it continued to grow until 2014, when, from a total of 64 Evangelicals, 45 were Pentecostals (~ 70%).

Figure 8
Number of Corporate Pentecostals and the Total of Evangelicals Elected

The situation is similar regarding the State Assemblies. The lower Pentecostal share of the total of elected Evangelicals occurred in 1998, when there were 30 Pentecostals from a total of 46 elected Evangelicals. This represents a percentage of roughly 65%. From 1998 onwards, this share only increased, reaching 79% in 2014, when 59 Pentecostals were elected from a total of 75 Evangelicals. Therefore, it seems safe to assume that, whether in the Chamber or in the Assemblies, Pentecostal corporate participation is growing and is responsible for much of the Evangelical representation in Brazilian legislatures.

A second question to be investigated concerns the electoral success of Pentecostal churches that have adopted the corporate representation model. On the one hand, the growing Pentecostal corporate participation suggests that these churches do obtain some electoral success. On the other, the degree of this success is an issue usually ignored by the literature. The few existing studies on the subject corroborate the hypothesis of Pentecostal churches’ high electoral success. Fernandes (1998), for example, found through a survey applied in the metropolitan area of Rio de Janeiro that 87% of the Assembleia de Deus church-goers reported to have voted for the church’s official candidates. The percentage would reach 95% in the case of the faithful of the Universal Church (1998, p. 126). Fernandes’ work is often cited as evidence of the high degree of adhesion of the Pentecostal faithful to their official candidates.

Despite the low number of evidence of Pentecostal voters’ support for Pentecostal candidates (and, by extension, the electoral success of the latter), most of the literature on Pentecostals and politics in Latin America assumes that this support is high. This would be an important explanatory piece in the more general argument on how Pentecostal politicians reproduce clientelistic practices to have access to state resources (Chesnut 1997; Gaskill 1997; Bastian 1994; D’Epinay 1970). According to this view, Pentecostal leaders would use “herd votes” to provide electoral support for candidates in executive elections, which, once elected, would reward the leaders with benefits. Or, similarly, the electoral support of the “herd” would be used by Pentecostal leaders to win seats in the legislature, where they could, once again, benefit their churches with parochial benefits (pork). Although this argument is based on a series of empirical hypotheses that still need to be tested, I draw attention here just to one aspect, namely, the electoral success of Pentecostal candidates. The above argument is based on the assumption that Pentecostals candidates enjoy high electoral support from their congregations. If this is true, then it is expected that candidates of Pentecostal churches exhibit a high degree of electoral success.

However, Figure 9 shows a picture that, if does not flatly deny this assumption, certainly does not confirm it. The graph below shows the total number of presented and elected candidates per church (includes Chamber and State Assemblies). A first observation to be made is that although this analysis has focused on six churches, three of them (IIGD, IMPD, and Maranata) had a much inferior number of candidates than the others. In 2014, IIGD and IMPD had a significant growth and presented eight and 13 candidates, respectively, but only elected four and seven. In other years, both churches could not present more than six candidates. Regarding Maranata, the church came to elect four out of six candidates presented in 2006, repeating the same performance in 2010. In 2014, however, the church presented five, but only one was elected.

Figure 9
Number of Candidates (Total and Elected) per Pentecostal Church

Source: Evangelical database.

AD, IEQ, and Iurd present a different picture, but, on the whole, also do not corroborate the assumption of a high degree of electoral success. As the largest Evangelical church in Brazil, it is not surprising that AD presents a greater number of candidates than the others. Moreover, it is remarkable the growth of the number of its candidates, whether elected or presented. Of the three churches, it is the only one that shows a trend of growth. In 2014, AD presented 109 candidates and elected 57, just over half. However, it was the only year in which it occurred; in no other poll AD elected more than 50% of the candidates.

The analysis of the IEQ’s frame reveals considerable variation of the presented candidates throughout the period. It came to present 30 candidates in 2002, but only 17 in 2014. Considering the period as a whole, the church maintained a steady performance: it elected six of a total of 17 candidates in 1998, and eight of a total of 17 in 2014. Similarly to AD, IEQ did not obtain more than 50% of electoral success in any dispute.

