There have been recent discussions about Bush’s proposed budget (see here or here for instance.)
There seems to be some major freaking out about the transition cost of Bush’s Social Security Plan. But as Arnold Kling points out the transition cost is not a new government liability. It is simply moving an off-book debt onto the audit books where it belongs:
In fact, the cost of privatization is zero.
Look at it this way. There has been a brouhaha over accounting for stock options in private companies. Companies deny that options are an expense. But as Warren Buffett puts it, “If stock options are not compensation, then what are they? And if compensation is not an expense, then what is it?”
The promises to future Social Security recipients are to the government what stock options are to private companies–off-the-books costs. You could say, “If promised Social Security payments are not obligations of the U.S. government, then what are they? If an obligation of the U.S. government is not debt, then what is it?”
If we decided tomorrow to recognize Social Security promises as obligations, then this would be a huge “cost” in accounting terms, but no change in economic terms. That is what the “cost” of privatization is all about.
This seems correct to me. Take 4 age cohorts: A=0-20, B=21-40, C=41-65, D= 66+. Right now we have a cost of retirement in Social Security for all of them. We pay as we go. Under a privatization plan we pay for D with current accounts just as before. I presume we would phase C into an intermediate system. Part of their Social Security will be paid under the current system, part of it under the new system. The total extra cost over their lifetime is zero because they were going to be paid Social Security anyway. Cohorts A and B are entirely in the new system. Total extra cost is still zero, but unless the economy completely tanks for decades (in which case there was going to be other problems in paying Social Security) they will have a much higher return on their ‘investment’ into the fund. That means either that they will have more money for retirement, or the government can lower the mandatory contribution amount which functions the same way as a tax.
But none of this is an extra ‘cost’.
Now my prefered solution would be to make Social Security a real safety net program instead of a huge citizenship program. If you are of retirement age and are poor, the government provides assistance. The middle class and rich ought to provide for their own retirement. We obviously can’t implement that immediately, but we could warn people in their 40s not to expect Social Security unless they are poor. This is something we could easily afford. But in any case it is important to remember that identifying a liability is not at all the same as creating a new cost.
Sebastion:
I posted a more detailed comment on RedState. Although he makes some good points, Arnold Kling forgets the time value of money. Even taking each of his assumptions to be true, there are indeed additional “costs” associated with the plan (above and beyond the de minimis transaction costs one might expect from such a large reorganization).
The basic reason why people say that substituting Social Security would entail transition costs of $2 billion is because under the current system, B and C are paying for D’s retirement, and A and B will pay for C’s, and so forth; whereas with private accounts each generation will pay for its own, and the government will have to foot the bill for D and any transitional benefits C gets.
You argue that this is not a new cost, since the government is already liable for these benefits. Suppose I grant that point, the basic problem remains: it is now the case that the government has a revenue stream that will pay for these costs, namely the payroll taxes paid by B and C. If the money now paid by B and C stops going to D’s benefits, but is instead paid into personal retirement accounts, then that money has to be made up from somewhere, unless D’s benefits are cut. It’s about $2 billion, and whether you describe it as a new cost or as a cut in revenue, the point remains the same: it’s money we will have to borrow, and will be added to the already swollen deficit.
Of course, one could avoid this little problem by not cutting payroll taxes. In this case, B and C would be taxed as they now are, but would be allowed to invest any additional money they might have kicking around into personal retirement accounts. But this is the current system: I have just such an account, and no one stops me from investing in it. It is certainly not what Bush is proposing. The only difference would be that at some point in the future, Social Security benefits would disappear.
The bottom line, I think, is that it might be true that Bush’s Social Security reform entails no new “costs”. But if it is true, this would not be because there isn’t $2 billion that we have to find somewhere under his plan but not under the current system, but just because the reason we have to find it isn’t best classified as a “cost”, but as vanishing revenue.
Von and Hilzoy make points that are correct and complementary. The issue is this: granting all assumptions, B is going to be stuck with twice the bill of A, C or D. D has already paid what it is going to pay. C is paying and reaping the benefits paid in by B. A is going to support itself. But the pressure is now on B to fund C, and to fund it’s own retirement.
Long term, the average increase in cost over several generations will be zero. But there will be this hump that will be extremely unpopular, and somebody will have to pay for it in the short term through paying for a previous generation’s benefits as well as their own.
