Safe assets and Triffin's dilemma

"Providing reserves and exchanges for the whole world is too much for one country and one currency to bear."

Henry H. Fowler

U.S. Secretary of the Treasury

There has been quite a bit of puzzlement in some quarters as to why I savaged BIS over the whole idea of a limitless supply of government-produced safe assets purely to meet the needs of the financial system. Firstly, let me make it clear that I do not believe that any asset is ever really "safe". Nor do I believe that global investors have any right whatsoever to expect national governments to provide them with what amounts to unconditional backing for their deposits. The first duty of national governments is to their people, not to the needs of a global financial elite. And it is quite wrong of global financial elites to pressure governments into issuing debt that they do not need purely in order to provide them with liquidity. Nor should global financial elites impose austerity measures that are not warranted by the economic situation, purely to reassure themselves that the assets they rely on for liquidity cannot ever become unsafe.

So that is my political stance, if you like. But this post is concerned with the particular effect on the US of producing global safe assets in much the same way as it has hitherto produced the world's reserve currency.

Firstly, a reminder of the global safe assets scheme. The BIS paper envisaged that certain governments - notably the US - would provide enough debt assets to meet the needs of the financial system, making up the current (considerable) shortfall. This debt issuance would be considerably more than actually required to meet government spending obligations, and it would mostly be of short tenor (Pozsar). This would leave the US government with surplus funds. However, because investors would expect the effect of actually spending that money to fuel inflation, which would reduce the value of the debt, the BIS authors effectively recommended that the government should not actually spend the funds raised from debt issuance. On the contrary, to ensure that the government could meet its interest and refinancing obligations at all times, the BIS authors expected the government to run primary surpluses - in other words keep public spending commitments below tax income. In addition to the surplus funds from debt issuance, therefore, there would also be surplus funds from the excess of tax income over public spending. These funds would presumably be placed on deposit at the central bank.

I pointed out in the post that from a national macroeconomic position this makes no sense whatsoever, and would lead to progressive indebtedness and impoverishment of the supplying country's private sector. This is where people lost the thread, and I assume that was because I did not explain the economic effects of such excessive debt issuance in an austere fiscal environment. So let me explain here using the sectoral balances equation that I looked at in my last post.

The basic sectoral balance equation is this:

Private sector (savings - investment) = Government (spending - tax income) + External (net exports - imports)

In algebra:              (S - I) = (G - T) + (X - M)

The first thing to note is that if government debt is bought by foreign investors, those investors must have currency with which to buy it. So the government must also issue more currency than the country itself needs, and the excess must find its way to overseas residents. This would have to be done by running a substantial trade deficit. Therefore in the equation, (X - M) must be significantly negative.

The excess US dollars that find their way into the pockets of overseas exporters are exchanged for US government Treasuries through financial sector intermediation, and are repatriated to the US. This creates a surplus in the capital account equal (more-or-less) to the trade deficit. At this point we simply have potentially inflationary expansion of the US government's balance sheet. But our investors don't like inflation, and the US government doesn't want investors to lose confidence in US Treasuries. So the BIS authors assume that the US government holds down government spending and keeps taxes high in order to generate a primary surplus (T > G). Therefore (G - T) is also negative. To keep it simple I will allow the uninvested proceeds from debt issuance simply to disappear from the equation (because this equation assumes that G includes all money received from debt issuance, which in this scenario is not true).

We now have two negative terms on the right-hand side of the equation. Therefore (S - I) must also be negative. To show the effect of this on the US private sector, I'm now going to use the enhanced version of this equation developed by Monetary Realism. The enhanced version of this equation looks like this:

S = I + (S - I) = I + (G - T) + (X - M)

where I is private sector direct investment and (S - I) = dNFA, the change in private sector holdings of net financial assets. Note that according to the first equation, those net financial assets are government liabilities (debt and cash) and foreign liabilities (debt and cash). Now we already know that the three terms (S - I), (G - T) and (X - M) are negative. So the combined effect of the large trade deficit and the primary surplus is that the private sector's holdings of net financial assets have fallen. In other words, household and corporate financial savings have been raided.

