Anchor: Petrobras — Lava Jato strategic-pricing reasoning and procurement cartel

1. Episode summary

Between 2006 and 2014 Petróleo Brasileiro S.A. (Petrobras) — the Brazilian state-controlled oil company — operated inside a genuinely hard strategic problem: domestic refined-product prices were held below international parity by government directive, producing structural downstream losses; at the same time the firm was funding a capital-intensive pre-salt deepwater upstream development programme requiring unprecedented procurement scale. The firm's internal reasoning about how to bridge the financing gap settled on a structure that tolerated — and operationally depended on — a procurement cartel among major Brazilian engineering contractors, with contract overpricing recycled as kickbacks to political parties, elected officials, and Petrobras executives in exchange for contract allocation. The strategic question the episode turned on was whether the firm would construct a financing bridge that preserved institutional integrity. It did not. The scheme was surfaced by the Lava Jato ("Car Wash") Federal Police investigation from March 2014, with subsequent court findings, Brazilian Congressional Parliamentary Commission of Inquiry (CPI) reporting in 2015, and a US Department of Justice Foreign Corrupt Practices Act settlement in 2018. The reasoning was hard in genuine domain complexity and wrong in the specific institutional solution the firm constructed; the eight-year execution of that solution was itself active criminal conduct, not merely the downstream mechanical effect of the reasoning choice. The episode is a fraud case under R3.

2. Sources

Primary:

1. Brazilian Federal Public Ministry (Ministério Público Federal) and Federal Police, Operação Lava Jato investigation filings and court-ordered testimony, 2014–2018 — available through Brazilian federal court records (13ª Vara Federal de Curitiba).
2. Câmara dos Deputados (Brazilian Chamber of Deputies), Comissão Parlamentar de Inquérito — Petrobras, final report, 2015.
3. Senado Federal (Brazilian Federal Senate), CPMI Petrobras joint parliamentary inquiry record, 2014–2015.
4. United States Department of Justice and US Securities and Exchange Commission, FCPA settlement agreements with Petrobras, 27 September 2018, including Deferred Prosecution Agreement and statement of facts.
5. Petróleo Brasileiro S.A., 20-F filings with the US SEC, 2006–2018, including restatements following the Lava Jato disclosures.
6. [CITATION NEEDED: specific Operação Lava Jato plea-bargain (delação premiada) testimony references for procurement-cartel-operation detail — exact case IDs.]

Secondary (with justification):

1. Transparency International and Organisation for Economic Co-operation and Development anti-bribery reviews covering the Petrobras episode — used to place the case in a non-US anti-corruption enforcement context.
2. Brazilian and international financial-press coverage (Folha de S.Paulo, O Globo, Valor Econômico, Financial Times, Reuters), 2014–2018 — used for chronology and the sequencing of plea-bargain testimony.
3. Tribunal de Contas da União (TCU — Federal Court of Accounts), econometric audit findings on cartel-caused damages to Petrobras contracts, 2014–2015 (Portal TCU, "TCU aponta que cartel causou prejuízos de R$ 18 bilhões à Petrobras") — used for quantification of contract overpricing and for structural evidence on the failure of external governmental audit to detect the scheme during 2004–2012.

Tertiary (flagged):

1. Business-school teaching cases on Petrobras / Lava Jato — used for frame only. Flagged tertiary.

3. OTA narrative

Observe. Observation of the firm's strategic environment was substantially accurate. The domestic price-control gap, the pre-salt capital demand, the political environment of state-controlled enterprise governance in Brazil, and the capability of the domestic engineering-contractor base were all correctly perceived internally and externally. The observation phase is not where the failure originated; a well-run peer state-affiliated NOC would have perceived the same landscape. Observation was Correct; under P2 it carries zero weight in the failure frame.

