Philadelphia Reflections

Health Savings Accounts: Classical Model

New topic 2015-09-03 22:42:59 description

Classical Heath Savings Accounts (C-HSA), Brought Up to Date

We begin this second edition with a shortened description of the Classical variety of Health Savings Accounts. There had been over fifteen million subscribers to this idea by the time of writing the First Edition, and since then many more people have joined. HSAs are no longer a novelty, requiring a great deal of explanation.

However, newcomers may need some basic information, and after all, there are three hundred million other people who remain to be enlisted. For them, we probably need to explain that the classical Health Savings Account was first conceived by John McClaury and me in 1981, endorsed by the American Medical Association a year or two later (I was a member of their House of Delegates at the time), widely publicized by John Goodman in his book, Patient Power, and taken up by the Cato Foundation, among many other friends. It was enacted as a pilot program under the title of Medical Savings Accounts in 1996 and finally enacted permanently as Health Savings Accounts in 2003. Bill Archer, Chairman of the House Ways and Means Committee's Subcommittee on Health is generally given credit for maneuvering the law through Congress, but a great many others helped a lot.

The actual content of the law is very simple, containing only two main parts. The first of these is a tax-exempt saving fund, and the second is a linkage to a high-deductible indemnity health insurance. Anyone between the ages of 21 and 66 is entitled to join, and after age 66 any remaining balance of the HSA changes into an IRA (Individual Retirement Account). The distinction is, expenditures from the HSA are entitled to a second tax deduction, provided they are spent on legitimate medical expenses. For the whole duration of the period of eligibility, HSA is the only "qualified "retirement account entitled to two tax deductions, the first when money is deposited, and a second if it is spent on medical care. This configuration of double-deduction is common in Canada, but in the US it is unique. The underlying reasoning was that Medicare would replace it at that age, and in addition, there were laws prohibiting payments between two federal programs for the same purpose. It later turned out these age limitations interfered with suggested expansions of the program described later in this book, so one of the modifications now suggested is some appropriate amendment of them.

It is useful however to remind people that Health Savings Accounts are quite different from Health Spending Accounts, which are entirely different. These other accounts contain a "use it, or lose it" spending limitation to the current year, leading to surges of unnecessary spending on prescription sunglasses, etc. at year-end. These spending accounts are definitely not the same thing as savings accounts, but they would come close if the "use it or lose it "feature was removed.

Furthermore, the various vendors of Health Savings Accounts have proved to be ingenious in providing alternatives within the rules. There is even a website displaying five examples of typical variations, and no doubt more will appear. Generally, those variations respond to changes in the economic scene, since many investors remember the high-interest rates of a decade ago, whereas others focus on the negligible rates currently available, but look ahead to a return of high ones. No doubt short sales against anticipated falls in short-term funds rates will appear, and all of the complexities of Wall Street may eventually make an appearance. Investors in small towns are particularly warned to consult a wider range of vendors by reviewing the choices on the internet, reading some national journals devoted to the subject, and similar means of looking beyond limited choices. The original idea was certainly to permit the subscriber to shop around for what suited him best

We, therefore, venture a recommendation, with the usual warning that past performance does not guarantee future results: we favor the purchase of low-cost index funds mirroring the entire American stock market. For a price of fewer than ten dollars a trade, widespread diversification can be obtained and maintained in what John Bogle calls "passive investing". That market has maintained a gross return of 11-12% for a century, and even after inflation is very likely to net 6% or 7% to the investor who adopts a "buy and hold" strategy. Few brokers will assume any risk for small investments unless the customer commits for long periods of time. One available strategy is to over-deposit when the account is opened, but even there the law currently limits deposits to $3300 a year, and many brokers have a $10,000 minimum. These limits, like the age limits, are becoming obsolete, and in any case, should be expanded to a lifetime limit rather than an annual one. Furthermore, it seems useful if some impartial body were given the authority to re-examine all such issues annually, particularly emphasizing the customer's ability to change agents when dissatisfied. Remember that investment agents are not required to put the customer's interest ahead of their own; they are not "fiduciaries". So be cordial, but act with care.

Good Ol' Health Savings Accounts .

Garlands of Unexpected Good Features. So the first part of a Health Savings Account is just that, a tax-exempt savings account, obtainable in the same way you get an IRA or a Roth IRA, although a few eligible outlets were slow to take ours up. And the second combined feature was to require a high-deductible, "catastrophic", stop-loss health insurance policy -- the higher the deductible, the cheaper the premium gets. Somewhat to our surprise, the idea had the greatest appeal to younger people, who immediately recognized the value of compound interest rather than simple interest, sooner than we guessed they would. So most of the early adopters are between the ages of 25 and 40, and most of the appeal has been in the savings accounts rather than the high deductible. They are wrong about that last part. You need both to make it work.

With regard to the catastrophic coverage, which spreads the risk, the more you deposit in the account, the higher is the deductible you can afford, so you save money going either up or down, but by going up, you get into a virtuous circle and the returns can be quite surprising. The industry term for this kind of insurance is "excess major medical", which the two of us wanted to avoid because of its implication it was somehow frivolous or unnecessary, when in fact it is central to the whole idea. Linked together, the two parts enhanced each other and produced results beyond the power of either, alone. The savings account was first envisioned to cover the deductible, but nowadays it also commonly attaches a special debit card to purchase relatively inexpensive outpatient and prescription costs without a lot of insurance processing and delay. That led to further administrative savings to the subscriber if he shopped frugally for optimum proportions of deductible insurance. Right now, it's a little uncertain what the current Administration will permit in the way of catastrophic health insurance, so, unfortunately, it is just about impossible to give concrete examples of what the ultimate cost will prove to be. But we do know that in the old days, a $25,000 deductible was available for $100 a year. Nowadays, a $1000 premium is more likely. When we get to explaining first year and last year of life insurance, it will become clear this premium can be appreciably reduced, once the marketing costs subside. We then got another surprise: a great many young people paid the deductible in cash, in order to preserve the compounding power of what they left in the account.

