Brad Setser

Follow the Money

Cross border flows, with a bit of macroeconomics

Print Print Cite Cite
Style: MLA APA Chicago Close


Doesn’t a smaller (external) deficit mean less dependence on (external) creditors, including China?

by Brad Setser
July 28, 2009

There is a common argument that the US depends more on China now than before because the US needs to issue so many Treasury bonds to finance its fiscal deficit.

I disagree, for two reasons:

First, the trade deficit is down significantly, so the amount that the US needs to borrow from the rest of the world has fallen. That means less dependence on external creditors. The fiscal deficit — obviously — is much bigger now than it was a year ago. But inflows from the rest of the world can finance a private sector deficit as well as a public sector deficit. Private borrowing in the US is way down – and that has pulled total US borrowing from the rest of the world down even as the fiscal deficit rose. After crises in Asia in the 1990s and in Eastern Europe and the US in this decade, there should be little doubt that external deficits that have their roots in excessive private borrowing are also risky.

In my view, the US was more dependent on central banks in general and China in particular for financing back in 2006, 2007 and the first part of 2008 — when the US trade deficit was larger than it is now and emerging market reserve growth was higher than it is now.


Second, the majority of the fiscal deficit isn’t being financed by foreign central banks. That’s a key change. Indeed, the rise in the Treasury’s issuance of long-term debt has come even as central bank demand for long-term debt has fallen. That key fact gets lost amid the general sense that the US must be relying more on China now than in the past because the US government is borrowing more.

The evidence? Three additional charts.

The US government clearly is borrowing a lot more — as we all know. The stock of outstanding marketable Treasuries increased by close to $2 trillion over the last 12 months of data (June 2008 to June 2009). Some of those Treasuries were issued to finance the Fed’s actions as a lender of a last resort — and some were used to finance the TARP. But more and more are being issued to finance the fiscal deficit.


Give US debt managers a bit of credit. Over the last several months, they have sold more notes and fewer bills, financing more of the deficit with longer-term borrowing. That is good debt management. In the month of June, they sold close to $220 billion of notes while reducing the stock of outstanding bills by $60 billion. And they did so in a market where central banks remain reluctant to buy long-term US Treasury bonds, so most issuance is being bought by private investors.


When I look at a chart showing note issuance v estimated official demand, I don’t see any evidence that US dependence on foreign central banks — or China — or financing is rising. Rather I see evidence that the US is relying more on private demand for long-term Treasuries to finance its deficit.

To be sure, it is important to look at the market for short-term Treasury bills as well — and there is now doubt that central banks are buying more bills. China alone added about $200 billion to its stock of bills over the last 12 months of data (May 2008 to May 2009 in this case), taking up about 1/5 of the increase in the outstanding stock of bills over this time.


But the price of bills is largely set by US short-term policy rates; central bank demand there has less impact on key market rates than central bank demand for longer-dated Treasuries.

A few caveats.

Global reserve growth has fallen by more than Chinese reserve growth; China clearly still matters for the financing the US external and fiscal deficit — though it is not currently playing the same outsized role it played back in late 2007 and early 2008.

The data over the last 3 months isn’t quite as clear as the data over the last 12 months as reserve growth has resumed, and there are signs that central banks are starting to buy Treasury notes again. Though they still need to prefer the shorter-dated notes, i.e. two and three year notes.*

And clearly China holds enough Treasuries (and Agencies) that if it ever started to sell — really sell — it would move the market. Any single investor that controls a $1.5 trillion dollar portfolio — actually now more like a $1.6 trillion dollar portfolio — can move the market if it abruptly changes its behavior. When central banks stopped buying Agency bonds last fall it had an impact.

But we shouldn’t ignore the adjustment that has happened. A smaller trade deficit means that the US relies less than before on inflows from the rest of the world . That’s progress.

The goal should be to maintain that progress — and not go back to a world where the US was running a far larger external deficit than it is now. Those deficits came at a time when private markets wanted to finance the fast-growing emerging world, not the slow-growing US. The gap was filled by truly enormous inflows from the world’s reserve managers — and China — into US markets. Those inflows were far bigger than in the inflows we observe now. Just look at global reserve growth then. But those inflows were into a host of different markets, not just the Treasury market — so they weren’t as visible. But the total flows were huge, as they had to be to support a large US trade deficit.

That was a world where the US structurally relied on a small set of other governments for large amounts of ongoing financing. A smaller external deficit means a bit less reliance on external financing. And maintaining the improvement that was brought about by the crisis requires more “balanced and sustainable” growth going forward.

