I’ve been asked by several people what this animal might be, and whether we ordinary folk can buy them.
Since the interview I read with Julian Robertson and FT calls them a kind of put on long Treasuries that isn’t active until five years from now, I’ll try to explain, knowing full well that I am just guessing what the exact terms of their trade might be. He also said his risk was limited to the purchase price, and that it provided very high leverage.
So, starting from those bits of information, and subject to the disclaimer that there are really several ways to achieve these general goals, I’ll take a shot.
My sources tell me that there are hedge funds coming in to the largest dealers to execute two kinds of interest rate trades five years in the future. One behaves like Julian Robertson’s description of the economics, and that is a Cap. The other popular trade is a yield curve swap, which would only conform to his description of downside equal to initial investment if it was a “swaption,”, which is a European style option that allows the holder to convert at maturity into the swap.
Let’s start with Caps, and then go on to the swaptions later.
Caps – in the interest rate swap world – are contracts with a counterparty that pay nothing at all until the reference rate (usually LIBOR) exceeds a certain level. On any determination date (often every three months, but for mortgage bond holders, every month), if the reference interest rate is above the strike rate, then the counterparty pays the purchaser the difference between the market interest rate and the previously agreed rate (the strike rate), with the payment multiplied by the notional face amount of the contract.
I can almost see eyes rolling back into heads, and this is just the beginning, so let’s use some example numbers to get a feel for how it works.
Let’s say I own a 10-year, $10 million floating rate bond that pays me L+50 (LIBOR plus half a percent, or 50 basis points) every three months, but that bond’s rate is capped at 8%. My problem is what to do if LIBOR goes above 7.5% (not very likely right now or even for the next few years).
I’ll start out assuming that the bond is worth par, or $10 million. In other words, the credit risk of the issuer of the bond is judged to be worth 50 basis points per year.
I might want to buy insurance that covers me for rate risk, so that I get L+50 no matter how high LIBOR goes. Since 3-month LIBOR is now under a half a percent, and since 10-year the swap rate (proxy for LIBOR farther out the yield curve) is under 4%, I might only have to pay a dozen basis points or so to buy that insurance policy. In other words, I might pay around $12,000 to protect the future interest rates on my capped floating rate bond.
So how big is the payoff on that, if things get really bad? Let’s say that two years from now Fed Funds are at 8% and three month LIBOR is at 8.25%. Worse yet, let’s assume the market has a very positively sloped yield curve, driven by high inflation expectations. At that point, if I hadn’t bought the interest rate cap as insurance, my bond, with eight years left in it, would be paying only 8% in an environment where short rates are higher than 8%, and long rates might be 10%, 12% or even higher.
Doing a back-of-the-envelope estimate on the value of an eight-year 8% bond in a 10% rate environment would make that bond worth 12% to 15% less than its face amount, for a mark-to-market loss of $1.2 million to $1.5 million.
For this example, by combining the bond with the interest rate hedge (the cap), I still have a combination that pays me a steady L+50, which would still be worth par, as long as the credit of the bond issuer and the cap provider have not declined and credit spreads are stable.
Using the estimate of 12 basis points as the original cost of the hedge, in this example, there would be effective leverage of 100-1, or even as high as 125-1.
Still with me?
OK. I don’t think Julian Robertson or David Einhorn are particularly worried about the next two years, and even if they are, buying the insurance so it starts five years from now instead of today might save a few pennies. Of course, if you own 10 or 20 years of meaningful interest rate protection two years from now that doesn’t start for a few years at that point, that insurance policy itself will be worth huge bucks if you wanted to sell it, assuming the prior scenario actually happened that quickly.
But let’s parse what Robertson said a little further.
It was clear that he wasn’t insuring against 3-month LIBOR, or any other short maturity rate. He was insuring against long rates going wacky, probably the 10-year Treasury Note rate or even the 30-year rate.
The 30-year swap rate is probably the most interesting rate to defend against out in the long end of the curve, because technical supply-demand issues have pushed that rate lower than the benchmark 30-year Treasury rate.
