Americans Live Longer and Work Less

Today I saw an article on Hacker News entitled, “America’s CEOs Want You to Work Until You’re 70”. I was particularly surprised by this article appearing out of the blue because I take it for granted that America will eventually have to raise the retirement age to avoid bankruptcy. After reading the article, I wasn’t able to figure out why the story had been run at all. So I decided to do some basic fact-checking.

I tracked down some time series data about life expectancies in the U.S. from Berkeley and then found some time series data about the average age at retirement from the OECD. Plotting just these two bits of information, as shown below, makes it clear that Americans are spending a larger proportion of their life in retirement.


Perhaps I’m just naive, but it seems obvious to me that we can’t afford to take on several additional years of retirement pension liabilities for every living American. If Americans are living longer, we will need them to work longer in order to pay our bills.

23 responses to “Americans Live Longer and Work Less”

  1. JVM

    It is true that social security is becoming slightly more expensive, it’s important to note that the lion’s share of the entitlements crisis comes from the projected growth of Medicare costs:

    The problem with raising the retirement age is that it will be burdensome on the working poor (consider blue collar workers doing backbreaking work), but it won’t actually be a huge help with Medicare spending, because most health expenses are incurred later in life. Unfortunately, the actual solution will likely involve massively scaling down the services offered on Medicare.

  2. George

    Life expectancy for people with more education and higher incomes is going up a lot, but life expectancy for the poor and uneducated is not nearly as high and in some cases is going down. The people above the median income are exactly those people who can afford to fund their longer years of retirement. I believe the life expectancy gap between the affluent and the poor is growing in the USA and that is partly driving the over all mean increase.

  3. meh

    A comment on your data:
    You’ve plotted life expectancy from age 0 (I believe). Try plotting the life expectancy from age 65, when it matters. The numbers will be different. Between 1950 and 2009, life expectancy from birth increased 10.3 years. However, life expectancy from reaching 65 only increased 5.3 years.

    See Table 22 of

    A comment on your reasoning:
    There’s quite a bit of literature out there, and quite a few academics and think tanks have looked at this in far more depth than your three plots. See note 1 for articles from EPI (a liberal think tank), each with academic references (if you care to, you can easily find your own conservative ones). It’s particularly enlightening to look at how long this very debate has been going. See note 2 for an EPI briefing paper from 2000 on raising the retirement age.

    It’s easy for the young and affluent to propose working longer. The affluent don’t need Social Security, and the young haven’t had a lifetime of backbreaking work (if poor), nor have they faced being laid off from a professional job in their 50s or 60s (if middle class). The reason that the CEO proposal is particularly noxious is because it’s extremely one-sided. There are no proposed sacrifices for higher-income individuals who don’t need Social Security.

    Some obvious one-sided alternatives that wouldn’t touch the retirement age:
    (1) enact a wealth-tax that funds social security and medicare (or just raise the top brackets to higher than the Clinton era).
    (2) Increase the FICA withholding cap. Social security is currently only withheld on the first ~$100K of wages.
    (3) Increase FICA withholding.

    All of these have drawbacks, but they’re different drawbacks than raising the retirement age. There are multiple levers to pull here. It’s a mistake to ignore any of them, and it’s all too convenient for CEOs to propose the one that affects them the least.


  4. Michael Waskom

    Just to back up these two points with some data:

    Although I’m not sure “life expectancy” is the right way to think about it since I think most of the big gains that we see on your graph are from decreased rates of death early in life (from disease, etc.). The trend we should probably care about is life expectancy conditional on reaching the retirement age.

    Also raising the _Medicare_ eligibility age is particularly self-defeating, since it removes relatively healthy individuals from that risk pool. And, if those people do run into health trouble, they’re either going to (a) die tragically or (b) wait until they can get Medicare coverage and then be dealing with a problem that will in all probability be more expensive to treat then if they’d just had proper health coverage to begin with.

