Social Security: Cost, Solvency, Debt and TF Ratio
The Table above is extracted from Table VI.F9 in the 2013 Annual Report of Social Security. It is radically simplified because I want to make a very simple set of points, ones that have little or nothing to do with the proper policy approach to Social Security or to the adequacy of the model that produces the specific numbers at hand. And fair warning to long time contributers/commenters, this post will be strictly moderated to avoid hijacking. I don’t care if it seems less important than 80 cents a week or the impending doom facing Social Security because of inadequate modeling. This isn’t about that and off topic comments will be treated accordingly.
That aside what we have here is a simple projection of Social Security total Cost in current (non-inflation adjusted) dollars over the standard 75 year actuarial period. And not surprisingly that total goes up year over year due to a combination of inflation, population growth, and shift between demographic cohorts. What doesn’t change a lot, at least after mid-century is cost as a percentage of GDP which stabilizes around 6%. If that was relevant to this post, which it isn’t.
Now what does it mean to say that Social Security is ‘solvent’? Well in terms internal to Social Security financial reporting ‘solvency’ means ending each year in the given period with a Trust Fund balance equal to 100% or more of the NEXT year’s cost. As most of us know by now the Social Security Trust Funds are required by law to maintain its reserves/balances in interest earning securities guaranteed as to principal and interest by the federal government. Which in practice means Treasury Bonds, and in the case of Social Security so-called Special Issue Treasuries. These Special Issues are legal obligations of the Treasury and backed by Full Faith and Credit of the United States, there is nothing ‘Phony’ about them. On the other hand simple inspection of this table reveals something rather interesting: under conditions of ‘solvency’ as defined they would never need to be redeemed on net. Because the requirement to maintain a 100% minimum reserve (aka a Trust Fund Ratio of 100) means that a minimally solvent system would have to roll over 100% of its balance each year and then augment THAT by the amount of projected increased cost. Which makes the Trust Fund an odd kind of obligation or burden on Treasury or to future taxpayers, although it represents real borrowing from the economy and does require debt service under ideal circumstances it would never need to be paid down. In fact any such repayment which drove the balance under 100% of next years cost would make the Trust Fund ‘insolvent’ by definition. Which leads to the rather counterintuitive conclusion that a Trust Fund with a balance of ‘only’ $33 trillion in 2090 would be ‘bankrupt’ and ‘flat broke’. This despite the fact that the ‘shortfall’ represented less than 1% of cost. To say the least this does not resemble the kind of budget balancing households or even corporations do.
Now to another curious point. The Special Issue Treasuries that comprise the Trust Funds are counted by the Bureau of Public Debt as ‘Intragovernmental Holdings’ and indeed make up over half of all such holdings. And in turn ‘Intragovernmental Holdings’ are with ‘Debt Held by the Public’ the two components of ‘Total Public Debt’ and its close analogue ‘Debt Subject to the Limit’. What this means is that under conditions of Social Security solvency as defined the U.S. would never be ‘debt free’ or even want to be. That is while we might want to drive ‘Debt Held by the Public’ down to much lower levels than today this doesn’t mean that we would ever drive ‘Total Public Debt’ down to that degree. In fact by 2090 the U.S. will need to be maintaining $33.3 trillion in Public Debt simply to ensure that Social Security is ‘solvent’. Even though none of that $33.3 trillion would ever need to be paid off on net.
Is any of this important in a real world sense? Perhaps not. Except for the fact that it turns a lot of concepts of ‘Public Debt’ and ‘Debt Subject to the Limit’ on their heads. Because we could balance the General Fund tomorrow and even start to pay down ‘Debt Held by the Public’ and yet never see net reductions in ‘Total Public Debt’ simply due to the need to maintain a 100% Trust Fund Ratio. That is by 2074 we need to be $17.038 trillion in debt, or roughly the same amount we are in today, simply to make Social Security ‘solvent’
I didn’t think solvency was legally required (the projections that say that if nothing is done, in 2033 the Trust Fund is depleted and benefits would have to be cut to 77% or whatever, suggest that a 100% TF ratio is not required). Since Social Security always has money coming in, predictably, there’s no real need (once the baby boomers are gone) to have a full year’s reserves (especially since the government could always make up for any shortfall). It might be a good idea, but maybe not always.
uhh..
is the observation that “solvency” is not the same thing as “actuarial solvency” off topic, or do words still matter?
on the other hand: “the U.S. will need to be maintaining $33.3 trillion in Public Debt simply to ensure that Social Security is ‘solvent’. ”
is certainly turning words on their head.
