I note that recently this blog has been getting a more political bent, but I’m trying to avoid it with a recent flurry of role-playing/fantasy related posts. I fully intend to restrict the political analysis to a) health-related stuff, b) stuff I can apply stats to and c) very occasional commentary on really big stuff. The post below is a bit exploratory, it by no means represents my finished thoughts on the topic, so comments (particularly from those ideologically sympathetic to the HSA idea) would be welcome.

So recently I stumbled again on that old fancy among libertarians, Health Savings Accounts (HSAs) (also called “Medical Savings Accounts”), which are offered up by some free-marketeers as an alternative to current insurance systems, public or private. You can find a summary of the idea as put forward by the Cato Institute here, and an explanation at the “Institute for Health Freedom” here. Reason magazine has an amusing attempt to justify them on the basis of the experience of a minority of wealthy people in South Africa, and also an amusing article about one person’s attempt to set up a Health Savings Account for his own employees, which was stymied by government (of course). Amusingly, his opinions don’t match the WHO review of HSAs at all. That review (pdf) is available from the (surprisingly empty!) wikipedia page on health savings accounts. They appear to have been tried with varying degrees of success in 4 countries: Singapore, China, South Africa and the USA.It looks like China is pressing ahead with them.

How a Health Savings Account works

In an HSA, your employer puts a fixed amount of money (all the examples say $3000) into a savings account in your name from which you can only draw money for health care. They then also purchase you a catastrophic insurance plan, which typically has a very high deductible ($1500). In the US at time of writing these catastrophic insurance plans appear to cost $3000 or so, and this makes the total cost of the employer-provided insurance $4500, which is apparently equivalent to the average amount that employers paid at the time of writing (1993 for Cato). The whole amount (savings account and catastrophic insurance plan) is tax deductible. The same rules apply if you’re self-employed or your employer is not offering health insurance; you put your $4500 of pre-tax money into the same scheme, and it is tax deductible.

Then, when you get sick, you have this $1500 to pay for standard procedures, like visiting the doctor; but if any thing “catastrophic” happens you’re covered by the employer plan (except for the deductible, of course). Any money you don’t spend at the end of the year is rolled over into the following year’s savings account, which accumulates over time. When you are sick, you negotiate the cost directly with the doctor/hospital, etc.

Justifications for Health Savings Accounts

There seem to be three primary justifications for health savings accounts:

  • They remove the current favourable tax considerations given to employer-based health insurance
  • They offer patients choice
  • They restrict costs

I won’t talk about the second justification, since choice is obviously good. However, I want to talk a bit more about the logic of the first and third.

Better tax arrangements

Under this justification, one of the main reasons that there are 35 million uninsured Americans is that their employer doesn’t offer them health insurance, and they can’t afford to pay for it themselves because health insurance purchased individually (i.e. not through employers) is not tax deductible. Using post-tax income makes your insurance cost more, due to the loss of tax deductions. I think there are two problems with this justification:

  1. It has nothing to do with health savings accounts. You could reform these tax arrangements tomorrow without changing a thing about the way the insurance industry works, and if this argument is correct then suddenly your uninsured 35 million will vanish. I can’t see a particular reason why there should be a relationship between the tax deductibility of non-employer insurance and HSAs
  2. I don’t know that the evidence is that strong that the tax arrangements are the main reason Americans can’t afford health insurance. Health insurance is offered by employers in lieu of income, and it’s possible that employers not offering it will pay more in cash instead. In this case, although the employee has to pay more after tax for their health insurance, they also have a higher after tax income than a person in an equivalent job who gets health insurance, which makes individual insurance more affordable for these people compared to their employer-insured cohort. Furthermore, if this is not the case – i.e. those not offered insurance by their employer are not paid more – then the lack of insurance must be at least partially related to the existence of two tiers of job. If this is the case, then introducing a different system of insurance will not fix the problem, since the problem is caused by income inequality. The solution here is to make health insurance cheaper or to reduce inequality between workplaces, not to make it tax deductible.

There is a related claim that insurance costs are lower for people with HSAs, and so more people can afford to take them out, but this claim is based on a RAND study (cited in the Cato Institute paper) that compared people paying out of their own pocket with those receiving free health care. This study is obviously useless – you can’t compare those two groups of people in America. The comparison here, in any case, is between people paying out of an insurance system and people paying out of their own pocket. In fact, the WHO review found limited evidence of cost reductions in the US on this comparative basis – 12% compared to insurance, in a study with dubious methodology that covered only a very small number of non-randomly-selected employers.

