Bonds
Bondlife: Lions and Tigers and Convexity Hedgers (Oh My!)
By Tom Graff
Street.com Contributor
12/10/2010 5:09 PM EST
This week brought a violent selloff in bonds, culminating in the terrible 10-year Treasury auction on Wednesday, which priced at a 3.34% yield. We heard all sorts of excuses for why bonds sold off, from foreign selling to bad auctions to bond vigilantes worried about tax cuts. But the real violent selling came at the hands of convexity hedgers.
Understanding who these traders are, why they buy or sell, and why they are known as the worst market-timers going will help you understand how to trade through this period. In this week's Bondlife, I'll show you how to recognize and profit from convexity hedging. Plus, we'll look at why financials are still cheap vs. industrials, and how munis have sold off but still aren't cheap. We'll look at when to buy. Interest Rates: Behind the Curtain
Let's say you run a leveraged mortgage-bond portfolio. You might be a REIT, like Annaly Mortgage (NLY) , a hedge fund, a dealer desk, a bank, or even Fannie Mae or Freddie Mac. Say you want to hedge the interest rate risk of that portfolio. With a corporate bond, this is an easy process: You short-sell a Treasury, an equivalent futures contract or a swap contract. But with a mortgage bond, the proper hedge isn't easy to determine. That's because since the mortgage is prepayable at any time, your actual cash flows are unknown.
We have complex models to try to estimate what the prepayment stream will look like, but even if those models are perfect (which they aren't), you still have the reality that prepayments are interest-rate sensitive. If interest rates fall, more people will refinance, i.e., prepay. If interest rates go back up, fewer people will refinance.
For example, take Fannie Mae 4.5% 30-year mortgage-backed securities. The average borrower within those pools has an underlying mortgage rate of 4.96%. In the last six months, when the 10-year averaged 2.72%, these pools have paid principal at a constant prepayment rate of 19 (or 19CPR -- equal to about a 19% annualized principal payment rate). That would give each dollar invested in the mortgage pool an average life of 4.37 years. In other words, the investor in this pool can expect to have half of his principal repaid to him in the fourth year. If you are hedging the interest rate risk, a simple hedge would be to enter into a four-year swap. More often, mortgage hedgers use 10-year hedges (either swaps or futures) and then just size the hedge to be approximately equal to the expected duration of the security.
Now let's say that interest rates rise a bit. How would this bond pay? Well, let's take the previous six months (i.e., January to June) when the 10-year averaged 3.67%. During that period, these pools repaid principal at an average rate of 6CPR. That equates to an average life of 10.18. Quite a difference. Notice that because four-year bonds naturally yield much less than 10-year bonds, the mortgage will not only lose value because interest rates rise, it will also lose value because of a change in the expected maturity. This is called convexity, and it is the main problem in hedging mortgage-backed securities.
So let's pretend that the 10-year Treasury futures contract has a duration of around 6, and let's say that hedgers use only 10-years to hedge their mortgage book. To hedge the FNMA 4.5% bond with a four-year average life, it would take about a 2-to-3 hedge ratio (i.e., duration of the mortgage-backed security divided by duration of the hedge). So for a $100 million MBS portfolio, the hedger would have a short position in $67 million of futures. Now let's say that 10-year Treasury rates rise from 2.72% to 3.00%, and let's say that mortgage hedgers project that this will result in a prepayment decline from 19CPR to 14CPR. That causes the average life to extend from 4.37 years to 5.76 years. Suddenly the hedge ratio goes to about 1 to 1, requiring another $33 million in hedges. Then the 10-year rises another 20 basis points to 3.20%. Now that projects to a 10CPR, or a 7.48 average life. That's another $25 million in required hedges.
I've made several simplifications here, granted, but that is the big picture of what goes on. And while all that may be interesting, the key is understanding how to profit from this convexity hedging. The first trick is to recognize it. When hedgers are dominating bond trading in a given day, you'll usually see several telltale signs. - Futures contracts underperform cash Treasuries on heavy volume. That is because hedgers tend to use futures.
- Swap spreads widen. Swaps are another MBS hedger favorite, because a swap hedges both the asset price risk and the liability cost side.
- On-the-run Treasuries (i.e., the ones most recently auctioned of a given maturity) underperform off-the-run issues. This is because anyone who is shorting cash bonds almost always uses the on-the-run to do it.
