Philadelphia Reflections

The musings of a physician who has served the community for over six decades

Related Topics

Health (and Retirement) Savings Accounts: Steps To Lifelong Health Insurance
If you are a fast reader, we will begin with a ten-minute summary of Health Savings Accounts. At first, it covers future revenue, then spending projections follow. No matter how medical care changes, cost and revenue must remain in balance.

The Arithmetic Behind Our Claims, With Commentary

Knowing how to calculate is less important than knowing what to calculate. Any computer's search engine has several free examples of compound interest calculators. It is also useful to know a few short-cut ways of calculating approximate answers in your head, just to keep a check on computer calculations going off the rails. Be careful, because calculators can have bugs (or the computer's memory cache may not unload), but the commoner mistake is to calculate the wrong components.

The chief short-cut to know is a $100 fund of money earning 7% will double to $200 in ten years. Its related fact is, in fifty years a hundred dollars at 7% will equal (2,4,8,16,-->32) times a hundred dollars, or $3,200. With compound interest, the result is not linear. The 7% in 10 years shorthand is a subset of the "rule of 72", which says you can divide 72 by the compound interest rate, to get the number of years it takes to double. So while 7% doubles in 10 years, 6% doubles in 12 years, 10% doubles in 7 years, etc.

It's true, $100 at 10% will double every seven years, (as in 2, 4, 8, 16,--> 32x) in 35 years, or to $3200, reaching the same value as 7%, but in fifteen fewer years. Our economy contains more examples of 7% interest charges, than of 10%. The difference between 7% and 10% is surprisingly great and surprisingly meaningful in reverse. The 10% figure is often applied to impaired credit, or galloping inflation and therefore higher rates are often a warning signal, not an opportunity. Approximations permit the reader to recognize anomalies or to verify accurate calculations without repeating them.

Both approaches, short-hand and the Internet calculator, are recommended for verifying the following calculations, involving advance predictions which cannot be precise. One method of signaling approximation is to round off the answers to three or more digits of zeroes (2,400,000 is most probably an approximation, for example). We prefer fully calculated numbers, not to imply an accuracy which isn't present, but to allow the reader to follow arguments by recognizing components as they jump from line to line without elaborate identification.


Summary of the Math, With Caveats. This proposal is not free, it mainly substitutes investment income for government borrowing. The government thus benefits, but the patient benefits as well because he ends up owning the investment. This is an essential feature, since without it the consumer may see the proposal as benefitting the government, without benefitting himself. The calculations are necessarily rough ones, and 7% may prove optimistic, but the system is self-balancing. Borrowing is reduced, so in the example, the consumer gets the same product for roughly half its present net cost. "Roughly half" is approximate; the actual saving will be determined by market forces, both when he buys the investment, and later when he sells it.

To be explicit, this proposal involves passing early savings into the individual's own savings account, and immediately investing them. The consumer can measure his savings growth and later can watch his own bills being paid with them. He keeps any savings for his own retirement, and even more, if he is frugal enough to generate extra savings. Furthermore, he lives longer, which is largely only possible if his government has the money freed up to spend more billions on basic research to cure the diseases he might have died of. It is, therefore, possible to describe this whole system as a shrewd investment, growing out of cold, hard self-interest. Which, as the owner of the investment, visibly improves the individual's private benefit. So to return to the point we were making in the last chapter, hold off on ridiculing the idea of a newborn babe contributing money to it, it just looks that way at first.

To resume the arithmetic discussion, a deposit in a Health Savings (and Retirement) Account of $1650 1 yearly from birth to age 20 (from diverted Medicare premiums not yet authorized by Congress to be diverted) would result, invested at 7% in an escrow balance of $72,192 at age 20. If, at age 20, contributions of $825 yearly 2 for 40 more years (Source of revenue: diverted average Medicare withholding tax) were then substituted, both contributions firmly placed in escrow against college expenses and the like, might result in a balance of $1,260,000 at age 60. If investments realize less than 7%, this balance will be less. At age 65, 7% interest alone for five more years, generates a balance of $1,772,351. After subtracting a Medicare buy-out at age 65 of $134,000 3 plus $30,000 (income tax), a balance is left for retirement of $1,608,351. That hypothetical balance could provide a pretty reasonable retirement fund starting at 65, and eliminate additions to the Medicare debt, and by age 84 (average life expectancy) still allow for extraction of $18,000 4 to a designated member of the child or grandchild generation, at birth.