That's bad enough. But look also at the effect on I, which is direct investment by the private sector in the domestic economy. I is reduced by the combined total of the primary surplus and the trade deficit. So not only have domestic savings been raided, domestic direct investment has also been clobbered. Money has been drained from the domestic private sector. It has been divided between a ridiculous government surplus and cash backing for future claims of nervous global investors.

Of course, the private sector can maintain investment if it cuts consumption spending, since savings are the excess of disposable income (after taxes) over consumption spending. This would force imports to reduce, but since our investors would like that (because most of them assume that a large trade deficit is a Bad Thing even though it's necessary in order for them to have currency to buy their safe assets) it would also be likely to increase the value of both the currency and the debt.  But notice how contractionary all this is. What it amounts to is permanent austerity for the domestic population. Progressive impoverishment, simply to preserve the perceived safety - and therefore the value - of government debt (and by extension the currency). Either starve the domestic economy of investment, or starve its people. Or both.

It could be argued that the over-production of both debt and currency to meet global needs is pretty much what the US is doing already. But there are two very important differences. The US not only runs a large trade deficit, it also runs a fiscal deficit: the currency repatriated through the sale of government debt to overseas residents is used to meet government spending commitments in excess of tax income. So although (X - M) is negative, this is offset by the fiscal deficit (G > T). Yes, I and/or C aren't as large as they would be if the US was running a trade surplus. But they aren't disastrous - as they would be if the US tried to run a primary surplus while maintaining debt issuance.

Over-production of government financial assets by dominant world nation (s) to meet the needs of the financial system is nothing new. It has been going on for centuries. Yes, in the past it has mainly been currency that was in demand: the financial system's need for government debt as a safe asset is a new phenomenon and arises from the fundamental change in the nature of money that I've discussed previously. But the problem is an old one. It is known as Triffin's dilemma, and it was the proximate cause of the failure of Bretton Woods.

As the IMF explains, the US had been running balance of payments deficits throughout the 1950s, which flooded the world with dollars. The world became dependent on a plentiful supply of dollars for liquidity and reserves. But the buildup of reserves exchangeable for gold by central banks threatened to drain the US's gold reserves. The economist Robert Triffin outlined the problem as follows:

  • If the US stopped running trade deficits, the world would suffer a crippling shortage of liquidity which could push it into a deflationary spiral, leading to instability
  • If the US continued to run trade deficits, confidence in the dollar would decline, leading to its rejection as the world reserve currency and breakup of the Bretton Woods system, leading to instability.
In other words, the US had to run a trade deficit in order to maintain dollar liquidity, but it had to run a trade surplus in order to maintain dollar confidence. Clearly it couldn't do both. What actually happened was the second of these. Bretton Woods collapsed in 1971 when Nixon suspended the convertibility of the dollar to gold, preferring that to the loss of the US's entire gold reserves. The dollar has not been rejected as world reserve currency, mainly I suspect because there really isn't an alternative. But eventually there will be, and then the US dollar will fall from its pedestal and the US will be left with the burden of an extraordinarily large trade deficit. I do not look forward to that day with any pleasure.

Proposals for over-production of US government debt suffer from the same problem with knobs on: the US must run both trade AND FISCAL deficits in order to maintain excess debt production without impoverishing its population, but it must run at least a fiscal surplus and ideally a trade surplus as well in order to maintain confidence in the debt. This is not a sustainable solution to the global safe assets problem.

There can be no solution to Triffin's dilemma that relies on nation states. The only possible solution would be a supra-national body providing both reserve currency and safe assets to the world financial system. The present proposals for safe assets production do not go far enough.

