“For almost a decade through our Diversity Matters series of reports, McKinsey has delivered a comprehensive global perspective on the relationship between leadership diversity and company performance. This year, the business case is the strongest it has been since we’ve been tracking ..”
(McKinsey, 2023, p. 3)

The above quotation is a typically enthusiastic statement of the positive effects of diversity on company performance: the “business case for diversity”.

So, what is “diversity”?

Diversity

Diversity, in the context of this topic, means the participation of different kinds of people in a team, an organization, or a company.

The idea that diversity is a good thing that one should strive for has its roots in America, a society beset with racial issues and tensions between whites and blacks since the founding of the country in the 18th century. In the mid-1960’s there was an attempt by the US Government and Congress to solve the country’s racial issues by putting an end to discrimination against blacks. The Civil Rights Act was passed by the US Congress in 1964. It forbid discrimination based on race, gender, religion, etc. in hiring for jobs and other areas of life (Hanania, 2023, p. 39). Diversity was a welcome byproduct of this legislation: increasing the representation of blacks and women in organizations necessarily increased diversity. Thus, by forbidding discrimination, the Civil Rights Act enforced diversity.

Soon, however, courts started to interpret the law in a way that any practice by a company which resulted in less hiring of or some other disadvantages for blacks – i.e. measures which had a “disparate impact” on blacks – was seen as against the law. In the famous case Griggs v. Duke Power Co. in 1971 the Supreme Court ruled that the Duke Power company was not allowed to use intelligence tests to find employees for work in higher paying jobs because blacks scored lower than whites on this test. That is, the test resulted in “disparate impact” on them.

Though the ruling allowed companies to cite “business necessity” to defend such tests, they would have to prove that such a necessity existed. The result was that, rather than risking going through this – and then potentially losing court cases – companies simply set up quotas for blacks, increasing their numbers and thus increasing racial diversity (Hanania, 2023, p. 41).

Universities have been left leaning at least since the 1960’s, and they didn’t need much convincing by legislation to try to increase their diversity. For example, the University of California at Davis introduced so called “affirmative action” in the 1970’s – racial quotas, setting 16 % of the student places aside for minorities to be admitted to the Medical Faculty. This led to a court case, Regents of the University of California v. Bakke, when Alan Bakke, a white man, sued the University because he was denied admission into medical studies. The University defended racial quotas partly by emphasizing the value of diversity in its student body. While the Supreme Court ruled that racial quotas were illegal, considering race as part of the entry requirements was allowed (Hanania, 2023, p. 13).

The debate about how to increase diversity is still ongoing and it became more complicated over time. For example, universities, in their eagerness to increase the numbers of their black students, have discriminated against other minority students – in particular students of Asian origin. A number of Asian students sued Harvard University to stop this practice. In 2023 the US Supreme Court ruled in favor of these students and forbade this “affirmative action” in university admissions as unconstitutional.

Parallel to the efforts by the US government to decrease discrimination – and thus increase diversity – there were radical movements which blamed racism and sexism for existing racial and sexual disparities in the USA. One of these, Critical Race Theory (CRT) has played an important role since the 1990’s in the creation of “Diversity, Equity and Inclusion” (DEI) which can be seen as a practice and implementation of CRT in institutions and organizations.. CRT is inspired by the earlier Critical Theory of Herbert Marcuse, a Marxist ideologue, who saw US society as a struggle between oppressors and oppressed. CRT retained this view, focusing primarily on “institutional racism”. Later, CRT proponents also introduced sexual and other “oppressions”, in a so called “intersectional” framework.

The ultimate oppressors in the view of CRT are white (and male) and the oppressed are blacks, other minorities (and women). CRT tends to see every disparity between races – e.g. in income or participation in prestigious jobs – as a result of institutional racism. It demands “equity” – the equality of outcome, independent of “input”. This makes equity very different from the equality of opportunity.

Equity and diversity are intimately interconnected. One example of equity has been the student admission system in elite US universities like Harvard: study admission requirements were set lower for blacks than for others. As a result, the number of black students – and thus racial diversity – became a lot higher than it would have been under an equality of opportunity system.

There are many other recent examples of equity.

US universities are getting rid of grading students and are lowering standards – because the grades of black students were far worse than grades of other groups.

