Federal Reserve SOMA Holdings of the Long Bond
It is well known that the Federal Reserve in its three rounds of QE (and especially the first two) aggressively bought Treasury Notes and Bonds or collectively the ‘Long Bond’ of 10 Years and longer. But there has been relatively little discussion of what that means for such things as Federal Government Net Interest etc. So rather than advance my thoughts I figured we could start with the numbers in the following web tool:
System Open Market Accounts
The link should take you to the Summary Tab which shows that total holdings in these accounts are $4.221 trillion comprised of $1.738 trillion of MBS’s (Federal Mortgage Backed Securities) and $2.347 trillion in Treasury Notes and Bonds plus some smaller amounts in TIPS and Agency Securities that are more a rounding error in context. Of interest is that the Fed’s SOMA holds no T-Bills at all. That is the holdings are all Medium to Long (2 years and longer) and not Short (1 year or less).
If you click on the Notes and Bonds tab you will see a breakdown of Fed SOMA holdings of each issue. Now while Notes and Bonds are not individually labelled it is easy to distinguish 10 Year and shorter Notes and 30 Year Bonds simply by inspection of the relative coupon rates of different Treasuries maturing on the same date: rates above 6% being Bonds issued 15, 25, and 30 years ago and rates in the 1-3.5% range being mostly 10 Year Notes. If we then inspect Fed holdings of those higher yield Bonds we see that they comprise up to but never exceeding 70% of the total issue.
I’ll let you all inspect the Tables as you will but would like to ask a question. The numbers show that the Fed’s SOMA holds something like 2/3rds of ALL existing high yield Treasury Bonds. Which means that these accounts are also collecting that percentage of projected Net Interest going forward for that part of Federal Debt actually held in Bonds. Since the Fed has no need to ever sell those holdings against its will and can simply hold them to maturity and since it rebates ‘profits’ to Treasury this would seem to mean that simply examining Net Interest in either nominal terms or as a percentage of GDP will lead one to erroneous conclusions about the ability to the Government to service its debt. That is just examining the total for Public Debt and taking its average coupon doesn’t do much until or unless you segregate out both Intragovernmental Holdings AND the Fed’s SOMA holdings and THEN adjust for the fact that these holdings include up to 70% of the highest yielding instruments.
I have put up these numbers before but as usual had the comment threads hijacked into other channels (cough, cough **Social Security** cough). But I really really hope SOMEBODY will address these issues more directly in terms of macro finance.
I can make one basic comment from mathematics which everyone can completely ignore. But it is fundamental and part of aggregate statistics. Money appears in the numerator of just about all economic modelling. Money thus has the fundamental multiplier quality. Hence, the model can be constructed in one of two ways: 1) The model is only valid in the limit as the economy becomes infinitely divisible. or 2) Money is ultimately generated by something called a Weiner process which is matched to the model.
So, there is always assumed to be a currency function that has the following properties: It is maximally divergent in the model, and optimally dense, and part of a connected log additive network. Otherwise money obeys the infinitely precise Euler exponentiation, and all process are infinitely dense.
Currency is the distribution of numbers available to the multiplier function; mathematically, it is a self referencing number line. You are stuck.
I bring this up because the ‘Will’ of the central banker is that Weiner process. Until we know the match between that ‘Will’ and some good definition of all the players, then it is mathematically impossible to say anything, really, at all. Are the remits to Treasury a tax? Is Treasury a unique member bank? What thing modifies a nominal zero mean Weiner process that generates inflation. What other things in the model use ‘time’? How does the model conserve connectivity? In other words, it you are not infinitely going to zero then you have to have all the components of something called a ‘ring’, a specific type of group that can divide and multiply angles around a circle.
This guy nails it exactly. A friggen hero, his post appeared just after this one. He uses the math correctly, and discovered the ECB was not a Weiner process, and so really screwed everything up.
