Primus Guaranty and the Viability of the CDPC Model

Includes: PRD, PRSG
by: James Cullen

Most people who have followed my writing during the summer have seen my coverage of Primus Guaranty, a credit derivative products company [CDPC] that sells credit protection via credit default swaps [CDS]. On July 22nd, I placed my entire portfolio into Primus, via a combination of its common stock (PRS) and senior debt (NYSE:PRD). I saw (and see) Primus as a company with very clear risk exposures thanks to the finite and short-lived duration of its swap portfolio, and the fact that a stress-tested loss estimate nets an equity value in the mid-single digits (i.e. about $5/share), which holds even if the company were to enter a run-off.

I've tried to clear a number of misconceptions about Primus, the most important being that the company does not face the liquidity risk that took down Bear Stearns, et al. This is not an accident of chance, but due to the capital structure of Primus – because their AAA/Aaa-rated subsidiary that sells CDS protection is a CDPC, they do not post collateral regardless of how bad the mark-to-market losses on their portfolio become. Even a ratings downgrade could not result in counterparties demanding collateral; it simply isn't how Primus does business.

This "continuation" structure is the huge advantage of the CDPC model; not needing to post collateral makes writing business more economical and can make the system safer as a whole by preventing liquidity-driven runs on sellers of CDS. But that only holds true if the CDS sellers don't take advantage of the collateral exemption and write too much business that goes bad, ultimately leading to insufficient capital to fund their obligations. Because Primus – the first CDPC – received its rating in 2002, there isn't any historical data on how the model performs in periods of market stress.

From what Primus has been saying about business recently and what I've gleaned from the helpful notes by the financial guarantor research team at William Blair, the major concern is that CDS buyers become wary of the CDPC model, and are no longer willing to transact business with sellers who don't post collateral. Ultimately, it's a question the market is going to answer – is there a significant difference in the value of CDS protection from a CDPC, and does the counterparty rating count for anything? Monoline-guaranteed bonds have traded at a discount to those backed by Berkshire Hathaway (BRK) Assurance Corp., so it's conceivable that on the other side of the credit derivatives market mess, some sellers might get slightly better rates than others (we're talking in basis points, but that counts at 30x leverage), because their balance sheets are perceived to be more solid.

Of the CDPCs listed to date that have been rated by Moody's, only one has been placed on review for possible downgrade - Athilon (the second Aaa-rated CDPC, behind Primus) – on July 9th. Moody's has essentially stood behind the CDPC model so far, but that is no guarantee they'll continue to do so in the future, especially if one or more has a spectacular and highly-visible meltdown. Most of their concerns are limited to the CDPCs that were launched later, because they lack the existing book of business that generates consistent and recurring premium cash flow streams; this is not a problem for Primus, because they already have an adequately leveraged swap book generated good premium income along with an existing healthy cash position.

The most interesting point I gleaned from Moody's note on CDPCs (keep in mind, this was published in March) is that their estimate of the one-year rate of defaults on investment grade corporate debt is 0.0466%; my stress test model has Primus' realized credit losses roughly 30x higher at 1.32%, though that isn't entirely comparable because the actual losses net out recovery, whereas the straight default rate merely shows what percentage of issuers default. This is significant because the recovery rate on some defaulted senior corporate debt can be very high – one person told me they aren't highly concerned about the concentration of Primus' swap portfolio in financials because they expect the senior debt to remain intact, I imagine through a combination of white knight guarantees or government backing.

On an apples-to-apples basis using a 24% recovery (the lowest recovery rate on senior debt ever in a single year), the 1.32% rate I establish that Primus can survive between now and the middle of 2010 is 120x higher than the loss rate Moody's estimates translate to. The DCF value of Primus common stock, should no credit losses take place and no new business be written, is $15/share – a triple from the current price. There will likely be some losses, but the probability is that Primus is sufficiently capitalized to weather a spike in defaults and still retain value down through the common stock, which is why I'm invested the way I am.

If you would like a copy of the Excel model I use to value Primus, send me an email at jcullen – at – and I'll reply as soon as possible.

Read the Primus Guaranty Stock Report for more.

Disclosure: Long PRS and PRD.