Summary:
It can make sense for life insurers to offer a new UL policy that also has a cancer treatment benefit.

Hypothesis: Life Insurers Should Pay for Costly Cancer Drugs

I came across an interesting op-ed recently in the Wall Street Journal, written by two professors at USC. Their argument is that life insurers should pay for breakthrough costly cancer treatments because the gain in mortality will offset the cost of the drug.

Their calculation compares the NPV of a much later death benefit payment from a cure of the patient’s cancer, in addition to the continuation of annual premium, paid against the cost of the drug treatment. They conclude that in many cases it would be NPV positive for the life insurer to pay for the treatment, not to mention the social good of saving a life.

The argument is compelling on the surface but after giving it some additional thought I see several potential flaws in the premise.

Conclusion: A Good Idea, but Flawed

Let’s start with the most obvious problem. The professors suggest that the insured and insurer have aligned interests because the patient surviving improves the profits of the insurer. While that is true if the treatment is free, it may or may not be true when the drug is costly and the life insurer is paying for it.

An Impossible Position

They show in the article that paying for treatment can be NPV + for the insurer, but it will not always be. What does the insurer do when it is NPV – to pay for treatment? Pay anyway? Not pay and be accused of running a “death panel“? Do we really want life insurers deciding who lives or dies based on whether they bought a large enough policy to justify treatment?

This is an untenable situation and, in and of itself, is enough to end the thought experiment. But out of my sense of completeness, I will also address some other issues that are problematic.

Cost Shifting Issues

Those of you who are regular readers may recall my earlier post on one of my favorite topics, stationarity. Injecting life insurers into the health insurance process certainly qualifies as making the historical data non-stationary relative to the new environment.

Very simply, if today a health insurer covers say $75K of a $100K treatment, tomorrow, once they realize life insurers may pay the remainder, they will decide to only pay $65K, or even $50K. Health insurers will adjust their bargaining position to reflect the new subsidy available. This is no different from how universities increase the price of tuition when the government increases financial aid (just think what college will cost if it becomes “free”).

The end result is the NPV benefit to insurers will shrink to a barely tolerable amount, if that. By “if that”, I mean that if we play things forward, once life insurers start paying for cancer drugs, it will be hard to take the benefit away. Thus, it is likely health insurers will drive the co-pay to the point where it is NPV negative most of the time.

Hyperinflation Risk

Another issue is addressing the unknown. What if cancer drugs become perfect cures and the costs goes into the millions? Life insurers can’t leave themselves exposed to paying for these breakthroughs. This can be addressed through policy language which limits the $ amount of payment but we’ve seen insurers get into trouble not thinking about what could go wrong in the future so caution would be required.

Lapse Risk

Next, there is the persistency issue. After you’ve paid for my cure, will I continue to pay my life insurance premium? Who knows? But remember, the NPV calculation assumes that I do. If I decide to lapse, the math changes.

Pandora’s Box

Finally, there is the Pandora’s Box risk. What prevents this idea from spreading beyond cancer? Will lobbies for other diseases start pressuring life insurers to pay for their treatments? Is paying for an insured’s statins to reduce heart attacks, NPV positive? Quite likely, at least for those who show high cholesterol on their policy application exam. Should life insurers pay for transplant surgeries, particularly for young people who would have 50+ years of additional life expectancy? Where does it end?

An Alternative: A Hybrid Policy

OK, so I’ve told you all the reasons this noble idea can’t work. I could just end things here or I can try to write a happy ending by coming up with a clever fix. I think I may have found the latter!

If you follow the life insurance industry closely, then you’re likely aware of how life insurers have tried to address the open-ended risks in long term care policies. They came up with a new product, the life-LTC hybrid.

A hybrid policy is essentially an either/or approach. You buy a universal life policy but it has the option to forgo some or all of the death benefit to pay for long term care needs. This solves two problems: first, it caps the liability of the insurer vs. traditional long term care policies. Second, it removes the buyer’s objection that they are spending a lot of money on a product they may never use.

Could we devise a similar product to address cancer treatment within a life policy? I think we can!

The Benefit

I envision this being a fairly straightforward design. You start with a basic universal life policy. You add a rider that allows part of the death benefit to be drawn down to pay for qualifying cancer treatments.

To put some numbers around it, if you have a $1M death benefit and your cancer drug costs $50K, you can have your life insurance pay for your treatment, but your future death benefit will only be $950K. I will introduce some bells & whistles below but that is the base product.

Note, the major improvement here is I’ve eliminated the “death panel” issue. It is not up to the insurer to judge whether the treatment is a good IRR for them. It is up to the policyholder to decide if they want to borrow from their death benefit to pay for treatment. The patient is now empowered rather than captive to an actuary’s decision.

The Premium

My original thought was the cancer rider would require additional premium like the LTC hybrids. However, after thinking it over, I’m not so sure (I’ll welcome any actuarial input).

I think this rider can be offered for “free” because, as the USC profs suggest, the improvement in mortality over the population is an NPV +. Additionally, it’s not really free. It’s being paid for in the form of lower death benefit.

In other words, if the patient dies after treatment, you are still paying them the same $1M you would have anyway. You just changed the timing slightly. In the cases where the patient survives, you are paying the DB much later and receiving additional annual premium in future years.

It’s also worth mentioning there is a powerful marketing angle to not charging upfront for the rider. In an age where many people feel they no longer need life insurance, offering “free cancer treatment” can open the market up to new buyers.

Other Features

The two remaining problems are lapse risk and inflation of cancer costs. I think I can address both with some clever policy features.

The inflation/innovation risk is fairly easy. There just needs to be a sub-limit on the available cancer benefit relative to the death benefit. Perhaps 10% of the death benefit is available for cancer treatment costs. If innovation drives drug prices meaningfully higher than that, the insurers can always respond with a new product that appropriately prices for the higher risk.

The lapse risk remains in my base version. Insureds may only be a few years into a policy and have just received a benefit worth multiples of what they have paid in. It could make sense to lapse and buy a new policy elsewhere with a full death benefit again.

The easiest way to deal with this would be to require policyholders to agree to keep their insurance in-force for a certain period of time post claiming a cancer payment. I don’t know if there are regulatory hurdles to this though (let me know in the comments if there are). Maybe a softer version would require the policy to remain in-force until the policyholder is in remission.

Regardless, I have an idea that feels less compelled and would encourage voluntary persistence. You can earn your full DB back over time. Think of how auto insurance allows the policyholder to earn a deductible waiver if they remain accident free for a period of time. You can have the insured earn back some of their death benefit each year when they pay their annual premium (amount can be derived actuarially).

Better yet, they can get some DB restored each year they remain cancer free. Envision sending policy statements that say “Good News! You’ve been cancer free for five years so we’ve restored your full policy benefit”.

Problem Solved?

I believe I’ve answered all the objections I raised to the original proposal:

1. Insureds are empowered to choose whether to pay for their treatment with their life insurance rather than be dictated to by a death panel.
2. Health insurers can’t abuse their market position because consumers will still care how much they are paying since it impacts their death benefit.
3. Drug cost inflation is now largely borne by the patient and can otherwise be addressed by sub-limit.
4. Lapses post treatment can be reduced by restoring the death benefit over time.
5. While I didn’t directly address the Pandora’s Box element, if other patient groups want a hybrid policy, insurers should be able to create one.

Conclusion

Life insurers should offer hybrid UL-cancer drug policies as it will improve profits and increase demand!

Do you like the idea? Did I miss something that makes this not viable? Feedback is welcome either in the comments below or offline.