Arnold Kling  

Stop Calling it the Retirement Age

PRINT
Should I Take to Drink? The L... Government in the Mortgage Bus...

Ezra Klein writes,


there is one thing that both parties increasingly seem to agree on: You should work longer.

Klein's thinking is that people do not want to work longer, so raising the Social Security "retirement age" is a bad idea. His conclusion does not follow from his premise. To avoid this sort of error, we need to stop using the term "retirement age" to refer to the age at which one becomes eligible for government benefits.

Instead, think of it as the age of government dependency. If Klein does not think we should increase that along with longevity, then he is arguing that as lives get longer and longer, taxes on working-age citizens should get higher and higher.

Keeping the age of eligibility low may not necessarily help people to retire sooner. It may have the opposite effect. By keeping taxes high, it may reduce employment and wages for people of working age, thus making it harder for them to retire when they would like.


Comments and Sharing


CATEGORIES: Social Security



COMMENTS (22 to date)
Fmb writes:

If there were no ss, it could still be the case that the age of retirement-savings-dependency would not go up much even as we become longer-lived-- perhaps people will want to live that way, or perhaps it's hard to maintain non-zero marginal product much longer than we do now. It certainly seems likely that the proportion of productive years will go down.

As this happens, the need to save for retirement out of earnings "should get higher and higher.". Yet this need not create any disincentive to work.

A well-designed ss system could therefore have the properties you worry Klein advocates and yet not have the drawback you fear. One key is a strong connection between taxes paid and benefits received.

Which is not to say that your title isn't correct, nor that the existing system is adequate to prevent the disincentive effects you fear.

me writes:

[Comment removed for supplying false email address. Email the webmaster@econlib.org to request restoring this comment. A valid email address is required to post comments on EconLog.--Econlib Ed.]

D. Boon writes:

The phrase "longer and longer" in this context is a bit misleading. A look at actuarial tables shows life expectancy at birth has increase roughly 3x more than life expectancy at age 65 since 1930. More people are reaching the age of "government dependency", but we don't hang around that much longer now than we did in 1930.

Ask why the cash flow into SS is not sufficient to maintain the current retirement age given that the contribution rate is 14% of gross wages. Perhaps someone is collecting too much "rent"?

And why raise everyone's retirement age when my and a lot of other folk's contribution to SS is capped? Remove the income cap first, then let's talk about raising the retirement age. Workers benefiting from the income cap are more likely to have jobs that are less physically demanding and we have the means to fund retirement outside of SS.

And what about the elephant in the room - Medicare. It is arguable whether SS has a fiscal problem. Nobody argues that Medicare is not a problem. But the defenders of the status quo here are not mere workers, but the medical/FIRE corporate interests.

d writes:

Raising the age would hurt the lower classes. Yes I know some people don't care. Just raise the tax cap and means test it, voila.

d writes:

Also, think of the alternative. Social Security is currently self-funding, threatened by demographics and recession. Turn those around and we're good.

What would you rather have, a bunch of poor folks dependent on welfare? Or a bunch of poor folks dying?

rjs writes:

ongoing longevity hasnt increased much for those who reach 65...the longevity at birth, which this argument is apparently based on, has increased because of declining infant mortality...

that said, for those not performing physically hard labor, there is no reason for retirement...

Jim Glass writes:

Klein's thinking is that people do not want to work longer, so raising the Social Security "retirement age" is a bad idea.

What people want only matters when compared to the alternative.

CBO says the cost of running SS is going to increase by about 2% of GDP out of general revenue (to run the trust fund). That's about a 15% across-the-board income tax increase by 2030, the largest in near 60 years in GDP terms. (And it will be on top of one twice as large to run Medicare, 50% total.)

Of course, seniors will not be exempt from this tax increase -- people over 50 have the highest incomes and wealth of all age groups.

So what Klein should be asking people is: "Do you want to pay the largest tax increase in 60 years (on top of another tax increase twice as large) on both your pre-retirement and your retirement income -- pensions, IRAs, life savings, *and* Social Security benefits -- to pay for your Social Security benefits, OR do you think people should work longer?"


Jim Glass writes:

The phrase "longer and longer" in this context is a bit misleading. A look at actuarial tables shows life expectancy at birth has increase roughly 3x more than life expectancy at age 65 since 1930...

ongoing longevity hasnt increased much for those who reach 65...

Life expectancy at age 65 has increased since 1940 from 12.8 years to 18.5 years -- by 45%, a tremendous amount for a retirement plan to finance ... especially as the workers who finance it have retired earlier and earlier.

