Arnold Kling  

Debating Social Security

PRINT
Stadium Obstruction... Debating Outsourcing...

It's Tyler Cowen vs. John Irons at the Wall Street Journal on line. (WARNING: the WSJ to override the article with a new article, which means that the link will not take you to the social security debate)


[Cowen:] Social Security has two parts conceptually: a welfare system for old people, plus a regime of forced savings for the young. I would prefer that Social Security evolve into a system of welfare for the elderly, thereby allowing us to junk the forced savings aspect...

A better fiscal idea [than privatization] is to raise the retirement age, thereby keeping the system solvent. One of the original virtues of Social Security was its minimum administrative costs and its relatively "clean" fiscal nature. Let's not lose these properties of the system simply to adopt a flawed version of the privatization vision.

[Irons:] Yes, Social Security is costly but it also has enormous benefits. Poverty among the elderly is significantly lower than it would be otherwise. According to the National Academy of Social Insurance and the Social Security Administration, "the poverty rate for persons over 65 was 8.5% in the year 2000. Without Social Security the poverty rate would have been 48.1%." Even for those in the middle class, it provides a needed supplement to many who would find it extremely difficult to get along without any public support. It also insures the elderly from running out of their own personal savings.

Meanwhile, Alex Tabarrok chimes in,


for current and past retirees benefit increases, a growing economy and longer life expectancy made social security a real deal. It's today's workers and children for whom social security is a raw deal. Even if the system does not go bankrupt, current workers will receive a very poor return on their "investment."

In refusing to cut benefits to current and soon-to-be retirees the costs of any reform are forced onto those people for whom the system is already a poor return. It would be fairer to spread the costs to all recipients especially to those who have benefited from social security the most.


Except that if you spread the cost over all future generations, we can prepare and save for it. If you reduce the size of grandma's check, she's hosed.

For Discussion. How well does the debate/blog format work in this context?


Comments and Sharing


CATEGORIES: Social Security



COMMENTS (30 to date)
ed writes:

I find John Irons' contributions to the WSJ discussion to be extremely strange. He takes issue with Cowen's calling for a smaller, welfare-like version of social security. Then he procedes to list the benefits of social security, all of which seem exactly like what we would want from a welfare program: lower poverty, guaranteeing "at least some income" for the elderly, ensuring "a modest income in retirement." These are the exact goals that Cowen's also wants to accomplish, but he wants to do it without social security taking over the world. Irons doesn't respond to Cowen's points at all.

johnthe lamb writes:

Here's a PLAN that might work:

At age 40 (for instance), all citizens have a choice:
1. Stay in Social Security, pay FICA, and receive the benefits, OR
2. Abandon Social Security, giving up any benefits claim AND forfeiting all taxes they've paid into the system. In return, they and their employer will continue to deduct FICA BUT it will go into a SS IRA.

The beauty is that people can do what they want; that there will be enough money to pay off current retirees; and that over time you can lower the decision age, eventually getting rid of Social Security once and for all.

All this without transition costs.

Mark Horn writes:
All this without transition costs.
But transition costs are a myth.
Axel Kassel writes:

Mr. Cowen's characterization of Social Security as partly "a regime of forced savings for the young" suggests a very peculiar notion of "savings." Money I save on my own accord is my property. Money taken from me through Social Security tax is not; the Supreme Court has explicitly said so. The widespread idea that people are paying into a personal account that they will one day tap is one of the more pernicious myths standing in the way of rational and equitable reform of this transfer program.

Rob Sperry writes:

I agree with ed, the debate lacked Clash. They seemed to be talking along side each other as opposed to with each other.

I think it would be wonderful if someone could figure out how to get high end folk with differing opinions to take each other seriously and engage each other and record it for others enlightenment. There is really no source of intellectual conversation on the Internet.

I am thinking about is the kind of conversation that spans several sessions where a lot of time is spent uncovering assumptions and clarifying definitions so that each side starts to really understand the points of commonality and conflict. Factual conflicts can be side bared for off line investigation, and the focus can be on values. This of coarse will lead to a discussion of meta-ethics which will lead to epistemology and eventually metaphysics. At this points its just me and 3 other people reading...sigh

But maybe I am overly skeptical?