If there is a church that could embody the prototype of Pentecostal corporate representation, that is Iurd. In this case, one would expect the highest degree of electoral success. The church’s political and media power have been emphasized not only by the press, but also by scholars (e.g., Mariano 2004; Oro 2003). Figure 9 shows a different picture for Iurd vis-à-vis the other churches. In 1998 and 2014, Igreja Universal was able to elect a considerable share of its candidates. In 1998, from 39 candidates, 31 were elected; in 2014, from 38 candidates, 27 were elected. Nevertheless, it is also worth noting that, until 2002, it was the church that elected most representatives to Brazilian legislatures. AD dominance would begin only in 2006 (this, of course, in absolute terms; in percentage, Iurd always elected more). However, even in the case of Iurd, it can be noted that, despite being the prototypical case of Pentecostal corporate representation, the church is far from electing all of its candidates, and, in 2006 and 2010, elected less than 50%.

It could be argued that the church’s goal is not to elect all of its candidates. This is an important point. It is clear that, even without having a high degree of electoral success, churches already benefit by electing some representatives. I agree with the assumption that churches seek different goals when supporting candidates and that some of these goals can be carried out without all of them being elected. In an interview, a state deputy from Partido da Social Democracia Brasileira (PSDB) told me that one of the reasons why Iurd elected deputies was to use their legislative immunity to “protect” some of the church initiatives.²⁵ The objection is valid, but misses the point. I am not saying that Pentecostal churches fail when they do not elect all of their candidates. My point is that their electoral performance is lower than the literature suggests, and that, therefore, the hypothesis of clientelistic use of Pentecostal churches loses much of its plausibility.

Finally, it is worth noting the relationship between the drop in the number of elected Evangelical candidates in 2006, visible in Figure 5, and the drop in the number of candidates of the three major Pentecostal churches, AD, IEQ, and Iurd in the same year, visible in Figure 8. All of the three churches elected fewer candidates in 2006 than in 2002, but no decrease was as big as that of Iurd, which elected 38 candidates in 2002 and only 14 in 2006.

Conclusions

Most of the literature on Pentecostalism and politics in Latin America to date emphasized how Pentecostal churches would act in a clientelistic way and contribute to the maintenance of existing structures of domination. This argument would rest on several assumptions, one of them being that Pentecostal leaders get high electoral support from their congregation. The Pentecostal “herd vote” would be used either to support candidates in executive elections, or to elect Pentecostal “official candidates” in legislative elections. In both cases, the churches (or their leaders) would be rewarded with pork. The little amount of evidence found in the literature on the degree of electoral support of congregations to their candidates would go in that direction.

However, this study reveals a more complex picture. First, the share of Evangelical candidates for Brazilian legislatures has remained stable since 2002. The number of elected Evangelicals, while increasing between 1998 and 2014, remains far below of what would constitute a proportional representation of the number of Evangelicals in Brazilian population. Second, although underrepresented, Evangelicals owe their growth in the Câmara dos Deputados and state Assemblies largely to Pentecostal “official candidates”. Any normative discussion on the political representation of Evangelicals and its relationship with Brazilian democracy should take this fact into account. Third, Pentecostal churches that have adopted the corporate model of political representation have a much smaller electoral success than what is usually assumed by scholars and the media. The electoral performance of the churches presented here suggests that, although they get support from their congregations, this support could hardly be characterized as unrestricted, nor even equivalent to 80% or 90% of the congregation, as found Fernandes (1998).