Think. The reasoning is one of the two root-cause phases. Faced with a genuinely hard reasoning problem — how to fund a very large capital programme inside a price-controlled revenue structure — the firm's internal reasoning converged on a structural response that tolerated corruption as a financing mechanism. The reasoning embedded the cartel-based procurement pattern inside operational decision-making: contract allocation routes, bid-evaluation rules, and executive decision rights were designed or tolerated in forms that assumed the cartel. This is worse than the reasoning adopted at peer NOCs in structurally comparable positions. Pemex and PetroChina, facing analogous price-control and capital-demand pressures, developed structurally different adaptations — not without their own governance problems, but not with the specific feature of institutionalising a contractor-cartel-plus-political-kickback loop inside core procurement reasoning.

Act. Acting is the second root-cause phase. Over eight years the firm executed the rigged-procurement-plus-kickback mechanism at scale: contracts were awarded to cartel members at overpriced terms, payments were made, construction was programmed around the cartel allocation, and kickback flows to political parties and executives operated on a persistent basis. The actions themselves — awarding rigged contracts, making and receiving kickback payments, participating in concealment — are the criminal conduct the Lava Jato filings, the CPI report, and the DOJ FCPA statement of facts document. This is not "technically competent execution of a wrong strategy" (which would zero under P2); it is active wrong performance of the Act phase over the eight-year episode. Think is upstream of Act, and the reasoning step converted the hard domain problem into the corrupt institutional solution — so Think carries the primary weight — but Act carries independent causal weight because the persistent criminal execution is a distinct operational failure from the reasoning choice that framed it.

---

Note to the Phase 2.3 rater: Sections 4 through 10 of the full anchor file are deliberately withheld from this workspace. You are being asked to score this case on the basis of Sections 1, 2, and 3 only, plus the methodology document and the Peer Reference Sheet. Do not attempt to locate or read the canonical anchor file, any other rater's file, the Phase 2.2 workspace, or any T-022 analysis or decision document. Section 3 (OTA narrative) is scoring-relevant scaffolding in the Phase 2.3 blind contract per the revised §9 of the methodology.

4. Modality evidence

Direction. The Brazilian federal government, as controlling shareholder holding roughly 54 per cent of Petrobras's common voting shares throughout the episode, exercised direct and documented influence over the strategic direction of the firm. Direction here is not merely a corporate-level choice by Petrobras management alone; it is the product of a dual principal structure in which the government as owner instructed the firm on domestic fuel pricing (holding refined-product prices below international parity) while simultaneously driving the aggressive pre-salt upstream expansion, captured in the Petrobras Strategic Plan 2020 and Business Plan 2008–2012 approved in the third quarter of 2007, which projected $30.9 billion in pre-salt capital expenditure through 2014 (Petrobras 20-F filings, 2006–2018; DOJ/SEC FCPA statement of facts, 2018). Those two directives — suppressed downstream revenue and accelerated upstream capital deployment — created the financing squeeze that the firm's internal reasoning then resolved through the cartel mechanism. The directional choice to maintain the price-control posture and the pre-salt investment pace simultaneously was attributable and datable, made by the federal government as controlling owner with Dilma Rousseff chairing the Petrobras board from 2003 to 2010 and later as President from 2011 onward, and it was the upstream precondition that made the corrupt financing structure appear operationally rational to those inside the firm (DOJ/SEC FCPA statement of facts, 2018; CPI final report, 2015).

Scoring note (modality_zero_rationale, added 2026-07-23): Direction is admitted under Step 1 via the datable 2007–2008 government price-control/pre-salt decisions, but is zero-weighted under Step 2 comparative judgment because this §4 subsection frames Direction as explaining why the problem was hard, not why the corrupt solution was chosen — causal weight sits with Structure/Processes/Culture. This departs from Direction's own ≤15% Boundary-clause default and is flagged here for author confirmation.

Direction evidence satisfies the three-prong test for admissibility: the specific choices (price-control maintenance; pre-salt expansion mandate) are discrete decisions datable to 2007–2008 at the latest, and are attributed to the federal government and the Petrobras board in the CPI and FCPA records. The evidence also supports that Direction here is a contributing modality but not the primary one — the financing-gap framing explains why the problem was hard, not why the corrupt solution was chosen over available institutional alternatives. The causal weight in the episode sits with the reasoning and execution failures, not with the strategic direction itself.

