The Fall of AIG: The Untold Story
The inside story of Hank Greenberg and the rise and demise of AIG.
One of the more improbable resurrections to emerge from the financial crisis is the sublime bit of Kabuki theater whereby Maurice (Hank) Greenberg, the 84-year-old Yoda who built and led the insurance behemoth American International Group for 40 years before being deposed in March 2005 by his own handpicked board of directors, recasts himself as both a victim of a ruthless witch hunt by then–New York State attorney general Eliot Spitzer and as the only man who could have prevented AIG’s collapse and saved American taxpayers the tens of billions funneled into the company since September 2008.
AIG turned out to be ground zero of the crisis, the most interconnected of all the financial services companies, in large part because it had foolishly decided to insure many of the risks in the system. Its collapse on September 16, 2008, caught nearly everyone by surprise. Indeed, the unraveling of the House of Hank was a quake of such magnitude that the Federal Reserve Board decided to bail out many of AIG’s trading counterparties at 100 cents on the dollar. Who they were and how much they received was kept a state secret for months. (Why that happened has been the subject of an ongoing congressional investigation.)
There is little question that the combination of the bankruptcy of Lehman Brothers Holdings and the near-death experiences of both AIG and Merrill Lynch & Co. that same September 2008 weekend led directly to the decision by Ben Bernanke, the Federal Reserve chairman, and Henry Paulson Jr., then the secretary of the Treasury, to demand that the U.S. Congress pass into law what became known as the $700 billion Troubled Assets Relief Program. “Of all the events and all of the things we’ve done in the last 18 months, the single one that makes me the angriest, that gives me the most angst, is the intervention with AIG,” Bernanke said on the television program 60 Minutes. “Here was a company that made all kinds of unconscionable bets. Then when those bets went wrong . . . we had a situation where the failure of that company would have brought down the financial system.”
Over the past year or so, Greenberg has told his side many times. When asked by Vanity Fair columnist Michael Wolff about his reaction to the unexpected collapse of AIG, in which he was personally heavily invested and of which he never sold a single share of stock, Greenberg was incredulous. “It was hard for me to believe that you could damage that company, no matter what kind of incompetence,” he said, referring to AIG’s management after he left the company. “I was wrong. There are people who are so incompetent they can destroy anything.”
Not everyone agrees with Greenberg’s version of events, of course — an opinion summed up succinctly by California Democratic Representative Henry Waxman at a postbailout congressional hearing: “Mr. Greenberg blames Mr. [Martin] Sullivan and Mr. [Robert] Willumstad [Greenberg’s two immediate successors] for the downfall of AIG. Many others think it is Mr. Greenberg who sowed the seeds that led to AIG’s failure.”
One of those who believed Greenberg sowed those seeds is, not surprisingly, Spitzer, who was the first to uncover signs of shady accounting at AIG and to prosecute. “The company was rife with impropriety,” Spitzer tells Institutional Investor. “Management of the company was unambiguously flawed. There was a domineering attitude from the very top that said, ‘Play games with the numbers to get the outcome that we want.’” Observes one former senior AIG executive who is also of the view that Greenberg deserves a meaningful share of the blame: “Hank left a huge hole in the place. There’s no doubt. I mean, if Hank wants to subscribe to the theory that the place fell apart after he left, he’s right. But it’s what he left [behind]. He left the house without any support. The infrastructure around Hank was particularly weak. What a surprise! Who the hell would want to work for the guy? He was a son of a bitch.”
Yet many aspects of the Greenberg narrative are compelling. There is no question that things seem to have gone very wrong at AIG in the three and a half years between Greenberg’s firing and the taxpayers’ September 2008 bailout of the company. It was almost as if Greenberg had designed a company so vast, complex and Escher-esque, operating in some 130 countries with roughly 92,000 employees, more than $100 billion in revenue and many millions of individual transactions every year, that only one person could ever run it properly: Hank Greenberg, the former U.S. Army captain who enlisted at 17, landed on Omaha Beach on D-day and helped liberate Dachau.