“33.3 trillion” is scare words. Dean Baker would be the first to point out that if the NY Times used this figure without explaining it in terms that the average reader could understand, it would be misleading.
So how about,
The Congress has borrowed money from Social Security. It owes interest on that money. In the normal course of things, IF the Social Security tax is set at a level that maintains Social Security’s ability to pay for “promised benefits”… at a cost that is so low I am forbidden to mention it….
the ability of Congress to borrow “excess” Social Security funds will be severely limited.
But as long as Congress does not decide to pay back all of the money it has borrowed so far, the interest on the money it has borrowed so far will be enough to maintain the Trust Fund at the “one year’s reserve” level forever with no actual cash being taxed or “borrowed from the public.”
And (But?) SOME of that interest can be cashed to pay for part of the benefits people will want to collect. This is interest that Congess OWES, not “money that Social Security borrowed..” it was borrowed FROM the people who paid for their Social Security. Keep that straight in your mind and you won’t go far wrong.
The amount of that interest actually paid in cash each year will be about eight tenths of one percent of payroll.
Pardon me if I mention Achilles and the Tortoise: An ancient professor of philosophy argued that Achilles could never catch the Tortoise. His argument consisted of words arranged in a way that appeared to mean something, but in fact hid an assumption that rendered the words meaningless. Two thousand years of philosophy professors failed to notice the word trick (you were invited to look at only those instants of time that occurred before Achilles caught the tortoise). Normal people didn’t worry about this. Real life was more interesting to them than the word games of professors.
It is rather a different thing when the word games are being used to fool people into thinking that Social Security represents a huge and growing debt. And it is tragic when our favorite professor has ensnared himself with his own “logic” (words) to end up giving aid and comfort to the enemy.
Mike B
The Trustees are mandated to report to Congress when Soc Sec fails its official tests for Acturial balance and recommend that they take action. Obviously they cannot force Congress to take action but their language deploys the concepts of “pass” and “fail”
Dale
Yes you are dangerously close to going off topic here. Words do matter. But oddly they don’t only matter along and only along lines important to you.
I am using ‘solvency’ in the way I am to make a point. Which point isn’t your point perhaps. But for once in the long series of Social Security posts I have put up over the years maybe we could make the topic what I am posting on? It is not like you didn’t take the opportunity to put up thirty plus comments on the last post. I think most people get it.
In other words this might in part be word games. But then again it is my game and my ball. And if it comes to it my whistle.
Also current law is not as specific on the effects at Trust Fund Depletion as most people, myself included often would have it.
What is clear enough is that the Commissioner and his fellow Trustees do not have the ability to borrow and so the flow of funds out would be limited by the flow in, which translates to a 23% shortfall. On the other hand there is equally no provision in law for the Commissioner to simply adjust the Schedule.
I had a conversation with Andrew Biggs, who between his stints as a privatizer at Cato and AEI had positions at Bush’s SSA including the number 2 slot, where he related the results of a top level staff conference on this topic. The conclusion was much the same as Treasury reached in regards to a total shutdown, checks would have to be issued in sequence as money came in making not for smaller checks but instead for delays until Congress actually took action to boost revenues or officially change the schedule. In other words chaos and not some orderly (if almost equally disruptive) reduction in check amounts.
This by the way would actually create a short term real ‘unfunded liability’, something I have earnestly assured you cannot exist. Well it can’t in the way and to the magnitude that the Bad Guys would have it and everyone agrees that we will never actually reach that point (Congress will do SOMETHING) but it is one of the oddities if you get right down into the weeds and thickets. Which I find interesting even if some deem it unimportant.
My will includes a provision to send the U.S. Treasury general fund enough money to pay my portion of the public debt (assuming the markets don’t crash severely in the meantime). Since as I learned from this post (thanks) about half that debt is held by the SSTF as its reserve, I guess I could have allocated half as much money for that purpose. I won’t change my will but from here on will privately consider half that amount being my share and the other half being in honor of Mr. Webb.
My summary in case there’s a test:
1) The SS system defines its solvency as having next year’s expected payout in reserve.
2) That reserve can’t be hidden under a large mattress, but must be invested in interest-bearing U.S. Treasury bonds.
3) Over time, with assumed inflation and productivity growth, the reserve amount will grow into a huge number (although by that time a dollar bill might be worth what a penny is now and the currency system might be changed so the number is not so large in the future currency – similar to old francs and new francs).
