I asked ChatGPT to “Give me a 4-word funny title for a blog post on AI for Investments” for this blog post, so you have now been exposed to the power (as simple as this request is!) of AI! I dropped the prefix, “Robots Handling Your Dough:” so there was a bit of human interface and editing!
Everyone seems to be talking about artificial intelligence (AI) these days, whether it is the ChatGPT app for generic queries, robotics, or any of the other customized AI apps for videos, pictures, or business applications. These applications, though not perfected, are amazing and capturing the imaginations of all of us. Using AI to help with stock selection to help “beat the market” is, likewise, out there and trying to make an impact.
Way back in 2017, an investment firm named EquBot started a new ETF that was a collaboration between themselves and the IBM Global Entrepreneur Program and IBM Watson to explore ways to integrate AI with investment solutions. EquBot launched the AI Powered Equity ETF (AIEQ) on October 17, 2017 to analyze millions of news, social media, and financial statement data points to strive to beat the broad U.S. equity market.
At D&A, we favor broadly diversified asset allocation to help minimize portfolio risk. The new AIEQ ETF seemed to fill a niche to possibly help diversify exposure to the broad U.S. equity market reflected by the S&P 500 (IVV). We felt that the AI model driving AIEQ would be neutral to behavioral bias and naïve capital market weighting and could serve as a good diversifier. Let’s see how it has done so far!
As seen from the chart below, through the end of 2021 the AIEQ ETF (the blue line) performed fairly well and even beat the S&P 500 (IVV, the green line) by a good margin (+7.00%!) in 2020. However, since the beginning of 2022, AIEQ has lagged significantly and ended up lagging the S&P 500 since its inception by a 6.00% annual deficit.
One of the criticisms of AI is that it is somewhat of a “black box” meaning we are not really sure how it works and the answers that come out of it may not be explainable. That is certainly the case here! I can hypothesize that since the AI model is neutral to behavioral bias, then perhaps it has underperformed recently since it needs a robust “fear or greed” behavioral component to help anticipate the market. Also, as we all know, the existence of unpredictable “external“ factors like the Russia/Ukraine war, a hawkish Fed, and domestic and global politics certainly impact the market in mysterious ways.
The one truism relating to AI is that they tend to get better over time as the models have more situations from which to learn. So, being patient is a virtue in this case and may be worth the wait.