Structure. The structural enablers of the scheme operated at two levels: the internal governance architecture of Petrobras, and the external state-enterprise oversight framework. Internally, the Diretoria de Abastecimento (Supply Directorate) and Diretoria de Serviços (Services Directorate) concentrated contract-allocation authority in a small number of executive director positions — Paulo Roberto Costa (Director of Supply) and Renato Duque (Director of Services) — who, together with Executive Engineering Manager Pedro Barusco, controlled the contract-award pipeline that the cartel exploited (CPI final report, 2015; CPMI Petrobras record, 2014–2015; Lava Jato court filings). Contract-award decisions were not subject to independent technical review or separated from the directorates that also maintained relationships with contractor firms; the same reporting structure that owned procurement outcomes owned the relationship with the political principals whose parties received kickback flows. This is a structural configuration in which authority was placed where it could be captured rather than where it could be checked.

Externally, the board of Petrobras throughout the period included government-appointed directors who occupied board seats as representatives of the controlling shareholder, not as independent monitors. The CVM (Brazilian Securities Commission) ruling in November 2020 acquitted external auditor PwC of wrongdoing and instead sanctioned KPMG (and a KPMG partner) for a different failure — not running a required impairment test on the Abreu e Lima Refinery in the 2010 financial statements (InfoMoney, "CVM aplica multas a KPMG e sócio por irregularidade em auditoria da Petrobras," Nov 2020) [corrected 2026-07-25: the case previously said this CVM ruling found PwC failed to detect 2012–2014 overpricing; the real ruling acquitted PwC and sanctioned KPMG instead, over an unrelated 2010 matter] — while the TCU's own post-facto econometric audit identified R$18 billion in cartel-caused damages to contracts concluded between 2004 and 2012 — damage that accrued across a decade without detection by either external audit or the federal oversight bodies (TCU findings, 2014–2015; Petrobras 20-F 2014, citing $2.527 billion write-off of capitalised overpayments). The structural configuration in which a state-controlled enterprise's board was populated by political appointees rather than independent monitors, and in which procurement authority was concentrated in directorates with direct political-party linkages, was the architectural arrangement within which the cartel mechanism became operational and sustained.

Processes. The operational mechanism that sustained the cartel for nearly a decade rested on specific process-level failures in Petrobras's procurement, internal audit, and financial-reporting routines. The "clube das empreiteiras" — the contractor cartel comprising at least sixteen major engineering firms including Odebrecht, OAS, UTC, Camargo Corrêa, Queiroz Galvão, and others — operated by pre-allocating contract awards in meetings among cartel members before bids were formally submitted, then submitting artificial bids by losing participants to create the appearance of competitive tendering (CPI final report, 2015; Lava Jato court filings; CPMI Petrobras record, 2014–2015). Petrobras's bid-evaluation process did not have independent mechanisms to detect bid-rigging: evaluation was performed by the same directorates that received kickback flows, and the 1–3 per cent kickback percentage on contract value was embedded in the overpriced contract amounts, passing through the capitalised-cost accounting as normal project expenditure (DOJ/SEC FCPA statement of facts, 2018). The internal audit function and the Comitê de Auditoria did not surface the overpricing pattern during the 2004–2014 period; after the scheme was exposed, independent auditors subsequently identified material weaknesses in the company's internal controls (Petrobras 20-F 2014; TCU findings, 2014–2015).

The absence of an anti-corruption compliance programme or internal-investigation function capable of receiving and acting on signals from within the directorate hierarchy was a process gap rather than merely a cultural one: no formal whistle-blower channel with independent reporting to the audit committee existed in operating form during the episode period, and the routine bid-audit procedures did not include statistical analysis of discount rates across contractor groups that could have flagged the 17-percentage-point discount gap between cartel and non-cartel bids (TCU findings, 2014–2015). Pedro Barusco's personal meticulous accounting of bribe transfers — maintained in a private file across years of cartel activity — and his eventual plea-bargain testimony to the CPI demonstrated that the operational mechanics of the scheme were transparent to participants but invisible to the firm's formal process infrastructure (CPI final report, 2015; CPMI Petrobras record, 2014–2015).

