Wednesday, April 1, 2009

What is the debt-to-GDP ratio if we include the Social Security and Medicare trust funds?

As deficits rise and the Baby Boomers retire, the burdens of government spending on the economy are becoming more obvious. The most common measure of where we are and where may be going is the ratio of debt to gross domestic product. This ratio, more so than the simple dollar value of the debt, indicates the economy's capacity to support government borrowing.

But an important question is how to measure the debt to GDP ratio. All measures include the publicly held government debt, which means Treasury bonds held by everyone from Wall Street investors to ordinary folks saving for retirement. However, it's not clear whether we should includes intragovernmental debt, which is primarily composed of the Social Security and Medicare trust funds.

Many on the left don't like to count the Social Security and Medicare unfunded obligations – which total in the tens of trillions of dollars – as true "debt." That's fine. As some have correctly pointed out, future Social Security and Medicare benefits in excess of what the trust funds can finance are obligations, not liabilities, and can be changed at any time. (It's ironic that many on the left don't wish to change them, but that's another story.) That is, implicit debt isn't the same explicit debt.

But folks on the left are also adamant that the Social Security and Medicare trust funds are true debt – as good as any debt issued to Wall Street, as (say) Dean Baker argues. Ok, that's fine.

But if so, shouldn't we count Social Security and Medicare debt when we calculate debt-to-GDP ratios? I can't see why not. And when we do, it tells a very different picture about the evolution of public finance in the U.S. The chart below is drawn from OMB historical data from 1940 through 2007. It shows both publicly held debt – the kind we usually think about – as well as intragovernmental debt, which includes the Social Security and Medicare trust funds as well as other government trust funds (the highway trust fund, etc.).

As of 2007, publicly held debt equaled 37 percent of GDP. Not a problem, except that CBO projects that the recession and financial crisis coupled with President Obama's spending plans, by 2018 the publicly held debt could more than double, to 82.4 percent of GDP. Most people would consider that a problem.

But then add to that the intragovernmental debt that many people insist must be treated as just as "real" as publicly-held debt. There don't seem to be good projections of future intragovernmental debt, but let's just assume that the ratio to GDP remains the same. (This isn't implausible; while the Medicare trust fund is winding down, the larger Social Security trust fund is projected to peak in the 2020s.)

In that case, by 2018 total government debt will reach 111 percent of GDP. That level would put us somewhere between where Sudan and Jamaica are today, and fifth from highest globally. While there isn't a strict threshold at which debt becomes unsustainable – what matters is the growth rate of nominal debt relative to the growth of nominal GDP – even the IMF and World Bank's debt sustainability framework adjusted for a strong country like the U.S. would frown on debt at that level.

Moreover, the total debt level as of 2018 would be only slightly lower than the historical peak of 121 percent in 1946, when we had just finished fighting a war around the globe. But unlike the debt issued to fight World War Two, this debt won't be incurred in the cause of freedom but in the cause of political convenience – a fight not to defeat foreign enemies, but political opponents who say that long-term spending must be reined in.


Bruce Webb said...

Andrew the difference is that under Intermediate Cost assumptions none of the principal in the Trust Fund needs to be redeemed prior to 2023 when Income including Interest fails to meet cost. This is a far different challenge than the close to $739 billion in Treasuries held directly by China or the billions more that Setser believes are held on China's behalf in offshore banks. If China stops buying treasuries or worse starts selling them the Treasury is in bad trouble.

On the other hand the Managing Trustee of the Social Security Trust Fund is not going to stop buying Treasuries nor is he going to start selling out his portfolio. Maybe in large part because the Managing Trustee shares a desk chair with the Secretary of the Treasury.

The fact that the Special Treasuries are not marketable, are always held to maturity, and can and will be freely rolled over until they are needed (if they are, I have not given up on Low Cost) is why we need not consider them in the same category as say the 10 year bond held by the public and foreign entities. The fact that the same guy controls both sides of the equation is just gravy.

We just don't have the same risk exposure from Intra governmental holdings.

Jim Glass said...

Geeze, I've been going over this in the comments over at

The gist of one person's opinion:

The Treasury reports these numbers for various debt measures...

[] Debt held by the public: ~ $7 trillion.
[] Promises made by law and represented by intra-governmental debt: ~ $4 trillion.
[] Promises made by law but not represented by intra-government debt: ~$42 trillion.

Total: ~~ $53 trillion (and rising rapidly! wow!)

