Measuring the Success of Brazil’s 2014 Anticorruption Law

In January 2014, following nationwide protests prompted by concerns about widespread corruption, the Brazilian Congress enacted a new Anticorruption Law. The 2014 Anticorruption Law was a landmark in part because it represented the shift from the traditional focus on individual bribe-takers to the bribe-paying corporations. Although a few prior statutes did address the conduct of alleged bribe payers, the Anticorruption Law both authorized much more stringent penalties on bribe-paying companies (including fines of up to 20% of a company’s gross revenue in the prior year, and even mandatory dissolution of the company in extreme cases), and also adopted a revolutionary strict liability regime for corporate corruption offenses.

More than six years have passed since the Brazilian Anticorruption Law entered into force. Has the law been effective? What do we know so far about its enforcement? On the whole, the enforcement numbers seem rather disappointing, suggesting that the law has not (yet) been deployed aggressively to sanction bribe-paying corporations in most parts of Brazil. Nonetheless, there are a variety of reasons why it would be premature to conclude from these numbers that the law has not been, or will not be, effective.

Let’s start with some statistics on the law’s enforcement. Administrative proceedings under the Anticorruption Law are called “PARs” (processos administrativos de responsabilização), so the number of PARs is, in effect, a measure of the number of cases (both open and closed) brought government agencies to enforce the act. A prior study found that by January 2018 (four years after the law was enacted), the Federal Government had opened 183 PARs and punished 30 legal entities. (This might be misleading, though, because the federal government also uses the PAR designation to refer to proceedings under other statutes). But looking only at federal enforcement statistics is insufficient, because the Anticorruption Law also authorizes enforcement by state and municipal governments. To gather more data about subnational enforcement, last year I led a research team that compiled information about the statute’s enforcement by the 27 Brazilian states (including the Federal District) and 26 capital cities, using a combination of Freedom of Information Act requests and publicly available data. Though our expanded database is still incomplete (in part because some jurisdictions simply did not respond to our requests), this database allowed a much more in-depth evaluation of the statute’s enforcement.

We found that as of January 2019 (five years after the statute came into force), Brazilian states had opened 132 PARs against 306 legal entities, and capital cities had opened 42 PARs against 49 legal entities. But these aggregate numbers mask substantial variation across jurisdictions. Of the 27 states, eight had not even enacted the necessary implementing regulations that must be in place before state agencies can enforce the Anticorruption Law in their jurisdictions. And of the 19 states that had adopted such regulations, nine had still not opened a single PAR. Moreover, only 24 of the 132 state PARs had resulted in punishment of legal entities. The story is similar with the capital cities: Out of 26 capital cities, 17 had not yet adopted the necessary implementing regulations, and of the remaining nine, the city of São Paulo was responsible for the vast majority of the enforcement actions: 34 of the 42 capital city PARs were in São Paulo, as were 22 of the 23 administrative orders that punished legal entities. Furthermore, the number of PARs and sanctions imposed is not necessarily a good indicator of the use of the Anticorruption Law against bribery, since the law, despite its name, proscribes other acts as well, such as bid rigging and related wrongdoing. Indeed, our research found that only a minority of PARs by states (21 out of 132) actually concerned bribery, while 94 concerned bid-rigging. And out of the 42 PARs opened by capital cities, only two related to bribery allegations, while 18 concerned bid-rigging.

