Blackstone May Do Its Cleverest CDS Trade Again


(Bloomberg View) -- One of the great credit derivatives trades was when Blackstone Group LP's GSO Capital Partners credit fund bought credit default swaps on distressed Spanish gaming company Codere SA, offered Codere some much-needed refinancing of its debt, but demanded in exchange that Codere make an interest payment on its bonds a few days late. By doing that, Codere was -- briefly -- in default, and so GSO's credit default swaps were triggered and it could make a profit on them. It is the only credit derivatives trade that I can think of that got its own segment on Jon Stewart's "Daily Show," which makes it presumptively the funniest credit derivatives trade ever, outside of the financial crisis. And now GSO is revisiting its greatest hits:

A battle is raging among hedge funds, with one group saying that the other has offered the company, Hovnanian Enterprises Inc., financing in return for taking steps that would trigger payouts on those derivatives.

The claim came in a letter from law firm White & Case, which said it’s been made aware of a proposal in which Hovnanian would pursue a refinancing deal with the main intention of triggering a credit event that would lead to a payout on the credit-default swaps, according to people with knowledge of the matter.

GSO is the group proposing the trigger-and-refinance deal, just as it was in Codere, while White & Case is "representing a group of funds led by Solus Alternative Asset Management, which both owns Hovnanian’s bonds and sold credit-default swaps that guarantee the company won’t miss a debt payment." And as in Codere, the default could be a silly little thing: "Triggering such a default could be accomplished by delaying interest payment temporarily, or keeping some of Hovnanian’s small holding of 2017 notes outstanding beyond their maturity date." And as in Codere, Hovnanian could use the money: It has about $54 million of bonds coming due on Dec. 1, and "has been facing pressure amid declining sales and a deterioration in its credit metrics that have pushed ratings firms to cut its grade to the lowest rungs of junk."

In generic terms these trades are lovely. You have a CDS contract that pays out when the company defaults. GSO owns the contract, which becomes more valuable as the company gets closer to default. GSO has found a way to crystallize that value -- by making the company default -- and then give some of it back to the company in the form of an attractive financing proposal. The company can, through the magic of derivatives, turn a bad thing (default) into a good thing (cheap credit). Scolds like White & Case and Jon Stewart think it is shady and goes against the spirit of fair play, which I suppose is true, but if you're so interested in the spirit of fair play why are you trading credit derivatives? For myself, I can't see these trades as anything other than cute and fun.

But there is a practical problem. CDS contracts are not, generally, binary bets. A $100 CDS contract doesn't just pay you $100 if the company is two days late on an interest payment, and zero if it is on time. Instead, if the company defaults, an auction is held for its debt, and some price is determined. If the defaulted debt turns out to be worth 60 cents on the dollar in the auction, then the $100 CDS contract pays out $40. If the defaulted debt is worth 95 in the auction, the CDS pays out $5. The CDS has a quasi-insurance function; it pays out based on the amount you lost on the debt (if you had bought it at par). I said about Codere:

The reason that Blackstone made money on its credit-default swaps is not just that they were triggered by this clever maneuver. Just triggering CDS is not a big deal, because CDS pay out based on the difference between the face value of a bond and its post-default trading value. ...

Instead, Codere's bonds were worth 54.5 cents on the dollar in the CDS auction, so Blackstone got paid 45.5 points on its CDS, which is a lot. ... It got that money not just because of the missed interest payment, but also because Codere credit, after the technical default and Blackstone's loan extension to Codere, remained really risky. Codere's other bonds still trade in the mid- to high 50s.

Hovnanian is in a rather different situation. "The company’s stock and bonds show no hint of panic," notes Bloomberg News, and in fact almost all of Hovnanian's bonds and loans seem to trade right around par -- between 92 and 108 cents on the dollar. That's well above "no signs of panic" levels and right into "everything is great." (I mean, relatively speaking: The bonds have high coupons and low-junk ratings, and Hovnanian's most recent bond offering came at double-digit interest rates. ) But the key point is that if the bonds are basically worth 100 cents on the dollar, there's not much room to make money by triggering a quickie fake default. Meanwhile Hovnanian's six-month credit default swaps trade at at nosebleed levels, requiring a payment of something like $20 for $100 worth of protection for six months.

So how can GSO make money by triggering a default?

Well, here is one answer. Almost all of Hovnanian's bonds trade near par. But it does have a small ($53.2 million outstanding) 2 percent bond due in 2021. That bond's coupon is way below-market; by comparison, Hovnanian recently issued a bond due in 2022 with a 10 percent coupon. So the 2 percent bond trades at about 78 cents on the dollar. And the way the CDS auction works is usually that the cheapest-to-deliver bond sets the price.  So if GSO can trigger a default, it will get paid -- say -- 22 cents on the dollar on its CDS, making it potentially a nice little profit.

