There’s a question I get asked very frequently at investor events: why did we choose to float Raspberry Pi, the company I co-founded, on the London Stock Exchange rather than in New York, where valuations appear to be much higher?
Surely we were leaving money on the table by IPO-ing in London? And I really enjoy being asked about it, because the smoke and mirrors around our perception of a valuation gap between the two exchanges are the same smoke and mirrors that are fuelling the current AI bubble; and it’s fascinating to talk about.
Easy bit first. There isn’t a valuation gap: and it’s simple enough to demonstrate that there isn’t without even having to do any maths. If British stocks were really attracting artificially low values, American investors would be doing a ton of arbitrage to take advantage of that fact. Looking around, it’s evident that that isn’t happening.
What IS happening is that some very specific US stocks are overvalued – massively, headshakingly overvalued – and they’re all in the AI sector, or adjacent. Those Uber and Tesla valuations are based on the expectation of full self driving (FSD) coming along soon so all those messy humans who drive you home when you’ve had one too many in the evening can be disintermediated by robots. The hardware companies whose valuations have gone into orbit are all providing infrastructure for AI, and the valuations are backed by the firm belief that any moment now, the whole of the midwest is going to be carpeted with datacentres. And so will space. While OpenAI hasn’t IPO’d yet, it plans to later this year, and the valuation after its most recent round in March was $852 billion.
I’m not the only person who believes this to be a bubble that’s going to crash hard; smarter investors than me are making that prediction too.
If you look at the market as a whole, the outliers make the whole US market look as if it holds far more value than the FTSE. But if you take two comparable stocks from both exchanges, things suddenly look a lot more rational. (It is not, for the record, rational to believe that full self driving FSD is just around the corner, or that Elon Musk’s nightmare tangle of operations is going to produce a bipedal robot slave butler with artificial general intelligence real soon now, especially given his history of promising things and not delivering on them.) But the US stock market is wholly captured by the glamour of the godlike AI it’s anticipating (I have a theory about this, as well: Silicon Valley folk grew up reading a lot of sci-fi, and a lot of what’s being predicted is the epitome of the sci-fi utopias we have been promised over the last few decades), even though some of this stuff is starting to feel a bit Emperor’s New Clothes.
The numbers and what they actually mean
The big fat headline comparison that drives the whole “London is cheap” narrative is the raw P/E gap. At the start of this year, the FTSE 100’s trailing P/E ratio was around 17.7, against the S&P’s 28.3. On the face of it that looks terrifying for those of us in London, until you look a little closer and realise that that comparison is doing absolutely no analytical work, because the two indices contain fundamentally different kinds of company.
The FTSE 100 is weighted heavily towards companies which move slowly, predictably and steadily in sectors like energy, finance and industry. The S&P 500 is dominated by very fast-growing tech stocks. Technology firms account for about 1% of the FTSE 100 (Raspberry Pi is a FTSE 250 company); they account for roughly 30% of the S&P 500. If you’re in a high-tech sector with fast growth rates, you’ll axiomatically cary a high P/E ratio. James Arnold at UBS calls this comparing British apples with American pears.
The Magnificent Seven now account for around 34% of the entire S&P 500. At the same time, they return only 27% of the index’s earnings. The gap between market weight and earnings contribution is key to figuring out what’s going on with the valuation puzzle. And the distortion gets deeper when you look hard at the cyclically adjusted price-to-earnings (CAPE) ratio, which smooths out the effect of booms and recessions by using average real (inflation-adjusted) earnings over the past ten years too: if you take Nvidia alone out of the CAPE calculation, the CAPE for the whole of the rest of the S&P 500 falls nearly three points. (Nvidia’s a particularly extreme case here; its CAPE ratio is currently 293, which seems to have absolutely no connection to any valuation norms of any other sector in any market.) Take out all of the Magnificent Seven stocks, and CAPE falls from 41 to 33.
So without the Magnificent Seven in the picture, the US market is already looking far less intimidating. This January, the forward P/E for the S&P 500 was 22.2. Without the Magnificent 7 it was about 19.5. Meanwhile, the FTSE 100’s forward P/E in January was around 13.35 – but remember this is a market with very little tech representation, and a relatively massive weight given to energy, mining and financials, which are all sectors which structurally attract lower multiples everywhere in the world.
David Schwimmer (not that one, but the one who’s the CEO of LSEG) has said: “The notion that you get a better valuation in the US – it’s a myth. If you look company-by-company and adjust for growth rates and other factors, London is at, and in some cases higher, in terms of valuation than the US.”
I poked around for some comparable companies outside the tech sector in both markets. Experian (LSE) and Equifax (NYSE) are very similar businesses: global consumer credit bureaus/analytics organisations with similar subscription revenue models, similar data moats and similar business models. Experian is larger by international footprint.
Experian has a trailing P/E of around 30x and a forward P/E of around 23x. Equifax’s trailing P/E is around 27x and its forward P/E around 20x, with Experian commanding a slight premium due to stronger global diversification. Which is to say: the analysts who model these businesses peer-to-peer see Experian, the London firm, priced at a premium, not at a discount.
What about another sector? I am, as you’ll see from the name of the website you’re visiting, a fan of cocktails, so I thought it’d be worth having a look at the spirits industry. Diageo (Smirnoff, Tanqueray, Guinness and a whole lot more) are listed in London; and the New York-listed Brown-Forman (Jack Daniels, Woodford Reserve) are probably the closest comparators in the sector. They both have some very strong, globally recognised brands and international distribution.
Diageo is at the time of writing trading at a P/E of around 24x. Brown-Forman’s current P/E is around 16x: lower than the London-listed company’s. (Caveat: there’s been a drop in the fashionability of brown spirits, which may explain some of the compression.) But the point stands: the UK-listed companies are not being penalised, and in fact it looks in this instance like the narrative is running in the other direction.
There are other exemplars in other sectors. Rolls-Royce trades at a higher multiple than GE Aerospace, a US peer; AstraZeneca (LSE) trades at a higher multiple than Merck (NYSE).
The most important nuance of all is that while aerospace, drinks, pharmaceuticals, finance and other sectors all have representation across both markets, there is no London equivalent of an Nvidia; there is no cloud hyperscaler; no massively scaled platform software company. So what many are reading as a valuation gap is no such thing: it’s a sectoral absence (and possibly, given worries about what happens when the AI bubble bursts, a sectoral absence that is, in the medium- to long-term, going to be extremely good for the UK market). The UK simply doesn’t have those companies listed, and when you compare what’s actually there, the valuation gap vanishes in a puff of logic.