Boonton writes:
Mr. Cowen's characterization of Social Security as partly "a regime of forced savings for the young" suggests a very peculiar notion of "savings." Money I save on my own accord is my property. Money taken from me through Social Security tax is not; the Supreme Court has explicitly said so. The widespread idea that people are paying into a personal account that they will one day tap is one of the more pernicious myths standing in the way of rational and equitable reform of this transfer program.

The savings in this sense refers to the economy as a whole. By running a surplus (or at least a smaller deficit) gov't contributes to the total savings in the economy. Since this type of savings is produced through taxation it is 'forced' in the sense that you have to pay taxes while other types of savings (by individuals or businesses) are voluntary.

Bob writes:

Boonton, you seem absolutely convinced that the SS surplus has not increased gov spending relative to a parallel universe where SS taxes are set to balance in- and out-flows each period. Can you share your logic and/or any empirical evidence that supports your position?

Brad Hutchings writes:

How well does the debate/blog format work in this context?

Not very well when the entries disappear the next day!

Boonton writes:
Boonton, you seem absolutely convinced that the SS surplus has not increased gov spending relative to a parallel universe where SS taxes are set to balance in- and out-flows each period. Can you share your logic and/or any empirical evidence that supports your position?

Bob, take a peek at the comments on http://econlog.econlib.org/archives/000642.html

Boonton writes:

To give a brief summary, I used the table from http://www.cbo.gov/showdoc.cfm?index=4985&sequence=11 which shows revenues, outlays, debt held by the public and SSI's surplus or deficit over last 40+ years as a % of GDP.

I did two regressions, the first followed this model:

Change in debt = B + B2 * Revenue

If revenue caused the gov't to increase spending then B2 must be either 0 or greater. If it isn't then increasing revenue is not offset by increasing spending (spending might go up but not as much meaning that gov't must save at least a portion of revenue). The regression indicated B2 was -1.2551 with a good level of significance. So increased revenue is likely to result in decreasing debt (as a % of GDP).

I did one more, this time comparing the change in revenue to the change in outlays. Since SSI is listed only as a surplus or deficit I added that to outlays. The equation would be:

Change in net outlays = B + B2 * change in revenue

If revenue causes outlays to increase then B2 must be positive. Instead I got B2= -0.356 with R2 of 0.1189 and a t-score of -2.29 which is just under 98% significance. (Please note I made an error in my other post, I added SSI's surplus to outlays when I should have substracted).

this indicates that increased revenue is actually associated with decreased spending. Like I said on the other list, this isn't that shocking. Most of the gov't's spending is on auto-pilot. When times are good revenue increases but spending contracts (fewer unemployment, disability claims...even with SSI more people will opt to keep working when times are good lowering payouts and increasing both SSI tax revenue and regular tax revenue).

Lawrance George Lux writes:

Boonton,
Your analysis is good, but it might not actually address the issue. The SS surplus has negated the need to supply revenue by added taxation or debt. Your regressions only show no immediate correlation between the SS surplus and outlays. It does not answer how much Spending was done, because Legislators knew they would not have to increase taxation. lgl

Boonton writes:

Lawrance,

I would answer by questioning how spending is done by the gov't. Only a small portion of spending involves politicians sitting around saying 'how much shall we spend this year'. Much of spending and taxing is done automatically. This explains the negative relationship between revenue and spending (more revenue associated with less spending) and it would imply that SSI's Trust Fund is a saving mechanism for the gov't.

I'm open to ideas to other ways to test SSI's effect on spending decisions....

