One of the issues that keeps amazing me is how people – including governments – apply statistics in daily life’s decisions. Probability – also known as chance – seems to be the only relevant factor in discussions. For instance, a probability of 0.05% is an event that may happen once every 2,000 years. A probability of 0.05% is also seen as “statistically insignificant” and a probability of 0.01% even as “statistically highly insignificant”.
These probabilities are for instance used in the required height of Dutch dykes to protect us from future flooding as some 1/3 of our country is below sea level. It is my understanding that (most) Dutch dykes are built while using a probability of 0.05% that the height will not be sufficient to prevent future flooding. Hence a probability that once every 2,000 years we have a major flooding.
It is also my understanding that the New Orleans dykes were built while using a 5% (!!) probability that the height would not be sufficient to prevent future flooding. Hence a probability that once every 20 (!!) years the New Orleans dykes would not be sufficient to prevent flooding. We all know what happened in 2005 after the Katrina hurricane hit New Orleans.
The other relevant factor in statistical discussions is Impact. Or in risk management terms: Risk = Probability x Impact.
A probability of 0.05% is also referred to as tail risk or even catastrophe risk. In layman’s words it is the risk that a catastrophe may occur once in every 2,000 years. However, it is not necessarily AFTER 2,000 years but the event may also occur AFTER – say – a 100 years. You just don’t know. It is like rolling a dice 2,000 times. Once every 2,000 times a certain type of event will occur (e.g., the dice will not show one of its 6 faces but will rest on a corner). That event could happen the first time or the 2,000th time but – more likely – will happen somewhere in between.
Remarkably the Impact factor tends to be rather neglected in environmental risk discussions.
The aftermath of Katrina must have costed a fortune. Unfortunately, by lack of data I am unable to show the penny wise, pound foolish behaviour in building the original New Orleans dykes. I have no clue what risk management probabilities were used in building the Chernobyl nuclear power plant but we all remember the 1986 catastrophe.
While risk management discussions always talk about Probability and Impact, tail or catastrophe risk is strangely enough accepted in cases of a 0.05% probability. Nevertheless, it is certain that a once every 2,000 years a catastrophic event will occur.
The Environment is not only for us. It is also for future generations. We only use it for some 100 years. Next generations would expect us to take their interests into account, especially in discussions that would affect their future interests (e.g., nuclear technology, fracking).
I agree that in most discussions the Probability factor may indeed prevail but only when the Impact is manageable in our lifetime. However, in some discussions Impact should prevail above Probability.
In situations that could and – thus – would involve a catastrophic risk that affects future generations, the Impact should be leading in discussions and decisions rather than its “statistically insignificant” Probability.
0 Comments