That sense of mission may help explain Greenberg’s systematic undermining of potential successors, including both of his sons, Jeffrey and Evan. Of the man who eventually did get his job and whom Greenberg handpicked, Martin Sullivan, a 34-year AIG veteran, Greenberg says simply in a recent interview with II: He was, in effect, working above his pay grade — meaning, well above his ability. This despite having nurtured Sullivan along in various important roles at the company and making sure he attended years of risk management meetings. The way it was supposed to work, Greenberg said, was that Sullivan was to succeed him at the annual meeting in May 2005, with Greenberg keeping a close eye on him for about six months — Greenberg was to remain board chairman under the succession plan. If Sullivan performed well, no harm, no foul. If not, he would be out. “It’s one thing for a guy to do well in what he’s doing,” Greenberg said of Sullivan in the same interview. “Some guys can’t reach the next level.”
(Sullivan could not be reached for comment, but in his testimony before Congress he blamed AIG’s troubles on the “unintended consequences” of mark-to-market accounting and an unprecedented “global financial tsunami,” contending that he “succeeded” in restoring confidence in AIG after Greenberg’s departure.)
But the dominoes started falling on Valentine’s Day 2005, five days after AIG announced its 2004 earnings of $11.04 billion, a stunning 19.1 percent increase from the year before. That same day AIG revealed it had received subpoenas on February 9 from both Spitzer and the Securities and Exchange Commission related to their investigations into AIG’s accounting for various “nontraditional insurance products” and “assumed reinsurance transactions.” AIG said it would cooperate in responding to the subpoenas. A month later, on March 14, AIG announced that Greenberg would step down as CEO to become nonexecutive chairman of the board of directors and would be replaced by Sullivan, who since 2002 had been AIG’s vice chairman and co–chief operating officer as well as a member of the company’s board. AIG’s spin was that it had “implemented its management succession plan” by picking Sullivan. As Greenberg was then 79 years old, the claim seemed plausible.
But in reality, Spitzer and the SEC were using the threat of a criminal indictment, which no financial company had ever survived, to put pressure on the board to dispense with Greenberg. According to the Wall Street Journal’s breathless April 2005 front-page account, Spitzer worked the phones late on a Saturday night from his Colorado skiing vacation to make sure the board took action. Then, the Journal reported, “Mr. Spitzer issued the ultimate threat: His office would indict AIG on Monday if action wasn’t taken.” That did it. The board fired Greenberg. (Disclosure: At that time, my sister-in-law was on AIG’s board; we have never spoken about AIG.) Under additional pressure from Spitzer, on March 28, Greenberg severed all ties with AIG and announced he would not serve on its board in any capacity.
But with the firing of Greenberg, the Spitzer criminal investigation and the potential accounting restatements, Standard & Poor’s, one of the three major rating agencies, decided on March 30 that it would shoot first and ask questions later. Late that afternoon S&P lowered AIG’s coveted AAA debt rating to AA+ for both its long-term counterparty credit and its senior debt. In making the downgrade from AAA, S&P analyst Grace Osborne cited the delayed 10-K filing and the uncovering of a “number of questionable transactions that span more than five years” that could result in a decrease of $1.7 billion to AIG’s reported shareholders’ equity. She acknowledged that the potential financial impact of the “questionable transactions” might be slight — “about 2 percent of shareholders’ equity” — but some air seemed to come out of the company’s vaunted sails all the same. New CEO Martin Sullivan wrote a letter to shareholders on April 4 in which he noted the various inquiries and stressed how committed AIG was to cooperating with the company’s regulators to resolve any problems or concerns.
Curiously, nowhere in Sullivan’s letter, S&P’s downgrade analysis, news of Greenberg’s departure or the various investigations were the words “AIG Financial Products” even mentioned. Indeed, it is probably a safe bet that as of April 2005 very few people outside of 70 Pine Street, AIG’s world headquarters in downtown Manhattan, or AIGFP’s Wilton, Connecticut, and London offices had even heard of AIGFP and its manager, Joseph Cassano, or how it was making increasingly large amounts of money for its parent, AIG. Needless to say, there was little, if any, disclosure to investors of the unprecedented risks Cassano was taking on.
If nothing else, Hank Greenberg was a smart and ruthless businessman. He diversified AIG beyond simply writing fire and casualty insurance to become the world’s largest underwriter of commercial and industrial insurance. He was also an innovator. He created insurance for directors and officers of corporations to protect them against their own blunders. He created environmental protection insurance and coverage for those threatened by kidnapping. AIG “built a team of skilled underwriters who were capable of assessing and pricing risk,” he often proclaimed. Greenberg also knew that AIG’s AAA credit rating gave the company a valuable advantage in the marketplace by allowing it to borrow money cheaply, invest it at higher rates of return and make money on the spread. If this kind of thing could be done outside the often onerous state regulations that blanket the insurance industry, even better.