Related links:

Coppola Comment       When governments become banks
                                    The illusion of safety
                                    Modern gods and human sacrifice
                                    Consumption booms and austerity
                                    Liquidity trap heralds fundamental change
BIS working paper       Global safe assets
IMF working paper      Institutional cash pools and the Triffin Dilemma
Investopedia                 How the Triffin Dilemma affects currencies
IMF                             System in Crisis: The Dollar Glut


  1. I don’t see a big problem here. A country whose currency / debt is in demand around the world needs to let the world accumulate its currency / debt while ensuring the real or inflation adjusted rate of interest on the debt is slightly negative. That way, the issuing country makes a profit out of its creditors.

    The only real problem comes if too many creditors want their currency / debt converted to another currency at once (a possibility Frances alluded to, I think). But if creditors do that, they shoot themselves in the foot: the value of the former currency drops. For example China’s big holding of US dollars / debt means China has an interest in maintaining the value of the dollar.

    1. Ralph, I don't think the prospect of a buyer's strike is the main problem here. It is the conflicting demands of the financial system for simultaneous liquidity and safety. These two demands together place an impossible burden on the real economy in the supplying country. Yes, there is benefit for the supplying government because its borrowing costs will be very low - may even be negative, as I've noted before. But the combination of fiscal austerity (to maintain the value of debt) and trade deficit (to maintain the supply of debt) is toxic and unsustainable.

    2. I don’t think the restrictions demanded by foreign debt holders are so burdensom that they result in your “austerity” and for the following reasons.

      At what level of inflation are foreign debt holders really going to take fright? Not 2% or 3%. But possibly say 5%? But does a country really need to incur 5% inflation to impart adequate stimulus? I suggest that if stimulus does not work at about 3% inflation, then bumping up demand still further and raising inflation to 5% or more won’t do much to reduce unemployment or “austerity”. I.e. in the 3-5% range, I suggest one is in the scenario where further increases in demand primarily boost inflation, rather than boost output.

  2. "And it is quite wrong of global financial elites to pressure governments into issuing debt that they do not need purely in order to provide them with liquidity."

    It is not solely about liquidity, the same people applying that pressure are the same people who will, ultimately, own and profit from the debts generated.

    1. The flow of safe assets in the financial system is rather like pass-the-parcel - you don't know who will be holding the parcel when the music stops. So yes, financial actors are interested in safe assets retaining their value, because they hope to benefit. But the financial SYSTEM's primary demand is for liquidity in safe assets.

  3. The first duty of national governments is to their people? Wow! You sound a lot like the Prince of Liechtenstein How refreshing to see you talk in such an erudite manner. Perhaps you will, after all, consider:

    The Market For Liberty

    1. Oh no, not you again. I know I've said it before, but please talk about the SUBJECT, not your beliefs. And please refrain from personal attacks: "how refreshing to see you talk in such an erudite manner" is bordering on rude. Oh, and please identify yourself.

  4. Hi Frances,

    Just a quick note about safe assets. I agree there is no such thing. What we have instead is a hierarchy of informationally insensitive debt (I expect you have read that Gary Gorton paper on the subject entitled 'Slapped in the face by the Invisible Hand').

    Banking works by taking a bunch of highly informationally sensitive liabilities (incurred by bank loan customers), removing all the relative information from those loans and then issuing a bunch of bank debt, they key feature of which is that one dollar of this banks debt cannot be distingiushed from another dollar of that banks debt.

    A safer asset than that bank debt is debt issued by many banks which has been insured by some government agency. This has the effect of removing even more information (and increasing the mutual information between) dollars of bank debt issued by different banks.

    An even safer asset than that, with even higher mutual information between two dollars of it, is the direct liabilities of the government, which has a printing press. Marginally more informationally sensitive (and hence slightly less safe) is a government bond, but with a flat yield curve the information is equivalent to that of a dollar of currency.

    Even more informationally insensitive than that is the blended liabilities of national governments using debt issued by a world bank. After that there would be a bit of a problem. The safe asset is that which is most informationally insensitive, and of course this is all relative.