For persons in transition with less than a full lifetime to work this out, a postmortem Trust Fund could continue to earn interest from average age of death (84) to the legal limit of a perpetuity, dedicated to paying Medicare shortfalls even though the subscriber is dead, until the age of 105 5 when further debts are extinguished. Meanwhile, the $18,000 legacy recirculates in a child or grandchild's Health Savings Account, replenishing the fund for a new generation. It should be emphasized this escrow fund is for Medicare, to which major illness is increasingly migrating. (We here suggest Congress waive the Catastrophic insurance requirement during the sixty years when no medical withdrawals are being made.) Current medical costs for younger people, balancing the Medicare migration, should diminish, and be revenue-neutral for the lifetime system. In any event, they do not include either surplus or deficit from Obamacare.

This condensed example is given to answer the question: is there enough money in the system? There probably is, and the trick is to find ways of getting it out. The difficulties are political and cultural, not financial. Other programs can join the system like pearls on a string, so long as they individually remain revenue-neutral, or subsidies are created to make them so. Success depends on achieving 7% returns from passive investing 6 , and revenue-neutrality among newly added programs. In a century, scientific discoveries might even reduce overall costs to the first year, and the last year, of life.


1 For its own reasons, this is entirely credited to Medicare Part B, and Part A gets none of it. As a further twist, it is deducted from Social Security checks rather than billed.
2 The allowable amount is $700 per year, with an income tax deduction at the end of the year.
3 This is my own guess as to the probable buy-out price of Medicare, age (65) to average life expectancy (84)
4 The is the CMS estimate of childhood costs, birth to age 25.
5 A perpetuity is defined as one "life" plus 21 years.
6 My own portfolio averages 6.7%. I have it from John Bogle personally that 7% is a stretch, selected because of its calculating ease.


It is not intended to follow this outline strictly. For example, the transition period to it must be somehow shortened, avoiding transitional bond issues whenever possible. One additional twist would be to take advantage of the population continuing to live longer. Half of Medicare expenses are devoted to the last four years of life, so the transition period can be cut in half by dividing the revenue into two escrow funds, one of which is funded by the compound interest for 84+ years, and reimburses Medicare for the last 4 years of life it has already paid for. This puts the burden in half for the second fund, a particularly useful feature during the transition period.

If the public will not stand for using old-age premium money to fund childcare, or if there is resistance to shifting obstetrics cost from mother to child, either the retirement must be reduced, or new sources of revenue found. If there is significant resistance to other generation-shifting of funds, it is usually less important financially than socially.

Without any government changes, passive investing is gaining on active investing, by almost a trillion dollars a year. The financial community can be expected to resist the shift to passive investing. Whether they will stiffen their resistance or adapt to the change, remains to be seen.

Further, if the ownership of index funds can be tolerated, perhaps a segment could be directed to venture capital funds for cost-reduction of medical corporations, or even to transforming medical industry profits into cost-reduction for its products. When the health industry consumes 18% of the gross domestic product, it must consider new financing methods if it is to avoid either government ownership or a purely for-profit orientation. Perhaps healthcare could consider its place as an export industry.

In the meantime, healthcare must pay its own bills. Almost immediately I would establish a large force of accountants, to throw some light on central issues which are now conducted as either state secrets or business secrets. For example, I do not understand why hospital prices to uninsured patients are allowed to be so unrelated to their underlying costs, frustrating efforts to becoming market-oriented. Insurance companies have almost figured out how to protect themselves, but the public is appalled and unprotected. Perhaps the motive is the protection of indirect overhead as a mechanism of cost-shifting. The only self-interested group I can think of are the trial lawyers, who base their settlements on "seven times the medicals". But I have trouble believing they have enough power over the state legislatures and judiciary to whip-saw the public into tolerating sixty thousand dollar charges which the insurance companies discount to six thousand, and the hospitals write off willingly, knowing their real inpatient cost is limited by the DRG7 . And which I suspect cost the drug company much less than that to manufacture. If sunlight is the best disinfectant, let's have a lot of it.

7 Here, I am quoting my own personal health insurance EOB.

And I would like to see the true costs after taxes, of health insurance to various clients, particularly big business. It is possible to see two tax deductions, one to the employer and one to the employee, leading apparently to incentives for business to prefer high taxes on itself. There must be more to it than that, and one component is overworked congressmen. Each one of them has a million constituents and spends a lot of time on the phone.

