« Powerline On Guckert | Main | Very Good News »

Transition Costs, Briefly

Tim Lee thinks I don't know what I'm talking about with regard to Social Security privatization's transition costs. Will Wilkinson, less inclined to be generous, just straight-up accuses me of lying. Reading Tim's critique, I don't even think I see us disagreeing about anything. The Bush proposal, as I understand it, will, in fact, improve the long-term actuarial balance of Social Security vis-à-vis the intermediate projection of current law. But it's a very long term. Between implementation and the dawning of "the long term" it will lead to the accumulation of a massive -- and problematic -- quantity of government debt. What's more, the feature of the Bush plan that leads to the improved long-term actuarial balance is not the benefit offset for people who establish private accounts, but the switch to price indexing of everyone's guaranteed benefits. There it is in a nutshell. I'm not even sure what we're disagreeing about.

February 25, 2005 | Permalink

TrackBack

TrackBack URL for this entry:
https://www.typepad.com/services/trackback/6a00d8345160fd69e200d83457f05869e2

Listed below are links to weblogs that reference Transition Costs, Briefly:

» More on Transition Costs from The Bit Bucket
Matt says we don't really disagree about transition costs. I argue with some of his other readers in the comments to that thread (and in the comments to this one). Those commenters have convinced me that the explanation in my... [Read More]

Tracked on Feb 26, 2005 11:30:23 AM

Comments

I commented on Tim's website that it seems to me it is Tim that has his math wrong. He seems to be saying that in the future the government will have smaller obligations but the same amount of revenues so that the deficits of today will be matched by surpluses tomorrow.

I may be missing something but that doesn't sound correct. Deficits today will result in debt but there will be no new revenues to pay that debt (using Matt's example).

Posted by: GT | Feb 25, 2005 9:51:12 PM

GT,

Actually, under price-indexing, benefits would be cut so severely that even with the lower levels of revenue there would eventually be a huge gap between incoming payroll tax revenue and outgoing benefits. It's around 2080 that the government really sees the upside of privatization.

This is, of course, preposterous. Assuming America is still something resembling a democracy in 2080, there's no way people will allow payroll taxes to be so high for such meager benefits. But that's the assumption on which privatizations relies -- borrow huge amounts now, because we can be sure that Congress in 2092 will follow the rules we're making in 2005. You bet!

Posted by: grh | Feb 25, 2005 9:57:59 PM

grh,

I undersatnd that but I was referring to Matt's hypothetical example where no benefits are cut. That's the one that Tim suggested Matt got wrong but it seems it's Tim that has made the mathematical mistake.

Posted by: GT | Feb 25, 2005 10:01:00 PM

Matt, I didn't straight up accuse you of lying. I said "Yglesias is either being shady or doesn't know what he talking about." Disjunctions aren't a straight up way of saying anything. And "shady" shady doesn't mean "lying." What I had in mind was the possibility of characterizing the issue (perhaps technically correctly) in a way that obscures more than it illuminates. I wouldn't accuse you of flat out lying. I hope you know that.

Posted by: Will Wilkinson | Feb 25, 2005 10:03:53 PM

Will, I'll gladly be corrected but as I noted in both your and Tim's blog the only substantive point Tim makes appears to be based on a mathematical error.

Posted by: GT | Feb 25, 2005 10:08:18 PM

why aren't you guys all out gettin' laid?

Posted by: praktike | Feb 25, 2005 10:23:50 PM

Good idea prak!

I'll ask my wife about that.

;)

Posted by: GT | Feb 25, 2005 10:26:17 PM

I think Matt is right. Think of it this way:

By creating private accounts, we are in effect cutting taxes and cutting guaranteed benefits. If we cut everyones taxes by 25%, and everyone's benefits by 25%, we will be basically in the same position that we are now, in terms of the actuarial balance. But that isn't what we are doing. We are cutting everyone's taxes, but only cutting benefits for future retirees. Current retirees keep their old (higher) benefit levels. Thus, transition costs.

Posted by: cs | Feb 25, 2005 10:32:49 PM

Good post, but the term "transition costs" misleads, by making it seem like a minor issue. Making the "transition" from a PAYGO to a privatized system means nothing less than that one generation has to pay for 2 retirements: their own retirement and their parents. That's what "transition costs" really means, if you intend to honestly & responsibly privatize the system. Without that willingness to pay the "transition costs", any privatization proposal is merely a dangerous and irresponsible sugar plum fantasy that the government can create wealth by issuing new Treasury bonds and then "investing" the proceeds in the stock market.

