Arnold Kling  

Social Security Reform

PRINT
Subsidies Raise Prices... Social Security Transition Cos...

The New York Times has two relevant op-eds. John Kasich writes,


benefits are growing faster than inflation. First-time Social Security benefits are now tied to wage growth, and wages are rising faster than prices. The result: over the next 75 years, benefits are expected to increase nearly eighteenfold, while prices will go up less than half that rate. In order to keep pace, our children and their children will have to work longer hours and pay more taxes. Between now and 2080, benefits will most likely exceed payroll taxes by $120 trillion.

How do we get out of this mess? To preserve the system for the long term, we must change the way first-time benefits are calculated. Growth in initial benefits should be linked to the consumer price index - not to wage growth.


See also the chapter in Learning Economics called "A Social Security Policy Primer."

Jose Pinera describes the Chilean experience with private accounts. Of particular interest is how they dealt with the transition issue.


There was no "economic" transition cost, because there is no harm to the gross domestic product from this reform (on the contrary, there is a huge benefit). A completely different issue is how to confront the "cash flow" transition cost to the government of recognizing, and ultimately eliminating, the unfinanced Social Security liability. The implicit debt of the Chilean system in 1980 was about 80 percent of the G.D.P.

We used five "sources" to generate that cash flow: a) one-time long-term government bonds at market rates of interest so the cost was shared with future generations; b) a temporary residual payroll tax; c) privatization of state-owned companies, which increased efficiency, prevented corruption and spread ownership; d) a budget surplus deliberately created before the reform (for many years afterward, we were able to use the need to "finance the transition" as a powerful argument to contain increases in government spending); e) increased tax revenues that resulted from the higher economic growth fueled by the personal retirement account system.

For Discussion. Could similar approaches be used for managing the transition in the United States?


Comments and Sharing


CATEGORIES: Social Security



COMMENTS (21 to date)
spencer writes:

I've always been a great believer in Samulesons great comment that any concept in economics that you could not explain to your father in law would eventually be proven wrong. Would you explain the difference between "economic costs" that you say do not exist and "cash flow costs", particularly in view of the point that your "cash flow costs" requires steps, a, b,c,d,& e to
finance the transitions.

I do not understand the difference -- do your?

Lawrance George Lux writes:

Long-term Government bonds is out, as Some claim We are absorbing two-thirds of the World's surpluses now. Cutting Government spending is out, unless you want to impeach George W. for being a Spendthrift. There will be no economic growth with increased Tax revenues, as the Stock Market and Financial markets suffer from oversupply of Dollar contributions, while Bush insists all this must not be taxed. A residual tax will work(?) possibly, if We cut out the Limit of FICA contributions, and if We raised the FICA tax rate to absorb the withdrawals into Private accounts. Oh, I forgot one--privatizing State-owned businesses--We could turn the White House into a Bed and Breakfast. lgl

Boonton writes:

Wait a second, if SSI's benefits are tuned to grow with wages (instead of prices) and wages are growing faster than prices then how can SSI's 'return' be negative?

The obvious solution is to cut benefits in the long run but do it now so people have years or decades to get ready. Yet we are told by the privatization crowd that would just make returns 'more negative'?

Boonton writes:

Let's also read this very carefully:

In 1945 there were about 42 workers paying into the system for each person receiving benefits. Today that ratio is 3.3 to 1, and by 2040, there will be just two workers for each beneficiary.

At the same time, Americans are living longer. That's good news, but it means retirees will receive benefits for longer, putting further pressure on the fund. Americans are also having fewer children, which means fewer workers will be paying the taxes that help finance benefits.

Furthermore, benefits are growing faster than inflation. First-time Social Security benefits are now tied to wage growth, and wages are rising faster than prices. The result: over the next 75 years, benefits are expected to increase nearly eighteenfold, while prices will go up less than half that rate

Today 3.3 workers per retired person.
Tomorrow 2 workers per retired person.
Tomorrow's wages 18 times larger than todays? Or will tomorrow's wages be only 9 times larger (half of eighteenfold)?

Either way if tomorrow's workers make more than 2 times todays workers in real terms then the decline of workers to retired people is not a problem.

spencer writes:

Also note that no one talks about the dependency ratio.

I do not know -- but will the increase in elderly as a share of the population be offset by a drop in children as a share of the population.

