New research shows the large economic value of effective supervisors and the surprising ways they make workers more effective.
What’s the value of a good boss? When the boss in question is a CEO, the question’s been the subject of a lot of scrutiny. But the role of less exalted front-line supervisors—the folks who directly oversee teams of workers—has been much less scrutinized. Do supervisors vary in quality? How valuable is a good one? And what makes a good supervisor good? These are questions that preoccupy people in their daily work lives but haven’t been the subject of much formal economics research. But Kathryn Shaw and Edward Lazear of Stanford’s Graduate School of Business recently teamed up with the University of Utah’s Christopher Stanton to explore the subject in a working paper called “The Value of Bosses.”
Their research comes from a study of a large company (it’s kept anonymous as part of the terms under which it was conducted) since that’s the best way to get the sample you need to analyze the data properly, but the results should be of interest to top-level managers at companies of all sizes. And, indeed, it’s smaller firms that probably have the most to learn from this sort of study, since they lack the scale to conduct an analysis in-house.
And the answer is that, yes, good bosses are a good deal better than bad ones. Replacing a supervisor from the bottom 10 percent of the pool with one from the top 10 percent increases output about as much as adding a tenth worker to a nine-worker team.
That it’s better to have a good manager than a poor one isn’t very surprising. But the authors’ two other findings are a bit more striking. The main impact effective supervisors have on their employees isn’t from motivating or supervising them per se, it’s from teaching them better work methods. But despite that teaching/learning dynamic, the most efficient matching structure is to assign the best workers to the best bosses rather than try to have the best bosses bring the weakest workers up to speed.
The research is drawn from a company that employs a large set of workers to perform what they call a “technology-based service job,” offering as examples of companies with TBS jobs “those with retail sales clerks, movie theater concession-stand employees, in-house IT specialists, airline gate agents, call-center workers, technical repair workers, and a host of other jobs in which an employee is logged into a computer while working.” These aren’t factory workers, in other words, but like the mass production workers of yore, their jobs involve production of discrete objectively measurable outputs. That’s different from personal service workers like hair stylists (or Slate) writers, whose outputs are heterogeneous and where productivity is driven by subjective quality as well as hourly output. This kind of large-scale routinized work is good to study since it generates lots of data.
Even better, the company works with relatively small teams (average size: 9.04 workers), and workers switch bosses quarterly. It’s that regular switching “that permits estimation of the effect of bosses on workers’ productivity” and makes the study so valuable even though it’s only of one company.
They find that the vast majority of the variation in worker output is driven by how efficiently they work when they’re working. Differences in “uptime” due to breaks and other disruptions are minor.
Running the numbers, it looks like workers produce 5.9 more units per hour when working for the top 10 percent of bosses than for the average supervisor. The gap between the 10th percentile and the 90th percentile is similar to going from a nine-worker team to a 10-worker team. And perhaps most interesting, about two-thirds of the productivity boost from working under a good supervisor persists even after the worker switches bosses. That leads the authors to conclude that “teaching”—imparting better methods or skills—is the biggest part of effective supervision. The good-boss effect does decay over time, however, so that after six months only about 18 percent of the boost remains. This raises the question of whether it makes sense to deliberately try to assign the strongest supervisors to the weakest workers in order to maximize the impact of teaching. In the case of this firm, at least, the reverse is true.
“Maximizing the value of bosses requires that the better bosses be assigned to the better workers,” they conclude, because workers increase their output so much when working with star supervisors.
Of course any boss out there should use caution before applying lessons gleaned from one firm to their own business. One moral of the story is that, where feasible, it’d be worth trying to repeat this basic study at your own firm. But smaller operations that simply don’t have the scale to repeat the data collection necessary might want to take a leap of faith about the generalizability of these findings. Pay attention to your supervisors—and your own work in a supervisory role—as more than just a routine element of efficiency. Make sure your most promising workers are paired up with the best supervisors. And make it clear that teaching and learning is the essence of an effective boss/employee relationship. Good bosses don’t just crack the whip or deliver a pep talk, they deliver concrete improvements in workers’ ability to get the job done.