Boonton writes:

Ok, So I took just plain outlays and compared them to the SSI surplus/deficit. The model will now be:

Outlays = Bo + B1 * SSIsurplus

We should expect a B1 that is positive if the SSI surplus causes non-social security outlays to increase. The result is the opposite:

B1 = -0.56048
t-score -1.25894
significance 0.215536
R2 = 0.0390

95% confidence bounds
Upper 0.340021
Lower -1.46098

I'm not going to argue that SSI's surplus causes spending to fall (although that's what you see if you just look at the B1 score) because of the low t-score of the regression. There is no clear evidence here that SSI's surplus has any effect on spending (regardless of the larger deficit or other tax revenue).

I would suggest that maybe SSI's surplus has let some congressmen opt for higher spending but remember they deal with raw numbers, billions of dollars. While that is significant in their world the effect of their inclination to wastefullness is swamped by the much larger effects of 'autopilot' spending and taxing and a growing economy.

Bob writes:

Refining the analysis would be pretty easy - replace outlays with discretionary spending and control for economic fluctuations. Of course, it is complicated by the potential that projected surpluses are as important as the current surplus, but that's quantifiable.

Perhaps there is an Econ Phd candidate out there who is interested in the issue? Or someone can point us to a careful study of the issue. I have a hard time believing that it hasn't been done but google hasn't helped me find anything useful.

Boonton writes:

I agree if we wanted to build a model of gov't spending but I think we have established that the SSI surplus is not causing gov't spending to increase in any way that is immediately noticable. If we accept this statement we are left with the conclusion that the Trust Fund is indeed a mechanism for forced savings.

skh writes:

From what Irons says:

"Without Social Security the poverty rate would have been 48.1%"

I'd have to venture a guess that I'd be starving if it weren't for the job that has fed me for 10 years. Is he actually suggesting that SS is the sole income option for retired folks? There are people in that category, and--like smokers who ignore not only warnings, but common sensical observations--they were/are remiss in their planning for their own futures. ROI is what? 1.5% historically? Hell, if the government would've issued some of those mythical IOU's the SS system would be better off by a factor of two, at least.

Irons ignores the fact that if SS weren't extant today, we might very well have a different system that didn't resemble a fraud.

Jim Glass writes:

Boonton, you're missing a few data points, IMHO.

But first, remember principles. You're claiming that spending is inelastic to revenue so that savings result when something like a SS payroll tax is added. Remember, you have to include *tax reductions* in that spending -- they reduce savings too -- and not merely actual tax cuts but also reductions in planned increases. I don't see such tax adjustments in your data.

You've also said that since spending (including tax reductions) is inelastic to revenue it is not restrained by deficits - the other side of it not growing with increased revenue and surpluses.

Now the data points you should add into your numbers:

1) FDRs own initial funded SSA of 1935 had 6 pts of payroll tax financing of which 5 points was planned to be "savings" for many years (as full benefits weren't to be paid for many years). But by 1939 the payroll taxes began to pour in and Congress explicitly said it was too much to save.

The liberals now said that saving by paying down the national debt (your explanation of how SS savings works)was bad because it reduced income taxes owed by the rich and instead paid off bonds with payroll taxes from low-income workers. They wanted to spend the money instead.

The conservatives said saving by paying down the national debt was bad because inevitably the liberals would get their hands on the money and spend it. They wanted to reduce taxes instead.

The liberals and conservatives then compromised by agreeing to do both -- spend more and cut taxes -- wiping out the 5 points of savings as soon as it arose, because it arose. (Over FDR's veto). The modern model of SS was born.

Please include this prompt wipeout of 5 points of new SS savings in your data.

2) Near every Congress after 1939 amended SS by increasing benefits, then afterward increased taxes incrementally by as little as possible to be able to pay the benefits. Please include all these expenditure driven tax increases in your data.

3) FDR's original SS was projected to have saved $500 billion in the trust fund by 1980 (1980 dollars) with the 6% payroll tax, and would have but for the spending increases and early tax cuts that wiped out all the savings. Yet by 1980 SS in fact in was bankrupt and literally within weeks of becoming unable to write any checks (in spite of the tax rate now being at 10%).

Thus we can quantify with a real number the amount of the planned savings that were to be obtained under the original SS law by paying down the debt with payroll taxes, but which were instead consumed by Congress on spending and tax cuts: >$500 billion in 1980, that is ALL of them.

