Monday, April 20, 2009

More on the SDRs as a Reserve Currency

In my last post on the SDRs Rebecca Wilder questioned whether the SDRs will actually make a difference to the global economic imbalance problem. As a follow up to her question, I asked a friend who formerly worked at the IMF and the U.S. Department of Treasury his thoughts on the matter. Here is what he had to say (note that "Fund" is short for IMF):
This is a very long, tricky subject that I think starts with what is an SDR really? An SDR is basically a claim on another countries' reserves held via your account at the Fund. So when a country needs more USD it cashes in the SDRs at the Fund and the Fund tells the US Treasury to provide the USD. The country then runs a debit on its account at the Fund, which it pays interest, while the US gets interest on the "surplus" SDRs it has. The main point of these accounts is that some countries have hard currencies, while others (lets call them Argentina) have currencies from time to time that no one wants - so it's just a more complicated way of giving more countries access to currencies of other countries that are in demand at that time.

This doesn't really seem to me to be a workable currency yet - (1) it's not freely exchangeable in the market; (2) you can't just park reserves you accumulate in SDRs - if China has surplus USD, it can't simply go to the Fund and say I'd like to swap these for SDRs please; the IMF would have to fundamentally alter the way SDRs work to get to a system where the SDR could be usable as a reserve currency; (3) who controls the stock of SDR outstanding? When the Fund "creates" SDRs, all it is really doing is putting in place an agreement across its members to allow their currencies to be tapped by other Fund members when needed. The Fund can't grow the underlying SDRs out of thin air - every country would have to agree to put more of their currency on call. For the US , the Treasury Department would likely issue debt to raise the USD to give to a country that wants USD for its SDRs. Other countries might simply have their central banks print the money - but ultimately it's finite.

Beyond that, Rebecca's point is a decent one - if China is determined to keep its currency undervalued, then the real question is what currencies does it do so against. But that begs the question of why China would need SDRs in the first place; it could simply set an fx reserve accumulation policy of buying some USD, some EURO, some yen, and some GBP; I guess that would help the US by spreading some of china's capital outflows across more countries - and hence allow the USD to depreciate more relative to the CNY - since they would engineer a broader nominal effective depreciation.

If China's motive is more feeling the need for fx to insure against sudden capital outflows, then the Fund could help in theory. In practice it's hard since until the FCL the only way the Fund could deliver assistance was via programs which both had stigma and because they carried conditions meant that access to IMF resources wasn't automatic - so not a great way to insure yourself if you are a country; much better to have your own stockpile of USD.

A reformed SDR could play a role in this I think by being say a unit of account for parking reserves; but really the way the IMF could fulfill its function as insurer of last resort better is instead of telling a country to go run CA surpluses, accumulate a bunch of USD and bring them to us (and then what? does the Fund buy US Treasuries on China's behalf? that doesn't seem to solve any problems) is instead the Fund should act as a central clearinghouse for Central Bank fx swap lines. So when demand for a country's currency falls, it could go to the Fund and get more of the currency that demand has shifted into. Of course we go back to square one - is the country's currency falling because of unsustainable policies which require external adjustment - or is it just a temporary capital account fluctuation?
In short, the SDRs are not a panacea for the global imbalance problem.

Update: for a pro gold standard perspective on SDRs see here.


  1. David,

    Larry White says what I should have said in the comments of your previous post, but instead muddled around discussing the SDR in the context of multi-metallic standards and scaring off other commenters. Here is Larry's quote:

    "Unlike gold, the SDR is not anchored to anything but a basket of unanchored national fiat monies (what Zhou calls “credit-based national currencies,” oddly given that their issue is based on fiat and not on credit in any standard sense). If the SDR were its own fiat unit it would remain unanchored.

    Unlike monetary gold, the SDR in either form is governed by no market forces making its quantity automatically respond to demand."

    This is fundamentally the point that I was trying to make. So long as the SDR is backed by fiat currency, I fail to see how this is any improvement over the current system.

    Indeed, this was the fundamental problem with Bretton Woods. It was basically a bimetallic system with one of the "metals" being the dollar. The problem is that although the dollar was tied to gold, it was still a fiat currency. Thus, when inflationary monetary policy took hold, the entire system broke down.

    The problem with the SDR is that it is completely backed by fiat currency and, what's worse, it is managed by the IMF, which is a political institution (as Larry White alludes to in his post).