So, if health insurance becomes 12% more affordable under this plan, it’s probably not going to make big inroads into the number of uninsured, and in any case the tax arrangements cited in defense of HSAs have no relationship to the HSAs themselves.

Cost containment

The other reason for using HSAs is stated as containing costs. As we know, health care costs are increasing in all developed nations, and every health care system is trying to find ways to contain costs. This is where the disingenuity starts from the libertarian sites, because of course almost every existing financing arrangement is based on some form of collective action or risk pooling, and everyone knows that libertarians hate anything like collective action. For example, here is the hideously disingenuous explanation for why health care costs are increasing in the US, from the Cato Institute paper:

Of every health care dollar spent in this country, 76 cents are paid by someone other than the actual patient–by the government, insurers, or employers. Consequently, in most situations patients neither benefit when they spend wisely nor bear the consequences of spending foolishly. With those incentives, it is no surprise that costs are soaring.

Elsewhere, they refer repeatedly to “spending someone else’s money.” Now, fair enough, that might count as a justification for rising health care costs – except that every modern financing system involves “spending someone else’s money” and health care costs are increasing faster in the US than anywhere else. In fact, the US has a much higher level of individual payment than other systems. In the UK it’s almost 100% – you almost never pay out of your own pocket, because either the NHS pays or a private insurer pays. The same is true of every other system. True, in Japan it’s possible that up to 30% of your costs are paid by you, but there is a cap on what you spend so it’s probably not the case that “only” 70 cents in the dollar are paid with “someone else’s money.” In general, however, the huge majority of health care expenditure is “someone else’s money.” It’s almost as if this fact has no relationship to increasing health care costs or, alternatively, is the cause of them.

In fact, the reason health care costs are going up is because health care gets more and more sophisticated, and as we get wealthier we demand better and better care, so we spend more. This is independent of the particular system in place, but it does appear that the more private insurance there is, the faster costs rise. Why libertarians think this is a bad thing, I don’t know. After all, in a private system increasing costs just represent consumer preference, right? Well, in fact, in health care they don’t. They represent the doctor’s decision, and the doctor will always decide the maximum treatment that the system can afford. Which brings us to the next bit of rhetoric attached to health savings accounts: patient power.

How Health Savings Accounts will lower costs

The key virtue of health savings accounts is supposed to be that they give the patient power, and that this power will lead to a reduction in costs. This is because the patient pays for non-catastrophic health care from their own money, rather than the insurance company’s or the government’s, and thus have more power to choose. So let us consider two scenarios:

  1. I am self-insured under an HSA, with a fixed limit on how much I can spend ($1500 x years enrolled), and I can choose my health care from a set of menus within this fixed price
  2. I am insured by the government under a universal provider, and I can choose any health care I want

Which of these gives the patient more power? Choice is not power when choice is constrained by finite resources. Economically it may be a good thing, but power it ain’t. A HSA only gives a patient “power” if it widens their choice relative to the existing system – which it may do given the caprice of the modern American private health system, but that is an issue that can be fixed by many mechanisms other than HSAs.”Patient power” is just another piece of disingenuous libertarian spin, because “constraining patient choice” doesn’t sound so nice.

So what is the specific mechanism by which giving the patient a choice under the HSA system lowers the cost of health care? The patient chooses their health care subject to their available budget, so presumably they shop around a bit, and choose the cheapest option.

What this system actually does is reduce total health care costs. It doesn’t reduce health care costs. Some patients – typically those earning less, or those who know they’re sitting on a condition that will be a time bomb in future, or those with a chronic disease – will spend less than other patients, and the main way they will do this is by choosing not to receive certain treatments, or choosing a treatment that isn’t as good. Let’s consider two examples.