Has anyone seen a projection of how many non-workers to workers we will have 2042?

Jim Erlandson writes:

Spencer: Transition Cost Story

My brother-in-law borrowed $1.0 million from me 15 years ago. In exchange, he promised to pay me $10,000 per month once I reached the age of 65 years. I have his IOU stating the above and have been carrying it as a $1.0 million asset.

My brother-in-law has no assets, no job, hasn't worked in 15 years and died today with no life insurance. I turned 65 today expecting to retire with $10,000 per month from my brother-in-law. My brother-in-law's IOU has always been worthless, but now it is official. Declaring it worthless has no real effect on my income for the year (my personal Gross Domestic Product). However, the $10,000 per month I was expecting in retirement (the cash flow I won't receive) is a very real problem.

I guess I need to keep working and saving for a much later, self-financed retirement.

I hope this helps.

Bob writes:

Boonton, your point is well known. It has always been the case that, given modest productivity increases, SS can be "saved" by giving all of society's increased standard of living to the retired. With productivity gains approaching the levels of the past five years you can even give a few crumbs to non-retired folks, but under the current system retirees get the lion's share (something like 80%) of aggregate gains.

You might as well argue that the real return for a current newborn will actually be great - all they have to screw the next generation more than they were screwed.

Eshan Shah-Jahan writes:

Since healthcare suffer from Baumol's cost disease, it doesn't make sense to index to price level. The point of indexing is to keep benefit levels approximately the same over time. Since healthcare costs grow faster than prices do, this amounts to a cut in benefits.

Now, cutting benefits may in fact be the best solution. If that's the case, we should be clear about what we're doing, instead of indexing in such a way that benefits will shrink.

Jim Glass writes:
Would you explain the difference between "economic costs" that you say do not exist and "cash flow costs"

Saving requires cash flow. No accountant in the world calls saving a “cost” – something that reduces your liabilities and strengthens your balance sheet is the opposite of a “cost”.

Also note that no one talks about the dependency ratio
The dependency ratio is a canard. The government spends more than seven times as much per senior than per child, and the gap is increasing each year.

So while people do talk about it, they shouldn’t.

Jim Glass writes:
Wait a second, if SSI's benefits are tuned to grow with wages (instead of prices) and wages are growing faster than prices then how can SSI's 'return' be negative?

Because taxes rise much faster than any of them.

I mean, if we want to calculate whether a return is positive or negative, we have to consider what the return is on don’t we?

Now, the Social Security Administration's actuaries say that every annual cohort retiring after 2000 will get a negative return from SS, compared to if their taxes had been invested in gov’t bonds. With today’s young workers gettting back as little as 50% of what they contributed. This is considering all benefits, including disability, survivor, etc.

That is using the legislated benefits that are 30% underfunded, of course. If that funding gap is closed by either benefit cuts or more tax increases – and with a paygo system there is no other option – then their return will be reduced to as little as 35% of what they contributed.

This figures to make SS very popular in future years, eh?

The obvious solution is to cut benefits in the long run but do it now so people have years or decades to get ready.
But of course the politicians have this "obvious solution" at hand right now. They aren't dummies.

All they have to do is stand up and say… “You young workers, we’re cutting your SS benefits back to one-third of what you will pay in SS taxes. Problem solved!”

Hey, it’s so simple, why aren’t they doing that?

Yet we are told by the privatization crowd that would just make returns 'more negative'

You are also told it by the SSA’s actuaries. And unless you know a type of math where a return of 50% of contributions isn’t negative, and a return of 35% isn’t less than 50%, you should be telling people that yourself!

if tomorrow's workers make more than 2 times todays workers in real terms then the decline of workers to retired people is not a problem.

Alas, with benefits indexed to rise with wages, rising wages leave you right where you started.

Boonton writes:
Boonton, your point is well known. It has always been the case that, given modest productivity increases, SS can be "saved" by giving all of society's increased standard of living to the retired. With productivity gains approaching the levels of the past five years you can even give a few crumbs to non-retired folks, but under the current system retirees get the lion's share (something like 80%) of aggregate gains.

Really all of it? If the 2 workers per one retired person have wages that are 8 times higher than today, that would roughly be equilivant to having 16 of today's workers per retired person instead of 3.3. Something is quite fishy about the numbers being bandied about, if you take them seriously there would seem to be plenty of room for retired and wage earners.