While it is certainly troubling that many Brazilian states and municipalities do not seem to be aggressively employing the Anticorruption Law to sanction bribe-paying corporations, we should hesitate before we conclude from this that the statute is not accomplishing its goals. Indeed, while the enforcement numbers may be helpful as a starting point, they cannot by themselves tell much us about the success or failure the Anticorruption Law. There are several reasons for this:

  • First, the Anticorruption Law does not apply to conduct that took place before the law came into force on January 29, 2014. That means, for example, that the Law doesn’t apply to the entities investigated for their involvement in the Lava Jato (“Car Wash”) scandal, which took place before the Law’s effective date.
  • Second, anticorruption enforcement in Brazil is not limited to the Brazilian Anticorruption Law. Prosecutions of individuals take place under other sections of the criminal code, and Brazilian prosecutors also make use of other statutes, such as Administrative Improbity Law, to take action against legal entities involved in corruption cases.
  • Third, focusing only on enforcement of the substantive terms of the Anticorruption Law overlooks another important feature of the statute: its authorization of leniency agreements. The Anticorruption Law established the legal framework for this settlement instrument, which can be and has been used to resolve corruption-related cases that do not fall under the scope of the Anticorruption Law. Indeed, many of the most important cases in the Lava Jato investigation, including those of Odebrecht and J&F, were resolved pursuant to the leniency agreements authorized by the 2014 Anticorruption Law.
  • Fourth, it is simply too soon to reach any conclusions on Brazilian Anticorruption Law’s effectiveness. Illustratively, the U.S. Foreign Corrupt Practices Act (FCPA) took more than 20 years to produce significant results. Prior to 2001, the US government had never initiated more than five FCPA enforcement actions in a single year.

So, while the numbers are suggestive—and do tend to indicate less aggressive enforcement of the Anticorruption Law than proponents may have hoped—we should be careful not to exaggerate their significance.

Still, there is room for improvement. One area that deserves attention concerns the fact that the administrative agencies in charge of enforcing the Anticorruption Law lack several investigative methods commonly used in corruption cases, such as telephone intercepts and plea bargains for cooperating witnesses. In Brazil, these tools are only available in criminal investigations, and since enforcement of the Anticorruption Law is administrative, those charged with enforcing the Anticorruption Law often depend on evidence produced in related criminal investigations. In practice, this means that the Anticorruption Law’s effectiveness depends on different enforcement agencies working together, or at least exchanging information effectively. Unfortunately, Brazilian agencies have been known to compete among themselves for the spotlight in the fight against corruption. Overcoming that problem – and encouraging agencies to work together towards a common goal – will be crucial to the long-term success not only of the Anticorruption Law, but of Brazil’s broader anticorruption struggle.

9 thoughts on “Measuring the Success of Brazil’s 2014 Anticorruption Law

  1. Luiz, thanks so much for obtaining this data and presenting your observations in such an informative and compelling way! It does seem too early to draw too many inferences regarding the law’s effectiveness. Do you have any theories as to why cities besides São Paulo lag so far behind in enforcement actions? Is it a matter of political will, resources, expertise, or other factors? Also, do you have any speculation as to why bid-rigging rather than other prohibited acts generated most PARs at the city and state levels? Is this just an accurate reflection of the prevalence of the bad acts, an issue of the burden of proof, or something else?

    • Thank you, Laura! I’m glad that you enjoyed reading it. When I concluded the research, I had some interesting conversations with some heads of local departments in charge of enforcing the Anticorruption Law. My impression is that some cities and states simply chose to empower their administrative agencies, probably following strong anti-corruption political agendas, while others did not. Regarding your second question, bid-rigging is generally easier to investigate than bribery, especially considering (i) that corruption investigations usually require the use of investigative methods that are not available in the administrative sphere and (ii) the broad statutory definition of bid-rigging under the Anticorruption Law, which can encompass minor negligent acts or mistakes in connection with public bids that nonetheless may be deemed detrimental to the public administration pursuant to a strict liability regime. It seems hard to think of a bribe paid or received because a party was negligent, so the strict liability standard does not really come into play in such cases.

  2. Luiz, a great and important message to prematurely judge the Anticorruption Law. I would also love to know what you think prospective outcomes of the Anticorruption Law are– though it’s too early to say that things haven’t worked, what might be some factors we can look to in predicting outcomes? How does this compare with anticorruption enforcement mechanisms in the past? How do the leniency agreements factor into how successful you anticipate this law will be in tamping down corruption?