To that end, in 1987, Greenberg created AIGFP by hiring a group of traders from the investment bank Drexel Burnham Lambert, led by Howard Sosin, who supposedly had a better model for trading and valuing interest rate swaps and for taking on and managing the risk that other financial firms wanted to sell. The market for derivatives was in its infancy, but growing, and Greenberg determined that AIG could be at its forefront. According to Greenberg, the overriding strategy at AIGFP was for the business to lay off most of the risks it was taking on behalf of its clients so that AIG was not exposed financially in the event of huge market-moving events that could not be modeled or anticipated. Greenberg says under his watchful eye the formula worked famously: From 1987 to 2004, AIGFP contributed more than $5 billion to AIG’s pretax income and helped the company’s market capitalization increase to $181 billion from $11 billion.
The business Sosin and his team created was nothing more than a hedge fund attached to a “large and stable insurance company,” Bernanke said of AIGFP in a March 2009 interview — with the added benefits that their access to capital was seemingly unlimited and costless, and that instead of getting the typical “2 and 20” deal, AIGFP’s traders got to keep a jaw-dropping 30 to 35 percent of the profits generated from other people’s money. This sweet arrangement allowed many of the 400 or so people who worked at AIGFP to get very rich.
Things at AIGFP were humming along so well that when Sosin and Greenberg had a falling-out in 1993 and Sosin quit, taking a reported $182 million in severance with him, the business didn’t miss a beat under his successor, Thomas Savage, a mathematician who encouraged his traders to challenge him about the efficacy of the risks the group was taking. Savage retired in 2001 and was replaced by Cassano, a former political science major at Brooklyn College. Cassano had been the back-office, operations guy at both Drexel and AIGFP before getting the top job. By then, following J.P. Morgan & Co.’s creation of a way to do so, AIGFP had started insuring the risk that corporations might default on the debt they had issued. By selling something that became known as “credit default swaps” to nervous investors, AIG agreed to pay off the defaulted debt at 100 cents on the dollar. AIG got the premiums; investors got peace of mind. The biggest part of the AIGFP insurance book, some $400 billion, was written on behalf of European banks looking to take risk off their books as a way of avoiding the need of raising additional capital to appease the European regulators. “This is a great irony,” explains a former AIG executive. “The European banks went out and were able to buy credit default insurance on their assets so that they didn’t have to keep as much capital on their balance sheets. So here was an insurance company in the United States with essentially no liquidity, no equity and no reserves providing equity relief for European insurance companies. Talk about the house of cards.”
Cassano was not a particularly benevolent leader. “Joe would bully people around,” a trader told writer Michael Lewis. “He’d humiliate them and try to make it up to them by giving them huge amounts of money.” Needless to say, the camaraderie and openness of the Savage era was lost quickly in the savage Cassano regime. But none of that mattered much at AIG as long as Greenberg was running the show, for he was every bit Cassano’s match in ruthlessness, but a better and more astute businessman. (Cassano could not be reached for comment.)
All that changed, though, when the AIG board dispensed with Greenberg in March 2005. “[R]eports indicate that the risk controls my team and I put in place were weakened or eliminated after my retirement,” Greenberg wrote in his October 2008 congressional testimony. Not only did the oversight of Cassano lessen, but the loss of AIG’s AAA credit rating meant that the counterparties who had paid the premiums to and bought the credit protection from AIG could demand that AIG post collateral, in cash or securities, should the value of the debt being insured fall. This made sense, of course, in that an insured party would want to know that its insurance company was good for the money. The posting of collateral in an escrow account makes the insured feel better but requires the insurer to have the cash to put into escrow.
At this moment, Greenberg now says, Sullivan should have shut down the credit default swap operation at AIGFP. “When the AAA credit rating disappeared in spring 2005, it would have been logical for AIG’s new management to have exited or reduced its business of writing credit default swaps,” he says. In its public disclosure AIG admitted that an extra degree of caution would be required in the wake of the rating downgrades. “[C]ounterparties may be unwilling to transact business with AIGFP except on a secured basis,” AIG disclosed in May 2005. “This could require AIGFP to post more collateral to counterparties in the future.”
