As the premier health policy blog on the net, I am honored that Matthew has asked me to occasionally guest post here. I also am regularly impressed with
  the level of thought and knowledge of so many posters of comments.
  Every once in a while, I think it is important to review, what I would believe
  so basic myths about our healthcare system. One of the greatest myths is the
  3% overhead myth for medicare. (This is often also cited as the 2% overhead
  myth.)
  Argument one in favor of the ‘medicare for all’ expansion of government
  regulated healthcare is the disparity in overhead between Medicare and private
  insurance plans. The Medicare overhead number is stated as 2% or 3%, depending
  upon where you read it. The overhead for private insurers is stated as 15%-25%.
  The disparity is used to explain the statement that, “if we just take the
  amount spent on wasted overhead and apply it to medicare-for-all, we could easily
  pay for everyone to be covered”. Where does that math come from? Here’s
  the brief skinny:
  $2 trillion in healthcare. 50% from private sector= $1trillion. Reduce overhead
  by 15-20%= $150- $200 billion available.
  Oh, if it was so simple.
  Let’s attack the 3% myth.

Proponents say overhead should be calculated as: admin costs/ payout for services.
  However, medicare recipients use much more care, on average, than younger groups.
  So, for example (made up numbers), medicare recipients might use $5000 per year,
  while commercially insured people might use $3000 per year. If both groups consisted
  of 10,000 people, it would take the same amount of oversight, management, etc.,
  yet the perceived overhead to take care of the younger group would be much higher;
  or put the other way, medicare overhead would be much lower.Solution: calculate overhead on a ‘per enrollee’ scale. This alone
  accounts for 50% of the discrepancy in overhead between medicare and private
  payers. If you do not believe me, just check with the Kaiser Foundation research
  saying the same thing.
  Proponents focus on the low administrative costs, on the one hand, while denouncing
  the amount of fraud by hospitals, providers. This perhaps is because medicare  does not spend enough on administrative oversight of the program. This happens to be exactly what the GAO and the National Academy of Social Insurance has  said within the past 6 years. So, if medicare spent more on administration,
  the discrepancy would be decreased even further.
  Proponents fail to account that for every regulation, costs are incurred by
  providers to comply. The 100,000 plus pages of medicare regulations function
  as an unfunded mandate on providers. Currently, the coding is based on ICD-9,
  which has over 24,000 codes. ICD-10, slated to go into effect within the next
  2 years has over 207,000!!!! And does upgrading for the change count toward medicare overhead? Of course not.
  Proponents say that private insurance company rules would create as much hassle
  for providers. Perhaps, but we will never know since private insurance rules
  are based off of the medicare guidelines.
  Thus, the ‘medicare 3% myth’, is, in reality, just that: a myth.

But like most myths, true believers will never pay attention to facts.

Michael Cannon has written a response to my response to him. Even ignoring the issue about my personal HSA, we’re really talking past each other. Cannon doesn’t think our discussion is fruitful, and in truth it’s not. He wants to discuss the vast majority of his paper which looks at the role of HSAs within our current system. To me our current system is so broken, the introduction of HSAs (at least in the limited form they now exist which is all we’re likely to get for now) is pretty irrelevant, and a minor incremental change—albeit one away from the compulsory social insurance that, he correctly states, I advocate. Frankly in the next five years neither of us is going to get our way…so this argument is about what comes next.

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. As I explained at length before, I think that a significant number of people would take the money and by no or minimal insurance coverage. So apparently does he.

Large HSAs would give workers far greater freedom of choice. Workers could use their HSA funds (and non-HSA funds) to purchase insurance from their employer or any other source. Alternatively, they could forgo insurance to build larger HSA balances.

Now lets just assume that over say 20 years people really do build up huge HSA balances, and so when they need the money for their individual health crisis in year 20, they can pay for it all themselves. Even accepting that this would happen and that young healthies (His “students”) could buy a cheap heavily underwritten very high deductible policy for the early years, my question is what would happen in Year One? The money that would cover the sick in a compulsory social insurance pool, would have been extracted and instead be sitting in the personal accounts of the “students”. So when the sick start incurring huge health care costs, the money to pay them must come from somewhere. Unless the people who get sick had already saved up the huge amount they need or were allowed to buy into the cheap underwritten catastrophic plans, (both of which are totally unrealistic and the latter of which would destroy those plans as a profitable business), then that money must come from the taxpayer, or the providers (in the form of non-payment for services rendered).

This is the problem that I just don’t understand about the individual account theory. This is after all about the crux of insurance, which Cannon believes can work in a voluntary, HSA-based system. I just wish someone promoting those accounts would explain why I don’t understand how they overcome that issue rather than continually ignoring it.

