I have long worried that as the carry trade gets less profitable (i.e. the difference between the interest rate on short-term and long-term debt falls), hedge funds would respond by doubling up their bets. From what Jim Melcher of Balestra Capital says, that seems to indeed be what has happened.
Mr. Melcher is talking his book, but his message still strikes me as important. From the FT:
Jim Melcher has anticipated nearly every market meltdown of the past 25 years and profited from them, including the stock market crashes of 1987 and 2000, the bond woes of 1994 and the emerging market crisis of 1998. ... While a market disruption may not come to pass, if it does, "it could be a very nasty scene", he says.
The manager’s bet revolves around the remarkable resiliency and popularity among hedge funds of the "carry trade" - the practice of borrowing money at low interest rates and investing it in higher-yielding issues such as junk bonds, emerging market debt and mortgage-backed securities. Hedge funds and other investors have made vast sums over the past two to three years on the spread between short-term US Treasuries and the longer-dated issues.
The spreads have narrowed considerably in the past year, making the trade less profitable. Rather than unwind their positions, it is widely believed that many hedge funds have simply used leverage, which sustains the outsize profits on a diminishing trade but heightens downside risks considerably. ... . "The carry trade has been a way for hedge funds and others to make easy money," Mr Melcher says. "But it’s become overcrowded, and any overcrowded position is dangerous when it unwinds." ... "I am worried about the fairly broad systemic effects here," he says.
The preponderance of hedge funds in the carry trade is partly what unnerves Mr Melcher. "Hedge fund guys have a very quick trigger finger. When things start going against them, they get out," he notes. If there’s any sort of rush to the exit, "there is no easy way out. And because our financial system is interconnected as it never has been before, everybody is going to get hurt".
DeLong’s hard landing scenario hinges, in part, on the assumption that hedge funds will bet wrong on dollar, causing financial distress. I would worry a bit more that hedge funds will get caught betting the wrong way in the fixed income market, whether betting to heavily on curve flattening (a falling spread between short-term and long-term treasuries) or spread compression (a falling spread between risky and risk-free assets; if you borrow short to buy long-term corporate debt, you are hoping to pick up the interest difference, and hoping that the price of the long-term bond rises/ the interest rate on the bond falls).
It seems that Randy Quarles -- the future Under Secretary of Domestic Finance at the Treasury -- is considering reintroducing the 30 year bond. I am all for it. And despite the denials, the prospective financing needs associated with partial privatization of Social Security probably has something to do with the idea. After all, the cash flow savings from "Social Security reform" that offset the upfront costs are VERY long term (and I would say, very uncertain). It is the sort of thing that should be financed with long-term debt. Moreover, even in the absence of partial privatization, with deficits as far as the eye can see (using realistic assumptions) the Treasury probably does need to reconsider its heavy reliance on two and three year notes.
Remember, lots of two and three year notes issued to finance the 2003 and 2004 deficits will start coming due soon (for data on the surge in two and three note issuance, see pages 7, 16 and 19 of this document), and they will need to be refinanced even as the Treasury is seeking to raise new money to cover the 2005 and 2006 and no doubt 2007 deficits.
That said, there is a big difference between replacing $20-30 billion of issuance with two-year notes with thirty-year bonds, and replacing $20-30 billion of ten-year notes with thirty-year bonds. We don’t know if the Treasury is reconsidering its overall plan (see p. 13) to shorten the maturity of the US debt stock. Lengthening the maturity of the debt -- or just no longer shortening it -- does not require resuming issuance of thirty year bonds. It would require issuing more ten-year notes, and fewer bills and short-term notes.
What is the link between the Treasury’s issuance plan and hedge funds making ever-more leveraged bets on the carry trade?
It is tenuous, but there potentially is one.
The yield on long-term bonds has pushed down by structural demand from pension funds looking to do a better job of matching their assets to their liabilities, no doubt. But given the volumes involved, my hunch is that two other factors are even more important: demand from foreign central banks across the entire Treasury curve, and the Treasury’s decision to radically shorten the maturity of its debt stock. The net result has been to squeeze the supply of ten-year notes in private hands, and a slow reduction in the available supply of long bonds.
But even as the Treasury is (rightly) reconsidering the risks that arise from the combination of a rising debt stock and a shorter-maturity on that debt stock, demand from central banks for Treasuries may be shrinking somewhat, albeit from an extremely high level. Central banks right now are all talking about diversifying the composition of their dollar reserves, and put less of their still growing reserves into the Treasury market.
That supports the case for resuming issuance of the thirty-year bond. It might bring new demand into the Treasury market even as central bank demand fades. Since central banks generally prefer shorter maturities, a shift in the Treasury’s issuance pattern would help to match the bonds the Treasury issues with the bonds the market most wants.
But it is at least possible that the Treasury could end up increasing the supply of long-term debt in market just as demand for those Treasuries starts to fade a bit. Who knows, that might even put pressure on rates and some hedge fund might get caught with too large a leveraged bet ... (Note the complaining at the end of this New York Times article). Is it likely a surge in long-term Treasury issuance will change the dynamics of the market? My guess: No, not on its own. But it might interact with something else in the market in an unpredictable way.
My bottom line: the Treasury should have started issuing more long-term debt - be it ten year notes or longer term bonds -- back in 2003, even if there was a cost advantage to issuing more short-term debt at the time. The Treasury does need to take steps to reduce its own refinancing risk. But it is at least possible that it may not be able to do so in 2006 without putting more pressure on the market than it expects ...
UPDATE. For those with an interest in the Treasury market, this presentation is pure gold.