2023’s equity market is a story of “Magnificent 7” outperformance, and for many that’s problematic.
A narrow rally is an unsustainable rally. Breadth is poor, and that’s not a sign of a market in fine fettle. And so on.
Morgan Stanley’s Mike Wilson will tell you all about it. “The outsized contribution of just seven stocks to the market cap weighted benchmark’s return (and earnings) has left this year’s equity market performance one of the narrowest on record,” Wilson wrote, in his 2024 outlook. “Such weak price breadth is not indicative of a healthy bull market.”
Context is key, though. In October of 2022, I mentioned in passing that I was buying Meta shares. Some readers were surprised (aghast, even) given my avowed disdain for Facebook and penchant for lampooning Mark Zuckerberg’s metaverse ambitions. My rationale was simple: Meta is a bluechip and at the time, it was down 75%. The risk-reward asymmetry was highly skewed. Meta went on to nearly quadruple within 12 months.
The point certainly isn’t to tout my stock-picking acumen. I’m a lot of things, but a stock-picker most assuredly isn’t one of them. Rather, the point is that 2023 is in many respects just a reversal of 2022 with an assist from the A.I. narrative (or the A.I. pipe dream if you’re skeptical).
“The massive outperformance of mega-cap technology stocks has been a defining feature of the equity market in 2023 [but] the dominance of mega-cap tech largely reflected a reversal of meaningful underperformance in 2022,” Goldman’s David Kostin noted in his own year-ahead outlook. “The Magnificent 7 stocks collectively returned -39% in 2022 compared with a -11% return for the remaining 493 constituents in the S&P 500.”
That doesn’t mean the Magnificent 7 don’t constitute a bubble (they very well might), but we shouldn’t forget that Meta (to stick with the same example) was trading at $112 this time last year, down from $382 a mere 14 months previous. That was one helluva selloff. From November of 2021 (the peak of the pandemic “everything bubble”) to November 2022, Amazon fell 55%. Tesla fell 75% in the 14 months to January 2023. A bounce in those names (and their brethren) was a foregone conclusion.
The question now is whether 2023’s outperformance counts as excessive even in the context of 2022’s almost existential declines. When viewed through the lens of market concentration, it certainly seems that way. As a share of market cap, the top seven comprise nearly a third of the index, the most ever.
“The 30ppt of outperformance of the seven largest stocks at the start of this year vs. the rest of the index YTD ranks as the second-largest annual difference since 1970,” Goldman’s Kostin noted.
If you ask SocGen’s Albert Edwards, that’s ominous. Or, actually, the Magnificent 7’s valuation premium is ominous. “So reminiscent of the 2000 Nasdaq bubble,” he said Thursday.
Edwards has, from time to time, warned that one ostensible mitigating factor for mega-cap valuations (their projected growth rates) may be naive. Kostin pointed to the Magnificent 7’s relative PEG ratio, which is in line with the 10-year median.
As the figure on the right shows, the Magnificent 7’s relative P/E ratio ranks in the 91%ile on a 10-year lookback. However, the earnings-weighted ratio (using long-term expected EPS growth) is middling.
But that assumes expected growth turns out to be… well, as expected. In the dot-com bust, such assumptions proved faulty, something Edwards isn’t inclined to let the sell-side forget. He called earnings growth expectations for the Magnificent 7 “hope” on Thursday.
While Edwards’s concerns may be valid, the problem with any long-term bear case for these companies is always the same: They’re synonymous with everyday life, and look poised to become more so over time.
2022’s selloff wasn’t anything to do with tech somehow becoming less pervasive as a fixture of human existence. Rather, the slump was a function of the sharp run-up in real yields weighing mechanically on multiples, impossible top-line comps and bloated cost structures (e.g., Amazon and Meta) tied to misguided extrapolation of the pandemic operating environment. Margins for the mega-caps contracted more than 500bps in 2022 versus 60bps of margin expansion for the rest of the index.
Kostin was keen to note that even the seemingly intuitive relationship between the mega-caps and real yields actually isn’t all that reliable, and where it exists, it may be a false optic. “Sometimes the market has treated the Magnificent 7 as a long duration asset, with the group underperforming as rates rise,” he wrote, citing 2022 as an example, before quickly noting that because earnings were falling concurrently, it was difficult to discern what was actually behind the group’s underperformance. Investors, Kostin suggested, might’ve “overstated the impact of rates” on the mega-caps last year and besides, the market could just as easily reward the Magnificent 7 for their “quality” characteristics in an environment where elevated rates imperil companies with weaker balance sheets and steep maturity walls.