Capability. Petrobras during this period was among the most technically capable deepwater upstream operators in the world; its engineering and project-management capabilities in pre-salt deepwater development were genuine, as demonstrated by the physical execution of major refinery and offshore platform construction programmes of unprecedented scale and depth (Petrobras 20-F filings, 2006–2018). The capability gap the episode surfaces is narrowly institutional rather than broadly technical: the firm lacked the anti-corruption compliance, forensic financial-audit, and independent procurement-review capabilities that would have been required to detect and resist a sustained, multi-directorate, politically-connected bid-rigging scheme. This is a Capability finding in the sense that the required institutional competence — anti-corruption programme design, statistical bid-audit analytics, independent internal-investigation capacity — did not exist in operational form, not merely that it existed somewhere in the organisation and was not deployed. Post-scandal compliance reforms introduced from 2015 onward, including the establishment of an independent whistleblower channel and enhanced internal-controls architecture, confirm by retrospective contrast that these capabilities were absent during the operative period (Petrobras 20-F 2014; DOJ/SEC FCPA DPA compliance conditions, 2018).

Scoring note (modality_zero_rationale, added 2026-07-23): this subsection argues FOR a real Capability gap, so "no evidence" is not a valid zero-basis; the defensible rationale is Step-2 comparative subordination — the absent capability is read as a downstream symptom of Culture's normalised tolerance, analogous to Structure being read as downstream of Direction elsewhere in the methodology. This is a rater-supplied inference needing author confirmation; an alternative reading would redistribute a small non-zero weight to Capability instead.

There is a boundary question here between Capability and Structure: the absence of these functions could be read either as "the wrong people in the right boxes" (Structure) or "the required capability did not exist at all" (Capability). The evidence supports the latter. The internal audit function existed formally but lacked the independence and forensic scope required to do the job the situation demanded — not a misplacement of existing capacity but an absence of the capacity required. Low-confidence note: the evidentiary record on the specific institutional capability state of Petrobras's internal audit and compliance function during 2006–2014 (as distinct from the structural board-governance configuration) rests primarily on secondary and post-hoc sources; primary source granularity on the pre-2014 compliance programme design is limited.

Culture. The cultural evidence in the Petrobras episode has two distinct components that the Fraud Case Structure-Culture Rule requires to be scored separately from the structural contributors already attributed above. The first is the normalisation of the kickback-and-cartel mechanism as a standard operational routine among senior Petrobras executives: Pedro Barusco's testimony to the CPI described receiving bribes from contractor companies since 1997 and characterised the arrangement as systematic and methodical — he maintained organised personal records across years of payments, indicative of a behavioural norm rather than an isolated aberration (CPI final report, 2015; CPMI Petrobras record, 2014–2015). The scheme operated across at least two successive directors in two separate directorates and persisted through management transitions, which is consistent with a shared normative tolerance for the arrangement rather than a single deviant individual's choice.

The second cultural component is the relationship between the political-party funding dimension of the kickback flows and the behavioural norms of the broader Brazilian public-enterprise political economy during the PT-government period. The DOJ/SEC FCPA statement of facts documents that between 2003 and 2012 senior Petrobras executives worked with contractors to inflate costs in exchange for kickbacks to Brazilian government officials and political parties, with the scheme having cross-party reach — members of PP, PT, PMDB, and other parties receiving flows (DOJ/SEC FCPA statement of facts, 2018; Lava Jato court filings). The cultural norm the evidence supports is one in which political-party financing through state-enterprise procurement was treated within the enterprise as a structural feature of operating in the Brazilian political economy, not as a red line requiring active resistance. That norm was upstream of the structural cartel arrangements: the willingness to participate and the absence of internal disclosure was a behavioural default, not merely a consequence of absent channels. Applying the Structure/Culture test: the formal channels for reporting concerns to the board and to external auditors existed in some form; the evidence is that cultural pressure and norm-alignment prevented their use — a Culture primary finding for the normative and motivational layer of the scheme, consistent with the Fraud Case Structure-Culture Rule's requirement to score both, with Structure capturing the architectural enablers and Culture capturing the behavioural perpetuation.

---