In this story, GSO's Hovnanian trade is essentially a cheapest-to-deliver play. This makes it in a way more interesting than Codere: In Codere, GSO triggered a trivial default, but made a lot of money on that default because Codere's bonds really were in a pretty perilous state. In Hovnanian -- look, Hovnanian is not a great credit or anything -- but GSO won't make money because Hovnanian's bonds trade at distressed levels. They mostly trade above par. Instead, GSO will make money by noticing and taking advantage of the cheapest-to-deliver mechanics of the CDS settlement auction.

Which is fine! If you are gaming the CDS market anyway, you should certainly be smart about the cheapest-to-deliver mechanism.  It is Codere with an added level of cleverness, and I like it.

But here is another possible answer that is madness. Most of Hovnanian's bonds trade near par, except that one little weird one. But what if ... what if you just built a new bond? GSO, after all, "has offered to provide the homebuilder with relatively cheaper financing on terms that would fit the definition of a credit event in credit-swaps contracts." That means that it will lend Hovnanian money at below-market rates. What if the rates were way below market? Sridhar Natarajan, Katherine Burton and Lisa Abramowicz report:

One theory among market participants is that at least a portion of the new debt would be structured with a low coupon, which would cause it to trade at a low price in the secondary market. Such a move would allow the holders of the CDS to deliver that discounted security in a settlement auction. The lower the price is set in that auction, the higher the payout for investors that bought the swaps.

What if GSO bought a package of bonds from Hovnanian that included a 10-year bond with a 1 percent coupon? That bond would be worth something stupid, call it 45 cents on the dollar.  If you issued those bonds, and then did the quickie default, then ... could you deliver those new bonds into the CDS auction?  What would stop you?  And if you could do that ... wouldn't your CDS pay out at 55 cents on the dollar?

Of course GSO wouldn't be especially jazzed to buy those bonds, at least not at par (it'd be fine to buy them at 45 cents!), but the point is that it needs to put together a refinancing package that combines:

  1. Being generous enough overall to convince Hovnanian to accept it;
  2. Not being so generous overall that GSO will lose too much money on it; and
  3. Having at least one bond that is so stupidly priced that GSO can make a ton of money on its CDS.

If step 3 is lucrative enough, then step 2 is less important -- the more GSO makes on its CDS, the more it can afford to give away on the refinancing -- and so step 1 is easier. 

To take entirely imaginary numbers for illustrative purposes, if GSO owns $100 million of six-month CDS on Hovnanian, and paid $10 million for it, and can engineer an $55 million payoff on it, then it will have $45 million of profit. It can keep half of that for itself and give half to Hovnanian, in the form of below-market interest on a new financing.

To be clear: All of these numbers are totally hypothetical and exaggerated for ease of exposition and also comic effect. No one is really going to do a 10-year 1-percent bond. But the point is that if you are building your own new bonds for the purpose of delivering into CDS, you do have a lot of flexibility.

If this is the story then it is Codere with an added level of gorgeous insane cleverness and I love it.

Should Hovnanian love it? I don't know. Sure it would get some cheap financing from GSO, but at the cost of defaulting on its debt, at least for a few days. That could bring immediate headaches (a stock drop, creditors demanding repayment), but also longer-term ones: If you default voluntarily in order to play games with your creditors, that doesn't exactly make you a more attractive credit risk in the future. Plus, come on, if you default to play games like this, someone is going to sue you. I'm not sure that if I were a Hovnanian board member I'd want to participate in this GSO art project just for its undeniable aesthetic appeal.

And of course White & Case, and its clients and friends who have sold credit-default swaps on Hovnanian and don't want to pay out for gamesmanship like this, do not like it. My view on the matter is that if you participate in the CDS market and your counterparties do something as aesthetically pleasing as this, you should just treat your losses as a modestly expensive form of art patronage. Someone just spent $450 million on a da Vinci painting that might even be fake. This deal could be much prettier, and you know you're getting a genuine GSO original.

This column does not necessarily reflect the opinion of the editorial board or Bloomberg LP and its owners.

Matt Levine is a Bloomberg View columnist. He was an editor of Dealbreaker, an investment banker at Goldman Sachs, a mergers and acquisitions lawyer at Wachtell, Lipton, Rosen & Katz and a clerk for the U.S. Court of Appeals for the Third Circuit.

  1. Peter Grauer, chairman of Bloomberg LP, is a non-executive director at the Blackstone Group LP, GSO's parent.

  2. Caa2/CCC+ at Moody's and Standard & Poor's.

  3. In July it priced $440 million of 10 percent five-year notes and $400 million of 10.5 percent seven-year notes. 

  4. Bloomberg reports a spread for six-month CDS of about as of Thursday afternoon. Bloomberg's CDSW page translates that into about points upfront, plus a percent annual coupon, so I am just rounding to points. Meanwhile five-year swaps are quoted at like a spread points upfront). And six-month swaps were trading at about a spread points upfront) just two weeks ago.

