Communicating Uncertainty

The language of uncertainty permeates our lives. Whether a starting athlete is listed as questionable to play in the weekend game, or snow is likely in the afternoon, we often make attempts to place quantitative sentiment on uncertain events. Back in 1964, analyst Sherman Kent compiled “Words of Estimative Probability,” a since-declassified brief for the Central Intelligence Agency that attempts to map numeric estimates to linguistic probabilities.

Kent’s “Estimative Probability” writings assert that there is a language for odds; in fact, there are two: 1) the precise mathematical language of the actuary or the racetrack bookie; and 2) a less precise though useful verbal equivalent.

In an effort to reconcile the two languages, six ranges for phrases were created:

1. Certain = 100%
2. Almost Certain = 93% (+/- 6%)
3. Probable = 75% (+/- 12%)
4. Chances Are About Even = 50% (+/- 10%)
5. Probably Not = 30% (+/- 10%)
6. Almost Certainly Not = 7% (+/- 5%)

Similar efforts have been made in medicine and finance. For actuaries, communicating uncertainty is more than just an exercise in vocabulary; in some cases, it is a necessity.

Actuarial Standard of Practice (ASOP) No. 41, Actuarial Communications (December 2010), §3.4.1, states, “The actuary should consider what cautions regarding possible uncertainty or risk in any results should be included in the actuarial report.”

Let’s dig a little deeper into the “consideration” of what to communicate with respect to uncertainty.

Let’s step away from actuarial practice for a moment and look at a different example to understand some implications of disclosing uncertainty in the world of investments.

Derek Harmon, Ph. D. (University of Michigan), and Ivana Naumovska, Ph. D. (graduate business school INSEAD) authored a paper in which they studied the degree to which founders of Special Purpose Acquisition Companies (SPACs) used uncertain language in their Initial Public Offering (IPO) communications with investors and how that affected their financing outcomes.¹

For all SPACs between 2003 and 2019, Harmon, Naumovska, and their team used natural language processing methods to capture the degree of uncertainty related words in each S-1 IPO filing (illustrated below). They then used this measure to predict IPO success:

Frequency of the top 10 uncertain words in the full S-1 forms

Harmon, Naumovska and their team found that the benefit (or cost) of SPAC founders communicating their uncertainty depended entirely upon the type of uncertainty they and their investors faced.

One type of uncertainty arises due to information asymmetry, where SPAC founders know something that their investors do not (e.g., their business strategies or information about
their top executives). In cases where SPAC founders had an information advantage over their investors, the team found that communicating less uncertainty led to better IPO outcomes.

However, a second type of uncertainty can arise when something is fundamentally unknowable to both SPAC founders and investors (e.g., the future acquisition target or future regulation). When SPAC founders did not have an information advantage over investors, and both parties faced the same unknowability, the team found the opposite—that communicating more uncertainty led to better IPO outcomes.

How should we understand these results? Shouldn’t expressing more uncertainty make SPAC IPOs seem riskier? It does, but only when founders have the information advantage over investors. Since you shouldn’t convey uncertainty when your investors expect you to know something, the logic flips when founders themselves cannot, in principle, know something. In these circumstances, acknowledging the uncertainty that is truly present in the situation conveys credibility and honesty, giving investors more confidence.

So, what is an actuary to do? How can we extrapolate the findings from SPACs to actuarial practice?

I think it illuminates that there are a few key questions in framing a discourse on uncertain events.

• What is the balance of information?
• Is there a knowledge advantage?
• Do both parties face high degrees of uncertainty?
• What is the appropriate quantity of information to relay under uncertain circumstances in order to ensure credibility and authenticity?

Returning to the initial discourse on linguistics and mathematics, there is one more question to ask. Are qualitative statements also accompanied by numeric displays? Confidence intervals, volatility measurements and scenario analysis are all attempts to give numeracy to possibility.

Finally, the volume of language about uncertainty may not be the only thing that matters. Harmon, Naumovska and their team point out that other dimensions, such as sentiment, how concrete or vague the language is, and what rhetorical strategies are used may play a role. For example, they offer that an overly positive but highly imprecise statement may be less credible than a negative but precise statement. The sophistication of the audience adds yet another dimension.

In uncertain times, technical prowess is undoubtedly useful, but being able to communicate effectively in a way that gives stakeholders confidence is arguably just as important. Through intentional disclosure of uncertainty, actuaries are well-positioned to strengthen confidence, enhance decision-making, and help generate more favorable outcomes—well, almost certainly.

The views expressed in this article are those of the author and do not necessarily reflect the views of the Society of Actuaries. The content in this article is provided for educational discussion and does not constitute professional, investment or regulatory advice.

Copyright © 2025 by the Society of Actuaries, Chicago, Illinois.