Why investors throw money at eccentric CEOs

“A firm has to answer two questions: What do we do, and how do we do it? The first defines the category you’re in. The second is independent of that and is more performative,” Goldberg says. “In this latter sense, identity is not something you have but something you do.”

The Weird Go Pro

There are many ways to enact atypicality, Goldberg says, from meme-worthy ad campaigns (think Old Spice) to the Hawaiian shirts worn by TJ’s employees. But the researchers focused on one particular signal to test their idea: the language used by executives in earnings calls.

By applying a deep-learning model to the transcripts of over 60,000 quarterly earning calls between 2008 and 2016, they identified firms whose self-presentation was markedly distinct from their competitors. Tesla, unsurprisingly, ranked as much less conventional than Ford. The mobile banking platform Green Dot, likewise, was far removed from JP Morgan Chase. On the flip side, Dell and Bank of America ranked among the most conventional.

When they looked at earnings forecasts, they found that stock analysts who had been on the calls projected higher earnings for atypical firms, likely boosting their stock at the time. The researchers dubbed this the “performative atypicality premium.”

There was just one problem: When they compared the consensus forecasts to actual financial results, it turned out that these companies’ price premiums were unfounded. The atypical firms performed below analysts’ expectations, resulting in negative earnings surprises.

“The analysts seemed to interpret unorthodox behavior as a signal of some intangible ‘it’ factor, a unique quality or vision,” Goldberg says. “But maybe when you get Elon Musk talking about smoking weed and going to the moon, that’s all it is — Elon Musk being weird.”

Conventionally Unconventional

To analyze the transcripts, the researchers used a word-embedding model, which is trained by predicting hidden words based on the context of adjacent words. Goldberg explains: “If you have the sentence ‘I took my ___ to the veterinarian,’ the model learns by reading lots of texts that dog is more likely than father or sofa.”

In this way, words are mapped as coordinates in a 300-dimensional space, with the distance between them corresponding to their semantic proximity. The researchers could then calculate the geometric center of all the words a person has spoken — their center of gravity, if you will — and see how far it is from other speakers

 The word maps also revealed that there were different ways of being different, and some were better than others. In particular, the more a CEO emulated the behavior of celebrated innovators, the greater the performative atypicality premium their company received — sort of the verbal equivalent of wearing a black turtleneck instead of a suit, à la Steve Jobs.

“It’s a bit ironic,” Goldberg says. “But novelty is especially rewarded when it conforms to popular expectations of what constitutes novelty. The key is to be atypical — in a typical way.”

In a sadly rare citation of Monty Python in a scholarly paper, the authors recall a scene from Life of Brian, in which the protagonist, who has been mistaken for Jesus, “tells his thousands of followers that they are all individuals. ‘We are all individuals!’ they respond in unison — with the exception of one screechy voice shouting ‘I’m not!’”

Also read: Confidence: How it helps to keep the marketing budget healthy and long-term

Genius or Goofball?

As a CEO approaches a certain level of eccentricity, you might think analysts would begin to have qualms. However, the researchers found that the farther out an executive’s performance was, the more (unjustified) endorsement they received. Those in the highest quintile of atypicality earned the highest premiums. There was no inflection point.

Goldberg points to the tragicomedy of WeWork: “The product wasn’t new; shared workspaces had been around for 20 years. But they had this messianic CEO, Adam Neumann, who talked about living forever and referred to his company as a capitalist kibbutz. He’d pass out in board meetings and be spotted walking around barefoot in Manhattan.”

Seasoned investors got swept up in the antics and the hype, believing that an office subleasing operation was revolutionizing the workplace. “It was a circus,” Goldberg says, “and they were like, ‘We don’t get it, so it must be genius.’ It was only when WeWork filed for an IPO and people looked at the numbers, they realized, ‘This doesn’t make any sense.’”

To be sure, WeWork is an extreme example. Carving out a unique corporate culture, like Trader Joe’s and Virgin Atlantic have, can be a source of lasting advantage, Goldberg says. “But if we interpret eccentricity or charisma as a signal of other qualities, like strategic vision or leadership — under the pressure to discover the next big thing — we end up making the same mistakes.”