Here’s my question: how do corporations moving from a defined benefit pension plan (a la SS today) to a defined contribution (a la 401k) make this transition, and are there lessons that can be applied? I am all too aware that DB pension plans are killing several major unionized corporations, some of which (United?) are looking to bankruptcy to absolve themselves of this unsustainable obligation – to the detriment of seniors who depend on that check.
I can only hope our government finds the political fortitude to get this fixed before entering such dire straits.
“Long term, the average increase in cost over several generations will be zero. But there will be this hump that will be extremely unpopular, and somebody will have to pay for it in the short term through paying for a previous generation’s benefits as well as their own. ”
Sure it will be unpopular, but this fix deals with three long term problems:
1) It fixes the problem of pay-as-you go ‘investing’ by allowing A and B to become self funding.
2) It allows us to avoid silly projections about how much medical technology will or will not extend lifetimes. If the system looks bad now, imagine tacking just 4 or 5 years onto the life expectancy over the next 50 years. Or imagine a major medical shift that allows a 10-year bump in life expectancy. This provides a hedge against that.
3) It breaks the pyramid scheme problems of the current system without leaving a lot of seniors out of the money they have been planning on their whole life.
3) is also my response to the time value of money problem. The system as is will have a breaking point altogether too soon. If we slowly phase in investment while slowly phasing out pay as you go, we avoid a huge crunch at the end. There is value in spreading out the liability over the next 30 years or so instead of waiting for some nasty 5-year crisis.
Of course my proposed solution is very different, but my point about on and off book liabilities remains.
OK, I’m in C. You can’t cut my payroll taxes without diminishing the amount paid to people in D — which is off the table. If you’re going to cut my benefits in the future, then I have to (a) continue paying the full freight for D and (b) replace the amount by which you’re reducing my benefits with some kind of private contribution.
Of course if you’re going to means-test benefits for C retirees, I’d be paying the full freight for the D’s and have to plan on no SS at all for myself.
Those C’s who believe in a welfare state might be willing to play along, but it’s not a majority among C’s, especially not among the C’s who vote (and tend to be better off than the C’s who don’t vote).
Sebastian: your original post was not about the merits of Social Security reform, but about whether or not it represents a new cost. My point was that if I grant your point about its not being a new cost, that doesn’t change the fact that it is $2 billion dollars that we would have had covered before, but wouldn’t have covered under Bush’s plan, and would therefore have either to borrow or to raise taxes for.
In a way, I think it’s the exact opposite of stock options. In the case of stock options, a company’s revenue and payouts remain the same; all that changes is whether or not you call stock options an “expense” (and whatever follows from that name change.) In the case of Social Security, however, we’re not just deciding to call something a “cost” or a “liability”, we’re diverting the revenue stream that would have paid for it. It’s a real change involving real money, not just a terminological point. And it matters to be clear about this, since it’s two billion dollars of real money.
As far as the merits of Social Security reform go, most economists I talk to seem to think that the system as it stands is nowhere near in crisis. (For instance, see here) The latest Social Security trustees’ report says will be able to pay benefits in full until 2042. There is, I think, a serious argument to be made that this is too pessimistic, among other things because it assumes that productivity growth in the US economy between now and then will be at 1.6%, which seems quite low. If this argument is right, then Social Security will be fine until after 2042. If it isn’t, then Social Security will be fine until 2042. That’s a long ways off, and it also encompasses much of the time during which the baby boomers will be drawing benefits.
Of course, when I say that Social Security is “solvent”, I don’t mean that we can keep on using its surplus to finance the other operations of government as we now do until 2042. That’s when it stops being able to pay for itself fully, not when it stops being able to pay for itself and whatever else we want to use the money for. But obviously the problems caused by the fact that Social Security money will no longer be available for general use is not solved by diverting the revenue stream that provided that money in the first place, especially when, as you note, our liabilities would still remain.
As it currently exists, Social Security only runs into problems if the economy grows at a forecasted annual dismal rate of 1.5%. Why are proponents of privatization schemes allowed to insert, in most cases, very, very optimistic annual growth rates to make the case for privatization? If you wish to make a comparison between SS, as it currently exists, and a privatized SS scheme–use the same forecasted annual growth rates.
