1 Introduction

In 2015, Brazil’s Federal Supreme Court (STF) banned campaign contributions from corporations. In the previous year, Brazil held its last national election which allowed corporate contributions to political campaigns. In aggregate, the 2014 election moved circa 4.2 billion reais—approximately 1.8 billion US$ according to the average exchange rate in 2014 (1 US$ = R$ 2.3534).Footnote 1 Of this total, Brazilian corporations donated just over 3 billion reais (about 1.3 billion US$), which represents approximately 73% of the overall amount (Mancuso et al., 2018). The Odebrecht group was among the top donors in that election.

Founded in 1944 as a construction company in the state of Bahia, the largest state in Northeast Brazil, 70 years later the Odebrecht Organization had become, according to Exame magazine’s “Biggest and Best” ranking, the fourth largest Brazilian business group in the private sector,Footnote 2 operating in several areas, such as agribusiness (ethanol, sugar, and electricity), construction, defense, engineering, gas, pipelines, naval industry, investments, logistics, oil, petrochemicals, environmental services (water, sewage, and waste), and transportation (roads, railways, subways, and airports). Among all Brazilian construction companies, the Construtora Norberto Odebrecht ranked first in 2014, according to the traditional “Engineering Ranking—500 Largest Construction Companies”, published annually by the magazine O Empreiteiro.Footnote 3 Marcelo Odebrecht, CEO of the conglomerate, then ranked eighth in Forbes magazine′s list of Brazilian billionaires,Footnote 4 with a personal fortune valued at 14 billion reais (5.9 billion US$).

That year, campaign contributions from the colossal Odebrecht group amounted to approximately 88.9 million reais (circa 37.8 million US$).Footnote 5 The strategy adopted was to divide donations among different companies belonging to the group, although not equally. Seven Odebrecht Group companies made electoral donations in 2014: Braskem, Construtora Norberto Odebrecht, Enseada Indústria Naval, Odebrecht Agroindustrial, Odebrecht Ambiental, Odebrecht Óleo e Gás, and Odebrecht Serviços e Participações. In total, just over 17,000 Brazilian private companies made campaign donations in 2014. The Odebrecht Group companies occupied prominent positions on the list of donors as all seven ranked among the quartile of largest donors, while Construtora Norberto Odebrecht ranked seventh in the general position; Braskem at 13th; Odebrecht Oil and Gas at 52nd; Odebrecht Agroindustrial at 229th, Odebrecht Serviços e Participações and Enseada Indústria Naval tied at 362nd, and, finally, Odebrecht Ambiental ranked at 3,558th.

In this chapter we discuss the Odebrecht Organization as a source of official resources for the 2014 election campaigns in Brazil. We compiled the information about campaign donations from the Superior Electoral Court (TSE). The intent is to analyze the strategy of the business group in view of the national and international political science literature. To analyze the proposed case, we conducted a quantitative research, using descriptive statistics, social network analysis (SNA), and logistic and linear regression analysis.

For this purpose, this chapter has been divided into two sections, on top of this introduction and the final considerations. In the next section, we present a brief overview of the political science literature addressing the decision of private corporations to invest in political campaigns. We then analyze, in the following section, the official donations by the Odebrecht group in the 2014 elections, both to candidates and political parties. We must not forget that, according to the Brazilian law in force at the time, business actors such as the Odebrecht Organization were allowed to make electoral donations both to individual candidacies for different public offices (whether in plurality elections, such as president, state governor, or senator; or in proportional elections, such as federal or state deputy), and to party organizations, such as national and state directorates or common financial committees (the latter set up exclusively during the electoral period to organize fundraising activities and distribute donations among individual candidacies for different offices and/or among party directorates at the national or state level).Footnote 6 The limit for corporate electoral donations was not set in absolute values, but relative to 2% of the company’s gross revenue in the year prior to the election, which enabled large donations by major companies, such as the Odebrecht group conglomerate. Lastly, the concluding section summarizes our findings and proposes some ideas for future investigations.

2 The Political Science Literature on Corporate Electoral Financing

In capitalist and democratic countries such as Brazil, major business groups may engage in politics at different levels, whether at the macro level, in which major national issues are discussed; the meso level, which addresses sectoral issues; as well as the micro level, which debates specific issues pertaining to the specific interests of each company (Lowi, 1964).

The major business community has a myriad of tools to engage with politics at these different levels (Schneider, 2010). Some of these tools are collective and institutionalized, such as participation in business associations within formal spaces for debating public policies, often created and maintained by way of government initiatives to gather feedback from society in the policy-making process, among which consultations, hearings, or even a diverse range of collegial bodies such as councils, committees, commissions, chambers, workgroups, forums, etc. (Schmitt, 2020). Other tools for political action are individual, often activated by the initiative of the business community itself, such as lobbying, mobilizing personal influence networks, and financial contributions to electoral campaigns, which form the specific object of our study.