  1. HIV-positive health care worker: let’s say this person was infected from a needle-stick injury, to avoid any moral panic. They are going to be on anti-retrovirals for the rest of their life, and this will eat up some percentage of their $1500 a year. They know that some time in the future – probably 30 years away – they’re likely to get AIDS, and then they’re going to need every spare dollar in that account. So, from amongst the (100-x)% of their HSA remaining after anti-retrovirals, they aim to save a little every year. Any other non-AIDS-related health concerns in the current year need to be paid for from the remainder of their $1500. So, they’re going to purchase less healthcare overall – they will restrict the total cost of health care available to them.
  2. Non-catastrophic orhopaedic surgery: Presume you’re me, and you came off your bicycle in a stupid accident that broke the little bone on your elbow that controls movement. There is only one treatment for this – surgery to reattach the little bone. The alternative is to do nothing, and lose the mobility in your arm. If you do this twice in a year, and the second time your $1500 is not enough, you either pay extra or go without. Again, you haven’t reduced the cost of healthcare; you’ve reduced the total cost of health care

Of course the libertarian view is that HSAs will reduce health care costs through the magic of competition. There are several problems with this view.

Problems with arguments about competition and cost (in HSAs)

The three problems are, basically, as follows.

Health care consumers are not informed or rational

When you go to get treated, the informed consumer of medical care is not you: it is your doctor. They are the person who knows what needs to be done and how, and you are the person who says yes or no to their suggestion. In this situation, shopping around “for the lowest price” means asking the doctor if there is a way they can do it for less. In this case the doctor will tell you what you can get rid of and you will say yes or no. It is true that you could sit at home, phoning around all the doctors in your area to find out what price they offer their services at, but you’ll find that they won’t tell you until you go in and visit them, during which time they charge you for a consultation. Given these visits need to happen during office hours, you aren’t likely to do many. Doctor’s services aren’t like buying clothes, where the quotation is free and the price is on the item, and you get to see the effect before you buy it[1]. The conditions under which you exercise your rational informed choice are limited.

Available reductions in health care cost are marginal

Consider my example above of the non-emergency orthopaedic surgery. First of all, it’s not really non-emergency – I have maybe 2 days to work out what I’m going to do before the decision has been made for me by the natural workings of my body. Working out what to do involves getting an x-ray and a consultation, and takes time. It may require appointments, of which I can only get a few in 2 days. In fact it took me half a day to get a decision about what to do, so at best I could compare 4 hospitals’ prices, and the comparison would be expensive.

This situation has only one treatment: surgery. So what are the ways in which surgical costs can be reduced? You could choose to use a cheaper (i.e. generic) antibiotic after surgery, and you can choose to get the treatment in a hospital with very bad hotel services (mixed-sex wards, no food, etc.). But the main costs of the surgery – anaesthetic and surgery – are going to be unavoidable. Although competition in these areas will push prices down, there is a significant lower limit on how low they can go – doctors don’t get cheaper than a certain amount, and anaesthetic equipment ain’t cheap. So it’s likely that the costs that patients are able to skimp on will be the ones that don’t really affect the bottom line that much.

Individuals cannot negotiate healthcare effectively

Negotiating the costs of health care is tricky, and it’s made doubly tricky by the fact that you don’t have a choice. Consider my elbow example, and suppose that due to a previous health problem I only have $200 remaining in my account. Note first of all that I have very limited ability to shop around here, since shopping around will cost me in consultation fees. The doctor tells me that they can do the treatment for $300, but I only have $200. So what do I do? I pay the extra $100 “out of pocket,” i.e. from my non-HSA. The alternative doesn’t bear thinking about – I lose the use of my arm, essentially – so I pay. How did that interaction contain costs? It didn’t, is how. Basically, when I “shop around” the doctors are going to be asking me an implicit question: how much do you value that arm? And the answer is, “how much money can I get together by Tuesday?” However, when my health insurance company shops around on my behalf, the question becomes “how much will you provide orthopaedic surgery [insert casemix code here] for?” The doctor then sets a reasonable fee, and the health insurer says “I’m offering you probably 1000 of these services a year. If you want them, you shave 10% off that,” and the doctor agrees. Shopping at a doctor isn’t like when I bought those pants at muji and they fell apart after a month, so I don’t buy muji clothes anymore. I only have 2 elbows, so it’s unlikely that I’m going to ever get the chance to refuse service to the doctor a second time, and although if I subsequently discover the doctor down the road was cheaper, if I tell all my friends that they shouldn’t use the doctor I did, the chances are that none of them will break their elbow anyway, so the doctor will never suffer from my recommendation. I know many people who have shopped at muji, and not many people who have broken their elbow.