Greenberg is able to tick off the litany of mistakes made at the time by Sullivan, Cassano and company. Rather than curtailing the sale of credit default swaps, AIGFP exponentially ratcheted up its issuance of them, and no longer just on corporate debt but on a whole new explosion of risk that Wall Street was madly underwriting through the issuance of
increasingly risky mortgage-backed securities tied to subprime and Alt-A mortgages. Then there were the credit default swaps written on collateralized debt obligations, or CDOs, with huge exposures to subprime mortgages. And then there were the swaps sold to hedge the risk of CDOs made up of CDOs, and that of synthetic CDOs, where just the risk was bought and sold and there was no underlying security. AIGFP also wrote insurance for something called “multisector CDOs,” a collection of more than 100 bonds backed by everything from residential mortgages to auto-loan receivables. In other words, a real dog’s breakfast of risk.
As of December 31, 2007, AIGFP had a portfolio of credit default swaps totaling $527 billion, of which $78 billion was written on multisector CDOs, most of which had some exposure to subprime mortgages. Indeed, it has become widely accepted that without Cassano and AIGFP around to insure the risk that Wall Street was taking in underwriting these increasingly squirrelly assets, the debt bubble might have run out of air far before it did in 2007. In its defense, which shows up, among other places, in an August 2009 court filing by law firm Weil, Gotshal & Manges in a shareholder lawsuit, AIG argues that it wrote credit default swaps on only the supersenior portion of the multisector CDOs — that is, those payments made first. Incredibly, AIGFP believed it had little real exposure “because the supersenior tranche has priority of payment ahead of even the AAA tranches, [and so] is regarded as having a better-than-AAA rating.”
Additionally, AIG and Weil Gotshal argue that in December 2005, AIGFP stopped writing new credit default swaps on multisector CDOs that included subprime mortgages, because of “concerns over deteriorating underwriting standards for subprime mortgages,” and “[a]s a result, AIGFP’s exposure to multisector CDOs with more risky subprime mortgages and mortgage-backed bonds originating in 2006 and 2007 was limited, a fact which later provided additional comfort that losses suffered by other market participants would not spread to AIG.”
While that last bit of legal spin is certainly debatable, what soon became crystal clear — and a major problem for AIG — was that as the value of the toxic securities it had insured before 2006 fell, AIG’s counterparties ratcheted up their demands for collateral. At the forefront of these increasingly strident collateral demands was none other than Goldman Sachs Group. Unlike every other Wall Street firm, in 2006, Goldman became nervous about the growing risks its traders perceived in the free-for-all that the market for mortgage-related securities had become. The traders’ concerns eventually bubbled up to the 30th floor of 85 Broad Street, where Goldman’s top executives had their offices, and in December 2006 the firm decided to “get closer to home,” in the words of CFO David Viniar. The firm started to make huge bets that the mortgage market would fall by shorting individual mortgage-backed securities and the newly created ABX index, which was a basket of mortgage-related securities, and by creating synthetic CDOs and selling the risk related to them to the likes of AIG.
Goldman also began vigorously writing down the value of the mortgage-related securities it held on its own books. Though this caused the firm to take large losses on the securities in early 2007, before its bets that the securities would fall really had the chance to pay off, Goldman was zealous about its daily mark-to-market exercises, in large part because of the firm’s mantra: “long-term greedy.” After December 2006, Goldman’s marks were increasingly aggressive on the low side and began to upset some of the big buyers of mortgage-backed securities, like the two Bear Stearns Cos. hedge funds run by Ralph Cioffi and Matthew Tannin, and forced them to mark down the value of the securities in their portfolios, ushering in a wave of catastrophic losses.
By April 2007, Goldman’s change in sentiment began to have serious implications for the rest of the market. The two Bear Stearns hedge funds traded with Goldman. One of the two funds, the Enhanced Leverage Fund, published its net asset value for the month of April 2007 as down 6.5 percent. That’s plenty bad, especially after months of increasing NAVs, but that was before the fund had received the month’s marks from Goldman, which came in a week after marks from other dealers had been received. When the marks from Goldman came in at 50 to 60 cents on the dollar — meaning that Goldman believed the securities were now worth that much in the market and had marked its own books the same way — rather than the 99 cents, 98 cents or 97 cents reported by other dealers, Cioffi and Tannin had no choice but to average Goldman’s marks with the others they had received and report that new number for April to investors: down a whopping 18.97 percent. Investors in the two Bear hedge funds sprinted for the exits. By June the two hedge funds had been closed, and by July they were liquidated. Investors collectively lost roughly $1.5 billion.
