How a Credit Swap Innovation Unwound After the Financial Crisis
Tom Jasper founded and ran Primus, a unique financial company that sold credit protection on corporate debt and subprime mortgages via credit default swaps. Read time 30 minutes.
Like all senior Salomon Brothers executives in control positions, I left the firm soon after Travelers purchased us in 1997 and merged us with Smith Barney. While my last position at Salomon Smith Barney was as its global treasurer, I had been deeply involved in the interest rate swap market at various points in my 18 years at the firm.
Tom explains: An interest rate swap (IRS) is a contractual agreement between two parties where one pays the other a fixed interest rate and the other pays a rate determined by a floating interest rate index. When I was involved in IRS, the most common floating interest rate referenced was three-month LIBOR, the London Interbank Offered Rate, which doesn’t exist now. The dollar amounts exchanged are determined by multiplying the floating and fixed interest rates by a “notional” amount.
As an investment banker at Salomon Brothers in 1983, I matched up entities wanting to pay floating with entities wanting to pay fixed. European and Asian banks had issued bonds in the fixed-rate Eurobond market, but really wanted to pay a floating rate to match their floating-rate assets. American companies had issued commercial paper at the cheap, short end, of the yield curve, but really wanted the certainty of long-term fixed-rate funding. In a laborious process, I found a bank willing to pay a floating rate to receive a fixed rate it would use to pay its Eurobond. We paired the bank with an American company willing to pay a fixed rate for a floating rate that would more closely match the fluctuating cost of its commercial paper as the company rolled from one issuance to another.
I led the next evolution of the IRS market as co-head of Salomon’s IRS trading desk. Instead of matching fixed and floating rate payers, Salomon stood ready to enter into swaps as principal. We earned a spread between the fixed rate we offered and the fixed rate we bid and managed any unevenness in our book with interest rate futures and long and short positions in US Treasurys. While at Salomon, I was also a founder and co-chair of ISDA (International Swap Dealers Association) which facilitated the market’s growth by developing standardized IRS documentation. At the end of my time at Salomon, as its treasurer, I used IRS to hedge parts of our balance sheet. I was also chairman of Salomon Swapco, a triple-A-rated entity known as a derivative products company (DPC), set up to offer Salomon’s IRS customers a highly rated counterparty.
ePrimus
When I left Salomon Smith Barney, Jay Shidler, a very successful venture capitalist, asked me to help create a highly innovative approach to selling credit protection in the form of credit default swaps. Jay had a vision to create the first credit derivative products company or CDPC. Whereas Swapco was a DPC that entered into interest rate swaps, Jay’s triple-A-rated CDPC would sell credit protection via credit default swaps.
Tom explains: In a credit default swap (CDS), one party pays another for credit protection on a particular reference entity. The dollar amount of protection purchased is known as the notional. The credit protection buyer pays the premium (expressed as a percentage) times the notional to the credit protection seller. A list of credit events including, for example, the bankruptcy of the reference entity, defines the situations where the credit protection seller must pay the credit protection buyer. The amount the credit protection seller pays is determined either by the cash method or the physical delivery method. In the cash method, the price decline from par of the reference entity’s debt obligations is multiplied by the notional amount. In physical delivery, the protection buyer delivers a reference entity debt obligation to the protection seller in par amount equal to notional and the protection seller pays the notional amount to the protection buyer.
Jay had spent considerable time with the credit rating agencies trying to convince them to rate a CDPC triple-A, and he needed someone to help develop the business model and run the company. I was doubtful. Jay was known for completing highly innovative deals in competition with larger venture capital firms, real estate companies, and banks. His partner, Rob Denton, was very analytical, which was a good match for Jay’s outgoing enthusiasm. Both were smart. But Swapco had been set up with the resources of a bulge-bracket investment bank, not to mention the expertise of Noble Prize winner Myron Scholes.
A major accounting firm was already helping Jay build a capital model and the rating agencies were willing to consider the idea. Jay and Rob were off to a good start, and I would be at the cutting edge of yet another swap market innovation. I began working for Jay and Rob as a consultant.
Jay believed the secret sauce was to position the CDPC as a fintech e-commerce company. Its name would be ePrimus Credit Technologies Inc. to capitalize on the buzzwords of the booming e-commerce marketplace in 1998. It would be a one-stop online shop for credit protection where the protection buyer could design its own terms and receive a price with “one click.” He had spoken to several quantitative firms about building such automated capacity, and one stood out, so we flew on Jay’s plane for a meeting in San Francisco.
