Brad Setser

Follow the Money

Cross border flows, with a bit of macroeconomics

Print Print Cite Cite
Style: MLA APA Chicago Close


Central banks aren’t always a stabilizing presence in the market

by Brad Setser
January 5, 2009

Over the last couple of years it was often asserted that sovereign investors — due to their long time horizons — tend to be a stabilizing presence in markets that they invest in. The argument was generally made about sovereign funds, but it presumably applied at least in part to central banks as well. They have a somewhat shorter time horizon than most sovereign funds, but they also presumably care a bit more about financial stability.

Alas, there is now one clear case where an abrupt shift in central bank purchases destabilized a market.

Central banks stopped buying Agency bonds in August and never resumed their purchases. The US Treasury now says that Agency bonds are “effectively” guaranteed — and they certainly have more Treasury backing than in the past. But that wasn’t enough to convince the world’s central banks. They now want nothing less than a full guarantee.

The latest (year-end) data from the New York Fed on foreign central banks’ custodial holdings shows that the magnitude of the shift out of Agencies and into Treasuries in the later part of the year. Central banks went from buying $250-300b of Agencies a year (judging from the growth of their FRBNY portfolio) to net sellers of Agencies in a rather short period of time.

The 3 month (proxied by the 13 week) change, in billions of dollars, in the Fed’s custodial holdings is even more dramatic in some ways. It leaves no doubt that central banks have been large sellers of Agencies. The Fed’s custodial holdings of Treasuries rose by $250 billion over the last three months of 2008 while the Fed’s custodial holdings of Agencies fell by $150 billion. Try annualizing these numbers. In the fourth quarter, central banks were buying Treasuries at a $1 trillion annual pace and selling Agencies at a $600 billion annual pace.

It seems hard to argue against the proposition that a sudden increase in central bank’s risk aversion has contributed to the distress in a key part of the US market.

Central bank reserve managers’ core concern, of course, isn’t stabilizing the US debt market. It is making sure that they have enough liquid assets to meet their own country’s liquidity needs. It used to also be to make a bit of profit on the country’s– before it swung to making sure that they didn’t take credit losses. The net result, though, was a lot of pressure on the Agency market in the fourth quarter — and a lot of central bank demand for Treasuries just when private demand for Treasuries also soared. Suddenly risk adverse reserve managers sold what was cheap and bought what was dear, magnifying rather than dampening market moves.

This experience should also to some way toward settling another debate: are central banks’ purchases and sales big enough to impact prices in large, liquid markets?

Many argued that the Treasury market was so deep and so liquid that it could absorb even large central bank sales without too much trouble. Central banks (net) purchases were large relative to the Treasury’s (net) sales, but they weren’t that large relative to total Treasury market turnover. That led many to argue that if central bank sales ever pushed a bond away from its fundamental value, private buyers would step in — preventing any large move in price.*

The Agency market isn’t a perfect analogue to the Treasury market. But it is quite large — the outstanding stock of Fannie and Freddie bonds (counting Fannie and Freddie guaranteed MBS) is over $5 trillion. Agency issues aren’t quite as homogenous as Treasury issues, but they don’t differ that much from each other either. Both Freddie and Fannie have lots of outstanding bonds in the market — and, at least until recently, the Agency market was also considered to be fairly liquid. It still doesn’t seem to have been able to absorb the big swing in central bank demand.

Agency spreads widened significantly when central banks pulled back (the expansion in the supply of debt with an implicit if not explicit guarantee also played a role …). They only came back in when the Fed indicated it would start buying Agencies …

And it sure seems like the enormous increase in central bank demand for Treasuries is one — though certainly not the only — reason why Treasury yields are so low right now. The obvious risk here is that the US government will infer too much from the fact that Treasury yields collapsed even as Treasury issuance soared. The current surge in central bank demand for Treasuries is unlikely to be sustained, if for no other reason than global reserve growth has slowed and central banks have a finite supply of Agencies to sell. I personally think this risk is manageable, as I expect a meaningful rise in US savings (an a fall in investment) will free up domestic funds for the Treasury (or start to flow into the banks, allowed the Fed to scale back its loans to the banks and scale up its Treasury holdings). But it is a risk.