The subject of corporate electoral donation raises important normative and methodological issues, already thoroughly explored in other works and which we will not explore in detail here. Generally speaking, the main normative issue concerns the discrepancy between the democratic value of political equality and the asymmetry of influence engendered by political investments from major corporations, especially when there is no clear rule in place to impose transparency and restrictions to corporate donations, potentially leading to negative consequences such as corruption, distortion of political competition in favor of better-financed candidates, control of the government agenda by the wealthiest, invisibilization of the poorest, and erosion of public trust in democratic institutions (Przeworski, 2011; Norris & van Es, 2016).

Methodologically, the main issues concern mutual determination or reciprocal causation. These issues refer to the two goals which, according to Mueller (2003), political donors may pursue when opting to support certain candidacies: to affect the electoral outcome or the behavior of victorious candidates. Hence, the first question is whether the correlation between funding and performance at the polls exists because donations increase the candidates’ odds of winning or because the candidates’ prospect of winning attracts more electoral contributions. In other words: if we adequately consider all factors affecting electoral performance and campaign financing, would election donations still carry any significant effect on the outcome of the polls? To address this issue, political scientists have continually strived to improve their research designs and explanatory models so as not to neglect relevant variables. Thus far, the literature has indicated a series of interfering conditions in the electoral effect of campaign financing, such as the candidate’s condition of being an incumbent or challenger (according to Jacobson (1978), the effect is greater for challengers), man or woman (for Speck and Mancuso (2014), the effect is greater for women), or the nature of the electoral competition, whether more or less fierce or whether or not there are incumbents in the race (according to Przeworski (2011), the effect seems to be greater in fierce competitions as well as those without incumbents).

Another methodological issue concerns the potential correlation between electoral financing and the behavior of the winners and, if such correlation does exist, whether business donations influence the behavior of elected representatives or whether the projected behavior of elected representatives determines business donations. Once again, to adequately address this challenge we must rely on research design and models that do not neglect relevant variables. A prime example of the divergence surrounding this issue is the controversy between Ansolabehere et al. (2003), for whom campaign donations have minimal effect on how legislators vote, and Stratmann (2005) who argues in the opposite direction, i.e., that electoral financing does indeed affect the vote of parliamentarians.

Without disregarding the implications of the normative and methodological issues described above, in this chapter we shed light on another challenge, namely, deciphering the logic behind corporate electoral financing, or, in other words, understanding a corporation’s decision to engage or not in electoral donations and, when doing so, how much to contribute and how to distribute donations among candidates and parties running for different offices.

The literature on the Brazilian case has cast attention to some possible determinants in the donor behavior of corporations. We will explore the possible impact of these factors on Odebrecht’s electoral contributions in this chapter as we seek to characterize the profile of this business group as a source of political campaign revenue. We will focus on five elements, two of which linked to candidates (political capital and region) and three linked to parties responsible for launching the candidacies (ideology, belonging to the government’s ruling coalition in Congress, and size of the political party).

Starting with the individual factors, some have argued that a candidate’s political capital is an important element in a company’s decision to finance political campaigns (Araújo et al., 2015). The reasoning is that companies would rather direct resources to candidates with more political capital, who have verified political competitiveness and a more familiar political activity profile. Prior electoral success is an important indicator of political capital as well as a relevant—albeit not foolproof—predictor of future electoral success (Speck & Mancuso, 2013; Avelino et al., 2018; Araújo Júnior & Pires, 2020). As a result, companies would rather invest resources in incumbents than to bankroll challengers (Lemos et al., 2010; Marcelino, 2010; Mancuso & Speck, 2015). Brazilian legislation allows indefinite re-elections for legislative offices, but limits re-elections for executive offices (president, governor, and mayor) to a single subsequent period.Footnote 7 Thus, when a previously re-elected president, governor, or mayor can no longer run for re-election, the expectation is that the company will invest resources in candidates who “inherit” the political capital of the incumbent, i.e., candidates who belong to the same party as the incumbent or supported by the same political group. In theory, the strategy of targeting candidates with more political capital would allow a corporation such as Odebrecht to make safer political investments, minimizing the risk of wasting resources on defeated candidacies and maximizing the proportion of money allocated to successful candidacies, thus paving the way for important channels of political influence.

The second individual factor addressed in this work is the candidates’ region of origin. The hypothesis is that the companies’ geographical origin matters, as they tend to foster stronger political ties with candidates in their region, thus making them favored beneficiaries of campaign contributions. The Odebrecht group is originally from Northeast Brazil, and thus the company would have stronger ties with candidates from this region, favoring them when it comes to making donations.

