Let's put aside that executive compensation debate for a moment and consider instead what Sreedhari Desai is studying. Does sky-high pay lead to worse treatment of workers? Does it make leaders, literally, meaner?
Desai says the answer is yes. "Increasing executive compensation results in executives behaving meanly toward those lower down the hierarchy" she writes (with co-authors Arthur Brief from the University of Utah and Jennifer George of Rice University) in the paper "When Executives Rake in Millions: Meanness in Organizations," which was presented at last week's International Association of Conflict Management meeting in Boston. (It's under review with Administrative Science Quarterly.)
Desai's a bit of a polymath. She holds degrees in metallurgical engineering and finance, is working on her doctorate in management and is a fellow at Harvard Law School's Program on Negotiation. "When I delved into organizational behavior, I started looking at Adam Galinsky's work," says Desai. "It made me want to look at income disparity, power and moral disengagement. Does this income gap help the leaders to feel comfortable setting up policies that hurt the people at the bottom?"
To test this, Desai got her hands on employee relations research from Kinder, Lydenberg, Domini & Co. (KLD.). This data set is part of a bigger project on corporate social responsibility. Desai scored firms using the KLD data. For example, companies that had recently paid fines for employee mistreatment were docked points while firms that offered profit sharing gained points. After an exhaustive tallying, she subtracted "strength" points from "weakness" points, and gave each company a "meanness score." Then, while controlling for several variables, Desai cross-checked that score with Compustat data on general firm information including executive compensation.
The results showed a clear pattern. The higher the executive compensation, the higher the meanness score. In an ideal world, Desai would have been able to measure the gap between executive and lower-level employee's average compensation, but that data was impossible to obtain. However, consulting National Bureau of Economic Research economists, Desai was assured just measuring based on the executive compensation was a reasonable proxy.
Desai saw a correlation in the existing data. Her next step was to test this correlation in the lab. Her experiment reminded us of the ones that Dana Carney at Columbia did to see how power affected people's ability to lie. In Desai's experiment, she had subjects solve anagrams and told them that they'd be compared to other people taking the same test. In reality, there was no other person.
Subjects were always told that they got one more anagram right than the other person (signaling to them that while they outperformed the other person, they weren't noticeably superior performers). For their performance, they were awarded 65 points. Then half the subjects were told that the other person was awarded 60 points (almost the same amount, or low income inequality) and half were told the other person was awarded 15 points (not nearly as much; high income inequality).
In round two the subject became the manager of the other, fake person, who ostensibly would be solving some puzzles. This time, the newly minted managers were told the puzzles were mazes, so that subjects wouldn't have done the puzzles and wouldn't know, really, how difficult they were. The subjects were told that they and their "employee" would earn profits relative to the imaginary employee's performance.
In every case, the "managers" were told that their employees performed "average" after this round. Then, prior to round three, Desai gave the managers the opportunity to retain or fire the employee. Desai also had the subjects answer questions about their perceived power. The results again were clear. "The higher the income inequality the more they perceived power and the more likely they would fire average-performing employees."
As "disheartening" (Desai used this word several times during our chat) as the results were, Desai speculates that they're not terribly surprising results. She thinks they simply conform to a long line of well-known research on power, starting with work by the late David Kipnis from back in 1972, through Albert Bandura's work from 1999, and more. Power holding theory suggests that power is control, and, basically, humans are world class rationalizers who find ways to insulate themselves form the mean, sometimes unethical or inhumane things they do. Money may be one of those insulators.
Desai's not sure where else she'll take this research, but she's also working on a paper about how volatility "thickens the glass ceiling." That is, in uncertain times, management "cocoons." It tries to avoid conflict and difference and therefore promotion and hiring decisions at these times become more homogenous. Thus, women are promoted and hired less in those uncertain times, Desai theorizes.
Sounds worth watching.
This blog first appeared on Harvard Business Review on 6/29/2010.