Bear’s hedge funds turned out to be the proverbial canary in the coal mine. Wall Street firms such as Lehman Brothers, Merrill Lynch and Bear Stearns itself were stuffed to the gills with many of the very same securities that Goldman was busy writing down and selling off at prices far below others’. Meanwhile, over at AIG — the firm that had insured about $75 billion of these securities through AIGFP — the displeasure with Goldman Sachs was becoming increasingly well known. Goldman’s lower marks forced AIG to grapple, albeit reluctantly, with its portfolio of insurance tied to the securities that Goldman kept insisting were worth less and less. “When the credit market seized up,” Sullivan later testified before Congress, “like many other financial institutions, we were forced to mark our swap positions at fire-sale prices, as if we owned the underlying bonds, even though we believed that our swap positions had value if held to maturity.”
By late July 2007, Goldman had begun putting pressure on AIG to pony up more collateral to cover the declining value of the mortgage securities AIG had insured. For instance, in an August 2 e-mail, AIG managing director Thomas Athan wrote AIG executive vice president Andrew Forster about a particularly difficult conversation he’d had that day with Ram Sundaram, a senior Goldman mortgage trader. “Tough conf call with Goldman,” Athan wrote Forster. “They are not budging and are acting irrational. They insist on ‘actionable firm bids and firm offers’ to come up with a ‘midmarket quotation’” as a way to determine how much collateral AIG needed to post.
Athan did not want to put the details of his call with Sundaram in the e-mail but informed Forster that the collateral issue needed to be quickly decided and that they needed to get back to Goldman by noon the next day. “I feel we need Joe to understand the situation 100 percent and let him decide how he wants to proceed,” Athan wrote. “I played almost every card I had, legal wording, market practice, intent of the language, meaning of the CSA” — a reference to the Credit Support Annex, a codicil to the International Swaps and Derivatives Association’s Master Agreement that governs collateral calls in credit default swap transactions — “and also stressed the potential damage to the relationship and GS said that this has gone to the ‘highest levels’ at GS and they feel that the CSA has to work or they cannot do synthetic trades anymore across the firm in these types of instruments. They called this a ‘test case’ many times on the call. It seems Ram has put himself in a bind that the firm is watching him here to see how he works this out and anything other than getting collateral close to liquidation levels will be considered a failure. Someone (like Joe) might need to convince a senior person that there is an alternative way to look at this situation.” Athan suggested they hold a conference call the next morning. “BTW,” he signed off, “This isn’t what I signed up [for]. Where are the big trades, high fives and celebratory closing dinners you promised?”
(The Wall Street Journal has reported that Athan and Forster, along with Cassano, are the targets of a U.S. government criminal probe.)
On an August 9 conference call, Cassano told AIG’s investors about AIGFP’s credit default swaps, “It is hard for us, and without being flippant, to even see a scenario within any kind of realm of reason that would see us losing $1 in any of those transactions.”
In reaction to Cassano’s assertion, on August 13 the Wall Street Journal reported on its front page that “AIG Might Be Deceiving Itself on Derivative Risks.” But Cassano pooh-poohed the story. “Hopefully people just ignore it,” he e-mailed Forster. “It is a real non-story.”
Of course, it wasn’t. The questions intensified about whether AIGFP had the cash to meet mounting collateral calls from counterparties. The rating agencies wanted to know, as did AIG’s outside auditors at PricewaterhouseCoopers and AIG’s internal auditors at 70 Pine. One of them, Elias Habayeb, the CFO of AIG’s financial services group, who reported to AIG’s CFO, was particularly nettlesome to Cassano. For months he had been pushing Cassano for more information about the value of the swaps AIGFP had written and about the collateral availability. “There must be something in the air or the coffee at 70 Pine,” Cassano complained in an August 31 e-mail to Habayeb about the numerous inquiries he was suddenly getting from headquarters. He then went on to detail AIGFP’s sources of liquidity and urged Habayeb not to bother Forster and others with further questions. “We have taken many steps to insure our liquidity during this market disruption,” he wrote.
Inside AIGFP there were also concerns raised about the unit’s ability to provide the necessary collateral. Joseph St. Denis, AIGFP’s vice president for accounting policy from June 2006 and a former assistant chief accountant at the SEC’s division of enforcement, got worried when he came back from vacation in early September 2007 and heard about a multibillion-dollar margin call on supersenior swaps that AIG had written. “I was gravely concerned about this,” he said in congressional testimony after the bailout, “as the mantra at AIGFP had always been (in my experience) that there could never be losses” on the supersenior credit default swaps. He was uneasy about whether AIGFP could be in a “potentially material liability position” as a result of the increasing collateral demands. After St. Denis resigned on October 1, 2007, he said in congressional testimony that he told AIGFP’s general counsel he “had lost faith in the senior-most management of AIGFP and could not accept the risk to AIG and myself of being isolated from corporate accounting policy personnel, especially given the situation” with the supersenior credit default swaps.
