I have always been fascinated by novel “medical tests” that can forecast things about my distant and future health prospects that are currently hidden from the naked eye, many years before any visible symptoms emerge. I personally have taken a fair share of these assessments especially as I continue to age (chronologically) thru the human lifecycle. In the language of probability theory, and pardon my geek Greek, I prefer having the largest possible sigma field of information about my stochastic mortality rate. Not everyone agrees of course, especially if there is little you can do to change the course of these gloomy predictions.
Now, up until recently I thought there was an ethical or moral duty to disclose this information — if I learned of any bad news — to insurance companies so they could charge the proper premium for my risk coverage. After all, if you are about to purchase life, health or critical illness insurance, even if you don’t suffer from any symptoms of pre-existing conditions now, you are riskier (to the company and your cohort) and should pay more. To me it just feels like the right thing to do, although I understand the topic is controversial, perhaps not legally required and in some cases outlawed altogether.
But a fascinating study out of the University of Amsterdam’s Research Centre for Longevity Risk, where I was recently invited to visit by the director Professor Torsten Kleinow, has made me question some of these prior beliefs. In a series of articles with his research co-authors, Oytun Hacariz and Angus Macdonald, published in the Scandinavian Actuarial Journal, they delved into specific genetic testing and how important – or perhaps not – it is for life insurance pricing.
In fact their research suggests a number of intriguing implications, which actually goes beyond life insurance pricing. Apparently, given all the different factors that can kill you, knowing that you have an elevated risk of dying from only one overall factor doesn’t necessarily increase the total chances you will die. Curious and perhaps morbid consumers (like me) who take the time to ‘test and learn’ about all that is likely to kill them might actually have a higher (not lower) life expectancy and should be paying less (not more) for their life insurance policy.
Yes, this all sounds very counterintuitive…
Here is some simple math to understand the underlying intuition of their very odd claim. Assume that you are subject to ten different uncorrelated causes of death or ‘diseases’ each with a 0.5% mortality rate. Yes, that’s rather simplistic, but bear with me. Adding them all up the total mortality rate and chances of dying in the next year are 5%. From a financial perspective, this implies that a simple one-year term insurance policy that pays $1,000,000 to your beneficiaries should cost approximately $50,000 today. This number is probability times magnitude, a.k.a. insurance pricing theory 101. Ok, these do ignore profit loadings and some other administrative costs. So in practice the policy premium might be anywhere from $40,000 to $60,000 depending on business considerations that have less to do with your death statistics and more to do with competitive pressures and economic conditions. In fact, the empirical range of insurance prices can be as wide as plus or minus 20% of purely mathematical calculations.
Now imagine that you have taken a novel test – be it genetic or other — that indicates your risk factor for one of those ten diseases is in fact double, that is a 1% mortality rate versus the baseline 0.5% for the rest of the population. Let’s further and depressingly assume that there is absolutely nothing you can do about that 1%. It’s virtually incurable in the sense that no action you take today can improve those odds tomorrow. You are now stuck with double the mortality rate from (say) the ACE disease.
Now, double the risk of dying is quite scary, but remember that this is double the risk of dying from one of ten possible causes. Alas, the grim reaper has many cards up his sleeve or tools at his disposal and there is no guarantee he will draw aces for you.
Your total mortality rate – which is really all that matters for life and death – is now 5.5% instead of 5% if we merely add them all together. Using the basic insurance pricing logic, the life insurance premium you should be paying to account for your true risk should now be $55,000 which is indeed 10% higher. However, it also might fall within the wide business pricing margin noted above. Again, your odds of dying from ACE have doubled, but that doesn’t mean you should be paying anything near double for your life insurance premium. Arguably, it might only have a negligible impact on the actual premium. You certainly won’t bankrupt the company or fellow policyholders if you keep the results of this “double the risk” test secret. To put it crudely, double the risk of something is not double the risk of everything. And, this assumes all ten risk factors are uncorrelated and additive…
Alas, here is where Professor Kleinow and his colleagues’ argument gets really interesting. Remember that in the above noted toy example there are nine other factors that might kill you, each with their own 0.5% mortality rate. Now evidently, and here is the light bulb moment, those conscientious consumers who get themselves tested and who learn they are more susceptible to dying from ACE disease go on to medically monitor themselves much more frequently. They take better care of their overall health and embark on preventative medical interventions, even if futile in combating ACE disease.
Guess what that leads to…
Now, although I was clear that ACE might not be curable and all this added medical attention might not alter the inviolable 1% (versus 0.5%) chances, there is a corollary and very positive side effect from all these regular doctor visits. Namely, the mortality rate from the other 9 factors – some of which are malleable – tends to decline. Yes, those nine other factor numbers might drop from 0.5% to (say) only 0.4%. Think about it in slightly different terms. You just got a traffic ticket for running a stop sign, which then improves your overall driving on many other dimensions (at least for a while.) Your chances of speeding, running a red light, and parking illegally all decline (at least for a while.)
Ok. What are we left with? You now have a total mortality rate of 1% for the incurable ACE disease plus 9 other factors times 0.4% = 3.6%, leaving you with a total or aggregate morality rate of 4.6% versus the 5% for the population. What does this mean in English? Yes, you have a gene mutation that makes it more likely you will die from the ACE disease – information you might be hiding from the insurance company – but your overall mortality rate drops to 4.6%. You are safer for them. The insurance premium you should pay is actually $46,000, not $50,000. That is $4,000 lower than what you are being charged absent any of this information. You certainly are not cheating or deceiving the insurance company by not disclosing the test results. In fact, you are saving them money and subsidizing the ignorant masses. They should thank you.
The implications of the research by Professors Kleinow and co-authors Hacariz and Macdonald, goes far beyond life insurance premium setting and perhaps gets to the heart of the question “Do I want to know about things I can’t do anything about?” The answer – to me at least – is yes. Not only might I learn as a by-product of the testing about factors on which I do have some control, the process of monitoring the progress of the incurable ACE disease might have its own therapeutical effects.
More tests, anyone?
M.A.M.