The above frame seems contradictory, but it is not. Pentecostal churches are not as electorally successful as it is asserted. Even though they do have some electoral “strength”, they do not elect all, or, in most cases, not even half of their candidates. However, this statement is not inconsistent with the idea that the corporate model of political representation is largely responsible for the representation of Evangelicals in Brazilian legislatures. The consequences of the findings summarized here are diverse and may indicate paths for further research on the topic. I suggest three possibilities. A first question to keep in mind refers to the change in the parliamentary profile caused by the growing number of Evangelical politicians, and to what extent the growth of Pentecostal official candidates would not be a determining factor in the diversification of the Brazilian Chamber. A second issue concerns the cases of electoral failure of Pentecostal official candidates, its possible explanations, and, more generally, the contextual variables that could impact the support given to these candidates by voters in their churches. Finally, a third question deals with the political behavior of Evangelical politicians. There are still no researches that corroborate or refute the assumptions prevalent today on the legislative activities of Pentecostals politicians.

REFERENCES

• BASTIAN, Jean-Pierre. (1994), Protestantismos y modernidad latinoamericana: historia de unas minorías religiosas activas en América Latina México, Fondo de Cultura Económica.
• BOAS, Taylor. (2013), “Serving God and man: evangelical christianity and electoral politics in Latin America”. Paper presented at the annual meeting of the APSA, Chicago, IL.
• BORGES, Thiago D. P. (2009), “Identidade política evangélica e os deputados estaduais brasileiros”. Perspectivas, 35: 149-171.
• BUENO, Natália & DUNNING, Thad. (2014), “Race, resources, and representation: evidence from Brazilian politicians”. Available at SSRN: http://ssrn.com/abstract=2498662
  » http://ssrn.com/abstract=2498662
• CALVO, Ernesto; GUARNIERI, Fernando & LIMONGI, Fernando. (2015), “Why coalitions? Party system fragmentation, small party bias, and preferential vote in Brasil”. Electoral Studies, 39: 219-229.
• CLARK, William; GILLIGAN, Michael & GOLDER, Matt. (2006). “A simple multivariate test for asymmetric hypotheses”. Political Analysis, 14 (3): 311-31.
• CHESNUT, Andrew R. (1997), Born again in Brazil: the pentecostal boom and the pathogens of poverty New Jersey, Rutgers University Press.
• DANCYGIER, Rafaela; LINDGREN, Karl-Oskar; OSKARSSON, Sven & VERNBY, Kare. (2015), “Why are immigrants underrepresented in politics? Evidence from Sweden”. American Political Science Review, 109 (4): 703-724.
• D’EPINAY, Christian L. (1970), O Refúgio das massas: estudo sociológico do protestantismo chileno Translated by Waldo Cesar. Rio de Janeiro, Paz e Terra.
• FERNANDES, Rubem C. (1998), Novo nascimento: os evangélicos em casa, na igreja e na política Rio de Janeiro, Mauad.
• FRESTON, Paul. (1993), Protestantes e política no Brasil: da constituinte ao impeachment Campinas. PhD dissertation, State University of Campinas.
• GASKILL, Newton. (1997), “Rethinking protestantism and democratic consolidation in Latin America”. Sociology of Religion, 58 (1): 69-91.
• KINZO, Maria D. G. (1997), “Governabilidade, estrutura institucional e processo decisório no Brasil”. Parcerias Estratégicas, 1 (3): 9-25.
• LIMA JR., Olavo B. (1999), “Reformas de sistemas eleitorais: mudanças, contextos e consequências”. Dados, 42 (1): 17-61.
• MACHADO, Maria D. C. & BURITY, Joanildo. (2014), “A Ascensão política dos pentecostais no Brasil na avaliação de líderes religiosos”. Dados, 57 (3): 601-631.
• MANSBDRIGE, Jane. (1999), “Should blacks represent blacks and women represent women? A contingent ‘yes’”. The Journal of Politics, 61 (3): 628-657.
• MARIANO, Ricardo. (2004), “Expansão pentecostal no Brasil: o caso da Igreja Universal”. Estudos Avançados, 18 (52): 121-138.
• MARIANO, Ricardo & ORO, Ari P. (2011), “The reciprocal instrumentalization of religion and politics in Brazil”. Annual Review of the Sociology of Religion, 2: 245-266.
• NORRIS, Pipa & LOVENDUSKI, Joni. (1995), Political recruitment: gender, race and class in the British Parliament Cambridge, Cambridge University Press.
• ORO, Ari P. (2003), “A política da Igreja Universal e seus reflexos nos campos religioso e político Brasileiros”. RBCS, 18 (53): 53-69.
• PERISSINOTTO, Renato M. & MIRÍADE, Angel. (2009), “Caminhos para o parlamento: candidatos e eleitos nas eleições para deputado federal em 2006”. Dados, 52 (2): 301-333.
• PITKIN, Hanna. (1967), The concept of representation Oakland, University of California Press.
• ROLIM, Francisco C. (1980), Religião e classes populares Petrópolis, Editora Vozes.
• SCHOENFELDER, Rosilene & PAZ, Joice L. (2006), “A Igreja do Evangelho Quadrangular nas eleições de 2006: a disputa pela vaga de deputado Federal”. Debates do NER, 7 (10): 27-37.
• WÄNGNERUD, Lena. (2009), “Women in parliaments: descriptive and substantive representation”. Annual Review of Political Science, 12: 51-69.
• WEBER, Max. (2001), Metodologia das ciências sociais, Parte I Tradução de Augustin Wernet. São Paulo, Cortez / Editora da Universidade Estadual de Campinas.
• YOUNG, Iris M. (2000), Inclusion and democracy Oxford University Press.

