Why We're Not Going To Eat Our Grandchildren

I often putz around at Matthew Yglesias' site. In the comments to this post, I was inspired to a frenzy of back-of-the-envelope calculations.

My frenzy was an attempt to estimate total US expenditures on the elderly, as a percentage of GDP -- both in 2004, and as projected for 2054. My assumptions were that total expenditures would be the sum of (1) Social Security, (2) money the elderly get from other sources (pensions, savings, etc.), (3) Medicare, and (4) other spending on elderly health care.

It's important to note here that, in the present, US public expenditures on the elderly (Social Security plus Medicare) are just about the same size as US private expenditures on the elderly (pensions, savings, etc. plus private health care). Thus, if this holds true in the future, the private sector side will be just as much a "problem" as the public side. Yet we only hear about the public side, and the private sector side is treated as though it doesn't exist.

I hope Matthew didn't mind my use of his bandwith. In any case, I'm also going to stick my frenzied calculations here as well. For the short Social Security book I'm working on, I'll actually get real numbers rather than these rough estimates. And I'll add graphs! But in the meantime, if anyone wades through this and sees anything wrong, please mention it. (You may also have to read the whole thread on Matthew's site first to understand it.)

But the point of all this is that the whole "old age crisis" is bullshit. America is not going to turn into a nation of youngsters crushed under the unbearable weight of their grandparents. Here's why, according to my rough figures:

With no changes to Social Security or Medicare or anything, the per capita GDP for everyone under 65 will be 68% HIGHER IN 2055 THAN TODAY... THAT IS, AFTER THE ELDERLY HAVE TAKEN THEIR SHARE IN SOCIAL SECURITY, INVESTMENTS, AND HEALTH CARE.

Those are the consequences of the "demographic tidal wave" we face -- the per capita share of GDP going to 2055's youngsters (ie, those under 65) will be 68% higher than the per capita share the under-65 get today. (And this is under pessimistic assumptions.) Wow, what a massive injustice.

Again, my book will have non-rough numbers by people who actually know what they're talking about. But in the meantime, here are my calculations

P.B. Almeida,

Here's my attempt to put some numbers on this. Of course, it's extremely rough. Plus, I'm tired right now, and not as familiar with the health cost numbers, so I might have made some errors. Still, it should be good enough for a blog comment.

Of the total amount of increased economic output that will flow to old people over the next half century or so, a portion of that (20%? 25%?) will be in the form of Social Security checks.

My best guess is 15.5%. (This is under the SS and Medicare intermediate assumptions.)

2004 SS % of GDP
4.33

2054 SS % of GDP (same page)
6.49

SS increase 2004-2054
2.16

If we assume 2004 and 2054's retirees have the same proportion of asset income, pensions and "other" as 2001's retirees (95% of SS), the asset income, pension and "other" % of GDP come to:

2004 Other Over-65 Income % of GDP
4.11

2054 Other Over-65 Income % of GDP
6.17

Other Over-65 Income increase 2004-2054
2.06

2004 Medicare % of GDP
2.68

2054 Medicare % of GDP (same page)
10.03

Medicare increase 2004-2054
7.35

Medicare currently accounts for about 50% of health care spending for those over 65. Because of the new prescription drug benefit, let's make a generous (from your perspective) assumption that this will increase to 66% in the future. Thus:

2004 over-65 non-Medicare health care % of GDP
2.68 (2.68 X 1.00)

2054 over-65 non-Medicare health care % of GDP
5.02 (10.03 X .5)

over-65 non-Medicare health care increase 2004-2054
2.34

2004 total elderly consumption
4.33 + 4.11 + 2.68 + 2.68 = 13.80% of GDP

2054 total elderly consumption
6.49 + 6.17 + 10.03 + 5.02 = 27.71% of GDP

total % of GDP elderly consumption increase 2004-2054
27.71 -13.80 = 13.91

% of increase attributable to Social Security
2.16 / 13.91 = 15.5%

% of increase NOT attributable to Social Security
84.5%

% of increase attributable to health care
(7.35 + 2.34) / 13.91 = 69.7%

There are many ways of looking at this. Here are two.

1. Let's say we slice SS benefits (both guaranteed benefits and returns from any personal accounts) in half from 6.49% of GDP to 3.25%. This is a gigantic cut, to a level far below that of today. But it would only reduce the % of GDP going to the elderly by 11.73% (3.25 / 27.71).

By contrast, we could achieve exactly the the same thing by reducing health care costs by only 21.59% (3.25 / (10.03 + 5.02)).

So, in terms of total elderly consumption in 2054, every percentage point of reduction of health care costs gives us 2.3 times as much as an equivalent percentage point of reduction in Social Security benefits. ((10.03 + 5.02) / 6.49)

2. It's probably more sensible to look at the increase from today's levels, rather than the absolute projected amounts in 2054.