Estimated Cost of Medicare Buy-out.

The data in this analysis are usually condensed to a single average individual, in anticipation of use in individual Health Savings (and Retirement) Accounts. The number of individuals enrolled in Medicare (55,504,005), obtained from the Internet figures published for 2015 by Centers for Medicare & Medicaid Services (CMS), is divided into the total 2014 budget of the program ($373 billion) =$6,720 Medicare cost per person per year for 20 years average life expectancy at 65, or restated as roughly $134,400 per person per Medicare lifetime. As a check, Medicare cost is said to be half of lifetime medical costs of $350,000 or $175, 000. This is the ballpark guess of a Medicare buyout cost.

The accuracy of this figure would improve by excluding 9 million recipients under the age of 65 eligible by reason of disability, not by age. However, figures were not provided separately to distinguish costs of the disabled from the 46 million who became beneficiaries by attaining age 65. Assumed here to be roughly the same, disability probably costs more.

An important side point is not the exact amount of cost, but rather that this composition includes everyone, rich or poor, who is over 65. These figures therefore definitely include the indigents, and if anything includes a disproportionate number of them. This should be attractive to single-payer enthusiasts, who wish to extend Medicare coverage to everyone, and who find that employing extra eligibility based on age rather than income simplifies their task of persuasion. The calculations which follow are therefore not a "rich man's proposal" at all because Medicare is not a rich man's proposal. Its fault is being funded by pay-as-you-go for several decades.

A second non-obvious fact of regional variability is readily inferred by comparing $6,720 per year (the overall Medicare average cost) with the $18,000 annual premium for New York employer health plans cited in the newspapers. Quite obviously New York prices are higher than those in a small rural district, but not three times as high. The New York employees, like the Medicare beneficiaries, must include a large subsidy of non-paying clients by paying ones, because of New York laws and demography, and are thus "conservative" estimates of revenue requirements in a limited sense. That the average is an overestimate is still further suggested by the $5700 annual premium charged for Medicare Part F for the highest income bracket, including a profit margin for the contractor. (Medicare sells coverage at wholesale prices and allows the Part F vendors to discount them, while still making a profit by constraining provider reimbursement.) Counting on the tacit reinsurance of the "risk corridor", some contractors lost money, but evidently, not all of them did. It should be comparatively easy to calculate a more accurate estimate of costs by excluding under-65 beneficiaries, but the prediction-- that wholesale is less than $6720-- will almost certainly be borne out.

The underlying principle of our proposed example is equality with the present Medicare public charges. That's a rhetorical boundary, not an actual one. And an example, not a rule. Congress can set these charges where ever it pleases. Eventually, the charges must reflect an underlying need for them. For the present, current charges will suffice as an example.

Current revenue comes in two approximately equal components. The first comes from the withholding tax on the employment of people aged 25-65, the amount of which is specified as 2.9% of the wages, 1.45% from the individual paycheck, and 1.45% from the employer. From the Medicare annual report, it comes to an average of $700 per year for 40 years = $28,000. The other half, comes from the Medicare premiums (applied to reduce the benefits of Social Security checks by $1400 a year) from age 65 to the date of death, on average of 20 years = another $28,000. Some accountant should explain attributing all this to Medicare Part B, and why Part A is not attributed to hospitals at all. It seems entirely arbitrary, but you never know. Since contributions to Health Savings Accounts are tax-deductible, their value increases. We use the 18% flat-tax figure offered as a weighted average of real income taxes, so the HSA annual deposits are effective $825 and $1650, ignoring the attributions. That's another example; does it make sense? To repeat, total average lifetime spending by Medicare is $6720 per year times 20 years = $134,000 ($2580/yr left for retirement). Is this a reasonable price for a buy-out? I doubt it, but even this figure is surmountable.

Investment Income.

There are two ways to increase investment income: invest the same at a younger age, or invest more over a longer period of time. Depending on the sequence of investing, these previously collected sums should produce enough to buy out Medicare, with some left over to finance a moderate retirement -- without increased debts to bondholders. Of course, you can't spend it twice, but this is in an escrow account. The alternative is to have a separate account, and it is hard to say which is simpler. Although its internal distributions and cross-subsidies remain perhaps a little unclear, the over-65 retirement benefits are likely overestimated. The reader must weigh these approximations for himself, keeping one principle foremost. If the approximation does not comfortably dwarf the projection, it is too close for comfort.

