The cost to Brazil · study
Operation Lava Jato, conventions and private investment
- Authorship
- William Fabiano Dias Filho, Marco Flávio da Cunha Resende, Lucas Carvalho
- Institution
- Cedeplar — Federal University of Minas Gerais
Center for regional development and planning of UFMG (a federal public university). - Year
- 2024
- Period analyzed
- 2002.1 to 2023.1 (effect measured from 2014.2 to 2023.1)
- Document
- Operação Lava Jato, convenções e investimento privado · original ↗
Method
Equação de investimento privado estimada por ARIMAX, com dummy de intercepto e dummies de interação para o período pós-delações; abordagem pós-keynesiana (convenções e incerteza).
Results
| Coefficient of the intercept dummy in equation 2A (p = 0.014) | -1.25 coeficiente | source |
Study results are estimates from the authors' model, not measurements.
Assumptions
- Lava Jato fostered a pessimistic convention that depressed private investment and its sensitivity to its determinants.
Limitations
- The dummy captures everything that changed from 2014.2 onward (economic policy, commodities, political crisis); the authors admit possible overestimation.
- Presented in conference proceedings; not an indexed journal; publication date not given.
- The paper reproduces the DIEESE (2021) results in Table 1, without replicating them.
Divergences and other studies
It is the only academic work, among those found, that tests the operation's effect on aggregate private investment with quarterly data; its result converges with ANPEC 2025, and both depend on the same type of dummy or cutoff date.
Extracted evidence
A Cedeplar-UFMG study (Dias Filho, Resende and Carvalho) estimated, by ARIMAX with quarterly data from 2002.1 to 2023.1, that the Lava Jato intercept dummy (post-plea-agreement period, from 2014.2) has a negative sign and is significant at 5% (coefficient -1.250; standard error 0.499; p = 0.014) in the private investment equation, and that the interaction dummies indicate lower sensitivity of investment to credit.
Source: Operação Lava Jato, convenções e investimento privado — Proceedings of an academic event (even3); authors from Cedeplar-UFMG, 2024 · Section 5, Tables 4 and 5; Final considerations · open original ↗
Study (primary) · full text examined · accessed on Oct. 24, 2026
Limitations: The authors acknowledge that other factors (economic policy errors, the fall in commodities in 2015) contributed to the pessimistic convention and may cause the model to overestimate the influence of Lava Jato: the dummy captures everything that changed in 2014.2. Not an indexed journal (proceedings). The text does not give the publication date.
The Cedeplar study mentions the DIEESE study (2021) and reproduces, in Table 1, DIEESE's results (4.4 million jobs, R$244 bn in value added, R$47 bn in taxes, R$85 bn in wage bill, R$20 bn in social security and FGTS, 3.6% of GDP); it adds its own calculation of the Rasmussen-Hirschman linkage indices with the IBGE matrix of 2010 and 2015.
Source: Operação Lava Jato, convenções e investimento privado — Proceedings of an academic event (even3); authors from Cedeplar-UFMG, 2024 · Section 3, Tables 1 to 3 · open original ↗
Study (primary) · full text examined · accessed on Oct. 24, 2026
Limitations: It is not an independent replication of DIEESE: it reproduces the numbers and uses them as context. It is valuable for showing that DIEESE's result circulates in the literature, not for confirming it.