The global insurance industry reportedly spends nearly $200B annually on technology,1 never knowing which technologies will survive into the future. To help visualize this future, the research firm Gartner publishes a graphic each year called the Hype Cycle that represents emerging technologies’ maturity, adoption rates and real-world relevance. The curve essentially compares public expectations over time to show how new technology has larger-than-expected impact in the long run, but smaller effects than initially assumed in the short run.2 A quick rummage through our garages or basements will probably remind us that not all innovations ultimately ascend Gartner’s “slope of enlightenment” that elevates technology from the “trough of disillusionment” (failed experiments and divestments) to the “plateau of productivity” (mainstream adoption). Therefore, it is instrumental to an organization’s success to be able to sift through hype and place informed bets on which technologies will prevail.
Actuarial science may not be the first profession that comes to mind in connection with hype, but actuaries can play a critical role in curating it. Most of the topics on recent Hype Cycles — including blockchain, artificial intelligence and autonomous driving — are absent from actuarial exams and many actuaries’ everyday work. However, these topics are ubiquitous at recent CAS conferences and in publications such as the Actuarial Review, so the absence is not due to unawareness. Actuaries even have their own Standard of Practice (ASOP No. 13) devoted to conducting trend analyses to project future values. If actuaries astutely characterize technology trends as well as they are capable of doing, they have the potential to be powerful strategic voices in their organizations. If they instead consistently misestimate technology’s potential, they risk losing relevance to professionals more closely associated with new technologies, such as data scientists. As actuaries attempt to master the science of hype, they should consider the following “laws.”
Law #1 — Provide measures of hope
One of the greatest strengths actuaries bring to hype is a fluency in speaking quantitatively about trends. ASOP No. 13 presents several different ways actuaries can do so, such as using point estimates, ranges or probability distributions. Such premeditation is not necessarily typical of what one would hear when consuming hype, which may contain vague or out-of-context statistics with the potential to lead one to the wrong conclusion. For example, it is often noted that autonomous vehicles (AVs) could reduce or eliminate the 94% of car crashes caused in whole or in part by human error.3 Such a reduction would logically diminish the demand for auto insurance, which indemnifies policyholders for liabilities and damages related to such crashes. While this possibility may cause some auto insurers concern, the actual reduction in demand would probably not be a full 94%. Actuaries can help their organizations more precisely estimate how much AVs are likely to reduce demand and when.
James Lynch, FCAS, who is chief actuary at the Insurance Information Institute and a former journalist with the Miami Herald, has been covering AVs for nearly a decade. Lynch points to hyped hypotheses that AVs will spell the end of accidents or cause products liability coverage to supplant auto insurance as testable. “There is substantial data in the public realm regarding how long people hold onto vehicles, how long technologies take to mature and how effectiveness of vehicle safety features varies by manufacturer,” he says. “If an accident occurs because someone failed to maintain an AV, could product liability cover that? Will public policy evolve quickly enough, if at all, that you will start to see more of these vehicles en masse? Will people be able to afford them?” he asks. “The more questions you ask, the less likely you see . . . the hyped possibilities happen overnight.”
Law #2 — History repeats itself
ASOP No. 13 offers detailed guidance regarding the use of historical data to analyze trends. The past may seem less relevant to understanding technology than claim frequency or severity. However, many hyped technologies are clever repackagings of technologies actuaries know and understand. For example, AVs synthesize technologies including radar, LIDAR and global positioning. Similarly, blockchain makes use of cryptography, backup and distributed computing. Understanding hype’s ancestors not only breaks a larger problem into more digestible sub-problems, but also helps form a more refined view of viability. The Lindy Effect states, “Future life expectancy of some non-perishable things like a technology or an idea is proportional to their current age, so that every additional period of survival implies a longer remaining life expectancy.”4 In other words, if a predecessor technology has failed to thrive, the apple may not fall far from the tree.
The Lindy mindset is especially useful in interpreting instances where hype descends, in part, from other topics that are currently trending. For example, blockchain is largely a response to security challenges of transacting with virtual currency. However, as insurers and others look for ways to generalize this technology to their businesses, they may consider what their own relevant security challenges are. Many of these challenges may relate to pursuits with connected homes, vehicles and workplaces and the larger internet of things. Lo and behold, these technologies also reside on Gartner’s hype cycle — and are far from ascending the “slope of enlightenment.” One could thus consider a single new technology a response to challenges largely created by adopting other, not fully proven technologies. Actuaries may assign a different credibility to hype when its ancestry is more versus less speculative.
Law #3 — Connect the dots
Shifts in the zeitgeist can materially influence the direction and magnitude of a trend. ASOP No. 13 advises actuaries to consider economic and social influences in their trend analyses. The sharing economy — a trendy topic at many actuarial conferences in recent years — provides an instructive example of why. At their core, on-demand services such as ridesharing use technology to align supply with demand quickly and at scale. Postmates, whose platform connects consumers with local couriers, is one of the companies born in the past decade whose technology helps power this sector. However, Mustafa Rahman, software engineer at Postmates, observes, “The basic approaches underlying many on-demand services are not new. Integer programming has been around since at least the 1940s.” Actuaries pondering what “the next sharing economy” is should start by considering what changes led to the original sharing economy boom.
Rahman, who previously worked at CSAA Insurance Group, cites a number of factors that helped contribute to the rise of companies like Postmates. “Conditions became more favorable for execution,” he says. “Streaming APIs sped up processing.5 Enough data accrued over many years to make old problems solvable. Celebrities started using services like ours and promoting them over social media.” In other words, gradual changes stoked the embers of the sharing economy, but chance occurrences poured gasoline on the fire.