  5. Robert Young

    meh and Mr. Waskom bring some reason to the data. The whole “we’re living too long for Social Security because it was setup for age 65 life expectancy”, as if the designers of SS intended for no one to live past 65, is sophistry. If one goes back to 1935, and run forward to today, one finds that life expectancy at birth vis-a-vis life expectancy at 65 is even more skew than from 1950. The reality is that what’s pushed life expectancy at 65 is the reduction of early death. Antibiotics and vaccines are the largest part of that. Reduction in smoking, too. The increase in obesity may soon offset those gains.

    Mostly, there’s the issue of unintended consequences. Due to de-industrialization, more of the US economy is service based, and health services are a larger part of that. The advances in US medicine, post WWII in particular, has been funded by employer based health insurance and Medicare. Without that funding, sawbones would live hand to mouth, just as they did before then. On one hand, the US pays more for so-so health care; on the other, US corporations in the health business rake in fat profits from this waste. Reduce Medicare, and you clobber both employment and profits. The cash for SS and Medicare is largely a pass through to for-profit corporations. SS cash is spent on consumption. Medicare on for-profit HMOs, pharma, and hospitals. In all, a wealth transfer from the many to the few. As it always has been in the USofA.

    Besides, what’s wrong with health care being the major discretionary expenditure? Have we all forgotten what Your Good Mother told when you were growing up: “if you’ve got your health, you’ve got everything”? Kiddies, who haven’t got any sense yet, would rather have a new iPhone, of course. Yet another reason to not let them make important decisions

  6. JVM

    In support of JMW’s views, here are some great infographics from Nate Silver:

    1. meh

      @JVM: There’s no question that entitlement programs are driving spending. The argument is whether raising the retirement age, alone, is the answer. More specifically, JMW referenced the CEO proposal to raise the retirement age. That particular proposal is especially vexing because it’s being pushed by people who won’t suffer at all from its consequences. The Ezra Klein article that @MichaelWaskom referenced does a fine job of stating this.

      Raising the retirement age may have to happen, but it should be paired with an increase in revenue (saying, eliminating the FICA cap) to make the whole thing more fair.

      1. JVM

        If you read my earlier comment you can see that you are preaching to the converted. But the 538 analysis does show that SS has been growing a lot recently, although I think it’s supposed to level off pretty soon while Medicare continues to soar.

  7. Tal Galili

    Question: does anyone know why around 1950 the variance of the lines appear to diminish significantly?

    1. David Mangen

      I suspect that the variance decreases around 1950 because of more widespread use of antibiotics, with the resulting decrease in the incidence of pandemics and the like.

  8. Stephen

    If we were an agrarian society that lived hand to mouth then yes we would have to work longer. Since we are not then we should just all be paid more for the period that we can work – no net winners or losers.

  9. Robert Young


    It is in the post WWII period, funded by employer based insurance (and funding from newly minted National Institutes of Health, et al) thence by Medicare later, that early- and mid-life diseases were dealt with. Again, vaccines and antibiotics, the latter spurred by the war effort, were the main reason. The simple fact is: there was little professional medicine available to citizens before post WWII. WWII materiel production spawned The Great Migration (look it up), which moved millions from rural medicine (largely an oxymoron) to urban medicine. History can teach us a lot, if only we read it.

    The emphasis on the provision of insurance to workers starting with WWII materiel production is vital to understanding the change. Due to wage/price controls during the war, corporations added insurance coverage to circumvent the controls. Whether they expected to jettison the insurance coverage once the war ended, may or may not be documented. In the event, health insurance from corporations became the rule. Once there was a steady, and much larger, flow of funds available, medicines and medical procedures which were only science fiction emerged.

    Does anyone rationally believe that heart bypass or heart transplant surgery would have been developed if health care had continued to be bought as bread and eggs? It would not have been possible; not even the wealthy could have afforded to fund the R&D, or even the unit cost that the small number of procedures that the handful of wealthy individuals could afford. It’s a vicious circle: with too few units to spread cost, not even the wealthy can afford it (even the rich need Obamacare). Unit cost is driven down by increased production. With direct payment for health care, only the very wealthy could fund the entire gamut of R&D and production, yet without widespread production, not even the moderately wealthy could afford it.