A related point, it seems to me that the most dangerous & foolish assumption made by the privatizers is not that stocks will always go up, but that Treasury bonds will always be cheap, that the government will always be able to borrow money at a mere 2 or 3% after inflation, no matter what. It wasn't so long ago that Treasury bonds were 14%, but everybody, even Prof. Delong and Kevin Drum, seems to have forgotten that. The bottom line is that the "Borrow at 3 and invest at 7" logic underlying most private accounts proposals is incredibly Naive & irresponsible, even if you believe stocks always go up "in the long run".

Posted by: roublen vesseau | Feb 25, 2005 10:52:42 PM

Matt,

The specific sentence I was disagreeing about was this one:

"The debt, however, won't ever go away unless some huge some of money is found elsewhere."

Under Bush's plan, the transition-cost debt will go away, because the carve-out from future benefits will reduce expenditures to the point where the system will begin running a surplus. Indeed, the long-run impact of Bush's personal accounts on Social Security's finances are a wash.

Now if that's what you meant, then I guess I just misread your article. But then I also don't understand your oft-repeated assertion that Tanner et. al. have been changing their story on the subject. We've always said that the transition will have large short-term costs, but that in the long run the transition costs wash out due to lower required benefits. Those are perfectly consistent statements.

Now, it's true that personal accounts by themselves don't eliminate the unfunded liability. No one--including Cato or the administration--claims that they do. But accounts do raise the benefits workers can expect to receive from a given level of taxes, thereby making benefit cuts in the traditional system less painful. Thus, when combined with benefit cuts, personal accounts allow us to have a solvent Social Security system that provides workers with higher benefits than you'd have if you tried to close the gap with benefit cuts alone.

Posted by: Tim | Feb 26, 2005 1:12:07 AM

hey I just posted a comment at tim's blog that might make this clearer. Note that opting into private accounts means investing money-- so if you're choosing whether to opt out or not, it is a choice between (A) giving current retirees X dollars or (B) giving your future self more than X dollars.

Thus, it is quite plausible that you could pay comparable benefits without diverting all your SS taxes-- thus you really are making money for the system, and you have revenue to pay down the debt at the end of the story.

Posted by: mk | Feb 26, 2005 1:29:19 AM

Let me be even more explicit since this is important.

Under the current system, you pay X to give current retirees X.
Under the would-be new system, you pay X to give your future self more than X. (because it's an appreciating investment).

So you really do get some extra money. (But wait, it's only really "extra" if it's more than the interest rate-- since the gov't pays interest on the initial outlays)

So if you invest in something that does better than bonds, then yeah you get some extra money. If you get enough extra money, you can pay down the debt so it's all a wash financially.

Devil's in the details, I don't know how much extra money there is.

Posted by: mk | Feb 26, 2005 1:38:51 AM

mk: You've got the right idea, but the picture's actually rosier than that. The way the math works out, the money you invest only has to match the rate of return on bonds in order to allow the state to pay of the debt without making you any worse off.

The baseline assumption is one in which the state is already on the hook for benefit payments both to today's retiree and to the worker who's paying the taxes. Under personal accounts, you pay X to your future self, and the state borrows X to pay the current retiree. X appreciates to some larger value, while at the same time, the bond the government issued to allow you to invest in your personal account also accumulates interest. Then, when you retire, your personal account pays you X plus the investment return. The government, meanwhile, has a debt of X plus the bond interest to pay off. But the government, knowing you've got that personal account, can cut your traditional Social Security benefits by the amount needed to pay off the bond, and as long as your personal account returns more than the bond that financed it, you end up better off. The only way personal accounts are a net loser for either the worker or the state is if personal accounts underperform bonds. In that case, the interest on the bond is larger than the return on the personal account, and so if the state deducts the amount needed to pay off the bond from your Social Security benefits, you end up worse off. But historically, stocks have outperformed bonds consistently and by a wide margin for more than a century.

Posted by: Tim | Feb 26, 2005 2:34:30 AM

Ack, it just occurred to me that my last comment wasn't quite right. As generation 1 is paying back the bonds, generation 2 is going to need to borrow bonds to finance its personal accounts. So I think you're right-- if the personal accounts only return as much as government debt, then there will never be an opportunity to pay down the transition cost borrowing.

Fortunately, in the real world, equities have done better than government bonds for more than a century.

Posted by: Tim | Feb 26, 2005 3:57:30 AM

"Fortunately, in the real world, equities have done better than government bonds for more than a century."

This seems like a dubious claim. What data are you using?

Stocks are down about 40% so far this century. Good thing we didn't start this plan in 2000!

Posted by: monkyboy | Feb 26, 2005 4:18:03 AM

This is a great deal for people who will get both personal accounts and the remainder of their due benefits financed by debt.

It's not so great a deal for the unborn foetuses who get to pay for the baby-boomer ancestors they never knew.

But I think I'm finally getting a sense of Bush's overarching ideology and vision. It's all about giving future generations a chance to own a treasured piece of history: a vintage, Bush-era IOU to the federal government.

