New research says great managers are really great teachers. What are your thoughts?

Middle managers have a rough time of it in pop culture. They're the petty villains of all our great workplace satires, and from the Peter Principle to the Dilbert Principle, put-upon employees have spent decades concocting theories for why their supervisors all seem so, so inept.   

So maybe it was time for someone to rush to middle management's defense. And that someone, as of this week, turns out to be Stanford economist and former George W. Bush advisor Edward Lazear.

In a new study out on NBER, Lazear, together with Stanford's Kathryn Shaw and the University of Utah's Christopher Stanton, has finally brought some math to bear on the great existential question of whether bosses actually do anything useful. And contra Dilbert, "The Office," and "Office Space," they conclude that talented managers not only exist, but that they do something very specific: they teach. 

Studying the role of mid-level managers poses a challenge to economists for a couple of reasons. Bosses, by definition, manage people. And not all people do jobs that can be easily quantified. By extension, it's hard to boil down their supervisor's performance to an easily analyzed set of numbers. Meanwhile, it's not always clear when a team of workers is successful because of their boss, or in spite of them. 

Here's how Lazear, Shaw, and Stanton get over those hurdles.The study examined data generated between 2006 and 2010 by more than 23,878 workers and 1,940 bosses employed at an anonymous, "large, service-oriented company." The workers were assigned to a single repetitive task that was carefully tracked and timed by computer. They also switched managers roughly once every four months. In short, it was the sort of work environment you might see at an in-house IT department or a customer service call center, and it allowed the researchers to see if there were any supervisors who consistently generated higher productivity among different batches of workers.  

And as it turned out, some bosses achieved much better results than others. Mathematically, taking a boss in the 10th percentile of productivity and replacing them with a supervisor in the 90th percentile was the rough equivalent of adding an extra worker to a nine-person team. 

"If bosses were mere decorations, one would expect no variation" between them, the researchers write. "The fact there is wide variation...implies that there is a substantial productivity effect that bosses confer on their teams."

So what were the good bosses doing right? The researchers considered two possibilities. Either a boss might have been really good at motivating their team (i.e., they were a cheerleader, or maybe a drill sergeant) or they might have taught employees lasting skills (i.e., they were a coach). By looking at how well workers sustained their productivity after switching supervisors, the team concluded that teaching accounted for about two-thirds of a boss's impact on his workers' productivity. 

All of this, of course, comes from a deep dive into the workings of a single corner of a single company engaged in the post-industrial equivalent of basic assembly work. Who knows how a similar study, if it could even be conducted, would hold up with creative professionals, or higher skill workers. 

Still, it's findings seem worth remembering. What's the secret to being a good boss? Teach. 

Any thoughts on what makes a good boss? A terrible, Dilbert-esque one? Leave your thoughts below.

***

The paper contains another interesting but somewhat speculative conclusion that I wanted to at least share. The researchers claim that, in terms of productivity, the average boss is 1.75 times more valuable than the average worker, which just happens to be commensurate with the differences in their pay. But they arrive at that figure through the somewhat circular assumption that if an employee was promoted to management, it must have been because they were more productive than their peers. The paper tries to account for the possibility that some people might be promoted for the wrong reasons, but count me as skeptical. I guess I'm not ready to give up on The Dilbert Principle entirely.