Regarding political parties, the first factor is the ideology of the political group, whose impact on electoral financing has been discussed by the available literature (Samuels, 2001; Mancuso et al., 2016). Candidates and collective organizations belonging to right-wing parties would be more likely to receive donations from Odebrecht, since parties that adhere to this ideological line are traditionally more favorable to business interests.Footnote 8

The literature on political financing has also addressed another variable regarding political parties, namely if the party belongs to the government’s ruling coalition in Congress (Samuels, 2001; Lemos et al., 2010; Mancuso, 2012). The expected effect is ambiguous. On the one hand, the company may favor ruling candidates and political parties out of genuine affinity with the government’s agenda, or simply to please a political coalition which, in the company’s view, is likely to remain in power. On the other hand, the company may consider candidates and parties from outside the ruling coalition, whether due to the company’s own rejection of the government’s line of action or due to suspicions that government supporters will fare poorly in the next election. Thus, we also included this variable in our study to verify whether Odebrecht favored one particular side of the political competition or if it divided resources between ruling and opposition parties.Footnote 9

Finally, the scholarly literature has also considered the size of the political party as another explanatory factor (Mancuso et al., 2018; Mancuso et al., 2021a, b). Several indicators exist for measuring the size of a party, such as the number of affiliated members or the number of municipal and state directories, among others.Footnote 10 The size indicator adopted in our research is the number of seats held by the party in the Chamber of Deputies, on election day.Footnote 11 The representation of Brazilian states in the Chamber of Deputies is not exactly proportional to the population in each state, as the Constitution establishes a minimum of eight and a maximum of 70 federal deputies per state, which leads to overrepresentation of sparsely populated states, mainly from the North region of the country, and underrepresentation of the most populous states, especially the state of São Paulo (Nicolau, 2017). Nonetheless, we justify this methodological decision since, among all existing elections, this election provides the best comparison for the level of support obtained by different political parties at the national level.Footnote 12 By including this factor in our analysis, our objective is to assess whether Odebrecht preferred to donate to candidates and collective organizations belonging to larger parties. As stated earlier, past success in politics often serves as a predictor of future success. In other words, if a party was strong enough to win a significant number of seats in an election, one would reasonably expect it to have advantageous conditions to remain among the largest parties in the next election. As a result, candidates and collective organizations from these parties would attract more funding from pragmatic companies willing to invest their resources in the most promising alternatives.

In addition to these variables, the international literature has also explored the relationship between corporate political financing and the driving motivations and strategies of the business community during election periods. Political scientist Iain McMenamin (2008, 2009, 2012) has provided important contributions to the study of the logic behind corporate political donations. In general lines, the author identifies two possible motivations for the behavior of corporate donors: (i) ideological motivation, which prompts companies to concentrate donations on candidates and parties with similar ideological positions to the companies’ owners, generally further to the right of the political spectrum; and (ii) pragmatic motivation, in which the company defines its donations not according to the ideological convictions of the owners, but rather two criteria: political (to establish ties with those already in power or leading the polls) or economic (to improve the company’s relative position in economic competition). Both motivations manifest themselves through different donation patterns. In the first case, the company donates to parties regardless of whether they are ruling or opposition or their likelihood of winning elections. In the second case, financial support favors ruling parties or parties with likelihood of electoral success. Other authors have explored the same division between pragmatic and ideological donations (Snyder Jr., 1992; Francia et al., 2003; Stratmann, 2005; Goerres & Höpner, 2014). Speck (2016) associates the concepts of ideological versus pragmatic donations to two different moments of exerting influence on the electoral process and elected representatives. Given that it is impossible to maximize both objectives at the same time, the author argues that companies face a dilemma, which manifests itself in the financing strategies during plurality elections. Ideological donors favor the strategy of influencing the electoral process and, to achieve this goal with maximum efficiency, they invest their financial support solely in one party. This strategy, however, risks failing to obtain good relations with the future government in case of defeat of the party supported by the company. Pragmatic donors, in turn, specifically prioritize access to and influence over the future government, regardless of the party’s ideology. Consequently, they finance all feasible parties during the electoral campaign. The cost of this strategy, i.e., supporting opposing parties, is that companies refrain from influencing the electoral outcome. We′ll return to the donor dilemma when we explore Odebrecht’s donation pattern for president and governor elections.

In summary, our analysis explores to what extent characteristics such as the candidates’ political capital, party size, and the relationship with the ruling government influenced the Odebrecht Organization’s donation strategies in the 2014 elections, which would indicate the company’s pragmatic rationale. In turn, if other motivations have dictated the Odebrecht group’s donor logic, we would also find a significant impact of other factors, such as ideology, the company’s relationship with the Rousseff administration, and geographical region. Likewise, when focusing on plurality elections, we interpret the balanced allocation of donations among several candidates as a preference for pragmatism, while unilateral financing of a single party signals a more ideological stance.

