EBRI Gets It Wrong — Again
The Employee Benefit Research Institute has put out another study of Consumer Driven Health Plans (CDH plans). They do this every six months or so, and every time they illustrate the limits of internet surveys. Yet compared to two other surveys, EBRI seems to be underestimating the number and size of CDH plans by a considerable margin.
In this case, EBRI tries to estimate the number of HRA and HSA accounts, the balances in the accounts, and the amount of money rolled-over from one year to the next. They began with a random sample of 2,007 from an internet list of people who have agreed to be surveyed, of which 21% responded — 94 people with a CD Health plan. So they went back and generated (oversampled) another 879 people with a CDH plan.
The whole methodology is suspect. Sampling can work well when you have a pretty good idea of the universe being tested. A manufacturer who produces 100,000 widgets might randomly sample 100 of them to test for quality. If 4 of the sample is defective, we can pretty well know that 4,000 of the universe is defective. But if you don’t know the universe, you can’t estimate how many defective widgets exist by testing a handful. Your sample may tell you that 4% are bad, but 4% of what? One million, ten million, one hundred million?
In this case, you may find that in your sample the average HSA has an account balance of $5,000, but it is impossible to know what the total amount of money is held in HSAs based on that. You also need to know how many HSAs exist, and EBRI has never been very good at that.
Another problem is that the respondents are self-selected. Who is more likely to respond to an internet survey of account holders? People who are very happy with their HSA or people who are very unhappy? We simply don’t know.
This is not just a theoretical concern. In this case, EBRI was very wide of the mark compared to two other similar reports that have come out about the same time. One was done by William Boyles, publisher of the Consumer Driven Market Report, and the other by Eric Remjeske of the Devenir investment firm. In both cases, they surveyed actual banks that manage HSA accounts.
This is not an enormous undertaking, there are a couple of dozen facilities that have 70% or more of the market, and if there is one thing banks are good at it is counting money. Anyone who would like to know how much money is held in HSA accounts would be foolish not to start with the account managers.
So, what do we find in these reports? That EBRI is way underestimating the number and size of CDH plans.
In viewing the chart below, keep in mind that EBRI is combining both HRA and HSA accounts, while CDMR and Devenir are looking solely at HSAs. (HRAs are notional accounts. There is no cash money held anywhere.)
EBRI always gets it wrong on consumer directed health care.
EBRI’s never had a good word to say about empowering people — whether in health care, or any other area. They believe in bureaucracies. That may be why they are in Washington.
Agree with Bruce and Tom. EBRI has always been on the other side.
Here’s what I don’t understand. EBRI is supposed to be representing industry. Yet they sound like they are a subdivision of a government agency.
EBRI has never had a good word to say about HSAs, HRAs, IRAs, Roth IRAs, or 401(k)s. On the other hand, they are a big fan of Ponzi schemes — Social Security, Medicare, and unfunded employer pension and post-retirement health care promises.
Blocking ctmoenms is not cool.Both car insurance and health insurance are real insurance ; they’re a hedge against the risk of unexpected losses. Health insurance covers relatively few expected events.The problem with health insurance is that it removes the consumer’s incentive to conduct a cost-benefit analysis.I’ve got high cholesterol. I should lose weight, exercise, and eat more fiber. If I do those things, I probably won’t need to take medicine. I’m lazy, though; part of the reason is probably that I know that if I end up taking medicine, it’ll only cost me $25/month. If I had to pay for the Lipitor myself (god only knows what it costs), I’d probably take care of myself.There are lots of other examples. Doctors test for conditions even when the risk is low because the marginal cost to the consumer is effectively $0, and the doctor gets a modest benefit if not a cut of the money for the test, then at least a little more protection against a malpractice suit.So health insurance at least contributes to spiraling medical costs. How do you discipline those costs?HSAs obviously have their own problems. Our medical system doesn’t provide transparent pricing information; there would be a lot of waste during the transition.It would have a bad effect on wealth distribution. Not just financially, but we’d see more poor people with significant, untreated medical problems. Probably a higher mortality rate for the poor, particularly babies born to poor families. (Here, poor means those who wouldn’t qualify for Medicaid but who would have little disposable income for medical treatment.)Eventually, employees who get employer-subsidized health care now would have some or all of the savings returned to them in the form of higher salary. I’m not sure they (we) would be worse off in the long run.I don’t know that HSAs are fair or practical. I do believe though that if we don’t figure out some mechanism for getting consumers to make cost/benefit decisions, the HMOs will eventually do it for them (more than they already do).