I knew the firm and it was an impressive group. We were shown into a large conference room and introduced to their team of quantitative analysts. I could sense the heat coming from the tops of their heads as their brains whirred. The lights went down and the presentation began, but it was nothing but a jumble of complex math equations and Greek letters. I asked the person leading the presentation to explain how the pricing model worked in simple, practical terms, not as a mathematical equation. He looked around the room for help as he struggled to answer my question. All he could say was that the model was a highly efficient way to price credit. I looked at Jay and thanked the group as we walked out of the room. The trip to San Francisco had been a complete waste of time from my perspective. Jay was frustrated and agreed with me.
I had worked with many smart quantitative analysts (quants) during my career at Salomon, and I was comfortable around numbers. I had no problem understanding the broader concepts of how a model worked, what factors it analyzed, what its risk tolerances were, and what results one could expect. But a quant had to be able to explain his model in simple terms before he started spouting equations. As Richard Feynman said, “If you can’t explain it simply, you don’t understand it well enough.” The San Francisco team couldn’t explain it simply.
I got Jay and Rob to agree we did not need to build “one click” automated price capability. Instead, we needed to develop operating guidelines and a capital model based on the rating agencies’ own ratings of the companies on which we would be selling credit protection. This was, after all, the model rating agencies already used to rate structures exposed to corporate credit risk, from collateralized loan obligations to Salomon Swapco.
The operating guidelines we developed detailed the credit risk ePrimus would assume. We would sell credit protection via CDS only on companies rated investment grade (triple-B or higher) and we would hold that risk rather than turn around and buy credit protection from somewhere else. We had single name and industry exposure limits. The required capital formula made sure that even under the harshest default scenarios of reference entities, ePrimus would have enough capital to pay its CDS obligations. To protect that capital, it had to be invested US Treasurys. The operating guidelines were almost a hundred pages, and we had to engage an accounting firm to attest every week that we were in compliance.
We began interviewing investment bankers to help us raise capital. Our challenge was explaining why ePrimus was an e-commerce company; the basis for its triple-A ratings; its business rationale, risk-return model, target customers, structure, team, and profitability forecast; and the terms for start-up capital. We spent a month working with a well-regarded investment banker developing the pitch and writing a comprehensive private placement memorandum. The firm was a small boutique investment bank I thought would give ePrimus the attention it deserved as we approached the capital markets for funding.
We were very excited and optimistic when the process of trying to raise ePrimus’s start-up capital began in mid-2000. ePrimus was proposing a lofty valuation comparable to other e-commerce companies that were already trading in the market. Unfortunately, our timing was bad, as the bloom had come off the dot-com rose. Valuations for e-commerce companies were dropping fast. Notwithstanding, we thought we still had a good chance because we were trying to raise capital from sophisticated investors in the private market, not the public market, and we forecast profitability within a year.
I led the pitch in investor meetings. There was the normal give-and-take, but the feedback was mostly positive. We continued to be optimistic and hopeful. But then weeks dragged into months. We kept hearing that it was an interesting business model and a great management team, but it was too early stage and unproven. Finally, our investment banker told us, in the jargon of the day, that it was over, at least for his firm, “ePrimus has run out of runway.” This was crushing news and I immediately called Jay. It was a Friday, and to his credit, he did not overreact. Instead, Jay told me that he wanted to take the weekend to think about our next step. On Monday, Jay had a new plan built around a more traditional Bermuda insurance company model.
Primus Without the e
Primus Guaranty Ltd. would be a Bermuda-based insurance holding company. Primus Financial Products (Primus) would be a triple-A-rated credit derivative product company selling credit protection in the form of CDS. Primus would be based in New York City, but revenue from the sale of CDS would roll up under the Bermuda holding company for tax purposes. Avoiding US tax would significantly enhance profitability. Another important differentiating factor was that Bermuda insurance regulators were comfortable with insurance companies selling CDS, whereas many US state insurance regulators were not.
This was an important learning moment for me. I watched Jay not take no for an answer and create a viable plan B over a weekend. This was another example in my career of finding the right partner. The smartest guys in the room aren’t necessarily the investment bankers. They prefer easy deals with robust fees and guaranteed profits. Our banker told us it was over without offering to work with us to figure out how to get it done. It took Jay Shidler’s creative genius to make it happen.
