Is insurance even an appropriate business model for paying for health care?

This is a serious question that is almost entirely ignored in the current health insurance reform debate. The current debate has been largely concerned with answering a different question:

Can government do a better job at health insurance than private health insurance companies?

It’s possible that answering this question is important. However, that’s only true if insurance has long term viability as a means of paying for health care. I’m going to argue against that.

First, what is insurance? Insurance is a method of pooling resources to pay for low probability, uncorrelated, costly events. Low probability means that payouts don’t happen all the time. Uncorrelated means that one payout doesn’t automatically imply another payout. Costly means … costly. The low probability, high cost is fundamental to the insurance industry (it makes pooling resources worthwhile). Having events be uncorrelated is important for the statistical calculations, but not theoretically necessary: the important thing is that an insurer can estimate the expected value of payouts and bill premiums that are greater (over time) than the expected values. For uncorrelated events, the law of large numbers allows the insurers simply multiply the estimated probability times the estimated cost. For correlated events (or for smaller numbers), this becomes more difficult. Then the insurer can use the pooled resources to invest and make a profit, holding enough liquidity to make the expected payouts.

There are nontrivial differences between health insurers and any other insurer.

Let’s look at auto insurance. The first difference is the distribution of the risks. For auto insurance, they are fairly easy to assess and correct. If you are under 25 and male, you have a higher risk for auto accidents, and the insurers can charge you a higher premium to offer you coverage. If you’ve had an auto accident, you are also associated with higher risk of repeat accidents. If you get caught speeding a couple times, same thing. The statistics don’t quite work out so well for health insurers. Some of this has to do with the ability to gather data, but some of it has to do with the very nature of the risks in the industry. Your health insurer can require a physical, but they are barred by law from using certain risk factors that are statistically relevant.

A second distribution problem is the distribution of the payouts. The liability costs of auto accidents follow a very different distribution than medical costs. Most auto accidents have a predictable cost. Sure, you can hit a Ferrari, but that is statistically insignificant next to the odds of having heart disease or cancer, where the future costs can become effectively infinite, particularly if newer, expensive treatments improve survival, leading to more costs. Both insurer industries have limits to payouts. In the auto insurance industry, that doesn’t automatically mean that an auto insurance policy holder will go bankrupt as a result of every accident (sucks if you hit a Ferrari, though). For any chronic health problem (which, from a treatment perspective, cancer counts as a chronic, and not an acute problem), any limits to total payouts will lead to bankruptcy or foregone treatment. There are “fat tails” on the statistical distributions of health care costs. As an extremely relevant aside, the medical costs problem is non-linear: the existence and dominance of insurance payouts forms a positive feedback loop for costs. The distribution of medical costs mean that insurance is worthless for certain problems (the insurer limits will always be hit). This is a fundamental problem of the health insurance industry (and note, it is in no way a fundamental problem of the health care industry). There is also one fundamental difference between the two insurance industries. An auto insurer can always opt for replacement. A replacement car is simply an option against repair. The same cannot be universally said for health insurers. Sure, you might have the option to replace a kidney or use a dialysis machine (especially now that portable options exist), but there’s a distinct technological advantage to replacements in the vehicle domain. Note that there are supply/demand problems in the auto repair vs. health repair industries. It is far less costly to train a mechanic than it is to train a surgeon or general practitioner. This is a different problem, but it’s still worth considering.

The last difference I want to examine between health insurance and any other type of insurance how it is applied. People use health insurance to pay for everything. Checkups? Co-pay. Imagine if you used your auto insurance to cover oil changes. Or worse, gasoline. Maybe you’d be willing to drive a lot more if your auto insurance company picked up the tab — imagine what that would do to gasoline prices. Then imagine the resulting increase in insurance premiums.

Let’s look at a bit of history on health insurance, and speculate as to why it still exists. For this, I want to first make the claim that health insurance is unsustainable in its current form. Prices can’t rise forever on any good or service. Some people point out that the current health insurance system will make us all go bankrupt. The health insurance industry would collapse before then (they’d all become investment banks, since that’s most of what they do, anyway). But health care would march on. Doctors and drug manufacturers want to get paid. They won’t stop their goods and services just because insurance doesn’t pay them any longer. Of course, this couldn’t happen overnight. Current billing procedures at most hospitals are atrocious: nobody tells you what your visit will cost until after you’ve had the procedure. In any other industry, this would constitute contract fraud. Incidentally, in some states, vendors have to tell you what a prescription fill will cost by law. I live in such a state, and it’s actually funny how few can comply. But, despite these problems, the health care industry would survive. My greatest concern is that current health insurance reform will prolong this failed business model instead of ease the transition to a sustainable model.

The modern health insurance industry largely came into existence as a result of wage control laws after WWII. Employers wanted to attract better employees, but couldn’t offer better salaries. Instead, they identified a loophole and offered health insurance as a benefit. Seeing as the sole intent was the exact same as salary, you’d think that health insurance would be taxed as income. Of course this wasn’t the case, because our tax law has always been full of such loopholes. By the time congress realized what was happening, untaxed employee health insurance was already an entitlement. This severed any traditional economic price control on the industry. Coincidentally, I’ve been told that this also marked the period when the billing for the health care industry started going bad. This provided the necessary conditions for a price bubble, but the bubble wouldn’t actually come for a few decades. Medicare and medicaid also formed a part of this foundation, they started price increases, but not what I’d describe as the bubble in costs. Another aside: having a large pool of publicly subsidized benefits that have the property that benefits always increase without making concessions to actually pay for the benefits certainly did contribute to the present bubble. And I mean “concessions to pay” in two ways. First, recipients don’t pay directly, which has had measurable overuse. Second, most of the very recent benefit approvals haven’t even been paid by current taxpayers. We’ve started deficit spending out subsidized health care! But more important than increasing payouts without increasing taxes commensurately was probably the bull market of the last three decades, and the conversion of health insurance to investment banking. Just as separating the payment from the user hurts cost controls in health care, separating the funding agent from the risk taker hurts prices in investment banking. It’s called other people’s money, and it’s the moral hazard financial equivalent of heroin addiction. This probably warrants its own post.

The most insidious part of the reform debate is the fact that while some benefit will come at the expense of the health insurance industry, the net benefactor will not be the people receiving care (and there are well reasoned arguments that they will actually be worse off). The prime benefactor will be the drug industry. There’s a reason that the number one contributor to pro-reform advertising is big pharma. The drug companies see the writing on the wall. Increased cash flow into their industry is unsustainable with the current insurance industry. Their profits will necessarily decrease when the current unsustainable trajectory changes course. However, when the public at large pays big pharma, the end of their profit bubble moves considerably further into the future. You’d think that Medicare Section D was a large enough transfer of wealth from the US taxpayer to big pharma (remember, the future outlays from that travesty dwarfs Social Security). No, there’s some risk that huge cash cow will get sacrificed by fiscal realities. It’s not that our government wants to cut benefits, but when greater than 100% of GDP must be allocated to a single benefit, realities such as basic math may force their hand.

The current health insurance reform debate is addressing the wrong questions, when it even tries to address them. It seems that those advocating reform are more concerned with engaging the mentally insane protesters than addressing the real questions raised as to the sustainability or desirability of their plans. Any reform based on attempting a better insurance provider is doomed to fail. This is not because government can’t be a better insurer than private insurance companies, but because our current concept of health insurance is broken and doomed to fail.