Transition Costs.

In any event, the entire Medicare program must at most borrow $78,000 times 55,500,000 beneficiaries divided by 20 years of the collection -- each year -- or roughly $200 billion of a 373 billion budget -- for a transition from one system to a better one. Moreover, alternative proposals for transition are suggested in later chapters. Even Wall Street veterans concede the U.S. Treasury has an outstanding reputation in bond issues for wars and depressions, so their advice is worth consulting. Once again, that figure is high to the degree the 9 million disabled on average have different costs than the 46 million elderly, low to the degree inflation becomes rampant and largely unpredictable in sum. As far as frightened investors are concerned, the escrow rules should actually lessen the risk of panic illiquidity. Whether they see it that way or not, they are investing in the durability of the entire long-term American economy by investing in total stock market indices. It is safe to respond the Medicare deficit is made wildly unsustainable by pay-as-you-go financing, whether privatized, or not, and the proposed approach might actually help quite a bit. Furthermore, starting with a contingency fund of only $100 yearly from birth calculates to reach the same number with a little margin of error built in. But it includes no pension, and probably less elimination of debt.

The Health Saving Account law already provides potential revenue relief to the beneficiaries in two main ways: a tax deduction for deposits, and the freedom to invest deposits, tax-free, in any way agreeable. Furthermore, it collects no tax on withdrawals for medical purposes. (There is one exception which might be changed: taxes are not collected from medical expenditures, but are collected on HSA to IRA conversions.) Medicare, on the other hand, collects an average of $700 yearly for 40 years (25-65) in salary withholding, and $1400 a year for 20 years in Medicare premiums. That's about half its cost, another half is a deficit which must be borrowed. Using the rule of thumb of 7% doubling in ten years, an investment return of 7% in an HSA of the same money which resulted in an unsustainable deficit by pay/go, would be worth $825 per year (700 from Medicare + $126 extra value from the tax exemption) or $185,000 at age 65 (or even more, if Medicare costs are substantially less than $6720 for over-65 subscribers). That would buy out Medicare with $161,459 or so to spare, which after taxes would be $132,396 by subtracting the real average income tax of 18%. That much ought to provide approximately $8,000 annual retirement for the 20 years 65-85, the average life expectancy at birth. We can do better than that, but some money will have to be spent on the transition costs, so it's a maximum, providing an individual HSA can earn 7% on its investment. And worse than that if it can't. At least, we can sustain the point that, alone among health insurance programs, HSAs transform any surplus after taxes into retirement income in the form of Individual Retirement Accounts, (IRA), and provides a way to generate potential surplus which is then likely to be under-stated.

Future Projections.

It will be noticed we have left two features of HSAs unexploited, and one fly in the ointment unmentioned. In the first place, you have to be younger than 25 to take full advantage of what we have described, and you have to wait 40 more years to receive any tangible benefits. Secondly, the $3400 per year voluntary contribution deposit limit to HSA is unused and we suggest those who can afford it, spend it all on looming transition costs. These are both artificial limitations on the simplified example, a standing invitation to exploit new features. For example, look at what a hundred or so dollars would do, if deposited at birth. Of course, you can add more, up to a $3400 limit per year. The result would probably be in the millions. Age and contribution limits should first be expanded, and the employment requirement eliminated. By expanding the time available for recovery, these features would actually enhance the safety of the investment.

An important stickler is, the law requires a high-deductible Catastrophic backup throughout the life of the HSA but makes scant mention of overlapping insurance. When age and employment limits are removed, this needs to be addressed by "or it is equivalent", and possibly then more specifically, since it leaves the HSA at the mercy of ambiguities its competitors can manipulate.

The Importance of Solving Childhood Health Costs.

The HSA act was originally intended to appeal to young people, whose current medical needs were often small ones, and whose distant retirement looked utterly remote. Since the Affordable Care Act has considerably confused matters for this age group, we have decided to avoid discussion of the working age group, at least until after the 2016 Presidential elections give us some indication of where things are likely to go. The proposal does cover Medicare, retirement, and newborns -- substantially everything except what Obamacare covers. When the future of ACA becomes clearer, it will be necessary to write another book to integrate the issue of filling in such gaps. The HSA law does permit individuals to contribute up to $3400 annually to the accounts, but solvency seems temporary unless the age group 25-65 can produce a surplus to subsidize other groups. Here, we only contrast the HSA with about 100 to 250 dollars per year, from birth. That makes calculation easy, and small, but it grows relatively slowly and well might have to be supplemented at later ages, particularly during the transition period.