Please include this data point in your numbers.

4) At the same time as #3 was happening, the Reagan deficits were arising. In specific response to these deficits, Congress adopted the "paygo" deficit-control spending caps that blocked all discretionary spending not financed by tax increases. This sharply reduced growth in discretionary spending to about 2% a year from Bush I until Clinton II.

Please look up in the legislative record the spending proposals that were blocked by the deficit-control spending caps and include it in your data. Quantify the difference the caps made.

Also note the tax increases from Bush through Clinton, include them in the data.

5) With the end of the deficits in 1997 the spending caps came off. As Penner said: "It was the surplus". Discretionary spending vroomed up, as we all know, and we all know about all the tax cuts that followed.

Remember, you are trying to measure the relation between changes in *revenue* (not "SS surplus") and spending including tax cuts. So measure the sudden increase in discretionary spending and tax cuts post-1997 that consumed the short-lived surplus, in relation to the unexpected surge in revenue that created it.

Wiping out all that money so quickly with tax cuts and spending increases, it was like 1939 all over again! It seems politicians don't ever change, eh?

Now, doing all the above does seem like a lot of work. But I'm sure professional economists have done all that work many times over. Or at least students writing theses have.

So wouldn't it be easier just to give a couple academic research citations supporting the idea that government spending is inelastic to revenue? After all, you can't be the first person the idea has occurred to -- you can't be *all* alone on this.

rick doll writes:

Potential Format for Social Security Reform

We currently pay 15% of our salary into Social Security. There is concern over how to keep commitments to current recipients. Further, there is the sense that people 50 under are being short-changed.

I am still in the process of researching how the actual numbers would work, so proposal is conceptual in nature--

5% to fund of personal choice
5% to low-risk fund with guaranteed return
5% to cover cost of current recipients

It is my understanding that two very serious things are happening to the Social Security funds...
1) The Government has borrowed them for purposes other than Social Security, and
2)The Clinton administration mergeed the Social Security fund with the General fund which is how the apparent "surplus" was achieved.
If this is the case, then the commitments to current and near-term recipients could be met by simply paying back the money to the Social Security fund. After a few years the 5% that is going toward current commitment could be phased in to the personal choice funds.
The long term result of allowing people to own their accounts and get market-rate returns would be a wealthier group of retirees instead of a group that is dependent and is penalized for making extra money. It would also allow for families of the account-holder to receive the funds regardless of age or relationship.

If anyone has better information on the merging of Social Security funds with the General fund or the borrowing of the Social Security fund, I'm open to looking at it.

Boonton writes:


This is missing the point. we looked at total gov't spending compared to total gov't revenue as a portion of the economy. Looking at billions of dollars sounds like you are watching what's important but its not. Looking at the actual dollar values causes you to miss the fact that this is all taking place in a dynamic growing economy.


4) At the same time as #3 was happening, the Reagan deficits were arising. In specific response to these deficits, Congress adopted the "paygo" deficit-control spending caps that blocked all discretionary spending not financed by tax increases. This sharply reduced growth in discretionary spending to about 2% a year from Bush I until Clinton II.

So what? As I pointed out several times already discretionary spending is only a portion of total Federal spending. What does it mean if a portion of spending was well controlled if the 'automatic pilot' spending is swelling?

I think you need to clarify what you mean by 'tax cuts are spending'. For example, revenue as a % of GDP has fallen from 20% under Clinton to 16.5% under Bush. Am I supposed to consider that a 3.5 point increase in outlays? If so what amount of revenue should I associate that with? 20% or 16.5%? I dont see any sense in changing the analysis in this manner, the CBO definitions of revenue and outlays are fine for this analysis in my opinion.

