Artificial intelligence (AI) is coming to the investment world.
With the help of deep learning techniques, AI researchers have made significant strides in natural language processing (NLP), speech recognition, and image recognition. Computers can now see, hear, and understand human beings. They have also demonstrated shrewd decision making.
What does this mean for investment management professionals?
Computers are gaining an edge over humans through such common standards of intelligence as language skills, mathematical skills, and memory. That margin will only grow wider over time. Will that edge translate into better investment skills?
“The biggest advantage of a computer [over a human being] is its practically unlimited memory,” Eric Chang of Microsoft Research Asia explained. The 152-layer-deep neural network his team at Microsoft developed can tell what’s in a picture with more accuracy than humans.
Training such complex models requires a tremendous amount of data, more and more of which has become available in recent years. “Data alone is not enough though,” Chang said. “Our focus is on getting insights from the data.”
Investors have used image recognition programmes to find oil tankers on satellite imagery. Some have been able to get a better gauge on oil supply by analysing the tankers’ tonnage, routing, and port arrival times.
Many analysts listen to quarterly conference calls from corporate management to detect clues that they can use to estimate corporate earnings and build valuation models. With the help of voice recognition programmes they can zoom in on a small number of companies where AI raises a red flag based on changes in management’s speech patterns.
Furthermore, AI algorithms can help a bank client evaluate its risk exposure to a potentially problematic borrower. An example of this is using NLP and knowledge maps to go through a whole host of information including regulatory filings, legal proceedings, and online information about related transactions. On one such occasion, the programme detected over 800 accounts related to the problematic borrower. The banks originally thought there were four.
Microsoft applies its image recognition models to understand investor personalities. They can then harvest that data to build more customised portfolios, demonstrating how deep analysis can inform better decisions.
Better investment decisions come, in part, from more precise asset pricing. More-in-depth analysis provides more accurate inputs for valuation models. For example, if the information on oil tankers gives you an edge over your competition in forecasting oil prices, it will also help you better model revenues and costs for oil companies and airlines. If your program succeeds in catching CEOs in their mistruths on conference calls, you’ll likely capture alpha by selling those companies when you hold them and avoiding them when you don’t.
Behavioural biases will continue to influence our investment decisions, often to our detriment. For example, investment managers are often prone to herding, or following the crowd. At the height of the tech bubble, for example, too many investors chased a stock simply because management added a .com to the company name.
But machines won’t follow the next machine. Unless we program them to do so.
The End Game
Given AI’s superior brain power and lack of emotions, the market could eventually be dominated by a small number of AI programmes, maybe even a single one.
AI also has support in academic circles. Campbell R. Harvey of Duke University believes AI will assume a major role in investment decision making and that the proliferation of AI and big data will result in “15 to 25 investment management superpowers that can harvest all that data.”
So the big question is when — not if — AI will supplant human investment managers.
At the 71st CFA Institute Annual Conference in Hong Kong, David Pope, CFA, will discuss natural language processing (NLP) and corporate earnings sentiment analysis.
If you liked this post, don’t forget to subscribe to the Enterprising Investor.