Schools in British Columbia, Canada introduced a new “proficiency” grading system which describes all student performance levels as good to different degrees. According to advocates of this system, this avoids “labeling” students.

In King County, Seattle, USA, fines for not wearing bicycle helmets were abolished because blacks were not wearing helmets more often and were thus fined more often.

In New York City, fines for jaywalking were also abolished, for the same reason: people of color, in particular blacks, apparently jaywalked more and were thus fined more.

In each of these cases the “outcome” was equalized, while the “input” – the study performance of black and non-black students, and the helmet-wearing and jaywalking by blacks and non-blacks – was different.

These examples also illustrate some of the drawbacks of equity. For example, in the last 5 years, 200 people died in New York while jaywalking. This number will presumably increase, now that jaywalking is not unlawful any more. Head injuries will probably increase in King County, after the helmet-wearing requirement has been abolished. And, as a cautionary tale illustrating the effects of “affirmative action”: Patrick Chavis, one of the black students admitted to the Medical Faculty of the University of California based on its racial quota, went on to become a doctor infamous for his incompetence and negligence and for the large number of lawsuits against him because of malpractice.

Looking for a Business Case for Diversity

Diversity: political vs. practical aspects

In the Regents of the University of California v. Bakke case mentioned above, the University of California argued that its quota system, in addition to helping a historically disadvantaged minority, had additional benefits as well. In particular, the University argued that increased diversity in the classes benefitted teaching. It is likely that the original motivation of the University in introducing racial quotas had nothing to do with this diversity and its educational benefits. It was an “affirmative action” measure, a political program. The argument about the practical – educational – benefits were probably brought up post hoc, to increase their chances of winning the case and keep the racial quotas.

A similar thing seems to be happening in the business world. The original intent to increase racial diversity in hiring was rooted in the Civil Rights Act, as a political program. Initially, companies were more or less forced by this Act to increase their hiring of racial minorities by court decisions. The “business case” for diversity – a practical benefit of diversity – was emphasized later: maybe in order to sell the diversity idea better, its advocates have tried for years to show that diversity is good not only for “disadvantaged” minorities, but also for the organizations which hire them. In order to prove this, there has been an intense search for business advantages of diversity for companies – with mixed results, as will be shown below.

Researching diversity

The “diversity” research literature has grown immensely over the last few years. The following is a small selection from this literature.

Consulting companies, like McKinsey and the Boston Consulting Group, have been at the forefront of this search for a number of years. In several reports, McKinsey has described positive effects of ethnic diversity in executive teams on company financial performance (e.g. McKinsey, 2020, McKinsey, 2023).

Based on an analysis of 1265 companies, McKinsey reported in 2023 that they found large positive effects both of gender diversity and ethnic diversity of executive boards:

“Across our widened dataset, companies in the top quartile for gender diversity are 39 percent more likely to outperform peers, continuing an upwards trend from 15 percent in 2015 ..
When it comes to ethnic diversity, too, a consistently strong business case is apparent over time. Our analysis indicates a 39 percent increased likelihood of outperformance for the top quartile in ethnic representation versus the bottom quartile.”
(McKinsey, 2023, p. 11)

The term “outperform peers” above is defined like this:

“Outperformance is calculated as the likelihood to place above the median profitability of other companies in the same industry and region.”
(McKinsey, 2023, p. 11)

Some academic research found similar trends. For example, sociologist Cedric Herring found positive effects of racial and gender diversity on the business performance of several hundred US companies (Herring, 2009). Business performance was measured by four variables: sales revenue, number of customers, market share and profits. All of these increased when racial diversity increased. Gender diversity also increased all but one of the performance variables (market share). Herring concluded:

“The findings presented here are consistent with arguments that diversity is related to business success because it allows companies to “think outside the box” by bringing previously excluded groups inside the box.”
(Herring, 2009, p. 220)

Herring’s research became an often cited evidence for the positive effects of diversity on business performance of companies.

