urban institute nonprofit social and economic policy research

Q&A with C. Eugene Steuerle: Estimating Social Security and Medicare Taxes and Benefits over a Lifetime

The estimates you and Stephanie Rennane provided for the lifetime value of Social Security and Medicare benefits and taxes have been reported in a number of newspaper articles. In addition to what you provide in the introduction and notes to those estimates, can you walk us through some of the calculations here? How did you reach these numbers?

Why a 2 percent rate of return?

Your estimates were broken down not only by earnings levels but by different generations. How did you determine the real value of benefits and taxes for different cohorts?

Do your tax estimates take into consideration the matching Medicare contributions made by employers?

Do your benefit estimates account for Medicare premiums paid by most beneficiaries?

How about for very high earners?

Why do the estimates differ for men and women?

How did you adjust for life expectancy?

Why are these “expected” benefits?

Why is there such a big gap between taxes paid and benefits received?


The estimates you and Stephanie Rennane provided for the lifetime value of Social Security and Medicare benefits and taxes [see the estimates here] have been reported in a number of newspaper articles. In addition to what you provide in the introduction and notes to those estimates, can you walk us through some of the calculations here? How did you reach these numbers?

Sure. We used the example of a typical worker, employed every year from age 22 to 64, who retires at age 65. We calculated the lifetime value of taxes and benefits on the basis of what’s called “present value,” allowing us to compare the value of money paid or received in one year with that of money paid or received in a different year.

Let’s start with taxes. One dollar paid in taxes today doesn’t buy what $1 paid in taxes might buy next year. Further, we assume that the $1 paid today is put into something like a savings account that grows over time. The rate of interest we applied is the rate of inflation plus 2 percent.

A few people commented that the money they paid in taxes would be worth a lot more had they been able to stash it in a savings account. That’s why we adjusted for inflation plus 2 percent—about what a person could have realized in savings.

We also received a few questions from people who said they didn’t pay as much in taxes as we estimated—and that’s correct if they’re only adding together the actual dollar amounts (the nominal value) of taxes paid.

We made similar adjustments to the calculation of benefits. How do you compare a dollar in benefits today to a dollar in benefits tomorrow? If that money hadn’t been paid out to a beneficiary, it would be accumulating value. So, to calculate the lifetime value of benefits, we “discount” back the value of the benefit by the same 2 percent real interest rate. So, if we assume inflation is, say, 5 percent, then $1 in benefits next year would be equivalent to, roughly, 93 cents today (after deducting inflation and a 2 percent real interest rate). You would need to have about 93 cents in a retirement account today to cover that $1 dollar worth of benefits tomorrow. At a zero inflation rate, you would need about 98 cents.

Back to the top.


Why a 2 percent rate of return?

Well, that’s a very fair question. There’s a big debate among economists over what rate to use when doing these types of calculations, and it can vary by the type of program and the type of risk protection provided. We used a 2 percent real rate because it’s as good a rate of return as one can expect from a private annuity and not far from a long-term real rate of return on a fairly protected mix of government securities. In fact, given all the protections in Social Security against default and inflation, the interest rate might be a bit high for such low-risk policies.

If you use a much higher rate to discount benefits and perform comparisons over time, you create a problem because, in developed economies, per capita personal incomes rarely grow by much more than that over long periods. And you don’t want a much lower rate of return because you want to take some account of economic growth. So 2 percent real return seems to us to be a reasonable compromise. The Congressional Budget Office has used a 3 percent rate of return for some of its calculations—a bit high, but in the same ballpark.

Back to the top.


Your estimates were broken down not only by earnings levels but by different generations. How did you determine the real value of benefits and taxes for different cohorts?

We calculated, over time, the present value of benefits and taxes for each cohort at age 65. We adjusted the amounts to reflect their real value in 2010 dollars. This way, you can compare the package of benefits and taxes for someone retiring in 1960 with the benefits and taxes of someone retiring in 2030. You’ll see that the real value has gone up over time partly because we’ve become a richer society, partly because benefits often grew even faster than incomes.