Speaking of strange amounts of currency, I am so tempted to tell the story of my friend Mario’s visit to Italy 50 years after his father came over on a boat at the age of six, and negotiating a fee with a taxi driver when his (Mario’s) uncle suddenly showed up and began yelling at the taxi driver (in Italian), “50,000 lira! This is my nephew, the son of my brother who went to America on a boat when he was six years old and I haven’t seen for 50 years, and now his son comes back and you gonna charge him 50,000 lira?” At which the taxi driver waved his hands in protest and exclaimed, “I didn’t know! I didn’t know!” – but that might be off-topic.
Jim V
maybe not as off topic as you think.
but i do hope you were also kidding about leaving your money to the debt. please don’t do that. the debt is as close to an imaginary thing as it is possible to be under the law and business as preacticed by governments.
leave it to a good cause, or a hell of a wake.
at a rough guess, the cost of OASDI in Webb’s chart is about 30 times as high in 70 years as it is today.
A days wages today is about $150. 70 years ago it was about $5.
roughly, more or less. but not one person in a hundred will think that when they hear “we will need to be 17 Trillion Dollars in Debt to keep Social Security solvent.” That’s about 1/3 the size of the current Trust Fund in comparable dollars.
but that might be off topic.
Let’s assume we were able to balance the budget and pay off the debt. Would not the $17 Trillion intragovernmental debt from SS be balanced by a $17 Trillion surplus in the debt held by the public, giving a total,public debt of zero?
Jerry, hmm no.
Total Public Debt = Intragovernmental Holdings + Debt Held by the Public
“Pay off the debt” and ‘Debt Held by the Public’ goes to zero.
$17 trillion – $0 = $17 trillion.
The government can pay down debt but has no way to run an actual net cash surplus. Nor would there be any point in doing so. Instead reducing Debt Held by the Public to its lowest possible level consistent with having the U.S. remaining the world’s effective currency of record and vehicle for ‘flight to safety’ means opening up that much more room for investing in American human productivity in all its forms. But this process can only go so far.
I have asked the question on numerous occassions and never gotten even an attempt at an answer. Given the multiple roles played by the American Long Bond and 10 Year Note in the world economy how much of Debt Held by the Public is truly structural and so should NOT be paid off? How many dollar equivalents does it take to allow world commerce in commodities of which many of the most crucial are still priced in dollars? Could the dollar simply be a vehicle of account while not actually existing in circulating form? Or do we simply revive the $100,000 bill and supplement it with the $1,000,000 and $1,000,000,000 bills so that people can still execute oil contracts?
Jim might well have been just being facetious but actually brought up a key point, there is a number for what I call ‘Real Debt’ out there. But nobody seems interested in putting it into black and white (or in accounting terms black and red).
_____________
And no Dale your comment is not totally off topic. Unlike the one that claimed that this whole line of inquiry made me the modern equivalent of a 50s era Com-Symp (“Biggs and Krasting laughing at us”). If you like I can put up a ‘Social Security Open/Alternative Thread’, in fact that was my initial intent when I planned out this particular post.
Bruce,
What you are saying then is that the government must run a deficit sufficient to absorb one year of SS cost — $17 Trillion in 2074 — if the debt held by the public is zero.
Jerry
to make sense of the number you must remember that the 17 Trillion is in 2074 dollars. That’s about 2 Trillion in today’s dollars.
Then you need to remember that the 17 Trillion is “one year’s cost of Social Security benefits”… the amount of money that 50 million (more) people are going to need to live on for that year. Money that they paid themselves, if you acknowledge the “time cost of money.”
Then you need to remember that the 17 Trillion is about 16% of “wages below the cap.” It is as if you decided to set aside about 16% of your wages to have an “emergency reserve”in the bank.
Then you need to remember that “wages below the cap” is only about 30% of GDP. So “the country” is setting aside about 5% of its wages to have an emergency reserve in case the economy goes bad for not one year… but more like ten years… that is how long it would take to spend down “one year’s income” in a “normal” recession that reduced tax income about ten percent a year.
then you need to remember that it is NOT Social Security “causing the debt.” Social Security did not, does not, borrow the money. The Congress borrows the money FROM Social Security. Social Security cannot force Congress to lend it money.
then you need to remember that congress has ALREADY BORROWED THE MONEY. The 3 Trillion it has already borrowed from Social Security will grow to 19 Trillion by 2074 from interest if Congress DOES NOT PAY IT BACK.
In fact, over that time Congress WILL pay back some of it in the form of interest payments that will… in a sustainable solvency scenario… make up about one percent or a little less of what Social Security spends on benefits…. reducing the need for the payroll tax to be about one percent higher.