Other research, however, was unable to find positive effects of diversity. University researchers Jeremiah Green and John Hand recently tried to replicate the McKinsey results. As McKinsey refused to provide their own dataset for them, citing confidentiality issues, they had to use a different dataset: ethnic diversity of executive teams and financial performance data from 497 of the S&P 500 companies. They justified this decision by writing that, since “the firms in the S&P 500® are large US public companies, they likely match well with McKinsey’s set of US firms”. They did not find a positive effect of diversity:

“.. in contrast to McKinsey’s results, we do not find a statistically significant positive correlation between McKinsey’s measures of the racial/ethnic diversity of the executive teams of firms in the S&P 500® Index at 12/31/19 and either the likelihood of financial outperformance over 2015–2019 or financial outperformance per se over 2015–2019.”
(Green & Hand, 2024, p. 28).

And, when a group of university researchers from the Netherlands tried to replicate Herring’s results in 2017, they found that Herring made substantial errors in the statistical analysis of his data. After correcting those errors, these researchers found no statistically significant effect of racial diversity at all, and only one statistically significant positive effect of gender diversity – on the number of customers. The researchers concluded:

“In the years following its publication, Herring’s article has become an influential source of empirical support for the value-in-diversity perspective. Our replication shows that the data Herring analyzed are not consistent with this perspective. The overall pattern of findings suggests that diversity is nonconsequential, rather than beneficial, to business success.”
(Stojmenovska, Bol & Leopold, 2017, p. 6)

A number of recent meta-analyses summarized the results of studies on gender diversity on company boards and financial performance of the company.

Corinne Post and Kris Byron, from two US universities, meta-analyzed 140 studies from a number of countries. These 140 studies investigated 90,070 companies for the effect of gender diversity on two measures of company performance: accounting returns and market performance. They found that

” .. it appears that firms with more women on the board have higher accounting returns ..
it appears that, in general, female board representation is not significantly related to market performance.”
(Post & Byron, 2015, p. 55)

Thus, they found a positive correlation between gender diversity and accounting returns. However, this correlation (r = 0.047) is commonly called very weak or negligible. Between gender diversity and market performance no significant correlation was found.

In another meta-analysis, Jan Luca Pletzer and his colleagues from German, Dutch and Belgian universities attempted to replicate the results by Post & Byron (2015), but using more stringent criteria in selecting the studies for meta-analysis. For example, while Post & Byron (2015) used non-published working papers in addition to peer-reviewed articles, these researchers relied only on studies published in peer-reviewed scientific journals. This limited their dataset to only 20 previous studies. Their results showed no statistically significant relationship between gender diversity on company boards and financial performance:

“.. the overall mean weighted correlation between percentage of females on corporate boards and firm performance was small and non-significant .. These results indicate that the mere representation of females on corporate boards is not related to firm financial performance if other factors are not considered.”
(Pletzer et al. 2015, p. 1)

Companies are composed of groups or teams. Thus, research about the diversity and performance of teams – e.g. in decision making or problem solving – even if not directly investigating the financial performance of companies, is relevant for a business case for diversity.

One often cited meta-analysis evaluating studies about team diversity and performance was done by university researchers Katherine Williams and Charles O’Reilly. After evaluating 80 scientific articles about the effect of different types of diversity – race/ethnicity, sex, age, educational background – on group performance, these researchers came out with a negative conclusion:

“the preponderance of evidence suggests that diversity is most likely to impede group functioning .. by itself, diversity is more likely to have negative than positive effects on group performance”
(Williams & O’Reilly III, 1998, p. 120).

A more recent meta-analysis (Wallrich et al. 2024) of the effects of different types of diversity on team performance is a comprehensive evaluation of the team diversity – team performance research. The authors found a bewildering variety of relationships between diversity and team performance.

The authors broke down “diversity” of the employees into three dimensions: demographic (race, ethnicity, gender, age, etc.), cognitive (personality, education), and job-related (function, length of time being employed).

In addition, they identified the following task dimensions: complexity, creativity required, and cooperation required.

They looked at 11 previous meta-analyses of diversity’s effects on team performance in organizations, written between 2000 and 2021, doing a kind of meta-meta-analysis. This was their conclusion:

“Three key results emerge from that work: (1) any overall relationships between diversity and performance appear very small (i.e. |r| < .1), (2) while job-related diversity tends to have positive associations with performance, demographic and cognitive diversity tend to have negative associations, and (3) effect sizes are highly heterogeneous.”
(Wallrich et al. 2024, p. 10)

Related to Point 1 of their conclusion about the meta analyzes, “r” is the correlation – either positive of negative – between diversity and performance. As mentioned above, a correlation of this magnitude is often called “very small” or “negligible”. The authors describe it as “unsubstantial”.