But while the savings account allowed someone to keep personal savings for himself, the insurance spreads the risk of an occasional heavy medical expense at what ought to be a bargain price for bare-bones insurance. You needn't spread any risk for small expenses because you control them yourself, but no one can afford some of those occasional whopper expenses. There's no reason why you couldn't set the deductible level yourself, weighing your own ability to withstand bigger risks. In practice, the actual savings were reported to approach 30% (compared with "First-dollar" health insurance), quite a pleasant additional surprise. Because of the younger age group of the early adopters, much of this saving was achieved in the out-patient area.

(Let's start using the present tense to talk about it, although right now it's hard to know what politics will permit.) So, hidden in this bland dual package are lower premiums, less administrative red tape, less moral hazard, but complete coverage. Right now, that's somewhat subject to change. It provides complete coverage in the sense that the insurance deductible can be covered by the savings account, but contains the option to be saved, invested or used for small outpatient expenses. Furthermore, the account carries over from year to year and employer to employer. So it eliminates job-lock, use-it-or-lose annoyances, and allows a healthy young person to save for his sickly old age.

In one deceptively simple feature, many of the drawbacks of conventional health insurance had been removed. The bank statement from the debit card can even do the bookkeeping. The first part of the two-part package, the savings account, creates portability between employers, opens up the possibility of compound interest on unused premiums, eliminates pre-existing conditions even as a concept, and creates a vehicle for transferring the value of being a "young invincible" forward into age ranges when the money really is likely to be needed for healthcare. Maybe some other features can be added later, but introducing an unfamiliar product is always greatly assisted by having it appear simple. The HSA only has two features, and yet they solve a dozen pre-existing problems.

To return to its history, nearly 15 million accounts have been opened, containing $24 billion. John McClaughry and I (neither of us received a penny for any part of this) were seeking a way to provide a tax exemption to match the one which employees of big business get when the employer buys insurance for them. That is, Henry Kaiser inspired us to do it, but at the last moment, someone slipped in a clause prohibiting the HSA from paying the premium. That alone remains undone, of the plan to restore tax equity to health insurance premiums. It should be reinstated. Although we got the general tax-free savings idea from Bill Roth, we did him one better by giving a deduction at both ends, provided only -- you must spend the money on healthcare to get the second tax relief. An additional novelty at that time was a high deductible, which permits a "share the risk" feature unique to all insurance, but invisibly limits it too expensive items, consequently hospital items. It wasn't the original idea, but it turns out you get spread-the-risk and limits to out-of-pocket patient costs in the same package. The absolute delight in discovering these features, one by one, is surely a major reason for such sales success without much marketing.

Volume control versus Price Control in Helpless Patients.We did know of a third automatic advantage, not fully exploited so far: it seems possible the hateful DRG system (with its codes restructured) could become a useful tool for dealing with a major flaw in the Medicare system. Professional peer review has become pretty good at controlling the volume of inpatient services, but prices still escape effective control. No amount of volume control can, alone, address the price issue. Controlling vital services for helpless people is a delicate matter.

Quite a few of those inpatient services match (or contain) identical items in the outpatient area. The outpatient area faces outside competition from other hospitals, drugstores or vendors, as the inpatients do not. Instead of letting helpless inpatients generate unlimited prices for the outpatients, why not let competition in the outpatient area define standards of prices for helpless inpatient captives? Outpatients and inpatients overlap in the ingredient components, considerably more than most people suppose. Inpatients may have higher overhead because of the need to supply their needs at all hours, but a standard extra markup around 10% ought to take care of that. No doubt some services are unique to the inpatient area, but a relative value scale is then easily constructed, whereby unique costs are linked to equal-cost services which are exposed to competition. Ultimately, provable relationships to market prices might even discipline big payers demanding unwarranted discounts. This last is a deal breaker, provoking suspicions of abused power by a fiduciary. The government in the form of Medicaid is often the worst offender, so we need not imagine laws alone will prevent discounts so long as law enforcement remains crippled. Every business school teaches that discounts below cost are the path to bankruptcy, but business schools have apparently not had enough experience with governments to suggest an effective remedy.

Other than two variations (double tax deductions, and incentives if used for health care), a Roth IRA would be nearly the same as an HSA, with independently purchased Catastrophic backup. We do need some workable standard for out-of-pocket limits, but the assured presence of low-cost, high-deductible insurance provides security for another needed feature :

Using individual accounts with year-to-year rollover , we could strengthen the notion of frugal young people pre-paying the healthcare costs of their own old age. To make that complete, we need permanent insurance, not term insurance.

For all we knew, there weren't any frugal young people, but we were certainly pleasantly surprised. And catastrophic insurance added the ability to share the opportunity of that feature -- subsidizing the poor at bearable prices. As we will shortly see, it also offers an incentive to save for retirement. Think of it: almost nobody can afford a million-dollar medical bill, but almost everybody can afford low premiums. Catastrophic coverage offers the only chance I know, of approaching both goals at once. And it offers the fall-back, that if you are lucky and don't get sick, you can use it for your retirement.

As the only physician in the room, I also pointed out another pretty gruesome fact: either people end their lives have a lot of sicknesses, or they end up paying for a protracted old age. Only infrequently, do real people encounter both problems. It can happen of course; breaking a hip after long confinement in bed would be an example.

People end their lives with sickness, or else they must pay for protracted old age.

Still More Good Features. Including these self-canceling needs in a single package allowed some flexibility between them -- something badly needed for a century. We cannot go on passing a new regulation for every quirk of fate; a good program must allow some latitude. Extended longevity tends to be hereditary, and so separate policies (sickness care and long-term care) are more expensive individually than the two combined because the patient can out-guess an insurance company. Health Savings Accounts balance an incentive to save for one's own future health costs "at the front end" with reasonable cost limitations "at the time of later service", even though two time periods are decades apart. That's obviously superior to just increase the sickness subsidy at the back end, because, among other things, the patient will later have even more clues about his impending future. If cost reduction goes too far at either end, it amounts to an incentive to spend carelessly. Saving becomes fruitless.