We hired a new investment banker to help us find a strategic partner and raise start-up capital. This time we selected a bulge bracket firm with significant Bermuda and insurance company experience. Starting in late fall 2001, I led the team on a roadshow that took us to many cities in the US as well as to Bermuda and Europe. It was intense and exhausting because we knew this trip was make-or-break. The US economy was in recession from the dot-com bubble burst, and the stock market had experienced a significant correction. The September 11 terrorist attacks exacerbated economic uncertainty. Even though it didn’t appear to be a good time to raise funds, we felt like we had a strong product to sell, so we remained hopeful. Ultimately, we were successful in attracting XL Capital, a Bermuda-based reinsurance company, as our lead investor; two other insurance companies, Aegon USA and Radian Group; and one private equity firm, Pacific Corporate Group/CalPERS. We raised $270 million of equity and mezzanine capital.
2002
I became Primus Guaranty’s CEO, and our timing could not have been better in mid-2002. The CDS market was booming and CDS spreads widened out significantly after WorldCom’s bankruptcy in July. There was nothing like Primus in the credit swap market. It was not a CDS dealer but a one-way seller of credit protection. Dealers were very interested in Primus because it was a new source to lay off credit risk, and a triple-A-rated one at that. Once they understood how our capital aligned with the risk we took, many of them were on board.
Our first credit swap counterparty was Bear Stearns and our first transaction in the summer of 2002 was the sale of five-year credit protection on highly rated (AA/Aa2) General Electric Credit Corporation (GECC). The price was north of one hundred basis points. A one percent premium on a highly rated company like GECC was very attractive. The guiding strategy for the business was to maintain the company’s triple-A ratings while building a well-diversified portfolio of risk at attractive returns to our shareholders.
By the end of 2002, we had sold $4 billion notional of credit swap protection. Primus was now a recognized source for credit swap protection, particularly on less liquid reference entities which did not trade daily in the CDS market. There was limited pricing transparency on these reference entities, which meant that Primus had pricing power in this sector, whereas in the actively traded market, there was good price transparency and availability from many different CDS dealers. Fifteen dealers might quote IBM at a small premium. In that scenario, Primus had little pricing power. We focused on less liquid reference entities where we had pricing power and an advantage as a buy and hold investor. We could quote a fair price to a risk manager on these names knowing they would be hard-pressed to find a competing bid from a dealer.
Counterparties who wanted collateral
We had swap documentation in place with twenty counterparties, but there were still a few credit swap dealers at major banks that would not do business with us. Notwithstanding our triple-A ratings, capital, and portfolio management approach, these dealers wanted Primus to post collateral to them based on the market value of the CDS. If a reference entity’s credit quality deteriorates and its credit spread (the difference in the yield of the reference entity’s debt over a benchmark like Treasurys) increases, the market value of the CDS increases. The amount is a liability to Primus and an asset to the dealer. Some dealers wanted us to agree to post collateral equal to the CDS’ value to them in that circumstance.
It was an easy pass, especially given that our rating agency operating guidelines did not allow collateral posting. Our capital was there to pay any credit loss we incurred. It was not there to post as collateral to counterparties. Adherence to our ratings guidelines worked to our advantage when, in 2007-08, credit spreads widened completely beyond real credit risk and other CDS market participants defaulted because they did not have enough capital to post as collateral on positions that ultimately did not incur any credit losses.
2003 to 2005
At the end of 2003, Primus’s portfolio exceeded $6 billion, a 50% increase in just one year. We reported earnings of $15 million to our shareholders and there had been no credit losses. Other companies began copying our triple-A CDPC business model. This was flattering, but it also pushed us to go public as soon as possible to maintain our edge as the founding and leading CDPC. We did so in October 2004, trading under the symbol “PRS.” We got to ring the opening bell at the New York Stock Exchange.
An important aspect of Primus Guaranty becoming a public company was that we had to adhere to generally accepted accounting principles (GAAP) in our quarterly financial reports. GAAP required that we report earnings based on the fair value of the company’s swap portfolio. This was standard accounting policy for derivatives like credit swaps, but Primus’ buy-and-hold business model meant that we were going to show significant earnings volatility, and this would not please public shareholders. We appealed to our auditors that our credit swaps were a risk transfer vehicle, like an insurance policy. Insurance companies were buy-and-hold investors, like Primus, and did not mark to fair value. We got nowhere with this argument. It was frustrating, but there was no flexibility with GAAP.
I developed a non-GAAP measure of Primus performance called Economic Results which took out the mark-to-mark volatility and focused on the core earnings power of the company. The SEC is not a fan of non-GAAP measures, but I added this to our financial statements and encouraged shareholders to focus on that performance measure. I also convinced the Primus Guaranty board to compensate employees based on Economic Results.
At year-end 2004, Primus’ swap portfolio was $10.5 billion, an increase of 68% from the previous year. Primus Guaranty reported economic results of $22.4 million, a 37% increase from the previous year. Capital grew by 85%, and we were now doing business with 38 counterparties. We broadened Primus’ risk parameters and received rating agency approval to sell credit protection on synthetic collateralized debt obligations (SCDO).