• 1
  I use the terms “Protestant” and “Evangelical” in an equivalent manner.
• 2
  See for example: “Presidente da CNBB diz que igrejas não são ‘currais eleitorais’” (Folha de S. Paulo, 29/08/14); “‘Apóstolos’, ‘Bispos’ e ‘Pastores’: os novos coronéis dos currais eleitorais” (T1 Notícias, 12/08/14); “Igreja Católica não tem curral eleitoral” (O Estado de S. Paulo, 07/09/14); “Bancada evangélica terá sete vereadores na Câmara Municipal do Rio, 14% do total” (iG, 11/10/12); “Voto não pode ser condicionado à fé evangélica, defendem igrejas históricas” (Congresso em Foco, 03/10/14); “Partidos tentam evitar veto de igrejas em SP” (Folha de S. Paulo, 16/01/12).
• 3
  Carlos Alberto Libânio, better known as Frei Betto, told the newspaper Folha de S. Paulo that Brazilian secular state would be threatened by an Evangelical caucus that wants “to confessionalize politics”, creating “[a] form of highly dangerous Brazilian fundamentalism”. The growth of the Evangelical caucus, far from being the legitimate product of the population’s religious composition, would be explained by people who “totally lose consciousness (...) and become little lambs of anyone who wants to manipulate them”. Evangelical churches would transform “their faithful in lambs, which, threatened by a theology of fear, end up following the pastor’s voice in what he says” (“‘Temo que a presidente Dilma renuncie’, diz frei Betto”, Folha de S. Paulo, 09/08/15).
• 4
  See for example: “A força dos evangélicos” (Revista Época, 05/09/14); and “Vinde a mim os eleitores: a força da bancada evangélica no Congresso” (Revista Veja, 23/03/13).
• 5
  The ideal type is based on the accentuation of certain characteristics of a given phenomenon (Weber, 2001).
• 6
  It should be noted that although, in terms of number of faithful, the Catholic Church has remained hegemonic in Brazil until the beginning of the second half of the twentieth century, the situation in Brazilian public debate was, since the late nineteenth century, very different, as evidenced by, for example, the conflict between Catholic clergy, liberals and freemasons.
• 7
  Of all the largest Brazilian Pentecostal churches adepts of the corporate representation model, the only one who answered my questions about supported candidates in elections was the Igreja do Evangelho Quadrangular. I thank the advisors of federal deputy Jefferson Campos (PSD) and city councilor Carlos Evaristo (PSD), both from IEQ, for help in obtaining information for this research.
• 8
  The scandal involving Mr. Bueno was reported by newspapers and internet sites. See for example: “Pastor diz que vereador envolvido em escândalo não é membro da igreja” (Midiamax, April 21, 2015).
• 9
  Author interview on May 17, 2015.
• 10
  As one of the anonymous reviewers correctly remarked, political advertising in churches is forbidden in Brazil (Lei 9.504/97, Art. 37). The prohibition does not specifically address churches. It is rather a comprehensive ban directed to public use places, such as cinemas, stores, shopping centers, and places of worship. The law is vague, leaving considerable room for interpretation of what would constitute political propaganda in religious temple. Still, Pentecostal churches’ political activities often seem to be beyond what would be considered legally acceptable.
• 11