Greece has a problem with debt that must be addressed on way or the other. This column proposes a way to estimate a ‘fair’ level of fiscal consolidation in Greece. The author’s central argument is that contagion risk made the Greek crisis worse by preventing early debt restructuring. If restructuring took place in 2010 instead of 2012, Greece’s debt to GDP ratio would have been 30 percentage points lower today. To bring Greece’s debt under 120% of GDP, it would be fair for Greece to run a 3% primary surplus over the next decade or two. This is less than the current target of 4.5% but still requires a significant effort.
Matt I was asking for “macro” and not “meta”.
So maybe a little more Krugman vs Cochrane than Euler cum Nietzsche.
But the more the merrier I guess. Maybe PGL will follow you over here from E.V. and so revisit his old stomping grounds. Because I know he is a huge fan of yours. And maybe can make sense of the Weiner process as it relates to the Will of the Central Banker. Because FSM knows I am baffled.
My response was to your first comment but the second leaves me at an equal loss. What exactly is the relation between the situation in Greece and the composition of the Fed’s SOMA Account holdings?
The thing we need to know above all else is whether that $4.2 trillion of paper will ever be sold back , restoring the Fed’s balance sheet.
If the answer is yes , then it seems to me that it’s all a wash , or at least close enough for gov’t work.
If the answer is no , then the accounting of interest , maturities , and such is a sideshow , since we’ve now experienced the sweet taste of the forbidden fruit – debt monetization – and the more compelling question will be how many more bites , and how big , we’ll take going forward.
fwiw, yesterday zero hedge had a chart from morgan stanley showing the net issuance of government debt in 2015 will be negative; central banks will buy it all…
as to whether that $4.2 trillion of US paper the Fed holds will ever be sold back, if one believes that QE was stimulative, then unwinding QE would be the opposite, and it’s hard to imagine a scenario wherein they’d believe the US economy would need to be cooled off to that degree..
You’re only baffled?
You are one of the lucky ones.
Rjs an important point and feeding into a question I have been asking for a long time.
What part of U.S. Federal Debt and specifically Bonds and Notes are being held by public and quasi public entities (Central Banks, Sovereign Wealth Funds) not as investments to be bought or sold but simply as reserves? And as such are really not “on the market”?
Take Marko’s question. Is there really any reason why the Fed would or should sell of its $4.2 trillion in government paper as opposed to just letting it mature off that balance sheet? And specifically as regards its holdings of Treasury Bonds and Notes? And if not should U.S. taxpayers really look at that as debt exposure?
I ask this because one of the narratives being sold by Austerians to the American people is that we one) have $18 trillion in debt hanging over our heads and that two) huge amounts of that are held by overseas investors like the Chinese who could alternately A) sell of their holdings and so drive up rates on competing new auctions by Treasury or B) simply refuse to continue buying those new issues and so in effect doing the same thing. Meaning that is is absolutely imperative to keep the Confidence Fairies satisfied lest we have some sort of massive dump of our debt instruments and simply unable to issue new.
But this narrative breaks down if it turns out that say the Chinese hold around a sixth of U.S. debt and for world trade and currency considerations have to hold onto a large portion of that in the form of reserves and that furthermore the numbers show that the Chinese are not actually buying Treasuries on net and look likely to keep their $2 trillion holding of Treasuries steady. Like the Fed. And since between the Chinese and the Fed we are talking about something like a third of total U.S. Debt Held by the Public and a much higher percentage of that debt which was issued as Bonds and Notes (as opposed to short term Bills) how likely is is going to be that Treasury has some difficulty issuing new Bonds?
And given the evidence of that article from ZeroHedge the answer seems to be “not too likely”. Meaning you don’t have to go all MMT to understand that the U.S. is not Greece or still less on the verge of being Weimar 2.0.