Finally I'll ask you to clarify what you theory is regarding spending. Are you asserting that revenue drives spending or surpluses? The evidence shows that revenue does not cause spending to go up and it, in fact, has a small tendancy to cause spending to go down. This is logical if you understand that most of Federal spending is on automatic pilot and will tend to fall when the economy is doing good. Likewise, revenue will tend to rise automatically in a good economy without any action by Congress.

skh writes:

Boonton, your assertion that evidence shows that revenue is not tied to spending is in direct conflict with the basic Keynesian model. If, as you say, spending has a tendency to decrease with a revenue increase, we'd have a near-zero level of government spending...since revenue has almost ALWAYS increased. That's not the case, of course, because the more revenue that gets in government coffers, the more money the government will spend. I'd say that is axiomatic. And common sense.

Boonton writes:

skh,

It's neither axiomatic and as for it being common sense, well little more has to be said. You say revenue is always increasing as is spending, yea that's true in nominal terms but that's an observation that is very uninteresting for the following reasons:

1. Inflation is positive each year.
2. GDP growth is positive (almost all the time).
3. Population growth is still positive.

Looking at spending & revenue as a % of GDP filters out the first two effects since we are able to compare these things as a portion of the economy. $1,000 Billion is one thing in the economy of 2004 and quite another thing back in 1973.

skh writes:

Very true. But, how can you argue that the SS system doesn't increase governmental spending when there isn't anything left in the SS accounts? The government has used SS as a buffer to its over-spending since its inception. Regardless of spending/GDP ratios, if Social Security did NOT increase government spending, then we would have all net receipts from SS taxes remaining in the SSA. That is not the case. Congress has taken those same receipts and used them as General Revenue.

Really, though, whether or not the discussion is about SS taxes or income taxes or tariff proceeds, the result is the same: Higher revenue=higher spending. No evidence is available to refute that, but much evidence is available to support it. Nominal or real, absolute or percentage, the flip side of your coin is reality.

Boonton writes:
Really, though, whether or not the discussion is about SS taxes or income taxes or tariff proceeds, the result is the same: Higher revenue=higher spending. No evidence is available to refute that, but much evidence is available to support it. Nominal or real, absolute or percentage, the flip side of your coin is reality.

That's cute skh except for the fact that I actually got the numbers and they don't match your assertion.

skh writes:

No, you got numbers to support your assertion, just as I could get numbers to support what I contend to be the truth. The difference is, my assertion doesn't hinge upon a Catch-21. Seriously. What you tenaciously defend could be compared analogously to Medicare/Medicaid. If the government had never began that particular bit of solcial engineering, medical expenditures would be a mere PERCENTAGE of what they are today.

There is not one bit of evidence extant that you can provide to actually PROVE that collection of SS taxes hasn't increased government spending. There is, however, the troubling bit of evidence: government has been collecting SS taxes for over 50 years--to the tune of trillions of dollars--and doesn't have a single one of those dollars in a "lock box," a "trust fund," or any other financial instrument that would ensure monies be available to SS recipients. Every one of those dollars has been spent. There is no IOU.

Maybe the federal government could have sold some debt and managed to spend as much as they have since SSA inception. But we don't know. What we DO know is that they have managed to spend everything that they've collected. It's torturing reality to claim otherwise.

skh writes:

Oh, BTW, sometimes observations suffice to reach a conclusion, and numbers are superfluous. Something about Occam. Or a razor.

Boonton writes:
No, you got numbers to support your assertion, just as I could get numbers to support

You didn't and you couldn't because I did get the relevant numbers and they don't back up your story. Whether or not the gov't should have kept raw cash in the Trust Fund or gov't debt is a seperate issue.

On that front using the surplus to reduce gov't debt was a proper method of 'saving' the money. Putting raw cash into the fund would have been equilivant to decreasing the economy's money supply which would have been deflationary but would have been neutralized by the Federal Reserve. By paying off debt, the trust fund effectively put the money back into the economy while at the same time reduced the gov't's indebtedness...a close analogy would be someone who knows he will have a $10,000 bill years from now and prepares by paying down his credit card balances.

skh writes:

I'll take that as a concession that you agree that SS has increased government spending overall. You admit that those trillions the feds received from payroll taxes was churned back into the economy which is equivalent to government spending (actually, it is EXACTLY government spending). I thought that was the point of contention. You'll get no debate from me on the relative merits of the government doing so because I agree that it is better for them to "spend" it rather than "save" it.