3 The Odebrecht Group’s Political Donations in the 2014 Elections

In the 2014 Brazilian elections, the last to allow corporate campaign financing in the country, official campaign contributions from the Odebrecht group amounted R$ 88,895,700.00. Of this total, R$ 22,575,700.00 was allocated to individual candidacies (25.4%) and R$ 66,320,000.00 (74.6%) to collective party organizations, such as national or state-level directories, and common financial committees.

Campaign contributions create ties between donors and recipients, forming relationship networks between them. SNA allows us to observe and interpret such relationships.Footnote 13 The actors of a financing network are nodes (or vertices) in a graph. Here, the nodes are the Odebrecht group, in the capacity of financing agent, as well as the candidates and party organizations that received donations from the company. The donations form the edges (or ties). This configuration enables us to comprehend some network measurements employed in this research, which we present and discuss below.

The degree of a node equals to the number of relationships it establishes with other nodes in the network. Therefore, the Odebrecht Group has a 285 degree, equivalent to the number of actors to whom it donated in the 2014 elections. Given that this is a directional network, it is an out-degree as the company only donates. In turn, the 207 candidates and 78 party organizations, given that they only receive in this network, have different in-degrees equivalent to the number of deposits they received (Table 1).

Table 1 Frequencies and measures of centrality, by office and party organization
Full size table

The degree, however, is an insufficient measure to analyze the network since, for example, a greater number of donations received by an actor may add up to less than the amount of resources received by another actor from a single donation. We must therefore also consider the weighted degree, which is the sum of the donations that exit from or arrive at a given node. Thus, the weighted degree of the Odebrecht Group is the total sum of donations, i.e., R$ 88,895,700.00. Among recipients, national and state directories have the highest weighted degree, followed by presidential candidates.

3.1 Odebrecht’s Donations to Candidacies

As shown in Table 2, the Odebrecht group’s donations to individual candidacies were divided as follows: 56.7% for national candidacies for the presidency of the republic and 43.3% for statewide candidacies, whereas among the latter, candidates for governor received 17.8% of the total; for federal deputy, 13.7%; for state deputy, 9.4%; and for senator, only 2.4%.

Table 2 Odebrecht donations to candidacies, 2014
Full size table

Table 3 shows, in aggregate, the distribution of Odebrecht’s total donations to candidates according to their ideology and relationship with the federal government (i.e., whether they belong to the government’s ruling coalition in Congress). We may already identify some typical indicators of political pragmatism: the Odebrecht group distributes its eggs among various baskets, establishing ties with government and opposition candidates, without prioritizing ideological lines.

Table 3 Measures of the Odebrecht Group’s Financing Network—Candidates
Full size table

We may better visualize Odebrecht’s relational dynamics with political actors through graphs of the financing network. Figure 1 shows the general graph of the relationships between the company and the candidates it finances. The chosen layout is the Circle Pack, which distributes nodes from a hierarchical structure of the research variables. Here, the order of the hierarchy is as follows: 1) ruling-opposition; 2) ideology; 3) proportion of the weighted in-degree.

Fig. 1
A diagram with text in a foreign language depicts the Odebrecht Group Financing Network. A small circle is at the bottom, from which several circles in various sizes and shades branch out. They are labeled with numbers.

Source: Data from the TSE, prepared by the authors

Odebrecht Group Financing Network—Candidacies.

Full size image

The size of the nodes represents the weighted in-degree, i.e., the sum received from the Odebrecht Group. On the left side of the graph are the actors who comprise the support base of the Dilma Rousseff administration; on the right side, the opposition. The hues of the nodes express the positions in the left–right spectrum, as indicated by the caption.

The visualization reinforces the information from the tables above, i.e., we find a balanced distribution of resources between ruling and opposition candidates/parties. Furthermore, the company clearly has few ideological restrictions, cautiously balancing its donations to parties and candidates from the left, center, and right. It is worth adding that candidates from the left, center, and right are present both in the ruling coalition and in the opposition.

We shall now address the Odebrecht Group’s financing pattern according to political office. We will begin with the plurality elections, i.e., for president, governor, and senator.

