Macro-corruption and the Lava-Jato Case: A Criminal Networks Analysis
Luis Jorge Garay-Salamanca, Eduardo Salcedo-Albarán, and Guillermo Macías
Social Network Analysis (SNA) has become an approach that facilitates organizing, systematizing, analyzing and visualizing criminal structures. The set of methodologies and protocols sustained on SNA, referred herein as Criminal Networks Analysis (CNA) provides an accurate and succinct way to analyze information about facts and social dynamics between both individuals and groups. It has stood out for being a fundamental tool in developing strategies for dismantling networks, especially in extensive cases that are difficult to visualize, such as those corruption schemes in which it is necessary to understand actions and interactions between full-time criminals and gray agents. This article defines and describes some strategies used to analyze the criminal structure known as “Lava Jato”, a transnational macro-corruption network. Likewise, the main results of this analysis are also discussed.
In March 2014, the Federal Police and the Public Prosecutor’s Office of Brazil initiated an investigation, still ongoing, to identify and dismantle corruption and money-laundering schemes. Since its beginnings and until mid-2018, in the framework of this operation 51 investigation stages have been developed, revealing the participation of private and state-owned companies, high ranking officials, politicians, businessmen, financial operators and drug-traffickers, in an operating area of 47 countries. Considering the quantity and variety of the involved agents, the large amounts of money, the number of documentary evidence, the number of declarations and the number of involved countries, this case goes beyond most conceptual and methodological postulates within the judiciaries, on a global basis.
On the one hand, from a conceptual perspective, the notion of organized crime adopted by most nations in their criminal codes, limits the composition of a criminal structure to a group of three or more individuals who operate in a hierarchical organization with strictly defined roles, over a period of time, to carry out a criminal goal that needs to be realized through serious illegal behaviors, defined by specific laws in each country. These definitions are often an obstacle for analyzing complex structures because they restrict the compositions of such structures to full-time criminals who play defined roles and are subordinated to a single leader that concentrates the decision-making of every action performed by the organization. Furthermore, the restriction, and its subsidiarity to certain criminal classifications considered as severe by the Penal Code, hinder the understanding of the scope of modern criminal structures.
On the other hand, from a methodological perspective, limitations in the established concepts, cause several epistemological obstacles for investigating these phenomena. First, by leading researchers, investigators, and prosecutors, to restrict their analysis to full-time criminals, these limitations cause a biased image of the operation and scope of the structure’s actions and effects. Second, they also restrict the analysis of behaviors and acts to serious criminal offenses, since the penal definition of association per se is subsidiary to the commission of crimes such as kidnapping, drug trafficking, human trafficking, arms trafficking, prostitution and, just in some cases, corruption. Due to this conceptual restriction, activities that are not necessarily illegal but essential to achieve a criminal objective, are often not considered within the criminal plot, which affects the determination of responsibilities. Third, it distorts the image of the criminal structure, since investigators and prosecutors are always compelled to identify and create a pyramidal model of the investigated structure, composed by agents with strict and defined roles, which does not occur in the reality of decentralized, resilient and changing criminal networks. As a result, functions, responsibilities or facts are often mistakenly assigned to members of the structure, which later cannot be accredited or sustained during the trial.
Considering the aforementioned, this analysis of the “Lava Jato” structure was based on the concepts and postulates of Networks Theory, which states that a social network is a group of collaborating and interrelated entities. Hence, its analysis is based on a model in which nodes represent individuals or companies with moral agency, and lines or arcs represent interactions or links. To this end, all the nodes/agents that intervene in the interactions are identified, according to the evidence stated in the criminal investigation, mainly, sentences and plea bargains. In this sense, not only strictly criminal interactions and nodes/agents are systematized and analyzed, but all those that can be gathered from the analyzed sources.
Each element of this syntactic structure is entered into a systematized database by means of algorithms that facilitate assigning categories and generating adjacent matrices that represent the database. Then, those matrices are used to calculate centrality indicators such as the direct centrality degree and betweenness, to identify the most relevant nodes/agents. In this case, this procedure allowed understanding how the nodes/agents, considering their capacity of moral agency, interact during a period to achieve illicit objectives through practices of corruption and massive money laundering.
