Infectious Greed

by Frank Partnoy

  When Johnsen told various banks Gibson might be interested in a swap, he was besieged by proposals, in the same way a homeowner who indicated an interest in refinancing would be bombarded by mailings and early-evening calls. Banks, having tasted the profits from equity derivatives and other similar transactions, were aggressively seeking new customers. Gibson was a perfect candidate for a new custom-tailored swap. It was big enough to do a deal, but not sophisticated enough to understand how to evaluate it properly.

  Gibson had an ongoing relationship with Bankers Trust, run by two senior commercial bankers. But the two salesmen who flew out to see Johnsen in November 1991 were much younger and much more sophisticated than these commercial bankers. Two of Charlie Sanford’s nerds—Gary Missner and Mitchell Vazquez—made just the right pitch to Gibson. Jim Johnsen remembered the salesmen as “forthright and honest, not like hard-sell securities salesmen who call you up at home during dinner.”45 On November 12, 1991, Gibson agreed to do two interest-rate swaps with Bankers Trust, each for $30 million.46

  Although these two swaps initially appeared to be plain vanilla, they actually were a bit more complicated. The first was a “two-year” swap, in which Gibson agreed to pay a low fixed rate of 5.91 percent and receive a floating rate. The second was a “five-year” swap, in which Gibson agreed to do the opposite: pay a floating rate and receive a higher fixed rate of 7.12 percent.47

  The effect was to create two different payment periods for Gibson. For the first two years, Gibson would pay a low rate (5.91 percent). For the three years after that, it would pay a floating rate. Bankers Trust would hedge its position in the market, so that it pocketed a fee, but did not put its capital at risk.

  This trade had what the dealers called attractive optics. It looked simply marvelous. Gibson reduced its interest cost by several percent for the next two years. It took on some risks in years three through five, but rates might be low then, too. Who wouldn’t do this trade?

  In fact, this was a fairly common trade at the time,48 and it turned out to be a winner. Interest rates declined during early 1992 and the swaps moved in Gibson’s favor. Seven months later, in July 1992, Johnsen got a call from Bankers Trust with good news: Gibson could close out its positions right away, at a profit of $260,000. What would he like to do?

  What would you do, in Johnsen’s situation? Could you determine whether $260,000 was a fair payment? One way to do it would be to create a computer model, and input various data and assumptions regarding interest rates, credit risk, and other variables to come up with a valuation. But few corporate treasurers—including Johnsen—were capable of this in 1991. He could have called another bank for a second opinion. But that risked offending Bankers Trust, and he wouldn’t necessarily get a better price. Besides, as long-standing advisers to Gibson, Bankers Trust wouldn’t fudge the valuation of a couple of simple interest-rate swaps, would they?

  Johnsen should have made a few calls, but he apparently didn’t think it was necessary. He decided to trust Bankers Trust and closed out the swaps for a $260,000 gain. That decision would cost Gibson dearly. The actual value of the swaps to Gibson at the time has been disputed, but it was at least $550,000 and probably closer to the amount Gibson ultimately claimed in a lawsuit: $750,000.49 The difference of several hundred thousand dollars was gravy for Bankers Trust.

  Gibson’s willingness to close out the swaps at such a bad price meant that Gary Missner and Mitchell Vazquez had found a sucker. If Gibson couldn’t evaluate a reasonably simple trade, it would have no chance with a complex one, and that meant Bankers Trust could charge an even higher fee for a more complex deal. Missner and Vazquez were doing exactly what Charlie Sanford had indicated he wanted: generating big profits, with little capital at risk.

  From November 1991 to March 1994, the nerds of Bankers Trust lobbed salvo after salvo of complex swap proposals at the greeting-card maker. Many of them hit, and before it was all over, Gibson and Bankers Trust would enter into twenty-nine derivatives transactions, in which Gibson placed huge interest-rate bets, and Bankers Trust booked huge profits.