Note: all data on official and Chinese inflows comes from the TIC, but the monthly TIC data has been adjusted to reflect the survey revisions — i.e. flows have been reattributed from private investors in the UK and Hong Kong to China and official investors in proportion to the survey revisions. There is no survey data after June 2008, so I have estimated the magnitude of the likely revisions based on past patterns, and a bit of judgment. Given how much time I have spent looking at this data, I hope I get a bit of latitude here. If I used the 2008 revisions to estimate the 2009 revisions, China’s current holdings would be revised down slightly. I believe though that the flows that China accounts for a higher share of the UK flow now than in 2008 – and thus the 2006 and 2007 revisions provide a better guide to the likely 2009 revisions.

* The indirect bid at today’s two year auction was on the weak side, but there is a host of anecdotal evidence indicating ongoing central bank interest in short-dated notes as a higher-yielding alternative to bills. Central banks’ interest in the last two-year auction, for example, was especially strong.


  • Posted by Twofish

    Something to notice from your charts is that the fraction of external financing has decreased, but that’s because the total amount of debt issued has dramatically increased.

    The other thing to point out is that what happened in the last few months is a one time event rather than something that can be sustained for very long periods of time. The huge increase in treasury purchases was made possible only because everything else in the world was collapsing (i.e. US stocks dropped by close to 50%).

    Once the economy starts to improve, the demand for treasuries is going to start dropping as people move out of Treasuries into other things.

  • Posted by Michael

    First, the good news:
    “Give US debt managers a bit of credit. Over the last several months, they have sold more notes and fewer bills, financing more of the deficit with longer-term borrowing.”

    Now, the bad news:
    “When I look at a chart showing note issuance v estimated official demand, I don’t see any evidence that US dependence on foreign central banks — or China — or financing is rising. Rather I see evidence that the US is relying more on private demand for long-term Treasuries to finance its deficit.”

    Why is the shift from foreign central banks to the private sector bad news? Yes, it’s true that China or Japan (or even Russia, Norway, etc) could move the market quickly if they started to sell their Treasuries. But the single most stable fact throughout the pre-crisis, crisis, and current eras is that they don’t tend to sell their Treasuries. Private investors, on the other hand, have been exceedingly unstable in their willingness to hold onto any depreciating (or potentially depreciating) asset during every era. Combine that with Twofish’s observation that the ballooning growth in total U.S. gov’t debt is itself a destabilizing factor and it’s not so clear we’re in an improving situation.

  • Posted by WStroupe

    Yup! ‘To have and to hold’, or more accurately, ‘To KEEP BUYING and to hold’ is what the U.S. desperately needs, especially as it runs up huge budget deficits to deal with this crisis. All that piling into short-dated Treasuries is a big strain on the Treasury, ’cause it’s too refund/rollover intensive, and if for some reason investors decide in significant numbers NOT to roll over their short-dated Treasuries, but want refunds instead so they can put their money elsewhere, then the Treasury could potentially be flooded with refund demands it can’t meet. So getting investors long on Treasuries is absolutely necessary. How you gonna do dat if China balks at deepening its long position on Treasuries? It’ll move the market, allright. Move it to stay short, if at all. And China won’t have to sell off in order to see this happen.

    It’s all OK for now – the extremely precarious U.S. Treasury balancing act is working better than I thought it would. But only a proverbial sneeze by China or in the form of some unexpected phase of this crisis or by some Washington policy blunder (or more likely a combination of the above)could push things off the ropes to the hard ground below.

    All the while, as the posters above point out, U.S. financing is deteriorating as far as quality goes.

  • Posted by Twofish

    One thing that you can do if you have about eight major players is to get everyone on the phone in a crisis and agree to coordinate action. You can do this with central banks, and you can also do this with major US banks. If you have the owners of Treasuries distributed among thousands of people, then you can’t do that.

    There is one area in which the system seems to have worked. If it had been any country other than the United States, then the events of the last few months would have resulted in capital flight and the total destruction of the economy.

    However, because Treasuries were the savings of last resort, when people started running, they ran right into the arms of Uncle Sam. This allowed the US to appreciate the currency and undertake a massive stimulus package and avoid what happens to most countries when they get into this situation. It’s also very different from what happened in 1973, when people ran out of Treasuries into gold.

