There was an opinion piece published in May by PropertyCasualty360˚. The title of it is, “Plethora of challenges face the insurance industry on the road ahead”. Most of the piece was fairly predictable. But this caught our attention:
Another troublesome challenge for the insurance industry is overcoming cognitive biases, such as recency or legacy biases. These biases cause policyholders to believe that because nothing bad has occurred, nothing bad will occur — until of course it does. Underappreciation of loss occurrence leads to underinsurance and a lack of resilience. Both are unfortunate, yet preventable.
We get the concept of cognitive bias. But we also recall the notion that insurance is not bought. It’s sold. And if inadequate levels of coverage are being sold because of policyholders’ cognitive biases, doesn’t that constitute greater claims liability for the insurance companies that underinsure their policyholders? There’s a kind of self-defeating circularity to that logic. Rather than cognitive bias, it seems to suggest cognitive dissonance.
According to Psychology Today:
Cognitive dissonance is a term for the state of discomfort felt when two or more modes of thought contradict each other. The clashing cognitions may include ideas, beliefs, or the knowledge that one has behaved in a certain way … The theory of cognitive dissonance proposes that people are averse to inconsistencies within their own minds.
In the case of insurance, cognitive dissonance may occur between these two modes of thought in the mind of a policyholder: (1) I live in a flood zone; therefore, I may be susceptible to considerable water damage. (2) I don’t want to pay for adequate insurance coverage to fully indemnify that considerable water damage.
It’s almost a cliché that the insurance industry is slower than most to adopt emerging technologies. But it does adopt emerging technologies, even if it does so at its own justifiably conservative pace. And if there’s anything the insurance industry has in abundance, it’s data, particularly experience data. So, as the insurance industry continues to adopt and assimilate machine learning and AI, it will become better able to predict loss potential. And those predictions will become increasingly accurate as more and more experience data informs the AI.
As that predictive capability continues to grow, insurers will be more able to plausibly have conversations like this:
Policyholder: I know I’m at risk. But I don’t want to pay for adequate coverage because I haven’t needed it yet.
Insurer: Yes. We understand the premium for fully adequate coverage seems high. But what will it cost if you don’t have that protection?
If insurers can keep policyholders from grinding their psychic gears now, they can keep them from grinding their teeth later.
Maybe then underappreciation of loss occurrence can become sincere appreciation for insurance.