    • Thank you, Jaylia! These are crucial questions. As a general takeaway, we cannot underestimate the deterrence component of the Anticorruption Law. In addition to the traditional fear of the statute’s heavy liabilities, it also promotes enhanced deterrence via the incentive for legal entities to implement and maintain anti-corruption compliance programs. Anticorruption enforcement relies on ripening an ethical business culture that the number of enforcement actions simply cannot measure, no matter how high (or low) they are. Either way, enforcement of the Anticorruption Law still needs some tinkering and perhaps even some reforms, but the statute is inequivocably finding its way. Just last month, the Office of the Comptroller General of the Federal Government announced its highest penalty yet in a corruption case governed by the Anticorruption Law (https://www.time24.news/2020/10/vivo-responds-to-the-anti-corruption-law-and-is-fined-almost-r-46-million.html), and there’s definitely more to come. Still, it is hard to compare with anticorruption enforcement mechanisms in the past because the traditional trend of corruption enforcement (criminal action) persists as the core of anticorruption enforcement. The Anticorruption Law does not change how corruption cases can or should be investigated, it merely opens a door to seek direct action against legal entities under a strict liability regime (there is no corporate criminal liability in Brazil, and the other statute that allows civil action against legal entities required proof of intent). The difference is that all these liability regimes co-exist. Hence, the government has the means to go after both individuals and legal entities engaged in corruption instead of limiting its action to the individuals. Finally, I believe the idea of a leniency agreement regime is remarkable yet poorly executed in the anticorruption scenario. A recent blog post discussed the shortcomings of the leniency agreement provisions quite extensively (https://globalanticorruptionblog.com/2020/02/03/the-shortcomings-of-the-leniency-agreement-provisions-of-brazils-clean-company-act/), and I intend to write a separate post addressing one particular flaw of this regime (that the statute does not allow the parties to negotiate assumption of responsibility). This is a particular aspect of the new legislation that has not been as effective as it was expected to, and this scenario is unlikely to change without significant reforms.

  3. This is an incredibly interesting post, thank you for writing it! I would like to hear more of your thoughts on the leniency program under the Anticorruption Law. Is it like the leniency program we have under the DOJ Antitrust program where, in an anticompetitive conspiracy, the first company who breaks from the pact is granted leniency or certain defenses? This goes back to the zoom conversation we had about the efficacy of leniency programs in general and how to best get cooperators to assist in the investigation.

    • Thank you, Vincent! I’m glad you enjoyed reading it. You are correct – the leniency program under the Anticorruption Law resembles the DOJ Antitrust program in many ways, including the theoretical restriction of its benefits to the first company that states its interest in cooperating. I invite you to think about whether this is a limitation that also makes sense in anticorruption enforcement, because it has been consistently “ignored” by Brazilian agencies. A recent blog post described some of the shortcomings of the Anticorruption Law’s leniency program (https://globalanticorruptionblog.com/2020/02/03/the-shortcomings-of-the-leniency-agreement-provisions-of-brazils-clean-company-act/), and I intend to write separately about another feature of the regime that I consider to be skewed if the program is intended to be effective (the parties’ inability to negotiate assumption of responsibility).

      • You may be addressing this in your next post, but based on what I read in the blog post you linked, it seems less that the agencies are ignoring the leniency agreement and more like the companies themselves are not trying to engage in any leniency agreement because anticorruption enforcement seems scattered in Brazil. To a certain extent, U.S. criminal antitrust actions can be considered a subsection of anticorruption (like rent-seeking, bid-rigging, etc.) and the benefit in the U.S. is that there is only one federal government agency that a conspirator has to deal with. The post dealt with the Clean Company Act, so I am curious if other laws in Brazil face the same problem where multiple agencies have the authority to enforce the anticorruption statutes.