If we hold SS benefits steady at today's level of 4.33% of GDP, in 2054 costs would be 2.16% of GDP lower than with scheduled benefits.

Under these projections, health care costs are assumed to increase by 9.69% of GDP (7.35 + 2.34), from the 2004 level of 5.36% to 15.05% total (10.03 + 5.02).

Thus, if health care costs for the elderly only rise 240% in GDP terms ((15.05 - 2.16) / 5.36) instead of 280% (15.05 / 5.36), then we're in just as good shape as if Social Security increases by 0%.

Put another way: the increase in health care costs is four and a half times as large a problem as the increase in Social Security costs (9.69 / 2.16).

Again: I may well have made a mistake, in arithmetic or otherwise. If so, I'd appreciate it if you'd point it out. But as it stands, I see no reason -- particularly with proven incompetents running the country -- to tinker with America's most successful government program ever. Not when the increase in health care costs is four and a half times as large a problem. (Although there is one thing I'd endorse re SS -- raising the payroll cap to cover 90% of wages. That in itself would eliminate much of the shortfall, according to the CBO's projections.)

And now, I will go truly crazy. It would be much more efficient to just search for a good study of this, rather than estimating it myself. But who cares?

Let's assume that nothing whatsoever is done, and the percent of GDP going to the elderly DOES rise from 13.80% of GDP today to 27.71% in 2054. That seems pretty bad. So what are the consequences of that for the non-elderly, according to the SSA intermediate projections?

(I'm fading on this, so again: please let me know if you see any mistakes.)

2005 population
301,935,000

2005 over-65 population
36,739,000

2005 under-65 population
265,196,000

2005 GDP
$11,911,330,000,000 (12,090 / 1.05)

2005 GDP going to elderly
$1,643,764,000,000 (2005 GDP X 0.138)

2005 GDP going to under-65 population
$10,267,570,000,000 (2005 GDP X 0.862)

2005 per capita GDP for under-65 population
$38,720.85 ($10,267,570,000 / 265,196)

2055 population
388,938,000

2055 over-65 population
83,105,000

2055 under-65 population
305,833,000

2055 GDP
$27,528,190,000,000 (109,843 / 3.9902)

2055 GDP going to elderly
$7,628,061,000,000 (2055 GDP X 0.2771)

2055 GDP going to under-65 population
$19,900,130,000,000 (2055 GDP X 0.7229)

2065 per capita GDP for under-65 population
$65,068.62 ($19,900,130,000 / 305,833)

AND SO:

With no changes to anything, under the intermediate Trustee assumptions -- which of course, seem likely to be pessimistic -- plus my assumptions, the per capita GDP for everyone under 65 would be 68% HIGHER IN 2055 THAN TODAY (65,068.62 / 38,720.85 = 1.68).

So, if nothing is done, in 2055 people under 65 will suffer the crushing burden of having 68% more per capita GDP than those under 65 today. Of course, it seems quite likely we'll find a way to slow down health care cost inflation, making the 2055 picture even brighter for 2055's under-65s -- even under the intermediate projections.

Speaking of which, you ask:

if you've got some ideas [re health care costs] I'm all ears

Okay: how about we switch to a health care system like Sweden's?

In 1995, 17.3% of Sweden's population was over 65, and they spent 7.2% of GDP on health care (for everyone, elderly and non-elderly).

We're a younger society than Sweden. The SSA projects that we won't get to 17.3% over 65 until 2025. Yet we spent 14.2% of GDP on health care in 1995.

So, if we switched to a system like Sweden's, we could theoretically cut our health care % of GDP in half by 2025 -- while providing coverage for all citizens, including the increased percentage of elderly. This saving of 7% of GDP would be more than three times the scheduled increase in Social Security costs.

Sounds like an incredibly sweet deal to me. What do you say?

And finally, you write:

I question your logic here. A bigger pool of financial assets in 2050 implies more growth along the way (than otherwise would be the case) and higher overall standards of living for everybody.

You're not questioning my logic. You're questioning the logic of the Social Security Administration and the Congressional Budget Office. None of their projections show increased economic growth anywhere near high enough to allow retirees to get higher benefits from individual accounts while also allowing non-retirees to have a higher per capita GDP.

Also, when I suggested raising taxes on the asset income of the elderly, you said:

Um, yeah, we could do that. And your point is?...

My point is that this would reduce the % of GDP going to the elderly just as cutting SS benefits would. Yet you speak of the "voracious and growing appetite" of SS while not speaking of the "voracious and growing appetite" of the elderly's asset income. It's either both or it's neither.

Whew! Well, that was exciting. As I say, I've posted this on my own website here. In the near future I'll doublecheck this, while also looking for a good study on it done by people who actually know what they're talking about.