Tom explains: Unlike a cash CDO, which buys a portfolio of corporate bonds or loans, SCDOs gain credit exposure by selling credit protection via CDS, most often on a well-diversified portfolio of 100 or so reference entities. It balances this risk by buying credit protection via tranches responsible for different levels of portfolio credit losses. A first-loss equity tranche might be responsible for first losses up to 15% of notional, an attachment point of 0% and a detachment point of 15% in SCDO jargon. The last-pay triple-A-rated tranche might be only responsible for losses exceeding 90% of the portfolio.
Primus would underwrite each reference entity in the SCDO and sell protection on single-A or double-A tranches with 15% attachment points. Given the history of investment grade defaults, we never expected any credit losses and collected a nice premium. Primus had an advantage over monoline insurers like MBIA, which were rating-agency restricted to selling protection on triple-A tranches. We also received rating agency approval to sell credit protection on Euro-denominated single-name credit swaps. These new product lines complimented Primus’ single-name domestic CDS business.
We formed Primus Asset Management (PAM) to get involved in the SCDO market as a manager. In 2004, PAM formed a $1 billion notional SCDO composed of a diversified portfolio of one hundred $10 million notional credit swaps on investment-grade reference entities. PAM received a fee for originating, structuring, and managing the portfolio on behalf of its investors. The SCDO was tranched into triple-A, single-A, and unrated equity tranches and sold by investment banks to global investors.
We had another banner year in 2005, our first as a publicly traded company. At year-end, our credit swap portfolio was $13.4 billion, with 535 reference entities with an average rating of A-/Baa+ across 40 industries in 25 countries. We now had the highly diversified portfolio that Primus’s operating guidelines called for, and the rating agencies approved us to sell a limited amount of credit swaps on less-than-investment-grade double-B-rated reference entities. This addition enabled Primus to broaden its platform and to be even more attractive to its counterparties. With a large portfolio covering a multitude of reference entities, it was inevitable that there would be some credit issues. We began to aggressively mitigate risk by terminating swaps when we became uncomfortable with the credit risk. This resulted in Primus realizing small credit losses.
2006
Entering 2006, Primus continued to have the wind in its sails. During the year, we grew Primus’s credit protection portfolio 18% to $15.8 billion. We also grew assets under management in PAM by $800 million, including launching our first collateralized loan obligation (CLO) and our third SCDO. I had been interested in the non-investment grade credit (rated double-B to triple-C) for a few years. I felt that PAM should manage more CLOs. We didn’t have enough expertise in this area, so I hired an experienced team to get us into this business.
A difficulty with CLOs was that PAM had to buy the loans and hold them in a temporary “warehouse” fund until the loan pool was of a certain size and diversity and the CLO could be launched. One of our challenges in getting into the business was that warehouses were generally financed by a bank and finding a bank was difficult for a first-time manager like PAM. To convince a bank to provide the warehouse facility, PAM assumed a first-loss position in the pool of loans being purchased.
We also launched PRS Trading, a relative value credit hedge fund we capitalized with $50 million. PRS engaged in relative value directional strategies in credit swaps to profit from changes in market prices over one month to a year. This was different from our core five-year buy-and-hold strategy at Primus. I hired an experienced credit trading portfolio manager to run it.
2007
At the beginning of the year, I forecasted significant growth in Primus’ credit swap portfolio, given the outlook for wider credit spreads. At the same time, I felt that the company had to be highly focused on managing its credit risk, with the goal of improving its portfolio credit quality.
I had been watching the growth in subprime residential mortgage-backed securities (referred to as asset-backed securities or ABS for reasons too complicated to describe here) for a couple years. Was this a market where Primus could add value and make money? Could we understand the risk associated with ABS, and would that risk fit our portfolio? Could I hire a top-notch manager? These were important questions. One of Primus Guaranty’s board members warned me that selling credit swaps on ABS would be a difficult and risky business, particularly given the current issues in subprime mortgage lending. Many ABS were poorly structured and underwritten and home prices had been falling since the middle of 2006. This board member wanted to make sure I was aware that the industry had bad actors. I listened closely to his advice but decided to move forward. That turned out to be a big mistake.