  4. Great and extremely interesting post! I agree with you that it may still be early to tell the effectiveness of the 2014 Anticorruption Law. However, I would love to know whether we could predict a timeline of when we will begin to see major movement. You mention that the FCPA took more than 20 years for significant results to be produced. I would love to learn more about why that occurred and the applicable lessons we could learn from the FCPA. For example, were there critical deficiencies that were later addressed and if so, what can Brazil do to cure any similar deficiencies? Conversely, were there extrinsic factors that led to the delay in significant results with the FCPA and if so, are those factors present within the 2014 Anticorruption Law that could or could not be addressed? Admittedly, these are questions we may never find the answers for.

  5. This is a great read. I agree with you that it would be premature to judge the total effectiveness of the Brazilian Anticorruption Law so soon, especially since it took twenty years for the United States’ FCPA to gain traction. Since, as you stated, the number of cases brought is not sufficient to show the effectiveness of the Brazilian Anticorruption Law, maybe other data could help shine light on whether the anticorruption efforts are effective. The first that comes to mind for me would be looking into economic data to see if the economy of Brazil has changed. Corruption leads to many negative effects within a domestic economy. Thus, if we see positive economic trends, this may be able to indicate to us a trend in the effectiveness of the government’s fight against corruption.
    One set of readily available data are the published figures on economic growth of Brazil. While there are many factors that go into aggregate economic growth within a country, many of these factors are directly affected by corruption. We could look to the change in GDP to see if the economy is growing at a quicker rate since the Anticorruption Law was enacted or if it is moving closer to the same rate. GDP per capita would also be a strong indicator because it would take into account potential population growth within the country, as well.
    Another piece of data to look at would be the foreign and domestic investment within Brazil. Corruption leads to a decrease in investment within the country. This seems pretty obvious, countries that have more corruption are less desirable to invest in. The corruption itself raises the cost of potential investment, but it also raises the uncertainty of that investment because the corruption makes everything more unpredictable. We could use this information to see if there is a decline in corruption. Investment in Brazil increasing could mean corruption decreasing.
    Lastly, it may be helpful to look at the public infrastructure and tax revenue. When corruption is a large problem in a country, public expenditures are diverted to private uses. Further, if there is corruption in bidding or in maintaining the codes and standards of public infrastructure, there would a visible difference in the quality of the public infrastructure. Overall this leads to less money going into public infrastructure and the quality of the infrastructure itself being lower. In theory, effective anticorruption laws should reverse these effects, so looking into the public infrastructure both in the amount invested in it and its quality could be helpful.
    Corruption can also lead to decreased tax revenue because much of the money exchanging hands in the country is not done through legal means, which means it is untouched by tax policies. Corruption can lead to tax officials even getting paid off instead of the taxes themselves getting paid. Increased tax revenue could be a sign that more money is exchanging hands through legal methods and that corruption is on the decline.
    With all of these statistics, finding any way to focus on the most impoverished people or most impoverished areas as a whole would also strengthen the prediction because corruption disproportionately effects impoverish people and impoverished areas within the country. Impoverished people are disadvantaged by petty corruption that may be harder to detect through other data. Petty corruption can lead to the poorest people not being able to receive the proper protection from police and can also lead to a lack of access to education and medical care.
    Impoverished areas as a whole are also affected. These areas may lack access to fresh water or proper sewage systems. These areas may not receive the funding they need for these basic needs because corrupt governments tend to focus their public spending on projects that are more capital intensive because doing so allows for greater kickbacks. Focusing on data that covers impoverished people or impoverished areas could be useful for seeing the effects that a law has on corruption, especially smaller petty corruption, within a country.
    Again, I really appreciated your blog post and really appreciate you highlighting how long it may take to see results from anticorruption laws. Examining economic data could help shed light on the potential effectiveness of the laws being enacted. My analysis of economic effects of corruption was informed by the report Corruption and Poverty by Chetwynd et al.

Leave a Reply to Laurel Zigerelli Cancel reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.