Friday, March 31, 2017

Health Insurance Should Be More Like Auto and Life Insurance

Health insurance would be a hell of a lot more affordable and the market would be a lot more competitive if its policies had some features that are common to auto and life insurance.

Longer Policy Terms

I have a life insurance policy whose term is measured in decades. It will pay my wife about three years’ worth of my annual salary if I die during that term. They cannot cancel the policy or change my premiums if my health status changes and the likelihood of a payout suddenly increases dramatically. My premium is locked in for the term. This works fine, because the premium already has integrated into it the chance that I will get sicker and die at some point during the policy term. Since my policy is priced based on my all-cause mortality risk at the beginning of the term, if I suddenly get cancer my insurer says, “It’s fine, I planned for this. You’re covered, buddy!” They might secretly wish I would cancel my policy or forget to collect, but the only reason they have customers in the first place is that they actually pay out as promised when their policyholders die. They certainly don’t want a reputation of weaseling out of their contractual obligations.

We’re talking about an insurance policy here with a huge payout (six figures), a very long policy term, and premiums at just over $20 per pay period. In other words, it costs me just over $500 annually, easily affordable for a middle class family. A catastrophic health plan that does the same would probably have a higher frequency (major health episodes are considerably more frequent than actual deaths) but lower severity (the average surgery is probably closer to four or five figures, not six). The premium, which is based on the expected cost (frequency of a claim x average severity of a claim, plus operational expenses), could be higher or lower than for a life policy, but should be in a similar range. The example of life insurance suggests that catastrophic health insurance with long policy terms should be affordable. Such insurance policies would not be budget-straining disasters like existing health “insurance” policies, which cover a lot of elective and routine expenses.

Liabilities Don’t Transfer To A New Insurance Policy

If I maim someone with my car today, my current auto insurance policy owns that liability. If the guy I maimed has a series of surgeries related to his injuries for the next ten years, my insurer still owns that liability forever (at least until his expenses exceeds the limits of liability on my auto policy). If I shop around next year and switch insurance companies, that liability does not follow me. My new policy will cover future maimings that occur during the policy period, but not past maimings. I will be charged slightly more because I now have a bad driving history, which is predictive of future accidents. But it will be nowhere near the cost of paying my victim’s medical bills. This is how health insurance should work. It’s absolutely crazy that it doesn’t actually work this way. If I get a diagnosis for a chronic disease or a specialist decides I need an expensive surgery, the liability for the related costs follow me if I switch to anther insurer next year. If people are dragging their existing liabilities around with them, insurers will be hesitant about insuring them and try to find any reason to say “No.” The adverse selection problem looms large in the health insurance market not because of some inherent market failure unique to health insurance, but because liabilities that have already been incurred follow people to their new insurer.

If you have very long-term and hard-to-cancel policies, this should rarely be a problem. Changing health insurers should be a once-or-twice-in-a-lifetime event, not an annual shopping experience. But when it does happen, your new insurer is only insuring for new changes to your health status, unrelated to past diagnoses. Your previous insurer owns the liability for every expense already incurred.*

Once again, this kind of policy is affordable for most households. I pay around $8 per pay-period for the medical liability portion of my auto insurance (slightly more for the other coverages I get), just shy of $200 per year.  A catastrophic health insurance policy would likely be higher in claims frequency, so the premium would surely be higher than this. But so what? Conservatively multiply by ten and you still get something that’s quite a bit more affordable than the average existing health policy.

I discuss this topic in greater detail here.  