  2. Josh:

    Just went over with my class the differences between the pre-WWI classical gold standard and the interwar gold standard. Among other things, we discussed B.Eichengreen's point that as more people became enfranchised it becomes incredibly difficult to follow the rules of the game with an international gold standard. Short-term domestic economic concerns trump any external economic balance questions. From that perspective it is not surprising that the classical gold standard worked well--with less enfranchisement--while the Bretton Woods eventually collapsed. For this reason, I am not sure there will ever be an international monetary system that isn't compromised for domestic concerns.

  3. David,

    I simply do not think that Eichengreen is correct. Okay, maybe that is a little too harsh, but if he is correct, he is only half correct. Allow me to explain.

    Under a fractional gold standard there is sufficient room for maneuverability for monetary policy in the short run. However, monetary policy is severely constrained in the long run. Thus, the fractional gold standard contains the proper institutional framework to promote good monetary policy.

    I don't know if anyone still reads it, but Jurg Niehans The Theory of Money is one of the best texts I have read on monetary theory. His chapter on commodity money alone is worth the cover price. He makes the point above by saying the a fractional gold standard prevents bad policies and bad economics (and this is someone who favored a fiat currency) through this long run constraint. In other words, there is room for countercyclical monetary policy, so long as it is subsequently corrected.

    The Bretton Woods system is different from the prewar gold standard in that the dollar was essentially a fiat currency although it was pegged to gold. However, this system theoretically combined the best features of a bimetallic standard and a fractional gold standard in that:

    (1) the dollar and gold could basically serve as buffer stocks for one another.

    (2) there was room for short-term monetary policy.

    Eichengreen is correct in saying that the enfranchisement caused domestic concerns to take precedent. This is the basic story that the system was inherently fragile. However, this is not the case. What caused the system to collapse was bad monetary policy. The ultimate irony is that a system that characteristically prevents inflationary policies actually collapsed because of inflationary policies. However, this is not because of enfranchisement, but rather bad long-term monetary policy of the 1960s and early 1970s.

    Thus, the system collapsed not because of a desire to satisfy domestic concerns, but rather because of long-term inflationary policies. Now, one could spin this as enfranchisement, but such policies are bad as long term strategies whether the currency is backed by/pegged to gold or fiat.

    Basically, the point that I am trying to make is that bad monetary policy is bad monetary policy and that short run domestic concerns, such as cyclical unemployment, are not sufficient to collapse a gold standard (whether the prewar standard or the Bretton Woods system).

  4. Thanks for the tip on Jurg Niehans' The Theory of Money. I am interested in learning how there can be short-run maneuverability of a gold standard.

    Here is my question: are bad economic policies themselves partly the product of the political pressures inherent in a democracy. I am not saying it is impossible to have good policies with full enfranchisement, but on the margin it got to be more difficult to make tough economic choices. Think, for example, of the difficulties with free trade policies.

    Paul Volker demonstrated you can make tough choices in the face of immense opposing political pressure. I wonder, though, whether he would have been able to make the tough calls if he had had to deal with the 24-7 media coverage we have today.

  5. David,

    I should clarify that there is short run flexibility under a fractional gold standard, in which the amount of monetary gold is only a fraction of the money supply. This does not apply under a pure gold standard. I don't know how well this fits with the interwar period, but it certainly fits the postwar era up to the collapse of Bretton Woods.

    Also, you write,

    "Here is my question: are bad economic policies themselves partly the product of the political pressures inherent in a democracy. I am not saying it is impossible to have good policies with full enfranchisement, but on the margin it got to be more difficult to make tough economic choices."

    I am not refuting this claim. However, the point that I am making is that under the Bretton Woods system, there was the flexibility to conduce short run expansionary policies so long as this expansion was corrected over the long term. The problem is that the 1960s and 1970s were not examples of short term expansionary policies, but rather perpetual policies.

    Gold is simply the nominal anchor. If the inflation forecast is the nominal anchor, would this not similarly be prone to enfranchisement?

    My fundamental point is that we know that short term monetary policy can be expansionary. However, in the long run, we simply end up with inflation. The Bretton Woods system essentially put a long run constraint on monetary policy. However, the U.S. simply inflated its way out of the constraint and the system.

    With an independent central bank and a clear recognition of what monetary policy can and cannot do, I simply do not understand how enfranchisement matters. In other words, there was as much of a problem with the idea of the role of monetary policy during the Bretton Woods era as there was with enfranchisement. Just look at Friedman's 1968 AEA Address. At the time it was see as somewhat controversial. Today, much of it is generally accepted.

    Thus, I think that bad policy is much more to blame than enfranchisement.