As for the notion that consensus may be too optimistic on the group’s growth prospects, Kostin conceded the point and didn’t shy away from the dot-com parallel. “In 2000, the eventual underperformance of the mega-cap market leaders occurred when those companies failed to meet elevated growth expectations and multiples collapsed,” he wrote. “If the Magnificent 7 repeat this dynamic and disappoint expectations in 2024, those stocks’ valuations will likely ‘catch down’ towards the remainder of the index and underperform.” Kostin cited possible A.I. disappointment and antitrust proceedings as potential catalysts for lower growth expectations.
But as it stands, growth estimates for the group are so far ahead of the so-called “S&P 493” that it almost doesn’t make sense to make the comparison.
The figure on the left suggests sales will grow at a CAGR of 11% through 2025 versus a mere 3% for the rest of the S&P 500. The figure on the right points to net margins for the Magnificent 7 that’ll be double the market for the foreseeable future.
Of course, that’s an exercise in question-begging: The risk is that those estimates turn out to be misguided.
In that regard, Kostin’s crystal ball is no more useful than anyone else’s. “The key determinant of whether the mega-caps will continue to outperform next year will be how realized sales and earnings growth compare with current expectations,” he wrote. His rear view mirror is working fine, though. “In recent years the trajectory of earnings has explained the performance of the Magnificent 7 relative to the rest of the market,” Kostin added.





I would not bet against the magnificent 7 or the other 493 right now. I spent a couple days at a conference last week and also watched a few presentations on what vendors in my space are doing with AI and all I can say is that the innovation in generative AI is happening fast and the impact will be felt across industries. I suspect we’ll start seeing that impact in the earnings estimates for 2024, and it’ll expand more broadly in 2025.
From an employment perspective, AI is going to do to white collar jobs what outsourcing manufacturing to China did to blue collar jobs. I know historically innovation has created new opportunities, but there is going to be a lot of displacement which will have some big winners but a lot of losers.
One obvious example is programming. Can AI produce quality code right now? Maybe not on its own, but a 10x programmer with AI? Now that 10x programmer might become a 20x or 50x programmer compared to last year’s standard. An average programmer might still be able to multiply their effectiveness by a smaller factor, but the fact that they are likely to be more productive reduces the company’s headcount needs, organizational complexity, and competition for talent. With fewer people who need to be involved, the efficiency gains are compounded.
Another example is sales. If every sales rep effectively has their own assistant to take notes, automate follow ups, identify the right prospects, and tailor content, that significantly reduces the administrative overhead. Now I don’t need as many people and I can keep my best reps. Suddenly, I’ve got lower headcount cost, my deals close at a higher rate, and I don’t need as much support staff. Again, the efficiencies compound and margin and growth both go up.
That’s at a micro level, but what about the macro? Well, I think we’ll initially see some major productivity and profitability gains across the economy in the near term, but what happens if people are displaced en masse? I doubt UBI would become a reality soon enough to offset what will likely be a significant increase in unemployment. We’ll also see new markets that are built on hyperpersonalization (think travel agents or book writing or video games), but I’m not sure how much employment that will create. I would not be surprised if we are once again struggling with near zero inflation and ZIRP.
I’ll just end by saying that the CEO of the company that was hosting the conference said during his keynote that they don’t think their innovations will eliminate jobs, but I’d take the opposite side of that wager…or just bet on the market since that’ll likely benefit from the combo of AI’s impact on productivity and the return to ZIRP.
IBM recently commented on the takeup of their AI offerings (built on the old Watson base). Like you, they pointed at coding, customer relations and human resources management. All important functions, but a far cry from predictions by academics and sell-siders speaking of AI revolutionizing every aspect of life. Or, so we should hope!!
On a related note, what a time to be a tech worker! First “outsourced to India” and now this. And to think coding and such was once widely seen as a lucrative and stable career.
But at least one wiser person than I sides with Mr. Dayjob:
https://finance.yahoo.com/news/ai-create-largest-displacement-human-193018798.html
The fact is that since 1/1/2022, SPY is down about 5% and the Magnificent 7 are, in aggregate, up about 4% – almost entirely due to Nvidia more than offsetting the declines in Tesla and Amazon. The remaining 4/7 are currently priced pretty close to their prices as of 1/1/2022.
From 30,000 feet, a number of plausible stories could be written based on this fact pattern.