Risk and Uncertainty
All project managers have experienced a situation that goes something like this:
- You try to explain your project’s “contingency” budget line item to a bean counter (or lay person).
- The bean counter thinks that the project is hoarding beans, has not bothered to fully flesh out the scope and budget, and refuses to approve the contingency amount unless it can be notionally applied the current set of line items.
- To which you respond – for the third time and to the same blank stare – that if what the contingency was going to be used for was knowable, it would no longer be contingency. It would become a line item in the budget.
Contingency is how project managers deal with uncertainty. But before further discussing uncertainty, let us first define risk.
Risk is not the same as uncertainty.
Risk is the product of the probability of the occurrence of that risk event times the consequence of that occurrence (where both probability and consequence are quantified).
Risk = Probability x Consequence
Risk can broadly be mitigated. Risk quantification generally fits well within a modelling or probabilistic evaluation framework.
Nevertheless, one problem with modelling exercises in general is that they can create a false sense of accuracy, for example due to overstated precision or understated complexity. Because of the precision or accuracy that models insinuate, outputs are often treated as sacrosanct. However, the means by which outputs were derived (e.g., data selection methods, model assumptions, earlier but unpublished modelling results) can provide valuable insights for complex risk-benefit analysis and problem solving, which will often exceed the information value of the easy-to-comprehend outputs themselves.
If models are trained on a certain dataset or time series, they will not accurately forecast data points that do not conform to the assumptions ascribed to the dataset in the first place. This is often the case with economic models or financial models, which are dynamic and continually adapting systems.
Financial crises or material construction accidents by definition fall outside of the prescribed expected bounds of their models otherwise they would never be allowed to happen. If one knew that a crisis or serious accident was coming, mitigation measures would be taken that would not allow such an event to exceed a predetermined threshold (i.e., making it neither a crisis nor an accident).
Such data points represent uncertainty.
Risk management is a necessary, but not entirely sufficient, set of guidelines to inform strategic decision making.
Uncertainty is not the same as risk.
Uncertainty represents outcomes that fall outside of expectations. It comprises ill-defined vectors that influence outcomes in unpredictable ways. Uncertainty is the proverbial “unknown unknown", to paraphrase a former US Defense Secretary. It is the possible but unquantifiable.
Examples of uncertainty are hail-mary touchdown passes, technological breakthroughs, financial crises, or a regularly occurring perfect storms.
Uncertainty is not easily or even practically quantifiable. Another way to think about uncertainty is as data points that do not fit into a rudimentary a normalized bell curve.
Good planning and project management processes should adopt strategies to proactively recognize and respond to uncertainty.
What to be done with uncertainty?
For solutions, I turn to finance and investment industry  (which employs strategies for protecting against uncertainty).
- Build a portfolio, targeting negative correlation or no correlation between the price paths of the assets.
- Seek out “anti-fragile”  assets – which is to say assets (or strategies or facilities) which improve under duress (e.g., electrical capacity is anti-fragile, while electrical energy is not).
- Avoid intellectual certainty – if everything seems to be a nail, question whether or not it is because you are in possession of a hammer.
Practically speaking, any strategic planning, project management, project execution, or system / resource planning exercise will benefit from the acknowledgement and incorporation of uncertainty analysis.
At Midgard, we pride ourselves on our independent and original problem-solving skills.
 My personal journey to my current position started out in central banking, portfolio management, derivatives risk measurement, and commodities trading and deal structuring before landing in project development. Do not ask me to design you a bridge, but I know my way around the personal investment landscape.
Recommended reading: “Antifragile: Things That Gain From Disorder”, by the brilliant and irascible Nassim Nicholas Taleb.