Much to Cassano’s consternation, Goldman wasn’t going away either. Despite the early August debate between Sundaram and Forster, Goldman asked AIG to post $1.5 billion in collateral on its $23 billion of credit default swap exposure to the insurer. Some time later AIG put up $450 million. In October, Goldman asked for an additional $3 billion in collateral; AIG put up $1.5 billion. By the end of November 2007, Forster had catalogued the extent of AIG’s collateral disputes with nine banks and Wall Street firms in an eight-page memorandum. He wrote that because of the “extreme illiquidity” in the market — acknowledged by all the counterparties, he contended — the discussions had been “friendly” rather than “disputed,” although that assumption seems inconsistent with the August confrontation with Goldman. “All of the dealers feel that as the market is under extreme stress that prices should perhaps be lower but none have any real idea as to how to best calculate the price or if indeed that statement is true,” Forster wrote. “The market is so illiquid that there are no willing takers of risk currently so valuations are simply best guesses and there is no two-way market in any sense of the term.”
Despite the caveat, Forster’s analysis of the trades AIGFP had with each firm was quite specific, as were the value of the securities involved. With Merrill Lynch, he noted, AIGFP had close to $10 billion in exposure, in 20 different trades. He noted there was an 8 percent “price threshold” — meaning the price would need to fall to 92 cents on the dollar — “before any posting [of collateral] is required,” based on the underlying bond price, not the value of the credit default swap. As of the end of November, Merrill was asking for $610 million in collateral, but, Forster pointed out, “we are disputing the call with them and they agree prices are too illiquid to be reliable.”
With $23 billion of exposure, Goldman was the Big Kahuna for AIGFP: AIG’s counterparty on 51 different positions, in 33 trades. Goldman had also negotiated more-favorable terms than its competitors, whereby AIGFP had to start posting collateral mostly when the value of the underlying securities fell to 96 cents on the dollar. “They have made collateral calls totaling $3 [billion] on 38 positions covering 23 different transactions,” Forster wrote. He then listed the 38 positions and Goldman’s discounted prices for them. There was no mention, as there had been with Merrill, of any disagreement or potential dispute resolution with Goldman. Forster did take note of the price discrepancy of one CDO — Independence V — which Merrill had underwritten in 2004. As of November 2007, Merrill valued Independence V at 90 cents on the dollar, while Goldman valued it at 67.5 cents. Cassano sent Forster’s analysis to William Dooley, the head of the AIG division under which AIGFP operated, with the note that despite the disputes, the dealers were working with AIGFP in a “positive framework toward seeking resolution.”
Greenberg tells II, “Goldman had the lowest marks on the Street, by everything I hear. There was no exchange [of views]. Where was the price discovery?” AIG board members and executives also viewed Goldman’s collateral requests with a high degree of skepticism. Some thought it was nothing more than a way for Goldman to get “a free loan from AIG,” according to one AIG executive. Goldman would announce that it had marked its AIG swap book down by, say, $5 billion and then insist, “Send us a check for $5 billion.”
In short order, Cassano’s house of cards collapsed. On November 7, AIG reported third-quarter net income of a little more than $3 billion but revealed deep in the announcement that AIGFP had suffered a “$352 million unrealized market valuation loss” on its supersenior swap portfolio and that October’s loss alone on that portfolio could be an additional “$550 million” pretax. The next day Cassano told investors that there was “opacity in the market” and that valuations for the securities underlying the swaps were all over the map. But “rest assured,” Cassano said, “we have plenty of resources and more than enough resources to meet any of the collateral calls that might come in.”
On November 29, Timothy Ryan, a Pricewaterhouse partner, told Sullivan in an e-mail that in “light of AlG’s plans to hold [an] investor conference on December 5,” Pricewaterhouse believed there was a growing possibility of a “material weakness” in AIGFP’s credit default swap portfolio that could result in potentially huge future losses. Such a warning needed to be disclosed to investors, according to SEC rules, but AIG did not disclose the warning for months. In the days leading up to the December 5 call, Cassano was wondering whether he could or should open up the AIGFP black box to investors. “As you know the exposure we have within these transactions is a mark to market, potential replacement cost,” he wrote in a November 28 e-mail to his colleague Robert Lewis, a risk officer. “[A]ll the entities that we have swaps with are highly rated (triple A primarily) and in an event of default or non payment we generally become pari passu with the bond holders. Once again an extremely remote risk with little real exposure.”
