Contrarian Investment Strategies

by David Dreman

  4. The EMH offspring, CAPM, states that risk is defined solely as volatility. The only way to lower risk is to lower volatility. But the risk that caused the 1987 crash was not volatility. The panic was caused by a serious lack of liquidity as the futures market dried up, as well as enormous excess margin. Both are measures of risk that are not part of EMH risk theory and therefore are excluded from it. This prodigious flaw in risk measurement is vitally important to your investment decisions. (Solutions will be discussed in chapter 14.)

  5. EMH theorists thought that increased leverage would facilitate more efficient markets by strongly encouraging futures trading on stocks at low margins. These academics played an important role by having their recommendation to drastically lower stock future margins about 90 percent below stock margins accepted by the regulators, the SEC, and CFTC. They apparently did not consider that in a major downturn low leverage could unleash massive selling that could cause markets to plummet as they did both in 1929 and again in 1987.

  What lessons did EMH theorists, as well as true believers, including then Fed Chairman Greenspan, learn from this crash? As noted, none. The theorists tried to explain away all the causes, although their explanations ranged from skimpy to downright foolish. As a result, the mistakes were repeated and contributed to two more market collapses, as we will see next.

  2. The 1998 Long-Term Capital Management Debacle

  You’d think that after the 1987 crash EMH would have undergone some serious revisions if not outright rejection by Wall Street analysts, as well as questioning by its adherents. But that didn’t happen. By 1998, a new crisis was beginning to reach the boiling point, owing to the actions of a monster hedge fund’s investment strategies. The firm was “too big to fail” long before the term became popular. The mixture of EMH theory, the increasing belief in the almost godlike investment capabilities of rising efficient market superstars, and serious amounts of money produced a crisis that was stunning in both its speed and its scope.

  Long-Term Capital Management (LTCM) was the world’s largest hedge fund. Beginning operations in March 1994, by early 1998 it had acquired assets of more than $100 billion, as well as more than $1 trillion in derivatives. More than fifty of the largest banks and investment banks in the United States and abroad scrambled to do business with it in spite of the almost impossibly tough terms it demanded, concessions that no other hedge fund was given. By 1998, it was the envy of Wall Street, more than quadrupling its original investment in less than four years.

  LTCM, led by its chairman, John Meriwether, was considered the crème de la crème for the exceptionally high level of its traders, but mostly because it hired some of the superb efficient-market superstars, led by Robert Merton and Myron Scholes, the latter of whom was instrumental in developing the Black-Scholes model. They shared the Nobel Prize in Economics in 1997. In addition, a group of brilliant Ph.D.s, several of whom were excellent traders, surrounded the two masters. On paper there wasn’t another firm with this kind of dazzling cast, as its record from the start seemed to prove.

  LTCM concentrated on “relative value” trades in both domestic and global bond markets. If two high-quality bonds, for example, had almost the same maturity but one yielded 25 basis points (one-quarter of 1 percent) more than the other, LTCM would buy the higher-yielding bond and short the lower-yielding one. Or if a highly rated bond in Spain provided a higher yield than one in the United Kingdom, again with nearly the same maturities, the fund would buy the Spanish bond and short the U.K. issue. These were called “paired trades.”

  The fund was also highly diversified, owning and shorting thousands of issues. Another feature of LTCM’s strategy was to buy somewhat riskier and more illiquid bonds in its trades than the ones it shorted to increase the spread. This increased the return on each paired transaction, as less liquid and slightly riskier bonds commanded a somewhat higher yield than bonds with the same risk and liquidity. LTCM believed that this was a low-risk strategy because interest rates tend to rise or fall in sync, so spreads move only a fraction as much as the value of the bonds themselves; LTCM should fare well no matter whether bonds rose or fell or even if the market crashed. One of the academics calculated the risk in a spread transaction at only about 4 percent of owning a bond outright.

  The paired-trade strategy that was a good part of LTCM’s portfolio was in essence a nickel-and-dime business. Myron Scholes, aside from being an outstanding academician, was also the firm’s top marketer. As he explained it, “the fund would be earning a tiny amount on thousands of trades as if it were vacuuming up nickels that nobody else could see.”7 In essence it would be like a giant penny arcade making big money on the volume of nickels and quarters thrown into the wide variety of game machines on the premises.

