• Health Insurance in Retirement, Part I: How Much Health Care Will You Use?

    This post originally appeared on The Finance Buff and has been cited in the Carnival of Personal Finance #205.

    I’ve been asked several times to provide guidance on retiree health costs and insurance (e.g., Jim’s request). These are complicated issues, and a proper approach includes an assessment of one’s expected health care requirements, the extent to which one needs to insure against health risks, and one’s tastes for various health insurance design features. I cannot fully address these issues, but I can suggest ways of thinking about the problems. This and a subsequent post are introductory guides to the issues, not the final word.

    Total health care spending is the sum of your out-of-pocket costs and those spent on your behalf by your insurance carrier. One should expect total health spending to grow in real terms as one ages. The first source of higher spending is health care inflation. Health care costs and health insurance premiums are likely to continue to rise faster than general inflation, as they have for decades. According to a Kaiser Family Foundation fact sheet, health spending has grown about 2.4 percentage points faster than GDP since 1970. The June 2009 Council of Economic Advisers report titled “The Economic Case for Health Reform” provides 10-year time series of medical and overall inflation that show the former above the latter for all but a brief period in the mid 1990s (see Figure 16 on page 31).

    The second source of higher total spending on health care is that one is likely to require more health care later in life. According to a 2007 study published in Health Affairs, compared to the working-age population (19-64 years olds) twice as much is spent on health care for individuals 65-74 years of age, four times as much is spent for those 75-84, and six times as much for individuals 85 years old or older.

    Some of the additional spending on retirees as compared to workers could be due to near universal or more complete coverage for the elderly (a moral hazard effect). On the other hand, this tendency for the insured to use more services is offset somewhat by the fact that Medicare pays lower prices than many insurance policies for working-age individuals.

    Nevertheless, by putting the inflation figures together with the age-based multipliers indicated above one can estimate future expected total health care costs based on one’s current total costs. Here’s a spreadsheet that does so. To use it, you will first need an estimate of your current total health care costs for the year. You can obtain it if you have all your explanation of benefits (EOB) statements from your insurance company for the past year. Or, likely your insurance carrier can tell you what they’ve spent on you over the last year and what your out-of-pocket liability was. (Because health care providers charge different prices depending on insurance coverage your annual figure is different from what it would be if you were uninsured or insured by another company. It is a very rough guide to total expenses.)

    Since there is a large variance in health care costs I wouldn’t put too much faith in the estimate you get from the spreadsheet. But it is, at least, a place to start. Indeed, predicting overall health care expenditures is a hard problem. Statistical models developed by scholars have relatively low predictive power. Predicting ten percent of the variation in expenditure is considered good (e.g., Medicare Advantage’s risk adjustment model). That means ninety percent of the variation is unexplained by the model or chalked up to random error. An individual ought to be a better predictor of his or her health expenditures than a model that cannot include measures unobservable to the researcher. (How much better? I don’t know.)

    Expenses for some specific services are more predictable. Drug expenses, for example, are persistent because individuals tend to use the same medications year after year. The best statistical models of drug spending can predict about 55% of the variation in next year’s drug expenses, leaving 45% to random error.

    Aside: The relative predictability of drug utilization led Charles Kahn in 2000, then president of the Health Insurance Association of America (now part of America’s Health Insurance Plans), to say that stand-alone drug coverage would be like “providing insurance for haircuts.” As with haircuts, most people know how much of which medications they will use, making drug spending challenging to insure on its own. (Yet, today many insurers offer stand-alone drug coverage through Medicare.)

    After getting a rough sense as to how much more health care one might use in the future, the next question is how much of that one should insure against. More importantly, since one is unlikely to predict one’s health spending with great accuracy, one should consider carefully how much of the spending above that estimated amount one should insure against.

    This is, of course, a very complicated question. It matters a great deal how your health expenses are distributed over services. If you use no prescription drugs, for example, you may not want to buy coverage for drugs. Also of relevance is one’s willingness, ability, and need for risk. A very wealthy individual might decide to self-insure against much of their health care needs. (I know one very wise and wealthy health economist who has decided he will not purchase Medicare Part B coverage, opting to self-insure for the costs of outpatient care.) Most individuals will want some coverage for hospital, outpatient care, and prescription drugs (there’s also long-term care to consider). How much of each is a personal choice.

    Of course, one’s choice should also depend on one’s options and their costs. I recently summarized some of options retirees might have. In my next post on retiree health insurance I’ll say a bit more about how to think about those options.

    Here are links to online calculators related to the issues discussed above:

    Comments closed
    • One of the major challenges of retirement planning is determining health care expenses. Fidelity has done a good job of compiling the total cost for a “average” 65 year old at $551 per month. However, your mileage will vary significantly. More info, including a cost calculator at

    • TimK,

      Fidelity’s estimate on the page for which you provided a URL is interesting. Its prices are a few years out of date and it does not suggest how to inflate them for a future date. The estimate is based on an assumption of enrollment into a particular configuration of insurance products (FFS Medicare A & B, Medigap F, and a PDP). That’s one way of doing things but not the only way and not necessarily the least expensive. My next post will say more about this.

      Having said all that, estimating retirement health care costs is extremely difficult. Fidelity’s approach is not without merit. It would be interesting to see it expanded to provide average costs for a variety of other insurance configurations and suggest how they might inflate into the future (based on past trends, for example). In principal I could do all that work and one day I might. But it is a lot of work. (I wish I had an assistant!)