Matias López, Graduate Institute of Geneva
The most accepted academic framework for understanding corruption accounts for it as the result of economic incentives: If elites are able to extract rents from access to government, they are likely to foster public spending in order to boost returns from corruption, sabotaging governance all together. However, such an understanding of rent seeking overlooks the role of party programs in mediating the incentives of corruption. Reporting from a project on elites’ policy preferences in Brazil, I theorize that strong programmatic ties prevent the economic effects of corruption as modeled in recent literature. In the case of Brazil this is highly consequential because anti-corruption policy has targeted programmatic parties more than non-programmatic parties, which in effect may have increased the negative effects of corruption, even if it is true that the average level of corruption was reduced. I address this possibility by looking at elites’ preferences regarding public spending and how they may or may not relate to corruption.
Introduction
Corruption in Brazil is often referred to as endemic, as it continues to reproduce despite the fast pace in which the country institutionalized its democratic regime (Hagopian 2016). After being celebrated for its success in reducing poverty and fostering economic growth, the PT (Workers Party) became stigmatized for corruption in government contracts in alliance with other parties. Among other crimes, a wide scheme known as Petrolão funneled billions of dollars to a cartel of partisan and business elites. The way Petrolão used to work is illustrative of how corruption greases the wheels of Brazilian politics. Party appointees from PT, PMDB, and PP inside Petrobrás received bribes in order to privilege a cartel of companies, including giants such as Andrade Gutierrez, OAS, and Odebrecht (on the latter see Pohlmann et al. 2019). Such companies then channelled part of their inflated profits back to political parties in the form of regular and off-the-books campaign donations, forming a vicious circle.
The reason Petrolão was exposed is because of Lava Jato (Operation Car Wash), a judicial task force implemented in Brazil between 2014 and 2021. Using a hardball approach, it jailed a share of the country’s elite, including the former president Luis Inácio Lula da Silva who at the time was favourite in the 2018 presidential elections. Because the cartel of companies was active before PT’s rise to power, and also in states’ governments run by the opposition, investigations ended up revealing the corruption of PT’s main rival, the PSDB, as well. Although involved in corruption, both PT and PSDB are highly programmatic parties in opposite poles of the ideological spectrum. By targeting corruption in the two most programmatic parties of Brazil, Lava Jato cleared the way for non-programmatic and populist alternatives (see Santos and Tanscheit 2019). In its aftermath, both PT and PSDB diminished their electoral appeal and Jair Bolsonaro, a far-right populist leader, was elected president against all odds.
It should also be noted that leaked communication between Lava Jato judges and prosecutors revealed how authorities within the task force colluded in order to meddle with the 2018 elections, exposing their own corruption. Many attribute the downfall of Brazil to Lava Jato’s political project. Regardless, the arguments put forward here portray that the negative effects of Lava Jato might have been felt even if the taskforce had been lawful and politically neutral in its usage of constitutional hardball to fight corruption.
In this note I hypothesize that fighting corruption may have increased the negative effects of corruption. I developed this hypothesis while modelling survey data on elites in the context of the project “How elites shape unequal democracies” at the Albert Hirschman Centre on Democracy. Models of corruption in academic research predict that the extraction of illegal rents by elites leads to inefficient investment, as funds are inflated in order to allow for more extraction (Fukuyama 2014, Aidt 2009, Uslaner 2014, Rothstein and Teorell 2014). Consequently, corruption is assumed to block economic growth. Corruption is also assumed to be detrimental to democracy, as corrupt elites have incentives to erode oversight and accountability. However, fighting corruption can also be costly depending on how it is implemented.
First, the detrimental effects of corruption should not be the same for cases of programmatic and non-programmatic parties. Both PT’s and PSDB’s main leaders, while involved in corruption schemes, should be unlikely to compromise their long-term projects in benefit of personal economic gains. This should not be true for non-programmatic parties. With no clear long-term programmatic goals, the elites in smaller non-programmatic parties have stronger incentives to loot the government. For instance, the current president Jair Bolsonaro, a prototypical non-programmatic politician, is known for having hired façade members of staff while in Congress in order to cash in their salaries. In his 30 years of service, Bolsonaro only passed two bills. The inefficiency of his cabinet can be attributed to the fact that its budget was indeed going to Bolsonaro’s own pocket. Without the counterweight of party programs, the negative effects of corruption are more acute and compromise both economic activity and democratic development.
Why elites steal?
Several researchers assume that elites steal because they can. Because corrupt elites profit from unsupervised public spending, researchers predict that elites will find incentives for inflating budgets as they benefit from corruption (see also Fukuyama 2014, Rothstein and Teorell 2014). For this reason, corruption is acknowledged as ultimately having negative consequences for democratic governance and economic development (Aidt 2009, Alesina and Angeletos 2005, Uslaner 2014).
Recent scholarship has started to question this purely economic perspective. By analyzing clandestine administrative books in a recent bribing scandal in Argentina, Figueroa (2020) notes that the distribution of elites’ intake of bribes peaks in the weeks prior to elections and declines after elections. This distribution is highly unusual in a purely economic perspective, as elites should pose as honest on the verge of elections and take the opportunity of stealing after winning. Figueroa argues that the pattern found makes sense from a political point of view, because partisans need more resources during elections. This suggests that rents obtained illegally through bribes were channelled mainly to politics, and not to elites’ own pockets. It becomes apparent that partisan elites do not steal only because they can, but also because they need resources to stay ahead in the electoral game.