  5. Well, it's more complicated than that, but that is broadly speaking what you'd expect in the ordinary case. Here is an article about an early CDS settlement auction, and here is a helpful Credit Suisse research note on CDS auction mechanics:

    Since the protection buyer has the right to deliver any of a specified list of (Deliverable) Obligations, it is in the buyer’s interest to deliver the cheapest; this is therefore what the dealers bidding in the first part of the auction and any protection sellers settling their CDS physically risk being delivered. The value of this cheapest to deliver will therefore drive the auction recovery price. It will do so to a greater or lesser extent depending on the size of its (freely available) notional outstanding relative to the size of the Net Open Interest in the auction.

    Determining the cheapest to deliver Obligation is thus an important aspect of valuing CDS, and its identity and importance can differ significantly depending on the circumstances leading up to the Credit Event and the type of Credit Event. We are back to the question of Obligations and Qualifying Guarantees discussed earlier. The issue can be most pertinent in a Restructuring Credit Event. In a Bankruptcy or leading up to a Failure to Pay, all Obligations should trade down to their recovery value, which should be fairly uniform. In a Restructuring, however, depending on which and how many Obligations are Restructured, there exists the real possibility that Restructured and non-Restructured Obligations trade at very different levels, both leading up to the auction and afterwards. A good example would be if there were a very steep yield curve (due to credit and/or interest rate risk): a long bond might be deeply discounted, substantially lowering the effective recovery from that implied by a shorter, Restructured Obligation.

    That said, the auction itself can create weird dynamics in which the resulting price is not really based on the standalone market value of the cheapest deliverable bond:

    At its most basic, the key is understanding who the buyers/sellers of the Obligations are and who is exposed to the auction setting high/low. If the Reference Entity was in a lot of indices and/or CDOs, there is likely to be a much larger CDS notional outstanding than otherwise. If lots of CDS contracts are held on basis, then many auction participants are likely to be selling bonds or loans into the auction, leading to a NOI to sell without necessarily any obvious buyers. If correlation desks were active in the name, there may be some auction participants with large net long positions (protection sellers), looking for as high a recovery as possible. ...

    The Thomson auction on October provides a good illustration. In the first stage, the 2.5Y bucket IMM set at with a Net Open Interest to sell million. The Cap Amount was so the possible range for the Final Price in the second stage of the auction was to for this bucket. Since Thomson had been in lots of iTraxx indices, the notionals outstanding were very large. One market participant clearly had net sold a significant amount of protection because it bid the maximum, for the entire NOI, above the value of the Obligations. The loss that market participant made on paying too much for the Obligations was presumably offset by the gain made on its CDS position.

    One could imagine that happening here: A protection seller might overbid in the auction, figuring that paying too much for the small amount of percent bonds available is worth it to save money on the CDS payout.

  6. I mean, a very little profit if it actually paid for of six-month CDS. (Or a loss if it paid way more than that for longer-dated CDS.) But if GSO bought a couple of weeks ago when the short-dated CDS was trading in the area or lower it would more than double its money with this (hypothetical) trade.

    It's worth mentioning that "despite buying up front-end protection, GSO still has a net long position on the company’s debt." In the text we talk about this CDS trade as a pure trade, because it is prettier that way, but in reality GSO is just an investor in Hovnanian's credit that happens to be optimizing its investment with derivatives.

  7. But it does, at least hypothetically, take us a bit beyond the story of Codere. The Codere trade worked because Codere really was distressed, so there was a certain rough justice in GSO getting paid off on its CDS contracts to lend more to Codere. The Hovnanian trade works because it has one off-market bond. Hovnanian is also quite high-yield, so the same rough justice sort of applies. But the cheapest-to-deliver version of this trade could work on companies that are  not  distressed, but that have one off-market bond: A quickie fake default there could generate profits for CDS investors without them having to take much "real" credit risk.

  8. That's just Bloomberg's bond calculator (page BC7) for a bond with a percent coupon, 10-year maturity and 9.5 percent discount rate. (Why 9.5 percent? Why not? In July, Hovnanian issued seven-year notes at a percent coupon, which now trade at about an percent yield.)

  9. This is reminiscent of the Greece CDS auction where Greece planned to issue new bonds worth far below par and then have them be deliverable into the auction.

  10. The International Swaps and Derivatives Association Credit Derivatives Definitions  define "Deliverable Obligations" for CDS based on general characteristics like seniority, currency, bond versus loan, etc. So you just have to make the obligation fit into the right categories.

  11. There was about million of net notional CDS outstanding on Hovnanian billion gross) as of the week ended Oct.

To contact the author of this story: Matt Levine at

For more columns from Bloomberg View, visit

©2017 Bloomberg L.P.

BQ Install

Bloomberg Quint

Add BloombergQuint App to Home screen.