Vision is required for technology to boom even when conditions serve as accelerants. Actuaries possess aptitudes to provide this vision for their organizations. David Clark, FCAS, senior actuary at Munich Re, points to associative thinking — defined by Illumine Training as “the process of linking one thought or idea to another” — as one such aptitude. “The key is seeing what’s trending in one field and reflecting on whether and how it might apply in another,” Clark says. “Consider parallels between sociological analyses of students progressing through the education system and actuarial analyses of the life of a claim. There are strong similarities there, but it takes effort and intellectual curiosity to connect these dots.” Luckily, actuarial trend analysis is quite literally an exercise in connecting dots.
Law #4 — Consider the source
ASOP No. 13 urges diligence in communications and disclosures related to trends. Actuarial standards generally counsel that actuarial reporting should be sufficiently clear for other qualified professionals to appraise the reasonability of the analysis. Such clarity is not typical of hype. “The social institution of hype is a kind of specialization of labor, where we ‘trust’ some parties and they do the homework,” says David Wright, ACAS, managing director at Beach Re and host of the Not Unreasonable podcast. Susanne Sclafane, FCAS, executive editor at Wells Media Group and previously a practicing actuary, adds, “I worry sometimes about the fact that if the media reports on something, it feeds a frenzy that might not be real.” Fortunately, actuaries are experienced in separating truth from fiction.
In a 2017 article that appeared in the Actuarial Review, Stephen Mildenhall, FCAS, offered, “Actuaries write headlines about risk.”6 This quote evokes the similarities between actuarial science and journalism. ASOP No. 23 on Data Quality advises actuaries to consider the extent of any checking, auditing or verification performed on data they rely upon. This extent can be difficult to determine with hype. Sclafane observes, “Some articles might cite Bloomberg or The New York Times as a source. Those sources, in turn, might link to McKinsey or Deloitte as a source. Those might even cite other sources. It often takes hours to find the original source. Only then can you begin to determine whether the result is worth citing, by reading the original document to determine exactly how a number was derived, and how old the related study is.” Actuaries cannot speak confidently about the trends they hear about without having confidence in the underlying information.
Law #5 — Look in the mirror
ASOP No. 13 suggests that actuaries consider the effect of known distortions that could influence how they perceive trends. Ironically, one such distortion could be their own preconceptions. “Actuaries are probably less likely than others to trust hype, particularly in their domains of expertise,” says Wright. “But they have weaknesses to tropes like the rest of us, especially ‘intellectual superiority’ kinds of cognitive traps,” he says.
Actuaries cannot speak confidently about the trends they hear about without having confidence in the underlying information.
Clark similarly hypothesizes that actuaries may instinctively lean bullish or bearish on hyped technology depending on the circumstances:
- Bearish: “There is a tendency for actuaries and others to dismiss sci-fi-like buzzwords such as ‘artificial intelligence,’ or AI, after hearing about them ad infinitum,” says Clark. “But by dismissing AI as a whole, one may also be dismissing everyday automations that could help their companies substantially.” Clark points to autocorrect as one of the earlier examples of AI, asking, “Could our predispositions against AI lead us to miss the next autocorrect?”
- Bullish: Technical types may have a blind spot for hype presented in a technical vocabulary. As an example, Clark points to techniques that gained rapid popularity in the profession. “Markov chain Monte Carlo [MCMC] methods are a clever solution for integrating over highly dimensional spaces and have rapidly gained popularity by the standards of ‘technical hype,’” he says. “At the same time, much of the related technology is still computationally intensive; it requires compromises in how variables are defined, and the models may not always converge. It may not be wise to throw MCMC at a problem when simpler methods are available.”
To help manage one’s preconceptions, Wright points to the concept of “steelmanning,” that is, developing the best version of an opponent’s argument. “When looking at what others are talking about, do whatever it takes to answer the question, ‘How can this be right?’,” he says. This approach is oriented towards helping avoid either of the misestimating tendencies cited in our introduction.
Getting to the bottom of hype requires a lot of work. If it were easy to do, more people would probably be analyzing hype, and fewer people would be hyping. Actuaries’ skills and standards of practice make them uniquely qualified to be their organizations’ curators of hype. If they rise to this challenge and keep their standards relevant in a rapidly changing technology environment, they may find themselves possessing a rare form of control. “Maybe the only part of hype that’s controllable is people’s feeling of understanding something complicated or new without needing to do the work,” concludes Wright. “Actuaries can help simplify explanations of complicated things.”
1 Source: 2017 Celent Study, https://www.celent.com/insights/980614747.
2 This tendency is called “Amara’s Law.”
3 Source: National Highway Traffic Safety Administration, https://www.nhtsa.gov/press-releases/usdot-releases-2016-fatal-traffic-crash-data.
4 Source: Nicholas Nassim Taleb, Antifragile (2017), https://books.google.com/books?id=5E5o3_y5TpAC&pg=PA514#v=onepage&q&f=false
5 Streaming application programming interfaces (APIs) are sets of functions and procedures that essentially send data over the web to subscribed parties whenever a particular event happens.
6 Source: “The Coming Revolution in Actuarial Modeling,” https://dev-cas-actuarial-review.pantheonsite.io/the-coming-revolution-in-actuarial-modeling-election-day-lessons-for-the-predictive-data-analyst/.
Jim Weiss, FCAS, CSPA, works for Crum and Forster as an actuary in Morristown, New Jersey. He currently chairs the CAS Education Task Force and is an editor for Actuarial Review.
Khanh Luu works for AIG as an analytics manager in New York City and is a CAS candidate.