In one key respect though, central banks have remained a stabilizing force: they abandoned the Agency market, not the dollar market. This underscores an important point, one that I wish I had recognized earlier. Central bank reserve managers can influence the US market in two very different ways:

a) By shifting out of one asset and into another asset. Selling Agencies and buying Treasuries for example. Or selling equities to buy Treasuries.

b) By shifting out of the dollar

A country like China that is effectively pegged to the dollar can shift the composition of its US portfolio around without putting much pressure on the dollar or its peg. Moving from Agencies to Treasuries is consequently a fairly low cost option for China. Sure, China gives up a bit of yield. But it doesn’t have to abandon its exchange rate regime. It is an option that China can exercise at a low cost to itself.

Moving away from the dollar, by contrast, would require adjustment in China – not just adjustment in the US … and that hasn’t been something that China has been willing to do.

NOTE: I replaced by initial charts — which used the monthly data — with charts that show the 52 week and 13 week change in the custodial holdings. Thanks to Paul Swartz for help with the more detailed charts.

* The argument could also be made in reverse. If central bank purchases drove a bond’s price up too much, private investors would sell — so the bond’s price (and yield) wouldn’t move away from its “fundamental” value. That led some to argue that central bank purchases couldn’t have much of an impact on say the Treasury market.


  • Posted by sackerson

    So, who’s been buying the Agencies? And have they got themselves a bargain?

  • Posted by interesting

    great graphs here. actually curious if we can get some of Central Bank net buying/selling of Gold the last 6 months?

    Graphs indcating treasuries (and increased demand of treasuries) look very similar to those of real estate prices (and not too long ago certain commodities)… I remember buying yahoo at it’s IPO selling at 180 and then watching it double from there “thinking i’m world’s biggest fool”…

    Like Mohamed El-Erian recently said “Get out of Treasuries, they are very very expensive”.
    May not be these year, but at some point fortunes will be made shorting the U.S. 20-30 year note.

    The graphs made by Brad clearly indicate we have a major bubble. Inflation or Deflation I’m doubtful the world will have the surplus nor demand to buy 30 years at 3%. What goes up fast, will likely come down hard.

  • Posted by Christian

    I’m not sure if it fits the data well when looked at in more detail, but I seem to remember Russia holding a significant part of its dollar reserves in agencies.
    Since they had to sell a good fraction of their reserves lately, might it be fair to assume that Russia accounts for a significant share of the actual selling of agencies?

    I think some of that also applies to South Korea?

    If a central bank actually expects that it might have to use up a large fraction of its reserves, doesn’t it make sense to sell “riskier” assets like agencies first, rather than “safe” treasuries?

    Some of the central banks may simply not be able to afford having a long term horizon (anymore).

  • Posted by Twofish

    As far as who has been buying agencies, until recently it was Treasury although now the Fed is getting involved.

    It’s all quite fascinating watching the financial system evolve. Basically what we now have is that the Federal Reserve Bank has gotten into the mortgage and insurance business with the People’s Bank of China as a business partner.

  • Posted by Mrjacket

    How significant to this matrix is the Fed paying interest on deposits?

  • Posted by Indian Investor

    It would be interesting for me to know the implications/inferences, in terms of suggested policy changes, from the analysis of PBoC Treasury & Agency purchases.
    I think I’ve understood what’s happening in Gaza. Interested readers can read my brief article today on “the oil route” and the Gaza strip conflict at my blogspot.
    I’d like to have a similarly clear and simple understanding of the analysis of PBoC policies.

  • Posted by Indian Investor

    Ok, this is off topic but I’ll risk a couple of sentences since there’s a war on.
    There is a plan to transport oil from the Caspain Sea/South Caucasus region to the Mediterranean port of Ceyhan, and thence to the Red Sea port of Eilat, through to supply markets in East Asia. Hamas appears to have been targeting an oil pipeline inside Israel with rocket attacks to disrupt this plan, and this has led to the Israeli invasion and planned occupation of the Gaza strip.