    So I’m not so sure that the system is such a bad one for the US. The things that people were worried about in 2006 (massive trade deficits and loss of jobs) are trivial and insignificant compared to what could have happened if there were *any* credible alternative to US Treasuries. If there were any asset that was considered safer than Treasuries, then people would have massively dumped US Treasuries, and the US economy would have become a smoking wreck (see Argentina).

    Also trust and guarantees are a funny thing. You can swear up and down that you offer no guarantees on something, but people won’t believe you.

  • Posted by WStroupe

    Twofish said, “If there were any asset that was considered safer than Treasuries, then people would have massively dumped US Treasuries, and the US economy would have become a smoking wreck (see Argentina).”

    I mostly agree with you, except for this very important caveat:

    It isn’t totally the actual SAFETY of Treasuries as an asset. It’s a COMBINATION of their inherent safety as an asset along with the fact that the market for Treasuries is still very deep and very liquid.

    Put another way, if I buy Treasuries I can sell the damn things in a second if I have to – either to go into something like hard assets, which I might consider safer than any financial asset, or simply to go somewhere for a higher return. So the idea of safety of Treasuries is more and more a factor of the depth and liquidity of the market for Treasuries, and less and less a factor of their inherent safety as an asset. Witness China’s growing nerves over how safe they are past the short term.

    So I think it’s fair to say that global investor psychology as respects Treasuries themselves (past the short-dated bills) is significantly deteriorating, but the full extent and significance of that deterioration are being masked by the fact that the MARKET for Treasuries is still very attractive, for its depth and liquidity.

    But at some point, if you extend the trend line, we’re gonna see fears over the inherent safety of these U.S. financial creatures begin to negatively impact the market itself for these creatures. You mentioned alternatives to Treasuries – none yet. But if something breaks apart in the Treasuries market, perhaps due to continued extremes on the short end, then alternatives rooted in the emerging markets could rise pretty fast.

  • Posted by Cedric Regula

    Gee Brad, I dunno.

    I guess maybe everyone did always make too big a deal about “our dependence on FOREIGN financing”(translation foreign=mostly China), when we should be concerned about our dependence on foreign AND domestic financing.

    Besides, CBs have been the pussycats of creditors anyway. Our other choice is Wall Street, and that is where American culture believes is the best place to keep the non-violent, criminally insane.

    Besides, CBs don’t have to pay taxes, but when Americans do it, first you pay your regular income tax rate on the yield, then subtract a jiggered up CPI inflation rate, and more often than not you end up with a negative real return. Maybe that’s what gets us mad.

    The the fact that someone is going farther out on the yield curve is because 6 months ago everyone was piling into bills at zero percent, and now they are being enticed by that juicy 2.5%-3% on a 5 or 7 year note. (note that yoy CPI did go up 1.5%, so we only had a deflation scare..didn’t really happen.)

    But looking at the auctions lately they only sell a total of $22B in 10 and 30 yearlings combined, at a rate of once a month. That has alternated with $100B of note auctions in the in between 2 week auction.

    Until this week when the 7 year or less week ballooned to $235B, partly due to $68B in bills added to the lot.

    So the ratio between 7 and less and 10 or more is running somewhere from 5:1 to 10:1, depending on which auction you think is the anomaly.

    Then we can look at national debt vs CA deficit. The CA deficit may average 500B/year over the past few years if I remember the charts correctly.

    Pre-crisis public debt of marketable treasuries was around $6T and we added around another $2T. That’s $8T and its average duration is now 4 years. So on average we need to roll over about $2T a year. Plus add the new coming fiscal deficits.

    That’s a lot more financing than the trade deficit, and we know it WON’T be China doing it with or without a big surplus to re-cycle.

    So I don’t see what comfort some may take in being less dependant on China.

  • Posted by Cedric Regula

    Oh yes, one more point. The Fed has issued its not so cryptic anymore message to US carry traders…”fed funds will stay near zero for a considerable period of time.”

    So they will be selling some 30 yearlings to institutional traders who have access to zero interest savings accounts, money markets that allow them, and even the Fed discount window, if they don’t feel too embarrassed about going there anymore.

    With enough free funds, lots of money can be made, even on long term treasuries.

    Then of course the Fed signals the end of the party and they all sell.

    So there is no one holding longer term treasuries with the intent of holding to maturity. It’s all traders, probably all wearing an eyepatch and a parrot on the shoulder too.

  • Posted by bsetser

    Cedric — your numbers are a bit off. the precrisis stock of marketable treasuries (with june 08 = pre-crisis) was around $4.7 trillion, bv $6. 7 trillion today. and a a decent chunk — something like $700-800b in normal states of the world — of that stock would normally be held by the fed. the pre-crisis current account deficit was closer to $700b a year than $500 b a year, or the current pace of less than $400b a year in q1 and q2 08 numbers are annualized.

