How fast can Big Tech grow?

Best News And Media AffiliatesNov 29, 2023 2:35 am

0 0 4 minutes read

Unlock the Editor’s Digest for free

Good morning. RIP Charlie Munger, an icon of value investing and a very funny man. He urged investors, above all, to keep it simple. Many readers will know Munger’s famous line that you’d be wise to swap the term “ebitda” with “bullshit”. Another one of our favourites: “If you want a formula [for investing], go back to graduate school. They’ll give you lots of formulas that won’t work.” Send us yours: robert.armstrong@ft.com and ethan.wu@ft.com.

How fast can the Magnificent 7 grow?

Torsten Sløk, Apollo’s chief economist, wrote earlier this week in his useful daily email (sign up here) that the shares of the Magnificent 7 tech stocks “are beginning to look similar to the Nifty 50 and the tech bubble in March 2000” (the Nifty 50 was a group of super-hyped stocks that inflated a bubble in the early 1970s). Sløk’s a big shot, and people took notice. He presented this table, showing trailing-year price/earnings ratios for the three eras:

Sløk is quite right to bring up this comparison, and to sum it up in simple terms. Sometimes nuance is the enemy of insight. But we are valuation nerds, and can’t help but point out that p/e ratios have to be seen in the context of growth rates, and the Magnificent 7 are very fast-growing. The right question to ask is not “what is the right p/e multiple?” but instead “what growth rate of profits is it reasonable to expect for these companies?”

Consider this table:

Lines two to four of the table show the p/e ratios for the Magnificent 7, using consensus earnings estimates for this year, next year and for five years hence in 2028. Even using 2023 estimates, which are likely to be quite accurate given that we are in November, the p/e ratios come down a lot compared with Sløk’s trailing-year numbers — from 53 on average to 40 (mostly an effect of Amazon’s and Nvidia’s strong expected 2023 earnings). Still, that is still nearly twice the level of the S&P 500 as a whole, at 21. But the further out you go in time, the earnings estimates get higher and the ratios look correspondingly more reasonable. Using earnings estimates for five years from now, the average ratio falls to a normal-seeming 18 (unfortunately, none of the usual sources provide estimates for the S&P five years out, so a comparison with the index is impossible).

Cynical readers will point out that using p/e ratios based on earnings estimates for distant years is an old and dirty trick played by brokers who want to sell suckers on expensive growth stocks. The cynics are correct. A 2028 p/e ratio is only as good as the growth estimates backing it.

Are consensus growth expectations for the Magnificent 7 reasonable? Line five of the table above shows consensus estimates for five-year compound annual earnings growth rates for each company. For context, I have shown the actual compound growth rate of earnings for the past five years, in line six. The consensus expectation for the group is that it will grow at about half the rate in the next five years (18 per cent) as it did in the past five (34 per cent). But the group’s profits are now growing off a much larger base: in dollar terms, the expected profit growth is much greater than the retrospective.

It is useful to look at expectations for the Mag 7 against those for other large US stocks. Below is a selection of seven of the very largest:

As you see in line four, the 2028 p/e ratios are not that different: 16 vs 18. But embedded in that is an expected growth rate for the Mag 7 that is twice as high (18 per cent versus 9 per cent). Is that realistic? It is a core Unhedged tenet that growth is hard to predict and usually reverts to the mean. But this is so mostly because markets are competitive, and it is clear that the Mag 7 have incredibly strong competitive positions which will help them sustain high growth rates (so long as the markets they play in keep growing). I would not want to bet that the Mag 7 would revert to a growth rate of profits near the S&P 500 average (mid-high single digits) in the next five years.

Certainly it is reasonable to expect solidly double-digit profit growth annually. Take a few points of GDP growth, add a few points from operating leverage, a few points from stock buybacks, a few points for playing in markets that grow faster than GDP, and a point or two for taking market share, and you are over 10 per cent earnings growth. Looked at this way, earnings growth expectations for Apple and Microsoft (9 and 12 per cent annually, respectively) do not look crazy.

The other five companies have a higher bar to clear. Amazon’s very high expectations are almost surely based on the idea that it has been systematically overinvesting and under-earning for years, and that management will turn the dial towards profit before long. Could happen, but no certainty. Nvidia investors will be hoping that the biggest users of AI chips (that is, Apple, Microsoft, Google, Amazon and Meta) will not start building their own AI chips anytime soon. Tesla — well, who knows what Tesla investors expect.

Sløk’s point about the collective valuation of the Mag 7 is better framed in terms of expected growth than valuation multiples, but either way, expectations look peaky, in aggregate. Is the analogy with the 1972 and 2000 bubbles justified? I’m not sure. I’d frame the question in two parts: are the markets the Mag 7 are playing in growing faster or slower than the markets that, for example, Intel, Dell, Cisco and Microsoft were playing in in 2000? And are the competitive positions of the Mag 7 companies stronger or weaker? Keen to hear readers’ thoughts on either point.