Back to the top.


Do your tax estimates take into consideration the matching Medicare contributions made by employers?

Yes. The tax calculations assume that the individual pays both the employee and employer portion of the tax. Most economists believe that workers essentially bear the burden of this tax since the more employers pay, the less cash wages workers receive.

Back to the top.


Do your benefit estimates account for Medicare premiums paid by most beneficiaries?

Yes. We calculated Medicare benefits net of premiums. That means the average government benefit is reduced by the cost of the premiums people pay.

Back to the top.


How about for very high earners?

We have not performed calculations for very high earners. The cap on the Medicare tax was repealed for very high earners after 1993. A high earner turning 65 and receiving Medicare in 2010, for instance, would have potentially paid that higher Medicare tax for the past 17 years (Social Security Administration table).

A fairly small percentage of beneficiaries are also subject to a high-income additional premium. We have not performed that calculation either.

Note, however, that these caveats would have little effect on most of the examples presented in our tables, since we provided calculations only at levels at or below the Social Security maximum wage.

Finally, another caveat. For average wage, we use the definition provided by Social Security—the average wage of those in the work force. The lifetime wages of the average earner is below the lifetime wages of a worker who always worked at the average wage. That’s because many workers do not work every year nor pay tax every year. In that sense, we overstate the taxes paid by the “average” worker.

Back to the top.


Why do the estimates differ for men and women?

Mostly because women outlive men and can expect to receive benefits longer. However, we didn’t take wage disparities into account. The typical woman earns less than the typical man. So even though the average man has a shorter life expectancy, he might still end up with higher lifetime benefits.

Back to the top.


How did you adjust for life expectancy?

We used actuarial tables to adjust for the probability of living in all years after age 65. So, say the benefit is $20,000 a year. We assume that at age 65, 100 percent of the cohort having lived to 65 gets $20,000. At age 66, say, 98 percent of the group is still alive and receives the $20,000 benefit, and at age 67, 96 percent receives it. Roughly speaking, we multiply the benefit each year by the expected probability of being alive, convert to present value, and add the numbers up for each year to determine the expected lifetime value of benefits. This is similar to what an insurance company would do in determining how much to charge up front for an annuity that would stretch over one’s future life.

Back to the top.


Why are these “expected” benefits?

Because we’re trying to figure out what happens for an average person with an average life expectancy. Total lifetime benefits depend largely on how long a person lives. With Medicare, you have the additional variable of health. Sick people receive more health services than healthy people. In that sense, what we do is similar to valuing an insurance policy for an average employee. Even if an employer is paying an average of $10,000 per employee for health insurance, it’s worth a lot more for the very sick employee than for the fairly healthy one.

Back to the top.


Why is there such a big gap between taxes paid and benefits received?

I talk more about that in my recent Government You Deserve column. Setting Social Security aside for the moment, Medicare is facing a serious financing problem. You can perform these estimates a lot of different ways and still come to the same basic conclusion: Medicare taxes are far from enough to cover Medicare benefits. In fact, Medicare taxes support only slightly more than half the total cost of benefits. The amount pulled in from Medicare taxes has never been adequate, and the system is scheduled to rely increasingly on funding from general revenues.

Some argue that, at some level, all the benefits are covered by taxes—and that’s true. In the grand scheme of things, total taxes over time cover total benefits (although taxes can be collected not just directly, but indirectly, such as taxing bondholders through unexpected inflation or defaults). But what we’re trying to point out through these numbers is that relying on deficits and income tax revenues to make up the gap in Medicare funding means cutting other things that government does or leaving heavy burdens for future generations to pay. We’re all responsible for dealing with our large future projected deficits both as a whole and within our social welfare systems; we’re not “entitled” by any reasonable calculation to leave the costs of fixing them to younger generations.

Back to the top.

Email this Page