Table 2 shows that donations to presidential candidates represent 56.7% of the total sum donated to candidacies. This data reveals the primacy that the group attributed to elections for the highest office of the republic. Table 1 and Fig. 1 show that the Odebrecht group was selective and donated only to three of 11 presidential candidates: Dilma Rousseff (PT), Aécio Neves (PSDB), and Eduardo Campos/Marina Silva (PSB).Footnote 14 Broadly speaking, the candidacies represented different positions in the ideological spectrum: Dilma and Eduardo Campos/Marina Silva further to the left, while Aécio was further to the right. As for size, the PT was the largest party in the country while the PSDB and PSB were medium-sized parties. The distribution of resources was not equal but incorporated all candidates with some prospect of electoral success: R$ 8,250,000.00 (64.5%) went to then-president Dilma Rousseff, candidate for reelection; R$ 4,000,000.00 (31.2%) to Aécio Neves, the main opponent; and R$ 550,000.00 (4.3%) to Eduardo Campos/Marina Silva. The group aimed accurately and the ranking of three candidates paralleled their funding: Dilma won the election against Aécio in the second round, and Marina Silva finished in third place in the first round. Odebrecht pragmatically divided its investments among the three main competitors, with different ideological positions, but concentrated the largest donation on the victorious campaign. When distributing its investments, Odebrecht was mindful of the electoral polls, which consistently showed Dilma Rousseff leading the race,Footnote 15 and knew that every president running for reelection in Brazil until then had won the electoral race.Footnote 16 The story was no different in 2014 and, as a result, nearly two-thirds of Odebrecht′s resources went to the winning candidacy.

The Odebrecht group allocated 17.8% of their total sum of investments to elections for governor. Once again, the group was very selective, donating to only 22 of the 176 candidacies across 27 Brazilian states which held elections for governor (12.5% of the total candidacies). In this scenario, the donations followed a more proportional distribution than in the presidential race—18 candidates (81.8%) received R$ 200,000.00; three candidates (13.6%) received R$ 100,000.00, and one candidate (4.5%) received R$ 120,000.00. Of the 22 candidates funded by Odebrecht, nine were elected, which corresponds to 41% of victorious candidacies, a considerable level of success. Nine other candidacies came in second place (41%) and four (18%) finished in third place. The nine victorious candidacies amassed R$ 1,620,000.00, equivalent to 40.3% of the total sum invested in races for governor. The 22 gubernatorial candidacies funded by Odebrecht were in 12 different states. In four states, Odebrecht financed only one candidate for governor. In the other eight states, the company simultaneously financed two or more opposing candidates. The group’s prevailing strategy of distributing eggs across several baskets is once again clearly observable in the governor campaigns. If the company risked undermining its influence over the electoral process by recurrently backing both sides of the contest, on the other hand it improved the likelihood of securing good relations with the newly elected government, whichever it may be, by ranking among its campaign funders.

Seven candidates for governor who received financial donations from Odebrecht (31.8%) were either incumbents or belonged to (or were backed by) the party of the re-elected governor in 2010. The remaining 15 (69.2%) did not fit any of these criteria. Therefore, Odebrecht did not seemingly follow political capital as a criterion for distributing resources among gubernatorial candidates—at least not according to the indicator adopted in our work.Footnote 17 Eight candidates financed by Odebrecht were from the Northeast (36.4%), six from the Southeast (27.3%), five from the Midwest (22.7%), and three from the South (13.6%), which suggests a relatively balanced geographical distribution, even with the absence of financed candidates in the North. Regarding ideology, 13 candidacies funded by Odebrecht were launched by right-wing parties (59.1%), eight by left-wing parties (36.4%), and one by a center party (4.5%). This distribution pattern shows that Odebrecht donated to gubernatorial contenders with diversified and sometimes opposite ideological profiles. Regarding affiliation to the federal government’s ruling coalition, exactly half of the candidacies were either situationists or oppositionists. As for party size, eight beneficiaries belonged to large parties (36.4%), 11 to medium-sized parties (54.5%), and only two were affiliated to small parties (9.1%). Therefore, we find that the group has a clear preference for candidates affiliated to larger parties to the detriment of smaller ones.

Odebrecht invested only 2.4% of their total donations in elections for senate. Brazilians elected 27 senators (one per state) in 2014.Footnote 18 In a more selective approach, Odebrecht benefited only 11 out of 185 senatorial candidates (5.9% of the total), in eight different states, with the same amount of R$ 50,000.00 for each candidacy. Seven funded candidates were elected—a high percentage of 63.6%. Three funded candidates arrived in second place (27.3%) and only one candidate finished in third (9.1%). In five states, Odebrecht financed only one senatorial candidate. In the three remaining states, the group financed two opposing candidates concurrently. In this case, we identify a certain predilection for unilateral financing. However, given the high success rate, the company seems to have opted for candidacies with a high prospect of success.

