Much has been made of the November 15 deadline for many hedge fund clients to submit redemption requests. While we wait to see what the eventual impact of this will be, let’s consider how the changing nature of leverage has altered the bond market. Bear in mind that fixed income arbitrage was among the most popular hedge fund strategies. In addition, it was a core strategy for many dealer prop desks as well as the foundation of the CDO market. The decline of leverage may permanently alter the nature of fixed income spreads.
Fixed income arbitrage is, at its core, fairly simple. Start with a bond that yields x% (over and above some hedge in some cases), assume one can borrow y% of the par amount at a cost of z%. If the math of all that works out to a reasonable IRR on the residual, the arbitrage works.
These arbitrage accounts were the marginal buyers in the fixed income markets. As long as the arbitrage remained attractive, yields (or yield spreads) would remain within a narrow band. When an investor wanted to sell a bond, even if there was no long-term buyer at the ready, arbitragers would step in and provide liquidity. In this way, leveraged buyers were ensuring that the market remained efficient. Spreads would only meaningfully widen when credit risk increased.
We know it went too far. CDO^2 and SIVs were the most obscene examples. But even reasonable arbitrage strategies, like TOBs and CLOs became too large as a group. They stopped just being the marginal buyer and became the whole market.
That marginal buyer is gone, and isn’t likely to come back any time in the foreseeable future. Admittedly, it isn’t as though leverage is being pushed to zero in the fixed income markets, but haircuts (i.e., the amount of margin that must be posted) are now such that levered buyers cannot force efficiency. Take something simple like Fannie Mae 5-year bullet bonds. Should have a very small spread versus Treasuries given the government backing of the GSEs, but instead the spread is currently around 1.45%. It seems like an arbitrage.
But in order for an actual arbitrager to realize a decent IRR on the trade, it probably needs to be leveraged 20x or so. Now maybe one can actually get that amount of leverage versus Agency collateral, but what happens if the trade initially goes against you? The potential margin calls would kill you. Its a difficult arbitrage to actually realize.
The consequences for investors are far reaching. First, bonds will be permanently less liquid. Dealers will be going away from making money via bond arbitrage and go back to making money on transaction flow. In order for that to be viable, the bid/ask spread is going to have to stay much wider than in 2006. Odds are good that trading volumes will also remain much lower than was the case in 2006.
Second, yield spreads will probably remain more volatile than in years past. This is because real money buyers, like mutual funds and banks, will become the marginal buyer of bonds. The technicals of real money demand will suddenly become much more important in determining short-term spread movements.
Third, new debt issues will need to have a greater concession to secondary trading in order to get sold. Take a look at Thursday’s 30-year Treasury auction to see what I mean. Primary dealers aren’t able to keep Treasury auctions orderly in a world where Treasuries in general are in hot demand. The result? The 30-year Treasury priced about 2.5% lower in price than where it was trading the previous day. By Friday morning, it had regained all that it had lost. The same thing will happen to new issue corporate, municipal, and even agency bonds of large size.
In the short term, what should investors consider in bonds? If banks, mutual funds, and other long-term investors are going to be the new marginal buyers, buy what they are going to want. That is high quality, high yielding bonds. This means longer-term or bonds with option risk, such as callable agencies, municipals, and agency mortgage-backed securities.
Finally, if liquidity is going to remain challenging, buy bonds you are comfortable owning for the long-term. If you do buy a corporate bond, assume that it will cost you 3-5% of the bond’s value to sell at short notice.