  The confidence so many felt in the strategy was premised on work by Robert Merton, who was considered one of the outstanding experts on risk. He had made major breakthroughs in EMH theory in risk, which eventually earned him the Nobel Prize. Merton was convinced, as were almost all EMH believers, that volatility was the sole measure of risk; this was the bedrock of LTCM’s approach. He was equally sure that volatility remained stable over time. If a stock or bond had a certain volatility, it might swing higher or lower temporarily but would always return to its calculated rate. Since that was the case, LTCM could use volatility as its most important tool. If volatility was higher than it should be for the bond bought in a pair trade and lower than normal for the one sold short, it would return in each case to its normal level. The higher the volatility on the bond purchased and the lower the volatility on the bond sold, the greater the added return to the trade when the risk swung back to its normal level. That this assumption was never proved was of no consequence.

  The strategy was also premised on the belief that volatility was mathematically predictable. A large amount of testing was done on volatility, as well as the chance that something could go wrong. Using the cardinal Merton principle that volatility did not change, LTCM was convinced that it could calculate precisely the odds of what the best, average, and worst days would return, as well as the odds that the portfolio could take serious or mortal losses. Even worst-case calculations showed that the firm could sail through the most punishing financial hurricanes.

  The belief that volatility was mathematically predictable was not held only at LTCM; it was a core belief throughout the Street. The Black-Scholes pricing model was anchored on the predictability of volatility, and trading rooms on Wall Street and in every major financial center globally monitored volatility as if it were the Holy Grail. Every major trading floor was staffed by bright young Ph.D.s who had studied under Fama, Merton, Scholes, or other strong EMH believers, and they didn’t question the unproved premise that yesterday’s prices were reliable and predictable indicators that would determine today’s value. When Genius Failed, Roger Lowenstein’s first-rate book on Long-Term Capital Management, quotes Peter Rosenthal, LTCM’s press spokesman, as glibly stating “Risk is a function of volatility. These things are quantifiable.”8 “Meriwether, Merton, Scholes and Company had no more earnest belief.”9 The LTCM portfolio contained thousands of paired trades with precise data on the volatility and expected return of each.

  THE FAIL-SAFE STRATEGY

  What the professors and traders didn’t focus on was leverage or liquidity. Since they had volatility completely under control and that was the only source of risk, liquidity and leverage did not matter a hoot. So focused were they on volatility, and so strong was their belief that volatility was risk, that leverage and liquidity as independent risk elements were considered inconsequential. They were not even a punctuation mark in the EMH-inspired formulas LTCM used.

  The question most often asked by LTCM academics and portfolio managers was how much volatility (“Vol”) it took to optimize returns. The answer that almost consistently came back was that the Vol could be raised, with little addition to risk. The leverage went from 20 to 1 to 30 to 1, but Vol, they believed, could still
go higher.

  The gods continued to smile on Long-Term Capital Management. In 1994, it earned 20 percent; in 1995, its net return was 43 percent. By the spring of 1996, it had $140 billion in assets under management, and its capital had tripled to $3.6 billion. After only two years of operations, its assets were larger than those of Lehman Brothers or Morgan Stanley and were approaching those of the giant Salomon Brothers. In 1996, it earned 41 percent net of fees or $2.1 billion, more than giant blue-chip growth companies such as Walt Disney, McDonald’s, and American Express.

  But there were other somewhat less appetizing aspects to this sizzling growth. By early 1996, the firm’s leverage had, as noted, climbed to 30 to 1. Leverage works well, as we saw, when markets are going up, as they were before the 1929 and 1987 crashes, but when markets go down, it’s death. At 30-to-1 leverage LTCM had to lose only 3.3 percent of its market value to have its capital wiped out entirely. Meanwhile, its return on total assets was less than 2.5 percent, before counting its large derivative positions in the many hundreds of billions of dollars. Including those, it was probably less than 1 percent.10 A ninety-day Treasury bill at that time yielded 4.5 percent, which was approximately 4.5 times LTCM’s return on total assets. Almost all of LTCM’s return came from levels of leverage that defied gravity.

