Professor John Antonakis blows apart a multitude of modern management theories. Mike Levy reports from the trenches.

A man who criticises the basis of Myers Briggs, 'emotional intelligence', NLP and 'good to great' had better be sure of his facts. Professor John Antonakis is sure that he is. Catapulted into the media last year by his paper published in one of the most prestigious scientific journals, 'Science' - where he showed that little children could predict election outcomes merely on looking at the faces of politicians, the author of ‘The Nature of Leadership’ is able to defend his corner as a passionate advocate of evidence-based work on leadership.

"Despite the mass of hard evidence showing that sticks and carrots are weak predictors of future performance, leaders are still too quick to resort to these limited methods."

"There are too many snake-oil merchants in the guise of consultants, trainers and management gurus," says the ebullient (not to say pugnacious) Antonakis. "Very little of what they claim is supported by hard evidence; most of which has been totally ignored by those making a lot of money by selling models and techniques that simply don’t work." Antonakis, who is professor of organisational behaviour at the University of Lausanne in Switzerland, is not afraid to take on what he sees as the vested interests in supplying what techniques with dubious provenance. He is not surprised that his scholarly attacks on the great icons of leadership development sometimes do not find favour with the establishment. A recent letter criticising Jim Collins' work on leadership failed to get published in 'The Economist'. In the letter Antonakis says, "Finding the factors that great companies have in common does not mean that those factors predict performance. Collins did not use any advanced statistical procedures to account for survivor bias - there is no control group." Antonakis' argument is that studying similarities or traits in successful leaders is pointless.

The 'two testicle test'

Apart from statistical weakness, Antonakis criticises Collins' methods for making false assumptions. As he says, "Suppose that the names of the CEOs of the five most successful Swiss companies are Hans, Ulrich, Dieter, Jakob, and Johannes. What do they have in common? They are all men, and probably Swiss German. Well, beyond these obvious factors they might have other things in common: that they wear clothes, and have two testicles, among other things. Where do we start to look for commonalities and where to we stop? Now, what if I said that the names of the CEOs heading the five biggest company debacles in Switzerland were Jonas, Karl-Heinz, Rudolf, Jörg, and Jürgen. Managers usually have their 'ah-ha' moment right there: being a man and Swiss German, wearing clothes and having two testicles does not do the trick and more: something else must explain the difference in success and failure. To find predictors of performance we cannot just study success!"

Antonakis uses this 'two testicle test' as a metaphor to criticise much of the work on leadership development. "Evidence-based practice has not yet hit the radar in the business world. This fact is really unfortunate; the more I study leadership practice, the clearer it is that companies are not using the rich vein of scientific work in this area. In my own research on leadership effectiveness I find only one leader in 10 to be very effective (in terms of leader style). It looks like the 'Peter Principle' is alive and well in our corporate boardrooms."

The professor's findings that 90% of leaders are ineffective comes from studying business leaders in many Western European countries and in the USA. "Most leaders under-attain on 'vision and development-based' as well as 'expert-based' leadership." The evidence, says Antonakis, also shows that leaders are still too quick to rely on ‘contingent rewards and punishments' – the bonus culture is a good example of the former. "Despite the mass of hard evidence showing that sticks and carrots are weak predictors of future performance, leaders are still too quick to resort to these limited methods. Carrots certainly work where the task is very simple and the numbers cannot be easily fudged. But in complex organisations with interdependencies, the effectiveness of bonuses is very weak (if at all). Experimental evidence clearly shows that incentives can mess-up performance, creativity and other outcomes, particularly in high stakes situations. Incentives and performance are simply not correlated in many performance settings." The professor's contention is that in the business world weak and pseudoscientific-based leadership development models and practices have meant that talent does not rise to the top.

Absence of evidence

The absence of evidence (or evidence which proves otherwise) also leads Antonakis to cast grave doubts on cherished models and systems including Myers Briggs and NLP. A recent paper by Antonakis states: "There are hundreds of methods or approaches like neuro-linguistic programming (NLP), whose proponents claim to be useful for predicting leadership or for developing leadership skills. Alas, NLP continues to persist in the world of practice even though [research] psychologists stopped taking this construct seriously a while back." He is equally critical of work produced by Goleman and others on 'emotional intelligence' and methods to test it. "The evidence simply isn’t there."

Antonakis' critical gaze also falls on the whole area of leadership traits – what personality types can be predicted to make effective leaders? He gives two examples of popular models that have little or no validity: the Herrmann Brain Dominance Instrument (HBDI), and the DISC personality model. Says Antonakis, "I could not identify any research on this (DISC) model, though plenty of claims about its validity are on the internet. As with the HBDI, this model does not have the requisite research behind it to be used in industrial settings."

He also casts a critical eye on many MBA programmes: “They often rely on practice-based models and avoid too much theory and science-based learning.” Consultants and coaches also get the Antonakis treatment: “Most simply don’t have the tools nor the expertise to sift the chaff from the wheat of leadership development. Too many consultants draw on bogus leadership or selection techniques that do not work. The MBTI is one of the world’s most widely used personality instruments but it is a very lousy predictor of leadership ability.”

If so many of leadership development’s sacred cows are no more than golden calves, are there any models or writers that Antonakis would validate? In a second he replies: "The Big Five Personality Traits test and good old IQ." As for the latter, Antonakis avers that intelligence is a fundamental element in a leader’s likely success – though necessary it is not sufficient. "IQ and the Big Five are rare examples of models that have stood the test of time. Just like aspirin; simply because they have been around doesn’t mean that they are passé."

For a man about to bring out a second edition of his book 'The Nature of Leadership', this time co-edited with a heavy-weight in leadership development Professor David V. Day, Antonakis is very generous in recommending other writers whose ideas stem from hard evidence. "Before considering my book please tell all your TrainingZone.co.uk members to get 'The Halo Effect' by Phil Rosenzweig. Read it before you read any other book on management. It shows that most of what we hold dear in terms of best practice is, to put it lightly, excrement." He also urges us to read 'Predictively Irrational' by Dan Ariely, 'Blink' by Malcolm Gladwell and 'Fooled by Randomness' by Nassim Nicholas Taleb. The Antonakis-Day book will be on the shelves next year. It will, he says, bring readers up to date on the state of leadership research and will look at an area that is really pressing Antonakis' buttons: interdisciplinary work. "There is a lot of fascinating and entirely new work drawing on biology, genetics and evolutionary sciences. We also look ethics, gender, culture and other important topics from different scientific perspectives."

The professor himself is currently working on the role of charisma in leadership and developing predictive tools for voting intentions in political elections. What can we predict about the response to his views on NLP etc? We won’t need a rigorous evidence-based model for this.

Professor Antonakis is happy to take questions from members.