This is fundamentally nonsensical, since that spread relationship implies that a bank paying US dollars 30 years from now is a “better” credit risk than the US Treasury paying US dollars in the same time frame.
I may not think much of our last 30 years of fiscal management, but I am quite sure the US Treasury will still have access to a printing press 30 years from now, and even if those dollars are worthless that they might pay me, other dollars from some bank have to be even more worthless because they have the bank’s credit risk.
As an aside, this peculiar relationship between swap rates and Treasury rates has persisted for nearly a year. Maybe it’s a result of lack of issuance of 30-year Treasury Bonds (the Long Bond) vs. the liquidity of the swap market. Maybe it’s demand for high-quality long-dated payments by insurance companies and pension funds that can’t trust all the old standards like real estate, equities or top-rated corporate bonds. Who knows?
What we can be sure of is that, at some point in the future, if you own an obligation in 2039 to be paid US dollars, that obligation will be worth more from the Treasury than it would be worth from some bank.
However, returning to the problem of what Julian and David are buying, my best guess is a “five-into-ten” or “five-into-twenty” interest rate cap on the 30 year swap rate, also known as CMS30, or the Constant Maturity Swap 30-year rate.
Translating, they could be buying an insurance policy that will not be effective until fall of 2014, at which point it will run for 10 or 20 years, and periodically pay the positive difference (if any) between the 30 year swap rate and the cap strike rate just prior (two days) to each payment date.
Translating again, into numbers:
If they own $1 billion of these caps struck at 8%, and paid 2 points ($20 million) for them, a rate environment of 15% for long swap rates five years from now would turn into an asset that would pay them $70 million a year for ten or twenty years.
A quick call to a friend that trades these things (in billion-dollar increments) says that such caps would cost around 210 basis points today. Even more interesting is the cost of such a cap expressed as basis points in ongoing payments. Spread out over time, the theoretical cost is only 17 basis points per annum.
We talked about the language Robertson used in the interview, and since the market is also seeing demand in the yield curve swap, we looked at that. The most common form is the “2 year – 10 year” yield curve swap, since that captures the “belly” of the curve, where most debt is issued.
That swap is a contract where one party pays the other the difference between the two rates, or it could be structured as a contract that only pays when the difference moves away from today’s levels, with Party A paying Party B if the difference gets larger (yield curve gets steeper), and vice versa if the yield curve gets flatter (difference between 2 year rate and 10 year rate shrinks).
Five years out, the 2-10 swap gets much flatter than today’s curve, or “spot” rate. Today, the market was showing a USD 2-10 spot rate of 225 Basis points, a pretty large difference. We call this yield curve steep, especially when you consider the absolute levels of the two rates.
The 2-10 swap drops to 144 BP’s a year out, 87 BP’s two years from now, and only 23 BP’s five years from now, which means the market expects a lot of flattening over the next five years.
If you’re bearish on the economy and on the Dollar, you might conclude that the Fed will keep short maturity rates low for a long time, while international investors will demand high rates for longer maturities. That sounds to me like the logic macro hedgies might follow, in which case they want to buy something that performs well if the yield curve gets even steeper, or at least doesn’t flatten.
I should point out that a yield curve steepener or flattener is not inherently bullish or bearish. In fact, there are bull markets in which the curve flattens or steepens, and there a bear markets where both flattening and steepening occurred.
Having taken up all the time my buddy could spare, I decided to stop there, and not ask for the price on a five-year forward, out-of-the-money yield curve steepener structured as “one time look” (European) swaption.
The truth is likely that they are putting on both kinds of trades, the pure rate protection of a cap, plus the steepener, probably done as an swaption so ongoing payments are not required.
So, when Julian Robertson said “curve cap”, he was probably talking about two trades, purchase of a 5 yr/30yr swaption with an 8% strike that gives him the right, but not the obligation, to pay fixed at 8% and receive floating 30 year CMS, together with purchase of a yield curve steepener swaption that is in the money if the 10-year swap rate is something like 200 or 300 basis points higher than the 2-year swap rate five years from now.
If you’re not confused yet, we can start talking about who is long and who is short when deals like these are struck.