The cost to Brazil · study
Intersectoral economic implications of Operation Lava Jato
- Authorship
- DIEESE (marcas do DIEESE e da CUT; gráfico do slide 3 elaborado pela Subseção DIEESE/FUP)
- Institution
- DIEESE — Inter-Union Department of Statistics and Socioeconomic Studies
DIEESE (Inter-Union Department of Statistics and Socioeconomic Studies) is a research entity created and maintained by the labor movement. It is not a government body. The presentation also bears the mark of the CUT, a labor federation. - Year
- 2021
- Period analyzed
- 2014–2017
- Document
- Implicações econômicas intersetoriais da Operação Lava Jato · original ↗
Method
Choque de demanda numa matriz insumo-produto (MIP 2010 atualizada pelo método de Passoni, 2019). O choque soma (a) a diferença entre o investimento da Petrobras projetado por regressão linear múltipla (receita líquida, Brent, lucro líquido) e o realizado em 2015–2017 e (b) o valor de obras de construção paralisadas segundo levantamento em sites do MPF e do TCU.
Results
| Reduction in investments (Petrobras + construction) | R$172.2 billion | source |
| Value added not generated | R$244.6 billion | source |
| Potential jobs, summed 2014–2017 | 4,439,789 jobs | source |
| Wage bill | R$85.75 billion | source |
| Taxes | R$47.44 billion | source |
| Social security and FGTS | R$20.26 billion | source |
Study results are estimates from the authors' model, not measurements.
Assumptions
- Everything that the Petrobras regression does not explain with revenue, Brent and profit is attributed to the way the operation was conducted.
- The works listed as halted would not have stopped for other reasons (fiscal crisis, lack of credit, canceled plans).
- The coefficients of the 2010 matrix, updated, hold for 2014–2017.
- Consumption lost through income not generated (income effect) is counted in full in the impact.
Limitations
- Material in presentation format: the regression, the list of works and the values per work were not published.
- Nominal values, uncorrected.
- The jobs add up over four years; they are not distinct people laid off.
- More than half of the jobs (2.39 million) come from the income effect.
- It does not separate the effects of corruption itself (overpricing, poorly planned projects) from the effects of how the investigations were conducted.
- The study's own 'without Lava Jato' scenario keeps the recession: GDP of −2.5% in 2015 and −2.1% in 2016.
Divergences and other studies
See the other studies gathered in another entry in the collection.
Extracted evidence
The material is a 14-slide presentation, dated São Paulo, March 16, 2021, bearing the marks of DIEESE and CUT. Slide 3 credits the preparation of the chart to the DIEESE/FUP Subsection.
Source: Implicações econômicas intersetoriais da Operação Lava Jato — DIEESE, Mar. 16, 2021 · slides 1, 3 and footer of all slides · open original ↗
Study (primary) · full text examined · accessed on Oct. 24, 2026
DIEESE estimates a reduction of R$172,214.52 million (nominal) in planned investments from 2014 to 2017: R$104,321.71 million from Petrobras and R$67,892.81 million from civil construction.
Source: Implicações econômicas intersetoriais da Operação Lava Jato — DIEESE, Mar. 16, 2021 · slide 6, table 'Reduction of planned investments - Petrobras and Civil Construction' · open original ↗
Study (primary) · full text examined · accessed on Oct. 24, 2026
Limitations: Nominal values, not corrected for inflation. In the table, Petrobras has no reduction in 2014 (dash); civil construction has R$9,482.41 million in 2014 and the same value (R$21,286.56 million) in 2016 and 2017.
Per year, the estimated reduction is R$9,482.41 million in 2014, R$41,834.77 million in 2015, R$56,942.63 million in 2016 and R$63,954.71 million in 2017. Petrobras: R$25,997.49 million (2015), R$35,656.07 million (2016) and R$42,668.15 million (2017). Civil construction: R$9,482.41 million (2014), R$15,837.28 million (2015), R$21,286.56 million (2016) and R$21,286.56 million (2017).
Source: Implicações econômicas intersetoriais da Operação Lava Jato — DIEESE, Mar. 16, 2021 · slide 6, table · open original ↗
Study (primary) · full text examined · accessed on Oct. 24, 2026
The Petrobras part is the difference between investment actually made and investment projected for 2015–2017 by multiple linear regression, in which investment depends on net revenue, the Brent oil price and the company's net profit (annualized quarterly data, in US$ million). The slide calls the difference investments canceled 'as a result of the way Lava Jato was carried out, without preserving the productive and services structure.'
Excerpt: “em decorrência da forma como a Lava Jato foi realizada, sem preservar a estrutura produtiva e de serviços”
Source: Implicações econômicas intersetoriais da Operação Lava Jato — DIEESE, Mar. 16, 2021 · slide 5, chart and side text · open original ↗
Study (primary) · full text examined · accessed on Oct. 24, 2026
Limitations: The regression uses only three explanatory variables. What it does not explain is attributed entirely to the operation; the slide does not treat indebtedness, exchange rate, change in business plan or dividend policy (which slide 3 itself mentions) as separate causes. The coefficients, the fit and the period of the regression are not presented.
The civil construction part comes from a survey, on the MPF and TCU websites, of works halted as a result of Lava Jato actions and works in which the companies declared financial inability to continue.