A tax deduction is a tax deduction, but this one has two: An incentive to save, and a later option to spend the savings on either healthcare or retirement. That's nearly specific enough. Furthermore, it offers a choice between saving preferences -- you can have interest-bearing savings accounts, or you can invest in the stock market, or a mixture of both. The HSA automatically converts to a regular IRA (for retirement) at age 66 when Medicare appears; that should be optional for all health insurance, but isn't. The IRA up in Canada includes both front and back features, but in the United States the HSA is the only savings vehicle to have dual deductions, so it's more flexible. As the finances of Medicare become shaky, it may be time to provide additional alternatives. At least, we ought to consider extending age 66 to a lifetime coverage option.

This harnessing of two familiar approaches makes a deceptively simple package which ought to be considered in other environments, unconnected with medical care. In most public policy proposals, the deeper you dig, the more problems you turn up. In this one, we found the proposal already had hidden answers to most concerns we could discover. It's possible to fall in love with an idea that does that for you. It lets you sleep at night, secure in the knowledge you aren't mucking things up for people.

Yet another surprise. Overall, the Affordable Care Act has probably helped sales of HSAs, since all four "metal" plans of the ACA contain high deductibles, serving in a (rather over-priced) Catastrophic role. This may be a way of covering the bets in a confusing situation. The ACA is a needlessly expensive way to get high-deductible coverage because it pays for so many subsidies. Frankly, it baffles me why subsidies swamp the costs of Obamacare but are made unworkable for HSAs. Many of the details of the subsidies are obscure, including their constitutionality, so we have to set this aside for the moment.

One good motto is don't knock the competition, but we must comment on a few things. The Bronze plan is the cheapest, therefore the best choice for those who choose to go this way. But uncomplicated, plain, indemnity high-deductible, would be even cheaper if its status got clarified. The good part is, the current rapid spread of high deductibles suggests mandatory-coverage laws may, in time, slowly go away. At first, the ACA looked like a bundle of mandatory coverages, all made mandatory at once. But they may be learning a few basic lessons as they go. Mandatory benefits are an example of mixing fixed indemnity with service benefits, with the usual dangerous outcome. Like many dual-option systems, they create loopholes. The HSA seems to avoid this issue by effectively being two semi-independent plans, for two separate constituencies -- who are the same people at different ages. Once more, we didn't think of it, the features just emerged from the plan.

That's about as concise a summary of Health Savings Accounts as can be made without getting short of breath. But of course, there is more to it, particularly as it affects the poor. For example, there is an annual limit to deposits in the Health Savings Account of $3350 per person, and further deposits may not be added after age 65. They can be "rolled over" into regular HSAs when the individual gets Medicare coverage and supposedly has no further financial needs. So plenty of people have health care, but can barely support their retirement. These plans are absolutely not exclusively attractive to rich people, but it must be admitted, poor people start with such small accounts that companies can't operate profitably unless the client sticks with them for a long time. If people possibly can, they should scrape together at least one $3300 maximum payment to get a running start.

The problems of poor people can nevertheless be eased, within the limits of the plan's design. Since people will be of different ages when they start an HSA, it might be better to set lifetime limits, or possibly five-year limits, to deposits, rather than yearly ones. Some occupations have great volatility in earnings, and sometimes a health problem is the cause of it. To reduce gaming the system, perhaps the individual should be permitted to choose between yearly and multi-year limits, but not use both simultaneously. As long as the self-employed are discriminated against in tax exemptions, that point could certainly be modified. There remains only one major flaw, which we propose should be fixed:

Proposal 6: Congress should permit the individual's HSA-associated Catastrophic health insurance premiums to be paid, tax-exempt, by Health Savings Accounts, until such time as elimination of the present tax exemption for employer-based insurance is accomplished by other means.

Subsidies for the Poor? Here's my position. If poor people could get subsidies for HSA to the same degree the Affordable Care Act subsidizes them, Health Savings Accounts should prove at least as popular with poor people as the Administration plan. Mixing the private sector with the public one is always difficult. Why not make subsidies independent of the health programs? There is no point in having the poor suffer because someone prefers a different health system. Quite often, a subsidy program is mixed with a public program, in order to make its passage more attractive; that's not necessary.

Proposal 7:That health care subsidies be assigned to patients who need them, rather than attached specifically to one or another health system that happens to serve them.

Let's just skip away from all those digressions, and return to the poor in other sections. If the concern is, health care is too expensive, why in the world wouldn't everyone favor the cheapest plan around? Part of the answer, politics aside, is that young people have comparatively little illness cost, while old folks have a lot. Since Medicare, therefore, skims off the most expensive healthcare segment of the population, the fairness of any health subsidy program is difficult to assess. Evening out the tax deduction for the catastrophic portion equalizes the unfair tax deduction for self-employed and unemployed people. Perhaps the equality issue should be re-examined after each major revision since many moving parts get jostled, every time.

The government is going to have trouble affording the existing subsidy, so it may not endure, particularly at 400% of the current poverty level. But if we can subsidize one plan, we can subsidize the other, instead. The government would then be seen, and given credit for, saving a great deal -- by inducing destitute people to use HSA as an alternative option, equally subsidized by an independent subsidy agency. As for single-payer, the government for fifty years borrowed to continue Medicare deficit financing and got it to 50% universal subsidy without much notice. That's like boiling the frog too gradually to be noticed until it is too late. But suddenly expanding the 50% subsidy to the whole country at once, would definitely be noticed. Extending such levels to the whole country should anyway be buttressed with accurate cost data. Administrative cost savings are just a smoke screen. Total costs are the real cost. Other people also point out Medicare was financed after we had won some wars, but now we seem to be losing wars.

Expanding the Money in a Health Savings Account

Let's see how short and succinct we can make it. Our task is to take the maximum amount of savings we could possibly ask the public to accumulate, invest it more or less on autopilot, and see if it can generate enough money to pay for what we assume will be health costs a century from now. Some would say that's a fool's errand, but let's see what we can do.

We start with an assumption the average person can save $3350 per year from age 21 to age 66; that's $150,750, total, the most anyone can invest in an HSA. The actuaries at Michigan Blue Cross, verified by Medicare, estimate average lifetime healthcare costs to be $350,000. Some people state you can stop talking, right there, because that's too much money. Please be patient, we will address indigency later. For simplicity, we wish to reduce the question to whether we could turn $150,750 into $350,000 in forty-five years with compound interest at reasonable rates. The answer is yes. We can't predict whether those future predictions of costs are accurate, but if accurate, they can be achieved. We assume two things:

Long Term Strategies

Indigents. We assume there will be no more indigents than at present. Can the government afford to subsidize them in this model? The answer is Yes, but its present commitment is in another direction, so it isn't entirely likely, very soon.

