Rubinstein, a finance professor at UC-Berkeley, played an enormous role in the formation of “financial engineering” as some sort of half-baked academic “discipline,” and the concomitant rise of finance as an end in itself rather than a means to end. In 1976, Rubinstein - along with another Berkeley colleague, Hayne Leland - introduced the idea of “portfolio insurance.” The idea behind portfolio insurance was to program a computer to monitor market conditions. In the event certain “trips” or “triggers” take place, the computer will then automatically initiate trades to prevent further losses. Generally speaking, the idea makes sense – after all, computers are programed to fly airplanes and help run power plants. What could possibly go wrong?
Well, two things actually. The first flaw with this concept was the notion that market conditions – which can be dominated by human emotion – could ever be reduced to machine coded instructions to an unthinking computational engine like a 1970s-1980s vintage computer. Computers programed to fly airplanes and control power plants only need to capture the physics of flight and the laws of thermodynamics respectively. These computers and the programs that run them don’t need to account for the mood of pilots or the temperament of plant operators respectively to work properly. A program designed to work in a trading environment would need to model all the nuances of human behavior – how is this even remotely possible?
However, the flaw associated with the impossibility of modelling human emotion– as great as it was – was dwarfed by an even greater fallacy latent to the entire portfolio insurance concept. The same portfolio insurance trading program was provided to a large number of market participants. By having a large number of traders making the same trades at the same time, violent swings in the market were not prevented, they became inevitable. This is exactly what happened on October 19, 1987.
Because so many market participants were using the same portfolio insurance program, once the market started to move down in earnest, the same trades were initiated all over Wall Street. For example as the S&P500 started to drop a certain amount, the portfolio insurance programs automatically started to sell certain securities to raise cash and avoid further losses. As these securities were sold in massive volumes their price dropped. These price drops prompted the portfolio insurance programs to initiate automatic trades of still other securities. The portfolio insurance induced selling fed on itself, and, like a swarm of locusts devouring a pioneer’s wheat crop, the damage was done in just a few hours. Market participants stood by transfixed and appeared powerless to intervene or even understand what was happening as markets plummeted. By days end stocks were down 20%. This remains, by far, the largest one-day drop in history.
The role that his product played in the 1987 market crash and his failure to understand this in advance notwithstanding, Prof. Rubinstein was undaunted. Rather than trying to more fully understand the Frankenstein’s monster he helped create, he has since started the “Masters in Financial Engineering” program at UC-Berkeley. He has dedicated his life to the creation of still more, even bigger financial monsters. As it turned out, October 19, 1987 – as violent as it was - proved to be a mere harbinger of things to come.
In spite of the enormous losses produced – in part by programmed trading as well as the transparently bogus belief in the power of computers to flawlessly manage highly-leveraged investments – the financial services industry would re-order itself to take advantage of the completely mythical and non-existent power of computers to reduce financial risk. The best example of this remains the changes enacted to the SEC’s “net capital rule.” Here, investment banks, financial services firms as well as their media and academia stooges successfully argued to the SEC that because of computers, the SEC’s “net capital rule” should be changed to allow banks to use far more leverage. Changes to the net capital rule were adopted in 2004 and these changes then played a major role in the financial crisis just a few years later.
See William Donaldson (#17) for changes to the SEC’s net capital rule which allowed banks to use much more leverage and the role computer models played in this decision. See Gary Gorton (#27) for another finance professor who erroneously believed computer programs could flawless manage complex, highly leveraged investments. See Henry Paulson (#38) for his efforts to have the net capital rule changed to allow greater leverage in the banking system.