  • Posted by Indian Investor

    By studying at Princeton or Harvard you can only become a hired analyst. To get really rich you have to know what’s really happening, and they won’t tell you that at Yale.

  • Posted by bsetser

    Christian — Russia has been running down its short-term Agency portfolio, and if it held that portfolio at the NY Fed, that could explain maybe $50b of the $150b decline in Agency holdings. Korea has also been drawing on its reserves (at least til Dec) though it started doing so earlier than most (July i think). But here too it is too small to explain the full drop. The fall in korea’s reserves (not all $ and not all Agencies) is around $40b.

    The data for china suggests modest Agency sales and huge Treasury purchases — in the absence of any immediate liquidity needs. its impact has been felt mainly b/c it was among the few central banks buying Agency MBS for its own account (others outsourced this). And i wouldn’t be surprised if some other large CBs were reducing their agency holdings as well.

    2fish is right, the US gov has been buying Agencies (and lending v Agency collateral). PIMCO also rather publicly indicated it liked Agencies for the yield pickup now that they were effectively treasuries; central banks didn’t buy that argument tho. But there is also a bit of evidence indicating that the Agencies themselves adjusted their financing — issuing more short-term debt when it got harder to sell longer dated bonds (this refers to the Agencies own issuance) as CB demand at the short-end held up better than demand at the long-end. The WSJ had a good article on this in Dec.

    Finally, on topic comments please. This wasn’t a post on the savings glut; there will be one soon, which i am sure will generate plenty of criticism.

    And I personally think my basic view on the overarching issue has been fairly consistent:

    There is a need for gradual adjustment away from an unsustainable situation. Gradual adjustment implies ongoing Chinese financing of the US — not a sudden stop. Any abrupt end would be destabilizing; the Agencies provide an example of this. That analysis also applies to any asset class where CBs have been a big presence. But continuing on a trajectory where some parts of the world over-save, under-consume and over-produce and other parts under-save, over-consume and under-produce isn’t good either — so there needs to be a transition.

  • Posted by DJC

    Will Bernanke and Paulson’s plan to re-inflate the US debt bubble economy work? Not a snowball’s chance in hell.

    The United States lived in Lever-Lever Land too long. Like Peter Pan, the country has refused to grow up. The object of the stimulus plans offered by the present and the next US administrations is to return to Lever-Lever Land, that is, to debt-financed consumption. It won’t work. Leverage is for the young, who borrow to build homes and start businesses. The financial crisis forces Americans to act their age, that is, to save rather than borrow and spend.

    America’s leaders haven’t yet had the required moment of clarity. Its financial leaders still think the problem is a mere matter of confidence. These were the same people who swallowed their own sales pitch.

    Aging workers, who soon will predominate in the American workforce, missed their chance to accumulate savings for retirement and education during the boom years of 2002-2008. Instead, they borrowed cheap money from foreigners and gambled on real estate. Many analysts have drawn attention to the link between America’s zero-percent personal savings rate and the current account deficit. Americans’ home equity probably is worth half of what it was three years ago, and fall a great deal further. If they had a retirement savings plan, it is probably down by 40% or so. If they still have a job, they need to save as much as they can and make up for lost time. All the stimulus in the world won’t persuade them to spend now that they know that they can’t retire on the price of their houses.

    America will endure a lost decade more depressing than Japan’s during the 1990s.

  • Posted by Indian Investor

    @ Dr. Brad Setser:
    After a prolongued study of your various writings on different topics I’ve come to the conclusion that I have actually stumbled on a Marcus Brutus like character.
    There are two minor but very important points which especially convince me that you are probably unaware of the real themes. The first is that you wondered why Singapore had been included for a Fed swap line. The second is that you weren’t clear why it was important to negotiate FDI during the 2nd strategic economic dialogue with China. Most importantly I went through the writing style and language in your work on ‘Sovereign Debt and Sovereign Power’ and in some sense I’m convinced that you actually believe that current account imbalances are the root cause of the credit crisis.
    You wrote a paper on oil price shocks along with Dr. Roubini. I think this is a good time to have a look at it again. International crude prices have bottomed, from whatever little I can make out from Bangalore.