But remember this is interest on money that Congress BORROWED. Money it used to buy things the Congress wanted WITHOUT RAISING TAXES to pay for them.
So the 17 Trillion is not “needed to keep SS solvent.” SS could function just fine on strict pay as you go… with a tiny reserve and raising and lowering taxes from time to time to take care of the booms and busts in the economy.
But Congress is glad to be able to borrow the money and not pay it back. Remember that “rolling over interest” is the same as borrowing the money to pay the interest.
So while the 17 Trillion might make you gasp, and will be used by the Krastings and Petersons to make you think Social Security is running up a huge debt that will be a staggering burden on our children..
It is no such thing. It is a simple consequence of the normal way money works… especially inflation… and represents a FREE LOAN to the people who pay the INCOME tax. Which very conveniently can be kept on the books… with no money changing hands… and act as a “reserve fund” for social security to smooth the normal ebbs and flows of income vs benefits.
Getting you to run around screaming the sky is falling because someone showed you a big number has been the nature of Petersons game since as long as I can remember.
Please tell the professor to put down the gun.
I am not contradicting him. I am explaining his lecture to the class.
It’s what I used to do for a living.
No that is to confuse debt and deficit.
Remember that in order to build and maintain a 100% Reserve Social Security has to run a surplus as defined. As it happens that surplus is ‘invested’ in securities that are scored as part of Public Debt. But whose ‘purchase’ if anything reduced the need for Debt Held by the Public.
Once Social Security is at Sustainable Solvency (has a 100% reserve projected for each year forward) maintaining it simply requires taking some of the interest accruing on those Intragovernmental Holdings/ Public Debt and retaining it in the Trust Fund. That portion of the interest retained to maintain 100% of next years cost does not require an actual cash flow from Treasury, instead it just comes in the form of newly issued Special Issue Treasuries (fiat money if you will).
Now under normal economic conditions maitaining Trust Fund balances at 100% of cost does not require taking ALL the accrued interest (though it wouldn’t matter much if it did), instead there is some, perhaps 40-60% left over. Now since there are reasons not to OVERFUND the Trust Fund, that remaining interest would need to come in the form of cash and so narrowly reducing the amount of cost coverage needed from FICA and tax on benefits. But only by maybe 2-3% of total cost, the other 98% being truly Pay-Go by workers and certain retirees.
There is absolutely no reason why this particular cash flow has to result in deficits in any sense, it would be perfectly possible for the General Fund to have positive cash flow from ITS receipts sufficient to pay what is in the end discounted interest on principal that it will never have to pay back. But as Dale would point out “They borrowed the money”.
But this is the paradox of federal debt and deficit reporting. They don’t move in lockstep. For example in FY1998 the General Fund ran a small cash surplus and Social Security ran a larger one. The result of the former was that Debt Held by the Public actually shrank while Intragovernmental Holdings grew by a larger amount. With the net effect that total Public Debt grew by the difference. http://www.treasurydirect.gov/NP/debt/current
There is no reason why you wouldn’t have the same outcome with Sustainable Solvency: a reserve growing at the same rate as Cost, an addition to total Public Debt of that amount of reserve growth, the interest which funds that growth scoring as a surplus for overall deficit calculations, the left over interest ALSO scoring as a surplus for that purpose, EVEN as it results in a negative cash flow from Treasury to Social Security.
Perma-surplus plus perma-debt plus perpetual negative cash flow. It only sounds crazy if you have internalized the idea that debt is a simple sum of deficits. Well that is true for Debt Held by the Public, it is not true for total Public Debt (and the folks who adopted these oh so similar terms for two different debt metrics have a lot to answer for in my view).
Bruce,
It isn’t the $17 trillion number that bothers me. I understand that it is a 2074 number. I just used it because it is the number that you used. What concerns me, or I misunderstand, is that if the debt to the public is zero, that the government has to run a deficit if SS is to meet its obligation to have one year of costs held in a reserve in Treasuries whatever that cost is, $1 or $17 trillion. What happens if the treasury refuses to issue bonds to SS? Do SS have another mechanism to hold a surplus?
It also doesn’t make sense to me that the government does not have a mechanism to run a surplus if the debt to the public is zero. It has to spend the money?
I’m sorry if this is getting off topic, but I am trying to follow up and understand what you are telling me.
Of course, also I realize the probability of the government paying off the debt to the public also hovers around zero, rendering this discussion academic.
