Social Security Debate Online

I didn't think much of Clinton either, but otherwise thanks for the dialog.

Well you were both talking past me, but it seemed to be an interesting, informed, relatively staid discussion.

You should've sucker-punched him at the end, though. To punctuate your point, I mean.

Wow, I'm actually going to answer one of Arnold's questions instead of debating with Boonton.

Does the WSJ format work to generate a useful discussion, or do the participants simply talk past one another?

I've read online discussions like this before, and I don't think they generate useful discussion. I think they do better than real live debates. But they're simply too short. There is never, a point at which either side says, "Oh. I see what you're saying." Both sides seem to be so completly afraid to admit any value in the other sides position because it suggests defeat. I'd like to know not only what's wrong with the other sides, position, but I'd like to know what's right with the other side's position. I'd like to see both sides agree to do that. But of course, neither side will for fear that the other side will proclaim victory.

Of course, witnessing my discussions with Boonton, I should talk.

I agree that the WSJ forum is too short to provide a useful forum for more discussion.
It can spell out what the real differences are
and I was surprised how close you and Max were.

I still do not understand your argument on life expectency. I think you should have a good discussion with an actuary. I am sure I misunderstood, but when I read your article on life expectency I came out withe the impression that you were projecting a fall in average life expectency to some 69 years. I am sure I must have misread you somewhere.

It was definitely a talking-past-each-other experience. Very frustrating. Fundamentally, I think you missed the opportunity to flog Max with the one really sound (IMO) argument for privatization: A system where each person is responsible for their own retirement savings is inherently more stable to changes in any of the CBO projections than the current system. You pointed out (rightly) that the demographic projections may be completely out of whack, and he countered with "if they could be wrong on that, they could be wrong on any part of this, and so therefore let's not discuss it." That, IMO, was key - the opportunity to point out that the current system is much more vulnerable to variances in those projections than a private account system.

Given that the format made Max moderate some of his more nonsensical claims found on his blog, I'd say the WSJ format was productive.

You did a good job whacking his 'total dependency ratio' sophistry, but you missed the opportunity to jump on this:

"If we were starting from scratch, perhaps we would not construct a pay-as-you-go system. But we're not, so all the speculation about pre-funding has no purpose...."

That's an error in reasoning. The promises that have been made about future benefits are sunk costs (which are equal to the excess benefits paid out to people who retired already). The only question is who is going to end up paying them.

We should ignore those sunk costs in thinking about the system of the future, and simply design it so its incentives will lead people to fund their own retirement in a rational manner.

Does the WSJ format work to generate a useful discussion, or do the participants simply talk past one another?---Yes it does provide the Talking points in a rational format.

I just read Feldstein's address to the AEA conference, Jan-7-9, and must personally again downgrade Private Accounts. The Paper, though, would enforce most of your Views on Private Accounts. What turns me off is the fact that All ignore the real problem: How is the General Revenue Budget going to repay the Social Security Fund? I am afraid We can't be talking about more Debt here.

Arnold,
Your marginal increase in GDP lacks validity, unless you can prove it means a real increase in hard Capital construction, not just financial Paper with already poor P/E ratios. lgl

Would you please demonstrate how no public resources were used to raise children in the 1960s. I seem to remember some public school and public health care resources being used in the 1960s. But I have been wrong before.

Anyone who trots out the "dependency ratio" argument loses by default on my scorecard. That was officially refuted by the 1994 SS Advisory Commission's finding that the gov spends 9-1 more on retirees than children, IIRC. Which will only be going up with Medicare. Talking points should be updates within ten years.

As to Max's:

"I would insist that the likely adjustments are well short of major ones..."

I'd quantify it and let the facts speak for themselves.

By 2030 -- long before 2042, 75-years from now, or an "infinite horizon" -- all federal income taxes will have to go up 63% from today's level as a percentage of GDP to fund SS and Medicare. That's due to the claims of these programs rising by 5.34% of GDP. (All federal income taxes today are 8.5% of GDP, so there you are.)

Of this 63% income tax increase, 22 points will be for the SS trust fund and another 13 points for the other often forgotten trust fund also funded with payroll taxes, the HI-Medicare trust fund.

So by 2030 we're going to have a 35% income tax increase due just to the trust funds alone.

Then there will be another 28% increase in income tax due to general Medicare obligations.

[Analysis and links to Trustee's data.]

Now, I don't know about Max, but I'd consider a 63% income tax increase (with more coming every year) by 2030 a major adjustment.