Point 2 of their conclusion about the meta analyzes is that the types of diversity usually emphasized – demographic, i.e. racial / ethnic diversity, gender diversity, or diversity of personality characteristics – have a negative effect on team performance. Job related diversity – like diversity of functions that team members have and diversity of the time of their employment in the organization – had a positive effect.

The authors did their own meta-analysis of 615 research articles about the effect of diversity on team performance, written in English and in a number of other languages, too. They focused at team-level performance – except when the team was at the executive level, where they also considered performance of the whole organization.

They found positive and statistically significant, but very small (“insubstantial”) effects of diversity on team performance. This was their conclusion:

“Overall, our results show that diversity (across demographic, job-related and cognitive dimensions) is positively correlated with team performance, but with an insubstantial effect size: on average, diversity (on one trait) explains far less than 1% of the variance in team performance.”
(Wallrich et al. 2024, p. 60)

They write:

“The “business case for diversity” is widely articulated, and many efforts towards greater diversity are justified based on its claimed potential to increase organizational performance. The results here show that this may be too simplistic – diversity does not substantially improve (or hamper) team performance across the board. While it may be worth noting that the evidence suggests that diversity may be more likely to provide (minimal) benefits rather than harms on average, the picture is more complex.”
(Wallrich et al. 2024, p. 63)

And, regarding the hype created by consulting companies and others about positive effects of diversity, they conclude:

” .. Raising expectations regarding universal performance increases, however, appears not to be intellectually honest and may potentially backfire when expected changes do not materialize, and the very foundation provided for diversity initiatives is weakened ..”
(Wallrich et al. 2024, p. 63)

After analyzing the effect of racial/ethnic diversity on performance of executive teams in S&P 500 firms, Sekou Bermiss, Jeremiah Green and John Hand, come to a similar assessment:

“As such, we interpret our results as not supporting the commonly claimed “business case for diversity” when that claim is evaluated at the one-year-ahead overall firm level and with regard to the race/ethnicity of executives in S&P 500 firms over the past decade. Our results suggest that despite the imprimatur often given to influential non-academic studies, caution is warranted in relying on such findings to support the view that US publicly traded firms can create improved financial performance if they increase the racial/ethnic diversity of their executives.”
(Bermiss, Green & Hand, 2023).

Discussion

Here are some quotations from academic articles, summarizing the effect of diversity on business performance:

” .. reviews of the literature (van Knippenberg & Schippers, 2007; Williams & O’Reilly, 1998) as well as meta analyses (Bowers, Pharmer, & Salas, 2000; Webber & Donahue, 2001), could not establish consistent main effects of diversity on performance.”
(Roberge & Van Dick, 2010, p. 295)

“The prevailing academic views of the relations between the racial/ethnic diversity in firms and
corporate business performance are complex and nuanced. In the context of top management teams, where tasks are complex and non-routine, racial/ethnic diversity has been shown to have a positive, negative or zero effect on firm performance .. Several meta-analyses about the link between top management team diversity and firm performance find that many studies show weak or inconclusive evidence (Jeong and Harrison, 2017).”
(Bermiss, Green & Hand, 2023, p. 2)

“Even though popular business books (Syed, 2019) and management magazine articles (Rock & Grant, 2016) tout the promise of diverse teams, empirical findings are mixed. For many facets of diversity and performance, there are large studies that arrive at contrasting results. For instance, ethnic workforce diversity has been associated with better (Moon & Christensen, 2020) and worse (Pitts & Jarry, 2009) performance of US federal agencies.”
(Wallrich et al. 2024, p. 9)

All in all, the picture one gets from the research about the association between diversity and business performance in companies is confusing and contradictory. As seen above, some research shows that diversity is positive for company performance while other research shows negative effects. Lots of research shows no statistically significant effect at all. The only type of research which shows consistently positive effects of diversity is done by consulting companies like McKinsey – which have a financial incentive for coming up with results like that, and thus should be treated with skepticism.

What are the reasons for this confusing state of affairs?