    Now, all this willful obscuring of genuine information in order to create safe assets increases mutual information between economy participants but it has a price which is to increase uncertainty in the overall evolution of the economy.

    Now all these safe assets have been created, their creators find that they are causing enormous economic problems. But they can't easily be unpicked, because this process is essentially an irreversible one, in a thermodynamic/entropic sense. They can however be deleted.

  5. The Fed explained: “With the payment of interest on excess balances, market participants will have little incentive [sic] for arranging federal funds transactions at rates below the rate paid on excess. By helping set a floor on market rates…”


    Libor, Sibor, & Hibor reflect USD's (Euro-dollar's, Yaun-dollar's, Yen-dollar's) inter-bank borrowing rates. Libor rates, etc., are the functional equivalent of the FOMC's effective federal funds rate. The E-D system doesn't operate under reserve & reserve ratios constraints. The size of the E-D system is many times that of our Federal Reserve System of commercial banks.

    The size of the U.S. domestic money stock is independent of Libor borrowing rates. Thus the size of domestic commercial bank credit is likewise unaffected (except through the FOMC’s mis-management).

    Libor wasn't out-of-alignment with other interest rates. The E-D system is an international, prudential reserve, money & credit creating banking system (without a Central Bank payment, clearing, & settlement system). The E-D system is unregulated & without Central Bank backstops & FDIC insurance coverage.

    The higher Libor rates represented greater cross-border liquidity strains & inscrutable counterparty exposure. The BOG’s delayed crisis response was to provide temporary: dollar liquidity lines & foreign-currency liquidity lines.

    The E-D system underwent an unchecked, cumulative & reinforcing, bank credit contraction (as all previous prudential reserve systems have). The E-D market’s collapse set off the global economic recession beginning July 2008.

    As the E-D banking system contracted, the dollar's exchange rate reversed. As interest rate differentials (from the Fed’s credit easing – not quantitative easing) became less important than counterparty risk, the dollar strengthened & U.S. commodity futures sold off.

    The Fed underestimated this international dollar shortage. Central bank liquidity swap lines were underfunded. Meanwhile the Fed overestimated future gDp growth. Both long-term & short-term money flows turned negative.

    The effective daily Federal funds rate was stable up until Lehman Brothers’ bankruptcy. Even afterwards, the daily volume of fed funds transactions was comparable up until the introduction of the payment of interest on reserves.

    Commercial bank credit began to contract 2 weeks after the BOG started paying the banks not to lend/invest. The FOMC’s zero interest rate policy (credit easing) failed to expand Federal Reserve Bank Credit. And the untried & untested interest rate “floor” dissuaded the commercial banks from injecting new money & credit into starved economic sectors.

    If you don't understand what happened during the credit crisis of 1966 (dis-intermediation solely for the non-banks), you can't understand the payment of interest on excess reserve balances.

    1. Thanks, that's really interesting (though not related to this post, of course). I did know some of it but you've added some things to think about.

    2. I think you are bit confused. Leaving aside the incorrect definitions of sibor and hibor, libor rates are not the "functional equivalent of the fed funds rate", they are rates on short-term unsecured interbank lending and the participant banks (which includes most US banks) are all regulated and required to maintain reserves.

      Libor rates spiked higher in 2008 because European and other banks lost access to their main source of USD funding (money market funds) following the collapse of Lehman Brothers. This did not cause the global recession, it was just another symptom of a collapsing bubble. 3m Libor was stable up until the collapse of Lehman Brothers too. The Fed may have underestimated the consequences of allowing them to fail, but they and other central banks reacted swiftly to address the liquidity shortage and a few months later 3m USD libor went from its 5% peak to 1%.

      I am not sure what you are trying to suggest in the last to paragraphs, but it seems to be that banks are not lending because the Fed pays them interest on excess reserves. When the private sector is deleveraging there is little demand for loans so it is not clear changing the IOR from minimal to zero would have any effect.