As for the notion of a SS “lockbox”–that’s simply sloppy thinking.
According to one of the Social Security trustees it won’t be in 2042 that Medicare and Social Security stop being revenue producers, it’s now:
Lord, send us a cure! The disease is already here.
Dave: That depends on lumping Social Security and Medicare together. The real problems concern Medicare, not Social Security, and they’re the result of rising health care costs. “Reforming” Social Security in the way Bush suggests would not address the problems of Medicare at all; it would simply add billions to the amount of money we need to come up with.
Dave Schuler,
Funny, I don’t recall Medicare being part of the discussion. Do you have the same figures for Social Security only?
No, you’re absolutely right: Medicare is most of the problem. I just stumbled across the quote today and thought it was marginally relevant. The greatest relevance, perhaps, is that with Medicare absorbing a rapidly increasing proportion of the budget there’s less leeway in looking for solutions for other problems including Social Security.
So far as Social Security goes I have no problem with investment (by the government) of part of the trust fund. This was proposed under the Clinton Administration. Privatization appears to me to assume that there is some investment or combination of investments which is assured to result in enough return to provide the benefits that Social Security has provided heretofore. I don’t believe it. And don’t read me the stats on ROI on equities. Check the standard deviation. Historically we have gone through long periods with little or no growth in equities. The growth has come in the form of sharp spikes like the one we just went through.
The issue of cost of the proposed change is beside the point if the proposed change doesn’t solve the problem it proposes to solve.
“The latest Social Security trustees’ report says will be able to pay benefits in full until 2042.”
The payments exceed the receipts in 2017. The trustees are ‘solvent’ until 2042 because they have huge numbers of government bonds which they intend to cash. The US government as a whole however begins to see a serious problem in 2017.
When the trustees cash the bonds, the money comes from the Treasury. Which is to say higher taxes.
Assuming that the Treasury debt is good, Social Security is ‘fine’ from the point of view of the trustees until 2042. From the point of view of the government as a whole, there is a serious problem as of 2017, when Social Security must begin relying on cashing government bonds in order to cover obligations. The problem gets worse from there as more people retire and live longer.
One way or another the debt is going to hit soon. Putting it on the 10-year projection now isn’t creating a new cost. It is the fruit of 40+ years of government lying about the structure of Social Security.
Assuming that the Treasury debt is good, Social Security is ‘fine’ from the point of view of the trustees until 2042. From the point of view of the government as a whole, there is a serious problem as of 2017, when Social Security must begin relying on cashing government bonds in order to cover obligations. The problem gets worse from there as more people retire and live longer.
Well, if US Treasury securities aren’t good, I’d suggest we have a problem that renders any problem with Social Security trivial.
As for the problem getting worse, that’s not exactly true for the long haul. The immediate “problem” is the baby boom generation and it’s not an indefinite problem.
Well, if US Treasury securities aren’t good, I’d suggest we have a problem that renders any problem with Social Security trivial.
Honestly Jadegold, do read whole comments or do you just pick a snark and glaze out?
I specifically dealt with the Treasury issue in the very next sentence. They are good, BUT THEY HAVE TO BE PAID BY THE US GOVERNMENT.
The plan is solvent FROM THE POINT OF THE TRUSTEES BECAUSE THEY HAVE SECURITIES THAT HAVE TO BE PAID BY THE US GOVERNMENT.
That means that when they start redeeming the securities, the US government has a problem.
And I’m absolutely certain I already said that.
Sebastian: actually, the year in which we start having to sell bonds was estimated (in 2003, can’t find the 2004 estimate just now) to be 2028. Until then, taxes plus bond interest cover costs. That’s not 2042, but it is 24 years away, long enough for all sorts of things to happen. I am not suggesting doing nothing — seriously working on the deficit would be a start — but I do think that it’s not an immediate emergency.
Moreover, it’s not as though the government does not have to pay interest on all its bonds, or redeem them when they mature, and it’s not as though we don’t budget for this. The fact that huge numbers of treasury bonds currently held by the Chinese government will eventually mature does not mean that we will face a huge crisis when those bonds mature (at least, not in and of itself, leaving aside the question whether our debt is too large generally.) We always pay off our bonds, and we always pay interest on them, and why this should be grounds for alarm in the specific case in which those bonds are held by the SS trust fund is not clear to me.
