7.6 Do Not Consider Consequences of Failures as One-Dimensional
95
Any attempt to rate the consequences as a multiple “OR” clause (e.g., consider the
worst among potential casualties, OR environmental, OR …) is to be rejected as
simplistic and prone to biasing the “safety” perception. Let’s start with an example
drawn from recent history: the cumulative damages effects of the Samarco dam
disaster (already partially discussed earlier in Sect. 2.2).
We have read a great number of articles related to the true cost of the failure of
Samarco dam. The dam failed in November 2015, killing at least 19 people, leaving
700 people homeless, and polluting hundreds of miles of the Rio Doce river, all the
way to the Atlantic Ocean. However, we had to wait for the BHP annual report to
see some details of the cumulative damages effects of the Samarco dam disaster.
In the document BHP Billiton detailed what the true potential costs might be and
the list was as follows:
• 2.3B$ in civil public actions filed by state prosecutors in Minas Gerais
• 3.1B$ in damages filed by public defenders in Minas Gerais
• 620M$ in damages filed by state prosecutors in Espirito Santo
• 6.2B$ in a public civil claim
• 43B$ for reparations, compensation, and “moral damages” in relation to the
Samarco dam failure.
Oftentimes we see Failure Mode and Effects Analyses (FMEAs (http://
www.riskope.com/2015/01/15/failure-modes-and-effects-analysis-fmea-riskmethodology/)) performed where the cost of consequences associated to each
risks scenario are described with a single value deriving from a “OR” clause. As
stated above, it is indeed common practice methodology (certainly not a recommended practice) to split the consequences in categories—for example reputational,
environmental, cost and number of lives lost—then ask the risk assessor to select the
worst and use that “single-value” for the determination of the risk. Over the years
we have shown to many of our clients why this approach is bogus. The Samarco
dam failure gives a perfect example:
• Cost of consequences associated to each risks scenario have a single value;
• Costs are additive once an hazard hits or a scenario occurs. Consequences are an
additive function of at least the following costs:
– direct and indirect,
– health and safety,
– environmental, image and reputation,
– legal, etc.
As we can see in the Samarco list, the legal, the reparations (moral damages), etc.
costs have now indeed been added. In the oil and gas arena, BP learned the lesson
in the hard way with their 62USD billion additive estimate for the Gulf disaster.
This is the reason why, when we perform any risk assessment, even a very simple one (http://www.riskope.com/2016/09/21/risk-assessment-comparison-systemincluding-5-macro-elements/), we use the consequences category listed above in an
additive way!
95
Any attempt to rate the consequences as a multiple “OR” clause (e.g., consider the
worst among potential casualties, OR environmental, OR …) is to be rejected as
simplistic and prone to biasing the “safety” perception. Let’s start with an example
drawn from recent history: the cumulative damages effects of the Samarco dam
disaster (already partially discussed earlier in Sect. 2.2).
We have read a great number of articles related to the true cost of the failure of
Samarco dam. The dam failed in November 2015, killing at least 19 people, leaving
700 people homeless, and polluting hundreds of miles of the Rio Doce river, all the
way to the Atlantic Ocean. However, we had to wait for the BHP annual report to
see some details of the cumulative damages effects of the Samarco dam disaster.
In the document BHP Billiton detailed what the true potential costs might be and
the list was as follows:
• 2.3B$ in civil public actions filed by state prosecutors in Minas Gerais
• 3.1B$ in damages filed by public defenders in Minas Gerais
• 620M$ in damages filed by state prosecutors in Espirito Santo
• 6.2B$ in a public civil claim
• 43B$ for reparations, compensation, and “moral damages” in relation to the
Samarco dam failure.
Oftentimes we see Failure Mode and Effects Analyses (FMEAs (http://
www.riskope.com/2015/01/15/failure-modes-and-effects-analysis-fmea-riskmethodology/)) performed where the cost of consequences associated to each
risks scenario are described with a single value deriving from a “OR” clause. As
stated above, it is indeed common practice methodology (certainly not a recommended practice) to split the consequences in categories—for example reputational,
environmental, cost and number of lives lost—then ask the risk assessor to select the
worst and use that “single-value” for the determination of the risk. Over the years
we have shown to many of our clients why this approach is bogus. The Samarco
dam failure gives a perfect example:
• Cost of consequences associated to each risks scenario have a single value;
• Costs are additive once an hazard hits or a scenario occurs. Consequences are an
additive function of at least the following costs:
– direct and indirect,
– health and safety,
– environmental, image and reputation,
– legal, etc.
As we can see in the Samarco list, the legal, the reparations (moral damages), etc.
costs have now indeed been added. In the oil and gas arena, BP learned the lesson
in the hard way with their 62USD billion additive estimate for the Gulf disaster.
This is the reason why, when we perform any risk assessment, even a very simple one (http://www.riskope.com/2016/09/21/risk-assessment-comparison-systemincluding-5-macro-elements/), we use the consequences category listed above in an
additive way!