The Economics of Negative Bond-CDS Basis Trades

The bond-CDS basis is the difference between the cost of a credit default swap and the yield on a corresponding bond. When you own a bond, you receive its yield (in the form of both coupon payments and price convergence to par) but also take the risk of the issuer defaulting. Buying CDS protection eliminates that risk, at a cost that is typically a percentage of the notional amount you’re insuring, paid quarterly. Until recently, the basis was typically positive, i.e. buying insurance was more expensive than the yield you were getting from the bond. Many more bonds now have a negative basis, implying that indeed you can earn a yield from the bond, use some of it to remove all issuer-linked credit risk, and still make money. (Buying CDS protection leaves you exposed to the risk that your CDS counterparty blows up. You also have interest-rate risk, which is discussed later.) You can also potentially make money on the mark-to-market of both the bond and the CDS.

Let’s say you’re a hedge fund and buy $10MM of the Eastman Kodak bond maturing on Nov 15, 2013, paying a 7.25% coupon and currently trading at 63 cents on the dollar. (Note: this work was done on June 15, 2009. I haven't updated any numbers.) You finance half the purchase by borrowing from your prime broker at 1-month Libor + 150bp, but pay the other half ($3.15MM) cash. You’re getting $725k annually in coupons, and pay $57k in interests.

Now you’re buying CDS protection for a Spread-DV01-neutral notional amount, which in this case, as we’ll see later, is $4.799MM. As of today (Again: June 15, 2009), the cost of that protection is 5% of insured notional i.e. $240k annually (but paid in four quarterly installments of $60k). There is also an upfront payment of 28.04%; the capital needed up-front is thus $1.35MM, but you should be able to get 10x leverage from your PB and thus only post 10%, or $135k, as margin.

You also want to keep some cash reserve, let’s say $1MM in this case, so the capital used in this trade is $3.15MM for the bond, plus $0.135MM for the CDS margin, plus a $1MM reserve, totaling $4.285MM. The carry is $428k, annually, since you’re receiving $725k in coupons, paying $57k in financing interests and paying $240k in CDS protection.

The terms and price of the CDS can be read off Bloomberg using their CDSW page. The following screen shot shows the 5-year CDS on senior debt, where the bottom line of the “Deal Information” pane indicates that the upfront payment is 28.04% and the running spread is 500bp. (These are of course the numbers used earlier.)



This bond-CDS basis trade can be structured, in part, using Bloomberg. One way is, from the CDSW page shown above, to click on the “Bond Hedge” red button in the top banner, then on “Bond vs CDS hedge.” This leads us to the HGBD page, which can also be accessed directly of course. The page calculates the notional amount on the CDS and the notional on the interest-rate swap (IRS) that will hedge the rate exposure on the bond-CDS package.



In the top panel, we recognize the $4.799MM notional amount used earlier for the CDS. The middle panel confirms that the exposure on credit default (Spread DV01, i.e. the change in Dollar Value of the instrument for an upward, parallel 1bp move of the credit curve) is 100% hedged, and that sensitivity to rates (IR DV01, or dollar value change for a 1bp upward parallel shift in interest rates) is hedged to zero. In the bottom panel, the price and Z-spread of the bond are summarized. (The asset-swap spread, not the z-spread, is used in practice.) The cost of CDS protection is also shown, assuming no up-front payment and assuming a 40% recovery rate -- of course, different scenarios should be tried out to project the impact of actual recovery rates on the economics of the trade. Converting the CDS cost to an all-running format, as opposed to the actual upfront+running payments, helps compare the protection cost to the bond yield. Here, the basis is 1,532bp for the CDS, less a z-spread of 1,774bp for the bond, giving a basis of negative 242bp.

So, the negative basis trade has a positive carry even net of interest rate risk. But what if the issuer defaults on its debt? Assume the firm defaults 1 day before the next CDS payment is due, that is 89 days after the basis package is put on. Your initial outlay was as computed earlier, $4.26MM. You made no subsequent CDS payment and received no coupon, so the only additional cash inflows are the recovered amount on the bond, and the CDS payment. Assuming again a 40% recovery rate, you recover 40% times the bond’s face value, i.e. $4MM, and the seller of CDS protection pays you (100% - 40%) times the CDS notional of $4.799MM, i.e. $2.88MM. You also free up your $1MM reserve and thus end up with $7.88MM just 89 days after laying out $4.26MM.

Hey, wait! These calculations do not include mark-to-market gains or losses on the bond, the credit default swap, or the interest swap. It’s hard to predict what their values will be in the future, but the “horizon” tab of the HGBD page projects these MTM changes based on the yield curve and on the credit curve of the issuer. On an 89-day horizon, the MTMs are shown in the bottom right corner of the screen shot below.



So, if all curves stay where they are, we would after 89 days be making $67k on the bond and $35k on the CDS, but losing almost $17k on the interest rate swap -- resulting in a net gain of $86k on top of the $7.88MM profit calculated earlier!

Let us now assume the issuer defaults after a year. In this scenario, you had the same outlay and the same terminal payment, but the 4 quarterly cash inflows would be $107k, equal to the annual $428k carry divided by 4. (Bond coupons are usually semi-annual, so the actual cash-flows would be slightly different.) Turning now back to Bloomberg to get projected P&L on the bond and the swaps, we see the following:



This example trade looks so juicy that we may think hedge funds are piling on all negative basis trades. (They were!) However, a quick look at the bond universe seems to show that, although many bonds currently sport negative basis, relatively few are expected to post mark-to-market profits according to the horizon analysis done above. Also, as demonstrated above, returns are higher if the issuer does default and defaults relatively early after the trade is put on, further narrowing the opportunity set.

You can try for yourself to find negative basis packages and look for the “best” opportunities for early defaults using the SRCH page on Bloomberg: go to Advanced Search and select a country (selecting a country is not mandatory but significantly trims down search time), then go to the “inventory” section and select a basis range (e.g., from -200 to -100bp) in the appropriate fields at the bottom right of the page. Among the numerous answers, you will want to do the horizon analysis to make sure the positive carry on the trade won’t be killed by MTM changes.