Or to ask the question in Marko’s terms:
Why NOT eat the Forbidden Fruit? If it doesn’t give us indigestion OR expulsion from Eden (by the Gnomes of Zurich/Davos) why shouldn’t we?
i havent seen details on what part of our notes and bonds are being held by other central banks, Bruce…but i imagine each one of them publishes something akin to the Fed’s H 4.1 release, so that info should be out there someplace..
the problem with all discussions on US or any sovereign debt is the use of that word debt, which carries the confusing connotation that it’s something that has to be paid back…that leads some to compute how much each taxpayer would owe, something i even did myself when Reagan was president and probably even did as recently as 2008…it’s a nonsense exercise, because it isnt something we owe; the way that notes and bonds function in the economy is more akin to the way a ten dollar bill in my wallet functions as money for my personal use…
so i dont see us as being “exposed” to our notes, bonds, and bills….not even to the interest on them…the money the government pays out in interest does not go down a black hole…it enriches the bondholders, who are just as likely to recycle it as any other money that’s circulating in the economy….all money is loaned into existence, so expanding debt by either the government or the private sector is absolutely necessary for economic growth…any attempt by the government to pay down the debt will result in a depression, just as it did the previous six times in our history that the government ran large surpluses (steve quoted randy wray on that)…
the idea that we are borrowing from China also has its origin in the thinking that our script is somehow debt, leaving us with this picture of Uncle Sam going hat in hand to the Commies to get the money to run the government…the trade deficit with China is the only reason that they end up holding a large portion of our script…every time Walmart orders another boatload of trinkets from a Chinese supplier, they pay for it with a bond or note (or an electronic transfer thereof) …to get rid of our notes and bonds, the Chinese would have to stop trading with us..
the idea of the Chinese “selling their holdings” is equally ludicrous; what would they want in it’s place, a stack of greenbacks? or would they prefer Euros, or Swiss or German notes with negative interest rates? yen, when the “debt” to GDP ratio of Japan is 260%? and if they do buy yen bonds, then Japan buys Treasuries..
so i dont see our “debt” as a problem that has to be addressed; if anything we need to issue more of it, to allow the economy to reach its potential…when i see a post by the Austerians, i’ll take it on for what it says, but i wouldnt want to spend my time building a counterfactual to a nonsense premise…
Krugman has also been having an epiphany of sorts about sovereign debt…he’s had a least two posts and one column around the theme of “we owe it to ourselves”…he got started on that by quoting Antonio Fatas, who i usually dont read, but who really nailed it with this post:
Those mountains of debt (and assets)
(edited to correct arithmetic – and no regular commenters don’t have the ability to edit, WordPress giveth and it keepeth away)
The numbers for foreign central bank holdings are certainly available at least to a first approximation, via this Treasury monthly publication
That is after a breakdown of the total foreign holdings of Treasuries by country (total: $6.112 trillion) Treasury breaks out ‘For Official’ at $4.135 of which $3.791 are ‘Bonds and Notes’.
So we have Fed SOMA at $2.347 plus ‘For Official at $3.791 for a total Bond and Notes holdings by (mostly) central banks of $6.1 tn.
This against total outstanding Bonds of $1.589 and Notes of $8.232
So $6.1 against $9.81 total Notes and Bonds plus evidence (from the Fed at least) of a bias towards Bonds and you start thinking that our total exposure to high yield legacy debt from 10 and 20 years ago is somewhat overrated. Seems like most of the debt rolled up via the Great Society and the Cold War is either being held by state actors if high yielding or has been rolled over to Notes and Bills carrying coupons averaging less than 2%.