[[Name Actor 1[Description Node/Agent 1]][interaction[verb/action]][[Name Node/Agent2[Description Node/Agent 2]]]
Figure 1. Syntactic structure for analyzing qualitative information
The Lava Jato’s Structure
In the particular case of "Lava Jato", the following dimensions were established for classifying interactions: (i) economic dimension, which defines subcategories related to the physical transport of money and financial transactions; (ii) political dimension, which refers those interactions established between political leaders, candidates, and some public and private officials; (iii) logistics dimension, which includes interactions related to networking and the modus operandi of the criminal scheme.
Likewise, the following categories were applied to classify the agents/nodes: (i) public sector, which includes public employees of any government institution; (ii) private sector, referring to those people who carry out licit commercial activities outside the public sphere; (iii) full-time criminals, which covers those nodes/agents who regularly carry out activities defined as criminal offenses; (iv) financial institutions, which groups agents/nodes that belong to the financial system. These classifications allow understanding the organizational and institutional role of each agent/node, in addition to the particular characteristics of the structure.
Subsequently, the following indicators were calculated: (i) direct degree centrality, to identify the number of direct interactions that each agent/node established and (ii) betweenness, which identifies the nodes/agents with greater capacity to arbitrate or intervene in the geodesic routes that indirectly link nodes. These indicators are represented in the graph below by applying additional analysis algorithms: The degree centrality indicator is represented through the size of each node, and the betweenness indicator is displayed through the position of the node within the plane of the graph, with the highest indicator in the nucleus and the lowest indicator in the peripheral radius.
After designing the model and calculating the indicators, the results obtained were as follows: out of 906 nodes/agents investigated, it was identified that 65% belong to the private sector, 19% to the public sector, 11% are categorized as full-time criminals, and 5% belong to financial institutions. The first category consists of (i) 252 Brazilian companies that were used to establish undue contracts, (ii) 170 businessmen who paid bribes to obtain undue privileges in contracts with the State, (iii) 109 offshore companies used to pay and hide bribes to public servants, (iv) 21 Brazilian consortia created by legitimate and front companies to facilitate bid fraud, and (v) 15 foreign accounts used for money laundering.
In the “Brazilian companies” section, 12 instrumental/screen companies were glimpsed, created to support illegal economic transactions. For instance, although the company MO Consultoría, created and controlled by the forex trader Alberto Youssef, did not offer real products or services, it carried out a large number of contracts with the public company Petrobras and with private companies like Sanko Sider, Sano Servicios, Consórcio Rnest-Conest, Galvâo Engenharia S.A., Consórcio SEHAB Ltda., OAS Ltda., among others, to justify economic transfers that blocked the tracking of money related to bribes.
The category of public sector consists of (i) 100 public servants, (ii) 24 congressmen, (iii) 16 Brazilian politicians, and (iv) 13 political parties. Likewise, within the public servants section, 7 high-level officials of the public company Eletrobras and 8 directors of the public company Petrobras were identified. According to the analysis, agents/nodes grouped within this classification received a percentage of 3% of the value of each contract, paid by the private companies that participated in the scheme.
The classification of full-time criminals grouped (i) 47 figureheads, (ii) 19 forex traders, (iii) 21 financial operators, and (iv) 8 drug-traffickers. Finally, the category of financial institutions consists of (i) 4 banks and (ii) 1 investment fund, which were used to insert assets operated by the criminal structure in the national banking system of Brazil after being transferred across other banking systems abroad.
Figure 2. Graph representing the “lava Jato” network with sources until June, 2017. Size of the nodes represents the direct centrality degree. Location of the nodes represents betweenness indicators. The “Petrobras” node/agent is highlighted and its substructure (nodes/agents directly interacting with the Public company) moved out of the entire “Lava Jato” network.
A Corrupt Self-Sustained System
The scheme operated mainly as follows: (i) important private companies in Brazil, cartelized in the "Club", paid bribes to political parties to influence in the appointment of key public servants in State companies and the executive branch; consequently, (ii) the bribed political parties, through their representatives in the executive and legislative branches, appointed and kept public officials in key positions, so those officials could later favor the cartelized companies during public tenders; (iii) financial operators created domestic and foreign companies to give an aspect of legality to the economic transfers required to pay and legitimize the bribes; (iv) forex trader carried out “cabo-dollar” operations; and (v) public officials and political parties involved in this scheme received from 1% to 3% of the total amount of each awarded contract.
In this sense, the model described herein revealed a system in which political parties and businessmen worked together to appoint public officials at key positions within the public administration, including the management of state-owned companies. The scheme brought high profits to their members, from the private companies that received more contracts, to the political parties that obtained more financial support during their campaigns. With the economic benefits flowing from the public budget to these players, this scheme consolidated itself as an autopoietic system: (i) political parties, (ii) private companies, (iii) public servants, and (iv) currency dealers coordinated their efforts to maintain a co-opting system with high profits and low costs that reproduced itself during at least 15 years.