Primus went through the analysis the rating agencies required to sell CDS protection on ABS, and after many months of back-and-forth, we received their approval. We would sell protection on the double-A tranches of subprime residential mortgage securitizations at 12%-13% attachment points. I thought Primus would be safe at that attachment point. The challenge for Primus in underwriting this risk was understanding the underlying quality of the mortgage pool. This was difficult and quite different from the SCDO market, where Primus could identify and underwrite each corporate name in the pool. Could I find a top-notch professional to get us into the business? I hired a portfolio manager who I thought would be a good fit. Rating agency approval and hiring a team took twelve months from the time we got board approval. That delay probably saved the company.
I kept a close eye on the portfolio, owing to all the negative housing market news. When our ABS CDS reached $80 million notional, I had a conversation with our manager about the significant price volatility of the swaps Primus had sold. I asked the portfolio manager where the bottom was and received a blank stare in return. It was clear that our expert had no idea how bad things might get. I immediately put a hold on additional ABS exposure. Little did I know that the exposure we already had would generate a $41 million loss. I had made a mistake getting into ABS CDS, and I took complete responsibility for it with the Primus Guaranty board and shareholders. I also decided to shut down the company’s hedge fund, PRS Trading Strategies, owing to its poor performance and its inability to raise third-party capital.
There were some bright spots in the company’s performance in 2007. Primus took advantage of widening credit spreads to increase notional 46% to $23 billion, focusing the growth on higher-quality reference entities, improving the overall portfolio quality to a strong single-A. We had 47 counterparties and there were no credit losses. PAM executed a second $400 million CLO, bringing total assets under management to $1.9 billion before the CLO and SCDO markets shut down.
2008
We all knew 2008 could be a difficult year for Primus Guaranty. We were going to focus on taking advantage of wide credit spreads to increase Primus’s notional. This was our crown jewel, and it continued to perform well. Our credit protection portfolio was highly diversified, including 591 reference entities spread across 40 industries in 27 countries. We also had more than $860 million in capital to support the business. That said, we would look to mitigate risk if we saw deteriorating credit conditions in an industry or a reference entity.
We began 2008 strongly, selling $1.2 billion in credit swaps at very attractive premiums and increasing Primus’ portfolio to $24 billion notional. But then Bear Stearns experienced a classic bank run in March 2008. The company had taken significant losses from bailing out two of its subprime mortgage hedge funds and its liquidity dried up owing to a lack of confidence in the firm. Eventually, JPMorgan bought Bear for two dollars a share, a fraction of its value. I was pleased the Fed had forced the sale as Primus had significant CDS exposure to Bear Stearns. With the sale, JP Morgan was now the reference entity on those CDS and we escaped losses.
Investors heavily criticized the rating agencies for giving triple-A ratings to ABS and misleading them into believing those securities were safe investments. Many ABS were now rated below investment grade and experiencing significant realized losses. Faith in credit rating agency ratings were such that Primus’ triple-A ratings meant nothing.
I got calls from counterparties that they would no longer buy credit swaps from us. Bear’s failure was the tipping point, and fear and uncertainty were gripping the financial markets. Unable to sell any additional credit swaps, I focused on preserving the value in our existing portfolio, with the thought that counterparties would reopen their lines to Primus once the market calmed down. After all, Wall Street has a very short memory. Or so I hoped.
Lehman Brothers
I was very close to Lehman Brothers. They had taken Primus Guaranty public and arranged most of its $1 billion in capital. But after Bear Stearns’ bankruptcy, the market speculated that Lehman would be the next bank to fail. I paid close attention to Lehman’s public filings as well as the gossip surrounding the firm. We had written $80 million of CDS on Lehman. Was it time to reduce our Lehman exposure? Many on the Primus team felt so.
Amid all the rumors circulating about Lehman in spring 2008, I received an invitation from the bank to attend a golf outing to the Old Course at Saint Andrews, Scotland. The outing was first class: wives were invited, private jet to Saint Andrews, and dinners at amazing venues. I will never forget teeing up with my wife Alexis on Old Course’s first hole and walking over the famous bridge to the eighteenth hole. Meanwhile, the eight-hundred-pound gorilla in the room was whether Lehman would survive. But what company on the brink of bankruptcy would organize such a posh event? Well, Louis XVI’s ostentatious spending was designed to keep France’s creditors lending money all the while the country was becoming insolvent. Lehman’s top management told a consistent story that the company would prevail. I was heartened by their sincerity, but perhaps the great golf jammed my antennae.
The mark on Primus’s $80 million notional CDS referencing Lehman deteriorated to negative $50 million. The market was clearly signaling that it expected Lehman to fail. But now there was no clear course of action. We could pay someone $50 million to assume our obligation and crystallize the loss. If we kept the position, two things could happen. Lehman could survive or be bailed out, like Bear Stearns. That would be great. Or Lehman could fail and we’d have to pay off under the terms of the CDS. But we didn’t think our CDS payout would be much different from the $50 million we’d have to pay someone to take over our position. If we held on, we’d collect CDS premiums until Lehman went bankrupt and maybe it wouldn’t go bankrupt. I decided to hold on and ride it out.