Outliving Your Income. We assume some people will use up their savings. If the average life expectancy, which is now 83, holds its present course, we assume this model can cope with it, even if the average life expectancy grows to be 93 in the next century. The arithmetic is quite favorable, but unfortunately, we don't know what new costs will be added in the meantime. Assuming there are some, we aren't counting them, so predictions about the future all contain this flaw.

We assume some other things, mentioned as we go along, essentially coming to the conclusion the model will produce a result which falls between the top and bottom curves in the graph. Please note the narrow range of variation in the early years and the widening upward range in later years. In particular, notice how a 3% (inflation) rate tends to stay flat well past any reasonable life expectancy, while more likely, investment income returns start to rise at age 60, and even sooner as the rate approaches 9% net of inflation. That seems to be a "sweet spot" the economy has discovered for itself in two hundred years of exploration.

We assume the equity stock market will follow the paths it followed since the Industrial Revolution. That is, it will produce an average of 12% gross return, with 3% of that eaten up by inflation, or 9% net of inflation. We then estimate our present conservative projections at producing at least 6.5% after costs, out of the remaining 9%. Dismissing inflation, we assume the stock market will operate between a 2% real return, and 9.5% real after inflation, leaving a 3.5% "cushion" for contingencies. When the Industrial Revolution ends, these basics may also change. We have a decade or so to try to get the investor's returns up closer to 8% safe level. And meanwhile, we must try to remain prepared for a bleak and bad depression, a "black swan", on average every 28 years, but individually unpredictable. In the meantime, we aspire eventually to pay for 100% of healthcare expenses, but promise to pay only a quarter of that. And finally, we assume medical care will change so much during the next century, that our calculations will need to be totally revised, long before then, with a so-called mid-course correction. With the understanding, that anything which pushes outside of the accompanying graph will have an obvious explanation, we assume future managers will make appropriate adjustments.

Single Premium Investment/Look carefully at the graph. It makes an unfamiliar assumption. It assumes a newborn baby started a Health Savings Account at birth, deposited $500 in it, and didn't touch the account again until he died. It is our assumption the average person could do that, perhaps with a stretch, and our further assumption that the government could do the same for indigent babies. There are times when neither the government nor many middle-class people could manage the necessary expenditures, but we set the value of $500 at birth as an extreme limit of what we think they both could do, on average most of the time. It's a number which is easily changed if the economy varies from our projection.

Let's dramatize the point we're making on a totally different scale, by temporarily appointing Warren Buffett as its role model. According to a story in the Wall Street Journal by columnist Morgan Housel, this is the way the best investor in history made his money. At the age of eighty-four, his personal wealth was $73 billion. Of that, he made $70 billion after the age of sixty. Some might retort, the trick is to make the first $3 billion by the age of sixty, but a more civil underlying moral is that compound interest really starts to work toward the end of life.

Just take another look at that graph; the particular power of compound interest works as efficiently with $500 as with $3 billion. It starts earlier with higher interest rates, in this case, age 40 at 12%, compared with age 85 at 5%, Mr. Buffett's numbers. Obviously, it pays to start early and to get higher interest rates if you can do it safely. And conversely, it's a bad idea to spend or squander your savings while you are young. Our preferred method is to raise the interest rate by reducing the attrition of middle-men, in the approach mentioned earlier. You might not reach 12%, but you have a fair chance of reaching 9% if you allow yourself fifteen years to work on it. In the meantime, be satisfied with less than 100% coverage by this method.

More seriously, why else did we pick this way to depict the future? Because at age 66, when Medicare takes over, all of the plausible curves have reached a point where they could match Medicare's expenditures, indefinitely. If Medicare went broke, or was otherwise unacceptable for some reason, liquifying the account would produce a sum matching Medicare's present rate of expenditure. And finally, the numbers become so astronomical at the far end, it seems entirely reasonable to transfer part of the account to a grandchild's account. That trick alone should greatly reduce the problem, and add 21 years for compound interest to do it. As we will see in a coming section, paying childhood health expenses in advance solves some otherwise difficult issues.

Eligible Health Expenses

Back when Health Savings Accounts were started, just about anything you could buy in a drugstore was a healthcare expense. However, if you go into a drugstore today, especially a chain drugstore, you can buy ice cream cones, cosmetics, shaving equipment, and so many other things, you have a little trouble finding the pharmacist. Therefore, a debit card receipt from a purchase of a drugstore item will cover many more things than the Health Savings Account originally contemplated.

Eligible-item Debit Cards

Evidently, a number of people exploited this loophole, and it isn't surprising the regulators responded with regulations. The drugstores could have responded by making debit cards only apply to eligible items, but many didn't. So now you need a prescription from a doctor for the item, and you are subject to a 20% fine if you use the accounts for a non-eligible service. Some congressman could devise a system which would serve the purpose without the red tape, and Congress might get rid of this nuisance on behalf of the (now) 14 million subscribers to HSA, and the already beleaguered clerks in the drugstores. Next thing you know, the medical supply stores will start selling ice cream cones, followed by more regulations to prevent such evasions. That's, of course, the problem with depending on the stores to give up the illicit sales, in preference to giving up the red tape.

Several websites have started broadcasting lists of eligible items, hoping you will buy such items from them over the Internet, and eventually, the competition will sort this out. Meanwhile, we have to advise you to check the websites before you shop.

Longevity, a Moving Target

In the last fifty or so years, American life expectancy has increased by thirty years, enough extra time for three extra doublings at seven percent. So, 2,4,8. Whatever money the average person would have had when he died in 1900, is now expected to be eight times as great, since he dies thirty years later in life. And even if he should lose half of it in some stock market crash, he will still retain four times as much as he formerly would have, at the earlier death date.