This analysis challenges traditional concepts of corruption and research methodologies, in the sense that the illicit structure not only consisted of paying sporadic and direct bribes to public officials to obtain contracts. Rather, the system was more complex and sophisticated, co-opting public institutions and private companies to obtain benefits in an enduring perspective. In this way, the economic power of the companies with stronger infrastructure in Brazil, the political power of the government and opposition parties, and the illicit/criminal power of forex trader experts in money laundering operations were coordinated. As a result, the corresponding structure articulated hundreds of individuals and companies, therefore establishing a network of macro-corruption that cannot be analyzed through the common judicial frameworks. Applying traditional concepts and methodologies to understand a complex structure such as “Lava Jato” would drastically reduce the possibilities of effective investigation, prosecution and sanction of transnational corruption and money laundering.
A case as complex as the analyzed herein, for instance, requires reconstructing the wide and systematic scheme of economic and political relations among various public and private agents, in order to understand the illicit financing of political parties and the money laundering process; by omitting these complex structures it would have been impossible to criminally prosecute illicit/criminal interactions between political parties, political leaders, businessmen, financial operators, and high level public officials.
Although some recent literature recognizes and addresses the link between corruption and the formal functioning of democracy –through issues such as electoral rules and processes, or the model and level of political-administrative decentralization-, the influence of corruption on democratic institutions is often omitted, specially when the detailed and extensive structure of political and economic interactions isn’t understood. To understand the true dimension and effects of this phenomena it is critical to acknowledge that networks such as “Lava Jato” configure scenarios of “macro-corruption” that usually evolve towards situations of capture and co-opted reconfiguration of State institutions. Therefore, the methodology and analysis discussed herein, defined as Criminal Networks Analysis, are of utmost significance to combat and conceptualize such criminal macro-phenomena.
 A gray agent can be defined as a person who operates in a social group or legal organization but institutionally holds a value system that favors the execution of criminal activities.
 Usually, illicit association or organized crime penal definitions are subordinated to conducts related to the following criminal types: kidnapping, drug trafficking, human trafficking, arms trafficking, prostitution and, in some cases, corruption.
 Piet Van den Bossche and M. Segers. "Transfer of training: Adding insight through social network analysis." Educational Research Review 8 (2013): 37-47. p. 36.
 Eduardo Salcedo-Albarán and Luis Jorge Garay-Salamanca. Macro-criminalidad: Complejidad y Resiliencia de las Redes Criminales. Bloomington: iUniverse, 2016.
 This modus operandi consisted in formalizing contracts for provision of consultancy, intermediation, and management services between MITSUI TOYO, CAMARGO CORREA, and PIRELLI with JULIO CAMARGO’s companies: TREVISO, AUGURI, and PIEMONTE. Through these simulated contracts, JULIO CAMARGO generated surpluses that stayed available in accounts, property of TREVISO, AUGURI, and PIEMONTE. These assets were transferred by JULIO CAMARGO through AGORA CORRETORA, located in Sao Paulo, to an account outside the country operated by JULIO CAMARGO; the outflow of resources was achieved with a distribution of benefits from TREVISO, AUGURI, and PIEMONTE, tax-free, with the goal of facilitating the delivery abroad. After the money was available in the intended destination -accounts in Switzerland, Montevideo, United Stated of America, and Italy-, JULIO CAMARGO invested in stocks. Through this process, JULIO CAMARGO conducted, along with foreign banks, operations of loans guaranteed by the stock portfolio; the money obtained with such loans was destined to certain accounts indicated by ALBERTO YOUSSEF. Among these accounts, some are property of ALBERTO YOUSSEF, but under the name of CARLOS PEREIRA DA COSTA; LEONARDO MEIRELLES’ accounts were under the following company names: DGX, ELITE DAY, and RFY -all located in Hong Kong-; as well as the offshore accounts property of NELMA PENASSO KODAMA and CARLOS ROCHA’s clients. Finally, the account holders delivered the money in cash to ALBERTO YOUSSEF in Brazil. This mode of operation was developed since the beginning of 2005 and until half way through 2012, and its total value is approximately R$ 27 million.
 Elizabeth Dávid-Barrett and Mark Philip. "Realism About Political Corruption.” Annual Review of Political Science (2015): 387-402.