Issues with the Diversity index

One reason could be the issues with measures of diversity. One measure often used (e.g. in McKinsey, 2020) is the standardized inverse Herfindahl-Hirschman diversity index. Applied to racial / ethnic diversity, this index has its maximal value – that is, 1 – if equal numbers of people in the group belong to different ethnic categories. E.g., a team where 25 % of the people are whites, 25 % Asians, 25 % blacks and 25 % latinos, would have a diversity index of 1. In contrast, 100 % Asians would have a minimal diversity index, that is 0.

A problem with this diversity index is illustrated by the fact that a team with 91 % blacks, 3 % Asians, 3 % whites and 3 % latinos would have the same index as a team with 91 % Asians, 3 % whites, 3 % blacks and 3 % latinos. In an industry where mathematical excellency is important – e.g. the IT industry – these two equally diverse groups could be expected to perform dramatically differently, given the well known average superiority of Asians in mathematical problem solving.

A related problem is that in investigations of racial / ethnic diversity the percentage of different ethnicities is often either not reported or not correlated with team performance. For example, one team could have Asians, whites and latinos in it, and another one Middle-Easterners, blacks and latinos. Even if the two teams have the same diversity index value, it is very well possible that they would perform differently.

A further, related problem is described in Wallrich et al. 2024, p. 65: In a hypothetical example Team A would consist of 33 % German, 33 % Dutch and 33 % Danish members. Team B would contain 33 % Congolese, 33 % Chinese, and 33 % Greek members. Note that Team A nd B have the same diversity index with regard to nationality. However, the distance in personalities, cognitive styles and IQ is probably much greater between the three nationalities in team B than in team A. In spite of the identical nationality diversity index of the two teams, we’d probably see differences between the performance of the two teams.

Last but not least, it is a question whether an ethnic diversity index for an organization should have some relationship to the proportions of the groups in the general population or the work force available for the company. For example, as mentioned above, a US organization having 25 % of the people are whites, 25 % Asians, 25 % blacks and 25 % latinos would have the maximal ethnic diversity index value, that is 1. But is this maximal diversity value justified, if the distribution of these four races is very different in the USA the general US population (60.1 % whites, 5.6 % Asians, 12.2 % blacks and 18 % latinos).

Correlation vs. Causation

One other reason for the contradictory results could be that much of the research is based on correlations between diversity and business performance. The problem is that even if a positive correlation is found, it does not necessarily mean that greater diversity caused better performance: correlation does not equal causation. The causality could be in the reverse direction: better performance caused greater diversity. Wallrich et al. 2024 found that very few research projects took this issue into consideration (Wallrich et al. 2024, p. 69). This is the also one of the issues that Jeremiah Green and John Hand found with McKinsey’s analysis:

“McKinsey measures firm financial performance over the four or five years leading up to the year in which they measure the race/ethnicity of the firm’s executives”
(Green & Hand, 2024, p. 25).

McKinsey was aware of this issue with their analysis, and acknowledged that they can’t claim causal influence of diversity on business performance. They write:

“Correlation is not causation, and we are not asserting causal links. As with many levers of business
performance, particularly at such a high level, causation would be challenging to demonstrate, likely
requiring detailed longitudinal studies.”
(McKinsey, 2023)

Nonetheless, the titles and the contents of their reports suggest exactly such a causal effect of diversity (e.g. “Diversity matters”, “Diversity matters even more”, “Delivering through Diversity”, “the business case [for diversity] is the strongest it has been since we’ve been tracking”, etc.).

Note that in their attempt to replicate McKinsey’s results, Green & Hand, 2024 did not correct this issue: their financial performance data preceded the diversity data chronologically, just as in McKinsey’s research. Thus, the discrepancy between McKinsey’s and Green & Hand, 2024’s results was not due to this inverted causation issue. In addition, when Bermiss, Green & Hand, 2023 did correct the causation issue (using diversity data which preceded performance data), they got the same result: no positive effect of diversity on performance. Nonetheless, some of the conflicting results about the effect of diversity could be caused by this confusion about causality.