  Information available on the church’s official website: http://www.impd.org.br/portal/index.php?link=institucional (last accessed on December 17, 2015).
• 12
  Information available on the church’s official website: http://www.igrejacristamaranata.org.br/?page_id=2064 (last accessed on December 17, 2015).
• 13
  A quick walk in poor suburbs of São Paulo reveals the large number of existing Evangelical churches. Although the best known of them are big – e.g., Igreja Universal, Assembleia de Deus –, some are so small that their places of worship are not discernible from a warehouse or garage.
• 14
  “O DIAP classifica como integrante da bancada evangélica, além dos bispos e pastores, aquele parlamentar que professa a fé segundo a doutrina evangélica” (“Atua- lização da bancada evangélica: Diap identificou 74 deputados”, Diap, October 6, 2014).
• 15
  I am not suggesting that it would be normatively justifiable to oblige them to do it. I am only noticing that they do not do it.
• 16
  The data from the Tribunal Superior Eleitoral [Superior Electoral Court] for candidacies previous to 1998 elections are incomplete. According to information given by a Court’s official to the author, data regarding previous elections can only be obtained with regional courts [Tribunais Regionais], and still, it is not certain that they are digitally available. Thus, I chose to focus my research to the 1998-2014 period.
• 17
  Of course, candidate names containing unintentional Evangelical titles were excluded. Example: José Carlos Bispo [Bishop] da Paz was a candidate for the Chamber in the 2006 election and adopted the name “Bispo da Paz”. Such case was not computed for the obvious reason that his candidate name is not a religious title, but his own name.
• 18
  I am not considering here "suplentes", i.e., highly ranked losers who are able to replace winning candidates who resign their mandates for higher appointed offices.
• 19
  “Bancada evangélica cai para metade no Congresso” (Portal G1, October 12, 2006).
• 20
  In November 2012, Rodrigues was convicted of corruption and money laundering by the Supremo Tribunal Federal [Federal Supreme Court]. He received a sentence of six years and three months in prison. See: “STF condena ex-deputado Bispo Rodrigues a 6 anos e 3 meses de prisão pelo mensalão” (UOL Notícias, November 26, 2012).
• 21
  “Imagens do suposto esquema de mesada a deputados têm até oração” (G1, November 30, 2009).
• 22
  Parties are presented in accordance with their initials at the time of the election. It is worth noting that, in 2003, the Partido Progressista Brasileiro (PPB) changed its name to Partido Progressista (PP); in 2006, the Partido Liberal (PL) merged with the Partido da Reedificação da Ordem Nacional (Prona) to create the Partido da República (PR); and in 2007, the Partido da Frente Liberal (PFL) became the Democratas (DEM).
• 23
  See the PSC statute: <http://www.psc.org.br/site/partido-social-cristao/documentos/estatuto.html>. (last accessed on January 15, 2016).
• 24
  See: <http://www.psc.org.br/site/partido-social-cristao/documentos/guia-rapido-eleicoes-2014.html>. (last accessed on January 15, 2016).
• 25
  In Brazil, politicians are prosecuted and judged solely by the Supreme Court, even when accused of crimes committed outside their duties (murder, theft, etc.).

Publication Dates

• Publication in this collection
  2017

History

• Received
  07 June 2016
• Accepted
  09 Nov 2016