Thoughts about risk analysis
I have been asked to post excerpts from my new book. It devotes a lot of space to the discussion of risk analysis, including risk appetite, tolerance, and criteria (including why I acknowledge the need to understand risk appetite, while definition of risk criteria is crucial to intelligent decisions).
These are from the chapter on risk analysis:
A single number for level of loss does not enable effective decision-making when one of the possibilities is unacceptable but the calculated overall level appears ok.
A [more complex] example is when there is the potential for (net) gain as well as (net) loss. Consider a situation where management is considering bringing a new product to market. Let’s say that break-even will be achieved if sales reach 10,000 units in the first quarter and the likelihood of different outcomes is estimated as follows.
- 10% likelihood of 5,000 or fewer sales – net loss of $300,000 or more
- 25% likelihood of 5,000 to 10,000 sales – net loss of $100,000
- 20% likelihood of 10,000 sales – break-even
- 20% likelihood of 10,000 to 15,000 sales – net profit of $100,000
- 25% likelihood of more than 15,000 sales – net profit of $200,000 or more
You can use models ….. to help calculate the likelihood of each of these results. Some (especially for financial risk) might use a model to put a single value on the range of potential consequences.
But, does it make sense for management to look at a single number (+$15,000 if you take the sum of the P X I calculations) when deciding whether to go ahead with the launch? I believe a world-class organization would make its decision by considering all the possibilities. Is management willing to take the risk of a $300,000 loss because of the potential for a $200,000 gain? Does it have the liquidity to sustain such a loss? Does the potential for reward justify taking the risk of a loss? That decision can only be made intelligently when all possible outcomes and their likelihood are understood.
By the way, ‘traditional’ risk management only considers the downside. That is not helping management make intelligent decisions, as is readily seen in this example.
Another problem with trying to put a single number on the level of risk is that the calculation of P X I ignores other attributes of the risk, such as the speed of onset, duration, and so on.
World-class organizations understand that if they are to make intelligent decisions, all relevant information about a risk needs to be obtained in the analysis phase and considered in the risk evaluation phase. The level of risk is not a single number; it is the composite of all information necessary to make an intelligent decision about whether to accept the risk and, if not, what action to take.
I always welcome your comments.
 Martin Davies of Causal Capital has an interesting perspective. He says that “Risk practitioners who evaluate risk as a single number will miss the shape of uncertainty”. A December 2014 post, http://causalcapital.blogspot.sg/2014/12/the-shape-of-risk.html, explains.