Model vs. Experience

How we make investment decisions is based on our conception of how the investments work and what our expectations are. Before 1900, in the West, investing and financial activities were seen as vulgar. Respectable intellectuals didn’t give themselves to contemplating the vagaries of the market. More recently, much research has been done to try to improve the stability and profitability of stock market investments. It seems incredible that the market crash of 2008 was as severe as the Panic of 1907. Could it be true that our knowledge and expertise relating to financial investments has made little or no progress over the last century of study and thought?

What happened between the Panic of  1907 and the present day? The Roaring 20s drew many people into speculating on the stock market, using increased margin and betting that prices would always rise. That ended abruptly in 1929, with the market crash, the policy mistakes of governments and the Great Depression. An entire generation swore off stock market investments for life, seeing it as too risky. And, indeed, it took about 20 years for the market to reach and exceed its previous highs.

Two things happened in the early 1950s. Alfred Winslow Jones, a journalist at Fortune magazine, set up the first hedge fund, using hedging techniques to produce returns that beat the top mutual fund by 87% over five years. His hedgeing technique focused on using long-short pairs to increase returns while decreasing variability. Separately, Harry Markowitz developed a modern portfolio theory, using a variety of assets to optimise a portfolio for improved returns at a given level of “risk” or reduced “risk” at a given level of returns. Because risk needed to be measured and calculated, this led to the equation of risk with volatility. And since volatility and probability involved mathematical variability, Gaussian statistics were adopted by the interested academics in order to develop a mathematical model for financial returns.

The problem is that the model didn’t match reality. In 1998, Nobel laureates Myron Scholes and Robert Merton came very close to causing catastrophic losses across Wall Street that their hedge fund, Long-Term Capital Management needed to be bailed out in an arrangement by the Federal Reserve. The model that LTCM was using was based on the application of Gaussian statistics, assuming that the fluctuations of the stock market match brownian random motion (movements are unrelated, steps are of equal size). It basically presents a world view where markets rise and fall unpredictably, but unspectacularly. As we all know, markets crash.

This is where time-tested experience can help. Warren Buffet has said that he appreciates modern portfolio theory being taught in business schools because it makes his job easier. He has decades of experience in choosing companies that are both profitable and undervalued. His teacher, Benjamin Graham, was a lifelong student of the markets. And there are many aphorisms that traders use to sum up collective experience. The difficulty, then, is in separating the superstition from reality. Do market crashes always happen in September and October? Should we really sell in May and go away?

An additional benefit to paying attention to commonly accepted aphorisms is that a rational person will treat them with some skepticism. The problem with a theory or a model that has been accepted is that it may be acted upon unquestioningly. Question models, question accepted wisdom, and always plan for the case where everything goes wrong, however unlikely. The principals of LTCM learned these lessons the hard way, but we now know to be skeptical of academic models that work only under certain conditions, that may differ from real life.