Honestly Jadegold, do read whole comments or do you just pick a snark and glaze out?
I read ’em, Sebastian. But, frankly, I’m seeing the same social security myths trotted out time after time.
I see the ‘Ponzi scheme’ argument, the argument that Treasury securities are worthless, that SS revenues must be stuffed into an old mattress, that the system will come crashing down unless…
What is consistently being ignored is that SS’ demise is predicated on the idea of extremely dismal annual growth for the next 25 years. Yet, when we consider privatization–annual economic performance magically goes to rates exceeding single year highs. Ignored seems to be the fact if annual economic growth averages what it has for the past 75+ years, SS remains solvent indefinitely.
I’m not arguing that Treasuries are worthless in any general sense, I notice (not argue because it is a fact not a conclusion) that Treasuries held by a subsidiary of the issuing party do not represent savings to the issuing party. That party is the US government. The trustees can claim ‘solvency’ based on US treasuries if they want, but when looking at the accounting of the government as a whole it just isn’t true. The treasury money has to be paid from somewhere. That somewhere is increased taxes.
Sebastian: we can do the accounting in either of two ways: counting US bonds held by a US agency (a) neither as assets of the agency nor as liabilities of the government, or (b) both as assets of the agency and as liabilities of the government. If (a), the Social Security Trust Fund has no assets and will have to start being paid for from general funds in 2014, but luckily this accounting change will reduce both the Federal debt and its interest payments by enough that the reduction in interest payments takes us through 2028 and the reduction in debt through 2042, without our having to budget for more money than we have already allotted under the old system. If (b), of course, my original figures stand.
Obviously, the deficit itself is too high. We are already paying (in my opinion) too much money in interest payments, money that could be spent a lot better. But the fact that bonds will mature and have to be redeemed, and that they have interest rates, is not a new and unanticipated expenditure; and it makes no more sense, to me, to say that taxes will have to be raised to pay for social security in particular than it would to say that we face an urgent fiscal crisis since my ten-year bond is about to come due, and hey, it has to be paid for somehow.
“and it makes no more sense, to me, to say that taxes will have to be raised to pay for social security in particular than it would to say that we face an urgent fiscal crisis since my ten-year bond is about to come due, and hey, it has to be paid for somehow.”
It is quite possible that if enough bonds are issued and then redeemed that they could cause an urgent fiscal crisis. Why is that mysterious at all? If I take out a $100,000 loan to be repaid in 10 years at 5% interest I have a liability on the books now that causes a financial crisis for me if I don’t attend to it for ten years. We have been not attending to it for 40 years. BTW we haven’t begun to discuss replacing the money that Social Security cash is putting into the general fund for general expenditures now. Nor have we started to talk about the much uglier Medicare disaster.
Conceptually I think it is stupid to have poorer young people pay the retirement of older rich people–I would make changes differently.
Having the young and poor fund the retirement of the old and rich would indeed be perverse, but it’s more in line with historical norms.
And thus ends the bolding.
Sebastian, if you implement a private accounts system the US government will probably borrow the “transition cost” or whatever you call it, and since at the same time the companies that will handle the former payroll taxes will have a lot of money to lend, the predictable result is that most of the private accounts money will be invested in US bonds (it has happened in every country that has implemented this system). That is to say, the same outcome as now but even more expensive (just check how much charge for administration costs the companies that are doing this elsewhere). Now, if the US has a long term demographical problem (that is to say, retirees will keep growing as a share of the population forever, instead of stabilizing at a sustainable level), paying this cost may be worth it, but is this the case?
The notion that there is a big problem because the trust fund holds treasury securities rather than equities and corporate bonds is an illusion.
Suppose SS did hold corporate securities. Then the rest of the government would have borrowed on the capital markets what it now borrows from SS. So it would have, at the approrpiate time, to pay off those bondholders instead of the SS trust fund. Same thing.
Ah, you say. But if it defaults SS still holds its investments. First of all, as others have mentioned, if the govt starts defaulting on its debt all that Microsoft stock won’t help. Second, assuming the govt is in financial difficulties, what is to stop it from appropriating the SS portfolio and using that to pay off its debts? Nothing.
Whatever you think should be done with SS, please drop the notion that there is something uniquely bad about the investment procedure.