Posted by: jake haisley | Feb 26, 2005 5:10:16 AM

"And "shady" shady doesn't mean "lying." What I had in mind was the possibility of characterizing the issue (perhaps technically correctly) in a way that obscures more than it illuminates. I wouldn't accuse you of flat out lying."

But you seem happy to accuse him of intellectual dishonesty...

Posted by: Petey | Feb 26, 2005 5:21:46 AM

Tim,

"the picture's actually rosier than that. The way the math works out, the money you invest only has to match the rate of return on bonds in order to allow the state to pay of the debt without making you any worse off."

You seem to be missing the central point that by pegging baseline benefits to inflation rather than wages, if a retiree has received the equivalent to the return on bonds that the clawback returns to the government, he or she will be significantly worse off under the Bush plan.

Posted by: Petey | Feb 26, 2005 5:25:53 AM

But all of this assumes that with real growth of only 1.9% vs the historic average of 3.4% the stock market would still average historic returns -- a very
questionable assumption.

Posted by: spencer | Feb 26, 2005 8:11:08 AM

Remember, the point of MY's original post was to debunk the claim that transition costs are just "prepaying" the future Social Security debts, and don't add any additional cost to the system. To make that point (if I remember correctly), Matt talked about a hypothetical case in which the future actuarial balance is zero and privatization is not accompanied by additional benefit cuts. As far as I can tell, nothing that Tim or mk or anyone else has said here contradicts Matt's reasoning.

Posted by: cs | Feb 26, 2005 8:24:40 AM

cs,

You are right. Tim attacked Matt for no good reason. Matt's point is perfectly correct, there are transition costs. What Tim should have said is that there are no NET transition costs IF benefits are cut by more than payroll taxes AND IF returns on private accounts are large enough to cover the difference between the two.

That is very different than what Matt said.

Posted by: GT | Feb 26, 2005 8:41:53 AM

"Ack, it just occurred to me that my last comment wasn't quite right. As generation 1 is paying back the bonds, generation 2 is going to need to borrow bonds to finance its personal accounts. So I think you're right-- if the personal accounts only return as much as government debt, then there will never be an opportunity to pay down the transition cost borrowing."

That's exactly right. That's why the transition cost is very real (as any actuary familiar with the system will tell you). That's also why it's very useful to think of the system from a cohort perspective.

Now, you can say the transition costs will be paid for with future cuts in benefits, but that's still a cost.

Second, as regards higher returns from equities: When you do the math, what you see is that the higher returns don't make as big of an impact as you might think (or some privateers imply).

Our model showed that even if you immediately plunked the entire trust fund into equities and started earning a 7% real return, you'd only push the date of insolvency back about 10 years.

Bush's plan wouldn't do anything remotely that aggressive. Only a small portion of payroll taxes would get plowed into equities. If that's what you're relying on to make up the transition costs, you're going to be waiting a loooooong time. (And never mind the extra risk you're taking on.)

Posted by: Mahan Atma | Feb 26, 2005 8:48:06 AM

Fortunately, in the real world, equities have done better than government bonds for more than a century

If the equities are obviously such a better deal (in the real world) than government bonds, why do people and institutions buy government bonds? They must be really stupid.

We are lucky to have the Bushies to enlighten us, but we should keep it a secret from those idiots who are still buying $500 billion worth of US treasury notes every year. Otherwise they will certainly stop buying and the treasury will have to start paying higher interest rates. Bummer...

Posted by: abb1 | Feb 26, 2005 9:10:16 AM

Here's one of our papers dealing with the effect of investing in equities (among other policy changes):

http://www.ssa.gov/OACT/stochastic/others.html

It's the Lee-Tulja-Anderson paper. It doesn't deal with private accounts per se, rather we just plunk a portion of the Trust Fund into equities to see what happens. (As far as the boost you get from a higher return on equities, the math is not fundamentally different than you'd get with private accounts.)

Also, there's a link to an on-line simulator I set up. It lets you play around with various scenarios, including investing in equities.

Posted by: Mahan Atma | Feb 26, 2005 9:16:46 AM

"If the equities are obviously such a better deal (in the real world) than government bonds, why do people and institutions buy government bonds? They must be really stupid."

Well, most investors buy some mix of bonds and equities, don't they? It isn't because people are so stupid, obviously, it's because there's a "risk premium" -- the bonds' lower return is the price you pay for getting the security.

You cannot deny, however, that equities have historicallly earned a higher return than bonds. If you look at the S&P 500, you get an average of about 7% real return over the past 50 years, compared with 3% for bonds. I've got the numbers on my computer somewhere, if you want to see them.

Posted by: Mahan Atma | Feb 26, 2005 9:21:33 AM

The comments to this entry are closed.