Yesterday’s post, A Dog in the Healthcare Fight, got good follow-up in a number of places, including Greg Mankiw, Megan McArdle, Arnold Kling, Tyler Cowen, and the American Scene’s John Schwenkler. A couple comments seem worth addressing.
The Atlantic’s McArdle generously says:
Kudoes to AEI for publishing a graph that seriously undercuts one of the major conservative arguments about healthcare: that the main problem is consumers who don’t bear their own costs. Veterinary spending is subject to few of the perversities that either left or right suppose to be the main problems afflicting healthcare spending. Consumers pay full freight most of the time. They are price sensitive, and will let the patient die if keeping him alive costs too much. There is no adverse selection. There is no free riding on mandatory care. Government regulation is minimal. Malpractice suits are minimal, and have low payouts. So why is vet spending rising along with human spending?
While I like to think I’m willing to challenge my own prior predispositions—and, in the case of Social Security privatization, have sometimes done so—Megan misses the distinction between the level of spending and the rate of growth of spending, which at least partially derive from different factors. (Bryan Caplan seems also to have hit on this point.)
As I pointed out, the RAND healthcare experiment showed that health spending is very sensitive to the share of costs borne out of pocket. As the out-of-pocket expenses have fallen from over half of health spending in 1960 to around 13 percent today, we’ve gone from an á la carte approach to healthcare to an all-you-can-eat approach: from low premiums and high co-pays to high premiums and low co-pays. As anyone who has been to an all-you-can-eat restaurant can attest, you can eat a lot if you pay upfront.
Moreover, the tax exclusion for employer-provided health coverage implies that a dollar of healthcare will be a lot cheaper for employers than a dollar of other compensation. If, say, the average combined payroll and income tax rate by a firm’s employees is 35 percent, this implies that you can buy health coverage at around 65 cents on the dollar—or, more ominously, you’ll continue to buy healthcare until each dollar spent brings you only 65 cents in value. Almost by definition, that’s wasteful. (And by the way, if taxes rise, incentives to overspend on healthcare will only increase.)
Importantly, however, neither of these factors really alters the growth rate of health spending. Assuming the out-of-pocket-share and tax rates stay the same, these factors affect only the level, not the growth of health spending. If we raised out-of-pocket costs and repealed the tax exclusion for healthcare, we could expect significantly lower levels of spending—yet those levels would continue to grow at much the same rate as before.
There are obviously other factors involved—importantly, rising incomes—but the key distinction here is that the reasons we spend so much today are distinct from the reasons our spending is rising. Moreover, I’d argue that the key elements of healthcare waste are in high spending today, not the factors that will raise spending over time.
John Schwenkler calls the pets chart “potentially misleading” because it doesn’t account for changes in the number of pets. I’d actually go further, since it also doesn’t account for changes in the age structure of the pet population (as I’ve argued elsewhere, aging is the key driver of overall entitlement costs for people) nor does it account for changes in the composition of the pet population (say, pet ferrets may costs less to care for than dogs, but other exotic pets might cost more). So this is clearly a less precise measure of health spending increases than are done for people, but designed to get at the same qualitative point: that health spending would rise even in a totally free market, so we should concentrate our policy attention on factors that affect the level of spending—in particular, factors that cause wasteful spending—and not fixate too strongly on the growth rate.