The use of artificial intelligence and data analytics in life insurance
The emergence of data analytics and machine learning is providing insurers and reinsurers with new insights into how they drive and monitor their business.
As I write this article, we are in the midst of an unprecedented time for all of us, characterized by widespread social distancing and working from home caused by the COVID-19 pandemic. This situation has affected personal health and safety, and continues to have a devastating economic impact across much of the globe.
I live in the insurance capital of the world, Hartford, Connecticut. I've been an actuary employed in the property and casualty (P&C) insurance industry for over 30 years now, having been employed by four insurers. I currently work in a consulting capacity with the global firm Milliman. I am also the current president of the Connecticut Captive Insurance Association (CCIA). My career focus has largely been in the commercial lines.
I have read numerous articles about the pandemic, focusing on the impact of the virus to our citizens, how our society has responded to them, and then more specifically on my area of expertise, P&C insurance. As I look at all of the P&C insurance issues that have been discussed since the onset of the virus, it is becoming ever more apparent to me that the P&C insurance industry would benefit from more direct collaboration and communication with its customers. As a result, it would improve customers’ understanding of how their risks are managed and mitigated, leading to a successful insurance marketplace and enhanced public support for the industry.
We all want insurance markets to succeed; insurance is a critical function supporting our economy. This means that, in exchange for the transfer of all or most of a customer’s risk to an insurer at a fair market price, the insurer is positioned to make a reasonable return on equity across all of its customers. In doing so, the insurance industry demonstrates leadership and takes actions that improve the lives of the people it employs and the customers and communities it serves.
Let’s recognize that the insurance industry is justifiably not able to provide a complete insurance solution in all situations. In particular, this very reality spawned the captive insurance industry. We also know that, in recent years, we have seen the rise of insurtech due to opportunities within the industry. As a result, agents and brokers, who are also key constituents in this ecosystem, are adjusting their business models. All of these elements working together will continue to help provide full and efficient insurance solutions for customers.
However, the onset of the COVID-19 pandemic has once again revealed a fundamental issue that needs more attention from the traditional insurance industry—that insurance buyers often believe they have adequate insurance coverage when they actually do not. This leads to widespread dissatisfaction with the insurance solutions currently available to them. As a result, I believe we need to think even harder about working more collaboratively across these three insurance sectors with a holistic focus on customer satisfaction.
The insurance industry was never meant to cover all losses from a pandemic. Insurance is all about the law of large numbers—many insureds pay a premium so that insurance can be affordable for the few who ultimately need it. In the case of COVID-19, there are segments of our economy where this principle has been largely invalidated (e.g., restaurants, nonessential small businesses, household contractors, many gig workers, etc.); these businesses are virtually completely shut down and many insureds are looking for business interruption (lost income) coverage under their commercial property policies. The traditional insurance market is not meant to handle a situation where all insureds are making claims at once, and thankfully, our federal government has stepped in to provide temporary financial support.
On the other hand, much of the industry can do better. For example, the indemnification process for large, adverse events is often very inefficient for its affected policyholders and claimants. With COVID-19, variations in policy language for business interruption coverage 1 will lead to significant amounts of costly litigation, leaving businesses with a suboptimal insurance experience. COVID-19 is certainly a unique situation, but viruses in general are not new, and some insurance forms specifically address their impact. To its credit, the industry does support a new government fund to help businesses and workers hurt by COVID-19 shutdowns,2 and there are numerous companies refunding premium to policyholders who have reduced exposure to risk.3
Looking back, the industry has done a lot to manage catastrophe risk in general. For most traditional catastrophic perils, such as hurricane, tornado, earthquake, and flood, insurers have continued to improve upon models to assist them with underwriting and pricing. They realize they cannot fully fund this exposure in the private insurance market, so they collaborate with the reinsurance market and, more recently, with the capital markets (via insurance-linked securities). Insurance for terrorist events was made available due to the industry’s vision in recognizing that a federal backstop was needed for the product to function. While I believe there is a lot more that can be done to enhance public-private insurance partnerships, each of these efforts has been a step in the right direction. However, at the same time, the industry has not always been quick enough to adopt newer forms of insurance (e.g., parametric insurance, which can provide widespread insurance based on the trigger of a certain-sized catastrophic event). And for COVID-19, might better public-private coordination with government have facilitated a liquidity solution such that the industry could make very quick and needed payments to its insureds, and then seek government reimbursement?
