Fed Treasury Holdings, ‘Real Debt’ and ‘Real Debt Service’

It is not difficult to determine a dollar figure for Total U.S. Public Debt. In fact you have to round UP from the Treasury’s Debt to the Penny and Who Holds it web application which tracks that number daily. As of end of business Thursday that total was $17,472,051,696,926.14. Which is a lot of money however you slice it but doesn’t really get to the issue of ‘sustainability’. First this number includes Intragovernmental Holdings amounting to $4,993,180,664,362.07 which while are certainly full obligations of the U.S. government and backed by Full Faith and Credit or in fact obligations that are under the control of that government and are largely in held as rolling reserves for various programs that ideally will never be redeemed in full. For example over half of that $4.9 trillion is the $2.8 trillion held by the Social Security Trust Funds which, if prudent and necessary steps were taken to shore up its ‘sustainability’, might never have to be redeemed on net and indeed would need to grow over time. So it would make some sense for calculations of ‘Real Debt’ to use the remainder, which Treasury tracks under the name ‘Debt Held by the Public’. Which to my mind is too close to ‘Public Debt’. But it is what it is.

And what is it? $12,478,871,032,564.07. But even this doesn’t get us to a good measure of ‘Real Debt’ because it doesn’t address the issues of Rate, Maturity and Term. To take this to an imaginary limit, what if every penny of that $12.5 trillion was in 30 year Bonds bearing a Maturity ranging from 2039 to 2044 and carrying a Rate of 0.025%. Well whip out your Financial Caclulator and do some PV calcs and that wouldn’t really be ‘Debt’ at all. And in particular it would incur no ‘Debt Service’ in the meantime. Now take it to a different imaginary limit. Assume every penny of that $12.5 trillion was held in 6 month or 1 year notes at a Rate of 6.0% but was all held by the Federal Reserve. Well that would imply a HUGE amount of Debt Service. On the other hand the Fed rebates all profits to Treasury so the net effect would be Treasury writing a ‘check’ to the Fed that would be ‘signed over’ back to Treasury. So that too means no ‘Real Debt Service’ in practical effect.

Now we don’t exist in a world near either of those limits. But that doesn’t mean we shouldn’t net out Public Debt that is held by either Government Trust Funds or the Fed from ‘Real Debt’. And equally important we need to understand what portion of total debt service is ‘Real Debt Service’ that is payable to non-Federal entities out of actual tax collections. That is we get exactly nowhere just stating: “The U.S. has $17 trillion in Public Debt”. Instead we need to have a grasp on the questions of “Debt to Whom? At what Maturity? At what Rate?” before we can put our fingers on ‘Real Debt’ and ‘Real Debt Service’. Answering those questions would require more than a single blog post and comment thread, but I suggest we start by looking at Fed Treasury Holdings under the fold.

The New York Fed maintains a website called System Open Market Account Holdings which gives totals and subtotals for that account’s holdings of Federal instruments including Bonds, Notes, Bills, TIPs, Agency Securities, and Federal MBS. And we can see that after several rounds of Quantitative Easing (QE) this adds up to a fairly staggering portfolio of $4,017,101,832,900. Not all of this is counted as Public Debt, that I think would be confined to the totals of Bills, Notes, Bonds and perhaps TIPs which together give us sub-totals of $0, $2,244,872,848,800 , and $95,389,037,400 for a approx total of $2.3 trillion. Which if subtracted from the $12.4 trillion of ‘Debt Held by the Public’ gives us a first cut at ‘Real Debt’ of $10.1 trillion.

Now that is a useful number to have. In particular it gives us something to directly compare to the numbers in the following table maintained by Treasury: MAJOR FOREIGN HOLDERS OF TREASURY SECURITIES which in turn tells us that the total so held is $5.9 trillion. Which in turn could spark some useful discussions. But I want to start a hare in a different direction and start looking at ‘Real Debt Service’ which requires asking “Who holds what percentages of Treasuries along the current Yield Curve?” And more precisely “What was/is the effect of Fed QE on ‘Real Debt Service’?”

Hopefully the experts will fill the following in but on my reading the short version of QE was “Buy long to drive down short”. But this also had the effect of “Buying old to drive down new”. And since the combination of “old” and “long” adds us to “high coupon rate” (because long bonds issued a decade ago have much higher coupon rates than newer ones) this suggested to me that QE had the inevitable effect of driving down what I am calling ‘Real Debt Service’ by soaking up the pool of higher yielding long bonds. But how strong is this effect? What proportion of the Budget line item(s) devoted to Debt Service are actually flowing to the Chinese and to overseas Oil Producers? Well we can get a start by examining the sub-Tabs on that NY Fed SOMA Holdings site.

If you click on the Tab called T-Notes and T-Bonds you will see along list of holdings sorted by Maturity Date with the third through fifty columns giving Coupon Rate, Par Value, and % of Outstanding. And it all makes for fascinating reading, at least if yu are into this sort of thing. Now as you cast your eye across the columns it is pretty easy to distinguish Long (<10 years) from Medium and Short, the former having Coupon Rates above 6%, indicating Long Bonds issued before the crash. And if we cross compare the issues with 6%+ Coupon Rates with the column giving % of Total Outstanding some eye-opening results appear. For example if we take the issue maturing on 8/15/2017 with a coupon rate of 8.875% we see that fully 56% of all such outstanding are held by Fed SOMA. Or take the 8/15/2020 with coupon rate 8.750% where the % of outstanding held by the Fed hits 70%, which on inspection seems to be the Feds self-imposed limit. And which is pretty consistent starting with the 2019s. Anyway I'll turn this over to you all. But to summarize my preliminary conclusions: One - the best measure of 'Real Debt' is as a first cut 'Debt Held by the Public' minus 'Federal SOMA Holdings' or $12.4 trillion minus $2.3 trillion. Two - any estimate of 'Real Debt Service' has to take into account that the roughly 15% of 'Debt Held by the Public' held by Fed SOMA includes up to 70% of the highest Coupon Rate instruments out there. Three - it just won't do to take 'Total Public Debt' and then multiply that by average Coupon Rate to get some idea of what it takes to carry U.S. Debt. That number is basically meaningless. As to a lesser degree is the metric 'Total Public Debt' itself. If we really want to get to a place where we can measure sustainability of national debt we just can't start from R-R type calculations of Debt as % of GDP. Instead we need to do some work examining that debt through the lenses of Rate, Term, Maturity and perhaps most importantly 'To Whom?'