A few lessons for Mary Anastasia O Grady
By Brad Setser
Mary Anastasia O’Grady’s column today is centered on a cartoon version of reality — a world where the “Clinonistas fueled moral hazard by bailing out Wall Street cronies, who were up to their ears in high yielding debt,” a world where the IMF “lost political support for unrestrained lending to insolvent governments — aka bailouts” after Argentina’s default in 2001, and a world the “IMF’s good housekeeping seal of approval on Argentina” was the only reason why “markets pumped in money so liberally.”
All three arguments are myths.
- The claim: “The Clintonistas” fueled moral hazard by bailing out Wall Street cronies”
The facts: The Clinton Administration pulled the plug on the IMF’s bailout of Russia in the summer of 1998 after only $5 billion of the roughly $15 billion IMF package ($20 billion including the World bank and others) had been lent out. As Rubin’s memoirs make clear, this was not an easy decision. Russia was considered too nuclear to fail for a reason (Argentina, in contrast, is “a dagger pointed at the heart of Antarctica, to use Kissinger’s memorable phrase). But Rubin strongly believed that there was no point throwing good money after bad; he prevailed against the rest of the NSC, which wanted to keep on lending. “Wall Street cronies” were left holding lots of exposure to Russia — whether ruble denominated GKOS or London Club debt. Russia’s default certainly should have eliminated any expectation that every country would get a bailout large enough to let Wall Street get off scott-free.
A bit of math. In 1998, Russia got $5 billion. At the time of its devaluation and default, Russia had at least $50 billion in GKOs/ OFZs outstanding (not all held by foreigners, but foreigners are also not the only ones bailed out by the IMF), around $30 billion in London club debt, and a few eurobonds as well. $5 billion in GKOS may have gotten off the hook courtesy of the IMF, but the $80 billion that remained and was restructured “learned” a lot about the risk of lending to emerging economies.
But there are other examples as well. The Clinton Administration did not bailout bondholders in Pakistan, Ukraine or Ecuador — all of whom had to restructure their debt. And in Korea, after the initial bailout package failed, the Clinton Administration decided to rope in the banks, and pressure them to roll over their loans — not to provide all the funds needed to let everyone get out. Compare that with the Bush Administration in Turkey: throughout 2001, international banks were cutting their lending to Turkey as the IMF was putting its money in. The correlation is almost perfect: the IMF put $13.2 billion into Turkey in 2001, the banks took $11.5 billion out. I call that a bailout. The Bush Administration, for ideological reasons, was opposed to putting any pressure on the banks.
- The claim: “IMF lost political support for unrestrained lending to insolvent governments.”
The facts: If there ever was a case of clearly lending to an insolvent government, it was the Bush Administration’s decision to push for a $8 billion expansion of the IMF’s bailout of Argentina in the summer of 2001. At the time, it was pretty clear — as clear as it is ever going to be — that Argentina was effectively insolvent, and eventually was going to have to forcibly restructure its bonds. The Bush Administration pushed for this loan, the rest of the G-7 was not so keen. $5 billion of the $8 billion augmentation was disbursed — so the augmentation alone provided about 1/2 of the $10 billion (net) that the IMF provided to Argentina (The Clinton Administration had approved a $15 billion loan in December 2000. However, that loan was comparatively back loaded — about $5 billion of that loan was not disbursed — and a large fraction of the IMF’s lending just refinanced payments coming due, so the net disbursements from this loan in 2001 were only around $5 billion.)
For a full account, see Paul Blustein’s new book on Argentina’s crisis.
Moreover, the Bush Administration kept on backing large IMF loans to quite indebted countries AFTER Argentina defaulted. Turkey got a huge package ($18 billion) in 2002. Brazil got an even bigger package — $30 billion — in 2002. Uruguay got $3 billion, which does not sound all that impressive until you realize that $3 billion is about 15% of Uruguay’s GDP. All three countries had very high levels of public debt. And if you doubt my argument, do look at what happened to the amount of IMF loans outstanding in the course of 2002 (data is, alas, in SDR).