Consistency of “Diversity” definitions

Another possible reason for contradictory results is that the terms used to describe diversity are not defined clearly and consistently in the different research projects. In other words, “diversity” might not mean the same in project A as in project B. For example, Reynolds & Lewis (2017) investigated the effect of “cognitive diversity” on team performance in a “strategic execution exercise”. They defined “cognitive diversity” as differences in “how individuals think about and engage with new, uncertain, and complex situations”. They found a positive effect of cognitive diversity on team performance in the exercise.

In contrast, as mentioned above, Wallrich et al. (2024) found a negative effect of cognitive diversity on performance, after examining the results of 11 previous meta-analyzes.

This discrepancy in the effect of cognitive diversity might be explained by the fact that the definitions for “cognitive diversity” were different. Wallrich et al. (2024) defined “cognitive diversity” as “educational level, degree, values, personality, intelligence, neurodiversity”. It is clearly different from the definition given in Reynolds & Lewis (2017) – and it covers a much broader set of features.

“Diversity” has both positive and negative effects

Williams & O’Reilly, 1998 write:

‘Milliken and Martins (1996, p. 403), in a comprehensive review of the diversity literature, concluded that “diversity appears to be a double-edged sword, increasing the opportunity of creativity as well as the likelihood that group members will be dissatisfied and fail to identify with the group”.’
(Williams & O’Reilly, 1998, p. 79)

Thus, diversity may have both positive and negative effects. Why?

Williams & O’Reilly (1998) writes that the research which found positive effects was often conducted in an artificial setting – in a laboratory or a classroom – while the research which yielded negative results typically took place in real organizational (e.g. company) work groups. In addition, diversity of personality and of ability tends to have positive effects, while diversity of race and gender tends to have negative effects. (Williams & O’Reilly, 1998, p. 79-80)

Herring (2009) also argues that there might be two effects of diversity: a positive one (increased creativity because of the diversity of perspectives) and a negative one (increased conflict and communication difficulties).

The effect of “Diversity” depends on the team’s task

If the team´s task is to come up with creative solutions to complex problems, there is evidence that there is a small positive effect of diversity. For other tasks not even such a small positive effect might exist:

“For teams that perform tasks which directly benefit from a wide range of perspectives, such as those tasked with research, creativity, and innovation, it might make sense to aim for greater diversity in order to boost performance – even though the average associations remain small ..
In other teams, expected increases in team performance do not provide a strong justification for increasing diversity.”
(Wallrich et al. 2024, p. 64)

The effect of “Diversity” on company performance might be non-linear

Most studies assume that, if there is a relationship between “diversity” and company performance, it is linear. That is, increasing diversity results in increased company performance. However, maybe diversity has to reach a critical mass value before it can have an effect. If diversity in the companies investigated don’t reach such a critical mass, then this could be one reason for not finding positive effects of diversity.

Thus, both Post & Boyd (2015) and Pletzer et al. (2015) mention, as a potential limitation of their research, the low number of women on company boards in the studies they have meta-analyzed. For example, the average percentage of women in the studies examined by Pletzer et al. (2015) was about 14 %. This might be below the critical mass value for female representation on boards, if there is indeed such a value.

Joecks et al. (2012) found such a critical mass:

“.. we find evidence for gender diversity to at first negatively affect firm performance and – only after a “critical mass” of about 30 percent women has been reached – to be associated with higher firm performance than completely male boards. Given our sample firms, the critical mass of 30 percent women translates into an absolute number of about three women on the board ..”
(Joecks et al., 2012, p. 1)

Ferrary & Déo (2022) also found a critical mass of female representation, at middle management and staff level in large French companies.

A non-linear relationship between diversity and firm performance sounds plausible. It is relatively straightforward to investigate for gender diversity because in that case the relative proportion of only two groups – males and females – and only one critical mass value (for females) need to be taken into consideration. It would be far more difficult for other types of diversity, like ethnic, personality, cognitive or age diversity. For example, for ethnic diversity one would have to assume critical mass values for each of the ethnicities in a group. In addition, it is probable that these critical mass values for the same ethnicity would be different in groups with different ethnic composition. Thus, the critical mass value for whites might be different if the group consists of white, black and Asian members than if it consists of white, black and hispanic members.

Moderating factors

The relationship between diversity and company performance can be influenced by many moderating factors. For example, Elena Meliá-Martí and her colleagues (Meliá-Martí et al, 2024) found a negative relationship between gender diversity on corporate boards and company performance. That is, increased diversity was associated with lower performance. However, this negative relationship was only present in companies which did not have a female CEO (or female board chair).