Kind of a blizzard of numbers but certainly telling a more complicated story than “Man we are handing $18 trillion of bankrupting debt down to our children”
A tiny bit OT (but hey it is my post) are some numbers from the Major Foreign Holders of Debt http://www.treasury.gov/ticdata/Publish/mfh.txt
As noted we have $6.112 total Foreign Holdings and $4.125 of that ‘For Official’. Which leaves a shade under $2 trillion in debt in ‘foreign’ hands. But going down the list of countries we have the following in place order:
#3 Belgium $331 bn
#4 Carib Banking Ctrs $331 bn
#7 Switzerland $184 bn
#8 Britain (which includes off shore Crown Dependencies like the Isle of Man and the Channel Islands) $175 bn
#10 Luxembourg $167 bn
#12 Ireland $127 bn
Now even granting that the Bank of England and RBS might hold big chunks of Britain’s $175 bn we are looking at a lot of money that is not included under actual ‘For Official’ here. But instead is just money in off shore banks. And not just for Third World Kleptocrats, you have to believe a lot of those Caribbean and Bene-Lux and Swiss holdings are held by American billionaires. (Which is to say First World Kleptocrats.)
I’m all for taking as many bites of the apple as necessary to do the job , what I object to is our collective unwillingness to discuss , openly and honestly , what we’re really aiming for. What is “the job” ?
When Krugman et al say ” we owe it to ourselves” , they’re not talking about mortgage debt , or student debt , or credit card debt. That sort of debt is , and will remain forever , debt that is owed by the 99% to the 1% ( pick your own numbers – the point being it’s debt owed by the masses to the few ).
The only limit to monetizing debt is inflation , and if you don’t monetize the debt of the masses it’s pretty easy to achieve that happy low-inflation condition. The problem is that you get a low-growth economy , where the net effects of monetization show up as higher asset values rather than gdp , and , again , it’s the asset values of the few that get the boost. It can’t be any other way , because if you raise the wealth of the masses they’ll “monetize” that wealth , spend it , and inflation will result.
So , the endgame of the neoliberal low-wage , debt-led economy is a low-growth , low-inflation economy that serves only the interests of the ever-more-wealthy few. QE-enabled monetization , as practiced today ( in Japan , if not here just yet ) , is simply proof that we’ve reached that endgame and see it as a desirable status quo , to be maintained indefinitely.
on that “handing bankrupting debt down to our children”, should we damn our parents for the government debt they passed on to us? i was born into a country with a debt to GDP ratio of 120% coming out of world war two…should we argue that our parent’s generation should have fought the war on a pay as you go basis? and during the 50s, the government ran deficits to educate the returning soldiers and to build the interstate highway system…would the country have been better off using the roads we had at the time, and without the GI Bill?
I don’t think Fed holding of the long bond have to be all that critical. The Treasury can issue very low amounts of bonds and notes. I don’t see any reason why the Treasury should pay significant amounts of the higher interest rates on the long bond.
In particular, I am fairly sure that the yield curve slopes up because private agents (especially those managing other people’s money) have short trading horizons. The long bond is risky if one has to mark to market at the end of the year and one risks being fired (or for an institution being considered insolvent). The US Treasury does not need to worry about this at all. For one thing, it marks debt to face value. So why pay people to bear a risk which is no problem for the US Treasury ?
Of course there is a hypothetical advantaget to the Federal Government from issuing long term nominal debt instruments. It can inflate that debt away. However, this is universally considered a bug not a teature with the need for precommitment to take away the temptation. If almost all US public debt had maturity of three months or less, there would be much less risk of fiscal dominance and (even) less reason to appoint inflation hawks to the FOMC.
On the other hand, hold to maturity is not a neutral monetary policy. We will not always be in the liquidity trap. To keep demand for Fed liabilities as huge as it currently is, one of three things is needed ( two are really the same). The price level could increase so that real Fed liabilities are normal — that is a period of high inflation followed by price stability would work. This is politically absolutely unacceptable. Alternatively (but not really alternatively) the Federal Funds rate could be kept near zero for decades. This would cause high inflation which is politically unacceptable. Or The Fed could increase reserve requirements. I think it would need a required reserve ratio greater than one (M1 is currently less than Fed liabilities). This is an extreme policy. It is allowed by the Federal Reserve act, but it is basically a tax which can be changed without Congress. I’m pretty sure that if the FED imposed a reserve ratio greater than one, some banker would sue saying that the Federal Reserve Act (as written and interpreted so that is allowed) is unconstitutional, and I’m pretty sure that 5 Supreme Court justices would agree.