Among the senatorial candidates who benefited from donations by Odebrecht, five (45.5%) were either incumbents or belonged to (or were supported by) the same party of the senator elected in 2006, while the other six (54.5%) did not have such characteristics.Footnote 19 Therefore, the political capital criterion, as adopted in this chapter, does not seem to have determined the business group’s donations in this particular case. As for region, five of the funded candidates were from the Northeast (45.5%), three from the Southeast (27.3%), two from the North (18.2%) and one from the South (9.1%), which indicates a certain regional dispersion of the group′s contributions, albeit still somewhat concentrated in Northeastern candidates. Six of the funded candidacies were launched by left-wing parties (54.5%) and five by right-wing parties (45.5%), revealing a balance in the ideology category. Three candidates were members of parties comprising the ruling coalition in Congress (27.3%), while eight belonged to parties outside the coalition, whether independent or opposition (72.7%). Regarding party size, three beneficiaries belonged to large parties (27.3%), seven to medium-sized parties (63.6%), and only one candidate belong to a small party (9.1%). Hence, Odebrecht clearly favored the larger parties.

Tables 4 and 5 provide a more detailed look at proportional elections, i.e., for federal and state deputies. As for selectivity, the group donated to only 97 of the 5,813 candidacies for the Chamber of Deputies ratified by the Electoral Court in 2014 (1.7% of the total candidacies). The company was even more selective in the state deputy elections, as it donated only to 74 of the 15,139 candidacies for the legislative assemblies (0.5% of the total). Therefore, Odebrecht carefully selected candidacies before allocating resources.

Table 4 Logistic regression, candidates for federal deputy, 2014
Full size table
Table 5 Logistic regression, candidates for state deputy, 2014
Full size table

As for parity, Odebrecht donated the same amount to most candidates for each office. In the elections for Federal Congress, 90 of the 97 candidates (92.8%) received exactly the same amount: R$30,000.00. Only five candidates (5.1%) received above the modal value. Among them, Samuel Moreira stands out, a candidate launched by the PSDB for the state of Sao Paulo, who received R$ 133,300.00, over 4.4 times the typical amount. Only two candidates (2.1%) received less than usual. Jander Filaretti, from the PTB in the state of Minas Gerais, raised the least funds, earning only R$ 3,100.00, i.e., 89.7% less than most candidates. In the elections for the state legislative assemblies, 64 of the 74 candidates (86.5%) obtained R$20,000.00 from Odebrecht. Nine candidates (12.2%) received over this amount, among them veteran Barros Munhoz, from the PSDB of Sao Paulo, who raised R$ 200,000.00 for his campaign for state deputy, i.e., 10 times more than the most common value. Only one candidate (1.3%) received less than the modal value: Carlos Avalone, from the PSDB in the state of Mato Grosso do Sul, raised only R$1,000.00 from Odebrecht, 95% less than the most common value.

As for the company’s “good marksmanship”, that is, their bet on victorious candidates, 63 of the 97 candidates for federal deputy who received funding from the Odebrecht group were elected—a considerable success rate of 64.9%. Regarding the company’s total investment in political campaigns for a seat in the Chamber of Deputies, 67.5% were allocated to winning candidates (R$ 2,083,300.00, from a total of R$ 3,086,400.00). As for candidates running for state deputy, 46 of the 74 candidates financed by the group were elected—a high success rate of 62.2%. The elected candidates amassed 68.3% of the sum invested by Odebrecht in electoral campaigns for a seat in the legislative assemblies (R$ 1,448,300.00 from a total of R$ 2,119,300.00).

Table 4 shows the results of a logistic regression in which the dependent variable is whether the candidate received campaign contributions from the Odebrecht group in the 2014 elections for federal deputy. The 97 candidates who received donations from the company were marked with a value of 1. The remaining were marked with a value of 0. We included five independent variables in the model, discussed in the previous section, that could influence the company’s donor behavior. The reference profile is an incumbent candidate, from the Northeast, right-wing, member of the government’s ruling coalition, and launched by a large party.

The results in table 4 reveal that three explanatory variables included in the model are significantly associated with candidates for federal deputy having received campaign contributions from the Odebrecht group (we consider the association between variables to be statistically significant when the p value is equal to or less than 0.100). First, the company has a clear preference for experienced candidates, as the likelihood of incumbents having benefited from the company’s campaign contributions is almost 8.9 times greater when compared to challengers. Second, candidates from the Northeast were more likely to receive contributions from Odebrecht than competitors from all other regions of the country. The likelihood of Northern, Midwestern, Southeastern, and Southern candidates to receive donations from the business group correspond, respectively, to 3.8%, 11.3%, 36.3% and 44.3% of the odds for Northeastern candidates. Third, the company favors candidates from the largest party associations: candidates from large parties received more donations than candidates from smaller parties, with a statistically significant difference when compared against candidates from parties without seats in Congress (on average, the odds of a candidate from a party without parliamentary representation receiving a donation from Odebrecht corresponds to only 16.9% of the odds of a candidate from a large party), and against candidates from small parties (in this case, the likelihood of obtaining resources from the corporation corresponds to 21.7% of the odds of a candidate from a large party). The ideology professed by the candidates’ parties as well as belonging to the government’s ruling coalition do not seem to affect the company’s donating behavior.

