The views expressed are those of the author at the time of writing. Other teams may hold different views and make different investment decisions. The value of your investment may become worth more or less than at the time of original investment. While any third-party data used is considered reliable, its accuracy is not guaranteed. For professional, institutional, or accredited investors only.
“Economic progress, in capitalist society, means turmoil” – Joseph Schumpeter
The stock market has been captivated by the potential of artificial intelligence (AI) to transform the US economy. By one measure, all the gains in the S&P 500 Index this year have come from so-called “AI beneficiary” stocks. Optimists argue that AI is as revolutionary as the internet was in the 1990s and could lead to a productivity renaissance that has been missing for the last two decades.
The full impact and final-use cases for AI are nearly impossible to know at this point, so we are keeping an open mind to the possibilities of the technology. In this piece, I will discuss underappreciated implications if AI is truly as transformative as the stock market seems to believe.
Most investors are using a mental model of the 1990s to think about how AI could impact asset prices. This Goldilocks environment of high productivity/growth and low inflation created one of the most dramatic bull markets in history. However, there are two key differences between asset prices now compared to their pre-internet starting point in the early 1990s:
US equities have done well for the past 13 years despite low productivity growth. In fact, there were a lot of economies that grew faster and were much more productive than that of the US after the global financial crisis. However, industry concentration has grown substantially in the US over the past few decades — a trend that increased in the last decade with the large tech companies. The market doesn’t pay high valuations for productivity and growth; it does assign high prices for dependable cash flows and strong moats.
AI could potentially disrupt this dynamic. It may decrease the cost of new entrants into what were previously conceived of as impervious business models, particularly in many of the highest-margin segments of the market, such as software, services, and capital-light models. As Joseph Schumpeter, the economist responsible for penning the term “creative destruction” said, “economic progress, in capitalist society, means turmoil.” The point is that AI might mean stronger US growth and a more productive economy, but that is not necessarily a good thing for stocks if it erodes highly valued moats.
A rarely discussed aspect of the potential AI revolution is energy. There is a strong, exponentially correlated relationship between gross domestic product per capita and energy consumption. In other words, if AI is truly transformative to productivity, it would mean a substantial increase in energy consumption, which the market has not focused on yet.
In fact, historically, major productivity advances have tended to coincide with energy discoveries in cheaper and denser fuels, for example, coal in the 1800s and oil in the late 1800s/early 1900s. Perhaps shale could serve that purpose now, but companies need to develop those resources if productivity is truly going to rise over the medium to long term.
We are, in fact, seeing the opposite with shale decline rates intensifying and production growth slowing. In other words, while the conventional wisdom is that AI will be disinflationary, and likely will be for labor, in the short term it could increase energy prices.
The market has so far traded AI’s potential with the 1990s playbook that was marked by strong equity prices, tighter credit spreads, and lower inflation expectations. This has led to a flatter yield curve. Many of the following may prove to be the opposite if AI is truly transformative:
Perhaps a prudent manifestation of this view is to consider five-year US Treasury Inflation-Protected Securities (TIPS). These bonds could benefit from higher headline inflation, but also a looser labor market caused by AI. This could mean the US Federal Reserve might be less aggressive, as its focus has been on wage-driven core inflation.
At the same time, there may be a compelling case for selling long-term corporate bonds, where the credit risk premium is too low given AI’s longer-term threat to business models. The yield curve could steepen given elevated long-term growth expectations. The yield differential between five-year TIPS and longer-term corporate bonds is currently at its lowest level since TIPS were first created as an asset class. In my view, this represents a very attractive entry point.
Stay up to date with the latest market insights and our point of view.