    The irony? The “democracy” the US imposed on the Axis powers, far more democratic than what we have here, was in no small measure why those countries have single-payer (and lower cost be far) system; which by all measures, does a much more effective job as a public health effort. If you run similar curves for Europe, be prepared to cry.

  10. Nicholas Dronen

    Yes, taking it for granted that the US will have to raise the retirement age to avoid bankruptcy is indeed naive in very many ways.

  11. Vincent Thomas

    This meme of “Americans are living longer” drives me nuts. I believe people infer from it that individuals are living longer and therefore we could reasonably raise the retirement age. As others have pointed out in reality it is that fewer Americans are dying younger. I don’t think anyone has suggested any changes to the aging process so why would it be reasonable to raise the retirement age?

    1. meh

      @VincentThomas: This is accounted for if you look at the conditional life expectancy from age 65, which is the average number of years a person will live once they reach age 65. There *has been* an increase in life expectancy from 65, but it’s smaller than life expectancy from birth. I don’t know of anyone that has proposed that the aging process has changed, but improvements in medical care and nutrition have decreased mortality from heart attacks, strokes, cancer, etc.

      1. Robert Young

        One can’t ignore the Baby Boomer effect, but it’s opposite to what is commonly asserted. It goes like this.

        Public funding for Old Folks Programs have been funded by the wave of BBs in their prime work years, since the funding is on current account. The BBs have been paying for the interventions which prolonged the post-65 life spans of their parents and grandparents. This is the legacy of a current account system. Since a true “investment” funding (by definition, buying shares and bonds of corporations, or your brother in law’s pizza joint) of such programs leads to perverse decisionmaking; e.g. if the Federal government depends on investing in AIG, how can it police AIG’s behaviour? Or any other for-profit business? Would the SS fund have been suckered into housing CDOs? The lowest overhead, and least perverse, option is current account funding. Which is more important: enforcing honest commerce, or maximizing government return on private investment? Which would you choose?

        The trust fund meme is just a ruse to shut up some ignorants.

  12. Robert Young

    Thanks to Dave “The Revolutionary” Smith’s post, I came back to review. D’oh!!! While the comments pointed to the life expectancy issue, we’ve all missed the most glaring error in the analysis:

    — Perhaps I’m just naive, but it seems obvious to me that we can’t afford to take on several additional years of retirement pension liabilities for every living American.

    This statement *assumes* that SS (not to speak of private pensions, but those I know of behave the same) pays the monthly stipend irregardless of retirement age. Of course, that’s never been true. One is allowed early retirement, but at an actuarially neutral stipend (in aggregate), i.e. less monthly moolah if one retires before full-benefit age. Moreover, the age 65 full-benefit point is well in the rearview mirror. It’s now 66 and will soon be 67. So, in sum, the age at retirement is irrelevant, per se.

    A picture is worth a thousand words.

    1. Dathan

      Actuarial neutrality only considers the effect on the individual — i.e., that by retiring a year earlier, the total outstanding benefit is reduced by the entitlement they would have earned for the extra year, and then adjusted downward so that the entire aggregate payout is unaffected by the longer payout period.

      It does not, however, consider the secondary effects: i.e., that the GDP is reduced by the difference between the individual’s full salary and the same individual’s pension — ergo, reducing the aggregate income of the rest of the country (due to expected spending at full income that the individual can no longer enjoy) by a proportional — though not directly equivalent — amount.

      Early retirement, therefore, IS a net drain on the system, despite its direct effect on SS (i.e., the retired individual’s benefits) being neutral.

      1. Robert Young

        It’s been a while since I checked here, so that’s why the late answer. The collateral damage of early retirement, and thus a reduced moolah stream in each of the early retiree’s years of life assumes that such income is a significant proportion of 1) the current account contribution of all workers (it’s not likely, and minor effect is likely accounted for by the SSA actuaries, remember it’s not an investment based program, it’s current account), 2) relatively negative effect of the macro multiplier of the delta between early and full retirement (again, minuscule), 3) the negative delta between normal income and retirement income (once again, not large, and likely accounted for by the actuaries).

        Moreover, the notion that folks take early SS retirement, as distinct from CEO golden parachutes after their companies are destroyed, voluntarily is a myth.