  Gibson’s second swap, in October 1992, made the first one look like a plain U.S. savings bond. In this deal, also for $30 million, Gibson agreed to receive a fixed rate of 5.5 percent and to pay a floating rate, squared and then divided by 6 percent.50 That’s right: squared, as in “to the second power.” The swap payments would be based on the London Interbank Offered Rate, known as LIBOR, multiplied by itself and then divided by 6 percent. Let me repeat: Gibson’s payments would be LIBOR multiplied by itself and then divided by 6 percent.

  Why on earth would anyone buy a LIBOR-squared swap? One possible reason was that they didn’t know it was a LIBOR-squared swap, or that they didn’t understand what squared meant. But that hardly seemed plausible. Jim Johnsen knew plenty of math to understand that, and surely he had been able to spot the little “2” in the numerator.

  No, the more likely reason was greed: the squared feature was a way to bet big on declining interest rates. If floating rates remained low, as they had been, Johnsen could turn his little treasury operation into a profit center. For example, if floating rates stayed at 3 percent, then Gibson would be obligated to pay just one-and-a-half percent (the math was 3 times 3 divided by 6). But if rates increased, losses would increase, well, exponentially. If rates hit 6 percent, Gibson would owe 6 percent. If they hit 10 percent . . . well, no one wanted to think about that. Moreover, even a small increase in rates in the short run—to, say, 3.5 percent—would create big losses for Gibson, because it would increase the chances that, during the term of the swap, rates would go higher, causing Gibson exponential losses.

  The squared feature magnified the bet’s risks and potential returns, much as a trifecta bet—which names, in order, the top three horses to finish a race—magnifies the risks and potential returns relative to a simple win, place, or show bet. Treasurers bought squared swaps for the same reason gamblers played trifectas. (Gibson wasn’t alone in gambling: one former Bankers Trust salesman even claimed that a client had bought a LIBOR-cubed swap—raised to the third power—although he wouldn’t say which client it was.)51

  Thanks to ISDA’s lobbying against regulation of swaps, Gibson could make this bet in secret, in an unregulated market. A swap that was being used to hedge debt—as Gibson might argue its loony swap was—did not need to be disclosed to shareholders at all. But even if the swap wasn’t obviously a hedge, the disclosure requirements were minimal. In any event, Gibson couldn’t possibly tell its shareholders about such a bizarre trade, and it did not.

  Unfortunately, rates began rising during the final months of 1992, and even a fractional increase was enough to cause Gibson some serious pain. LIBOR rose from 3.0625 percent to 3.375 percent, and by the end of the year—with that increase squared—Johnsen was facing a loss of $975,000.52 Or at least that was what the salesmen at Bankers Trust told him.

  Gibson wasn’t capable of discerning the swap’s actual value, but that didn’t matter much. Gibson wasn’t about to walk away from the gambling tables with a million-dollar loss. During the next fourteen months, Gibson did deals with Bankers Trust that it described as “even more volatile and risky than the ‘Libor Squared’ transaction.”53 These were the financial markets’ version of the superfecta, a horse-racing bet like the trifecta, but with all of the top four horses.

  In a so-called Treasury-Linked Swap, Gibson agreed, among other things, to receive the lesser of $30.6 million or an amount determined by the following formula: $30 million-(103 ✕ 2-year Treasury Yield/ 4.88%)-(30-year Treasury Price/100). Another trade, called a knock-out option, became worthless when rates hit a specified level (it “knocked out” at that point). Financial economists would spend the next several years attempting to create pricing models for knock-out options. Still another Gibson trade, called a Wedding Band, was too complex to describe without a mathematician and a psychotherapist.

  Gibson’s swap smorgasbord included the most sophistic
ated derivatives anyone had invented, and they all were hidden from regulators and shareholders. In its 1993 annual report, Gibson disclosed $96 million of swaps outstanding, without any detail as to their risks.

  These swaps were incredibly profitable for Bankers Trust. From late 1991 to early 1994, the more Gibson lost the more it traded, and the more Bankers Trust made. In all, Bankers Trust made about $13 million from the swaps with Gibson,54 all of which supposedly began as an effort to find a low-cost hedge for a simple fixed-rate debt.

