Pumped! Why the market loves crazy valuations (until it doesn’t)

There are two popular misconceptions about speculative bubbles. First, that those who participate in the bubble do not know they are in it. And second, that a bubble can only be identified as such with certainty after it has burst, due to new information having come to light.

These misconceptions find purchase not only in the minds of some investors who sit out the speculative mania (the “everyone else is dumb” approach), but also in those of certain academics whose neural circuits resonate with efficient-markets ideas and who have tried for the last two decades to demonstrate (unconvincingly) that the dot-com bubble was in fact consistent with the efficient market hypothesis (the “I have a theory and I want reality to fit it” approach).

We submit that during real speculative manias both theories are wrong. We don’t think it takes any great perceptiveness to say “this is unlikely to end well”. We therefore further submit that:

1. During a bubble, almost everybody knows things are a little crazy.
2. Those who participate in the speculation do so for reasons that are sort-of rational, and that most of the people who can’t see this are too intelligent to do so.

We might as well come off the fence: we think that developed market indices overall are expensive, and that there is quite clearly a speculative bubble now in several non-trivial corners of the market: early-stage software firms, pure-play electric vehicle firms, Bitcoin, penny stocks, IPOs of almost any kind, practically anything backed by Softbank, and the occasional epic aberration in near-bankrupt firms.

Consistent patterns of behaviour

These feelings may be subjective, but a review of past bubbles appears to back them up. A 2017 Harvard paper by Greenwood, Shleifer and You showed that certain behavioural patterns appear consistently in market bubbles. The title is 'Bubbles for Fama' and the authors find that:

This finding is interesting for three reasons. Firstly, it describes pretty much exactly the type of market environment we find ourselves in today (minus perhaps the volatility, although volatility is clearly elevated in the most speculative segments). Secondly, fundamentals are of no use when trying to distinguish between bubble and non-bubble episodes, meaning that an attempt to explain a fast-rising market using fundamental metrics probably focuses on the wrong things. Thirdly, these situations tend to end badly.

Now, we stated at the beginning that we think everybody knows when they’re in a bubble. It’s obviously hard to know what everybody thinks, but we can prove that the opposite is false: we can show that it’s not enough for people to know with certainty that they are gambling for them to stop. On 7 January, WhatsApp updated its terms and conditions to allow the app to share more data about its users with Facebook. In response, Elon Musk tweeted “Use Signal”, which is another messaging app run by a private foundation Musk sponsors. Unlike WhatsApp, Signal gathers no data from its users and all communication is encrypted.

There happens to also be a tiny loss-making company with no revenues called Signal Advance, and when Musk tweeted many people started buying its shares. By market close on 7 January, Signal Advance’s stock price had risen by 527%. If the story ended here it would be a funny speculative anecdote, similar to the confusion that took place in March 2020 about Zoom Video Communications and the unrelated Zoom Technologies.

However, as Matt Levine from Bloomberg recently pointed out (in his newsletter Money Stuff), from a social psychology point of view, things then became interesting. The following day (a Friday) at around 2pm New York time, messaging-app Signal tweeted a picture of Signal Advance’s stock chart with the following text: “It’s understandable that people want to invest in Signal’s record growth, but this isn’t us.” Signal Advance had still gained 91% at the close, which although improbable could be attributed to slow dissemination of information. Later that day, CNBC reported the story, so one could assume that by then the confusion had been cleared up. Yet the following Monday (11 December), after the market had digested this news for an entire two days of lockdown boredom, Signal Advance’s stock went up another 438%.

This cannot be explained by ignorance. The only reasonable explanation is that Signal Advance had become a ‘gambling token’, and had captured the attention of day traders not because of its fundamentals, but because it could now be pumped. One prominent venue on which pumping takes place, by the way, is the well-known sub-Reddit wallstreetbets, whose recent efficacy at sending the shares of small illiquid companies parabolic you can check out for yourself. GameStop, for example, has gone up seven-fold since August, and social media is full of people posting screenshots of how much money they’ve made on the stock (and on Bitcoin, Tesla, Blink Charging, etc…). It’s hard to imagine a more effective FOMO factory. The point is that ignorance does not explain this behaviour. There is a kind of de-centralised collusion to either deliberately squeeze the price of certain stocks, or to participate in the fun for as long as it lasts (remember Citi’s ex-CEO Chuck “As long as the music’s playing” Prince? He knew). There is a fairly convincing argument that this behaviour is rational. Sure, at some point the marginal buyer will buy the top and be left holding the bag. But until then, everyone else will make money. It’s dangerous, but it’s also fun. And some people will get rich.

There is another convincing argument that the people who can’t spot this speculative behaviour are those who are too intelligent to do so. One of the great downsides of being intelligent is that it is easy, even necessary, to find reasons to explain things – reasons that make you look right. And the more intelligent you are, the more those reasons sound convincing. The danger of doing this is that it neglects the fact that the real investor advantage at market extremes is not intellectual but behavioural. This is what Buffett meant when he said that investing is not about intelligence: “If you have a 150 IQ, sell 30 points to someone else. You need to be smart, but not a genius”.

Efficient-market academics are of course extremely intelligent, and so are many others who have opined on the topic of bubbles over the last two decades. The reasons they have constructed for explaining these phenomena are sophisticated, complex, quite convincing and probably wrong. And while it is perhaps easy to be amazed that the run-ups that Signal Advance and GameStop experienced can occur, the question we should really ask is: are we, in the mainstream asset management industry, concocting similar (albeit less extreme) reasons to hold on to our overvalued stocks? Where are we so enamoured with narrative that we have forgotten the gravitational pull of fundamentals? Which part of our portfolios will we look back at in the future and wonder “what were we thinking”? It seems like an appropriate time to ask.

‘Value’ remains an uncorrelated trade

We were going to leave it at that, but our industry isn’t known for being too subtle. So we might as well point out that, in this environment, ‘value’ is pretty much the only uncorrelated trade out there. The first chart below, kindly supplied by Alliance Bernstein, shows the correlation of three factors (growth, return-on-equity as a proxy for quality, and momentum) with value. The negative sign for all factors shows that value is negatively correlated to all three, and that it is particularly negatively correlated with momentum and growth.

Global equities 180 day rolling correlation

Source: Alliance Bernstein, January 2021

The second chart, also courtesy of Alliance Bernstein, will not surprise many readers. It shows that value is the only factor that is positively correlated with rising bond yields. The strength of this positive correlation has hit its extreme at precisely the same time that the negative correlations of the other three factors have hit their own extremes:

US 60 day rolling correlations vs. US treasury yields

Source: Alliance Bernstein, January 2021

The point is that value stocks currently provide significant diversification. They are positively correlated to higher bond yields when everything else isn’t; they are negatively correlated to the most popular equity factors; and it just so happens that they are cheap too.