I believe this has been born out in studies – the cost of health care paid out of pocket is higher than the cost of healthcare negotiated by a large insurer. The most obvious example of this in the US is the HMO, in which hospitals and doctors are screwed blind by the health insurance agency, and costs stay low. Interestingly, the most functional HMOs tend to occur in areas without much competition in insurance providers. Health is one of those areas where monopolies can be good for consumers, rather than bad.

The other aspect of competition and its effects on doctors is that, though they earn a lot of money, health care professionals in general are driven primarily by professional pride, not economics. It’s unlikely, for example, that you would be able to negotiate with your doctor to receive that elective surgery without anaesthetic, at half the price. “Come on doc, my mates are rugby players, they’ll hold me down.” There are lots of doctors in the US who won’t provide circumcision services, even though they’re lucrative and risk free. Doctors act on more than crude financial motives, and because they know that your health is at stake, they know that you’ll pay the extra if they offer it to you. And because you’re an individual, with no negotiating power, you can’t push that price down effectively, or even choose your treatment. The HIV patient mentioned above, for example, won’t be able to say to the doctor “I’ll take AZT instead of HAART.” Not only will the doctor refuse, but probably AZT is no longer available, because doctors will be refusing as a group to offer it once a better treatment comes along.

This is the part of health care cost containment that libertarians really don’t understand, because it relies on a basic understanding of non-financial motivations and collective action, neither of which are familiar concepts to them. Healthcare is the perfect storm of all these issues: non-rational consumers with weak negotiating power purchasing something that they didn’t want to buy in the first place, but really really need, often in a hurry. In this circumstance, having a large and powerful organisation negotiate the price ahead of time is a much better idea.

The tyranny of time

Consider two individuals, perfect rational consumers in a perfect HSA world. Label them “Peon A” and “Peon B.” Both are earning the same amount, and putting the same amount aside for their discretionary health care costs, at $1500 a year. Let’s suppose that they are earning and being taxed in such a way that their tax rate is 30%, so their HSA payments are 30% tax deductible – that is their $1500 actually costs them $1000, so they’re getting $1500 of healthcare a year for $1000.

Peon A has no significant health concerns for 10 years, and at the end of 10 years has $15000 in his health account, plus interest. Let’s assume the interest is $1000, just for the hell of it. So Peon A has $16000 to spend on health care. At year 10, Peon A discovers a non-emergency medical problem that will become painful after a year, and chooses to treat it at a cost of $10000, paid for from the fund. This $10000 was essentially purchased with $7000 of Peon A’s money due to the tax deductions – the other $3000 was born by the government, so in essence Peon A has paid $7000 for this problem. Also, Peon B has $1000 of interest, which is 1/15th of their account, that they can pour into this problem, so really they have only paid $6000 of their own income on it.

Peon B experiences exactly the same problem at the beginning of year 1. So they pay $1500 of it from their HSA, and the remaining $8500 out of pocket. The $1500 used $1000 of their own money, so in reality they have paid $9500 of their own money on the same treatment as Peon A. In addition, they lose 10 years of interest on that $1500, which means that in ten years’ time they will be able to purchase $(compound interest x 1500) less health care than Peon A.

Purely due to the vagaries of time, Peon B has paid between 36% and 58% more for the same treatment as Peon A (depending on how you calculate interest foregone). It’s possible that Peon B would have paid more due to health care inflation during that period, but we’re in a perfect HSA world here, so health care costs are being restrained to the rate of inflation, which we’re assuming 0% in this example (in order to work in static money – we could inflate all the values but it makes the calculations fiddlier). The only way that Peon B is not disadvantaged relative to Peon A is if the HSA is not working, and health care costs are increasing at least 3-5% more than inflation per year.

Now, it’s possible that the HSA could be arranged so that in subsequent years you pay back the $8500 at a tax deductible rate through your HSA – i.e. you put $1500 into the HSA and then siphon it straight to your bank account. But this will only work if you can guarantee that for the next 10 years you won’t need the money for any other health care, which of course you can’t, so Peon B has to siphon extra money into their account – i.e. reduce their take home pay – in order to benefit from the tax deductibility to the same extent as Peon A. In any case, what this option basically does is make all health care costs tax deductible, rather than making them only tax deductible if paid for from the HSA.

