Infectious Greed

by Frank Partnoy

  The on-the-run/off-the-run trade was easy to do. It wasn’t complicated, and it didn’t even require a computer. Traders at many investment banks made this bet, and although Meriwether was widely credited with doing the trade in substantial size, he didn’t discover it. In fact, the on-the-run /off-the-run trade has been very common during the past two decades, and Meriwether wasn’t even the first person to do the trade at Salomon.12

  Meriwether’s strategies became more sophisticated in 1984, when he met Eric Rosenfeld, then a professor at the Harvard Business School. In his study of options, Rosenfeld had discovered some flaws in the assumptions of the Black-Scholes option-pricing model, which traders at both Bankers Trust and CSFP had used. For example, the model assumed the volatility of underlying financial assets was constant. It also assumed there were no costs of transacting and no market discontinuities (in other words, no kinks). Most important, it assumed that the distribution of returns on assets was bell-shaped, like the normal distribution of grades in the courses Rosenfeld taught. This key assumption—the gospel among most financial economists—was the source of the random walk hypothesis, which said that the movements of the market up and down were essentially random.

  However, to anyone with experience in the options markets, these assumptions were demonstrably false. Volatility changed over time, options were expensive to buy and sell, and there were kinks throughout the markets, where the supply and demand of particular options were unbalanced. Moreover, returns were not normally distributed; there were periods of manias and dramatic crashes, which did not fit a bell-shaped curve. Asset prices followed a random walk some of the time, but not all of the time.

  Eventually, financial economists would abandon these assumptions and seek more-nuanced models of price behavior. Even Burton Malkiel, the well-respected economist and author of A Random Walk Down Wall Street, would conclude in a revised edition of his book that “More recent work, however, indicated that the random-walk model does not strictly hold.”13 But during the late 1980s and early 1990s, economists were not focused on assumptions, and traders using flawed models routinely misvalued options.

  Meriwether recognized that options were more complicated than futures. If Salomon could approach options pricing in a more sophisticated way than its competitors, it might obtain a sustainable advantage. He seized the opportunity by hiring Rosenfeld and several other academics who understood the flawed assumptions of the models. He even brought on Myron Scholes and Robert Merton as consultants. Scholes developed new models, and helped to create Salomon’s AAA-rated derivatives subsidiary, known as Salomon Swapco.14 Merton, who had been Rosenfeld’s mentor, offered big-picture advice about options arbitrage.

  Meriwether also hired Victor Haghani, Gregory Hawkins, and Larry Hilibrand (the aggressive trader who would receive a $23 million bonus in 1990). These three H’s became the Arbitrage Group’s most profitable traders. Haghani had a master’s degree in finance from the London School of Economics and found arbitrage opportunities in Japan, especially in convertible bonds. (Haghani had done research for Andy Krieger when Krieger was trading currency options at Salomon a few years earlier.) The other two were Ph.D.s from MIT:15 Hawkins was an expert in mortgages (more about them later); Hilibrand was an expert in everything. These were the people whose 1990 bonuses would enrage Mozer and other employees outside the Arbitrage Group; all of these people were closely bound to Meriwether, and they became a family.

  Meriwether’s “quants” put the nerds of Bankers Trust to shame. He gave them a special place in the middle of the trading floor at Salomon, where they stood like geeks in the middle of a high school party. This group of traders was known as the best finance faculty in the world.

  When currency options were introduced, Meriwether’s group began trading them at the same time Andy Krieger was trading currency options at Bankers Trust. Like Krieger, Meriwether’s traders looked for options that, according to the group’s computer models, were mispriced relative to economically equivalent portfolios. As options traders at other firms were using sledgehammers to bludgeon each other with one-way directional bets and misdirection strategies, Meriwether and Rosenfeld were using a scalpel to dissect minor mispricings in dozens of options markets throughout the world.

  The Arbitrage Group designed the first pricing systems using a highspeed network of personal computers, while other traders at Salomon were stuck waiting in line to use the firm’s one, relatively slow, mainframe. 16 Meriwether set up trading desks in London and Tokyo with similar analytical firepower, to exploit opportunities in options markets abroad.17

  Victor Haghani found arbitrage opportunities in the Japanese convertible bond market. The owner of a convertible bond has the option to convert the bond into a specified number of shares of stock, and effectively has the choice between two investments—bond or stock—depending on the price of the stock. In other words, a convertible bond can be thought of as two investments: a bond plus an option to buy stock. When a company’s stock price is low, a convertible bond is really just a plain-vanilla bond with fixed interest payments; the stock-option component has little or no value. But as the stock price rises, the stock-option component of the convertible bond becomes more valuable. When the stock price is high enough, the holder of the convertible bond is better off converting the bond into shares. At that point, the bond portion of the convertible bond evaporates, and the holder simply owns stock.