The lucky reason increased longevity might rescue us is the doubling rate started soaring upward at about the time it got extended by improved longevity in 1900 (when life expectancy was 47). In particular, look below at the whole family of curves. Its yield turns increasingly upward for interest rates between 5% and 10%, and every extra tenth of a percent boosts it appreciably more. Let's take a small example. Why don't we invest everything in "small" capitalization companies? Because there aren't enough of them to support such a large diversion to a frozen account. We are therefore forced to concentrate in large capitalization corporations, yielding only 11%. A few tenths of a percent extra yield might be squeezed out of this curiosity. Life expectancy is slowly but steadily lengthening. And so on. It's useful for the nation to realize that having everybody live longer is a good thing, just as long as too many extra people don't get sick with something expensive.

In the past century, inflation has averaged 3% per year, and small-capitalization common stock averaged 12.7%. That results in an after-tax growth of 9.7%. Some people consider 3% inflation to be good for the economy, many do not. The bottom line: many things have changed, in health, in longevity, and in stock market transaction costs. Those things may have seemed to have deviated very little, but with the simple multipliers we have pointed out, that upturn in income at the end of life becomes steadily magnified. If you do nothing at 3%, your money will be all gone in thirty-three years. That is if you leave your savings in cash. While it is true there are risks with all choices, the option of being a deer in the headlights is a poor one. There's a small but critical margin, and everyone must collectively struggle for very small improvements in it.

If you work at things just a little, you take advantage of the progressive widening of two curves, also shown on the graph: three percent (for inflation) remains pretty flat, but seven percent (for investment income) starts to soar much earlier. Up to 7%, there is a reasonable choice between stocks and bonds; but if you need more than 7% you must invest in stocks. Future inflation and future stock returns may remain at 3 and 7, forever, or they may get tinkered with. But the 3% and 7% curves right now are getting further apart with every year of increasing longevity. Some people will get lucky or take inordinate risks, and for them, the 10% (large-company stocks) investment curve might widen from a 3% inflation curve a whole lot faster. But except for desperate gamblers, every single tenth of a percent net improvement, will cast a long shadow. That means blue-chip common stocks are best, except during a black swan crash where all bets are off, but bonds are probably least bad.

Save it, or Spend it. You can't do both.

But never forget the reverse: a 7% investment rate will certainly grow much faster than 4% will, but if people allow this windfall to be taxed, gambled or swindled, the proposal you are reading will fall short of its promise. We are offering a way to minimize taxes, the other two risks are your own problem. Our economy operates between a relatively flat 3% and a sharply rising 4-5%. In other words, it wouldn't have to rise much above 3% inflation rate to be starting to spiral out of control. Our Federal Reserve is well aware of this, but the public isn't. A sudden international economic tidal wave could easily push inflation out of control, in our country just as much as Greece or Portugal if they leave the Euro. Another issue: As developing, nations grow more prosperous, our Federal Reserve controls a progressively smaller proportion of international currency. Therefore, we could do less to stem a crisis that we have done in the past.

To summarize, on the revenue side of the ledger, we note the arithmetic that a single deposit of about $55 in a Health Savings Account in 1923 might have grown to about $350,000 by today, in the year 2015, because the stock market did achieve more than 10% return. It might be more realistic to say $250 at birth rather than $55. but the principle is sound. You can't do it twice, but it ought to work, once. There is therefore considerable attractiveness to the expedient of extending HSA limits down to the age of birth, and up to the date of death. It's really up to Congress to do it.

If the past century's market had grown at merely 6.5% instead of 10%, the $55 would now only be $18,000, so we would already be past the tipping point on rates. You do have to leave some extra room. In plain language, by using a 10% example, $55 could have reached the sum now presently thought by statisticians -- to be the total health expenditure for a lifetime. But by accepting a 6.5% return, the same investment would have fallen well short of enough money for the purpose. Unlike the municipalities that gambled on their pension fund returns, that sort of trap must be anticipated to be avoided. Things are not entirely hopeless, because 6.5% would remain adequate if our hypothetical newborn had started with $100, still within a conceivable range for subsidies for the poor. But the point to be made provides only a razor-thin margin between buying a Rolls Royce, and buying a motorbike. If you get it right on interest rates and longevity, the cost of the purchase is relatively insignificant. That's the central point of the first two graphs. For some people, it would inevitably lead to investing nothing at all, for personality reasons. Some of the poor will have to be subsidized, some of the timid will have to be prodded.

This is more of a research problem than you would guess: a round-about approach is to eliminate first the diseases which cost so much, choosing between research to do it, or rationing to do it. Right now we have a choice; if we delay, the only remaining choice would be rationing.

Commentary.This discussion is, again, mainly to show the reader the enormous power and complexity of compound interest, which most people under-appreciate, as well as the additional power added through extending life expectancy by thirty years this century, and the surprising boost of passive investment income toward 10% by financial transaction technology. Many conclusions can be drawn, including possibly the conclusion that this proposal leaves too narrow a margin of safety to pay for everything. The conclusion I prefer to reach is that this structure is almost good enough, but requires some additional innovation to be safe enough. That line of reasoning will be pursued in a later chapter.

Revenue growing at 7% will relentlessly grow faster than expenses at 3%. As experience has shown, it is next to impossible to switch health care to the public sector and still expect investment returns at private sector levels. Repayment of overseas debt does not affect actual domestic health expenditures, but it indirectly affects the value of the dollar, greatly. With all its recognized weaknesses, a fairly safe description of present data would be that enormous savings in the healthcare system are possible, but only to the degree, we contain next century's medical cost inflation closer to 2% than to 10%. The simplest way to retain revenue at 7% growth is by anchoring the price leaders within the private sector. The hardest way to do it would be to try to achieve private sector profits, inside the public sector. This chapter describes a middle way. Better than alternatives, perhaps, but nothing miraculous. .

What HSA Does for Poor Folks: A Proposal

When I give talks about Health Savings Accounts, there often seems to be some person in the front row who doesn't seem to be listening. But he's usually the first to raise his hand with a question, which goes more or less like this: "Well, what does this do for poor people?" So here's my usual answer.

Poor people are all poor, but they are poor in different ways. For preliminary discussion, let's divide them into three classes.