From the data above, the emerging profile of the Odebrecht group, as a source of funding for candidates running for federal deputy, reveals a politically pragmatic business group which allocates the same amount of resources to the vast majority of selected candidates, not singling out beneficiaries by ideological criteria nor by affinity with the incumbent government, but rather by factors associated with the likelihood of electoral success and the geographical origins of the corporation, favoring candidates for reelection from major parties, especially from Northeastern states. Such pragmatism is probably the most important explanatory factor for the group’s high investment in successful candidacies who, having benefited from its resources, may, at least in theory, during the term of office become more accessible and open to company representatives or more receptive to its interests.

Table 5 presents the results of another logistic regression in which the dependent variable is whether candidates were beneficiaries of campaign donations from the Odebrecht group in the 2014 elections for state deputy. The 74 candidates who received donations were marked with 1, while those who were not beneficiaries marked with 0. The explanatory variables included in this model, as well as the reference profile, are exactly the same as in the previous model.

Once again, running for reelection is the factor most closely associated with receiving donations from Odebrecht: incumbents were 8.4 times more likely than challengers to rank among the company’s list of beneficiaries. Once again, the likelihood of candidates from large parties obtaining electoral contributions from the group exceeds that of candidates from smaller parties, as indicated by the sign of the coefficient, but in this case the difference is only statistically significant with candidates from small parties. The sign of the coefficient also suggests that the regional effect continues to favor Northeastern candidates, albeit less obvious in the elections for state legislative assemblies, as the difference only gains statistical significance when we compare against candidates from the Midwest. Contrary to elections for the Chamber of Deputies, party ideology seems to have some effect in the state legislative elections, as centrists are 2.2 times more likely than right-wingers to gain electoral funding from Odebrecht, while the opposite occurs with candidates from parties without ideological classification, whose likelihood of receiving resources corresponds to only 26.1% of the odds of right-wing candidates. As in the previous model, the candidates’ membership to the government’s ruling collation is not statistically associated with receiving donations from the construction company.

In summary, while the statistical association is less substantial in this case, whether due to fewer recipients or the greater number of non-recipients, Odebrecht′s pragmatic behavior nevertheless remains conspicuous, flavored by regional criteria. Once again, the construction company offered the same amount of money to the vast majority of financed candidates, predominantly consisting of incumbents launched by larger political associations, especially in the Northeast, regardless of their status as supporters or opponents of the Dilma Rousseff administration, yet in this case with a certain predominance of a centrist ideological profile. Although the company invested relatively small amounts in elections for the state legislative assemblies, allocating less than 10% of the total donations to these candidacies, the company’s political investment once again seems predominantly guided by the prospect of electoral success, combined with regional criteria. This seems to have been the group’s leading strategy to raise the proportion of contributions allocated to successful candidates, who could be eventually called upon as a potential channel of political influence.

3.2 Odebrecht′s Contributions to Parties

Table 6 shows the partitioning of donations by the Odebrecht group to collective party organizations (national directorates, state directorates, and common financial committees), aggregated by political party.

Table 6 Odebrecht Donations to Political Parties, 2014
Full size table

According to the data above, the company concentrated 74.2% of its donations to the three largest parties in the country at that time—PMDB, PT, and PSDB—which received, respectively, 29.3%, 23.8%, and 21.1% of the total donations. The company distributed the remainder among 19 parties. Ten political parties received nothing from the group.

Table 7 shows the distribution of donations among parties, aggregated according to their relationship with the government’s ruling coalition and ideology. Although we find a predominance of donations to ruling and right-wing parties, opposition parties and left-wing parties also received a significant sum of electoral contributions.

Table 7 Measures of the Odebrecht Group Financing Network– Political Parties
Full size table

The graph in Fig. 2 provides a visual representation of the relationships between the Odebrecht company and party organizations, with the same previous layout for candidates (Circle Pack) and the same hierarchical distribution for the nodes, i.e.: 1) ruling-opposition; 2) ideology; 3) proportion of the weighted in-degree.

Fig. 2
A diagram with text in a foreign language depicts the Odebrecht Group Financing Network, Party organizations. A small circle is at the bottom, from which several circles in various sizes and shades branch out. They are labeled with numbers.

Source: Data from the TSE, prepared by the authors

Odebrecht Group Financing Network—Party Organizations.

Full size image

The figure reveals a clear predominance of the largest and most important political parties in Brazil at that time. The figure shows the allocation of resources between ruling and opposition parties and according to their positioning in the left-right spectrum. Ruling parties are to the left of the graph while opposition parties are to the right, and the different hues indicate ideology.

