WHEN he took his first space walk, NASA astronaut Reid Wiseman had a revelation. “I used to think I was scared of heights,” he said. “Now I know I was just scared of gravity.” This summer might be a good time for investors to think about this a little: Gravity has a tendency to work on equity bubbles rather as it does on astronauts before they escape the Earth’s atmosphere. They all come back to earth. From the niche (think the Beanie Baby bubble of the 1990s) to the mainstream (the US housing bubble), the same thing happens over and over again. There’s a great story. Everyone loves the story. Everyone buys. The reality doesn’t quite match the story. The asset class collapses. Up a lot. Down a lot.
It isn’t hard to spot a bubble forming. But experienced investors will know it’s very very hard to spot when it might finally come a cropper. That’s partly because it doesn’t really need a catalyst; it doesn’t require political disruption, financial scandal or shifts in monetary policy (although there is plenty of that about, of course). The end, as Societe Generale AG’s Albert Edwards points out, is often remarkably simple: “A reversal in price momentum in an asset class that has risen sharply for a number of years (sucking in huge quantities of loose money) is often sufficient in itself to cause prices to crash.”
When there is no one left to buy or when a few of those who might have already bought become a little nervous, it all comes crashing down. In that sense, you could think of all bubbles as a type of (legal) Ponzi scheme: As new investors become thin on the ground and a few of those already in look to get their money out, the whole thing collapses pretty fast. Gravity can be brutal.
So here we are again. Back in 2022, it rather looked like the US tech bubble had been dealt with in the normal way: Following a sharp rise in interest rates, the Nasdaq 100 was 35% off its highs, for example. Then came ChatGPT and a wave of optimism that brushed rate worries to the side in a rush to embrace the idea that a new world is just around a very close corner.
Since the beginning of 2023, the US tech sector is up over 100% and the S&P by 50%. You can argue that this makes sense. Artificial intelligence (AI) could transform our world (finally some productivity gains and some real growth) and tech earnings really have risen strongly since 2022. But, it’s also true that we have been here before. Those who recall the 1990s will also remember, says Edwards, that the last big tech bubble was fueled by investment in what turned out to be excess capacity. Sure, the story was exciting and, sure, earnings rose. But the story took much longer to play out than expected. and earnings never rose quite enough to justify valuations.
Is that happening again? Of course. You can see the bubble in the concentration of the US market: The tech sector now makes up 35% of the S&P 500. You can see it in valuations: The US market is trading on an end-of-2025 forward price-earnings multiple of 19.8 times, well above historical averages, says SocGen quantitative strategist Andrew Lapthorne. That means “there is little wiggle room” for any downgrades. And you can see it in expectations. Look at it in terms of the 1990s Information technology revolution and you will get the idea. That revolution “boosted real potential growth by roughly a percentage point for a few years,” say the analysts at BCA, an investment-research firm. Model something similar for AI over a 10-to-20-year time horizon and you see that it is worth somewhere between $3 trillion and $10 trillion to the corporate sector (this isn’t an exact science).
But US growth stocks have already seen a $4.3-trillion rise in market capitalization since 2022 and the market has a whole a rise of $7 trillion. This suggests that “the US equity market is significantly overvalued” — unless we do really see evidence of a huge productivity surge accompanied by “persistently high margins,” says BCA. That might still happen (and those of us wishing to maintain our living standards must hope it does). But even if it does, will it be clear in time to make sure new money keeps pouring into this market this year — and will it really justify the excitable valuations?
There isn’t much short-term certainty in markets — and there is little on the immediate direction of the AI bubble. But there is some long-term clarity: Value stocks underperformed firmly from 2007 to 2020, but over the long term, cheap stocks do win. As the authors of this year’s UBS Global Investment Returns Yearbook say: “We have seen that over the long run companies selling at a low stock price relative to fundamentals — value stocks — have beaten stocks selling at a high stock price relative to fundamentals — growth stocks.”
From 1926 to the end of 2023, the outperformance of value overgrowth was 2.6 percentage points, annualized (12.7% vs. 9.8%) — and that’s despite underperformance pretty much every year in the 21st century. Look to the UK and, albeit with a shorter time frame (1955 to 2023), you see much the same (14.8% vs 9.8%); value investing has paid off. The important thing for investors to remember is that, while it makes long-term sense to always participate in stock markets, you don’t actually need to participate in the bubbles (however tempting). You can simply note them and buy something else: Should the productivity revolution appear, all sectors will benefit. Think of it as Wiseman might: The less far you have to fall, the less frightened you need be of gravity.
BLOOMBERG OPINION