Source: Implicações econômicas intersetoriais da Operação Lava Jato — DIEESE, Mar. 16, 2021 · slide 6, side text · open original ↗
Study (primary) · full text examined · accessed on Oct. 24, 2026
Limitations: The list of works and the values for each are not presented in the published material.
Applying the reduction of R$172.2 billion as a demand shock in an input-output matrix, DIEESE estimates, summing 2014 to 2017: 4,439,789 jobs, R$244,647 million in value added, R$47,441 million in taxes, R$85,754 million in wage bill and R$20,259 million in social security and FGTS that were not generated, and a 3.6% potential addition to GDP.
Source: Implicações econômicas intersetoriais da Operação Lava Jato — DIEESE, Mar. 16, 2021 · slide 7, table, TOTAL row · open original ↗
Study (primary) · full text examined · accessed on Oct. 24, 2026
Limitations: These are potential values from a counterfactual scenario. The total of jobs sums four years (361,212 + 1,130,165 + 1,526,917 + 1,421,496): a single position that ceases to exist over several years is counted in each year. It is not a number of people laid off. The 3.6% adds up annual growth differences (0.3 + 1.0 + 1.2 + 1.1 points).
Per year, the estimated potential jobs not generated are 361,212 (2014), 1,130,165 (2015), 1,526,917 (2016) and 1,421,496 (2017); the value added, R$14,832.79 million, R$63,507.34 million, R$83,110.48 million and R$83,196.74 million; the potential addition to GDP, 0.3%, 1.0%, 1.2% and 1.1%.
Source: Implicações econômicas intersetoriais da Operação Lava Jato — DIEESE, Mar. 16, 2021 · slide 7, table · open original ↗
Study (primary) · full text examined · accessed on Oct. 24, 2026
Of the 4,439,789 jobs, 2,050,136 come from direct and indirect effects (affected sectors and their chains) and 2,389,653 from the income effect (consumption that does not take place). In value added, R$124,859 million direct and indirect and R$119,789 million from the income effect.
Source: Implicações econômicas intersetoriais da Operação Lava Jato — DIEESE, Mar. 16, 2021 · slide 8, table · open original ↗
Study (primary) · full text examined · accessed on Oct. 24, 2026
Limitations: More than half of the estimated jobs come from the income effect, the part most sensitive to the model's assumptions.
In the sectoral distribution of potential jobs, construction accounts for 1,075,719 and commerce for 802,176; oil and gas extraction and refining, for 44,153; the 'others' group totals 686,344.
Source: Implicações econômicas intersetoriais da Operação Lava Jato — DIEESE, Mar. 16, 2021 · slide 9, table · open original ↗
Study (primary) · full text examined · accessed on Oct. 24, 2026
Limitations: Values checked against the slide image (automatic text extraction from the PDF misaligns the table). The full table is in the dieese_ocupacoes_setor series.
In DIEESE's scenario chart, GDP 'without Lava Jato' also falls: −2.5% in 2015 and −2.1% in 2016, versus −3.6% and −3.3% observed. In 2014, 0.8% versus 0.5%; in 2017, 2.4% versus 1.3%.
Source: Implicações econômicas intersetoriais da Operação Lava Jato — DIEESE, Mar. 16, 2021 · slide 11, chart 'Evolution of GDP - Scenario with and without Lava Jato' · open original ↗
Study (primary) · full text examined · accessed on Oct. 24, 2026
Limitations: The study itself attributes to the operation a part of the 2015–2016 recession, not all of it. The observed GDP figures are those of the National Accounts available in 2021, subject to revision (slide note).
In DIEESE's jobs chart, the 'without LJ' line moves away from the 'actual' line each year and, in 2017, the distance between the two is visually close to 4.4 million, the same value as the summed total in the slide 7 table.
Source: Implicações econômicas intersetoriais da Operação Lava Jato — DIEESE, Mar. 16, 2021 · slide 12, chart · open original ↗
Study (primary) · full text examined · accessed on Oct. 24, 2026
Limitations: Visual reading of the chart. The table presents effects by year and the chart accumulates them; the material does not explain whether each year's effect persists in the following years. This is a point to clarify with the authors.
Stated methodology: period 2014–2017 (last year of the consolidated National Accounts); 2010 input-output matrix updated by the Passoni (2019) method; demand shock formed by the affected works in civil construction and the amounts not invested by Petrobras; oil and gas extraction, refining and construction sectors.
Source: Implicações econômicas intersetoriais da Operação Lava Jato — DIEESE, Mar. 16, 2021 · slide 13 · open original ↗
Study (primary) · full text examined · accessed on Oct. 24, 2026
The opening slide summarizes: 'Lava Jato cost 4.4 million jobs and 3.6% of GDP,' and states that workers support the fight against corruption, which should 'preserve the productive structure and punish the guilty.'
Excerpt: “A Lava Jato custou 4,4 milhões de empregos e 3,6% do PIB”
Source: Implicações econômicas intersetoriais da Operação Lava Jato — DIEESE, Mar. 16, 2021 · slide 2, 'Key data' · open original ↗
Study (primary) · full text examined · accessed on Oct. 24, 2026
Limitations: The summary sentence speaks of 'jobs'; the study's own table speaks of potential jobs summed over four years.