When Lehman declared bankruptcy on September 15, 2008, the only question was how Primus should manage a Lehman closeout. Under cash pay, we would pay a net amount, the swaps’ notional times the decline in the market value of Lehman’s debt from par. Under physical delivery, we would pay the total notional amount and take delivery of Lehman bonds with face value equal to notional. I decided to take delivery of Lehman bonds, which ultimately was marginally better for Primus.
Unfortunately, Lehman was not the only reference entity in Primus’s credit protection portfolio to fail in September and October 2008. Primus experienced six credit events totaling $280 million notional. Primus took net realized credit losses of $145 million. In addition, Primus paid $4.8 million to improve our credit position on certain reference entities and took a $9.3 million provision on our ABS CDS portfolio. At the same time, Standard & Poor’s and Moody’s downgraded Primus to A+ and A1, respectively.
We were being battered from all sides of the credit swap market. Counterparties were not willing to transact with us. Our high single-A ratings were worthless as the market had no faith in credit agency ratings. The SCDO and CLO markets were also shut down, so there were no opportunities for PAM. It was a very stressful period for me and the team. The stock price was just above $1 a share, a far cry from its $13.50 IPO price.
Primus Guaranty reported a year-end 2008 GAAP loss of $1.7 billion and an economic loss of $82 million. The only good news after the extremely difficult third and fourth quarters in 2008 was that even with credit events, ratings downgrades, and negative earnings, Primus was not faced with the burden of providing collateral to its counterparties. At the time, Primus had capital available to support realized losses of $770 million while the mark on our CDS portfolio was negative $860 million. If Primus’s counterparties had had the ability to call for collateral, Primus would have been out of business.
Our position contrasted with that of AIG Financial Products (AIGFP). AIGFP was a major provider of credit swaps, primarily on structured credit pools, including many subprime mortgage pools. It relied on a guarantee from its triple-A-rated insurance company parent, AIG, to do business. AIGFP’s CEO came to our office with his entourage and said he wanted to buy Primus Guaranty. I told him to make an offer, and I would take it to my board. He said I would hear from him shortly. Just as he was leaving, I asked him whether the rumors were true that AIGFP was now willing to provide collateral on its swaps. He said yes, but only if AIG’s rating fell to single-A or below. “That would never happen,” he said quite confidently.
I told him Primus would never provide such a downgrade collateral trigger. I felt it would be very difficult to access the liquidity needed to fund such a claim on Primus’s capital no matter how unlikely. He shrugged and left my office. I never got an offer from him to acquire Primus Guaranty and a few months later, in September 2008, the credit rating agencies downgraded AIG to single-A. AIGFP’s counterparties demanded their collateral, but AIG didn’t have the liquid assets to post. This led to the largest government bailout in history, $182 billion. Without the government coming to the rescue, AIG would have failed.
2009
Because our counterparties stopped buying credit protection from us, we developed a plan to reduce and derisk Primus’s $24 billion CDS portfolio and return capital to shareholders. Primus’s embedded value was cash capital, plus CDS premiums, less operating costs, capital costs, and CDS credit payments. We committed to returning capital to shareholders every quarter as swaps matured. It would take three or four years to return Primus’s total embedded value. The board also authorized a $25 million stock buyback program. The stock was trading in the one- to two-dollar range, a significant discount to its $8 economic value, so buybacks were accretive to shareholder value. Finally, we committed to reducing operating costs, including dropping credit rating agency ratings and the costly accounting fees needed to produce weekly capital compliance reports. Primus couldn’t write any new business and thus did not need ratings. Reducing operating costs also meant layoffs. It was very difficult for me to let go employees who had worked extremely hard through the turmoil of 2008.
At the same time, we wanted to grow PAM by pursuing opportunities in the structured credit markets. That could include acquiring structured credit asset pools and/or managers using capital not allocated to Primus. The good news was that PAM continued to rank as one of the larger structured credit managers with over $20 billion in assets across a variety of vehicles. The other piece of good news was the dramatic rebound in the credit markets. There was record issuance of corporate bonds, and credit spreads had narrowed significantly. Nonetheless, there continued to be bank failures, CDS credit events, and a high level of uncertainty in the financial markets.
