Morgan Stanley and the Birth of Statistical Arbitrage
August 28, 2011
Richard Bookstaber, the first market risk manager at Morgan Stanley in the mid-1980s and the author of the 2006 book, A Demon of Our Own Design, writes that “Statistical arbitrage is now past its prime. In mid-2002 the performance of stat arb strategies began to wane, and the standard methods have not recovered”. Was it pure coincidence or absolute prescience on part of Bookstaber that in less than a year after the publication of this book the global financial markets would unravel.
This month’s issue of Bloomberg Markets (August 2011) profiles Peter Muller, the star quant trader, who founded the Process Driven Trading (PDT) group at Morgan Stanley in 1993. Muller and his PDT group at Morgan Stanley have made most of their money from an algorithmic (quantitative) trading strategy called Statistical Arbitrage, or Stat Arb. Not only at Morgan Stanley but at many other investment banks and hedge funds, Statistical Arbitrage has been a hugely profitable quantitative trading strategy for the past 25 years.
Of course, Statistical Arbitrage is not just a single trading strategy. As things stand today, it is an umbrella term used for a broad range of quantitative trading strategies that use sophisticated statistical and mathematical models to analyze price differences and price patterns between securities to generate a higher than average profit for the traders. The math concepts used in Statistical Arbitrage range from Time Series Analysis, Principal Components Analysis (PCA), Co-integration, neural networks and pattern recognition, covariance matrices and efficient frontier analysis to advanced concepts in particle physics such as “free energy” and energy minimization.
The genesis of Stat Arb can be traced from a quantitative trading strategy called “pairs trading”. And 25 years after its birth, this strategy, which exploits price discrepancies and correlation between a pair of stocks to buy and sell them and make money, still lies at the heart of Statistical Arbitrage. It is believed that the notion of pairs trading had been around for many years prior to 1980; apparently, Paul Wilmott has claimed that this trading idea was discovered at his shop in 1980. However, the formalization of the concept of pairs trading and its implementation as an acceptable quantitative trading strategy happened in Morgan Stanley in 1982-83. There is a bit of a debate over who exactly discovered pairs trading. Some, including Bookstaber, believe that it was Gerry Bamberger, who hit upon this idea while working at Morgan Stanley & Co. in the early eighties. Bamberger, a computer science graduate from Columbia University, left Morgan Stanley in 1985 and disappeared from Wall Street around 1987. Others believe that it was NunzioTartaglia, a brilliant quan trader, working with a small group of researchers at Morgan Stanley in 1985 that discovered pairs trading. Putting the debate to rest, let’s just say that it was Gerry Bamberger and NunzioTartaglia at Mogran Stanley who discovered pairs trading in early to mid-1980s.
In the early 1980s Morgan Stanley was assembling a team of computer scientists and traders to work in an independent, ultra secretive group, which would exploit the discrepancies in the stock prices to generate abnormal profits. It would be a well-planned assault on the Efficient Market Hypothesis. One of the members of that team was a computer scientist from Stanford named David Shaw, who would later founded the legendary Wall Street investment firm D.E.Shaw& Co. Bamberger, Tartaglia and Shaw, all worked as a part of this ultra-secretive group that was in search of the Holy Grail of trading. Morgan Stanley’s black box was born in 1985 that would earn the firm a lot of money, and of course, bolster its reputation man on Wall Street. Even though the term Statistical Arbitrage would only come into prominence in the mid-1990s when Wall Street traders would embrace a plethora of complex and esoteric mathematical model to exploit market anomalies, with the introduction of Morgan Stanley’s pairs trading black box the war against the theory of Efficient Markets had begun.
However, with the departure of David Shaw, Gerry Bamberger and many of other associates of Tartaglia, in the mid to late eighties this quant trading group at Morgan Stanley would fall apart in spirit, though traders would continue to use pairs trading on the firm’s trading floor. In 1993, the task of resuscitating the group would fall on the 29 year old Peter Muller, who would be hired by Derek Bandeen, a prop trader at Morgan Stanley. The group would be anointed with a new name, the Process Driven Trading (PDT), and Muller would recruit his own army of quants and computer programmers to work with him. Over the next decade, Muller’s PDT would make lots of money for the firm and establish Morgan Stanley as the leader in the field of Statistical Arbitrage.
Any comments and queries can
be sent through our
More on History of Finance >>
back to top