Michael Cannon comments on my post about his paper yesterday, noting in passing that I have an HSA. C’mon Michael you can understand that people will take advantage of incentives, even though the policy behind those incentives is bone-headed, can’t you? After all like most of your colleagues at Cato I think that getting tax relief on my mortgage is bad policy, I think that paying taxes to support the war on drugs is terrible policy. But no one exactly gave me the choice…

But onto the real discussion. In his blog Cannon says I didn’t read his piece carefully enough. Actually frankly I’m not very interested in the attempt to figure out how HSAs fit into our current broken system which occupies most of the piece, and I despair of any of their supporters taking them very seriously. They all say that they’re “partial solutions”, or “incremental”. Frankly our care system is so screwed up that whether we force more problems on the sick in their decisions about accessing care (which Cannon agrees that HSAs/HDHPs might do) is pretty irrelevant when we have 15% of the population who’d love to have that problem.

What I like about Cannon (and Tanner and Kling) is that they’re among the very, very few on their end of the spectrum who’ll have a theoretical argument about the insurance “market”. So let’s get to our core “mis”understanding

Also, Holt accuses me of ignoring the fact that risk segmentation results in reduced subsidies to the sickest insureds. Yet that is a central theme of the “students & professors” hypothetical (pp. 6-8).

I don’t accuse him of ignoring the reduction in subsidies! I accuse him of both understanding that it happens and believing that it’s a good thing! And the conclusion to that hypothetical piece is

Though the professors would lose the cross-subsidies they received under Plan A,those losses would essentially be temporary transition costs. The higher health insurance premiums for today’s professors would convey to today’s students the importance of saving for their future medical needs. Thus tomorrow’s professors would face greater incentives to save for their future medical needs. Because their current premiums would be lower, they would be better equipped to do so.

In other words, the market would send a signal to the “students” that the if they didn’t avoid having any health care costs in the future, and hadn’t saved all their lives to finance them, they’d be lying bankrupt in the gutter with “professors” who also haven’t saved enough to afford the costs they’re paying for the care they need now. This is a “transition” cost, and Cannon and several of his colleagues believe that a) we really can get to a place where individuals accounts saved for over the years can cover all health expenses, and therefore insurance (with its implied social cross subsidy) is unnecessary, and b) the transition costs are small. Given the current savings habits of Americans  the first assumption is laughable, but it’s the next point that’s the real problem.

If you go to the logical extreme and do away with insurance, a) those transition costs are huge and b) the “students” who get sick will not be able to save enough over their lifetimes to deal with their future costs. The problem remains the 80/20 rule. If you allow the 80% to put all their money in an individual account and not in the social pool there will not be enough money to pay for the care of the few who need it—even the ones who’ve scrimped and saved all their lives.

But don’t fear Cannon has a solution for that. After we’ve eliminated the cross-subsidy of social insurance, we somehow or other bring it back

And on page eight I write:

Though HSAs may reduce hidden subsidies to sicker workers, they do not preclude subsidizing those workers in other ways.

Strangely he didn’t include the very next sentence

Other options include government subsidies or private charity, including assistance from family and friends, churches, civic associations, and uncompensated care from hospitals and doctors.

Which if I’m not very much mistaken is what we’ve got already and what the providers and employers are bleating about at the moment. Cannon just thinks that we should be pushing policies that will make the current zoo worse, and return more money to the healthy people who don’t need it.

His justification for all of this (which he continually says is “socially desirable”) is that putting people into HDHPs will reduce their spending overall and drive out that darned unnecessary care they’re all demanding. But as apparently although he will admit it he doesn’t want to consider that most health care spending is not under the control of a patient spending their own money, even if they have an HSA/HDHP. The stuff that costs the most money is the flat-of-the-curve medicine being visited on the nearly dead. And Cannon apparently has no interest in figuring out how to reduce that because it requires a supply-constraint. To be fair to him, not many other people want to do that either, as it means beating up on a bunch of doctors and hospitals. But other countries manage it!

So for the nth time, if you want to have a rational, fair and cost-efficient health care market you need compulsory social insurance, hopefully progressively based, so that those people who end up with large healthcare costs don’t end up being bankrupted. Then you need incentives for providers that induce them to provide cost-efficient care over a population, rather than to do as much as possible to those who can pay, and ignore the rest—which is the recipe for driving up costs. Cannon’s analysis suggest that he knows this, but his solutions drive us towards the opposite state, which is why I’m wondering about the color of his planet’s sky.

Kaiser Family foundation has an excellent slideshow introduction about the CDHP—how it works, what it’s supposed to do, who pays, etc, etc. Of course all you need to know is in the line below titled “Total Firm Contribution” as to why employers might think they like the HDHP.