On December 5, still in denial, Martin Sullivan and Joe Cassano gave Oscar-worthy performances on the investor conference call regarding AIGFP’s credit default swap portfolio — and made no mention of Pricewaterhouse’s warning comment. “AIG has accurately identified all areas of exposure to the U.S. residential housing market,” Sullivan said that day. “We are confident in our markets and the reasonableness of our valuation methods.” Added Cassano, in yet another instance of stunning understatement, “It is very difficult to see how there can be any losses in these portfolios.” Cassano made no mention of the disputes AIGFP was having with Goldman, and presumably some of the other counterparties, about the posting of collateral. “We have, from time to time, gotten collateral calls from people,” Cassano said, according to the transcript. “Then we say to them, ‘Well, we don’t agree with your numbers.’ And they go, ‘Oh.’ And they go away.”
A week later, after a collateral call from Calyon, the big French bank, AIG’s Athan wrote to a colleague, “We are in uncharted waters for our firm.”
By early 2008, AIGFP executives were busy scrambling to make sure they had a response to Pricewaterhouse’s growing concerns. At the January 15 audit committee meeting, there was much discussion about whether proper controls existed to monitor independently the valuation of the AIGFP portfolio. According to minutes from the meeting: “Mr. Habayeb believes that he is limited in his ability to influence changes, and the supersenior valuation process is not going as smoothly as it could. Mr. Ryan said that the control functions are not included in the ongoing process and lose the ability to participate in discussions of the issues. He added that roles and responsibilities need to be clarified, and pointed out that collateral issues could have been escalated to the AIG level earlier in the process.”
One way to better estimate the risks in the AIGFP portfolio was to tweak the computer models related to the swaps. Gary Gorton, then a finance professor at the Wharton School of the University of Pennsylvania and now at the Yale School of Management, created the computer models used by AIGFP. (His extensive résumé, available online, makes no mention of his work for AIG.) On January 21, 2008, Kevin McGinn, AIG’s chief credit officer, wrote Gorton with his concerns about whether his models accounted for the risks to AIGFP’s supersenior portfolio, especially as the rating agencies were busy downgrading many of the mortgage-backed securities — once rated AAA — that AIGFP had insured. “[T]he agencies are hardly through with their downgrades, even for 05 collateral, as S&P has signaled,” McGinn wrote Gorton. By February 6 the case Tim Ryan had made to the audit committee about the potential for a “material weakness” at AIG had trickled down to Cassano. Five days later AIG announced that as of December 31, 2007, Pricewaterhouse believed AIG “had a material weakness in its internal control over financial reporting and oversight relating to the fair value valuation of the AIGFP supersenior credit default swap portfolio.” That announcement, naturally, increased the pressure from counterparties and justified, according to Cassano, “new calls from our counterpart[ie]s stating that they can no longer accept our pricing methodology,” which had “weakened our negotiation position as to collateral calls.” On February 28, AIG released its full-year 2007 financial results and explained that those numbers included a charge of $11.47 billion related to unrealized losses from the AIGFP swap portfolio. On March 1, AIG announced Cassano’s departure from the firm “with our concurrence,” in Sullivan’s words. Cassano was allowed to keep up to $34 million in unvested bonuses after his departure. He remained a consultant to AIG, at a cost of $1 million per month.
Ironically, on August 18, a week or so after AIG announced more bad earnings, Goldman’s research ax fell on AIG. Banking analyst Thomas Cholnoky issued a report with the advice “Don’t Buy AIG,” citing the risks to shareholders from “likely” further rating agency downgrades and capital-raising activities that would dilute shareholders. Among the reasons Cholnoky cited for his report was the potential for an increase in collateral calls. He made no mention of the fact that Goldman itself was leading this charge and had been for about a year. (He now says he did not know.) “The bottom line: Large scale cash outflows and posting of collateral could substantially weaken AIG’s balance sheet,” Cholnoky wrote. “We believe that the rating agencies would force AIG to raise a large, dilutive amount of equity capital to: (1) plug the holes left by such cash outflows, and (2) prevent significant downgrades to avoid any further triggering of collateral calls and termination payments.”