  • Posted by bsetser

    I do believe the current account deficit and the credit crisis are closely related — for my reasoning, see:

    As for oil, the 04 paper with Roubini frankly wasn’t my best. I would suggest instead my early 07 Peterson institute paper “oil and global adjustment.” I spent most of 06 trying to understand petrodollars and the impact of oil prices on global macro variables.

  • Posted by Majorajam

    To me one of the most damning critiques of analyses showing that official buyer stockpiling of treasuries, etc. was neither distorting nor potentially destabilizing is the seemingly binary decision making going on at these institutions. So an outright guarantee insures trillions of investment while a merely very strong ‘implicit’ or ‘effective’ guarantee merits little (even in the face of significant yield premiums). Surely the latter has to be worth something if the former does, at least in the mind of the mythical ‘utility optimizing investor’?

    To those of us that haven’t been pickled at the University of Chicago for a sufficient duration, the agency problems of official investors seem to be even more acute than they are for private institutions, for which they are debilitatingly acute. And here I’m thinking some boss holed up in some corner office in Shanghi not wanting to lose money on agencies because it’ll expose him to very severe career risk, (perhaps even execution), given that the agencies have lost face and you can’t trust a fallen woman. So a memo gets written and spreads on the entire US mortgage market jump 100 basis points. And this is the situation we’ve set up for ourselves. I wouldn’t remotely be surprised if the decision making process were more farcical than an episode of Family Guy.

    Frankly, this blissful explanation is just another legitimation of a system that was flashing red lights to anyone paying attention. Current account deficits of 7% of GDP, absorbing 80% of international net savings? Consenting adults! Total debt to GDP globally spiraling northward? A rational response to new macroeconomic stability! Consumer indebtedness, financial system leverage, unchecked derivative growth, deep structural fiscal deficits, Latin America style wealth distributions, perverse incentives for financiers and corporate governors, etc. etc. etc.? No problem, no problem, no problem! Only variations on the theme of free market boilerplate were required.

    As regards the price discovery mechanism not being distorted, the explanation doesn’t even pass basic micro. Surely a treasury purchase/sale is not unique in being bereft of consumer/producer surplus, which would act as a cushion against direct 1-to-1 displacement of treasuries for each buy or sale from a central bank? And that’s before central banks become as large a share of the markets as they have become.

    I can only hope that someday the people that furnished us with this nonsense over the last 10 years can find their true vocation at some zoo in Siberia polishing turds.

  • Posted by Majorajam


    Don’t know if this helps your case, but as I’m sure you noticed, Hank Paulson has come around to your point of view. Course, it could be worse- Greenspan’s yet to cite the linkage…

  • Posted by bsetser

    Majorajam — I saw Paulson’s comments and cringed. It doesn’t help me …

    and I was rather surprised, as I generally got the impression that Paulson was far more worried by restrictions on uS banks ability to invest in China/ the competitive threat to NY from London than by global imbalances …

  • Posted by Observer


    I hope you had a wonderful holiday season.

    I thought up a simple analogy and please let me know if it holds any validity. Suppose a guy drank 25 beers in one night, of which 23 were domestic brands and 2 were foreign. Should he blame the hangover entirely on the consumption of the foreign beers?

    Let’s look at the US aggregate demand. 2/3 of that demand comes from household consumption, which by far dwarfs the 7% current account deficit. So do should we place the entire blame on the lack of US savings on the purchase of clothing and plastic toys while turning a blind eye on people who were buying Hummers and paying for plastic surgeries?

