Algorithmic trading ups the pace on Wall Street

On Monday Feb. 5 the Dow Jones Industrial Average suffered its largest one-day drop ever, falling nearly 1,600 points in the middle of the day before closing 1,175 points below the prior day’s close to register a one-day drop of 4.6 percent. In the intervening days the Dow, along with the rest of the stock market indices, appear to have stabilized somewhat and traders have bid prices back up – at least in the short term.

The initial catalyst for the selloff was a sudden fear on Wall Street that inflation and interest rates might rise more rapidly than previously thought.  High interest rates cut into corporate profits and so a re-evaluation of what corporate shares are worth caused a number of investors to trim their holdings in equities.  But what happened next is the real story.

To understand why stock prices experienced such drastic swings over the past week, we have to go back almost two decades to when Wall Street firms like Goldman Sachs, Merrill Lynch and others invested heavily in computer technology and human capital to take the guess work out of investing and also make their trading more efficient.  The big Wall Street banks sent recruiters to prestigious institutions like MIT, Cal Tech and Stanford to find and hire the most talented mathematicians and computer scientists to study the stock market and design trading strategies that would bring mathematical precision to the decision-making process and tie them to the most powerful computers to execute trades in fractions of a second.

These mathematicians designed what are called “algorithms.”  Algorithms are a specific set of clearly defined instructions aimed at carrying out a task or process.  A simple algorithm can be found in your home’s thermostat.  The thermostat is programmed to turn the furnace on when the temperature falls below a pre-determined degree.  Trading algorithms, by contrast, are extremely complex with pre-determined factors running in the hundreds.  One of the purposes of an algorithm is to generate a “sell” signal when the odds that an investment will lose money moves past a critical point – just like your home’s thermostat generates a “turn on” signal when the temperature drops. The “input” is the temperature.  The “output” is the signal to turn the furnace on.

What happened in the market on Monday was that when some humans became a little worried about higher interest rates and sold stocks, an algorithm somewhere noticed that stock was being sold and used those transactions as an “input.”  The “output” was a signal to sell some of its own inventory.  From here, you can probably imagine what happened next.  Each output became another input for the vast network of algorithms, creating a chain reaction that caused stock prices to fall much faster and further than they might have if mere humans were in charge of making these decisions, and the next thing you know, the Dow Jones has lost 1,600 points.

Programmed trading, like all the other mind-boggling technology that has crept into our lives lately, is here to stay so we’re going to have to get used it.  Fortunately, our investment department has been studying this development for years and is prepared to respond to it in appropriate ways.  As with this instance, the best response is to not react impulsively but to notice what is going on and keep focused on the long term.  Markets go up and down, but over long periods of time they always to up.  With computers dominating trading on Wall Street, the ups and downs are getting more intense but the approach to long term prosperity hasn’t changed.