  Then, in mid-1997, the Asian economies began to melt, first Cambodia, then the Philippines, then Malaysia, then South Korea, followed by Singapore and Indonesia. Enormous amounts of stock were dumped. And markets crashed while currencies went into free fall. The Asian Tigers became terrified kitty cats. The contagion spread to South America, Eastern Europe, and Mother Russia.

  LTCM determined to make its signature bet in the crisis. Because volatility was very high, the company decided to short sell volatility options not only on the S&P 500 but on major indexes in Europe and other major markets.*32 Volatility, Merton and the group were certain, would revert to the mean, and this meant that volatility futures would fall sharply. And here the biggest error in volatility measurement LTCM ever made mortally wounded the fund.

  Merton was taking an enormous chance. EMH theory, the Black-Scholes model, and Merton’s models say only that a stock’s volatility is consistent over time. They do not say in what time period volatility securities or derivatives will revert to the mean. It’s like crossing a river that’s four feet deep on average, carrying a fifty-pound backpack: it might be three feet deep in some places and fifteen feet deep in others. Not good survival odds, but precisely the odds Merton the supermathematician had chosen.

  In a bad crash or a bear market, would it take six months, twelve months, or even years before volatility went back to its normal level, as Merton and EMH believers fervently believed would occur? The efficient-market theory had no answer for that. How could it have one? The belief appears to have come from an EMH theory, not rigorous testing. But LTCM was leveraged 30 to 1, so the return to normal volatility had to come very soon, because leverage—not recognized by EMH as a risk factor—will gobble capital up almost at the speed of light if the markets continue to go against it. Unfortunately, that’s exactly what could happen to LTCM. Merton and Scholes were making the same kinds of bets that any desperate amateur deeply out of pocket might make: LTCM could fail on the toss of a coin.

  Maybe the gods were not major EMH adherents after all. LTCM’s spreads on its paired trades did not narrow; they widened. Through 1998, a flight to safety occurred in almost every sector it was invested in. Remember, the firm was always invested in the riskier, not safer, bonds and was short of the safest bonds. To add to its problem, the volatility indices it short sold skyrocketed as volatility increased. In May, problems with the Asian Tigers flared up again. By early June, the Russian financial system was near collapse. World markets turned down sharply. LTCM was hit even harder because the dash to the safest securities rapidly widened the spreads between its long holdings and its short positions, while volatility increased further.

  LTCM reported a loss of 18 percent for the first half of 1998. More important, as its capital dropped, its leverage increased dramatically, making it less and less able to meet margin calls. Then, in August 1998, Russia defaulted on its debt and market fear heightened across the board. LTCM’s capital of more than $5 billion near year-end 1997 fell to the hundreds of millions within four months. The two risks not even considered punctuation marks by EMH theorists and by Merton and Scholes—leverage and liquidity—came to the forefront, just as they had done in the 1987 crash.

  As the markets tumbled, the enormous short positions the firm owned resulted in gigantic losses; the increasing flight to safety continued to cause the more risky bonds it held to drop significantly more than the higher-quality bonds it had shorted in the paired swaps. Too, the volatility swaps that Merton and his team had shorted heavily because they believed Vol was much too high soared higher. Calls for more margin poured in to cover the flood of new losses, and consequently the firm’s capital was hemorrhaging rapidly.

  As a result of the firm’s gigantic margin calls and its inability to sell most major positions because they were very illiquid, the fund’s leverage shot up to 100 to 1, which meant that even a 1 percent drop in assets would bankrupt it. But it couldn’t sell to protect itself because its positions were gigantic and always less liquid than the positions it sold short. In the good times its capital, enormously magnified by leverage, made it huge profits. But in those brutal markets, selling even a small fraction of any holding would reduce prices sharply (lack of liquidity again). LTCM was doomed!