On Their Way Up. Most Americans came here at the bottom and worked their way up. Poverty may once have been their condition, but it wasn't their ambition to stay there. Everyone, particularly the newcomers, can see that cheaper means more people can afford something they once couldn't afford. It's the job of Health Savings Accounts to make healthcare cheaper, but if you subsidize more than half of the population, and then set a threshold of 400% of poverty, you tend to hold people in place. You tend to make subsidies hard to surrender, which increases the number of subsidized people, and ultimately makes subsidies too expensive to continue. It may be well-intended, but it makes things worse. The Latin expression Primum non-nocere is the medical profession's motto, meaning "The least you can do for somebody, is not make their problem worse." On the other hand, by making healthcare cheaper, more people can afford it, so you've done a lot of good.

Like Tolstoy's Unhappy Families, Poor people are Poor in Different Ways.

Stuck at the Bottom. It's true a dismaying number of people are permanently unemployable. Not just unemployed, but unemployable. The Mayor of my little suburb tells me 8% of the school budget is devoted to "Special education", which mostly means mental defects of one kind or another. In spite of special education, a large proportion of mentally retarded kids will never be able to support themselves. And despite movies about Nobel prize winners with Lou Gehrig's disease, a lot of other people born with neurological conditions will never be self-supporting, either. My profession is working hard to reduce the number of permanently disabled, and quite often it is fiercely expensive to treat them, but we keep doing it. For the most part, these disabilities are easy to recognize, and with few exceptions, it is society's obligation to subsidize them indefinitely. But it is not the role of Health Savings Accounts to define, identify or treat these people. In fact, it would injure our performance to take on a non-financial role. Give us the money and we will expand it and then pass it along. We will even contribute toward its cost, but much prefer to have the government pay its own bills, and not disguise their taxes as part of our operating budget. Who made it government's duty? Our elected representatives in Congress did, and we try to follow their rules.

Temporary, Borderline, and Political. For a while, I acted as a referee on Disability Determination. Let me tell you, it's often pretty hard to tell who is malingering, from who is eligible among a host of different assistance agencies, and who has long since recovered from a disabling condition. It's therefore expensive to administer Disability Determination, and physically exhausting if you take it seriously. It isn't the proper job for a Health Savings Account, which would do it very poorly, dragging down the performance of what else we would really like to do perfectly. Which is to reduce the effective cost of healthcare, by adding an unexploited source of revenue in the financial field.

HSA Proposal for the Poor. The proposal, therefore, is we should start with what is least controversial in almost everybody's mind, which is catastrophic health insurance. This type of insurance comes closest to what everyone would agree we owe all our citizens. No money is expended on the basis of income or social circumstances, it is decided individually at the hospital door, usually by accident room physicians. Its volume of component services would be controlled by an improved DRG, and its retail price should be determined by outpatient costs determined in turn by the marketplace, or by a relative value scale when no comparable outpatient service exists. It may be advisable to reconstitute the PSRO (Professional Standards Review Organization).

It may be desirable to dispense this catastrophic insurance through an HSA, although it is not essential. If it is the only benefit provided, it still would be useful to provide an accordion mechanism for additional services for those who rise from poverty, who are reimbursed through special programs in a variety of ways, or who find ways to supplement the cost themselves. Ultimately, it should provide a vehicle for integration, temporarily or permanently, into the private sector.

Commentary: Agency for Mid-Course Corrections

Among the many things we don't know about the future, is the average longevity eighty years from now. The whole-life insurance industry prospered when they sold policies assuming American longevity of 47 years in the year 1900, and it turns out to be 83 today, still growing fast. If longevity should get shorter, as it recently has in Russia, life insurance would go out of business. Since we can't rely on projections, we have to rely on early observations and make mid-course corrections.

President Lyndon Johnson both underestimated how much Medicare would cost, and how politically successful it would be. He was in no position to multiply 50 million Americans times $11,600 per year per person, times 22 years per person. That simple sentence tells you all you need to know about current Medicare costs, but who knew? Nor could he know how fast longevity would grow, or how fast the cost would rise. But we can monitor the trend, extrapolate, and revise the extrapolation. Medicare was a medical success, which had to be paid for; and President Johnson's successors might have found that out a little sooner, and changed course. If we must find fault, failure to readjust early would be my candidate.

So, who is counting?

Quis custodiet custodies?

For reasons obvious or not, the nation would be well served to create a monitoring agency for the guidance of future Congresses in charge of the type of Health Savings Accounts we already do have, and maybe some related issues. When we start envisioning lifetime coverage, it becomes even more vital to have a permanent agency to sort out what is happening. This is particularly important when the Branches of the Federal Government are divided between the two parties. Informally, the subcommittees of the Appropriations Committees have assumed much of the burden of overseeing agencies. They are the only Congressional Committees to review every program every year. However, the Agencies have grown to be the largest bureaucracies in the government, and tend to become jealous of their independence, as the Appropriation Committees grow too burdened to bother with them. It begins to look as though we need more Congressmen if we want Congress to maintain closer control of the agencies. Each Congressman now represents a million constituents, and simply cannot do all we would like him to do. As much as anything, we need a core group who worry about issues in advance and have the prestige and access to make their views be heard. Rather than design a blueprint, let's review some issues that such a body might explore.

Proposal 8: Health Savings Account Age Limits Should be Extended, from the Cradle to the Grave. A few extra years might be a minor improvement in special cases. The real benefit would be to create a continuous account, which could grow over long periods of apparent inactivity.

Proposal 9: Instead of annual contribution limits, the limit for HSA should extend over several years, or even be a lifetime limit. When deposits must be skipped for health or occupational reasons, there should be an opportunity to catch up. Athletes and similar occupations tend to concentrate on earning power in a few years.

