THE ECONOMIST: Nailing the art of choosing bosses - the skills to look for

The Ig Nobel awards, an annual ceremony for laugh-out-loud scientific papers, celebrate the joyfully improbable nature of much academic research.

One of this year’s Ig Nobel winners, “Factors involved in the ejection of milk”, was published in 1941 and tests whether fear causes cows to involuntarily drain their udders.

Its authors drew their conclusions by placing a cat on a cow’s back and repeatedly exploding paper bags beside it.

“Genetic determinism and hemispheric influence in hair whorl formation”, another winner, asks whether hair tends to swirl in the same direction depending on which hemisphere you live in.

Sometimes you come across an academic paper that asks a deeply practical question in a refreshingly plain way.

“How do you find a good manager?”, a new study by Ben Weidmann of the Harvard Kennedy School and his co-authors, sits in this category. Answering that question well is important.

Other research, to say nothing of the experience of everybody everywhere, shows that variations in the quality of management help explain differences in performance between companies and even between countries.

Yet a survey conducted last year by the Chartered Management Institute in Britain found that four in every five people entering management had received no formal training.

And loads of bosses accrue managerial responsibilities for reasons unrelated to their ability to discharge them.

Another paper, by Alan Benson of the Carlson School of Management and his co-authors, looked at the career paths of thousands of sales workers in over 200 American firms.

They found that better sales performance increased the likelihood of people being promoted but was also associated with worse performance among their new subordinates.

The “Peter Principle”, the idea that people rise up the ladder if they do their current job well until they reach a job at which they are incompetent, appears to be alive and well.

How then should managers be selected?

The study by Mr Weidmann et al sought to answer that question by running a series of repeated experiments in which participants were randomly assigned to three-person teams of one manager and two subordinates.

Each member of the team, including the manager, had to complete a number of problem-solving tasks.

The manager’s job was to assign people to the task they were most suited to; monitor their performance and reassign them as needed; and keep them motivated. In the real world bosses do more things, but this captures a big part of their role.

The researchers found that a competent manager had about twice as much impact on the team’s performance as a competent worker.

More usefully, they also found out which traits were associated with good and bad managerial performance.

Teams run by people who said they really, really wanted to be managers performed worse than those who were assigned to lead them by chance.

Self-promoting types tended to be overconfident about their own abilities; in a huge shock, they also tended to be men.

If appointing a manager just because he sticks his hand up and says he can read people is not a great selection strategy, what would be better?

The researchers found that good managerial outcomes were associated with certain skills.

One in particular stood out: people who did well on a test of economic IQ developed by researchers at Harvard called the “assignment game”, in which you have to quickly spot patterns in the performance data of fictional workers and match them to the tasks they are best at. (Anyone can play the game online: you end up with a percentile score and a mild headache.)

Since the assignment game is similar to the experiment in the study, you would expect people who were good at one to shine in the other.

But for David Deming, also of the Harvard Kennedy School and another of the paper’s authors, that is precisely the point.

Management tasks can be identified, codified and incorporated into selection processes: that is a better way of choosing bosses than drawing only on those who thrust themselves forward or looking at how people perform in other jobs.

There are echoes here of a paper by Alessandro Pluchino of the University of Catania and his co-authors, who found that it was better to promote people at random than based on how well they did their current role.

That won an Ig Nobel in 2010.

Just because something is funny doesn’t mean it should be dismissed.