Can AI Invest Better Than Me?
Key Takeaways
Using AI in investment processes is not a new thing with origins dating back to the 80s. For years, active investors and traders have attempted to get an information edge on others through AI tools that retrieve and process voluminous financial data.
At the moment, using AI to pick stocks is all the rage. However, there already had been an investment Exchange Traded Fund (ETF) that utilises IBM’s Watson to identify stock market winners running for the past 5 years. But the performance is quite underwhelming, with returns around 20% below and at higher risk than the United G Strategic Fund, a core fund used in our VaR portfolios.
Currently, the ability of AI to gather and process information and come to a decision is no different from what the millions of investors in the world do everyday. The market responds incredibly fast, and unless AI is able to predict the future, don’t expect it to provide market-beating performance.
One machine can do the work of fifty ordinary men. No machine can do the work of one extraordinary man.
We recently wrote a piece pondering whether with Artificial Intelligence or AI, humanity was on the cusp of a new revolutionary technological leap. One of the takeaways on whether we should be jumping in and investing heavily in this new technology was to approach it with moderation and to use time-tested diversification to ensure the exposure to any particular sector is not too concentrated.
If you ask anyone “what is AI?” the most likely answer you will get is ChatGPT. However there are thousands of AI companies and their associated platforms out there, all built with specific industry purposes.
With the exuberance in AI and its projected rise (seen in the diagram below), it is likely that more money and research will be pumped into this burgeoning field in the near future. Spending on AI software is expected to reach around US$40Bn in under 3 years.
With the amount of computing power that we can essentially harness, a key question we could ask is whether we can use AI to help us invest better, getting a leg up on the competition. After all, the origins of AI is much older than ChatGPT itself.
A defining moment on the ability of computers to outthink humans occurred in the late 90s when an IBM supercomputer named Deep Blue beat chess grandmaster Garry Kasparov during a widely televised event. For technology proponents, it was seen as a significant moment — where AI was able to defeat one of humanity’s intellectual giants.
IBM later used an improved iteration of Deep Blue to defeat the best contestants on a game show called Jeopardy! This win was likely more significant as it showed the ability of computers to understand complex and nuanced questions, much like the ability required in having a conversation with another person.
When it comes to investing, AI also has been used for quite some time by active investors and traders who have attempted to get an edge on others through the use of AI processes to retrieve and process data. There have been tools used from anything like measuring social media sentiment to capturing key words and information from company earnings calls and reports.
In fact, there is currently an investment Exchange Traded Fund (ETF) that utilises IBM’s Watson — the same computer that defeated Garry Kasparov and the Jeopardy contestants — to pick stocks in an attempt to outperform the general market. The marketing blurb for the fund sounds extremely impressive (seen below) — essentially claiming to be equivalent to a team of 1,000 analysts with the ability to sift through millions of data points.
With such power at your disposal, it is likely that you would be able to perform extremely well right? Actually…. not really. In the diagram below, the AI-driven ETF (AIEQ) underperforms a globally diversified investment — in this instance our core fund used for VaR portfolios — the United G Strategic Fund. Using standard deviation (or volatility) as a proxy for risk, the AIEQ is also much more volatile with an average 24% volatility compared to the 17% from the United G Strategic Fund.
So is Watson only good for games? Why does it do poorly in investments? We have written about this before — doing well in investing often differs from other facets in life. You don’t have to be the smartest, nor the fastest, to be the best. Investment success often stems from patience and temperament.
You might remark: ‘Computers have no meddling emotions, and probably can be more patient than us!’ But the ability of AI to gather information, process and come to a decision is no different from what the millions of investors in the world do everyday. While Watson may have decided to “SELL” with its available data on a company, a different investor halfway round the world may have decided that “BUY” was a good option instead. It does not mean that everyone would come to the same conclusion even after reviewing the same data. Once other new information is released, the process of acting on that information (buying or selling) is immediately incorporated into market prices.
The example below shows how quickly information is incorporated into market prices. On 17 Nov 2020, the S&P Dow Jones Indices announced that Tesla would be added to the S&P 500 prior to the market open on Monday, Dec. 21, 2020.
You can see that immediately on the announcement date, there was a big spike in Tesla’s share price as investors rushed to hold the company. Investors around the world all acted on the same information as it was released. So unless AI had the ability to read the minds of the committee members at S&P, or predict this addition in advance (which it did not), then the computer would need to react to the same information available at the same time as everyone else.
AI has excelled in areas such as forecasting weather patterns, doing so much more accurately than compared to existing physics based models. However, AI is far less likely to successfully predict changes within complex systems with numerous inputs and variables such as stock and bond markets.
So like the quote at the beginning of the article, AI at the moment cannot usurp the disposition of the patient and diversified investor, i.e. the extraordinary man. Maybe in the future it will, or maybe it won’t, but at the moment this technology is not the holy grail to investment outperformance. Evidence based methodologies like broad diversification, focusing on lower priced assets with higher profitability have been and still are proven to give you the best results.
If you would like to find out more about how your investments are constructed and how to optimise your long-term returns, come and have a chat with us.