In mid-2009, I presented the board with the opportunity to acquire Cypress Tree Investment Management, a Boston-based, highly regarded CLO manager. I felt that the acquisition would significantly strengthen PAM’s platform and the board approved the acquisition. Once the deal was signed, I quickly integrated its platform into PAM. We developed ideas for organic growth, and Primus Guaranty provided a modest amount of seed capital to begin developing an investment track record.
During 2009, Primus terminated $2.8 billion notional in credit swaps and capped its exposure on an additional $2.9 billion. It also raised its attachment points in some SCDO tranches to reduce the likelihood of a payout. Primus paid $66.5 million to terminate and amend contracts and $126 million to cap exposures. Primus’ portfolio performed well in 2009, with only three credit events, whereas there were thirty-eight credit events in the CDS market that year. We bought back 7.9 million shares of Primus Guaranty’s common equity and retired $30 million face value of Primus Guaranty notes. At year-end 2009, we reported $1.46 billion in GAAP earnings and a $3.7 million loss in economic results. The stock price rebounded to over four dollars a share, so it was moving in the right direction, but it was still trading significantly below its $8.50 economic value. We believed strongly that the embedded value in Primus’ credit protection portfolio alone was in the mid to high single digits.
I have always been a glass half-full person, meaning I look for the positives, and this mindset served us well as we navigated through the financial crisis and market turmoil of 2008. Primus not only survived the crisis, but also expanded its business and retained its reputation. Surprisingly, 2009 turned out to be a very good year for us. We were even able to celebrate with a year-end Christmas party that, I believe, gave us all a sense of hope about the future, especially given that there had been no party the year before. It was a merry occasion for all, and we closed 2009 with high hopes for an even better 2010. Little did I know how quickly that optimism would fade.
2010
It was obvious to me that 2010 was a rebuilding year, and I made this clear to the board. It was quite a blow, then, when my proposed strategy was met with serious resistance. Some board members insisted we focus on amortizing Primus’s portfolio rather than generating incremental revenues from PAM. I let them know that I was confident we could achieve both, but that we needed to be realistic about how much progress we could make in a year.
But XL Capital had no interest in being patient, and our board chairman, Mike Esposito, an XL executive, was under tremendous pressure from XL’s board to produce results and to produce them quickly, especially given that XL had written off its Primus Guaranty investment. I began to doubt whether I had what it would take to rapidly lead the company to higher profitability. But after much soul-searching and intense conversations with people I trusted, I decided that I was in fact the person to lead the company to its next stage. I had to remain positive and not reveal my inner thoughts to the board or to the team.
I hired an Irish investment banker to help Primus Guaranty acquire a European asset manager. My idea was to merge that entity with Cypress Tree and our existing CLO platform so that we could originate CLOs in the US and Europe. PAM needed a larger footprint in the loan market, and trying to do that organically deal by deal would take too long. I traveled to Ireland and met with acquisition candidates.
While I was on the trip, I received a call from Mike Esposito telling me to shut down the search. This was another ominous sign of things to come. The board then asked me to hire an investment banker to take Primus Guaranty private. With the stock trading in the low single digits, and the economic value of the company well above its market value, this seemed like an ill-advised move. As a private company, Primus Guaranty would have serious difficulty accessing the capital it needed to grow and reach profitability, never mind being able to attract the talent necessary to run the business. I tried to understand why the board had come to this decision, but the only explanation I could come up with was that XL Capital wanted out.
My job now was to find a strategic partner that would buy out XL and become the lead investor of the newly private company. But I couldn’t get past the question of why the board would even consider selling the company when its stock was trading so far below economic value. I had to keep these concerns to myself and simply take my marching orders. So I hired a banker, and began the process of marketing Primus Guaranty, its team, and its vision for the future to potential new strategic partners.
There were multiple ways to value Primus Guaranty, its current book of business, and its vision for the future. The stock price that continued to trade in a four- to five-dollar price range was just one touchpoint. I was confident that between our banker and myself, we were presenting a realistic appraisal of the company’s value. But I also knew that it would be very difficult to get a strategic investor to pay anything close to economic value. The board was disappointed with the proposals from the handful of strategic investors to acquire the company and take it private. I kept hearing from the board, “Why would we sell the company at these levels?” Ultimately, and much to my relief, the board decided to abandon the idea of going private.
I was at a stalemate with XL. Our lead investor wanted to sell its stake and move on, whereas I wanted to use its capital to build a new robust business over a two- to three-year time frame. I recognized that I had little leverage over XL and my own board, as I could not promise an immediate uptick in the business, let alone its stock price. The board decided that the company’s highest value asset was amortization of Primus’ credit swap portfolio. The company should focus its resources solely on the Primus runoff. The board would not approve allocating capital or spending money on any other business, no matter how attractive it was.