  How did Charlie Sanford react to all of this aggressive behavior by his salesmen? Substantial bonuses, of course. In late September 1993, Sanford proudly announced in a memorandum that he was promoting Gary Missner—the senior salesman covering Gibson—to the position of managing director.55

  Gibson Greetings was not alone. Numerous Bankers Trust clients were actively buying and selling similar swaps: Air Products and Chemicals, Equity Group Holdings, Federal Paper Board, Jefferson Smurfit, Sandoz, Sequa, and others, including two Indonesian companies—PT Adimitra Rayapratama and PT Dharmala Sakti Sejahtera—that made some of the biggest trades of all. Bankers Trust was everywhere. As David Urbani, assistant treasurer at Air Products and Chemicals, put it, “If I mention to another company that we were thinking about doing some weird derivative, they’ll say you must have been talking to Bankers Trust.”56

  The most publicized losses on swaps with Bankers Trust were on those done by Procter & Gamble. An investor in P&G stock—like an investor in Gibson Greetings—had no clue the firm was gambling on derivatives. Shareholders thought they were buying a soap company. The company’s 1993 Form 10-K filing with the Securities and Exchange Commission reinforced this notion, describing its derivatives activity as minimal and low risk.

  Kevin Hudson, another smooth-talking Bankers Trust salesman, covered P&G. In the early 1990s, Bankers Trust and P&G had done little business together. Other than an early 1993 deal—linked to the Mexican peso—Bankers Trust had made few inroads with P&G’s treasurer, Raymond Mains.57

  In October 1993, Hudson began talking to Dane Parker, a junior treasury analyst who reported to Mains. P&G had been borrowing money in the short-term commercial-paper market to fund its consumer products business. In that market, there is a floating-rate index, like LIBOR, called the Commercial Paper index, or CP for short.

  A well-respected company such as P&G could borrow at roughly the Commercial Paper index rate. But P&G’s goal was to use swaps to borrow at less than the Commercial Paper index, hopefully by as much as 40 hundredths of a percent, or 40 basis points. In the industry parlance, Parker might call Hudson and say, “I want CP minus 40 basis points.”

  In the commercial-paper world, every basis point was precious. If P&G’s treasury could outperform its peers by 40 of them, Mains and Parker would be supermen. To do it, they would need to take on some risk, of course, and Bankers Trust had just the recipe: a complex trade resembling Gibson’s Treasury-Linked Swap.

  It seemed incredible that a big, sophisticated company like P&G would pay Gibson-like fees to gamble on interest rates. But early on, Parker demonstrated that he was potentially just as much of a sitting duck as Johnsen had been. Soon, Hudson was regaling his colleagues with stories about how “smooth” he had been during his visits with P&G. In one taped conversation, Hudson described the deal he had proposed to Parker to a friend, and bragged, “You’re looking at an $8 million trade.” In other words, he had estimated that Bankers Trust’s profit from this one deal would be $8 million. The friend was impressed and responded, “It’s like our greatest fantasy.” Hudson couldn’t help but agree. “I know. It is. It is. This is a wet dream.”58

  Charlie Sanford’s pressures were having their intended effect. His nerds were aggressively pursuing deals, and—because their customers couldn’t evaluate complex swaps properly—Bankers Trust was making unheard-of profits. During this time, a fee of one percent, called a “point,” was substantial. But Bankers Trust was earning multiple points on deals of unprecedented size. The companies were no longer clients of Bankers Trust; they were pigeons.

  As one former managing director put it, “Guys started making jokes on the trading floor about how they were hammering the customers. They were giving each other high fives. A junior person would turn to his senior guy and say, ‘I can get [this customer] for all these points.’ The senior guys would say, ‘Yeah, ream him.’”59

  During this period, Bankers Trust developed a training video for new employees. In one segment of the video—which Bankers Trust officials have maintained was in jest—a bank employee described a hypothetical derivative transaction among Sony, IBM, and Bankers Trust: “What Bankers Trust can do for Sony and IBM is get in the middle and rip them off—take a little money.” The employee then nervously added, “Let me take that back. I just realized that I’m being filmed.”60

  Belita Ong, a former managing director and senior derivatives saleswoman at Bankers Trust, believed that Bankers Trust had developed “an amoral culture.” She said, “You saw practices that you knew were not good for clients being encouraged by senior managers because they made a lot of money for the bank.”61 One salesman noted, “Funny business, you know. Lure people into that calm and then just totally fuck ’em.”62

  Ong was the managing director who had called the Bankers Trust nerds “mostly guys without girlfriends.” But sweet-talking Kevin Hudson, the salesman covering P&G, had a girlfriend—a fiancée even—another salesperson at Bankers Trust named Alison Bernhard. And he was telling her all of the details about how Bankers Trust was milking P&G, on taped telephone conversations at the bank.