Life expectancy at birth has increased from 63.6 years to 77.7 years, 22%.

Table 11 (.pdf)

sfHeath writes:

Jim Glass: thank you so much for providing the source of your data. So a citizen at age 65 in 1940 could expect to live until 77.8; a citizen at age 65 in 2006 could expect to live until 82.5; a difference of 4.7 years.

And as you say, life expectancy at birth has gone from 63.6 years to 77.7 years; a difference of 14.1 years.

So D. Boon was reasonably close when he used a factor of three ("roughly") comparing the life expectancy at birth in 1930 vs. today.

Which is more useful when examining social security policy, your numbers or his? Obviously, both are mathematically correct and they support each other, but which is more useful?

My first impression is that the increased life expectancy at birth (which is usually the number looked at when politicians talk about the history of social security) means that there are significantly more workers in today's system, comparatively speaking, than there were in 1940's system. Am I wrong about that?

Randy writes:

I have no interest in raising the retirement age because I have no interest in the Social Security program as a going concern. I want the political class to pay to me what they promised to pay to me, and I want my children to be able to opt out entirely. And I don't care if they have to cut every other political program to the marrow in order to do it.

Dan Weber writes:

"Screw everyone else. I want mine."

Allan Walstad writes:

d

Just raise the tax cap and means test it, voila.

If you're going to do that, why not just fund it out of regular tax revenue? The problem is that SS was sold as an "insurance" program. You pay in, your money is saved in a "trust fund," and it's there to support you when you retire. Raising the tax cap and means testing benefits yanks the fig leaf away a bit too quickly for the pols' comfort, I suspect.

If we are going to keep robbing working people to pay retirees' living expenses, then the best way is to abolish SS per se and replace it with a means-tested welfare program for old folks.

Better yet, stop the intergenerational robbery and leave people free and responsible to save for their own retirements, as is appropriate for a free society. Raising the retirement age is the best route to that conclusion. Simply raise the age to collect benefits by, say, 3 or 4 months per year, every year into the distant future. Those already retired are unaffected. Those within a few years of retirement only have to work a little longer (or they can bridge the gap with their own savings). But young people entering the work force will understand that preparing for their retirement is their responsibility, and they'll be better able to do so as the number of people collecting benefits decreases relative to the number paying in, so that the SS tax can be phased out.

Jim Glass writes:

Jim Glass: thank you so much for providing the source of your data. So a citizen at age 65 in 1940 could expect to live until 77.8; a citizen at age 65 in 2006 could expect to live until 82.5; a difference of 4.7 years.

5.7 years

Table 11: "Average number of years of life remaining"

Age 65
2006: 18.5
1939-41: 12.8

difference = 5.7, 5.7/12.8 = 44.5% increase

And as you say, life expectancy at birth has gone from 63.6 years to 77.7 years; a difference of 14.1 years. So D. Boon was reasonably close when he used a factor of three ("roughly") comparing the life expectancy at birth in 1930 vs. today.

Though he could also have said life expectancy at 65 has increased 44.5%, compared to 22% at birth.

Which is more useful when examining social security policy, your numbers or his? Obviously, both are mathematically correct and they support each other, but which is more useful?

My first impression is that the increased life expectancy at birth ... means that there are significantly more workers in today's system, comparatively speaking, than there were in 1940's system. Am I wrong about that?

The workers paying in to retirees receiving benefits ratio has plunged.

1945: 42 to 1
1960: 5 to 1
today: 3 to 1
future: falling further as the boomers retire.

This huge drop is why the early generations of retirees received $15 trillion more from SS than they contributed, were made richer by that much, while currrent and future beneficiaries will get from it $15 trillion less than they contribute, as per the Trustees -- be made poorer on a lifetime basis by that amount.

That's a $30 trillion swing. The famous "intergenerational transfer" -- cemented in place by the 1983 SS Reform that hiked taxes on and slashed benefits of the then-young by that much to preserve the larger benefits of the then 50+ers, who'd paid as little as 3% tax.

One can understand how receiving $15 trillion net plus made SS hugely popular with grandpa's generation -- but it ain't grandpa's Social Security any more.

John Thacker writes:

The reason that life expectancy at birth jumped more is because of a decline in infant and childhood deaths. However, only a small decrease in such increases life expectancy by a lot.