And voting taxpayers will have to agree to pay this increase well before 2030, of course.

Personally I think that anybody who believes a 63% income tax increase is going to go through without some major benefit decreases as a part of the deal is a bit loopy.

After all, we do have precedent to look at. In 1983, when the needed tax increase for SS was nowhere near this size, benefits were cut by 50% of the funding gap. (And Congress wasn't also raising income taxes by 28% then to pay for Medicare).

And when retiree benefits get cut as part of such a deal, why would anyone think only Medicare benefits would get cut while SS benefits get held sacrosanct?

What if Congress instead decides that retirees' spending cash should get cut so as to save more for their health care? As we hold that tax increase to only 30%?

Jim,

From the Trustees report you cited:

The annual cost of Social Security benefits represents 4.3 percent of Gross Domestic Product (GDP) today and is projected to rise to 6.6 percent of GDP in 2078. The projected 75-year actuarial deficit in the combined Old-Age and Survivors Insurance (OASI) and Disability Insurance (DI) Trust Funds is 1.89 percent of taxable payroll, down slightly from 1.92 percent in last year's report. The program continues to fail our long-range test of close actuarial balance by a wide margin. Projected OASDI tax income will begin to fall short of outlays in 2018 and will be sufficient to finance only 73 percent of scheduled annual benefits by 2042, when the combined OASDI trust fund is projected to be exhausted.

Social Security could be brought into actuarial balance over the next 75 years in various ways, including an immediate increase in payroll taxes of 15 percent or an immediate reduction in benefits of 13 percent (or some combination of the two). To the extent that changes are delayed or phased in gradually, greater adjustments in scheduled benefits and revenues would be required. Ensuring the sustainability of the system beyond 2078 would require even larger changes.

http://www.ssa.gov/OACT/TRSUM/trsummary.html#wp31181

A 15% increase of the payroll tax would take the rate from 12.4% to 14.26%. While any tax increase is unpleasent it is pretty modest. As I demonstrated several times, if your SS tax and benefit is taking a relatively constant portion of GDP the system is both self-sustaining and will provide returns equal to the overall growth of the economy.

If you look at the above mentioned report, look about 3/4 of the way down the page to chart D.
You'll notice that currently the projections (which are probably unduly pessismistic for Social Security) indicate that SS is currently generating a surplus and will max out with a deficit of 2% of GDP in 2078. We are currently running a deficit of 4% of GDP today!!! You'll also clearly see the problems are Medicare's hospitalization coverage (peaking at about 4% of GDP in 2078) and Medicare's medical coverage (peaking around 6+% of GDP in 2078). It is quite frankly absurd to get upset about a projected deficit that in 50 years will be half of the deficit we are currently running while glossing over a deficit that will be seven times as large.

SS projected demands on the budget are both questionable and manageable. Medicare/Medicaid is a different story which is not as rosy (which, BTW, is why you always insist on linking it to SS...without Medicare the SS story just isn't bad enough to merit the cries of crises).

The problems with bringing Medicare into the picture, though, are

1. The projections are not as reliable. SS projections requires demographic projections which are very reliable and economic projections which are not. Medicare projections requires demographics, economic and the cost of medical care projections.

2. The cost of medical care is highly erratic and subjective. Since it has been growing faster than the overall economy any straightline projection will show it becoming unmanagable. We cannot project the pace of new discoveries and the changes in supply and demand that will determine the cost of medical services far into the future. In the 1930's SS's actuarials were able to predict the portion of Americans over 65 in 1990 within a margin of error of 0.16 points (see http://www.janegalt.net/blog/archives/005130.html#41559). They could not have done the same with medical costs.

3. Privitization does nothing to reform or address Medicare whose underlying problem is that medical care costs are rising faster than both inflation and economic growth.

Since SS reform does nothing to touch Medicare I think it is dishonest and deceptive to continually lump SS & Medicare together when talking about projections. Any serious reform of Medicare would be totally different from a reform of Social Security and also completely independent.

"A 15% increase of the payroll tax would take the rate from 12.4% to 14.26%. While any tax increase is unpleasent it is pretty modest...."

Boonton, as always, you are ignoring the cost of the increase in income taxes needed to pay down the SS trust fund bonds -- which cost is not included in the number you choose there.

The Trustees give this number for the year 2030 as being 1.93% of GDP (not payroll) and continuing to rise thereafter.

A 1.93% of GDP increase in general revenue needs for Social Security indicates a 23% increase in GDP terms of the income taxes that provide general revenue, being that all federal income taxes today equal 8.5% of GDP.