Table 8 presents the results of a multivariate linear regression analysis, in which the dependent variable is the amount Odebrecht donated to political parties (cf. Table 6) and the explanatory variables are ideology, relationship with the government’s ruling coalition, and party size (see criteria in Tables 4 and 5). The reference profile is a right-wing, ruling, large party.

Table 8 Linear regression, parties, 2014
Full size table

The results once again indicate the Odebrecht group’s political pragmatism. The only explanatory variable with a statistically significant impact is party size. When compared against large parties, medium-sized parties received, on average, circa R$ 15.4 million less from the construction company, with further disparity in the case of small parties and parties without parliamentary representation, which received on average, respectively, R$ 18.3 and R$ 18.7 million less than large parties. The status of being a ruling or opposition party, as well as the party’s ideological position, were not factors that significantly affected the donor behavior of the business group. In other words, the construction company’s strategy concentrated its stakes on the largest Brazilian parties during the 2014 elections, wagering on their current strength as the optimal predictor of their future strength, hopeful that the company’s donations would prompt the country’s major parties to be more receptive to the interests of the business group.

4 Final Remarks

In sum, we may characterize the Odebrecht group’s donor behavior in the 2014 elections as largely pragmatic. The group was in fact extremely selective when choosing recipients for its campaign contributions, prioritizing candidates and parties with greater odds of electoral success, with a view to enhance the company’s future prospects of political influence. As for the criteria investigated in this chapter, Odebrecht mainly adopted the criterion of party size, as the company preferably donated to candidates and collective organizations belonging to large and medium-sized parties rather than smaller political associations. Another pragmatic criterion adopted by Odebrecht was their clear-cut preference for financing incumbent candidates in presidential elections as well as in elections for the Chamber of Deputies and for state legislative assemblies. Political pragmatism yielded good results. When financing candidacies, the company revealed exceptional marksmanship and obtained high success rates across all elections for different offices, consistently allocating a large share of donations to winning candidates. The company had a uniform approach when apportioning resources to candidates in statewide plurality and proportional elections, allocating a similar sum of resources to competitive gubernatorial candidates (many of which competing against each other), to the vast majority of candidates for federal and state deputy, and to all senatorial candidates. In the presidential election, the company concentrated its resources on the victorious candidacy of Dilma Rousseff, while also supporting the main contestants, allocating significant amounts to the runner-up candidate. In the plurality elections for federal and state executive offices, faced with the dilemma of influencing the electoral process or ensuring good relations with elected governments, Odebrecht mainly opted for the latter.

Other motivations had a lesser impact on Odebrecht’s donor behavior. The group generally did not prioritize factors such as the ideological profile of contestants or the parties’ association with the federal government. In most cases, the construction company financed ruling, independent, and opposition candidates and parties alike, positioned across different ideological alignments within the left-right spectrum. The regional criterion only becomes clearer in proportional elections, especially for federal deputy, when the group showed a clear preference for Northeastern candidates.

In this chapter, we sought to contribute to the political science literature on the logic behind the use of corporate donations to fund electoral campaigns. We hope this work will encourage similar studies on other business groups in Brazil as well as in other countries.

Our investigation focused exclusively on the Odebrecht group’s official donations in the 2014 elections. However, we remain tasked with the challenge of investigating illicit corporate financing of political campaigns, i.e., funding procedures that do not comply with the current legal system. This type of financing is at the root of major political scandals in Brazil and worldwide (Mendilow & Phélippeau, 2021). In fact, 2014 also marked the launch year for Operation Car Wash, a large-scale investigation by the Federal Prosecutor’s Office into a massive corruption scheme involving Petrobras, Brazil’s giant oil state company. The scheme involved, on the one hand, major companies supplying goods and services to Petrobras, especially in the engineering and construction sector, among which the Odebrecht group, and, on the other hand, some directors of the state-owned company appointed to office through referrals by political parties belonging to the federal government’s ruling coalition in the National Congress. Petrobras directors would then surreptitiously transfer a share of the unlawfully appropriated money to political parties, which referred their nomination, to use at their discretion, including funding electoral campaigns. Analyzing major scandals such as the Car Wash scheme allows for a better understanding of the intricacies and modus operandi of illicit electoral financing. In this same book, the chapter by Costa et al. (2020) provides an important contribution in this direction as they address how illegal political financing, as much as its relationship with Brazilian democracy, has been described by former executives of the Odebrecht group, who signed plea bargain agreements amid the Operation Car Wash investigation.

Furthermore, illicit electoral financing not only lies at the heart of major corruption scandals, but reemerges, through different prisms, amid customary practices in election periods throughout the country. Hence, our current research agenda also seeks to analyze the impact of this issue on the judicialization of electoral competition in Brazil (Mancuso et al., 2021a, b). We believe this is another promising path to overcome the persisting shortfalls in the study of corporate campaign funding.