Primus Guaranty: At the Center of the CDS Storm

The number of high-profile defaults of late from Lehman and Washington Mutual, combined with the technical defaults of Fannie (FNM) and Freddie (FRE) and a ratings downgrade of AIG that pushed the company to the brink of insolvency due to collateral calls have all thrust credit default swaps – CDS, for short – past subprime mortgage securitizations and into the center spotlight of the financial mess. I've had a front row seat for most of this, thanks to my investment in both the common stock of Primus Guaranty (PRS) and the senior debt (PRD).

A recap of the last few weeks:

Primus's CDS-writing subsidiary lost its Aaa-rating from Moody's, and the holding company debt was cut a full notch from Baa to Ba. The former is mostly a non-event because Primus has essentially written no new business; the latter shouldn't be a surprise given how wide the spread on Primus's debt had become.

The credit events have started to roll in. What begin with a technical default on $215 million in notional swaps on Fannie and Freddie that resulted in no losses thanks to the government backing turned ugly when $80 million in notional swaps on Lehman were triggered as that firm went under. The recovery value on Lehman debt appears to be in the high single digits, putting this at a $70+ million loss.

Washington Mutual was another credit event ($16 million notional), although the settlement auction set a value of 57 cents on the dollar, making the net loss estimate about $7 million. The most, um, interesting of the CDS reference entities Primus had was Kaupthing Bank ($68 million notional), an Icelandic financial institution whose nationalization is only part of the enormous collateral damage dealt to that country. Primus provided no estimate of recovery value for Kaupthing, although I've seen estimates of 20-30%, implying a net loss of $51 million by taking the mid-point of that range.

The last event was a technical trigger caused by a downgrade of part of the (proportionally small) CDS on ABS portfolio, with a notional value of $10 million. Primus said they don't expect any immediate redemptions of those swaps.

In total – excluding Fannie/Freddie – that amounts to notional defaults of $174 million, with a rough loss estimate of $133 million. The actual loss number is the important one to watch, and I've been using a quarterly loss provision of just over $65 million in my run-off analysis of Primus' swap portfolio that began with Q2 of this year. Since there were no real charges taken in Q2, rolling the unused provision forward and combining it with the Q3 loss provisions (which I consider fair, because losses tend to be lumpy and the book is closed) equals $135.7 million. The last thing I'm trying to claim here is some degree of prescience in forecasting credit default swap losses; my real point is that Primus is fundamentally more or less where my model estimated it would be, but the stock (45 cents/share) is trading at about 15% of the equity value implied in that scenario ($3.20).

I do make a simplifying assumption or two, like assuming that cash expenses (mainly employee salaries) and capital structure remain constant, but otherwise the difference between those two figures represents how much higher the market thinks forward looking credit default rates will be. Right now, the markets think investment grade credit defaults are going to exceed anything we've ever seen historically. This is a possibility, and it's the danger of essentially being short tail risk – although if you're a stockholder, there's a moral hazard trade here because your losses are limited to what you put in, while Primus's counterparties bear the full risk in an extreme default loss scenario.

Looking forward, my model assumes high defaults for corporate debt through H1:2010, with Primus' losses averaging $60 million per quarter during that time. Increasing those loss estimates to match what the market is pricing requires upping quarterly loss provisions to average $75 million. Again, anything can happen, but focus on what's priced in and opportunity might present itself.

From an outsider's perspective, I see three apparent options for Primus: accept a run-off of the swap portfolio and look to extract the maximum value from the remaining term of business, seek to establish a business with recurring revenues (i.e. structured asset management) to replace writing swaps, or try to resurrect the CDS business.

Working backward, it seems unlikely that the Credit Derivatives Products Company (CDPC) model is going to have rating agency or market support going forward. CDPCs have a huge economic advantage in that they don't need to post collateral if CDS spreads widen, which is great for the CDPC but not for their counterparties. One of the reasons for how smooth CDS settlements have progressed so far is that constant collateral posting and netting on the part of normal swap dealers helps prevent the losses from an undercapitalized party from becoming extreme. While I'm not suggesting that Primus is undercapitalized, I do believe that forward-looking risk management of credit default swaps (including calls for a clearinghouse) effectively renders the CDPC model to a relic of the halcyon days of the "Great Moderation." Perhaps demand will emerge for CDS products like bespoke tranche protection where a buyer would not require collateral if the seller was operating under lower leverage allowances, but that remains to be seen, and I'll continue assuming no additional business will be written until indications prove otherwise.

The next choice – growing an asset management business specializing in structured products – is not given much credence from people I've talked to, but Primus has mentioned it in the past as a possibility. The company's inability to add organic growth to its existing business is probably one reason for that, but it is also facing an extremely difficult environment for structured finance. While this might be a very solid opportunity for creating an ongoing stream of profitable business, I think the prudent course is to be wary of investing Primus' limited capital here unless the transactions will be quickly accretive.

The final choice, and the one I believe Primus's management is warming to, is trying to manage its capital structure in a way that maximizes value. The announced buyback of $10 million of the holding company debt was the first step, and since the debt still trades at a steep discount to par, a tender slightly above the market price once the repurchase funds are exhausted should be considered as well. Should the stock fail to react positively to a substantial buyback of the debt, there will be lots of flexibility to use funds at the holding company level to repurchase shares with the company's market cap being so low, but that's something I plan on exploring more after the earnings release on Wednesday.

Stock position: Long PRS and PRD.