  August was a cruel month for Long-Term Capital Management. It lost $1.9 billion, or 45 percent of its capital. It had $155 billion in assets, with a leverage ratio of 55 times its shrunken capital. Mathematicians calculated that the odds of the losses LTCM incurred in August were so freakish as to be unlikely to occur over many repetitions of the life of the universe.*33 11 Its position was untenable. The leverage could not be reduced because of the enormous size of its trades and the complete loss of liquidity. As with subprime bonds almost a decade later, any attempt to sell would see the already wafer-thin markets for the securities it held collapse.

  The markets were in a state of near-panic. Meanwhile, LTCM was surrounded by a pack of predators: bankers and brokers who knew that its liquidity and leverage problems would put it under. They began shorting LTCM’s positions, well aware that in a forced liquidation, when it was compelled to sell, prices would drop drastically and the bankers could cover their shorts at bargain prices. The parallels between the portfolio insurers in 1987 and LTCM are remarkable. Both were undone by a massive lack of liquidity when it was most needed, in a bear market.

  By September 1998, the Federal Reserve Bank of New York, under William McDonough, its talented president, was concerned about the plight of LTCM. Although he had no authority to inspect the books of a hedge fund, LTCM gave the New York Fed permission to do so. What he saw alarmed him. Not only did he see that the collapse of LTCM was imminent; he also realized that the enormous positions in many thousands of holdings, if dumped on the market to sell at any price, could severely disrupt the already badly shaken financial system and possibly lead to disaster. He and his chief aides called numerous meetings with the leaders of the largest banks and investment firms in the United States, as well as important banks abroad that held large LTCM loans against the securities swaps and other holdings of the fund.

  An agreement to take over all of LTCM’s positions would need the support of amost all of the major banks and investment bankers on Wall Street. In the end, after much haggling, its positions were sold to a consortium of sixteen banks that anted up more than $260 million each. The fund’s shareholders saw their investments drop 92 percent in five months. By comparison, over the four-year period that Long-Term Capital Management existed, the S&P 500 had doubled.

  Despite this massive failure of EMH, neither of the two Nobel laureates who had championed the fund’s strategy questioned the theory. Although Merton admitted that the
fund’s risk measurements hadn’t worked, he stated that the principles it had followed were right. What was needed, he argued, was more sophisticated modeling. He went back to teaching at Harvard and, perhaps ironically, was hired by J. P. Morgan as a risk consultant.

  In a speech a year later, Scholes concluded that the change in volatility must be attributed to a permanent change in what investors would pay for the more risky investments LTCM made. This meant that all the theory about volatility that had been used up to then must have changed permanently by events at the time LTCM went down, which is not unlike saying that the weight of the supports necessary to build an expansion bridge of a certain length and width had suddenly and permanently been altered by a sudden change in the laws of physics. Scholes never doubted the efficacy of EMH’s supporting volatility theory. Since the permanent changes he discussed were reversed within a few years, his theory may have a few minor holes in it.

  Indeed, no questions were raised by EMH supporters generally about LTCM’s premises on volatility or whether leverage or liquidity was also a possible risk factor. How could they be?

  3. The 2006–2008 Housing Bubble and Market Crash

  By now we are all familiar with the housing bubble of the early to mid-2000s. The interactivity of so much of our financial system and the complexity and the unethical practices carried out by far too many of the large institutions involved are well beyond the scope of this work. What can be sketched out briefly is the part EMH theory played in fueling this bubble to the gigantic proportions it reached. This crisis was not dissimilar to the destruction of Long-Term Capital Management but on a colossal scale.

  In spite of the 1987 crash and the LTCM debacle, investment institutions still had almost blind faith that volatility was the sole measure of risk. Leverage, although extremely high on subprime and other lower-quality residential mortgage-backed securities (RMBSs),*34 was of almost no concern to the portfolio managers or the risk-control departments of these institutions, who were all trained in the EMH and CAPM theories of risk. Nor, for that matter, was their liquidity questioned. There were many thousands of series of home mortgages, whose quality varied enormously, from good to very poor. Normally liquidity was low, because each mortgage issue usually contained a number of series of RMBS issues ranging from AAA, the highest rating, to series that were toxic at best. Most required substantial analysis to determine even a rough price range, so dealers’ spreads between bids and offers were normally large. Still, volatility was the only risk factor considered, and as housing prices boomed consistently higher, it was very low indeed.