The Soft Bias of High Expectations?

(Please take note that Professor Bainbridge has update his post; see my two updates at the bottom.)

Professor Bainbridge has an interesting post up regarding the effects (or lack thereof) of diversity on workplace productivity. The question is whether diverse teams — mixing people of various ages, races, and demographic backgrounds — are more, less, or equally as productive as homogeneous teams. Bainbridge states:

It turns out that there is considerable evidence that homogeneous work teams tend to be more productive [than heterogeneous work teams].

Now, Bainbridge has done some research in this area; I have not. But if this is his thesis, he doesn’t offer much evidence to support it. To the contrary, the evidence he does offer seems to refute it.

He cites Diversity, Discrimination, and Performance as support. The paper’s abstract states that “[d]iversity of race or gender within the workplace does not predict sales or sales growth, although age diversity predicts low sales.” But this is no support for Bainbridge’s sweeping proposition — it merely states that age diversity may decrease sales, but that other forms of diversity have no apparent effect on productivity.

Bainbridge also cites Corporate Governance, Board Diversity, and Firm Performance, which actually found that diversity on a firm’s board of directors to be beneficial: “After controlling for size, industry, and other corporate governance measures, we find significant positive relationships between the fraction of women or minorities on the board and firm value.” Obviously, this too is no support for Bainbridge’s conclusion.

The best support for Bainbridge’s position is Diversity and Productivity in Production Teams:

Our results indicate that teams with more heterogeneous worker abilities are more productive. Holding the distribution of team ability constant, teams with greater diversity in age are less productive, and those composed only of one ethnicity (Hispanic workers in our case) are more productive[.]

This result, however, is highly suspect. As the paper notes, “the findings for team demographics are not robust to alternative model specifications.” Indeed, the paper studied only “Hispanics [at] . . . a [single] garment plant” over three years. Given that the study is based upon a single manufacturing plant in a low-skill industry and upon workers who, by virtue of their ethnic background and presumed educational level, are likely to be less proficient in English, the paper’s tentative conclusion that diversity (and, possibly, only linguistic diversity) at this plant may harm productivity should come as no surprise. But there’s no basis to extrapolate this extremely limited finding to the general workplace.

This paper doesn’t — and can’t — address whether its finding holds true for other plants, or other industries, or other ethniticies, or other jobs (law, medicine, education, finance). In fact, there’s evidence that it doesn’t hold true for higher-skill jobs, such as company boards (supra).

Indeed, to the extent that we can draw any conclusions from these studies — and I’m not convinced we can yet — they seem to be that (1) diversity in age negatively impacts worker productivity and (2) a linguistically-heterogeneous workplace negatively impacts worker productivity (no surprise). In other circumstances, there is either no good evidence, or the evidence suggests that diversity has no effect upon — and may actually increase — worker productivity. This is a far cry from Bainbridge’s conclusion that “there is considerable evidence that homogeneous work teams tend to be more productive.”

UPDATE: As indicated by my exchange with Dave Schuler in comments, it appears that Professor Bainbridge is updating his post. (The post is currently — at 1:09 p.m. CDT/Indiana — a bit of a scramble.) I would note that Bainbridge’s additional conclusion — “that on balance the economic incentive of employers is to pursue workplace homogeneity” — doesn’t really address whether diverse workplaces are, on balance, more or less productive than non-diverse workplaces. Moreover, when Bainbridge is talking about the “economic incentive of employers,” he is really discussing the “economic incentive of management”; this shouldn’t be confused with (for it is frequently contrary to) the economic interests of the company or its shareholders.

UPDATE 2: Professor Bainbridge has fully updated his post, and it’s much more defensible. Bainbridge’s point that “an employer’s preference for racial sameness won’t always be motivated by racial animus” is perfectly legitimate, I think, on the grounds that he cites — essentially, that it’s easier to manage a homogeneous workforce. But, as a noted in my first update, this doesn’t quite answer the question as to whether it is in the company’s or shareholder’s economic interest to foster diversity, even though it may not be in the empoyer’s (read: management’s) interest. And I still think that the jury is still out on the question of whether diversity might ultimately be in the company’s or its owner’s interests, even though it increases the costs on management.