Since the HSA is increasingly accustomed to augmented retirement income, thought should be given to extending the idea to amounts of money which could encourage that use. Furthermore, there are special circumstances, like a partial Medicare buyout, in which a limit to deposits forces a choice between two desirable uses for the same money. If the individual has the money for more than one purpose, it seems wrong to force a choice. For example, it's considerably safer to over-deposit more than you believe you will need, planning to return the excess. As a practical matter, the usual danger is overoptimistic revenue projection. Someone who sells his business at age 63 might have enough cash, but still, encounter trouble with the $3300 per year limit because he once needed the income to run the business. It seems pointless to squeeze through such a narrow window, and much better if the window were at least enlarged to permit lump-sum deposits up to a $ 132,000-lifetime limit. With that sort of cushion, plus a stretch of reasonably good health at the right time of life, it would become considerably safer to take risks. At age 65, a lifetime of health costs is already in the past, but the curve of health expenses starts to bend upward at age 50, at a time when college expenses for children may be persisting, and the house isn't quite paid for. It seems a pity to cripple a good idea with pointless contribution limits that almost stretch far enough, but leave people fearful. If Congress develops a serious interest in lifetime insurance, the yearly contribution limit should be revisited. The optional side use for retirement should be examined in parallel, including its potential for being gamed.

Revisited by whom? Someone should be empowered to travel, and talk to people in the field. Maybe hold hearings, maybe just interview. A simplified goal is, therefore, to accumulate $80,000 in savings by the 65th birthday, intending to make a single-premium buy-out. That clarifies costs, but is it practical on a large scale? Remember that savings get a lot harder when earned income stops. With the current law, you would have to wake up and start maximum annual depositing of $3300 by your 50th birthday, to reach $80,000 by age 65, and you would need generous internal compounding to make it. But notice how easily $100-200 a year would also get you there, starting at age 25 (see below), even justifying somewhat less optimistic investment income returns until age 65.

Many more frugal people might skin by with looser rules; It even could rather easily be subsidized for poor people and hardship cases. If you are going to cover lifetime health costs instead of just Medicare, many more will need $80,000 to do so and have something left to share with the less fortunate. But to repeat once again, that still compares favorably with the $325,000 often cited as a lifetime cost. That's all we care to promise in public, but secretly we know it may not be enough. The plain fact is, if longevity or inflation get out of hand, someone must have the authority to raise the contribution limits and to do that, there should be some research by a trusted house actuary.

Proposal 10: Instead of the present annual limit of contributions to Health Savings Accounts of $3300 per year, Congress should permit a lifetime limit of $132,000, with annual deposit limits adjustable to bring accounts at their present age, up to what they would have been if $3300 annually had been deposited since age 25.

Proposal 11: Congress should reserve decisions to itself for changing the lifetime contribution level, and review final appeals from contract terms which seem to threaten imminent major adjustment to the general public lifestyle.

The Cost of Pre-funding Medicare. Rates of 10% compound income return would reduce the required contribution to $100 per year from age 25 to 65, but if the income were only 2% would require $700 contributed per year, and at 5% would require $300 per year. Remember, we are here only talking of funding Medicare, as a tangible national example,

It is this calculation, however rough, which has made me change my mind. It was my original supposition that multi-year premium investments would only apply up to age 65, and that would be followed by Medicare. In other words, HSA should only be implemented as a less expensive substitute for the Affordable Care Act. It seemed to me the average politician would be very reluctant to agitate retirees by proposing a plan to eliminate Medicare. They would feel threatened, the opposing party would then fan the flames of their fears, and the result would have a high likelihood of undermining the whole idea for any age group, for many years. Better, I thought, to take the safer route of avoiding Medicare, and confining the proposal to working people, where its economics are overwhelmingly favorable.

But when the calculations show how close this proposal under cautious revenue projections could come to failure, and when nothing else remotely close to it has been proposed by anyone, the opportunity runs the risk of passing us by. So, I have changed my mind. The moment of opportunity is too fleeting, and the consequences of missing it entirely are too close, to worry about the political disadvantage of doing the right thing. The transition to a pre-funded lifetime system will take a long time to get mature, and the political obstacle course preceding it is a daunting one. At least we should allow it as a demonstration option, where some fears will prove unwarranted, while others can be corrected.

So we make the guess of the average life expectancy where things will eventually flatten out, will then be about 91. (Be careful, most census figures are for life expectancy-at-birth.) But many people would have to be lucky in all details: a favorable investment climate for the right ten-year periods, plus a favorable health situation which avoids expensive illnesses just at the age when they begin to threaten. Some life-saving scientific advances would be a big help, too. Using a lower goal of $80,000 and an interest rate of 7% is considerably easier to conjure, but the barrier which might be reached first is the $3300 yearly contribution limit. Some unfortunate individuals might be forced to pay all medical expenses out of pocket in order to make the investment fund stretch, even before the average becomes affected. The individual who came up short might still remain considerably ahead of where he would be without an HSA, but we are using a precise match of revenue and expense, to simplify the examples.

Someone who sells his house or business at age 63 might have the cash, but still, have trouble because of the $3300 per year deposit limit. It seems pointless to squeeze through a narrow window, and much better if the window were enlarged to permit lump-sum deposits up to a $ 132,000-lifetime limit, adjusted for inflation and compound income returns. With that sort of cushion, plus a stretch of reasonably good health at the right time of life, it would become considerably safer to take the risks. At age 65, a lifetime of health costs is nearly in the past, but the curve of health expenses starts to curve up at age 50, at a time when college expenses for children may be persisting, and the house isn't quite paid for. It seems a pity to cripple a good idea with pointless contribution limits that almost stretch far enough, but leave people fearful. If Congress develops a serious interest in lifetime insurance, the yearly contribution limit should be revisited.

The simplified goal is, therefore, to accumulate $80,000 in savings by the 65th birthday, remembering that savings get a lot harder when earned income stops. With the current law, you would have to start maximum annual depositing of $3300 by your 50th birthday, to reach $80,000 by age 65, and you would still need generous internal compounding to make it. But notice how easily $100-200 a year would also get you there, starting at age 25 (see below) and less optimistic investment income returns until age 65. Many more frugal people might skin by with looser rules; poor people and hardship cases could more easily be subsidized. If you are going to cover lifetime health costs instead of just Medicare, many more will need $80,000 to do it and have something left to share with the less fortunate. But to repeat once again, that still compares very favorably with the full $325,000 which is often cited as a lifetime cost. We have already imposed an extra $80,000 internal savings requirement in order to include Medicare; here is the place it would be a hardship. That's about as far as concentrated thought will carry you. It leads to the conclusion that it might be better to modularize Medicare and let the public pick and choose what it wants to buy its way out of.

