The Academic Who Believes He Can Predict CEO Departures
Jonathan Margolis
There is some heartening news for British chief executives toiling over their annual letter to shareholders. A senior academic may be analyzing your words for telltale signs of your psychological state – and to predict when you might have to quit.
By burrowing into your emotions, he will try to assess whether you are about to be fired, or are feeling under pressure to leave voluntarily. He claims a 73 percent success rate in correctly predicting CEO exits.
Your impending sacking – or your permanent move to golf course or garden – will supposedly be given away by two key indicators: an ominous lack of future-focus in your writing; a surfeit of negative emotion; or both.
The academic behind the research is Dr. Qingan Huang, senior lecturer in strategic management at theUniversityofEast London’sSchoolofBusinessand Law.
Dr. Huang is a native ofGuangzhou,China, who adopted the name Angus inGlasgowwhile doing one of five university degrees. He gained his initial psychology qualification in China and the rest from UK universities.
His work on CEO sackings won him a PhD last year fromLondon’sCassBusinessSchool, and is under review for publication in the Chicago-based Strategic Management Journal.
He examined shareholder letters from most of the 600-plus companies listed on the FTSE All-Share Index, over six years from 2002. The decoding and analyzing work was aided by software called Linguistic Inquiry Word Count (LIWC), developed at theUniversityofTexas.
The psycholinguistic variables thrown up by LIWC back in London were then analyzed using the popular Stata statistical software package to match CEO turnover type (dismissal or voluntary step down) with future focus and negative phrasing.
It was after this number crunching that the 73 percent-accurate link emerged between 268 CEO exits examined and the language the executive used. Typically, according to his research, departures followed within a year of the shareholder letter in question.
“Emotion takes an important role in CEOs’ decision-making. Even taking into account irony, sarcasm and unusual sentence structure in texts, LIWC is a reliable method of gauging the emotional expression and cognition of the writers,” Dr. Huang said when I spoke to him recently at his office, close to London’s Olympic Park.
The software’s negative emotion dictionary comprises 345 words, such as “hate”, “worthless”, “enemy”, “fear”, “unfair”, “anger”, and “sadness” that reflect the intensity of CEOs’ negative feelings and emotion. And it hunts for the good stuff – future-focused emotion – by seeking 48 words including “will”, “might” and “shall”.
For evidence of CEOs getting things emotionally right and wrong in their shareholder letters, Dr. Huang cites two examples.
The first is the final shareholder communication from Ray Webster, CEO of Easy Jet from 1996 to 2005, who Dr. Huang says demonstrated strong future focus and subsequently stepped down at a time of his choosing.
He had grown Easy Jet’s revenues by showing shareholders that he was thinking about the next step, according to Dr. Huang, with a statement that exemplified his future focus: “Despite the tangible progress, there is still more to do to provide an improvement to the underlying performance,” wrote Mr. Webster.
By contrast, Brendan O’Neill, CEO of ICI from 1999 until a profit warning in 2003 prompted his departure, ended his last letter to shareholders with the statement: “When the tide turns, I know we will be ready” – a case of what Dr. Huang calls “optimism without strategy”.
Dr. Huang’s study is the only one I have found into the emotions of CEOs, and is relatively small-scale as data-mining exercises go.
But it is part of an increasing vogue for using scientific method to assess, exploit and perhaps ultimately engineer emotion for commercial benefit.
In December, I wrote about Crowd Emotion, a London start-up that supplies facial expression-reading technology to clients such as Honda, the BBC and Harvard Medical School. Among the emotion-tracking technologies on offer from others are voice analysis, and bio-sensors.
Company HR departments have been known to sense from employees’ sick day patterns when they are close to quitting their jobs. Retailers, meanwhile, use big data analytics to work out what is happening in customers’ lives based on their purchases.
All this is fascinating. Thinkers from Aristotle have tried to rationalize emotion. But the result in the big data age is either too creepy and intrusive – or an over-complex attempt to systematize what Basil Fawlty would call “the bleeding obvious”.
Now the cat is out of Dr. Huang’s bag, would sensible CEOs not simply dodge bullets by checking their writing for the wrong kinds of words? He says they could avoid the sack by improving their language skills – the corporate equivalent, perhaps, of faking an orgasm.
But then my reaction was probably a prime case of mistaking emotion for expertise. For while current thinking by Dr. Huang and other researchers holds that raw emotions are a superb data point, the evidence suggests that they are not always the best basis for action.