The board was telling me to shut down Primus Guaranty except for Primus Financial Products. The goal was to cut costs to the bare minimum and to continue to de-risk the portfolio and return capital to shareholders as credit swaps matured. My only mandate now was to honor Primus Guaranty’s debt and preferred stock obligations and take advantage of opportunities to buy back debt and equity when it was accretive to shareholders.
This meant shutting down or selling noncore businesses like our CLO and SCDO assets and laying off personnel not directly associated with the amortization of Primus’ credit swap portfolio. I did not enjoy firing people under any circumstance, but this was gut-wrenching. I had hired these talented people, and they had successfully created and built these businesses. I did not try to sugarcoat the board’s decision, and everyone knew I didn’t support it. I negotiated hard with the board for generous severance packages for employees we fired, and I took pride in placing some of the team with other investment managers.
My career, until then, had focused on building and running thriving growth businesses. I told the board that I had no interest in managing a company in runoff and that they should begin the search for a new CEO. Now I had to negotiate my own severance package. This was a difficult negotiation, but I advocated hard for myself, just as I had for my employees. I left Primus Guaranty at the end of 2010 and at age 62 suddenly found myself unemployed.
I believed that Primus Guaranty was missing a big opportunity, and that made my last day there hard. For the year, the company reported GAAP earnings of $255 million and an economic loss of $41 million. Primus’ credit protection portfolio had declined to $10.4 billion, down from a high of $24 billion. Most of the restructuring and loss mitigation the company would undertake was now complete. They even went as far as to publish the names of Primus’s remaining reference entities, allowing shareholders to form their own view of the portfolio’s risk and better analyze its amortization value. Nine years after I left the company, in January 2020, Primus made its final $3.2 million distribution to shareholders.
The Primus Guaranty experience was one of the highlights of my career. It’s not often that you’re given the opportunity to build an entirely whole new type of business from scratch, to do it successfully, and to do it with great partners, particularly Jay Shidler, the visionary behind Primus Guaranty. Together, Jay and I created the credit derivative product company (CDPC) model that was, in effect, an unregulated insurance company specializing in credit risk.
Nine other CDPCs were created, largely copying Primus’ model. By the time I left Primus, these ten CDPCs had over $100 billion notional in outstanding credit swaps. Primus was the largest. I must give credit to the rating agencies that patiently facilitated creation of the CDPC model out of DPC whole cloth. Their triple-A ratings, capital model, and operating guidelines put Primus in business. It enabled us to build a robust, well-managed risk business that survived the 2008 financial crisis when many other institutions failed.
I enjoyed building a team of highly professional people, mentoring them, and watching them succeed. I believe I was a good leader. Primus Guaranty also gave me the opportunity to build a board of directors and to develop a robust governance structure. I enjoyed working with Mike Esposito, the board chair, and all the strategic and independent directors. The board was very supportive of me, my vision, and my management team. For the most part, XL Capital was a great lead investor, and Aegon, Radian, and Pacific Corporate Group were also excellent partners.
Although I disagreed with XL’s position late in my tenure at Primus Guaranty, I understood their motivation, as they had their own shareholders to satisfy. I also value the support I received from several public shareholders who expressed their appreciation for my leadership, my vision for the company, and the team I put together. One of the few disappointments I had was that the stock never reached the valuation it merited.
I developed strong relationships with a broad group of talented people in the credit markets, and these relationships were essential to our success. Even in the dark days of 2008, Primus and its counterparties acted in a responsible and professional manner.
I enjoyed working with Jay and Rob. They were very different personalities, but both were smart and focused. They also had good senses of humor. When I sought Jay’s advice, he would ask, “What’s the problem we’re trying to solve, and how do we solve it?” Those may seem like obvious and unhelpful questions, but what he was saying is that if we drill deeply into the problem and really understand it, that’s where we’ll find the solution. I’ve used this phrase many times since then to keep myself focused on what’s most important, and it always helps me to take the next right action.
This story is excerpted from Wall Street Maverick: Musings on a Career of Innovation, Derivatives, Credit, and Risk. Tom is currently Managing Partner of Manursing Partners, a consulting firm. He also serves on the board of directors of three nonprofits and on the boards and audit committees of three Blackstone credit funds. Risk Magazine inducted Tom into its Hall of Fame in 2002.
To comment on a story or offer a story of your own, email
Copyright © 2026 Thomas Jasper. All rights reserved. Used here by permission. Short excerpts may be republished if Stories.Finance is credited or linked.