I generally don't make predictions, but one thing I can say with absolute confidence is that artificial intelligence will radically reshape our work and daily lives.
It will, for instance, warn us of diseases like cancer or Alzheimer's years before symptoms appear, and recommend preventative steps, extending the average human lifespan.
AI will also make education much more accessible, breaking language and geographical barriers by delivering customised content through virtual learning platforms. And it will autonomously control urban infrastructure, reducing congestion, pollution and energy consumption.
Another prediction I can make without hesitation is that most investors who try to capitalise on the growth of artificial intelligence will underperform investors in simple index funds, and many will lose very substantial sums of money.
History keeps repeating itself
You might be thinking that those two predictions are completely contradictory, but they aren’t. There are many examples over the last 200 years of game-changing technologies which greatly improved people’s lives yet also produced speculative bubbles that cost investors dearly when they eventually burst.
Here are just a few examples.
For early Victorians, the advent of railways must have been hugely exciting, making long-distance travel and trade faster, cheaper and more efficient. No wonder investors piled into railway stocks in the early-to-mid-1840s. Yet overbuilding, mismanagement and competition caused prices to fall. Many rail companies went bankrupt, wiping out investors' capital.
The years either side of World War I saw the early development of the motion picture industry. Cinemas opened in major towns and cities, providing cheap entertainment for working-class audiences. But only a few studios, like Paramount and MGM, succeeded, and many early investors in film production, cinemas and equipment companies suffered heavy losses.
The growth of civil aviation opened up a world of opportunities for ordinary people. The so-called Jet Age in the 1950s and 1960s saw many invest in airline stocks, and the sector’s popularity increased when deregulation began in the 1970s. But only a small number of airlines consistently generated large profits, and Warren Buffett famously described the industry as a "death trap" for investors for decades.
In the late 1990s, investors couldn’t buy enough internet stocks. And yes, I was one of many who bought a technology fund. We were absolutely right in thinking that the internet was going to change our lives beyond recognition, giving us access to information, communication and e-commerce. Yet prices crashed in the early 2000s and most investors lost out. The fund I was invested in fell more than 70 percent in value.
The period immediately following the millennium was an excellent time for homes and businesses to invest in solar technology; it was, however, a bad time to invest in shares in solar energy companies. The solar industry has been dogged by over-capacity, falling prices and aggressive competition from China. Several high-profile firms went bust, including Solyndra in the U.S. and Q-Cells in Germany.
It's not enough to be right
All of these examples have two things in common. First, all were high-growth industries that delivered transformative benefits for consumers; and secondly, all resulted in substantial losses for many investors who tried to ride the wave.
In other words, it wasn't enough for investors to predict correctly that these technologies would be game-changing. They also had to identify, in advance, the very small proportion of companies which would survive and thrive.
Regular consumers of financial media will know that fund management companies are very keen to promote artificial technology funds — AI-themed ETFs in particular. Unsurprisingly, they like to emphasise the potential of AI to change our lives for the better. But students of financial history will realise that, even if the positive impact of AI exceeds expectations, as it may well do, it will make little difference for investors. Why? Because picking the very few firms that will dominate this space, ahead of time, is extremely challenging.
Should you buy NVIDIA?
You may of course be wondering about NVIDIA. The California-based company provides a wide range of AI-related products and services, and its share price has risen by more than 5000 percent in the last decade. So why don’t people just invest in NVIDIA to ride the AI wave?
The answer, simply, is that NVIDIA has already produced outstanding returns. The growth in its share price reflects enormous confidence among investors that it will carry on delivering healthy profits going forward. The higher the price you pay for a stock, the lower the expected return. So the challenge for investors in search of higher returns is to identify today the NVIDIAs of the future.
This point was made by none other than Eugene Fama, the Nobel Prize-winning finance professor, in a recent interview with the Financial Times. Fama is best known for developing the Efficient Market Hypothesis, the idea that, because all known information is already embedded in market prices, identifying stocks that are either undervalued or overvalued is effectively impossible.
Asked for his opinion on the extraordinary rise in NVIDIA’s share price, Fama
replied: “The world is betting that AI is going to rule the world and that NVIDIA will have a near monopoly, but who really knows?… Most of the prices were too high (in the dotcom bubble), but some were too low as well. Some companies made up for all the mistakes that were made on the other ones.”
In other words, the premise proved to be correct: the internet create huge, vastly profitable new companies. But the problem investors discovered was that identifying the likes of Amazon and Google before they became household names was like looking for needles in a haystack.
Takeaways for investors
The lessons for investors are these. First, don’t be distracted by investment themes, no matter how compelling you think the story is. Whether it’s AI, self-driving cars or lab-grown meat, the logic for investing will always seem persuasive. But logical doesn’t mean profitable.
Secondly, think about the cost of investing. Thematic funds usually come with slick marketing, but also significantly higher fees than simple low-cost index funds or passively managed ETFs. Yes, the manager might pick some future outperformers, but they’re bound to invest as well in stocks that will go to the wall. After costs, your chances of beating a broad market index tracker over the long term are very slim.
One final thing all of this should teach us is the value of taking a long-term view and not trying to dip in and out of different themes, sectors and countries at the right time. Market timing is fiendishly difficult. With the benefit of hindsight, the ideal time to invest in artificial intelligence was in the 1980s and early 90s, before people were really talking about its commercial potential. By the time a technology is headline news, you’re almost certainly too late to prosper from it.
© The Evidence-Based Investor MMXXIV. All rights reserved. Unauthorised use and/ or duplication of this material without express and written permission is strictly prohibited.