    3. I think it is fair to say that when short-term interest rates are very low then IOER is an incentive not to lend. But then so are T-bills, and those are always available. The problem is banks' own deleveraging process, which is making them very reluctant to take on risk, coupled with regulatory requirements for higher levels of safe liquid assets. But I also agree that there is little demand for lending anyway.

    4. FLow5

      I've now read up a bit on the 1966 credit crisis and I agree there are some parallels. But there are also major differences, the first and largest of which is the fact that the world was using the Bretton Woods system of managed exchange rates, under which the dollar was effectively on the gold standard and all other currencies were tied to the dollar. That is very different from our present fiat currency system with fully floating exchange rates. What is quite surprising, given that, is how similar the Fed's actions were to those of 2008/9. But they didn't introduce interest on excess reserves in 1966. They capped the time deposit rate below the prevailing market rate, which severely reduced the flow of household and corporate money INTO banks because there were better rates available on the money markets and in thrifts.

      The introduction of IOR stabilised the overnight repo rate, which was in freefall. Had this not been done, the Fed Funds rate would simply have become irrelevant. We now have excess liquidity in the system, which means the natural repo rate is zero. IOR creates artificial demand for reserves and therefore enables the Fed Funds rate to remain positive.

      danishcrows is correct about the source of the disturbances to LIBOR after the fall of Lehman being due to money markets freezing. Specifically, it was because Reserve Primary MMMF broke the buck, which caused a run on tri-party repo.

  6. James seems to think that “debtor governments” should continue “borrowing” despite the “stigma” of pushing debt onto citizens who did not voluntarily sign up for whatever particular portion of (international) debt you may be thinking about - what do you make of this, Frances? Should a stranger be permitted to sign you up for some debt without your consent?

    1. I think I have made my views clear on the subject of national governments providing international safe assets.

  7. I think you need to ignore, and not respond to the comments from Anonymous. Even better would be to delete, or not show, his comments to begin with.

    1. I really don't like censoring comments. But he is getting annoying - and evidently not just to me.

  8. 1966 is the same paradigm. It's no where in the economic literature. It takes an Einstein to understand this. The payment of interest on excess reserve balances induces dis-intermediation (where the size of the non-banks shrink, but the size of the CB system remains the same). The remuneration rate permitted the CBs to pay higher rates on funding, than the NBs could competitively meet. The outflow of funds from the NBs reduces the overall supply of loan funds (but not the supply of money).

    IOeR alters the construction of a normal yield curve, it inverts the short-end segment of the yield curve –known as the money market. The money market is differentiated by its position on the yield curve (i.e., short-term borrowing & lending with original maturities from one year or less). @ .25% the remuneration rate on excess reserves is still higher than the Daily Treasury Yield Curve almost 2 years out.

    The liquidity run during the Great-Recession was compounded by a flight to safe-assets by both the CBs & the NBs…

    The non-banks are the most important lending sector in our economy — or pre-Great Recession, 82% of the lending market (Z.1 release, sectors, e.g., MMMFs, GSEs, etc.).

    Political efforts should encourage the flow of savings thru the NBs as was required to resurrect the housing market in 1966 (NBs are the customers of the CBs). IOeR does the opposite. It stops & reverses the flow of savings into real-investment.

    CBs don’t loan out existing deposits (saved or otherwise). The lending capacity of the NBs is dependent upon monetary policy, not the savings practices of the public. The CBs could continue to lend if the public ceased to save altogether.