The final nail in AIG’s coffin came from another questionable decision made under Sullivan’s leadership: taking the cash generated when institutional investors, seeking to sell securities short, borrowed stock from AIG’s massive $800 billion investment portfolio and investing it in what turned out to be high-risk mortgage securities. “There’s no sense lending [the securities] to [other dealers] unless you’re going to take the cash and invest it someplace and earn a spread,” explains one AIG former executive. “That’s what a AAA rating is for.”
But what should have been invested in liquid, low-risk securities like Treasury bonds went instead into the illiquid mortgage market. At one point, AIG had converted some $60 billion of the cash it had received from other dealers into mortgage-related securities. “It scared the shit out of me,” says one AIG executive. “My first reaction was, What happens if all those guys come back and say, ‘We want our money back. Here’s your damn securities—give us our cash back.’ All the cash had been put into securities that are now 25, 30 percent underwater. If they were government securities, we could turn around, sell them tomorrow and give them their money back. Well, we had securities that you couldn’t sell to anybody.” By the late summer and early fall of 2008, borrowed stock was flooding back to AIG and investors were asking for their cash back, further exacerbating AIG’s cash crunch. Buried on the last page of Cholnoky’s August 2008 report was his concern about AIG’s securities lending business. “[D]ue to AIG’s aggressive investment strategy into riskier classes, the current market value of the assets stood at $59.5 billion as compared with liabilities of $75.1 billion,” he wrote, noting that AIG had agreed to post collateral to make up for these losses.
Cholnoky’s report broke the camel’s back. (He left Goldman after the financial crisis and now works at Transatlantic Holdings, a reinsurance company.) “The Goldman report comes out, and there was a lot more pressure on the company,” explains a former AIG executive. “It’s a confluence of events.” It was the equivalent of a run on a bank. After September 16, 2008, when the federal government poured $85 billion of taxpayer money into AIG to keep it from falling into bankruptcy as Lehman Brothers had only hours earlier, and in effect took over the company, AIG eventually turned around and paid out $62.1 billion to 16 counterparties to fulfill its collateral obligations related to credit default swaps AIGFP had sold. Second on the list, according to TARP Special Inspector General Neil Barofsky’s November 17, 2009, report about AIG’s counterparties, was Goldman Sachs, which received $14 billion, or everything AIG owed it at 100 cents on the dollar. (Société Générale, a French bank, came in first with $16.5 billion, some of which was then paid over to Goldman, according to the New York Times.) There has never been a good explanation of why the money was paid out or why the Fed paid the counterparties 100 cents on the dollar when they surely would have received much less, if anything, had AIG filed for bankruptcy. Indeed, the Federal Reserve Bank of New York unsuccessfully tried to negotiate discounts, known as haircuts, with the counterparties. “Seven of the eight counterparties told [the New York Fed] that they would not voluntarily agree to a haircut,” Barofsky wrote in his November 2009 report. “The eighth counterparty, UBS, said it would accept a haircut of 2 percent as long as other counterparties also granted a similar concession to [the New York Fed].”
Furthermore, according to a September 2009 report from the U.S. Government Accountability Office, “AIG’s securities lending program continued to be one of the greatest ongoing demands on its liquidity” even after the September 16 bailout. As a result, the Fed created a vehicle dubbed Maiden Lane II and funded it with $19.5 billion to purchase from AIG the most troubled residential-mortgage-backed securities it had bought with the cash received from the securities lending program. As of September 2009, AIG owed $17.1 billion on this credit facility.
Reflecting back on all the carnage, Hank Greenberg knows exactly what he would have done if he had still been AIG’s CEO instead of Martin Sullivan. First, he would have stopped writing credit default swaps the moment S&P downgraded the company: “It was an ancillary business.” And what about Goldman Sachs, a firm for which he once had great respect and admiration, and its ongoing collateral calls? “Goldman had the lowest marks on the CDOs on the Street,” he says. “And they kept on calling for more collateral. I’d have told them to stick it. Period. You want more collateral? We’re not going to give it to you. You don’t like it? Sue us. In five years we’ll find out who was right.”
William D. Cohan is a former senior Wall Street investment banker and the bestselling author of House of Cards: A Tale of Hubris and Wretched Excess on Wall Street and The Last Tycoons: The Secret History of Lazard Frères & Co.