The dependency ratio is a canard. The government spends more than seven times as much per senior than per child, and the gap is increasing each year.
So while people do talk about it, they shouldn’t.

Children are simply not as expensive as seniors. While I would support means testing the money we spend on well off seniors I do not think those numbers should be reversed and if they were I doubt the money would be well spent.

Saving requires cash flow. No accountant in the world calls saving a “cost” – something that reduces your liabilities and strengthens your balance sheet is the opposite of a “cost”.

Reducing cash flow and adding debt to your balance sheet 'strengthens' it?

Boonton writes:
Now, the Social Security Administration's actuaries say that every annual cohort retiring after 2000 will get a negative return from SS, compared to if their taxes had been invested in gov’t bonds. With today’s young workers gettting back as little as 50% of what they contributed. This is considering all benefits, including disability, survivor, etc.

The problem with this type of analysis is that it assumes that every cohort *could* have purchased gov't bonds with their tax money and the return on gov't bonds would have remained exactly the same.

Wait a second, if SSI's benefits are tuned to grow with wages (instead of prices) and wages are growing faster than prices then how can SSI's 'return' be negative?


Because taxes rise much faster than any of them.

No, SSI's taxes are a % of your wages with a cut off around $88K or so. By definition your tax bill can only grow as fast as your wages. Granted tax rates can be increased but the actuaries make their projections by assuming the program will remain on 'autopilot' for the next 75 years.

Hey, it’s so simple, why aren’t they doing that?

When was the last time politicians did anything whose payoff would be 50 years plus from now? The strains you are describing are the natural result of attempting to finance 33% of the adult population as retired. That can be done if society is very rich (that is if Arnold is right that nanotechnology & such means we are on the cusp of a period of growth that will dwarf anything previous in our history). If not, then more of the population will have to work.

So bring down the cost of the program and then you can control tax increases and possibly even cut taxes.

Bob writes:

Boonton,

Where do you get an 8x increase in real wages? What kind of productivity growth are you assuming? 2% for 35 years makes real wages double. 3.5% for 40 years doesn't even get you 4x. If 3% is realistically optimistic, then you'll get your 8x in 70 years. The crisis comes sooner.

But to your point, if productivity grows 1.5% then you can offset the drop in workers per old folk over the next 35 years only if you give the old folk ALL of the productivity gains. Non-retired folks get NO increase in their standard of living. If productivity grows faster, there's something left over. If I recall correctly, my "old folks get 80%" was from a calculation using 2.1% (real) growth.

Boonton writes:

Bob, that number came from trying to make sense of the article:

The result: over the next 75 years, benefits are expected to increase nearly eighteenfold, while prices will go up less than half that rate

I see the piece talks about benefits and not wages but it says that benefits are tied to wage growth. Hence wages go up x 18 while prices go up x9. 18/9 is 2 so real wages double. That's pretty close to a productivity growth rate of 1% (rule of 72 would say 72 years to double).

If real wages double then 2 workers paying the current 14% tax in 2080 would be like 4 workers paying that same tax today.

Jim Glass writes:
'Saving requires cash flow. No accountant in the world calls saving a “cost” ... '

Reducing cash flow and adding debt to your balance sheet 'strengthens' it?

The question was how something that requires cash flow could not be a cost, what's the difference. Saving requires cash flow but is not a cost, that shows the difference between cash flow and cost.

But in response to your question, sure borrowing can strengthen a balance sheet. If I borrow against my home at 6% to pay down a credit card incurring 18%, my new debt certainly strengthens my balance sheet.

Similarly, General Motors just strengthened its balance sheet substantially by borrowing several billion dollars to pay down pension liabilities that were accruing at a higher interest rate than it had to pay on its new borrowing.

Similarly, various SS privatization proposals have been scored by the SSA actuaries and CBO as reducing or eliminating the net present value of SS's current $10 trillion unfunded liability -- because the borrowing eliminates future accruals of unfunded benefits by an amount that exceeds the cost of the borrowing. Strengthening the government's balance sheet.

To see the debt incurred in a privatization proposal, but refuse to see the corresponding-and-larger debt in the status quo that the proposal eliminates, is rather disingenuous.

One really mustn't choose to evaluate changes in a balance sheet by looking at only one side of it.

