The lesson of Long Term Capital Management

Over the years I have marveled at the fact that some of the most intelligent people in the financial markets repeatedly get blindsided by market action. Frequently it is because in the real world the markets do not act in accordance with their view of how the markets should act.

A great many intelligent people lost money when the markets crashed in 2000 and 2008 because in each instance they did not see the crash coming. Many fall back on “nobody” can predict the market when what they mean is that “they” failed to predict the market.

A great deal of the advice given by the Wall Street firms is conflicted. Even simple tools like asset allocation are grossly misapplied. Finding a better than average financial adviser can be hit or miss.

Many people agree that investing requires time, information, analysis and discipline. There is logic that suggests using computers and mathematics to make investment decisions has merit. Computers will certainly analyze more information in less time and can trade any account subject to a rigid discipline.

Success should be dependent upon analyzing the right information in the right way. Hiring really smart and accomplished people to decide which information to collect and how to analyze it would seem to enhance the chance of success. Except that it does not always work.

The most outrageous example may be the case of Long Term Capital Management (LTCM), a Connecticut based hedge fund that lost about $4.5 billion of investors’ money in 1998 and almost brought the markets down with it. The investors were some of Wall Street’s biggest banks and many of the individual executives who managed them.

LTCM was started in 1993 by Lee Meriwether, a very accomplished trader who had made substantial profits for Salomon Brothers. Showcased members of the team were Myron Scholes and Robert Merton, two economists who had devised a mathematical model for pricing options. Merton and Scholes won the Nobel Prize in Economics for that model in 1997 just before the downturn that wiped out LTCM.

LTCM performed arbitrage with its investors’ money. They looked for small discrepancies in the price of the same or similar instruments in different markets. They assumed that the markets would always efficiently close those gaps.

LTCM created sophisticated mathematical tools to identify those discrepancies and to evaluate the greater markets so they could estimate how those gaps would close. No one has suggested that LTCM’s math was wrong; it is just that the events that occurred were not in the database that they were analyzing.

In 1997 the government in Thailand devalued its currency. The ensuing defaults roiled the markets in Asia and caused a serious decline in the equity markets. Credit markets in Japan, a major US trading partner and the most important capital markets in Asia tightened significantly. It did not help that Russia defaulted on its own sovereign debt shortly thereafter.

Importantly, LTCM did not lose money when the devaluation occurred in 1997 but a year later. The LTCM fund was very profitable into 1998. Losses started to mount up when its mathematical models could not account for the shifting market conditions caused by the devaluation. They were useless to predict the effects of the often conflicting ways in which other Asian governments and central banks would deal with it.

The lesson to learn from LTCM is quite simple. Even the best mathematical models created by the smartest people should not be relied upon to tell us what the markets may do. No computer program can accurately predict the price of securities one month or one year from today.

Despite this fact, there are currently a multitude of “quant” firms that are developing and using ever more sophisticated mathematics to do just that. Most are focused upon making predictions of what will happen in the markets today not next month. I wish them luck but I would not give them any of my money to invest.

The markets will continue to evolve, globalize and expand. Developing mathematical models based upon how the markets have acted up until today will be less and less accurate and have less and less utility going forward.

Millennials think otherwise and are expected to invest trillions of dollars with robo-advisers who use mathematics in the same way. A substantial percentage of those funds will be lost the next time the market turns down.

Then the market”professionals” and pundits who currently sell and endorse robo-adviser programs will remind the millennials that “nobody can predict the market” because some things about the markets never change.