Matthew Holt writes:

The argument I want to have is a theoretical one about what would happen if we had essentially a completely personalized account-based system, as he advocates in his Large HSA proposal.

Holt raises important questions about what would happen under a system of large HSAs, where workers would get a large but limited tax break for cash that they (and/or their employer) deposit in an HSA -- tax-free cash that workers could use for health savings, spending, or insurance as they wish.

Holt's first concern is that "a significant number of people would take the money and buy no or minimal insurance coverage." That some would choose to drop health insurance is certainly a possibility. I have two responses. First, that is already an option. People can and do choose to "go bare" and use their money for savings or other spending. The large HSA approach could marginally increase the number of people who do that, but only if the newly "bare" actually put money aside for future medical expenses. (Actually, large HSAs could even encourage today's non-saving uninsured to start saving for their health expenses.) That brings me to my second response. If large HSAs do increase the number of people who "go bare," the only people they would add to the ranks of the uninsured would be savers. As those "health savers" build up large balances in their large HSAs, it will occur to them, "Gee, one serious illness could wipe out all the money I've got stashed in my HSA." How do people typically protect their assets from such unforseen losses? Insurance. So there's a built-in incentive for health savers to purchase insurance.

Holt's second concern goes like this: Were we to allow people to take all of their health benefits in the form of a cash contribution into a large HSA, and let them choose how to allocate those funds (among savings, spending, and insurance), that would begin a process known as "risk segmentation." As I describe in my paper, some people would "go bare," many would purchase less comprehensive health coverage, and many would migrate to the individual insurance market, where their premiums (typically) would be based on their individual health risk. What concerns Holt is that sicker people would have to pay more for health insurance, to the point where many sick people could not afford it.

My response is not that sick people should not be subsidized.  (I would prefer that they not be subsidized by government, but let's assume that all options are open.) It is that sick people should not be subsidized through the vehicle of "insurance." Attempting to deliver such subsidies through "insurance" destroys much of the good that insurance markets accomplish. Insurance premiums cease to deliver price signals about the costs of bad behaviors (e.g., smoking, obesity, waiting until you're sick before you buy insurance). Many consumers drop insurance rather than pay the higher-than-necessary premiums, which increases the number of uninsured and tempts government to force people to buy insurance. Most importantly, when patients are spending someone else's money, we lose a very important ally in the fight to curb wasteful medical expenditures: the patient. Instead of nagging providers about delivering value for the dollar, patients -- especially the high-cost ones -- line up with providers on the side of more spending. 

My preference is to let insurance markets do all they can do to improve efficiency, particularly by encouraging patients to pay directly more often. Some people will still require assistance, though with a more efficient health care sector their numbers should be smaller. We should subsidize those who remain directly, with cash.

I'm not sure how much of this Holt will find persuasive. Given that we agree that providers are riding the gravy train, I would think that having millions of patients nagging providers about value would hold some appeal.

He and I agree on something else. HSA supporters too often ignore these issues.