Jim Glass writes:
"Now, the Social Security Administration's actuaries say that every annual cohort retiring after 2000 will get a negative return from SS, compared to if their taxes had been invested in gov’t bonds. With today’s young workers gettting back as little as 50% of what they contributed."

The problem with this type of analysis is that it assumes that every cohort *could* have purchased gov't bonds with their tax money and the return on gov't bonds would have remained exactly the same.

It assumes nothing of the kind.

It simply applies a discount rate to account for the time value of money, which is an arithmetic necessity for determining a rate of return.

The SS actuaries use the government bond rate. If that's not up to your standards and you prefer to use a different interest rate, please help yourself.

Pick a discount rate you prefer. The arithmetic result will be the same: Those in the past received from SS much more than they contributed to it ($10 trillion more, actually, up to $300 billion per year) while those in the future will get back much less.

This isn't a matter of left-right dispute, it's just plain fact. To deny it is simply an exercise in denial.

"Wait a second, if SSI's benefits are tuned to grow with wages (instead of prices) and wages are growing faster than prices then how can SSI's 'return' be negative?"

'Because taxes rise much faster than any of them.'

No, SSI's taxes are a % of your wages with a cut off around $88K or so. By definition your tax bill can only grow as fast as your wages....

Geeze. Rates don't matter??

Today's retirees who are getting back more than they put in started off paying 3% tax, and paid single-digit tax rates for decades. Today's young start off paying 12.4% for benefits that were reduced for them (but not for older workers!) by the 1982 legislation.

And even those legislated benefits are 30% underfunded, so there will have to be more tax increases or benefit cuts or both for the young in the future, as a matter of arithmetic.

So the young start paying more than 4X as much tax from day one, for smaller benefits, that are themselves 30% underfunded and so will have to be further reduced -- as you yourself have recommended.

See how today's young get such a worse deal than today's retirees?

The Social Security that everybody so fondly remembers as being such a good deal is already gone, dead and buried.

Figuring up the return on contributions, past and present, is no more difficult than counting up the taxes paid -- real dollars -- and benefits promised, and applying a discount rate.

If you don't believe folks like the SSA actuaries are competent do that, well, I don't know what to say.

Jim Glass writes:
"Hey, it’s so simple, why aren’t they doing that?"

When was the last time politicians did anything whose payoff would be 50 years plus from now?

50 years??? How may times have I linked to that GAO report showing just the annual interest on the debt hitting 20% of GDP in 30 years on current policy, and compounding straight upward, with the 'end of government' then.

You know, 30 years is within the lifetime of a lot of recent retirees -- not to mention younger workers.

But you clearly don't like the GAO's numbers, nor the SS actuaries', geeze. Though I've never seen any alternatives presented here.

However, you are correct that politicians are very bad at making even easy responsible decisions for such programs -- yet another good reason to give citizens property rights in them.

Boonton writes:
It assumes nothing of the kind.

It simply applies a discount rate to account for the time value of money, which is an arithmetic necessity for determining a rate of return.

The SS actuaries use the government bond rate. If that's not up to your standards and you prefer to use a different interest rate, please help yourself.

In other words it assumes the bond rate would have stayed the same if 14% of wages were suddenly shifted into purchasing bonds on the private market.

In the real world bond yields change minute by minute depending upon supply and demand and its pretty hard to believe hundreds of billions of dollars could have been redirected over the last 30 years without altering the bond yields.

Boonton writes:
50 years??? How may times have I linked to that GAO report showing just the annual interest on the debt hitting 20% of GDP in 30 years on current policy, and compounding straight upward, with the 'end of government' then. You know, 30 years is within the lifetime of a lot of recent retirees -- not to mention younger workers.

Nonetheless 30 years is a huge period in the career of a politician. Consider that a 40 year old Senator would be considered young in order for him to reap the benefits of putting in a policy that pays off in 30 years he will have to be running for office in his 70's!

Considering how many people horribly underfund their 401K's I don't think human nature leaves much hope in this case...

kenny writes:

I do not know -- but will the increase in elderly as a share of the population be offset by a drop in children as a share of the population.

john top writes:

For Discussion. Could similar approaches be used for managing the transition in the United States?

No, because Chile at the time of the transition was a dictatorship, so there was little viable opposition to any of Pinochet's policies. While we're close to being a one party system, we're not there yet.

Comments for this entry have been closed
Return to top