Bruce,
OK, after reading your last post, I do understand that we don’t have to run an annual deficit equal to one year SS costs, and that it is possible to have a budget surplus, yet have an increase in the debt.
Okay, to achieve Sustainable Solvency the combined TFs must be equal to 100% of the next year’s scheduled benefits. And yes, if in 2090 benefits are $33.285 trillion, a TF of a similar amount would achieve what is defined as Sustainably Solvent.
But what is really going on in 2090? From V1.F9 we get non-interest income of $24.288 T and costs of $33.285T for a net of -8.997T. If the TF is then at 33.285T and earns 5.7% on that (TF projection) the % income will be 1.9T. This still leaves a hole for the year of 7.1T. For the TF to be really solvent it would have to earn 27% on the 33T, or it would have to be $158T to insure the needed 9T at the forecast rate of 5.7%.
So 2090 is a mess (so are the prior years). The minimum TF to achieve the definition of Sustainable Solvency leaves a hole. So how is it sustainable with such a big hole? Maybe it starts with the definition.
It would take a radical change in our willingness to pay taxes to actually run a sustained surplus. Still, if that came about, then we would need to change the laws regarding requiring SS to invest in (Special) Treasuries.
No big deal.
Or raise FICA about 80…oh wait, off topic. 🙂
And speaking of off-topic, Krasting is.
What certain friends and foes of Social Security don’t seem to grasp is that you don’t have to address the totality of the policy debate around Social Security in each and every post about it. The original Social Security series I wrote for AB consisted of 44 posts and a summary and only a few focused either on the amount of the actuarial gap or ways to close it. But just about every comment thread went right there, even if it meant swamping the point I was trying to make.
I don’t know that I am a Professor, but I sure as hell resent even colleagues jumping in to “explain” my lecture to MY class, apparently under the belief that they are too stupid to understand that technical language on any given subject doesn’t always exactly mirror ordinary language. This is particularly so when the lecture hall in question is always open to that colleague should he want to make his own presentation in his own way. As oppposed to “helping” me out.
And Krasting if your comment was on topic I would be forced to point out that any conceivable path that actually ended up with a 100% TF ratio in 2090 would perforce also produce non-interest revenue in excess of the amount currently projected in Table VI.F9. Making your $7.1 trillion remaining gap a case of apples and oranges.
But since you are off topic I won’t explain the simple logic and arithmetic that gets us there.
And Jerry I think the source of the confusion is fairly simple.
There are two ways to build Trust Fund principal. One is to have non-interest revenue exceed cost which then results in positive cash flow to Treasury. If there was no existing Debt Held by the Public you would eithr need General Fund deficits of that same amount to absorb the extra cash or reduce General Fund revenues to compensate or have those flows screw up money supply in ways not really under purview of Treasury. But the truth is the days of postive cash flow from SocSec to Treasury are in all liklihood over and any building of TF principal will have to come from the other method.
That method is retention of some or all the interest accruing on the existing Trust Fund. Although this scores as a surplus for budget scoring it does not in fact generate cash flow to Treasury. In fact under conditions of Sustainable Solvency a portion of accrued interests scores as surplus while the rest covers the non-interest revenue cash deficit. This isn’t just my take, the Chief Actuary ‘explained’ it in an article for the 75th Anniversary issue of the Social Security Bulletin http://www.ssa.gov/policy/docs/ssb/v70n3/v70n3p111.html
“However, the occurrence of a negative cash flow, when tax revenue alone is insufficient to pay full scheduled benefits, does not necessarily mean that the trust funds are moving toward exhaustion. In fact, in a perfectly pay-as-you-go (PAYGO) financing approach, with the assets in the trust fund maintained consistently at the level of a “contingency reserve” targeted at one year’s cost for the program, the program might well be in a position of having negative cash flow on a permanent basis. This would occur when the interest rate on the trust fund assets is greater than the rate of growth in program cost. In this case, interest on the trust fund assets would be more than enough to grow the assets as fast as program cost, leaving some of the interest available to augment current tax revenue to meet current cost. Under the trustees’ current intermediate assumptions, the long-term average real interest rate is assumed at 2.9 percent, and real growth of OASDI program cost (growth in excess of price inflation) is projected to average about 1.6 percent from 2030 to 2080. Thus, if program modifications are made to maintain a consistent level of trust fund assets in the future, interest on those assets would generally augment current tax income in the payment of scheduled benefits.
A cash flow shortfall, therefore, is only a problem if it is large and persistent enough to cause the trust fund reserves to decline over time toward exhaustion”