Leonard and Levine (2006)

Leonard, Jonathan, and David Levine. "Diversity, Discrimination, and Performance." Working Paper (UC Berkeley), November 2006.

From the authors’ abstract: “Employee diversity may affect business performance both as a result of customer discrimination and as a result of how members of a group work with each other in teams. We test for both channels with data from more than 800 retail stores employing over 70,000 individuals matched to Census data on the demographics of the community. We find little payoff to matching employee demographics to those of potential customers except when the customers do not speak English.”

LK notes from 2004 draft:

This study examines the business impact of workforce diversity through an analysis of a single company's 800 retail stores, which employ 70,000 people. The author notes that some proponents of the business case for diversity (the employees base should look like the customer base) make an argument that has also been used to justify discrimination. In 1993 the CEO of Shoney's, for example, sought to increase the number of whites in customer-facing positions because the restaurant chain's customer base was predominantly white. "Proponents of workplace diversity...often share the view of Shoney's CEO that customers prefer to deal with employees of the same race or sex." The author offers another possibility: "in our society, all demographic groups may prefer to be served by certain groups; either high status groups or (if people prefer to have service people fit stereotypes) by low-status groups."

The study examines these questions through a complex regression analysis. The key independent variable is the degree to which the ethnic makeup of a store's employees matches that of its neighborhood (based on census age and race data by zip code). The key dependent variable is sales growth, which is studied using a variety of transformations.

Finds that "sales are higher if employees speak the language of customers who do not speak English," but otherwise "our results do not support theories that employee-customer match increases sales." Also finds that "race and gender are not generally correlated with sales," but that age diversity predicted lower sales. In connection with this last point the author notes that the firm studied had an unusually young workforce overall ("a 28 year old is an unusually old employee in this firm").

Link: https://escholarship.org/content/qt2p3880ms/qt2p3880ms.pdf