In this blog about actuarial and data science in the insurance industry, we’ll take a high-level look at how AI and machine learning will help solve issues brought on by the COVID-19 pandemic.
Some events take place and stay in our memories forever. Many people have vivid remembrances of 9/11 as clear as if it had happened yesterday. I remember sitting at breakfast in a Holiday Inn in Atlanta that morning and can recall everything that happened that day, including his long drive back to Ohio. My friend Brad was in NYC and vividly remembers the events and the horrors, along with the fact that it took a week to get back to Dallas.
Events like this have a significant impact on every one of us. The arrival of COVID-19 in March 2020 is another example of this phenomenon. There wasn’t an exact date or time of arrival, but it is, and always will be, an event that changed our world, affecting everything from how we work to how we conduct our personal lives.
COVID also made a major impact on the insurance industry. According to the New York Times, the United States is approaching 80 million cases and over 900 thousand deaths. As a result, there have been significant changes to life expectancy, automobile and home office usage, healthcare, and most aspects of our lives. These changes directly affect the three major lines of business within the insurance industry in North America.
In our two-part series on the impact of COVID on the US insurance industry, we will focus on how actuaries and data scientists in life and annuity (L&A), property and casualty (P&C), and health insurance companies are dealing with the effects of the pandemic. We also explore what machine learning (ML) and artificial intelligence (AI) tools are out there to assist in making decisions for their specific line of business.
Changing the Game: The Impact on Life and Annuities
Insurance carriers base much of their risk profiling and the pricing of their products on experience or legacy data. Actuaries use this experience to derive patterns that develop risk profiles and pricing for individual types of products. Life expectancy, for example, has been stable for the past 100 years with slight increases as healthcare and safety innovations have gradually increased our longevity. In 2019 (pre-pandemic), the life expectancy for a male was 76.23 years and 81.28 years for a female. As of 2021, provisional COVID-19 deaths resulted in a decline in life expectancy of 1.7 years from pre-pandemic levels.
COVID-19 has quite literally changed the game, but it’s not the only factor impacting life expectancy. During the pandemic, many people had to put off non-emergency surgeries or were hesitant to visit their doctors. The long-term health effects of COVID survivors are also still unknown. All these factors impact mortality rates, and experiences before 2020 won’t tell us the full picture.
With a lot of unknowns surrounding life expectancy, actuaries that turn to AI and machine learning will gain new insights. Using ML algorithms and predictive analytics, actuaries can quickly evaluate many variables to identify patterns and relationships that might have otherwise gone unnoticed. For example, based on historical data, actuaries can infer which variables best predict how customers and policyholders will act in the future.
Automation, AI and ML will be crucial for life insurers who want to harness data analytics to gain a deeper understanding of customers and remain competitive in the market going forward.
Navigating Disruption: The Impact on Property & Casualty Insurance
Property and casualty insurers are experiencing similar problems with their products because of the pandemic. A simple example is automobile insurance – people are not driving as much as before COVID hit. Normally, people put 15 to 20 thousand miles a year on their cars during their daily commutes. Those miles have drastically reduced as cars sit in the garage while customers work from their home offices. Many policyholders received rebates from auto insurance carriers as a reaction to this, but how will carriers price policies going forward?
Another consideration is whether remote work is here to stay and, if so, what the impact is on homeowner’s insurance. People are now in their homes more often and have expensive equipment in their home offices. What does this do to their risk profile? And how does this impact products like workers’ compensation and liability?
The property and casualty space will continue to see varying effects on claim frequency and severity due to COVID-19. P&C insurers have millions of claims flowing at any given time for auto and homeowner’s insurance. The unknowns around these products and risk profiles will also lead to an increased use of analytics that requires augmentation with AI and machine learning.
Reimagining the Future: The Impact on Health Insurance
For the health insurance line of business, how does putting off doctor visits and minor surgeries or not getting vaccinated impact future health costs? How does the rise in depression during the pandemic affect rates? Carriers are working diligently to address these issues.
Using AI and predictive models, health insurers can create more accurate risk models that predict which individuals need what type of care. This results in more accurate predictions of group risk and pricing of health insurance policies. AI can also help forecast claims by weighing the estimates of prospective risk according to accuracy and viability. The use of machine learning by health insurers will not only make them more efficient but also help them remain competitive in their market.
Solving the Conundrum
In our next article, we will delve into what carriers are doing (or should be doing) to address these concerns and offer some ideas based on our experience in the insurance and analytics spaces. We’ll offer ideas around integrating more recent knowledge with the history already obtained by the carriers. We will also explore modern analytics processes, including AI and ML, which provide actuaries and data scientists with the tools and information required to position their respective companies to thrive and profit in this new environment.
Our white paper explains an effective system to ensure you always keep a human in the loop when implementing AI.
