Evaluating a risk is much different from evaluating an uncertainty. Both describe different states of knowledge about future events, but uncertainty presents greater challenges to risk managers, particularly in areas like loss estimations and operational resilience.
Brenda Boultwood
Risks can be quantified when both the objective probability and the potential future states are known. Knightian uncertainty is a future state where it is impossible to assign future probabilities to unknown future states. It represents a lack of predictability due to insufficient information, often stemming from a lack of experience with the potential future state.
In risk analysis, two issues always arise. The first is the measurability of probabilities and the second is the knowability of the potential future states.
In risk analysis, not accounting for Knightian uncertainty leads to an incomplete and potentially flawed understanding of a firm’s true spectrum of risks. This can result in an inability to prepare for truly novel events, an underestimation of future losses, an overreliance on quantitative analysis, poor decision-making, and inadequate operational resilience.
But how can we distinguish risk from uncertainty, and which practices can risk managers use to mitigate uncertainty threats?
It’s worth restating that risk is when objective probabilities and known outcomes are available. There are many examples of risk analysis, including the roll of a die, the likelihood and cost of a car accident, the calculation of actuarial probabilities for insurance, and the hazards faced by property owners in geographic regions that are more prone to natural disasters.
In a standard risk management process, risk objects define the scope of ERM analysis. A risk taxonomy establishes a common language of risk, while risk identification seeks to acknowledge the existence of a potential hazard that could impact decision-making.
A careful assessment of hazards related to a risk object should distinguish the risks from the uncertainties (see Figure 1).
Figure 1: The ERM Risk and Uncertainty Decision Point
The distinction between risk and uncertainty makes a difference in three areas: decision-making, economic analysis and risk management.
In decision-making under uncertainty, expected values based on probabilities and outcomes are not available. Instead, decisions are based on judgment, intuition and qualitative assessments.
Economic analysis, meanwhile, assumes rational agents with complete information about the probabilities and possible outcomes in a future state. (Knightian uncertainty challenges this assumption and suggests that many real-world economic decisions are made under conditions of genuine ambiguity.)
A standard risk management process, as depicted in Figure 1, focuses on identifying, assessing and treating quantifiable risks. Knightian uncertainty, on the other hand, highlights the existence of "unknown unknowns" that are difficult to anticipate and manage. Dealing with Knightian uncertainty often requires building consensus, gathering expert judgments and adapting to unforeseen events.
Effective uncertainty analysis requires distinguishing between risk and uncertainty early in the risk management process. This can be achieved, in part, through the use of six commercial best practice methods (see Figure 2).
Figure 2: Commercial Best Practice Methods for Uncertainty Analysis
Using the above methods, any organization can align stakeholders on the nature of the uncertainty, drive consensus about the possible consequences, and support good decision-making based on the information available. These approaches to uncertainty analysis allow for a conclusive assessment of the potential hazards impacting decision-making.
Let’s now take a close look at each of the six approaches frequently used for uncertainty analysis:
Brandon Norton
Subjective Probability Assignment. By far the most common approach is to allow risk owners to assign their “best guess” probabilities and outcomes to an uncertain analysis. Grounded in the work of British philosopher and mathematician Frank Ramsey, the assignment of subjective probabilities assumes that individuals have degrees of belief, even in uncertain situations.
Conviction Narrative Theory. When faced with truly unknown and unquantifiable futures, human decision-making shifts from probabilistic calculation to narrative construction, emotional evaluation and the development of subjective conviction. CNT provides a framework for understanding the cognitive and emotional processes that enable action in a world where traditional decision-making tools are inadequate. There are many situations in which a bank might use CNT, including when deciding whether to buy, sell or hold a position.
The Maximum-Minimum Method. When assigning a subjective probability, experts can agree upon a range at which an event is possible. Each expert can offer an informed opinion about the probability range of an event occurring, with a maximum likelihood and a minimum likelihood.
After assessing the probability range, the decision maker can choose the maximum or the minimum probabilities provided by the experts. Alternatively, another probability in the range could be chosen (perhaps somewhere in the middle), based on the preference of the decision-maker.
Black Swan Analysis. A black swan event occurs when historical data and statistical models fail to predict truly novel and impactful events. Such events have monumental consequences, and seem ex ante impossible but ex post plausible. The end of the U.S. dollar dominance in the global financial system would be one example of a Black Swan event.
Black swan analysis considers the types of unexpected events that could trigger such a shift, including geopolitics, negative sum tariff policies, and other U.S. government policy mistakes. One objective of such analysis is to build resilience against each potential trigger.
For an organization dealing with a black swan event, the most common approach to mitigate uncertainty is through constant threat assessments of unlikely but possible scenarios.
Grey Rhino Analysis. On the other side of the uncertainty spectrum lies the “grey rhino” – slow-moving events with high probabilities and extreme consequences. While the outcomes of these events may prove difficult to predict, the fact that they are highly likely to occur means they can be identified.
One example of a grey rhino event is a financial market crash. Many factors – including growing U.S. debt, high inflation and the likelihood of a U.S. government debt default – point toward an impending crash.
The critical issue is how organizations prepare for such an event. They can either continue to neglect the signs or take preventive action – through, for example, developing preparedness plans that outline a plan to evade catastrophe.
Grey rhino events can be extremely consequential, particularly if firms ignore warning signs and fail to prepare properly.
Resiliency. Planned resiliency can be effective in situations where risks cannot be quantified or even fully understood because of a lack of historical data or precedent. Organizational resiliency can be improved through enhanced flexibility, diversification, redundancy, psychological capacity, and iterative and adaptive small decisions.
An example of operational resilience in the banking system is regulatory adaptation of bank reserve ratios that banks must maintain to stay open. When smaller banks are found to be systemically important, reserve ratios for banks with less than $250 billion in assets could be adapted to provide greater resilience in the event of a banking crisis.
A risk assessment is a standard tool used by every firm to acknowledge the likelihood of future hazards and threats. However, it relies heavily on objective probabilities and known outcomes that are rarely available.
Consequently, financial institutions often use subjective probabilities to combat adversities and to arrive at a scenario that is most likely to occur. Through uncertainty analysis, risk managers can provide a more conclusive risk assessment to decision makers. Firms will benefit from this analysis by being better prepared for future hazards and threats.
Brenda Boultwood is the Distinguished Visiting Professor, Admiral Crowe Chair, in the Economics Department at the United States Naval Academy. The views expressed in this article are her own and should not be attributed to the United States Naval Academy, the U.S. Navy or the U.S. Department of Defense.
She is the former Director of the Office of Risk Management at the International Monetary Fund. She has previously served as a board member at both the Committee of Chief Risk Officers (CCRO) and GARP, and is also the former senior vice president and chief risk officer at Constellation Energy. She held a variety of business, risk management, and compliance roles at JPMorgan Chase and Bank One.
Brandon Norton is a Midshipman First Class at the United States Naval Academy studying mathematics with economics. His research focuses on uncertainty analysis and practical application of approaches to decision-making under uncertainty.