Any time you see these three events going on at once, it's likely that hedging activity is heavy. And of course, we saw all this week, particularly on Tuesday. But to really identify the mortgage hedger, look for a point in which yields either rise or fall to some level not seen in several months, especially where a new bond becomes "creatable." That is, you take the current mortgage borrowing rate and subtract 0.5%, rounding to the nearest 0.5%, then see if that is a different coupon than has been the recently prevailing issue.
For example, up until recently, the 3.5% coupon has been the most creatable, basically meaning that as long as the prevailing borrowing rate was 4.2% or less, 3.5% was going to be the most "creatable" coupon. With borrowing rates recently backing up to the 4.7% area, 4.0% is now the most creatable, and we're getting close to 4.5%. More importantly, borrowers with existing mortgages of 5% went from having a strong refinance incentive to having none. That results in the speed for those borrowers slowing dramatically. Hence, billions in new hedges are needed.
When mortgage hedgers come in, they are not valuation sensitive. In fact, if anything, as rates keep spiking higher, they get more and more desperate to sell. Hopefully, you are now seeing why this week's action was classic convexity hedger trading. Rates spike higher, and hedgers become panic sellers, causing rates to spike even higher, and they panic even more.
How to play it? First, know that mortgage hedgers are in and then they are out. At some point they have sold their hedges, and then that source of selling pressure is gone. Bonds will re-rally at some degree. Second, mortgage hedging tends to be concentrated in the intermediate part of the curve, five to 10 years. So whether you are trying to front-run the hedgers with a short-sell or playing the rebound, don't use a product like the ProShares UltraShort 20 Year-Plus Treasury ETF (TBT) . The hedgers aren't playing on the long end. Again, we saw that this week, as the 30-year Treasury outperformed on a yield basis in the selloff. Treasuries: Are You a Good Auction or a Bad Auctions?
This week was another example of Tom Graff's rule of Treasury auctions. You never get three bad (or good) auctions in a row. So when the three-year was bad and the 10-year really bad, it shouldn't have surprised readers when the 30-year auction went quite well. It also shows that auction results rarely are predictive of anything about future performance of bonds. In fact had you bought that terrible 10-year auction at 3.34%, you'd be sitting on a very nice profit right now, essentially having hit the low tick on bond prices. The way to trade auctions is leading up to them, not using them as an input into how you think about rates. Corporate Bonds: It's a Twister!
I saw a presentation by a hedge fund-of-funds this week where the presenter explained that more and more long-short equity funds are using credit default swaps to hedge their long exposure. Interesting, and not sure it's smart, but it is certainly a phenomenon I've noticed happening. However, I think that in terms of financials, it has created an opportunity for long investors.
If CDS are being used as an equity hedge, then that results in CDS being more correlated with equity prices. Which I have certainly observed. And we know if the CDS is moving wider, the cash bonds almost always will also move wider. However, in the last six months we have seen several events that should have affected financial stock prices much more than bond prices. Most notably, Basel III, which tightens bank capital and benefits bondholders at the expense of stockholders, and the foreclosure problems, which doesn't alter the quality of the banks' balance sheet but will result in increased costs and thus lower profits. The latter is neutral for bondholders, and the former is a real positive. Yet financials have severely underperformed industrials, particularly since June and July. We are adding to financial exposure here. MBS: What Would You Do With a Brain If You Had One?
MBS gave us a stark reminder that there are still not enough bonds. As the bond selloff got rolling on Tuesday, MBS severely underperformed. That makes sense, especially given the convexity hedging math described above. But as banks saw that higher absolute yields were available, they poured in, causing MBS to rapidly snap back and finishing the week with solid outperformance vs. Treasuries. I added to MBS on Wednesday, and it turned out to be a great trade. I'll keep using selloffs to add to this sector. Munis: I'm Melting!
Obama's tax compromise was hailed in most corners, but not in the muni bond market. Word is that the Build America Bonds program was something that Republicans specifically negotiated out of the settlement. It is now widely viewed as less than 50% (and I think less than 25%) that BABs get extended. This will cause much more supply on the long end of the muni curve in 2011. Expect munis to steepen from 10-years to 30-years. In BABs, there was massive tightening in the large, long-term deals. By Thursday, that sector was some 30 basis points tighter on the week, which I something like 4 points of performance. I've also added to BABs this week. Get 'em while you can