Another vexing insurance coverage inefficiency that will arise with COVID-19 is related to certain workers. Many workers deemed “essential,” who have exposed themselves to the virus (e.g., grocery store clerks), will have a very onerous standard4 to prove that COVID-19 is a compensable occupational disease under workers' compensation state laws. I appreciate that our industry policies are a contract of adhesion—the insurer has more power. But while these workers and their employers will act in the interest of social good, it does not seem fair that they may not get compensated for lost wages if they miss work due to contracting the virus. We can debate or litigate whether these workers intentionally exposed themselves to this (“intentional acts” are not covered by insurance), or we could instead provide an efficient solution (e.g., likely a combined public-private government-backed plan). This resulting lack of customer understanding and satisfaction demonstrates that the industry is providing these businesses with a suboptimal solution. Policyholders don't always know what they are buying, and operating under a “buyer beware” mentality is likely not the best message for the industry to be sending at this time. Let’s recall that workers' compensation was set up to be a no-fault coverage—where the worker gives up the right to sue the employer in exchange for wage replacement benefits and reimbursement of medical costs. But the likely amount of litigation, focused on whether a claim is compensable (“arose out of and in the course of employment”) will be both staggering and fraught with delays as the court system deals with increasing backlogs given that they too have been shut down. This is detrimental to our industry’s reputation.
There are many other situations outside the COVID-19 pandemic where P&C coverage did not respond in the manner the consumer expected it would, again resulting in costly litigation. The wind versus flood debate with Hurricane Sandy. The questions of whether Sept. 11 was one occurrence or multiple occurrences and when coverage is triggered in liability policies. While we come to expect litigation, there’s no reason this situation cannot improve.
Another newer market segment that seems to be following a somewhat similar path is cyber. To this point, insurers have dipped their toes into the cyber insurance water by offering very limited amounts of coverage to lots of customers. They’ve done a great job for their shareholders by making lots of profit on it to date. And some insurers have done a very good job in helping to mitigate this risk for their insureds through training and proactive claims resolution. But if a truly large cyber event hits, what will the perception be of the industry when so many customers will not be sufficiently compensated because the industry has not done enough to create a backstop? And what if they then restrict coverage as typically happens in such cases?
I know that there are other ways to transfer risk if the traditional insurance industry does not do it efficiently. The proliferation of captive insurance companies in the last 35 years suggests that commercial insureds are ready to take on their own insurance risks when insurers either don’t want to at any price or won’t offer coverage terms at a reasonable cost. Captives can be more flexible forms of self-insurance and risk pooling that incentivize the insureds to actively manage their own risks and reduce costs. In addition, very large entities like Amazon have discussed offering insurance directly to customers. And the insurtech movement looks to help change the entire insurance-buying experience (e.g., Lemonade, a New York-based insurer of homeowners and renters, which donates a portion of its profits to charity if no claims are filed). In essence, noninsurance companies have recognized the opportunity to service the insurance market
Let’s be clear that the “traditional” insurance industry is not alone in struggling with providing satisfactory solutions to the market. Tech giants Uber and Lyft have had significant difficulty finding how best to structure their insurance offerings to drivers, both from the standpoint of the driver’s liability to others and for the driver’s own protection. Amazon is still developing its insurance offering. And you may have heard that certain types of small captives (known as an “enterprise risk captive,” or ones that make the 831[b] election for tax purposes) are under significant scrutiny by the Internal Revenue Service (IRS) for the rates being utilized.
Traditional insurers, however, also face additional challenges from many other fronts—climate change, social inflation, legislative changes (e.g., extensions of statutes of limitations on sexual abuse cases, continuing mass torts, etc.). And the world is so much more transparent these days—questionable actions cause immediate and forceful customer reactions that can easily cause severe reputational risk.
Thus, in order to achieve a desired level of success going forward, I think insurers will need to work more seamlessly with their customers. Insurers will need to continue to work collaboratively with all parts of the insurance ecosystem to ensure consumers are better educated, treated fairly, and adequately protected.
Here are a few other ideas to get us thinking along these lines of collaboration:
I believe insurers should work more to provide customers a complete risk mitigation solution-- not one that leaves customers less than fully reimbursed and/or facing a road of litigation. The more I see, the more it seems to me that greater collaboration and the pursuit of innovative tools in managing customer risks will be the key to a successful insurance marketplace and will also enhance public support for the industry.
1Goldburd, M. (April 10, 2020). COVID-19, Business Interruption Coverage, and the “Physical Loss or Damage” Requirement. Milliman Insight. Retrieved May 17, 2020, from https://us.milliman.com/en/insight/COVID-19-business-interruption-coverage-and-the-physical-loss-or-damage-requirement.
2Evans, S. (April 15, 2020). $500bn Pandemic Risk Insurance Act (PRIA) reinsurance program in the works. Artemis. Retrieved May 17, 2020, from https://www.artemis.bm/news/500bn-pandemic-risk-insurance-act-pria-reinsurance-program-in-the-works/.
3Molina, B. (April 7, 2020). Auto insurers including Allstate, Geico to give back millions as coronavirus keeps drivers off roads. USA Today. Retrieved May 17, 2020, from https://www.usatoday.com/story/money/cars/2020/04/07/allstate-american-family-insurance-coronavirus-covid-19-premiums/2959870001/.
4Blais, A. & Fleming, C.M. (April 9, 2020). If You Contract COVID-19 at Work, Is That a Compensable Workers’ Compensation Claim? Milliman Insight.Retrieved May 17, 2020, from https://us.milliman.com/en/insight/if-you-contract-covid-19-at-work-is-that-a-compensable-workers-compensation-claim.