Author: Roger Lowenstein
Written by an American journalist Roger Lowenstein, When Genius Failed arrived at my doorstep last August during the lockdown period. I had bought it out of curios-
ity, to see how a bunch of geniuses, with the help of two Nobel Prize winners, ran a hedge fund so wildly successfully in the 1990s, until it wasn’t.
The fund was Long-Term Capital Management (LTCM), and the two Nobel laureates — Myron Scholes and Robert Merton — are the ones who developed the Black-Scholes model widely used to value options contracts today. However, at the heart of the story is John Meriwether, the founder of LTCM who was known as one of the finest bond traders at that time.
If you have read Liar’s Poker by famous financial journalist and writer Michael Lewis, you would know that Meriwether is a poker player with an enormous risk appetite. “King of Wall Street” John Gutfreund, then CEO of Salomon Brothers Inc, had walked up to his trading desk for a US$1 million poker game. Hence, the famous quote of the book: “One hand, one million dollars, no tears.”
How did Meriwether respond? He upped the stake to US$10 million and Gutfreund backed off.
Lowenstein starts his book by casting a different light on Meriwether. He was “the priest of calculated gamble”, not a reckless poker player. He grew up in a Catholic community in Chicago and was accustomed to a “pervasive sense of order”. He wasn’t crude and loud, but a reserved and intensely private person. He was smart and understood that one needs an edge to stand out in the market.
The edge Meriwether found at Salomon Brothers was the untapped brain power of the academic world. Banks hired extremely smart academicians and mathematicians during the 1980s, mostly for research purposes, as they were deemed unfit for trading. Quants, as these people were known as, wasn’t a glorious word back then.
Meriwether embarked on a hiring spree to take in several people from the academic world and formed The Arbitrage Group at the firm. And they made tonnes of money from trading bond spreads, which, to put it simply, is trading bonds by comparing the value of one bond to another.
Investment becomes science, not art
Meriwether was very successful during that period as the group made more than US$500 million annually for five years before he quit Salomon Brothers in 1993, mainly due to a trading scandal perpetrated by his subordinate.
Meriwether very quickly assembled his old team at Salomon and founded LTCM the next year. The hedge fund had a stellar start by making 28% in its first year of operation. It made another 43% after fee the following year. A billion dollars was raised from investors subsequently.
The interesting part of the book is its analysis of the various reasons behind LTMC’s success. Many people attributed the success to the enormous brain power of the geniuses and the guts of the traders. Well, it was only partly correct, according to the author.
For one, there were fewer people trading the bond market back then, rendering it less efficient while trading opportunities abound. LTCM’s cost of funding — the money it borrowed from banks to trade bonds — was extremely low, partly due to the confidence it inspired in the minds of the bankers that the fund would not fail.
Lowenstein talks about how LTCM applied mathematical calculation to historical data to “predict” bond prices and the odds of the fund suffering from losses through various “ratios”. Such a practice is commonplace today, but it certainly was not in the 1990s.
“Just as a handbook of poker might tell you that the odds of drawing an inside straight were 8.51%, the professors calculated that Long-Term would lose at least 5% of its money in 12% of the time (that is, in 12 of every hundred years)…”, the author writes.
The firm told its investors that the hedge fund would lose at least 20% of its portfolio only once in 50 years, which is an awfully long period.
How accurate can these predictions be? According to Lowenstein, these prediction models are based on the “law of large numbers” that is commonly applied by physical scientists in observing and analysing natural phenomena. “If one holds enough samples of a random event, they will tend to distribute in the familiar bell curve, with the most occurrences around the average and a sharp drop-off at either extreme. This is called the normal distribution or, in mathematical terms, the log-normal distribution.”
To paint a clearer picture, the author writes, “It is like the way cream spreads through a cup of coffee. Any one molecule’s trip is random, but as a group, the molecules distribute themselves in predictable fashion, from the centre out. The cream will never go all to one side.”
In short, the geniuses behind LTCM treated securities price movements as a series of natural and random events, with each of them happening independently, without correlation with another. A certain pattern would then show when one gathered enough historical data of these events.
The scariest part of these prediction models is that they were accurate most of the time. Coupled with LTCM’s two years of stellar performance, the fund’s traders started to feel invincible with the power of prediction that lies within these mathematical formulas. They placed ever larger bets on their trades and across various asset classes.
What they failed to see is that each asset class behaves differently and, more importantly, market movements can quickly converge during major crises with investors all rushing out the door at once. After all, how can mathematical formulas predict human psychology and emotions? That was exactly what happened during the Asian Financial Crisis, which then led to the collapse of LTCM in 2000, barely six years after it was founded.
Don’t put all your faith in one number
Most investors who have just started their investment journey by investing in specific instruments, such as unit trust funds, would have come across a common disclaimer: Past performance is not an indicator of future performance.
It is a simple sentence with a history that stretches back at least two decades to the demise of LTCM.
You would have also come across various ratios, such as risk index, that tells you how much your investment portfolio could fall in bad times. It could be right most of the time. Yet, be prepared for the worst when there is a black swan event.
After all, how can mathematical formulas consider human psychology and behaviour? For instance, many investors and experts were wrong about how far US president Donald Trump was willing to go against China in the trade war he initiated in recent years. Many were also wrong about the impact and duration of the ongoing Covid-19 pandemic on the global economy.
All is good if you can digest and analyse big data to help you make better investment decisions. But don’t be overconfident about complex mathematical formulas, as even geniuses have lost their shirts in predicting the markets.