Let's be frank about projections so far in advance. They are meant only to distinguish areas where funds are "comfortably adequate" from other numbers which are more dangerously "barely adequate". The imagined precision is retained as a way of following particular numbers through the argument. Combining the two, a modest sum of $100 added to the account at birth would slowly grow to $542 (at 7%) by age 25. But that relatively small amount would be added to the withholding tax contribution at age 25, and even singly might grow to be $8,127 by 65. By age 84, without any withdrawals, it would grow (single) to $31,450. Combined with $825 a year , however, starting at age 25 with $825 (withholding tax equivalent) added yearly, it would grow to $180,500 at 65, and subsequently without withholding tax addition, might re-grow (without withdrawals for retirement) to $166,000 by age 85. It would be in an escrow fund for Medicare, so it could not be spent for other purposes until age 65. There would be a withdrawal, however, to buy out Medicare at age 65, of $134,000. So the reduced value to $41,500 might re-grow in 20 years (age 65-85) to be $166,000. Using it up for retirement, however, would provide an annual pension too small to depend on for distant calculations. For the retirement to be a real incentive, it would have to last as long as projected longevity. The person would have to supplement it with private savings and investment along the way, a normal situation for rich people, but a frightening novelty for most people. Another step is required before such a lifetime venture is undertaken. Therefore, what is presented is not the best case. It will be improved upon later, especially with cutting the numbers in half with the Last Years of Life approach.

That venture would be a rearrangement of revenue streams. The earlier a sum is invested, the larger it will grow; in meaningful terms, it is worth far more for retirement if you start the same contribution twenty to sixty years earlier. In order to get the most out of the same investment, Congress might remove the employment requirement and extend the age limits, both up to death and down to birth. Maximum revenue could then be extracted if the financial equivalent of the Medicare premium began at birth, and twenty years later was followed by forty years of payroll deduction. Essentially, this means moving the equivalent of Medicare premiums to some form of a trust fund, and re-directing the payroll deductions to the Health Savings Account as described. These sixty years of investment leave five years unfunded, so if the Medicare age limits remain fixed at 65, the trust fund could begin as late as the fifth birthday.

Such rearrangement of the payment stream would generate a fund in the millions at age 65, allowing for much more comfortable slack in the investments, and supplying ample funds to transfer to a grandchild. Funding the health needs of children is the greatest single unaddressed problem in all health planning. It gets little attention because it seems so impossible to pre-fund the birth costs of a newborn. And it even stimulates misinformation, such as the mistaken belief that health costs of newborns are trivial. They are not as great as the cost of dying, but the financial resources of the parents are more strained. They spill over into grievances about the higher health premiums of women, in an era where feminism has had hidden effects on politics and even Supreme Court decisions. The transfer of Medicare premium funds to grandchildren has synergy with grandparent funding of obstetrics, and the Health Savings Account provides a vehicle to do them both. This particular caper is a political issue, requiring Presidential leadership; it is definitely not a financial issue just waiting around at the end of a lifetime of compound interest.

Any greater results would have to rely on additional donations to the contingency fund, or spectacular good luck in the stock market. This is therefore just about the limit of what a favorable arrangement could provide; except for outside contributions, any other arrangement would provide less. For example, switching Medicare premiums to fall between 65 and 85 (the current arrangement) would reduce the benefits by a third, as might a bad decade in the stock market. Suffering both reverses at once might wipe out all retirement benefits. But short of an atom bomb attack, it is hard to see how the average investor would lose money by doing it. Most of this plan requires changes in present law, although the growing national debt will create pressure to do something major to change the future trajectory of Medicare. The finance industry, already under strain, may resist. Consequently, the opportunities and drawbacks of Medicare Part F for employers also ought to be scrutinized, particularly as long as large employers recoup so much from their various unshared tax exemptions.

Long periods of unemployment could destroy this tempting dream, as could a protracted poor investment experience. The ordinary person would have trouble finding an institution which would accept amounts smaller than $100 without long-term contracts allowing the institution to trade short-term losses for long-term benefits. In the next section, we must, therefore, take a moment to reflect on the finance industry. And then we will return to this lifetime plan, by adding the Ends of Life approaches.

Originally published: Saturday, October 29, 2016; most-recently modified: Thursday, June 06, 2019