The Automation of the Stock Market: Algorithmic Trading

Photo by Nick Chong on Unsplash

Sit down, grab that pen and piece of paper, write down each miniscule change in a stock’s price, chart those graphs, call up your firm’s broker, and shout amongst a trading pit to negotiate a price. At least, that’s how the stock trading process used to be in the 1960s, way before the market became computerized. Then, in the 1970s, superior trading algorithms began to emerge along with the internet and the entire process became automated, requiring less and less human intervention. Finally, in the 21st century, it seems that artificial intelligence has taken over every field imaginable and at least for finance, it wants to stay!

According to Yahoo!Finance, algorithmic trading now constitutes 60-72% of all U.S. equity trades as opposed to manual trading, and its use is expected to increase by an overwhelming 10.5% in the next five years. 

What is algorithmic trading?

Algorithmic trading (or algo trading) is a computer program that follows manually defined instructions to automate a trade by relying on past market trends. For example, a trader may instruct the program to sell a stock when its short-term 50-day moving average (SMA) surpasses its long-term 200-day SMA, called a “golden cross”. Likewise, the trader may buy a stock when its 50-day SMA dives below its 200-day SMA, a “death cross”. By implementing algo trading, a trader no longer needs to manually monitor live stock trends as the order executes automatically with low latency, or delay, and consequently, the pitfall of emotional investing is completely avoided. 

A specific type of algorithmic trading that has grown in popularity is high-frequency trading (HTF), which relies on complex and powerful computer programs to execute a large amount of orders in the fraction of a second, beyond the capability of any human trader. Generally, the faster the HTF program can make a transaction, the more profit the trader earns. Why? Because HTF profits off of arbitrage, which is the simultaneous sale and purchase of an equity in different markets for prices that are almost identical. Then, HTF exploits these markets’ short-term variations in prices.

The process of algo trading may sound simple, but its implementation is far more nuanced. The instructions for buying and selling take a lot of trial-and-error to work out, especially considering that the sole reliance on past market trends will render algo trading vulnerable to black swans, completely unpredictable events in the stock market with severe consequences 一most notably the Financial Crisis of 2008一which may be better dealt with by human judgment.

Where does AI come into play?

However, the implementation of Machine Learning (ML), a branch of AI that imitates human learning without being explicitly programmed to, in algo trading may circumvent the black-and-white rules set by its traditional counterpart and spot new patterns and trends without requiring human analysis of market data. Granted, the implementation of ML still does not avoid the problem of black swans, but it can certainly be used to crunch data far too large for human analysis and spot general long-term trends rather than focusing on localized trends. Other benefits of implementing ML techniques include the ability to conduct sentiment analysis on news, company announcements, earning reports, and social media to gauge a general public opinion of upcoming stock trends through a method known as Natural Language Processing (NLP) which can classify text into positive, negative, or neutral tones after being trained on labeled datasets. 

Algorithmic Trading: Friend or Foe?

Of course, some critics who are wary of severe technological change will push back against the computerization of the stock, contending that these algorithms will completely replace single-stock analysts and investors, resulting in the loss of hundreds of thousands of jobs. Although it is true that algo trading is far less error-prone and more efficient than any manual trader and that larger companies are taking advantage of these algorithms at light-speed, there will always be a need for traders who are familiar enough with the market. It’s only that they may be expected to adapt their skillset to this new era of the stock market and learn how to program. For example, the demand for quant macro traders is only going to increase as the market becomes more competitive and technology-based. After all, technology is only a tool, not a conscious living being.