The valuation mystery: more clues

Good morning. After Gamestop — a barely profitable company whose revenue fell by a fifth last year — announced Friday that it has raised $933mn in an equity offering, its shares enjoyed a 25 per cent jump yesterday. Every time something this stupid happens, an innocent little finance fairy dies. Email me with reasons to believe the market makes sense: robert.armstrong@ft.com

Why valuations changed, part 2

Monday’s newsletter, about the regime change in US stock valuations in the 1990s, drew more reader comments and emails than Unhedged has ever received before. People are really interested in this topic, for reasons that are unclear. Whatever the reason, a lot of ideas passed through my inbox.

A majority of respondents took the view that the big step-up in valuations in the past 30 years comes down to one or more of three causes.

Probably the largest group of readers responded that it’s as simple as interest rates. Whatever model for stock prices you use, the risk-free rate is an important part of it. When rates fall, future cash flows become more valuable today.

My colleague Robin Wigglesworth referred me to a paper by the Fed economist Michael Smolyansky (which Unhedged has talked about it a bit before). The paper argues, among other things, that the change in interest rates accounts for the entirety of the increase in price/earnings multiples between 1989 and 2019. The argument is simple and brief. The price/earnings ratio can be analysed, using the basic dividend discount model, as follows: 

price/earnings = (payout ratio) * (1+growth rate)/(risk free rate + risk premium - growth rate) 

Between 1989 and 2019, Smolyansky points out, the 10-year Treasury yield fell from 8 per cent to 2 per cent. Using that yield as the proxy for the risk-free rate, that change alone, keeping everything else in the above equation constant, is enough to explain all of the observed change in the S&P 500’s trailing PE ratio over the same period (it rose from 12 to 20). 

This is very tidy. But everything else is definitely not constant. Specifically (as this newsletter never tires of pointing out), risk-free rates, growth rates and risk premia are closely and causally, if not very reliably, related. Very low rates often result from very low growth, for example.

This does not make Smolyansky’s basic argument hopeless. In a paper from 2017, Rob Arnott of Research Affiliates and two co-authors note that cyclically adjusted P/E ratios seem to have a curved or “mountain-shaped” relationship with both real interest rates and inflation. P/Es are highest when both are around 2-3 per cent, tailing off when they are higher or lower. Arnott explained this “sweet spot” in an email: 

This is presumably the range where risk tolerance goes up because economic uncertainty is low. Deflation or rapid inflation are terrible for P/E ratios. And if real interest rates are negative or very high, bad things are happening in the economy. For most of the last 30 years, we have been in that sweet spot.

As I pointed out on Monday, rates and inflation did recently rise above 3 per cent, and valuations rose rather than fell. But that may be down to long-term inflation expectations remaining under control even in the worst of the inflation surge.

Another large group of readers argued that the change in valuation multiples reflects a change in the financial structure of US companies. Many companies now have capital-light business models: technology, globalisation and the shift from manufacturing to services all reduce capital needs. Capital-light businesses have higher returns on equity; higher returns on equity mean that reinvested profits compound faster, leading to higher growth; higher growth merits a higher valuation multiple. 

Nick Sheridan wrote to point out that share buybacks, which started to take off in the 1980s after the SEC gave boards the official all-clear, push equity out of businesses and increase returns on equity. This might have increased financial leverage and risk — but was, happily, accompanied by an offsetting fall in interest rates.

Another reader pointed me towards a recent paper by Michael Mauboussin (who Unhedged interviewed here), which explains why the ROE point is not just about growth. The move to more capital-light business models makes valuation multiples less representative because the accounting treatment of these businesses tends to be different: 

Companies create value when their investments earn a return in excess of the opportunity cost of capital . . . [valuation] multiples provide no direct insight into the magnitude of a firm’s investments or whether they will generate a sufficient return . . . Today, the majority of investments are in intangible assets, including customer acquisition costs and branding. But companies commonly expense these investments on the income statement as they incur them. Accountants record these investments as selling, general, and administrative and research and development expenses. This reduces current earnings . . . the rise of intangibles means that both earnings and invested capital are understated, weakening the signal that earnings and multiples formerly provided

Of course there is an empirical question here. How much has return on equity at large US companies risen since the 1990s, and is it enough to explain a big change in multiples? If you have that data, send it along.

A final group of readers wrote to argue that the shift in valuations is driven by demographics. As the population ages and more people enter their prime earning-saving-investing years, demand for financial assets increases and their valuations rise. The point is intuitive enough (although there is a debate about whether it is demographic shifts or higher inequality that drives the savings accumulation). Arnott has views on this, too. He argues that

Markets command higher multiples (and deliver lower yields) when more people want to save than to dissave. That’s the 40 to 65-year-old cohort. As life expectancy rises, the imperative to save for retirement becomes stronger. The result is that investors are more willing to accept lower real interest rates and higher valuation multiples. Children and seniors, because they don’t work, impose dissaving on the working age population, which is why the 20 to 40-year-old cohort (parents!) doesn’t help market valuations (and children and seniors depress valuations)

All three explanations — which work pretty well together — demand we ask the same question. How long will these effects last? The shift to less capital-intensive business models might be here to stay. But rates and demographics (which are of course related) may not remain as equity-market friendly as they have been in the past three decades.

One good read

Some evidence of a private credit bubble.

FT Unhedged podcast

Can’t get enough of Unhedged? Listen to our new podcast, for a 15-minute dive into the latest markets news and financial headlines, twice a week. Catch up on past editions of the newsletter here.