  When Hudson finally persuaded P&G to do a swap with Bankers Trust, he immediately called Bernhard to tell her about it. “I just took the biggest trade of the year,” he told his fiancée.63

  In the swap, P&G would pay the CP rate minus 75 basis points on $200 million, giving it a cushion so it could lose a bit and still lock in the goal of “CP minus 40.” In addition, P&G would effectively sell some put options to Bankers Trust, giving the bank the right to profit if interest rates increased. If interest rates increased, P&G would pay the following spread:64 [98.5 × (5-year Treasury Yield /5.78%) - 30-year Treasury Price] / 100

  Don’t stop reading; you don’t need to understand precisely how to value this spread to understand what it did. There were two variables in the formula: the 5-year Treasury Yield (the interest rate on 5-year government bonds) and the 30-year Treasury Price (the price of 30-year government bonds). Yields and prices moved in opposite directions: in simple terms, the more cheaply you could buy a Treasury bond, the more yield you would earn over time. The spread formula consisted of a multiple of the 5-year Treasury Yield minus the 30-year Treasury Price, but it just as easily could have added a multiple of the 30-year Treasury Yield.

  Why use the 30-year Treasury Price instead of the 30-year Treasury Yield? And why multiply the 5-year Treasury Yield by 98.5 and then divide—first by 5.78%, and then, at the end, by 100? As in the Gibson Greetings swaps, these terms were there for what traders called optics: they made the spread look more attractive than it really was. For example, 98.5 and 5.78 were roughly the price and yield, respectively, of a 5-year Treasury bond. They focused P&G on the current market values of the bond instead of on the risks of the trade. In reality, these numbers were meaningless: a savvy investor would simply do the math and notice that 98.5 divided by 5.78% divided by 100 was about 17.04, meaning that the spread actually was magnified—or leveraged—about 17 times. The same was true of the use of the 30-year Treasury Price instead of the 30-year Treasury Yield. A savvy investor would convert the price to a yield, so as to compare apples to apples within the spread formula. Because the price was much higher than the yield, the 30-year Treasury variable also was leveraged, by a similar amount.

  In other words, P&G’s $200 million swap with Bankers Trust really was a $3.4 billion bet that 5-year and 30-year interest rates would remain low. If 5-year rates stayed at around 5.
78 percent and the 30-year price remained at its current level, the spread would be close to zero. But if interest rates increased, watch out.

  It is unclear whether anyone at P&G did this math. Former P&G chairman Edwin Artzt called his underlings who bought the swap “farm boys at a country carnival” and derided their inability to seek the advice necessary to understand the deal. In discussing two of these employees, Carol Loomis, a veteran reporter for Fortune, concluded that “plainly, neither understood the derivatives they bought from Hudson.”65

  P&G certainly didn’t seem sophisticated to anyone at Bankers Trust. When the profits were tallied, Kevin Hudson and Bankers Trust had made $7.6 million from this one trade. When Hudson told his boss, Jack Lavin, about the trade, Lavin reportedly said, “I think my dick just fell off.”66

  His fiancée had a different reaction, in a taped phone call:BERNHARD: Oh, my ever-loving God. Do they understand that . . . what they did?

  HUDSON: No. They understand what they did, but they don’t understand the leverage, no.

  BERNHARD: They would never know. They would never be able to know how much money was taken out of that.

  HUDSON: Never, no way, no way. That’s the beauty of Bankers Trust.67

  In late January 1994, P&G did another swap with Bankers Trust, with a similarly complex formula, this time betting that German interest rates would remain low. This trade was only leveraged ten times. In one year, Kevin Hudson had done deals with more than $25 million in fees. His fiancée was beginning to worry.