Was the Brumadinho dam failure caused by a "normalisation of deviance"?
This blog is part of a series 'Towards Mandatory Human Rights Due Diligence'.
Many experts have discussed the causes of the mining dam failure near Brumadinho, Brazil, in January 2019, which killed 272 people. Essentially, it was clear before the breach that the tailings basin had too much water, and it was only a question of time before the last drop, or trigger – maybe a heavy lorry passing nearby, or cows stepping on the dam – would cause the dam to burst (also called liquefaction). But there is more to the story.
Those affected ask for accountability. They say it was not an accident, but a crime.
The CIAEA (Extraordinary Independent Consulting Committee for Investigation), an expert commission created by the board of the mine operator Vale, published a summary of its findings in February 2020. It finds, "The dam’s fragile situation and the need to adopt risk mitigation measures were known." The question is: known by whom?
A history of instability
The report recounts the dam’s history of instability. Back in 2003, an auditor found drainage problems. In 2010, this auditor recommended a liquefaction study, which was only carried out in 2014, based on data from 2005. Since then, the dam had been expanded twice, which the data excludes. A 2016 study again showed unfavourable results, which provoked a dispute among experts about applicable methods: one would have shown an imminent risk of failure, the other concluded stability. The dam was certified as stable. A Brazilian TÜV SÜD subsidiary, hired in 2017, analysed a problem with the safety factor, which indicates the correlation between a dam’s resistance and destabilising forces. It was below the internationally recognized minimum standard. TÜV SÜD examined the dam again in 2018: the problem persisted. Nonetheless, TÜV SÜD certified the dam’s stability twice, in June and September 2018. It finally burst in January 2019.
Vale: Normalisation of deviance
The CIAEA report identified a "normalisation of deviance" in this series of events, which dissociates deviant information from real risks and urgency. In this instance, it meant that "regulatory compliance and [stability declarations] attainment were prioritized, regardless of the actual safety situation of the dam", although "the experience of other hazardous industries (…) teaches that mere regulatory compliance is rarely sufficient to guarantee the safety of highly complex structures." This suggests that what was important to Vale was the safety declaration certificate, and not the dam’s instability.
The report also concludes that while some at Vale knew about the dam’s fragility, the highest-level managers did not. Even if this were confirmed, it does not mean that Vale executives should escape responsibility. Their lack of knowledge was the result of managerial decisions.
CIAEA described Vale as having "a strong hierarchical culture that is resistant to the exposure of problems to higher levels of the organization", as well as a "siloed" environment. This refers to lacking interconnections between business areas, with "a tendency not to expose problems or risks or vulnerabilities to other areas", such as informing compliance and legal departments. In sum: Vale lacked a constructive error culture.
Does TÜV SÜD also suffer from a culture of normalisation of deviance?
TÜV SÜD issued stability declarations for the dam in 2018. In order to do so, they lowered the required minimum safety factor below what was measured, hence masking its deviance. Yet, CIAEA found that "the scientific article cited as a reference is in reality not intended to establish minimum factors of safety."
Why would a professional certifier do that? In Vale’s contracting policy with TÜV SÜD, CIAEA identified one rather notorious risk factor in the auditing business: conflict of interest. This also reflects TÜV SÜD’s approach to conflict of interest risk management: TÜV SÜD accepted a considerably larger consultancy contract from Vale near the time of the audit. This practice, according to CIAEA, "could lead companies to compromise their judgment in audits with the aim of maintaining a good relationship with Vale and entering into consulting agreements."
In practice, compromising their judgement could mean, for example, repeatedly issuing stability declarations for a dam despite knowing it was unstable. Such regulatory compliance at the cost of safety suggests TÜV SÜD also had a culture of normalisation of deviance.
In November 2019, Axel Stepken, the CEO of the German TÜV SÜD AG, explained in an interview that the company’s internal review of events exculpated it from responsibility, but recommended a reputational risk committee. The CIAEA-report differs. It suggests the dam failure was related to structural problems at Vale as well as TÜV SÜD.
"An important driver for non-recurrence is accountability" said Baskut Tuncak, the UN Special Rapporteur on human rights and hazardous substances, after visiting Brumadinho in December 2019. The case of TÜV SÜD shows: internal reviews are not enough. Mandatory human rights due diligence could be a powerful incentive for companies to engage seriously in human rights risk management. Human rights violations like the Brumadinho dam failure are irreparable.