Boonton writes:

skh,

think about what you're saying. Retiring debt is the 'exact same thing' as spending? Are you going to tell me that if the Federal gov't simply raised the benefits of current retirees to eliminate the SSI trust fund surplus that would be exactly the same as using the trust fund surplus to eliminate Federal indebetness? Your logic demands that conclusion!

skh writes:

Something like that, yep. The fedes have been using the SS account surplus to offset their spending for decades. It is a curious bit of account that is in violation of GAAP and would earn a company legal action against it. What I'm saying is that there IS NO SS SURPLUS. Look, you have compelling evidence for your point, but you aren't 'fessing up that the government spends more now than it would have if it didn't have the revenue stream from SS. That's just not true. To torture YOUR logic a bit, if the government imposed a $10,000/year tax on each citizen--just for the hell of it--you contend that federal spending would tend to decrease. Of course, reality has proven otherwise. Legislators will fund new pork-barrel programs, expand entitlements, and increase foreign aid, among other things. What ever they choose to spend it on, they will spend it...and more. There are sure to be a couple of outliers in modern US history to that trend, but it is all too common recently.

Since this thread is about to scroll off of the main page, I'll agree to disagree with you.

Boonton writes:
Something like that, yep. The fedes have been using the SS account surplus to offset their spending for decades. It is a curious bit of account that is in violation of GAAP and would earn a company legal action against it.

1. "offset spending" is a fancy way of saying reduce debt. Obviously debt represents spending that was done at one time (unless you borrow money just to put in a bank account).

2. GAAP rules that apply to corporate accounting do not apply to gov't accounting for good reason. I refer you to the long discussion we had with Jim regarding the 'Treasury Department Balance Sheet' and why this document, while valid GAAP-wise, is not anywhere nearly as useful as a company's balance sheet.

What I'm saying is that there IS NO SS SURPLUS. Look, you have compelling evidence for your point, but you aren't 'fessing up that the government spends more now than it would have if it didn't have the revenue stream from SS. That's just not true. To torture YOUR logic a bit, if the government imposed a $10,000/year tax on each citizen--just for the hell of it--you contend that federal spending would tend to decrease.

1. No SSI Surplus? That's easy to check, compare SSI tax receipts to SSI disbursements. The balance is the SSI surplus.

2. You've presented no evidence that gov't spending has gone up because of SSI. I've presented evidence that it doesn't. Even more important, I've presented a reasonable explanation for why this would hold, you haven't.

To review for a last time before this post gets pushed off to the archives:

Most of gov't spending and taxing is on auto-pilot. Tax receipts go up in a strong economy because people will make more money, hence pay more income and payroll tax. Spending goes down in good times because fewer people collect unemployment, welfare rolls go down, even things like people filing claims for SSI benefits or disability go down. (Another thing that *may* go down is interest rates which will reduce interest expenses).

Legislators will fund new pork-barrel programs, expand entitlements, and increase foreign aid, among other things. What ever they choose to spend it on, they will spend it...and more. There are sure to be a couple of outliers in modern US history to that trend, but it is all too common recently.

That very well may be but it isn't enough to impact total spending as a % of GDP. Maybe a few billion dollars of new roads, stadiums and other pork was built when the economy was booming and tax receipts were flowing in fast and furious. That wasn't enough to counter the decreased spending from fewer people collecting. That also wasn't enough to counter the fact that the whole economy was growing *faster* than spending was growing so spending as a % of the economy was falling.

This isn't totally disconnected from everyday life. People who become very rich will often increase their spending dramatically but if you become really rich very fast you very well may discover that you cannot increase your spending as fast as your riches. Paris Hilton may drop $20K every night but all that spending can easily be swamped by all the money her families hotels can earn if the economy picks up just a little bit more. She can easily end up saving more of her income than you do!!!!

Comments for this entry have been closed
Return to top