A study published in a medical journal concludes that there is a significant correlation between COVID vaccine hesitancy and the probability that someone will be in a serious car accident. Specifically, unvaccinated people had a 72% higher probability of being involved in a crash. Even after correcting for “age, sex, home location, socioeconomic status, and medical diagnoses,” there was a 48% higher relative probability of being in an auto accident among the unvaccinated. The analysis was based on a population of more than 11 million people.
When I came across this study, I was immediately reminded of another controversial statistical link to auto accidents: the relationship between an individual’s credit score and their propensity to make a car insurance claim. In most U.S. states, insurers use consumer credit scores as part of the algorithm that determines how much you will pay for car insurance. The research linking credit scores to the probability of having an auto insurance claim is well-documented (also see here, here, and here). The credit-auto risk connection has interesting implications for further explorations of the meaning of the vaccine resistance -auto risk relationship.
The discussion of the link between credit management and car insurance claims is typically explained by the hypothesis that people who are careful managers of their money are also likely to be careful drivers. This has been referred to as the good stewardship argument. The extension of this concept to explain the vaccine hesitancy link is straightforward. Simply put, people who believe that the guidelines proposed by experts are likely to be biased or misguided infringements on their freedom may naturally be inclined to ignore driving rules and norms. This is, of course, just an hypothesis. Perhaps the common theme is whether people tend to follow the rules (pay your bills on time, obey the speed limit, get a vaccine if the government suggests it).
From a insurance perspective, are the vaccination findings useful? There are insurance companies that glean information from social media profiles, and they can get considerably more information if you ‘friend’ them. The insurer can offer a premium discount if you are willing be be friends. Simply put, the more information that an insurance company has about you and your life, the better they can calculate how much to charge in premiums. Might insurance companies seek to access medical records as a part of underwriting? The insurance company could only have access to such information if you give them permission (due to medical privacy laws), but I see little reason not to give my COVID vaccination record to my insurer if they are willing to give me a discount. Aside from privacy questions, is it reasonable for the insurer to use my record of COVID shots in determining my insurance premium? Assuming that the research holds up, the answer is yes. It does not matter that we cannot nail down a causal relationship. There is no clear causal factor between credit scores and driving safety, but the correlation is sufficient reason to include credit history in setting premiums. Assuming the correlation between between vaccination history and auto accidents is reproducible and holds up with more data, it seems probable that some insurers will offer opt-in programs in which they will give a premium discount if you are fully vaccinated.
Life insurers can charge a higher premium to smokers and it is not optional to answer the smoking / non-smoking question on an application (as far as I know). Medical research has concluded that smoking reduces life expectancy. The medical establishment has also concluded that getting a COVID vaccination is the best practice. It seems entirely reasonable, on this basis, for health and life insurers to charge higher premiums for people who are not vaccinated. Some companies are already doing this. It is something of a leap from this type of causal link to charging higher premiums on the basis on variables for which the predictive value is entirely justified by correlation.
The use of correlations without a clear causal link in determining insurance premiums is a somewhat slippery slope. Insurers charge young men higher auto insurance premiums than women because the data shows that there is a correlation between gender and driving safety. There is no causal connection that explains why this is the case. Insurance companies are also allowed to use credit scores to determine premiums, depending on the state and application and there is no direct causal link here, either. If individuals allow an insurer access to their personal information, it does not seem like a stretch that medical choices (such as vaccinations) might be justified as part of rate setting. If insurers seek to go down this path, there will certainly be long and protracted legal debates–just as there were when insurers first sought to use credit scores. The ultimate extension of this type of process is where things get difficult. What if credible research concluded that people with tattoos tend to have more (or fewer) car accidents than people without tattoos? Would it be reasonable to use someone’s tattoos in setting auto premiums?
Geoff Considine on Finance and Economics 