Black Swans: The Limits of Statistical Modeling in Risk Management
In risk analysis, what has never
happened was almost impossible to foresee and almost nobody counted on it, they
are called black swans . Clear examples known to all of
us: the attack the twin towers from New York in 2001 or the
global pandemic by COVID-19 that we are all suffering in 2020. From an approach
from RAMS Engineering , we could speak of outliers or events
without examples or without similar previous cases. The black
swan , moreover, is used to being that accident that changes the
paradigm of everything that was known so far.
The term black swan became popular during the sixteenth century where all the swans known to Europeans were white. The explorers who arrived in Australia discovered a species of black swan . The term is now used to refer to events that occur even though their occurrence was thought to be impossible.
What are the typical characteristics of a black swan?
Black swans are completely atypical values or events, which are outside the scope of what is expected, of the usual expectations since, by definition, nothing in the past can make us think about their possibility of occurrence . In addition, it has an extreme impact which usually changes the status quo.
Despite what we could call the rarity of the occurrence of black swans , the conclusion is also reached, on many occasions and once it occurs looking back in time, that despite being an atypical situation, it could have been predicted if not warnings would have been ignored, mostly due to strong resistance s Beliefs, under a wide view and, in addition, its explanation, normally, is not very complex to develop once it has occurred.
It is also important to note that not all large and rare events are black swans . We give a clear example: the Fukushima tsunami that caused an environmental disaster due to its impact on the Fukushima nuclear plant. A priori we might think that it was a black swan but the reality is that the Fukushima area of the coast of Japan had already experienced tsunamis of the same magnitude several times years ago and with even larger wave sizes. In this sense, the Fukushima plant, it seems that it was clearly designed without taking into account these well-known events of the past.
The risk analysis of RAMS Engineering in the face of a black swan
Unfortunately, experience tells us that if there are black swans it is because a risk analysis process is not capable of detecting and mitigating them. Therefore, typical risk management tools such as Preliminary Hazard Analysis , Hazard Log , FMECA, will normally be ineffective for the detection of black swans . Although these methods seek to identify, quantify and mitigate hazards with rigorous and exhaustive methods, the set of possible future events ,They will be delimited and unknown unknowns will not be applied because we do not precisely know them at the time of risk analysis .
From a mathematical point of view, a black swan could be considered as anything that occurs very far from the center of a normal Gaussian distribution. As is known, the probability of occurrence beyond 6σ is a negligible 9.9x10-10; a 10σ event is less frequent than one day in the life of the universe ... which means it won't happen tomorrow.
How do we deal with black swans as risk managers?
As we have already discussed, by definition, black swans are not predictable. In any case, it is advisable to try to determine the level of exposure we have to them. For example, in a classic risk analysis process, we recommend asking ourselves the following questions:
- Is my system simple, complicated, complex, or chaotic? As the reader can imagine, if the answer is "chaotic", our probability of suffering a black swan will be greater than if the answer is "simple".
- What is the level of confidence we have in our system and its environment?
- Does the system, installation or product under analysis work in asconditions of high volatility, uncertainty, complexity or ambiguity (in conditions of the well-known acronym VUCA )?
We also recommend techniques that try to reduce the probability of not detecting hazards that turn into black swans , such as challenging and training people in the organization to increase their vision and imagination regarding possible safety problems and, above all, promote communication upstream of problems that are detected in the organization in terms of safety and, from the point of view of the installation or system, maximize stress tests and identify the horizons or limits of the installation itself before its collapse.
When we talk about the concept of challenging and empowering people to explore and try to read beyond the horizontals and status quo of our organization, we think of simple ideas, closely related to the concept of think outside the box , such as :
- Encourage the development of more enriched accident-related scenarios than those normally considered in the organization's risk assessments .
- Reduce complacency on security .
- Trigger internal discussions on new approaches to managing risks and hazards .
- Implementation of a " red team ". A concept invented in the military industry where an independent group brings a contradictory, challenging perspective to the organization to improve its effectiveness, imagining attacks from a competitor, future worst case and "nightmare" scenarios.
When we speak of stress tests, we refer to analyzes or simulations designed to determine the ability to cope with an exceptional event, which can be used as a robustness test in the face of catastrophes and as a complement to standard risk projections and analysis. In the financial sector, stress tests after the 2008 financial crisis have become mandatory and questions such as What happens if GDP falls by X% next year? What happens if the unemployment rate increases by X% next year ?, should be simulated, analyzed and answered to financial regulators.
Scenario analysis: clear strategy for predicting black swans
The scenario analysis follows a systematic process to create a set of plausible and competing narratives that describe possible developments in key areas of uncertainty. These narratives ( scenarios ) examine the social, political, economic, and technological forces that can affect a project or strategy. They are developed by a diverse team that includes decision makers, experts, and stakeholder representatives. It can be used to analyze high-impact, low-probability events that occur so quickly that social systems cannot respond effectively - the typical example: the possibility that tomorrow the earth will be hit by a meteorite.
One of the characteristics of the scenario creation strategy is that, as it is a totally hypothetical, unknown and multiple analysis of the future , it does not only propose a possible future but rather opens the range to different probable scenarios, not ruling out any of them. Opening possibilities for the future also opens the door to cover events that s Weird enough to be considered possible future black swans . Once the different scenarios are raised, it is the responsibility of the organization to act accordingly, making decisions, normally strategic, in the face of the proposals presented with the aim of "being prepared" for the arrival of any of these scenarios.
Increase resilience against what we do not know
Given some possibilities of trying to predict black swans , we must assume that it will be almost impossible to predict the next black swan that will appear in our lives. However, we can try to act to limit its impact, reducing the vulnerabilities of our systems. The concept of resilience , so fashionable today, applies perfectly in this case. The resilience we can consider as the ability of a system to adjust its operation before, during or after events (changes, disturbances and opportunities) and therefore maintain operations required in expected and unexpected conditions. We can consider the following as the 4 pillars to maximize the resilience of a system or installation:
- Learn . Know what happened and learn from the past
- Reply . Knowing what to do when something happens, either by following previous practices or being able to adapt quickly to the new situation.
- Monitor . Know what to look for. Being able to see the first signs that something is changing and take steps to avoid it in time.
- Anticipate . Know what to expect. Ability to anticipate what may happen in the future and what effect it will have on the organization.
At Leedeo Engineering , we are specialists giving support to our clients at any level required for RAM and Safety tasks, and both at the level of infrastructure or on-board equipment. Do not hesitate to contact us >>