The important thing is dependency ratio, which is why the large percentage increase in the small number (retirees) is more important than the small percentage increase in the big number (workers), even though the absolute change is larger in the bigger number.

sfHeath writes:

Jim Glass:

Sorry about my arithmetic error. However, your point about the workers to beneficiaries ratio comparison doesn't really answer my question, since there's no point in comparing the inception of the plan to the maturity of the plan in that regard. The originators of SS planned for a ratio of around 3- or 4- to 1 at maturity, which is precisely what it hit between 1965 and 1975, according to the chart you provide. 3.7:1 at 1970 is not that radical a difference from 2006's 3.3:1.

Nor does it answer my basic question, which is the comparison of the number of people working at 1930's and '40's to today, as the high ratios at the inception of the plan were much more about the significantly smaller number of recipients at inception vs. at maturity.

The ratio won't continue to go equalize after the boomer generation dies out, as in the sixties we got reliable family planning through the Pill and Roe v. Wade, so the birth rate has been basically a flatline since 1970. Look at p. 16 of this presentation. As long as those rights are protected, this country will never again experience a baby boom like that.

It looks like the chart you provided does answer my question, (I'm not certain because I don't know the exact path of the plan to maturity); in 1930 there were 35 million workers and in 2006 there were 161 million workers. Like I said, I don't know exactly how to make use of that number because I don't know what the factors behind that increase were. But shouldn't an examination of the supply of workers be part of a discussion of social security?

Sorry for the long post.

JKB writes:

The problem with raising the retirement age is that it isn't that simple. I know way to many people who in their mid to late 50s got dumped from a job and had to struggle just to get by until they reached SS age since no one would hire someone that old in their fields. Recently, Instapundit linked to an article about how 40 is over the hill in high tech and you just aren't going to find a job as a programmer after the mid-thirties. Not everyone can be a manager or superstar. Also, recently there was a meme in the blawger community about how the old professors wouldn't retire or take emeritus status and open up tenure positions for the younger academics. This keeps younger workers making less and struggling in their most productive years.

Yes, employees could assume more control over their employment but that means they become more mercenary. All compensation up front, no discount for expected job security and nothing that vests after age 40. That would mean business creation would be harder since it would require more cashflow in the beginning. Or productive employees bail at the first sign of trouble in the company.

So the question is: What changes are needed to ensure job creation to absorb the younger workforce while maintaining a larger older working population due to increased retirement age? Tough since recently we've been unable to create jobs to cover the workforce as organized much less keep more people in the pool.

How do you handle the industries that would rather keep their workforce young and unintentionally drive out those over say 40? This is important now that childbearing has moved up to the 30s as the expensive child rearing years are now hitting in the parents 40s and 50s with college costs, etc.

sfHeath writes:

John Thacker, isn't the dependency ratio basically flatline once the boomer generation ages and dies? In other words, isn't the graph on page 17 of this presentation saying that after 2035, the dependency ratio is long-term stable?

Babinich writes:

Arnold,

If the Social Security Trust Fund holds special issues available to only the trust funds how is the interest rate of those special issues (long term bond and Certificates of Indebtedness) determined?

Randy writes:

Dan Weber: "Screw everyone else. I want mine."

No. Just, I want mine. I have no objection to others having what is theirs as well.

Hugh Watkins writes:

Some fascinating info here. I don't know if anyone can help me with the following question:

Given current payroll tax rates and life expectancy, are people currently paying into SS paying enough to fund their own retirements at some reasonable pension level?

I realise "reasonable" is vague - but hey, so's Krugman's understanding of how pensions work - so I'm in good company.

Yancey Ward writes:

Babinich,

Here was a recent discussion of the SSA trust fund and the bond issues contained within.

SfHeath,

The dependency ratio is the issue- you are making this harder on yourself than you need to. There are only two ways to raise the ratio of workers to retirees- either raise the retirement age or means test the program more severely. Also, the further out you get, the less sure one can be of the future of this ratio given present policy. If income taxes have to go up 50% in the next 20 years to fund already promised entitlements, how sure can you be that the birthrate of the US won't start to approach that of, lets say, Japan or Germany of today?

Dan Weber writes:

Given current payroll tax rates and life expectancy, are people currently paying into SS paying enough to fund their own retirements at some reasonable pension level?

Pretend I work for 100 credits a year. FICA taxes are about 12%, so I could save 12 credits a year.

Saving 12 credits a year for 40 years at 4% interest, I have 1187 credits at the end. With 1187 credits, a 65-year-old man in New York can get an annuity paying 7.4 credits a month, or 88 credits a year -- which is exactly what they were living on before.

Comments for this entry have been closed
Return to top