Managing unknown risks

David Griffiths has a post struggling with how to correctly plan around low-probability, high-impact events (a.k.a. "black swans").

He writes:

... we can talk about historic data from past events, but implications and impact will not be transferable ... We can talk about frequency, but we cannot predict the next event or a likely time-frame for that event. We can raise awareness, but we offer nothing ‘tangible’.

David then refers to Dave Snowden's Probable – Possible – Plausible continuum as a better planning tool. I agree with him, but perhaps not for the reasons stated.

Statistical models of the world, whether Gaussian or power-law based, are essential for estimating costs and benefits of what we might call "estimable risks". When we talk about 1 in 100 year floods, we don't know when the flood is coming, just roughly how often they will occur. This allows an explicit cost/benefit tradeoff where we spend $X/year to reduce the cost of that 100 year flood to acceptable levels.

But this works for towns where they are expected to exist long enough to justify that timescale of planning. Try and get a CEO to plan for a 1-in-100 year event when they are highly unlikely to feel loyalty to their company beyond their working lives, or more likely, their tenure at the job!

To a company leader, therefore, black swan events present the same conundrum as Schrödinger's cat. You are either keeping cat food for a dead cat (if we mitigate), or have to handle being exposed to a hungry cat without food (if we do nothing and respond if it occurs).

Either way, we are exposed to an inestimable risk. Worse, we have hundreds of boxes, each with an animal requiring different kinds of food! So what now?

The key to planning for "inestimable risks" is coherence. The question is not "how likely is this to occur?", but "what is the boundary of results that we are prepared to live with?" By mapping out all of the plausible events that might occur, we don't have to mitigate against specific events, but simply take actions as required to either make a class of events implausible or be equipped to respond to them effectively. (Drawing a line on implausibility is another question for another day.)

Then, no matter which or how many of Schrödinger's boxes open, we can be confident that the organisation will be appropriately prepared.

Did you know...

Our expertise in complex systems analysis, combined with a deep understanding of technology and modern, agile management and leadership techniques makes knowquestion uniquely positioned to find strategic solutions to your tough problems. Contact us today.


David Griffiths (not verified) — Wed, 12/12/2012 - 18:51

Hi Stephen and an interesting post.

The only thing I would say is that my post was more leaning towards the concept of 'fat tails' and the apparent increase in low probability/high impact events; my assertion being that the mere existence of a fat tail tells us nothing (my, "so what" question) - to say nothing of the question of power law validity in the complex domain.

What is much more interesting to me is talking in terms of natural frequency, which is primarily used in the medical field to allow for clearer thinking in the decision-making process. This is where I slightly disagree, where you say, "The question is not 'how likely is this to occur?'". I believe that the 'likelihood' of the event directly effects the decision weighting process and therefore weights the terms of CBA and then energy (time) and resources (people, finance, tech) applied to solving a potential problem.

Good discussion, as always...


Post new comment

The content of this field is kept private and will not be shown publicly.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Allowed HTML tags: <a> <em> <strong> <small> <blockquote> <cite> <code> <ul> <ol> <li> <dl> <dt> <dd> <br>
  • Lines and paragraphs break automatically.

More information about formatting options