Let me put this way: outgoing Treasury Under Secretary John Taylor did not get a medal from the Uruguayans for saying no to their request for help.
- Claim: the “IMF’s good housekeeping seal of approval” explains why the markets dumped money at Argentina.
The previous arguments are easy to document by looking at the IMF’s balance sheet. This one is a bit harder — I certainly don’t know for sure why investors lent so much to Argentina in the 1990s.
But I do know this: anyone who lent to Argentina after Russia did so knowing that no country was guaranteed access to enough IMF money to avoid default, and did so knowing that if they held a long-term claim, like a bond, they probably were not going to be able to get out on the back of the amounts of money the IMF typically makes available. Holders of short-term creditors have a better chance of getting out on the back of IMF lending, or the country’s own reserves. Holders of a ten year bond cannot get out until the bond matures. They can sell their bond — but that just shifts its ownership. Ecuador’s 1999 default and restructuring made it absolutely clear that bonds were not exempt from being restructured.
The IMF also — as Paul Blustein documents — issued some rather public warnings (by IMF standards) in the spring of 1998, warnings that the markets brushed off. I think the IMF should have been a bit louder, but it is a bit rich to attribute all bad investment in emerging economies to the IMF, and not put any blame on the markets own tendency to reach for yield.
Or, to put it differently, is the current wave of lending to emerging economies a product of the Bush Administration’s willing to use the IMF to bail out Uruguay, Brazil and Turkey, which, in turn, has triggered a new wave of moral hazard induced lending? Or is it a product of low yields in advanced economies, cheap short-term money and the carry trade?
A final note: most of the data referenced here comes from the book Nouriel Roubini and I did on crises in emerging economies, which, as this weeks’ Economist notes, if nothing else, is “admirably thorough.”
In that spirit, a couple of notes of caution to those in the press who are now comparing the 32 cents creditors “recovered” on Argentina with the higher recovery value in Ecuador and Russia.
- Russia restructured both its “London Club debt” (dollar denominated, Soviet era syndicated loans) and its GKOs (ruble denominated treasury bills). Doing the calculations on the recovery value for the ruble debt is a bit difficult, but it clearly was a lot lower than the 50 or 60 cents on the dollar investors “recovered” in the London Club restructuring. I suspect GKO investors did no better, and may have done a lot worse, than Argentina’s bond holders.
Incidentally, Russia left its “true” eurobonds entirely out of its restructuring, though the distinction between the “London Club” debt and a eurobond is a rather fine one.
- Doing the calculations for the recovery value on Ecuador is extremely difficult, because Ecuador had so much collateralized debt. Lots of the “recovery” came simply from releasing the collateral in Ecuador’s pars and discounts. The holders of Ecuador’s uncollateralized Eurobonds (maturing in 02 and 04) did extremely well, no doubt (they got to swap at par into a new 12 bond that paid a 12% coupon). The holders of the uncollaterized payment streams on the Bradies (A collaterized Brady bond can be broken down into two components, the collateralized principal, and the uncollateralized interest payments, and the stripped spread on the uncollateralized component can then be calculated), in contrast, did not do very well at all. No doubt, Ecuador’s restructuring was more generous than Argentina’s. But about 1/4 (15 cents) of the overall 60 cents recovery value came from the release of the bond’s collateral (Assuming a market value of the restructured bonds + collateral + cash payments of about $4 billion, which works out to about a 60 cents recovery value — I have data on the terms of the new instruments, not on the market valuation of those new instruments.
My point: some claims made by Argentina’s bondholders rely on a selective reading of the history of past bond restructurings. Holders of GKOs did not get 50 or 60 cents on the dollar; holders of Ecuador’s uncollateralized Brady payment streams did not do all that well either (Holders of Russia’s eurobonds and Ecuador’s 02 and 04 bonds, in contrast, did quite well). Ecuador most certainly did not “negotiate” with a bondholders committee; at the time, bond holders complained loudly about the unfairness of Ecuador’s take it or leave it exchange.