Deductibles and Coverage Exclusions

Health “insurance” policies have been trending toward higher deductibles in recent years, so this is something that’s actually working right at the moment. A deductible is basically an announcement to your insurer that “I’m not going to bug you for the small stuff. I’m only going to use my policy if something big and financially destructive happens.” Coverage exclusions are another way of doing this, but unfortunately these are explicitly illegal in the health insurance market. You can’t buy a policy that excludes, say, routine doctors checkups, birth and neonatal care, or inexpensive medications (although a high enough deductible can act as a coverage exclusion for some of these). There are laws that mandate “You must buy this coverage if you buy health insurance; you may not opt out.” You should be able to buy a plan that excludes everything that is routine, low-cost, or elective, but still promises to pay for large bank-breaking health events. If voters want to subsidize childbirth or prescription purchases or some other form of non-catastrophic health care, they should do it on budget with actual tax dollars and an explicit government program. They shouldn’t do it off-budget in a round-about way by hobbling health insurance markets.**

Residual Markets

Auto insurance premiums are based on the risk being insured. If you are a young male with a DUI and several recent accidents, you will pay a great deal more than a middle aged woman with a great driving record. Risk is quantified and priced. There are a few people for whom the expected cost of an auto policy is unaffordable. A bad combination of factors can put annual premiums in the $5,000+ range. We want these people to carry liability insurance, because they are a serious risk to other drivers. But with premiums so high, we run the risk that some of them will drive without insurance and cause uninsured accidents. There are usually residual market mechanisms to handle these high-risk drivers. They are a small segment of the population. It depends on exactly how you count them, but at the very most it’s in the 10% of population range. Auto insurance is regulated separately by each state, and each state has its own mechanism for these very bad risks. Sometimes a risk is just assigned to a random insurer (assignment is based on market share) and the state says, “Charge him exactly this rate, and eat the loss if he has a claim.” Sometimes a state creates a special risk pool, and the profits or losses from the pool are shared based on market share or some special formula. The state of Maryland runs its own auto insurance company, which serves a lot of high-risk insureds. Here we have the hand of government, but it is a small addition to a private market with competitive prices and sophisticated risk-pricing. That would work just as well for a market in health insurance: allow a humming free market that will work for 90% of the population along with a tiny bit of government assistance for the very few who aren’t served well. It could involve assigning risks to random insurers, subsidizing the purchase of insurance through a general fund, creating special subsidized risk pools, or something I haven’t even thought about. There are many tools to work with here. We don’t have to hobble 90% of the private market to serve those rare exceptional people who it doesn’t work for.

Health Insurance Isn’t Special

I’m writing this post to make the point that health insurance is fixable, and the policy features that would fix it are already out there in the insurance marketplace. We already have insurance policies with extremely long terms and very large six- and seven-figure payouts. We already have insurance policies that cover medical expenses based on when the liability is incurred, rather than when the bills are paid. And it’s all pretty affordable. Both the auto and life insurance markets involve sophisticated risk-pricing. There are residual markets for extremely high risks with unaffordable premiums. We wouldn’t have the death-spirals we see in today’s health insurance pools, where adverse selection drives premiums through the roof as all the good risks leave. We wouldn’t have the issue of health insurers not wanting you as a customer. Risk pricing solves these problems.

* In property and casualty insurance, we sometimes talk about “incurred” vs “paid” claims or “accident year” vs “calendar year.” The idea is that we have a concept for when the liability arises, such as an automobile accident or a house fire; we have a separate concept for describing when the payments are made for that liability. For example, an injury was incurred in December 2010, but payments were made to cover the injured party’s surgeries in 2012, 2013, and 2015. The liability was “incurred” in “accident year” 2010. But claims were “paid” in “calendar years” 2012, 2013, and 2015. The insurer will cover those expenses regardless of which year they are paid in, even if the at-fault driver jumps to another insurer. This is a useful concept , and discussions of healthcare policy desperately need it.

**For example, a $1,000 deductible on your auto physical damage coverage means that you will pay the first $1,000 of any claim, and the insurer will pay the remainder. For all those minor scrapes and door dings that cost less than $1,000 to repair, you just eat the cost. If you hit a large animal and your car is a total loss, your insurer will pay you the value of your car minus $1,000. This allows for a huge savings in administrative costs because the insurer isn't using its claims teams to settle piddly $50 or $100 repairs. It also gets people used to paying out of pocket for the small stuff, which is a good habit.

No comments:

Post a Comment