This mechanism of corruption seems to fit the case of PT’s cartel of parties and private companies. Lava Jato prosecutors tried very hard to prove that Lula had enriched out of the schemes at Petrobrás. They did so by claiming that the OAS company had an apartment reserved for him in the company’s name and by claiming that improvements in a friend’s ranch were in effect disguised bribes to Lula. Both cases are very hard to sustain, as neither account for actual cash flow to Lula. On the other hand, if we assume political rather than economic incentives, the theory that Lula and other party members okayed corruption without benefiting economically from is plausible. The bribing schemes financed campaigns and disciplined non-programmatic allies in Congress.
What more recent research shows is therefore that political elites might steal or allow others to steal for non-economic reasons, and ultimately for programmatic reasons. In such a case, the effect of corruption on public spending, as anticipated by models based entirely on economic returns, would be severely mitigated as it would be counterweighted by programmatic calculations.
Economic preferences within parties
Parties are distributed through an ideological spectrum between left and right. One can assume that left wing partisans are more inclined to social spending and taxation in order to generate more equality. Meanwhile, right wing parties tend to prefer more austere administrations and less taxation. This roughly applies to Brazil when one looks at parties that are clearly programmatic, such as PT and PSDB. The question is whether party elites would deviate, or have deviated, from these preferences due to the prospects of corruption.
Given the multi-party nature of Brazil’s party system, neither PT’s nor PSDB’s national or local administrations could govern without a broad coalition of parties. A few other programmatic parties in the left, such as the Communist Party (PCdoB), were by far insufficient for a stable PT administration. Likewise, the commitment of other programmatic right-wing parties, such as DEM (formerly PFL), could not sustain PSDB’s administrations. Both parties needed to ally with smaller non-programmatic parties, also known as centrão parties in Brazilian politics. The PMDB, PTB, PP, PSD and other parties from centrão have very loose commitments to any particular economic program. As such, they allied with ideologically opposed administrations in exchange for influence and pork. These parties are central in the corruption schemes at Petrobrás and elsewhere in government.
In my current research I am looking at campaign donations to parties from the cartel of companies prosecuted by Lava Jato. I contrasted the latter with the average attitude toward spending and taxation in each party using data from the Brazilian Legislative Survey or BLS (Power and Zucco 2012). The survey asked participants to disclaim their preference regarding spending and funding relative to different policies[i]. I combined all policies in a squared additive index which accounts for parties’ average inclination to endorse more spending.
Figure 1 shows the distribution of mean preferences for public spending in contrast to the donations per seat received from companies framed by Operation Car Wash as corrupt. I validated my coding of parties as programmatic and non-programmatic using V-Party’s clientelism index[ii]. The distribution portrays how the amount of donations and the preferences for more spending go hand-in-hand among non-programmatic parties but not among programmatic parties.
Among non-programmatic parties, those that received more substantial donations from corrupt corporations, such as PP and PMDB, scored much higher in their preferences for public spending. Furthermore, the relationship between donations and openness to endorse more spending is almost strictly linear. Among programmatic parties, the trend line is much less effective in predicting the average attitudes of party leaders. Their ideology seems to better explain average attitudes toward spending, with the left-wing parties PT and PCdoB showing the highest scores among all parties.
Conclusion
Non-programmatic party elites seem more inclined to adjust policy preferences in accordance with their expectation of rents. By targeting the corruption of programmatic parties, mainly the PT and PSDB, anti-corruption efforts strengthened non-programmatic parties which ultimately formed Bolsonaro’s base in Congress. It may or may not be true that the net amount of funds in corruption schemes has overall decreased after Lava Jato’s spotlight on Petrobrás and other government agencies. However, it may also be the case that the persisting corruption is more detrimental now with the removal of programmatic guidance for public spending and investment.
Whereas PT famously implemented bribes in Congress in order to pass its policy agenda, Bolsonaro is now openly bribing MPs with slices of the federal budget with no clear policy goal other than preventing his impeachment. Meanwhile Bolsonaro’s sons multiply their assets. If anything, the removal of PT and weakening of PSDB might have strengthened, rather than weekend, the inner mechanisms of corruption that compromise development.
References
Aidt, Toke S. “Corruption, institutions, and economic development.” Oxford review of economic policy 25, no. 2 (2009): 271-291.
Figueroa, V. (2021). Political Corruption Cycles: High-Frequency Evidence from Argentina’s Notebooks Scandal. Comparative Political Studies, 54(3-4), 482-517.
Hagopian, Frances. “Delegative Democracy Revisited: Brazil’s Accountability Paradox.” Journal of Democracy 27, no. 3 (2016): 119-128.
Pohlmann, M., Valarini, E., Trombini, M., & Jorge, M. (2019). Systemic Corruption in Brazil: An Autopsy of the Odebrecht Case. Journal of Self-Regulation and Regulation, 5, 55-78.
Power, T. J., & Zucco Jr, C. (2012). Elite preferences in a consolidating democracy: the Brazilian legislative surveys, 1990–2009. Latin American Politics and Society, 54(4), 1-27.
Santos, F., & Tanscheit, T. (2019). Quando velhos atores saem de cena: a ascensão da nova direita política no Brasil. Colombia Internacional, (99), 151-186.
Uslaner, E. M. (2014). The consequences of corruption. In Routledge handbook of political corruption (pp. 213-225). Routledge.
Rothstein, Bo, and Jan Teorell. “Causes of corruption.” In Routledge handbook of political corruption, pp. 93-108. Routledge, 2014.
Notes
[i] The policies are public basic education, public higher education, infrastructure, social assistance, and health. The options for each were a) reduce spending to reduce taxes, b) reduce spending without changing taxes, c) keep spending and taxes as they are, d) increase spending without increasing taxes, e) increase taxes in order to increase spending.
[ii] data available at www.v-dem.net. The threshold in my coding validation was clientelism index>1= non-programmatic.
