The occurrences and recognition of risks

The occurrences and recognition of risks^[1]

To explain problems associated with different types of risks, we use a simple box diagram to categorize the risks that can be faced by any decision-maker. This risk classification problem is the same for either short and long-term traders or discretionary and systematic managers. Risks in a simple two-dimension framework can be broken into four major types. Risks can be classified under one dimension as recognized or out of mind. Risks can also be classified by whether there have been occurrences or not. This 2x2 combination can describe risks by their recognition or occurrence.

Most decision makers encounter experienced risks. These are risks which have occurrences and can be recognized. These are the risks that have been occurred in the past and can be measured or counted. While all situations may not be the same, we have a basis for correlation, possible cause and effect, and control. The systematic manager may take experienced information through market and economic data and look for measurable relationships. The discretionary trader may use these past experiences to learn what is the most prudent strategy given a set of events. There is no guarantee of success because these experienced relationships have statistical error, but a clear link can be formed between information and market action. Our VaR calculations are based on experienced risks. This is one of the key reasons why we choose experienced traders for the portfolio. When someone has more exposure to a variety of market events, it will be easier to measure risks.

Contemplated risks are those that can be recognized but may not have occurred. In this case, traders can extrapolate from similar occurrences and make a judgment on what may happen even if there is not a specific example. The skill of being able to contemplate risk is important with any trader. They can take past situations and be able to use them to make judgments about the current environment. Scenarios can be built around risks that have not yet occurred.

	Out of mind	Recognized
No occurrences	Virgin Risks	Contemplated Risks
Past occurrences	Neglected Risks	Experienced Risks

The more difficult risks are those that can be classified as out of mind. Out of mind or risk not recognized can be classified as either being virgin (no occurrences) or neglected (past occurrences not used) risks. A virgin risk is one that has not been experienced or considered. A neglected risk is one that has occurred but is not currently contemplated.

A neglected risk is a problem with not looking back in history for what may have happened or not using information that is available. A simple example from a quant perspective will occur when you do not use enough data to make a decision. This is the classic problem associated with a small sample size or look-back periods. Events such as a stock market crash that has occurred but may be overlooked or neglected by some traders because it occurred years ago would be a neglected risk.

A virgin risk by its very nature is truly uncertain because we have no reference for what or when it may occur. At worst, “things happen”. The impact of a sovereign debt default in Europe could be a virgin risk. There have been no occurrences and even if we recognize this risk we cannot fully comprehend its impact or implications. A more localized example may have been the IEA strategic release of petroleum. Traders did not anticipate this action based on anything other than truly extreme market price activity.

Risks that are hard to recognize or do not have information available for forming probability judgments create ambiguity. Ambiguity causes investors to avoid taking risk. This avoidance of ambiguity creates a breakdown in past market relationships and range-bound behavior.

---

[1] The section is based on work by Carolyn Kousky, John Pratt, and Richard Zeckhauser, “Virgin versus Experienced Risks” in The Irrational Economist: Making Decisions in a Dangerous World by Erwann Michel-Kerjan and Paul Slovi, editors