You can decide for yourself whether a 23% increase in income taxes is "pretty modest".

Especiallly in a world where Medicare will have simultaneously driven (according to the Trustees) an additional 40% increase in income taxes -- 63% total by 2030.

The solvency of SS, assuming the trust fund bonds are paid off with magic free money from somewhere, which is what you go on about there, doesn't mean nuts to a tree. It is irrelevant to everything in the real world.

What matters in the real world is the solvency of the government, including the charge it will have to take in costly real money to cover the operation of the trust fund, which will mandate a 23% income tax hike by 2030.

And also the willingness of voting taxpayers to swallow all of that tax hike, of course.

Because if voters aren't willing to pony up the whole amount of taxes needed to pay promised benefits -- like they weren't in 1983 -- then benefits are going to get slashed some more by then, just like they were in 1983, only on a much bigger scale.

Boonton, as always, you are ignoring the cost of the increase in income taxes needed to pay down the SS trust fund bonds -- which cost is not included in the number you choose there.

Indeed, if you assume that the bonds will be paid down immediately. Why would you assume that roughly 10% of the Federal Debt would be paid off with income tax revenue in a relatively short period of time? Like much of the Federal Debt it is more likely to be rolled over in the form of bond sales to the public. Most likely it will never be paid off but, assuming sensible management of the budget in the future, will not be allowed to grow faster than the ecnomy in the long term. If that is done it will become a smaller portion of the overall economy. I could just as easily make a scary illustration by asking how much of an increase in income taxes will be necessary to pay off all the combined deficits of the Bush administration in 10 years.

The Trustees give this number for the year 2030 as being 1.93% of GDP (not payroll) and continuing to rise thereafter.

Again the source you are asserting as authoritative shows SS increasing by 2% of GDP and holding steady thru 2078 (see chart d).

The annual cost of Social Security benefits represents 4.3 percent of Gross Domestic Product (GDP) today and is projected to rise to 6.6 percent of GDP in 2078.

By your reasoning if 2% of GDP equals a 60% income tax increase then how much income tax is being devoted to the current 4.3%? Odd isn't it how no one is telling us that Bush's 4% of GDP deficits are not equilivant to a 120% income tax increase. How do you explain this:

Social Security could be brought into actuarial balance over the next 75 years in various ways, including an immediate increase in payroll taxes of 15 percent or an immediate reduction in benefits of 13 percent (or some combination of the two).

As I demonstrated, a 15% in payroll taxes is minor. Bush's proposal to stop wage indexing of the first benefit check will result in a benefit cut of way over 50%. All that is needed is a small tweak in both directions and you are perfectly fine. There's even a good chance that you will end up perfectly fine anyway since the Trustees are using unduly pessismistic numbers.

Finally I still maintain it is dishonest to continually link SS to Medicare in an attempt to generate a false crises. The two programs are driven by different dynamics and proposals to reform one have no direct effect on the other.

Especiallly in a world where Medicare will have simultaneously driven (according to the Trustees) an additional 40% increase in income taxes -- 63% total by 2030.

Which there you go again. Why do you assume that Medicare comes first? That income taxes of 20% for SS will be coming on top of Medicare's 40% hike? Considering that SS is highly stable and relatively predicatable and manageable it seems bizaar to say it needs to be tossed overboard to save Medicare which is the opposite.

For an example of a sensible reform take a peek at the CBO's analysis of the Diamond-Orszag plan. The plan would do the following:

1. Increase the payroll tax 0.26% each year from 2012 thru 2075. The rate would rise from 12.4% today to 12.7% in 2025 to 13.9% in 2050.

2. Increase the amount subject to the tax and institute a 3% tax for those income beyond the cap.

3. Reduce benefits starting in 2012 and accelerating in 2023. The overal reductions would be around 2% in 2025, 12% in 2050 and 23% in 2105.

Go down to figure 1A.

The first graph depicts the current projections. Note that SS outlays are relatively flat after 2045-2055.

The second graph depicts the proposed reforms. Notice that SS's bite out of the economy is reduced to just under 6% (today it is just under 5%) while revenue is just around 6.5% generating a new SS surplus from 2055 to 2105.

Of course, if elements of this plan seem too extreme the surplus could be decreased and the benefits increased. Alternatively the surplus could be eliminated entirely allowing the tax increases to be slashed, especially far into the future were projections are much more likely to be wrong.

That link is http://www.cbo.gov/showdoc.cfm?index=6044&sequence=0