    See Gilbert's: Requiem for Regulation Q: What It
    Did and Why It Passed Away. From Gilbert's paper it is evident that all economists don't know the differences between:

    the supply of money & the supply of loan funds

    between means-of-payment money & liquid assets

    between financial intermediaries & money creating institutions

    that interest rates are the price of loan-funds, not the price of money

    that the price of money is represented by the various price (indices) level

    that inflation is the most important factor determining interest rates, operating as it does through both the demand for and the supply of loan-funds

    Bernanke is directly responsible for the Great-Depression. Bernanke conducted 2 separate contractionary monetary policies. Negative roc's in MVt in Dec 2007 had already predetermined the fall in gDp in the 4th qtr of 2008. The IOeR introduction was the coup de grâce.

    1. I've read Gilbert's paper and I'm afraid I don't agree with the conclusions you draw from it. Some of the things you say are correct - for example the fact that IOER props up the short end of the yield curve, and the flight to safety in the financial crisis. However, your view that the contraction of lending back to the commercial banking sector is "disintermediation" is an incorrect use of the term. Disintermediation occurs when commercial depositors and borrowers bypass banks, not when they return to them.

      I'm very well aware that commercial banks don't "lend out" existing deposits. In which case doesn't it occur to you that the reason they aren't lending to NBs is because their balance sheets are terribly damaged and they are trying to de-risk and repair them? I really think you attribute far too much to central bank monetary policy. Throwing money at banks doesn't make them lend. Neither would cutting interest rates on reserves. They aren't going to lend while the risk-return nexus is horrible for them.

      You also have a very US-centric view. The US's financial system is unique in the world because of its dependence on NBs. Everyone else uses commercial banks far more. And we are experiencing exactly the same problems with commercial banks not lending, even though we don't have such a large NB sector. The ECB's IOER is zero - but European banks still aren't lending. There is a massive contraction of bank lending going on across Europe.

      The financial crisis (Great Recession) started in 2007, not 2008. Run on ABCP triggered by the failure of Bear Sterns hedge funds in August 2007. We in the UK knew about this because it caused the failure of Northern Rock. Lehman was towards the end of the financial crisis, not at the beginning of it. The Fed was way behind the curve in dealing with the crisis in 2007, but then so was every central bank.

      As my background is in banking & finance, not simply economics, I think I do understand the things you list. And I think you are wrong.

    2. Actually, on further deep reflection I eat almost every are on to something that I don't yet fully understand but I think is really important. It's all about deposits, not lending.....I will mull it over and get back to you. In the meantime I would appreciate it if you would say more about your thinking on this, either on this post or preferably by email:

    3. Ok, I think I've got it. Commercial banks are borrowing from MMMFs (repo). But then instead of lending those funds on down the shadow banking chain as they would normally have done, they are parking the borrowed funds at the Fed for IOR. Earns them a few basis points in spread for zero risk. The effect is to prop up the repo rate, which gives MMMFs some income, but it doesn't help people further down the shadow bank funding chain if banks are short-circuiting the whole thing.... The Treasury curve short-end prop presumably is an arbitrage with the IOR, since T-bills are a substitute for reserves. Though there is excess demand for T-bills anyway because of regulatory requirements and a general move to collateralised lending.

      European banks were doing the same until the ECB cut its deposit rate to zero, which killed this off and clobbered European repo rates. However, the Euro MMMFs then shut their doors to new funds. Thing is, you see, that when the natural repo rate is zero or negative, no-one can make any money from lending, so no-one lends. IOR actually props up the repo rate, so I'd hazard a guess that without it things would be even worse. But banks round-tripping is not helping matters.

      But IOR is only half the story. The other half is QE. The combination of QE and IOR is poisonous. Banks are being provided by the Fed with excess reserves on which they can earn 0.25% for doing absolutely nothing at all. In that respect you are right that they are effectively being paid not to lend.

  9. Correction: "The lending capacity of the (CBs) not (NBs) is dependent upon monetary policy

  10. [...] Although the debate appears to be dying down, at least momentarily, Simon Wren-Lewis and Frances Coppola (see here, here, and here) have added worthwhile readings. [...]


Post a Comment

Popular posts from this blog

WASPI Campaign's legal action is morally wrong


The foolish Samaritan