  • Posted by bsetser

    if savings were 12% of GDP and investment 18% of GDP, foreign investors covered about 1/3 of all investment …

    and one insight of economics that has stood up well in the crisis is that prices are set at the margin — and the incremental changes in macro balances in the us in my view wouldn’t have been possible without unusually cheap credit. I wouldn’t say all of the crisis stems from foreign inflows — but i do believe that w/o non-market foreign inflows (remember us bonds didn’t perform well in fx terms until 08 … but the us still got financing) the credit crisis could never have reached the scale it did.

    My sense is that there are a number of you who argue that no one in the US had to borrow/ spend on the scale they did — which is true. but if spending fell, other variables had to adjust. one variable that might have adjusted in the trade balance/ china’s surplus (less demand = smaller surplus in china). but there are other possible adjustments. us int rates could have fallen, pushing up home prices and convincing another person to borrow … or the us governemnt, worried about slow growth, might have increased spending. one of the perverse dynamics of this period was that weak us growth meant a weak dollar — and that meant a weak rmb v the euro, a rising chinese surplus with europe (and pressure on china to tighten domestic policy to avoid overheating) and more chinese financing of the us. so some of the pressure that might normally have led to rebalancing ($ weakness = exports, pressure on rates and thus on home values and consumption) instead got deflected via the rmb-$ link …

  • Posted by Observer

    I agree that the low cost of credit was a condition for the bubble, but previous bubbles in history, including the equity bubble in the late 90s, were formed without significant non-market capital flows. I guess what I’m getting at is the fact that leveraging, or the lack of regulation against leveraging, is the key cause for producing unstable and unsustainable bubbles.

    By the way I found your paper on the Chinese banking bailout to be very informative. Thanks for sharing.

  • Posted by Twofish

    bsetser: My sense is that there are a number of you who argue that no one in the US had to borrow/ spend on the scale they did — which is true.

    That’s not quite my belief. My belief is that if the US borrowed and spent at the level that it did, it could have found far, far more useful things to borrow and spend on than what the money actually was spent on.

  • Posted by Ying

    Agree with Observer’s insight about trade. Balance of trade between US and China will hurt the low strata of working people in both China and the United States. After all, only low and medium class people in US visit Wall Mart to buy cheap Chinese goods. Only immigrants workers in China will pick up the work that nobody else wants to do. The upper class in both countries are not touched upon. Asking poor people to bear the consequences of the crisis and make adjustment is a bit of unfair. Any restructuring should provide assistance to these people.

  • Posted by Anticipation

    twofish>>which leads us to the next question, what are we actually hoping for with a say “800B Obama Stimulus”? borrow more and spend more on what was actually spent in the past?

    Thus the cycle continues in hopes to ignite the same motor which drove us down the path we are on and refuse to swerve off, until…

  • Posted by adiemuso

    ying i agree with you. the poor and middle gets marginalised..

    however, sad but true, the ideal case only happens in Utopia.

    look at how they managed to bailout the rich folks, be it in US or UK or even in Asia.

  • Posted by Fugger

    Central banks (in currencies other than their own) have removed huge amounts of liquidity from the money markets since August 2007 as they have become very risk averse. There is a nice irony in comparing their liquidity withdrawal with the liquidity provision by central banks in their own currencies. Of course, the central banks withdrawing liquidity are generally from different countries to those adding it. Examples of the withdrawal of liquidity include:
    – the switch from 3-12 month unsecured bank deposits (often placed via the BIS) to short term government bond repo; BIS provides some great data on this in its quarterly reviews
    – stopping lending their Treasury bonds in H2 2008, causing the fails in the Treasury repo market (a vicious circle as stopping lending led to fails led to reduced lending – all exacerbated by the lack of penalty for failing when interest rates are very low, which is now being addressed).

  • Posted by Chris

    What are the Y axes? Billions of dollars? Percent change? Something else?

  • Posted by bsetser

    Billions of dollars. I rarely graph % changes (except for export/ import graphs). Most of my capital flows/ current account graphs are in either $ billion of a % of US or world GDP.

  • Posted by Chris


  • Posted by London Mortgage Rates

    The trouble with old school stable financial establishments is that they don’t want any new kids on the block to overthrow their power with ‘radical’ new ideas.