Stressing my ignorance (especially of law) I don’t think just holding to maturity is a feasible policy.
Robert let me stress my ignorance back at you. Because I just don’t follow the logic. Which in all likelihood is one me and not you.
Seems to me that we have two actors here: the Fed and the Treasury. The Fed decided to inject liquidity into the system by buying long and selling short and in the process largely locked down most of the long bond. Now from one perspective this act is neutral as regards Treasury, Net Interest is based on coupon and not price and gets scored as a Budget outlay. On the other hand the Fed rebates its profits to the Treasury which gets booked as a receipt. So it seems to this naif that from a macro perspective this all offsets, that portion of Net Interest which ends up being rebated to Treasury not actually needing to be financed through the debt markets at all. Not on net. Instead the ‘Real Cost’ of financing Public Debt is reduced. So all to the good.
Now there seems to be some assumption that the Fed at some point has to unwind, if only to suck out the excess liquidity in the economy. But equally this liquidity could be sucked out by Treasury by issuance of new long bonds. Which if all actors understood that the Fed was in the position of buy and hold should not have their price/yield determined by the total existing supply. That is for all practical purposes it would be as if the Fed bought and retired those Bonds. As far as the secondary market would perceive it.
So maybe I am just not understanding why QE has created “huge” Fed liabililities at least as it relates to their purchases of the Long Bond. Because like Treasury they seem to be able to just mark it to face value and continue collecting the relatively high coupon yields.
In normal times the tendency is for people to see the Fed as the responsible party for taking away the punch bowl before the party gets too lively and the Treasury as the irresponsible host that wants everybody to drink up and have a good time. But right now the Fed is in the position of seeing all the guests refusing to step up to the plate/punch bowl to start with. And rather than trying to take the punch bowl away are trying to think of ways to spike it. Even though they are sitting in a liquidity trap that has choked off their supply of needed hooch.
So why not allow Treasury to sell long bonds into a market that is only demanding short term rates on those longer term instruments? And use the proceeds to directly invest in infrastructure that would drive wages and so inflationary pressures up to their target? While potentially having all kinds of productivity multipliers along the way? (Hoover Dam 2.0 anyone?)
I am just not seeing why this precludes hold to maturity. Sure it is not neutral. But that doesn’t mean that unwinding the portfolio is the only possible monetary adjustment.
Then again I was a Medieval Celtic scholar back in the day and willing to take instruction on these number thingies from anyone who can make sense of them to me.
Webb – the Fed chose high coupon/long maturities for a reason. By doing so they significantly added to the monetary juice of QE.
The Fed paid a premium for the bonds. That premium is (more or less) the NPV of the coupons discounted at the then prevailing market rate for the same maturity.
The extreme example is that the Fed bought a 6% 30 year bond when the market yield for 30 year paper was 3%. The “price” of that bond would be 159 (59% premium to par)
The Fed buys $10m of this bond. $15.9m is released to the money supply. The Fed’s balance sheet goes up by only $10m. (The Fed’s balance sheet is a measure of the par value of bonds owned.)
The Fed’s income statement for this would be interest income (6% on $10m) minus amortized premium (59/30 years = 2%), minus the cost of funding (IOER and printed money – the funding requirement for this is $15.9m).
When does the Fed incur a loss? When % expense on 15.9m exceeds 4% * 10m. When IOER reaches 2.5% the Fed is at breakeven.
Basically, the Fed will lose money on its holdings if the CBO/SSA forecasts on short-term interest rates are realized.
The Fed does publish the par value of bonds in portfolio. It does not publish the blended cost it paid for all those bonds (premium). Thus you can’t look at the bonds and the coupons in the Fed report and draw any conclusions about what the Fed’s future income will be.
I see no reason that would compel the Fed to sell any bonds in its portfolio. What’s on the books today will be there for a very long time.