Your best friend is getting married. You teach at a public school but his other friends are investment bankers. They organize a weekend bachelor party in Vegas, complete with flights and a hotel and fancy dinners and what-not. They can afford it. You can’t. You’d rather play poker and drink some beer. But you’ve got to go: he’s your best friend, and he’s only having one bachelor party.
In some under-appreciated respects, health insurance is like that bachelor party—only worse, since you’ve got to pay every month. If you earn a lot, you’re probably OK with a health plan that covers all medically necessary care, including expensive care of marginal clinical value (Vegas, baby!). If you don’t make as much, you might rather save some money and buy a plan that excludes glitzy treatments that aren’t much better than cheaper alternatives (poker and beer).
But you don’t have a choice in the health insurance market. Plans compete over cost-sharing and their networks, but they don’t compete over the scope of coverage. As Austin, Amitabh, and I explain in our Brookings paper:
For a number of historical, legal, and institutional reasons, … consumers are constrained to purchase health plans (or are enrolled in government health-care programs) that uniformly cover all medically necessary care. This “demand pooling” means that many Americans insure themselves against the risk of needing therapies they do not value.
The problem is especially acute for low-income Americans. As Hall and Jones (2007) have shown, the proportion of income that a consumer is willing to spend on health care grows more rapidly than income. A typical individual who makes $400,000 per year would thus wish to spend more than eight times as much on health care than someone who makes $50,000. Yet, instead of buying health plans that meet their variable demand, the rich and poor alike must buy plans that cover health-care technologies of questionable value.
The trouble is not mainly that the market doesn’t (and probably can’t) offer health plans of varying depths of coverage, as Austin discussed yesterday. The trouble is that uniformly expansive coverage sends a signal to technology developers that “if you build it, we will pay for it,” whether or not the innovation offers good value for the money.
Naturally, innovators respond to that signal by innovating. Most of the time, that’s really good. We want more Sovaldis to come on the market. The problem is that we’re so indiscriminate about what we pay for that innovators are indiscriminate about what they develop. We encourage the development of technologies that are 1% better but 100 times more expensive.
How can we encourage innovation that offers better bang for the buck? Austin will have more to say on that later today—or you can go read the paper now.