The first two are commonly added and officially reported as the gross debt of the US, the "national debt", (the subject of the debt limit, etc.), today ~ $11 trillion. IMHO that number is very arbitrary and misleading.

ISTM that...

[] The "debt held by the public" is the prime number that matters. That debt is incurred ... payment of it is guaranteed by the Constitition, so there is no getting out of it ... it has to be financed with taxes now and into the future ... and the credit rating of the US depends upon the ability to service it -- and a *whole lot* depends on the credit rating of the US.

[] The $53 trillion of total unfinanced spending promises of the US is also very meaningful to look at -- because they will be rolled into the debt held by the public as time passes, as the law stands. Thus, it sets expectations about future credit rating of the US, which matter a lot, especially to investors in long T-bonds.

But the bulk of the $53 trillion can be changed by policy, such as by modifying spending programs like Medicare and Social Security, and by modifying the pct of GDP collected via taxes, so it is (or should be) the policy-driving number. (If it’s not affecting policy then it should be the flashing red warning light number).

[] I just don’t see much independent value to the gross federal debt figure of $11 trillion. It’s neither fish nor fowl, being the debt held by the public plus a small subset of the additional liabilities that exist in the form intra-governmental debt, only $4 trillion of >$40 trillion.

Thus, it exaggerates by two-thirds the amount of Constitutionally guaranteed debt the US “can’t get out of”, but understates by near 90% the additional liabilities the US has promised to pay and which as things stand will roll into the public debt in the future..

Where's the logic in that? The nature of the intra-governmental debt is identical to that of the $42 trillion of spending promises not reflected by intragovernmental debt. They all just reflect an amount Congress has promised to pay -- but the promise is not binding, Congress can delay the payment or get out of it entirely by changing the terms of Medicare, Social Security, and its other programs that use these bonds as a tally counter.

Since the nature of this debt is indistinguishable from the rest of the $53 trillion (in excess of the debt held by the public) what’s the benefit of distinguishing it? And counting only 10% of it.

As a clear, practical example of what I'm talking about, the recent Frontline show on the national debt used the $11 trillion, gross debt number for it. When they were criticized by someone for using that instead of the (mere) $7 trillion, the producers responded on the Frontline web site to the effect that "the $11 trillion, by counting promises of future spending for Medicare, SS, etc, that are reflected in the intragovernmental debt, gives a better overall picture of unfunded spending committments."

But it gives a hugely understated picture of them! E.g. the intragovernmental debt covers Medicare only until 2018, when the Medicare trust fund bonds run out. Say I become eligible for Medicare in 2019. Is Frontline telling be the govt hasn't promised me Medicare benefits for 2019 and later because there are no intragovernmental bonds to show for them? Hey, I need to know now!

If you are counting "spending promises", there is no logic at all in counting only the ones memorialized by intragovernmental bonds, and so not counting Medicare promised after 2018 as if that spending promise wasn't made. The tally should be the full $53 trillion.

If you are not counting future spending promises, but only binding, interest-bearing debt, then one should count none of the intragovernmental debt. The tally should be $7 trillion.

So for me the numbers that matter are $7t and $53t. Run them out as a pct of GDP as you will.

Forget the $11t (and intragovernmental debt) it is a misleading red herring that grossly understates the "promises" it purportedly accounts for.


Andrew G. Biggs said...

So the trust fund debt isn't an issue until it needs to be repaid. Ok, that's fine. And at that time, repaying the trust fund bonds will compete with repaying bonds held by the public. Since some of those bonds are currently held by the public, and more will be as time goes on, in a perfectly rational market (which I admit we don't have) the burden of repaying trust fund bonds will be reflected in the pricing of publicly held bonds. Since trust fund bonds are priced based on market prices, they'd eventually carry the same risk premium.

Andrew G. Biggs said...


I agree with a lot of what you said. Dean Baker often claims that cuts in Social Security benefits are a "default" (his word) on the trust fund, when in fact that's not the case at all. There would be no need to default on the trust fund bonds since the government could instead simply reduce benefits.

All that said, for people who do put a lot of credence into the trust fund -- that is, people further to the left than I am -- it's important to be consistent. If the trust fund is a firm, legally binding commitment of the government -- as the left likes to emphasize, to counter people who says it's a 'phantom asset', etc. -- then it should be counted as such. If you do, then the debt/GDP ratio is a LOT higher than currently reported, and approaching very dangerous territory.

Jim Glass said...

Andrew, re your last comment, I agree with you completely.

BTW, glad to see you are back at it here. I thought you might be shifting your effort to your other site (not to mention your family!).