  • Posted by Cedric Regula

    I didn’t look anything up, so that’s very possible.

    But the current national debt was almost $11T and I thought $5T of that was n the SS trust fund(captive IOUs at 2%, so they don’t count from a near to intermediate term financing liability), which is how I guestimated $6T as the figure that needs constant refi. Plus this years new 1.8T.

    I do remember the CA running at $700B+ at least a couple years.

    As far as the Fed goes, they tell us that part of the exit strategy is not to sell $300B in QE treasuries, nor 1.5T in GSE stuff until maturity. That’s to reassure the market in those securities. So I guess if you believe that is not printing money because you will call the money back in 5 to 30 years depending on the mix of long term stuff, then I agree that lessens the immediate financing problem. But clearly we have achieved a new milestone in central banking technology.

    But the comparison still leaves a big gap even using accurate numbers!

  • Posted by Rien Huizer


    Nothing surprising here. Dpending on if you would like the US to have foreigners subsidize longer term interest rates (by adding to “natural”, domestic/private sector demand for longer term paper) there is good or bad news here. The good news is that treasury paper placement becomes more of a domestic phenomenon, with everything that goes with it (crowding out, possibly more market-conforming term risk premia etc). As more of the stock moves to private sector investors, the market may become more volatile, etc. But the depressing effect on longer term interest rates from foreign central banks was only there because they must have had portfolio preferences favoring longer term paper. Why is difficult to say, and unless we know the answer, we will not be able to predict when, if at all, the considerable reservoir of foreign USD investment now hed in US Tbills and short term notes, will adopt a less risk averse stance. Pretty soon the Chinese e a will be rolling over more than half a trillion per annum.. Although a billion is not what it used to be, rolling over that kinds of amounts by a very small group of investors is going to have pretty funny effects on the auction and distribution system. Tails and dogs…

  • Posted by anon

    There’s only one reason.

    The US current account deficit is down, so the external financing requirement and dependence is down. That connection is true by definition. That forces the increased fiscal deficit financing to be absorbed domestically and then some. That’s happening with an increase in the private sector saving rate.

    Everything else is just observed data on the source and mix of financing. None of it is required to understand what’s happening and why it’s happening.

  • Posted by a

    Dunno, but I always thought the pertinent ratios were

    1/ (amount of debt owed to China)/ USA GDP,
    2/ (amount of yearly interest owed to China) / USA GDP.

    For 1/ the numerator has gone up, while the denominator has gone down – so not good. For 2/ (although I don’t have the figures) I presume the ratio has gone down because the interest rate has descended so much that it more than compensates for the decrease in GDP. The problem for the US is, should the interest rate spike, it is going to have to sell a lot of goods to China to pay the interest it owes and so just to tread water in ratio 1/, and interest rates probably will go up should GDP start going up. In brief it looks like the US may be in a situation where 1/ is intractable. And that’s not a good situation to be in.

  • Posted by bsetser


    I would add 3/ the increase in the amount of debt owed to China/ US GDP. The flow as well as the stock.

    anon. basically agree. I probably should have put my last chart first. I actually may do that — better visuals = more attention. the complication is that if lots of capital (private capital) is flowing out of the US, the US financing need may exceed the US current account deficit and inflows from central banks may top the deficit.

  • Posted by bsetser

    Rien — good points; my operating assumption is that most of those bills will be rolled into longer dated notes and agencies (eventually). if that isn’t the case, there is great potential for instability. but there is also graet potential for instability if American investors don’t show up at a Treasury auction, as you note.

  • Posted by Pax Americana

    Ok so the deficit is down for the external deficit, considerbly so will it stay at about $350b in 09, moving out to 2010/11 is the American Consumption Model, as driver of world growth dead? and also the impact of a reducing effect on the economy of less consumer spending across the board.

  • Posted by Mark G.

    How much debt must the US sell next year, the year after and the year after that? How and when will the Fed unwind? Will inflation, if we see any, effect demand for US debts? More questions than answers going forward.

  • Posted by Rien Huizer

    Mark G.

    That depends to what extent Americans do their patriotic duty and buy bonds, at least for a while.

  • Posted by David Pearson


    The wonder is why you expect China’s trade surplus to persist.