  The Japanese government restricted the amount of stock companies could issue, in an attempt to limit supply and keep share prices high. There were no similar restrictions on issuing convertible bonds. However, few investors were buying Japanese convertible bonds—in part because they didn’t like buying both the bond and stock components—and they were relatively cheap. That meant the options embedded in the bonds also were cheap—cheaper than they should have been in an efficient market. The legal rules restricting stock issuance in Japan, and the reluctance of investors to buy convertible bonds, had created an arbitrage opportunity. If Haghani could somehow buy just the cheap stock-option component of the convertible bonds and then sell similar stock options to investors at a higher price, he could make virtually risk-free profits.

  Haghani bought the convertible bonds, so that he effectively owned a bond plus a stock option. He then eliminated the interest-rate risks associated with the bond component of the convertible bonds, using interest-rate swaps, so that he remained exposed only to the stock-option component of the convertible bond. Then, in the over-the-counter market, Haghani sold new stock options to investors. These new stock options were designed to mimic the risk that Haghani faced in the stock-option component of the convertible bond. In other words, Haghani would simply pass along his remaining risks to investors. The beauty of this deal was that, because the convertible bonds were cheap, Haghani could buy the stock-option component for less than investors were willing to pay for the new stock options. Effectively, Haghani bought a convertible bond made of two pieces, and then sold the pieces for more than the whole—like buying a cheese sandwich for $3, and then selling the cheese and bread for $2 each.

  Haghani made hundreds of millions of dollars for Salomon using this strategy. In 1992, when he was just 30 years old, Salomon paid him a $25 million bonus as a reward for finding these trades.18

  The Arbitrage Group did similar trades when the options embedded in bonds were misvalued because companies did not understand them. For example, when short-term interest rates were falling during the late 1980s, many corporate treasurers recalled high interest rates during the late Carter and early Reagan administrations, and wanted to lock in lower rates. They bought interest-rate caps—options that made money if short-term interest rates rose. But the treasurers didn’t have good models for evaluating the options, and, as a result, Salomon was able to make money selling overvalued short-term options.

  At the same time, companies were issuing long-term callable bonds, which they could redeem, if interest rates declined, by returning the in
vestor’s principal investment. The coupon payments of these bonds—the semi-annual interest payments investors received—were higher than those of non-callable bonds. In effect, the companies issuing the bonds were purchasing options from investors, paying them above-normal coupons instead of an upfront premium. But companies didn’t necessarily want to buy these long-term options, any more than they had wanted to be exposed to an increase in short-term interest rates. Instead of keeping the risk associated with these long-term options, the companies sold offsetting options to Salomon. Again, the companies did a poor job of assessing the value of these options, and Salomon was able to make money buying undervalued long-term options.

  Salomon didn’t want to keep the risk associated with the short-term options they had sold and the long-term options they had bought. Instead, traders sought to buy short-term options and sell long-term options to offset their risks; even if those options were fairly priced, Salomon could pocket the difference between buying low and selling high. Sometimes, Salomon’s traders found counterparties willing to take on these risks in the over-the-counter market, just as Victor Haghani had done. When traders could not find anyone willing to take on the offsetting risk, they bought and sold options on Treasury bonds that carried roughly the same risk as their options on corporate bonds. Managing these risks was a complicated task, but Salomon’s traders seemed to know what they were doing.

  Salomon’s options arbitrage extended outside the United States. When German investors were interested in buying call options on Boeing Co., the American airplane manufacturer, Salomon simply sold the options on its own, instead of asking Boeing to issue the options—which would have required extensive negotiations.19 Boeing didn’t even need to know about the trade. Salomon hedged its risks from the sale by buying Boeing stock—on the New York Stock Exchange—and by buying put options on Boeing stock—in the over-the-counter markets. As Andy Krieger’s misdirection-play showed, a call option was roughly equivalent to an investment in an asset—in this case, stock—plus a put option. That meant Salomon was hedged, and could pocket a riskless fee.

  Salomon also found a new way to arbitrage Nikkei 225 options, trading on the difference between the stock market in Tokyo and the stock-index future market in Osaka, Japan. Investors were so bullish on stock-index futures—contracts to buy and sell all of the stocks in the Nikkei 225 index, traded in Osaka—that those futures were worth more than the underlying stocks in Tokyo. Salomon could buy the stocks in Tokyo, sell the Osaka index, pocket the difference, and wait for the two to converge.20 If Osaka investors were too pessimistic, Salomon would take the opposite positions. (This trade utilized the same strategy Nick Leeson later would tell his bosses he was employing in the Singapore office of Barings Bank.)

  A final example of a non-U.S. strategy was when Salomon’s traders learned that, although the German tax system rewarded German companies for buying German stocks, foreigners actually wanted to buy those stocks more than Germans did. Salomon created a trade whereby the German companies bought the stocks, but entered into a swap agreement to pay out the stock returns to the foreigners. No one paid tax, and everyone was better off (except German citizens, who lost out on tax revenue, but that wasn’t Salomon’s concern).21

  The “magic” behind these various, complex trades was that they made money even though Salomon was neither selling deals to clients nor taking on much risk. The trades were driven by legal rules or by the actions of unsophisticated parties, or by both. Unlike previous methods of arbitrage, these innovations generated profits for Salomon that were sustainable over a period of years.