John Bailey said...

Actually, I think the real debt number has to include some estimates for bank deposit guarantees (FDIC) and mortgage guarantees (Fannie/Freddie).

The obligations number should include some provision for rising Medicaid costs. Using a CBO projection, I put that number at
$16 trillion (above current spending), but you could make up any number you wanted and come up with some kind of defense.

The more immediate question is how the immediate increases will be financed. The Administration projection shows the deficit doubling over the next 10 years while the CBO (also somewhat optimistic) projection shows it tripling.

The administration estimate places the increase in the next 2-3 years at $3 trillion (from Oct 08) while the CBO show $4 trillion.

I don't know how either of those increases happen.

Jim Glass said...

A look at the 2008 Financial Report of the United States Government shows the govt's liabilities increased by $3 trillion last year, a bit more than the official $455 billion deficit.

That's from $455b for the deficit, plus $550b for federal employee & veteran pension liabilities, plus $2,023 billion for SS/Medicare (and leaving out all the other little odd and ends)

For perspective, the operating cost of the entire US govt was $2.9 trillion -- so the govt during the year ran up an increase in liabilities that was larger than the govt.

One suspects this can't continue indefinitely.

The total number for this liability is now >$53 trillion.

I'd like to see the size of this debt charted against GDP -- or for that matter, world GDP. I think it pretty much matches world GDP at this point.

Tom Harris said...

Bruce, Jim, Andrew, John . . .

This is the first time I have seen this blog . . . and you guys are "beyond my raisins" on this issue. I just got to the debt (I guess ratios will be next). Anyway . . .

I have only been interested in and studying politics and economics for about a year; and it is tough for an Independent like myself to navigate through all of the partisan blarney, bickering, and blaming and to get to the facts of the issues and to arrive at rational (i.e. non-partisan) conclusions and solutions.

I recently (and for the first time) looked at the history of the national debt (I searched the Bureau of the Public Debt, U.S. Treasury Department and studied the history of our debt from 1791 through today).

I did not know what to expect; but this is the way it looked to me: in general, the debt has been creeping up the entire time (for over two-hundred years). Whenever we have a war on, it seems to jump up faster than usual. And it has reached an all-time high every single year since 1957 (i.e., it has gone up every year since then); and (from other sources, it seems that) we are poised to see the greatest increase in the debt in history during the next budget period.

The interest on this debt is seemingly (to me) going to be staggering and stultifying.

Last year (with Bush) there was (another) very significant increase (from a 9 trillion dollar debt to a 10 trillion dollar debt); and this year (with Obama) it appears that we will see the most significant increase ever (the debt is already up to 11 trillion and the fiscal year is only half over). If someone would do a rigorous (i.e., unbiased, non-partisan) statistical analysis of the data and would account for the party in power in the white house and the party in power in congress (often the opposite party), wars, recessions and depressions, lag-time between policy changes and effects, etc. . . . my hunch is that there would be no statistical difference between the two parties as to which party creates the most debt (unless, perhaps, it is when the party in power in the white house also has a significant advantage in both houses of congress--my hunch is that this conflation of power turns the steady creep of escalating debt into a sprint . . . as it seems to be doing right now).

If we go through with the current budget, it seems to me that taxes--in one form or another--will have to go up steeply and/or the inflationary act of printing extra money to pay this debt and interest will have to take place.

No wonder China is lecturing us about our historically large spending spree and budget plans--they own so much of our debt and will get paid back in dollars with reduced value (if we continue with the bail-outs and stimulus spending and go through with the current budget).

Anyway, this is the way it seems to me.

Btw, Democrats and Republicans need to stop the partisan posturing and deal with reality (perhaps we should have paid more attention to Ross Perot). The GAO, the OMB, and the Treasury are all--right now--warning us that our debt is projected to increase dramatically from our current historical high to unsustainable levels if entitlement programs (Medicare, Social Security, etc.) are not reformed.

My conclusion from this quick and preliminary study:

"Power corrupts; absolute power corrupts absolutely." Lord Acton

(and he did not specify that a particular party affiliation causes this phenomenon)

My question number one: is this a real problem (as it seems to be--and an even greater one if one assumes the $57 trillion figure cited by Jim Glass above to be the more realistic number to be considering) . . . or merely a paper phenomenon that has been going on for our entire history and will probably continue (and with no dire consequences)?

Second question: in what direction do we look for a solution (if this is a real problem and not just a paper phenomenon)?

Just musing,