    The country is one of the few (if not only one) in the world which actually controls, by fiat, the velocity of money. Its monetary stimulus, if it keeps up, will certainly generate inflation and faster real growth than the rest of the world. This is particularly true if you compare the growth rate of Chinese consumption versus U.S. consumption. Further, the terms of trade would likely continue to turn against China in the next few months (again assuming continued stimulus).

    In fixed exchange rate regimes (see Brazil), the presence of large monetary stimulus combined with control over velocity typically generates large trade deficits. How is China different? You could argue they are not stimulating consumption, and I would respond that if they are to continue stimulating, they will eventually have to.

    I think you overestimate the system’s stability. Its quite likely that six months from now China will be selling Treasuries as its trade surplus reverses.

  • Posted by David Pearson

    Oh and btw, if they don’t continue to stimulate, and assuming China is currently propping up global growth, then deflation is a likely outcome.

    So the choice for China is a trade deficit or deflation. Either has out-sized impacts on the U.S. and global economies. How is this a stable equilibrium?

  • Posted by Cedric Regula


    I looked the current numbers for the national debt.


    $7.26T held by the public (foreign plus domestic)

    $4.34T intragov (mostly SS and fed pensions)

    $11.6 total.

    They do have a report there also that breaks down the debt structure by maturity. But I heard it’s 4 years duration quoted by Rick Santelli, the bond and currency commentator on CNBC, so I’m going on the assumption that he is correct.

    So the average rate that this needs to be rolled over is then 7.26/4=$1.8T per year.

    So that is very much more than even the peak CA of $700B that we got from foreign investment. Not that that comparison really means anything, but just gives a sense of scale, and the total financing need did jump by a huge amount lately and is on a steep upward slope indicating that we are in untested territory.

  • Posted by a

    a: “1/ (amount of debt owed to China)/ USA GDP,
    2/ (amount of yearly interest owed to China) / USA GDP.”

    Actually, on mature reflection, probably a better proxy for the denominator is “total US exports” – because, in order to pay down the debt, we will need to be exporting goods and services. Now *that* denominator has really gone down, so the ratio 1/ has really zoomed up.

    brad: “I would add 3/ the increase in the amount of debt owed to China/ US GDP. The flow as well as the stock.”

    Now consider (increase in the amount of debt owed to China/total US exports).

    That ratio might still have gone down, but nowhere near as much as the one with US GDP as denominator.

  • Posted by mdow

    Brad – Impossible for me to agree with you more! And, just getting back to NYC today from our nation’s capital, I can tell you that you can expect more of the same good debt management from Treasury. I can also tell you that even though the Chinese demand is less important, it is still very robust–despite public comments from them that might suggest the contrary.

    One more point on Treasuries: the fresh supply of AAA non-government paper is virtualy non-existant, as the economy continues to deleverage. Investors who are force by mandate to hold a certain share in AAA names have little alternative these days. This is a considerable help to the Treasury as it tries to place all that paper.

  • Posted by Cedric Regula

    CNBC just reported on how today’s 5 year auction went.

    In technical terms, they describe it as a pooch with a long tail.

    What that means in laymen’s terms is the bid to cover was less than two, the yield went up from pre auction levels, and large institutional buying(CBs) percentage was in the teens.

    Paradoxically, we have been getting a sizable rally in 10 and 30 yearlings the past couple days.

    7 year auction tomorrow.

  • Posted by Cedric Regula

    Scratch that rally in the 10s and 30s. They just reversed 7 basis points the last few minutes.

  • Posted by WStroupe

    mdow said, “even though the Chinese demand is less important, it is still very robust–despite public comments from them that might suggest the contrary.”

    Very robust at what maturities? Can you give us a quick breakdown? I’ll bet they’re still excessively heavy at the short-dated side.

  • Posted by HZ

    I think the question is preference and financing of US fiscal deficit. US domestic investors have a lot more choices — you can get FDIC insured bank deposits at 2.5% interest rate so what is the incentive (at least for small savers) to buy T bills? Chinese central bank has little choice but to buy Treasuries. So it is the financing of the public debt not just the net domestic debt that is in question. However I agree with you the dependence is less not more when the C/A deficit is down.

  • Posted by bsetser

    Mark — thanks for the comments. Glad we agree. i’ll try to get in touch with you tomorrow.

  • Posted by WStroupe

    Well, demand at the last two auctions isn’t blowing the Treasury’s skirt up, if you know what I mean:

    Before we conclude we’re really getting “good debt management from the Treasury”, and we’ve got the cat in the bag, wouldn’t it be better to keep on the watch to see if that’s really the way it turns out?