The “future of retirement” seems very much on the public’s mind. The topic is surfacing in the press, in the institutional marketplace, and will be the focus of a Washington, D.C., policy forum I’m attending in May. Here are some of my thoughts. I welcome your comments and will try to incorporate them in my upcoming discussions.

It’s very clear that there is a future for retirement. But to me, it’s really a tale of two futures: the future of retirement income, and the future of retirement health benefits.

Retirement income

Financing a retirement income is essentially a problem of compound interest. The income you generate (whether you’re participating in a traditional pension plan or are contributing to a 401(k)-type plan) is a function of how much you or your employer contributes, over how long a period these contributions are made (namely, how long you work), plus investment earnings.

For a brief moment in time, a series of exceptional factors made these compounding calculations relatively effortless. An expanding workforce made it easy to finance traditional defined benefit (DB) pensions, as well as expanded Social Security benefits. For both DB and 401(k) plans, a spectacular market in equities and bonds in the 1980s and 1990s acted as a rising tide that lifted all boats. At the same time, the age of retirement continued to fall. These factors partially hid the fact that, as a result of gains in longevity, Americans were actually living longer, and so would need more money to finance their retirement standard of living. But it was easy to overlook this fact in the buoyant years.

Now, the demographics are reversing, and we have been reminded that equity or bond markets don’t always generate high double-digit returns. As a result, the trend of increased longevity has become ever more apparent. It is clear that we need more money to finance that longer period of leisure we call retirement.

When you look across the U.S. economy, there are retirement systems that have generally gotten the compounding interest calculations right, despite the seductive demographic and market trends. These are the well-funded private- and public-sector DB pensions, as well as 401(k)-type plans with high average balances. Then there are the laggards—poorly funded DB plans, both corporate- and public-sector, as well as 401(k) plans with low balances. The main difference, it seems to me, is that the successful plans (and successful households) never forget the importance of ongoing and strong contributions.

This is what the angst is all about when we talk about the “future of retirement.” The future of retirement income in America is really about a psychological adjustment to a more realistic set of compound interest calculations. For those with poorly funded retirement plans, the required adjustment can be summed up with just one word: “more.” More resources are needed in retirement—whether from saving at a higher rate, working longer, or a combination of the two.

Retirement health benefits

Retirement health benefits are not readily addressed by the mathematics of compound interest. First, there is the demographic headwind in health care: Unlike retirement income, which is generally lower than during the working years, health costs increase substantially as we age. In other words, you generally spend more on health care in your 60s than in your 50s, and in your 70s than in your 60s, and so forth.

Second, due to a boom in medical technology, health care costs are rising rapidly across the entire economy, both for workers and retirees, at rates like 8–12% per year. If you’re setting aside money for retirement health expenses, there is no natural investment vehicle for your savings that steadily earns 8–12% per year. (If there is, I’ll take two!) Given rapid health care inflation, your existing savings fall behind when they fail to grow at a high rate each year.

Today, there is much discussion about how retirement health benefits should be financed, and how budget reductions might affect Medicare (covering hospital, doctor, and drug costs) and Medicaid (covering long-term nursing care). While the financing issue is important, the real issue to me is the underlying demand for medical care itself. The increase in medical costs starts in the doctor’s office or in the hospital, when your doctor’s evaluation leads to a series of new tests, procedures, drugs, or other treatments for your condition.

Slowing the growth in retirement health care costs is, in the end, about slowing the growth in medical treatments and procedures. When they prepare long-term forecasts of retirement health costs, government and private economists implicitly assume a certain number of medical treatments for each of us from the time we retire to the time we depart this vale of tears. Part of “bending the cost curve” is about slowing this implicit growth rate—i.e., using fewer medical services than projected.

By the way, this doesn’t actually mean spending fewer dollars on health care. Even if we reduce the number of medical interventions per person, the aging of baby boomers means that health spending in total dollars will likely continue to rise. Getting by on less also doesn’t mean the end of technological innovation either. As I described it once to a health policy analyst, it’s about a world where instead of 10 blockbuster medical treatments in the coming decade, we discover only 5—and are happy with that.

Choosing the future

So is there a future for retirement?

The answer is most definitely yes—if we save more and work longer, and accept fewer medical interventions in the latter years of life. The answer is likely no, however, if we continue to believe that retirement means an “early out” package in your mid-50s, along with health benefits to pay for pretty much any new medical technology that comes down the road.

The latter option being financially unsustainable, it seems clear that the former is the only path forward.