With numerous, often very complex models, we tend to recognize, evaluate and avoid risks in life. Risk management, the systematic recognition, analysis, evaluation, monitoring and control of risks, is one of the most important management and controlling tools.
But are the results of these models always helpful? Do they protect us from undesirable events? Why did all risk models in the financial sector fail to recognize the risks in 2008 or only inadequately? In order to recognize the sense and nonsense of risk management models and tools, we must first look at what is behind the term "risk". Here is a brief excursion into decision theory.
Our life is a continuous sequence of decisions. 99% of these are made unconsciously, like routine decisions and reflexes. Only 1% of our daily decisions require conscious consideration. These can range from low cognitive effort, as in stereotypical decision-making processes (What will we have for breakfast today?), to complex (constructive) processes (production of item A or item B).
Every decision-making situation requires us to select ONE option from a set of defined and/or undefined choices (options) with known or unknown long-term and side effects. Depending on how much we know about the set of options and their long-term and side effects, a distinction is made between decisions under certainty and decisions under uncertainty.
Decisions under certainty
Even decisions under certainty are not always easy for us. We know all the available options, as well as their long-term and side effects. We have clear expectations that remain unchanged even if something does not happen once. For example, turning on the computer or turning the ignition key in the car.
In situations of security, the results of risk management tools make sense. Based on the precise data, the models can be fed correctly and the results of the calculations are meaningful for the current situation. But how often do we find these optimal conditions? And do we need the support of complex models in situations of security? Aren't it rather the situations of uncertainty in which we hope that models will provide us with guidelines and information for our decision-making?
Decisions under uncertainty
In these decision-making situations, we are only partially or not at all aware of the multitude of options and their effects. These are often unknown, new situations for us, so we have little experiential knowledge.
Since the feeling of security is linked to the perception of the ability to act and make decisions, we look for decisions under certainty.
Even when there is actually no security. We are happy to trust the statements of supposed experts and follow their judgments. We experience the feeling of security and are all the more surprised when the expected result does not occur. In science, this is referred to as decisions under false security. For example, we are inclined to trust the judgments of doctors and medical examination results almost blindly. Numerous private and institutional investors have also relied on the assessments of rating agencies. Both examples of decisions under false security.
Unfortunately, risk management tools do not help us at all in these situations. Fed with false data and facts, the statements of the models are worthless for reality. The only thing that helps here is to check the source of your information and its motivation.
On the next level towards uncertainty, we find decisions under risk. Possible long-term and side effects of the individual options are known. We also know the probabilities of the various outcomes occurring. Gambling is predominantly one of these decision-making situations. The results and probabilities can be determined on the basis of decision trees and the best choice can be made mathematically.
In decision-making situations involving risk, the use of risk models can be helpful. If it is possible to determine the probability of occurrence as accurately as possible, the models represent an approximation of the decision situation. Approximation, because model errors and simplifications of reality can lead to distortions and misjudgments. A "detail" that we tend to overlook. The more factors differ between the model and reality, the less meaningful it becomes. We also easily forget that the model results are always snapshots. If the framework conditions change, the calculations are worthless. In a fast-moving world like today, there is not enough time for data collection, input and calculation. So people often rely on outdated risk indicators and model scenarios and feel falsely safe. But now to the most important group of decision-making situations.
We usually have no knowledge of the number of options, their long-term and side effects and their probability of occurrence. Most economic decisions and those that affect our lives are decisions made under uncertainty.
We only know some of the possible options and can therefore only partially determine their long-term and side effects. Precise data on the probability of occurrence is insufficient or non-existent. How useful are risk management models in situations of uncertainty? Pointless.
As soon as we lack concrete observable and measurable data, the results of the risk management tools represent some of many future variants. Assumptions and estimates for unknown values are usually fed into the models, so that their results have little or nothing to do with future developments.
What these models do instead is give us a feeling of security. We think we are acting in a predictable situation of risk and overlook the uncertainty. We experience the feeling of controllability and are therefore inclined to choose riskier options. Our willingness to take risks increases.
"If people feel they can predict something, they stop preparing for worse cases. But it should be the other way around: the more you can predict, the better you should prepare for serious cases." (Structural engineer Nate Silver in the Standard Interview on September 2, 2013)
The great attraction of models and tools for risk measurement and risk management does not lie in their accurate statements, but rather in the fact that they give us a feeling of security in which decisions made under uncertainty become decisions made under risk.
The term "uncertainty" itself seems more threatening than the term "risk". When we talk about uncertainty, alarm bells ring and our willingness to make decisions is reduced. In research, this is referred to as the Ellsberg paradox. It states that when people have to choose between the options "risk" or "uncertainty", they are more likely to choose the risk. Because we can develop ideas about the future based on known probabilities. This gives us subjective certainty - certainty about the possible outcomes.
The Ellsberg paradox in a small example. Read the text and decide between variant A and variant B.
Variant A: If you draw a red ball, you win. If you draw black or yellow, you lose.
Variant B: If you draw a yellow ball, you win. If you draw black or red, you lose.
If you have chosen variant A, you are in good company. In numerous studies, the majority of participants chose this variant. The risk is known here: the chance of winning with a single attempt is 1/3. The uncertain situation (variant B), on the other hand, is avoided.
What to do in everyday life?
Trust your critical mind more than mathematical figures. Risk management tools used incorrectly mainly lead to our willingness to take risks increasing, but this does not necessarily lead to a reduction in risk. The following thoughts can be helpful in practice:
Check whether experts and models can actually provide competent information for the respective question. What (own) motives could experts and models be pursuing?
Consider the limitations, sources of error and estimation errors of models and tools. What effect do these have on the meaningfulness of the model results?
Think about the unknown: what options do you have if things turn out differently?
Practice distinguishing between situations involving risk and situations involving uncertainty. In situations of uncertainty, you should avoid using risk models because the models give you a false sense of security. Your willingness to take risks increases, and your flexibility in reacting to the unexpected decreases.
Finally, let's think outside the box: Uncertainty makes change and the impossible possible. Without uncertainty, we would perceive our lives as very boring.
References:
Becker, Dirk (2008). Womit handeln Banken? Eine Untersuchung zur Risikoverarbeitung in der Wirtschaft. Berlin: Edition Suhrkamp.
Böhmer, Gerd (2010). Neuroökonomie: Neuronale Mechanismen ökonomischer Entscheidungen. Mainz: Johannes Gutenberg-Universität.
Braun, Walter (2010): Die (Psycho-)Logik des Entscheidens. Fallstricke, Strategien und Techniken im Umgang mit schwierigen Situationen. Bern, Verlag Hans Huber.
Gigerenzer, Gerd (2008): Bauchentscheidungen. Die Intelligenz des Unbewussten und die Macht der Intuition. Goldmann Verlag.
Glimcher, Paul; Camerer, Colin; Fehr, Ernst; Poldrack, Russell (Hrsg.) (2009): Neuroeconomics: Decision Making and the Brain. Amsterdam: Elsevier/Academic Press
Holtfort, Thomas (2011): Intuition, Risikowahrnehmung und Investmententscheidungen. Essen: Akademie Verlag.
Luhmann, Niklas (1991). Soziologie des Risikos. München: deGruyer Verlag.
Talab, Nassim (2008): The Black Swan: The Impact of the Highly Improbable. Penguin Verlag.
Comments