        One may consider the Times a left wing rag, but there you are.

  13. Michael Lachmann

    Notice that the effect seen can be an artifact of the baby boomers starting to retire. As boomers start to retire,
    first the early-retiring baby-boomers retire, skewing the average age of retirement. As the baby boomers move through
    so that we also start seeing the late-retiring baby boomers retire the average retirement age will go back to normal. I haven’t read the article describing the statistic used for average age of retirement very deeply, but it seems to be
    “the average age of all persons who retired in a given period of time”, which would show this artifact.
    I am not claiming that what is seen is an artifact. Just that one has to be careful, because life expectancy and age of retirement are calculated in different ways. (life expectancy is average age of all babies born today) Thus, a statistic that wouldn’t show this artifact would be the average age of retirement of all babies born today.
    Here is R code that shows the effect. Note – no change in average retirement age for an individual, and no change in life expectancy, and yet average retirement age goes down.

    babiesBoom=c(22,25,25,25,21,18,15,15,16,17) # taken from wikipedia, roughly by eye, 5 year interval
    babies=c(rep(20,31),babiesBoom) # This is wrong. I just assume a constant birth rate before the boom
    f=approxfun(x=born,y=babies,method=”linear”) # smoothen to 1 year interval.
    x=sapply(1973:2010,function(year) {
    i=age>50 # cutoff
    alive=babies[i]*(1-pexp(age[i]-60,1/10)) # death rate from age 60 is exponential with mean 20
    retired=alive*pnorm(age[i],mean=55,sd=3) # age of retirement is normal with sd=3, but cutoff at 50
    sum(age[i]*retired)/sum(retired) # mean actual retirement age of those who are retired
    } )
    plot(1973:2010,x) # This clearly goes down, though only starting in 2000.

  14. paps

    Just wondering,what’s the average age of the participants in the dialogue?
    For one there is the average life expectancy but what about average life QUALITY say over 60?
    Could one imagine oneself at the age of 70 (if having managed to get by strokes,alzheimers,cardivascular disease etc) working say in a highly mentaly demanding job (say a programmer) or even better doing physical work (let’s say loading/unloading luggages in an airport).
    It’s amazing that all you people are discussing about the statistics and what is the proper base for it etc ,and loosing the simple facts of life.Take a look around you for a a person aged >=70 and TRY to put yourself in her shoes but doing the job you ‘re doing today.

  15. Duff Clarity

    “Perhaps I’m just naive, but it seems obvious to me that we can’t afford to take on several additional years of retirement pension liabilities for every living American.”

    If worker productivity keeps increasing the way it has over the last century, there is no problem at all in keeping the retirement age where it is, or even in decreasing the retirement age by several years.

    Machines now do most of what had to be done by humans a century ago. This trend shows no sign of ending.

    It is weird that in one of the richest countries in the world, at the point in time when it is at its richest so far, we hear all these voices claiming we can’t afford nice things.

    You’re wrong. We can afford it. We’re rich. We’re getting richer every day, and we are automating everything. We can afford to let old people retire. We’re getting to the point where we can let young people retire, too.

    We’re programmers, we are supposed to be working ourselves out of our jobs.

  16. Karl Bunday

    A friend who is a policy analyst saw that I shared this post on Facebook, and he commented, “The comments are much better than the main article.” The comments definitely get into more issues than the main article. But I’m rather surprised that the comments aren’t up to date with the latest research on increasing lifespan. (I know a local researcher who works on that issue with Kaare Christensen in Denmark. He also knows James Vaupel, another path-breaking researcher on increases in human lifespan.) Life expectancy is going up quite steadily at all ages,

    and I (a Baby Boomer) have every reason to expect to work to much older age than the retirement age of my Silent Generation parents (which is not even to mention that their parents worked well into their late seventies, back decades ago). A routine expectation that most working adults will cease working at an age certain defined by a taxpayer-funded benefit program is a distinctly bad idea, and I will be glad to see that idea become a curiosity of history, even if I am “forced” to work till the age that both of my grandfathers were still working at.