  As the Arbitrage Group’s profits increased, Salomon’s top managers were unable to monitor the risks the firm’s traders were taking, especially in derivatives, and the reports they received about trading risks were inaccurate and incomplete. Although it was not widely known at the time, Salomon’s supposedly sophisticated risk-management systems could not accurately value the firm’s trading positions. Salomon had the same problem as Bankers Trust: its own employees could not assess the firm’s risks. Bankers Trust and Salomon were two of the most sophisticated participants in the derivatives markets, yet during the late 1980s and early 1990s, those banks were not able to figure out what their own derivatives were worth.

  In 1993, Salomon’s managers decided to overhaul the firm’s internal risk-management systems, just as many Wall Street firms were doing the same. As employees in New York began reviewing these systems, they discovered tens of millions of dollars of mistakes. Forty Salomon employees spent the next eighteen months, full-time, trying to track down the firm’s hidden losses. Ultimately, Salomon found $87 million of “unreconciled balances” in New York, and another $194 million of errors in London.22 The losses had been building up undetected in an outdated accounting system since 1989. The mistakes resembled those Bankers Trust made during the Andy Krieger incident, but over a much longer period of time, with more than double the losses.

  Salomon’s accounting firm at the time was Arthur Andersen, the prestigious firm that would be involved in numerous accounting scandals later in the decade, including Enron, Waste Management, and others. If Bankers Trust, with Arthur Young, and Salomon Brothers, with Arthur Andersen, couldn’t control and understand these financial risks, how could anyone else be expected to do so? As one well-respected stock analyst, Perrin Long of Brown Brothers Harriman, said after Salomon finally announced the losses in 1995, “You have to ask, ‘What the hell is going on?’ ” Another analyst asked, “When does it end?”23

  Ironically, it turned out that poor supervision and controls were the Arbitrage Group’s major advantages over its competitors. Meriwether’s traders could allocate more capital to their bets when they spotted arbitrage opportunities, because no one above Meriwether was carefully monitoring the group’s risks and use of capital.

  Meriwether had a close and informal relationship with Salomon’s chairman, John Gutfreund. Unlike Charlie Sanford of Bankers Trust, Gutfreund did not impose controls on employees based on RAROC (the risk-adjusted return on capital). Instead, Gutfreund kept track only of revenues, not costs, even with respect to new products.24 If Sanford had permitted Andy Krieger to place billion-dollar currency bets, imagine what Gutfreund would permit Meriwether to do.

  Fortunately, Meriwether was a sophisticated risk manager, and while Gutfreund was ignoring the details about risk and cost of capital for Salomon as a firm, Meriwether was setting profit targets based on these factors within the Arbitrage Group. In rough terms, each of Meriwether’s traders needed to make at least $60 million a year for the firm; otherwise, the trader would need to find a job somewhere else.25 In other words, Meriwether recognized that the Arbitrage Group’s traders were taking such large positions, and potentially putting so much of Salomon’s capital at risk, that if a trader couldn’t make at least $60 million doing so, it wasn’t worth it for Salomon as a firm.

  Gutfreund’s laissez-faire approach had one other advantage. Because Meriwether was not subject to strict controls, he could leave his trading positions in place for long periods of time, long enough for price discrepancies to converge. At other banks, senior managers became nervous when trades turned bad, and forced traders to liquidate positions that eventually might have become profitable. In contrast, Meriwether could “let his bets ride.” The result was counterintuitive: the more out of control the Arbitrage Group became, the more money it could make.

  One example was Meriwether’s response to the stock market crash of 1987. For the first ten months of 1987, the Arbitrage Group had been up $200 million. After losses from the crash, they were suddenly back to even for the year.26 Many of the traders were terrified. Their computer models, which looked at standard deviation as a measure of the likelihood of particular events, had failed. (A one-standard-deviation event happened about a third of the time. A two-standard-deviation event happened only about five percent of the time. And so on.) According to the computer models, the 1987 stock market crash was a vastly
improbable twenty-standard-deviation event,27 less likely than a hundred perfect storms.

  But Meriwether kept his cool, and turned to the reliable on-the-run/ off-the-run trade. In the panic, investors had rushed to the safety of 30-year Treasury bonds, making them more expensive than off-the-run bonds with similar maturities. The arbitrage opportunity—previously competed away—suddenly reappeared. Unlike his peers at other banks (who were panicking over losses, worried their firms might not survive), Meriwether didn’t need special approval to do this trade “in size.” Meriwether bet that the gap between bonds would close, and made $50 million on that trade alone.28

  Gutfreund’s hands-off approach also encouraged traders to engage in strategies designed to skirt legal rules. In addition to the strategies already mentioned, the Arbitrage Group structured trades to profit from regulations, such as those that encouraged European banks to invest money in government bonds, rather than lend money to companies. Meriwether’s traders also structured deals in Italian government bonds to convert government tax subsidies into trading profits. When Paul Mozer had worked in the Arbitrage Group, he did several trades designed to reduce Salomon’s own tax obligations. These opportunities to avoid legal rules persisted so long as the legal rules persisted.