Of course, if Peon A and Peon B were paying that $1500 into a standard health insurance plan, they would both pay the same amount for their treatment, since it would be averaged over the 10 years for both of them as part of the insurance company’s risk pooling process. But under the HSA arrangement, Peon B has been disadvantaged to the tune of $2500 – a 4 week holiday in Hawaii, probably – simply by the vagaries of time.

This seems like a bad design flaw to me.

Some advances on HSAs

If we look back at the HSA concept, note that it involves a division of health costs into two parts: a catastrophic insurance plan and a health saving account. The catastrophic insurance plan could, of course, be provided by a single government payer, at reduced cost compared to choosing between private insurers[4]. In fact, you could  extend this model slightly to allow the “catastrophic” plan to cover all hospital-provided services (elective or not), and then have the HSA cover just physician visits and drugs. You could even then have the “catastrophic” plan extend further, to cover some fixed minimum of physician fees, and all the remainder – drugs and any “extra” services – paid for by individuals. You’d soon find that insurance agencies would step in to offer an account you could pay into to cover those extras – for example, your own room in a hospital, jumping any queues on the “catastrophic” plan. These accounts would be quite cheap to maintain, since they would only cover the above-mentioned marginal costs, not the core costs of hospital treatment, and so they’d probably be constructed as standard insurance plans (rather than HSAs), to make them more attractive to the consumer (i.e. to offer patient choice rather than the misnamed “patient power”). Eventually, seeing the numbers of people paying out of their own pocket for these plans, the government would offer a rebate – say, 30% – through taxes.

Then you’d have the Australian system, but for one small detail – in the Australian system, the hospitals are all block-funded from the government, rather than reimbursed through a government insurance system on a fee-for-service basis.

Alternatively, you could have a health care system funded entirely through block grants, in which all health services are provided free at the point of use. Then, to try and contain costs, you could introduce a system of “personal health budgets” in which people are offered the opportunity to construct their health plan from a fixed budget, to try and get them to think about how they arrange their care. This is being discussed in the UK now[5], but has attracted a lot of flak because the “personal health account” (i.e. HSA) while non-mandatory, obviously tries to get people to constrain their costs by dumping health services they need. It also attracts flak on the basis that, being non-mandatory, it’s essentially useless. Which brings us to the point of HSAs: they restrain costs by forcing certain people (i.e. the poor) to decline some health care services. The only way they will work is if this “option” to decline services is mandatory, i.e. you can’t escape it. That is, they are established to ensure that, the less you earn, the sicker you are; also, if you are born unlucky enough to have a chronic illness, then you will be able to afford less health care for other illnesses than someone in the same social class as you. Health care that for other people in your class is essentially “basic” will, for you, be “discretionary.” What is the biggest single discretionary cost for a woman in her 20s? That’s a nice choice to be faced with, isn’t it?

This is the cruelest element of these types of systems: they punish you twice for an unlucky birth, either because your parents are poor or you got a bad gene – or, worst of all, both. Sure, no-one’s choosing to punish you, but that’s the cruelty: if you establish a slightly different system, the other, luckier people can help you at very little cost to them – and it appears that in the long run those systems have lower health care costs overall as well.

fn1: libertarian rhetoric actually likes to make this facile comparison – “clothes are sold by private providers!” But the modern clothing shop is nothing like the medical model. If we bought clothes the way we bought medical services, the shop would actually be a small office, and when you go in an extremely well-educated person greets you. For a fee, you tell them what clothes you need, and they offer you a small selection of clothes that they think suit you. Measuring you will cost extra, as will testing the colours that best suit you, and the store in any case only has a limited range – but they can refer you to another shop where you can meet another highly-educated person, who will do another investigation of your needs and offer you a further range of more specialised clothes. You can’t return the clothes if they fall apart too quickly or the button comes off, and you can’t just waltz into a shop and tell them that you want to compare their prices with the guy down the road – you have to pay to do that. Also, you’re incapable of making the judgement about what suits you yourself, or taking the measurements – if you do, the chances are that you’ll leave the shop wearing the wrong clothes, fitting badly, but most likely they won’t fit at all, or may even kill you[2].

fn2: This is why if an argument depends on a cheap analogy you should probably not bother with it[3].

fb3: Religion and libertarianism seem to have this trait in common.

fn4: He asserts, without proof.

fn5: The NHS has actually been doing some interesting experiments with patient choice, to try and contain costs and improve efficiency.