21 Jun 2004  Research & Ideas

Racial Diversity Pays Off

Diversity has been a buzzword in organizations for at least fifteen years. How much is really known about its effects on performance? HBS professors Robin Ely and David Thomas investigate.

 

Increase the diversity of employees in your organization—for example, on the basis of race, gender, religion, nationality, and sexual orientation—and you'll automatically have a better company. Right? That's the prevailing belief anyway.

But according to HBS professors Robin Ely and David Thomas, the benefit of racial diversity alone, done right, pays off not just in a better company, but a more productive one.

In new research that focuses specifically on racial diversity, Ely said they found measurable performance benefits when work groups chose to learn from members' different experiences rather than ignore or suppress them.

Comparing 450 bank branches of a single organization, a large commercial bank in the northeast U.S., Ely and Thomas discovered that when work groups actively acknowledge and engage with their different members in a way that fostered learning, they performed better as well. In addition, interviews at five of the branches supported the findings.

Speaking to an audience of other Harvard faculty at the HBS Faculty Research Symposium on May 25, Ely discussed the bank study as well as the complexity of diversity research in general. She and Thomas have worked together on diversity research for the past ten years.

In an introduction to the session, Professor John J. Gabarro said gender, racial, and ethnic diversity have been the subject of much attention in recent years, including academic attention. Progress has come "slowly, and in some cases painfully and in other instances not at all," Gabarro said, "despite the fact that in the last thirty years women have joined both managerial and professional ranks in unprecedented numbers, as have American minorities. Yet neither group is even close to being on a par with either their male or white counterparts."

Ely's contribution, he went on to explain, is in developing theory about when and under what circumstances diversity leads to process gains or losses. The questions she asks, he said, include: How do organizations' structures and work practices moderate the effects of identity and racial diversity? What makes a difference in terms of high performance or low performance?

Rhetoric of diversity

According to Ely, the "diversity rhetoric"—that diversity is "good" for a company—gained force in the late '80s and early '90s, and is still rather popular today. Two books helped to spark the trend. The first, Workforce 2000: Work and Workers for the Twenty-First Century, by William B. Johnstone and Arnold E. Packer, boldly predicted that by the year 2000 only 15 percent of the workforce would be comprised of white males, while the remaining 85 percent would be women and people of color. The second book, The New Leaders: Leadership Diversity in America, by Ann Morrison, made the following optimistic statement, a clarion call to advocates of diversity: "Diversity is a worthy goal undoubtedly linked to the overall performance of any organization."

The authors found a large following who believed diversity was a reality and furthermore was something to be celebrated, said Ely. These people believed "that [diversity] was really going to advance the interests of corporate America." The blanket statement that diversity was a worthy goal undoubtedly linked to performance struck her as false, however.

"In fact, there is nothing undoubted about this link to performance benefits. In the empirical literature, there is very little to support the idea that if you bring a diverse group together you will necessarily get performance benefits from that group," said Ely. "I was interested in this 'certainty' and I wanted to put it to an empirical test."

She defined cultural diversity as group differences in identity, particularly concerning groups that are socioculturally distinct and have different power positions in society. Sociocultural distinctions may be race, ethnicity, gender, religion, nationality, and sexual orientation.

Theory and practice

According to Ely, there are two competing theories about the impact of cultural diversity on work group performance. The first, which she called Position A, is an optimistic point of view. It is the idea that cultural diversity in work groups increases the pool of available resources—such as social networks, skills, and insights—to enhance the group's ability to be creative and solve problems.

I was interested in this "certainty" and I wanted to put it to an empirical test.

The second, Position B, is more pessimistic, she said. "According to this theory, what happens when you bring groups together is you get social comparisons, in-groups and out-groups, where people have a preference for their in-group members." Such preferences, the theory goes, can lead to miscommunication, stereotyping, polarization—and performance losses.

Most empirical studies probably are more supportive of Position B, said Ely. A few support Position A, but they have largely been conducted in labs, not in real-life companies.

Ely and Thomas decided to look at the contingencies: Under what conditions should they see A or B? Most academics who had studied diversity identified conditions that could mitigate against Position B, but these conditions did not necessarily translate into performance benefits.

To explore the possibility of performance benefits emanating from diversity, the two professors wanted to find a company that already had a lot of diversity in its ranks. This led them to the commercial bank. They then collected human resources data on branch demographics, the results of annual employee attitude and satisfaction surveys, and objective, branch-level productivity measures, Ely said.

In general, Ely added, companies such as the one they studied tend to take on one of three perspectives about diversity:

1. Discrimination and fairness perspective. The work groups aspire to being color blind, so conversations around race are limited, she said. They believe there is no connection between race and the work, but racial bias can end up being destructive in the work group, said Ely.

2. Access and legitimacy perspective. There is diversity only in certain parts of the organization. People are effectively shunted onto segregated career tracks and told, "This is what you're good at."

3. Integration and learning perspective. Group members are encouraged to bring all relevant insights and perspectives to bear on their work.

"Every work group we saw [in the study organization] used one of the three," said Ely. "Some groups developed through these. We didn't see competing perspectives." In addition, she said, a strong individual could institute one or the other perspective. Only the third, however, was linked to sustained performance gains, she and Thomas found.

Work groups that adopted the integration and learning perspective revealed an awareness that racism in personal life and work can and do interact, she said. Employees whose work groups adopted perspectives 1 or 2—neither of which encouraged learning from others—felt like cogs in a wheel.

The third perspective, on the other hand, meant that employees bring "the whole person to work," said Ely. Openly discussing and learning from differences made it possible for the groups to create psychological safety, she suggested, adding, "The integration and learning perspective is a way of managing the fact of racial discrimination in the larger culture."