The Cost of Pre-funding Medicare. Rates of 10% compound income return would reduce the required contribution to $100 per year from age 25 to 65, but if the investment income were only 2% would require $700 contributed per year, and at 5% would require $300 per year. Remember, we are here only talking of funding Medicare, as a well-understood national example, Obviously, a higher return would provide affordability to many more people than lesser returns. When $100 competes for the investment income from 10%, it's much easier than $300 competing for 5% income. Let's take the issues separately, but don't take preliminary numbers too literally. They are primarily intended to alert the reader to the enormous power of compound interest, and the big difference made by relatively small changes in it. Let's go forward with some equally amazing investment discoveries which are more recent, and vindicated less by logic than empirical results.

A transition from term insurance to pre-payment of Medicare is greatly eased by forgiving the premiums and payroll deductions, which are roughly age-distributed, and can, therefore, be forgiven in a graduated manner for late-comers to the program. Most cost-redistribution of high-cost outlier cases should be handled through the catastrophic insurance, which is well suited for invisible and tax-free redistribution. Because of hospital cost-shifting, inpatients are temporarily overpriced but are quickly becoming underpriced as a result of hospitals gaming the DRG to shift costs to outpatients. This will in time affect the relative costs of Catastrophic and Health Savings Accounts and must be carefully monitored for mid-course adjustments. This changing horizon of cost shifting reinforces the need to create a special agency to keep track of it. And to report its findings to Congress, who can consider the broader political implications, once they know the facts.

Proposal 12: Congress should create and fund a permanent Health Savings Account Agency. It should have members representing subscribers and providers of these instruments, with the power to hold hearings and make recommendations about technical changes. It should meet jointly with the Senate Finance Committee and the Health Subcommittee of Ways and Means periodically. It should have extensive access to the appropriate Executive Branch department, to review current activity, detect changing trends, and recommend changes in regulations and laws related to the subject. On a temporary basis, it should oversee inter-cohort and outlier loans, leading to recommendations about the size and scope of inter-subscriber loan activity. At first, it might conduct the loan activity itself, with an eye toward eventually overseeing a commercial vendor.

Cost Sharing with Frugality.At present costs, statisticians estimate future healthcare costs of about $325,000 (in year 2000 dollars) for the average lifetime. We could discuss the weaknesses of that estimate, but even though it's breathtaking, it's the best guess available. Women experience about 10% higher lifetime health costs than men. Roughly speaking, how much the average individual somehow has to accumulate, eventually must equal what he spends by the time of death. The dying individual himself has little interest in what is left unpaid at his death, so Society must do it for him, in order to survive as a Society. At this point, we, unfortunately, must also work around one of the great advantages of having separate accounts.

On the one hand, individual accounts to create an incentive to spend wisely, but it is also true that pooled insurance accounts make cost-sharing easier, almost invisible, and tax-free. Cost sharing induces reckless spending of other people's money, however, while individual accounts induce frugality with your own money. Therefore, linking Health Savings Accounts with Catastrophic insurance provides a way to pool heavy outlier expenses, while the incentive for careful money management remains in the outpatient costs most commonly employed (together with a special bank debit card) to pay outpatient costs. Such expenses are much more suitable for bargain-hunting anyway because dreadfully sick people in a hospital are in no position to argue or resist.

But a cautionary reminder: linking individual accounts to frugality through outpatients, as well as linking heedless spending to insurance through inpatients -- induces hospital administration to game the system we have here imagined. There's no doubt a system can be gamed by shifting medical care to the outpatient area, but we must expect the DRG to be attacked, in order to reverse such incentives, which run in the hundreds of billions of dollars. A well-informed monitoring system simply must be created and funded, if we ever expect the decision to hospitalize patients to rest on whether the patient needs to lie down, instead of on what kind of payment system we happen to fancy. At the same time, the present DRG coding system must be considerably improved to withstand being subverted. These are not tasks which congressmen typically enjoy, but they must be done within the legislative branch if we expect it to function.

Standard Deviation within and between age cohorts.Furthermore, there is an important distinction between a mismatch of revenue to expenses caused by chance within one age group and a revenue mismatch between two age cohorts. To put it another way, somebody has to pay off these debts, and we must have a plan about who should pay them when revenue is not present in the account. Borrowing between subscribers within the same age cohort should pay modest interest rates to forestall gaming, but borrowing between different cohorts for things characteristic of their age level (pregnancy, for example) should pay no interest if at all feasible. Unfortunately, people sometimes abuse such opportunities, and interest must then be charged. Until the frequency of such things becomes better established, this function of loan banking policy should be part of the function of the oversight body, rather than the executive agency, which tends to want to retain the function. When its limits become clearer, it might be delegated to a bank, or even privatized, but the policy must be monitored by specialists who understand what is happening "on the ground". While it is unnecessary to predict the last dime to be spent on the last day of life, incentives should be understood by the managing organization, separating routine cash shortages from likely abusive ones. And looking at all such activity as potentially having been caused by payment design. Much of this sort of thing can be minimized by encouraging people to over-deposit in their accounts, possibly paying some medical bills with after-tax money in order to build the fund up. Such incentives must be contrived if they do not appear spontaneously. User groups can be very helpful in such situations. People over 65 (that is, those on Medicare) spend at least half of that $ 325,000-lifetime cash turnover, but just what should be counted as careless overspending, can be a matter of argument.

Proposal 13: Current law permits an individual to deposit $3300 per year in a Health Savings Account, starting at age 25, and ending when Medicare coverage begins. Probably that amount is more than many young people can afford, so it would help if the rules were relaxed to roll-over leftover entitlements to later years, spreading the entire $132,000 over the forty-year time period at the discretion of the subscriber.

Finally, an observation. The classical Health Savings Account will save a big chunk of money, but who gets it will depend heavily on the health of individual subscribers because it is term insurance, year to year. Two concepts loom over it: (1) The nation may want to distribute the good and bad luck more evenly, and (2) It would be much easier to cut down an over-funded project than to supplement an under-funded one. If we can think of some ways to improve the product, we should start with as generous a benefits package as we can easily devise. Therefore, the rest of this book is devoted to making the returns more generous, and the outcome more predictable, sooner.