Death of the Stock Analyst: Why they will be Replaced by Robots

I was recently at a technology CEO conference. At one point, there was a discussion about how some companies no longer want to go public. One seasoned CEO bemoaned the amount of time that goes into analyst meetings when you are a listed company. I pointed out that things might get better fast: the very technologies these CEOs are applying to other problems – neural networks, machine learning, and other tools of artificial intelligence – could completely eliminate traditional stock analysis and analysts. I believe there is no reason it can’t.

When it comes to things like deep machine learning, you may have heard the terms but not yet appreciated the results. But they are here. Today you can download free software to build your own artificial intelligence algorithms, in a few minutes, that can detect handwritten numbers with well over 95% accuracy. In the past few years, advances in synthetic vision have allowed driverless cars. (Every time your computer identifies you with facial recognition or a fingerprint it is using that technology.) And Google has processed one hundred million street addresses with an accuracy as good as humans, but faster. They claim they processed all the addresses in France in under an hour.

Investing is not immune to technological innovations. Technology has already changed so much in financial markets that they are barely recognizable from even ten years ago. Only a decade ago, large investment banks had trading floors that were the size of football fields – and stock, options and futures exchanges had bustling pits. Today, those venues and floors are nearly empty – or totally abandoned.

But it goes beyond people. Algorithmic and high-frequency trading has fundamentally changed markets. Every sale of one million shares is electronically broken up into thousands of 100 share trades to make big orders indistinguishable from an individual punting with a few thousand dollars. The real actions are masked by the sheer volume that comes from all those artificially small trades. If nothing else, the visibility of bulky trades, which might signal even larger trades waiting to hit the floor, is lost today.

So how does technology affect stock analysis? Analysts build large spreadsheets from financial data, trying to create a handcrafted forecast of the future results of a single company or business. It is all about being a mile deep and a foot wide: spending enormous effort to try to get the right data about one company at a single point in time to decide if it will outperform or underperform business expectations.

But technology takes data analysis to another level. An algorithm can analyze not just one company at one time – but rather can evaluate every piece of financial data ever reported for every company over all time to find patterns and identify opportunities across firms – and do it all in minutes.

What’s profound is while it used to be that only large, well-funded investment firms could do this depth of analysis, today, the data is available to everyone.

The end result? Algorithmic trading has made capturing trading profits harder, diminishing opportunities to capture alpha, thus, leading to the rise of ETFs and factor investing. Instead of seeking fund outperformance, investors are focusing on how to combine factor exposures using low-cost ETFs.

Some observers believe we will look back to this decade in 100 years and say the computational changes happening today are as fundamental as when mankind developed electricity, the transistor, the gas engine, or airplanes. Those technologies put the city lamp lighter, the blacksmith, and the Pony Express out of business.

The traditional stock analyst may be one of the next professions to become extinct.