If the company was led by a female CEO (or had a female board chair), then the relationship had a U-shape: initially, company performance decreased as diversity increased, but only up to a point. After this, continuing increase in diversity started to increase performance.

By the way, note that this U-shaped relationship is another example of non-linearity, mentioned in the previous section.

Conclusion

Imagine you are the CEO of a company. You notice more and more positive reports about “diversity” and DEI in prestigious newspapers. In addition, your Human Resources department is more and more vocal about wanting to introduce “diversity” and DEI courses for your employee. When you visit such a course, you are told that these are good things, leading to a better working environment. This is illustrated by pictures of ethnically diverse male and female employees, all of them smiling and apparently happy. There is even one smiling person of color, sitting in a wheelchair.

In addition, you are told that the varied perspectives of a diverse body of employees leads to better decision making and thus to increased business performance. Shiny reports by the McKinsey consulting company, containing many colorful graphs, are shown as the proof for this.

As a CEO, you are naturally interested in this “business case for diversity”. But then, being a skeptical person, you also read some other information, for example this blog post, which lists some of the evidence from scientific research about this. You notice that while reports by consulting companies are extremely bullish about positive effects of diversity, scientific results are much more varied and often contradictory. They give, at the bottom line, very little evidence for a positive effect of most kinds of diversity. You ask: where does this difference come from and who should one believe?

Who should one believe?

The main reason for the discrepancy between scientific results and consulting company reports is probably that the latter are not science.

For example, scientific articles about the effects of diversity always have a review about the current state of the science regarding this topic. Even authors who, in their own research find positive effects of diversity, typically mention other research which found opposing results. Herring, 2009 is an example. In contrast, reports by McKinsey and the Boston Consulting Group are unabashedly positive about the positive effects of diversity – without mentioning any conflicting results.

Similarly, scientific articles about diversity always contain a “Literature” section at the end which contains references to relevant research. For example, the Bermiss, Green & Hand, 2023 article – critiquing McKinsey’s reports – has around 50 references in its Literature section. The meta-analytic article Wallrich et al. 2024 contains a huge Literature section of around 720 references. In contrast, McKinsey’s reports have no Literature section at all, and thus no references to other research.

One of the basics of science is that it insists on the replicability of results. Scientific evidence for a hypothesis grows the more often the results supporting it can be replicated. It is common practice that if the data is in digital form, it is made available for other scientists, too, to allow replication of the results. McKinsey decided not to follow this practice. They refused to make their data available, citing confidentiality issues (Green & Hand, 2024, p. 6). This way they avoided the discovery of potential problems with their data or data analysis.

Scientific research articles undergo a peer review. Before the article gets published in a scientific journal it is sent to a number of other scientists who critique it and recommend it – or not – for publication. Reports by consulting companies don’t undergo this process and can thus contain anything that the consulting company thinks serves its interests.

So, do consulting companies like McKinsey lie when they publish glowing reports about the advantages of diversity, in stark contradiction with the scientific evidence which shows far more conflicting results and, on average, very little positive effect of diversity? Not necessarily. There is scientific research which also shows positive effects. However, as long as McKinsey and others don’t make their data available for replication, it is impossible to say whether their data is real and their data analysis is correct.

There is also a dishonesty issue: as mentioned above, although McKinsey recognizes that their research results do not allow them making causality statements of the kind “diversity positively affects business performance”, this is exactly what they are doing, report after report.

Last but not least, it is clear that McKinsey has a financial interest in coming up with the results that they found.

Does all this mean that one should always believe science? I don’t think so. Scientists are also humans. They can lie and produce fake results, as the recent discovery of apparent fraud committed by a well known Alzheimer’s disease researcher shows. The research about “diversity” and its effects are done by social scientists most of whom are left leaning. Supporting “diversity” and DEI is a leftist issue, and thus, it would not be surprising if this affected scientific research in this area.

For example, Robert Putnam, a well known social scientist, did research in 2000 on ethnic diversity and social trust among members of communities in the USA and found that the larger the diversity, the lower the social trust. Apparently troubled by these results, he did not publish them for 7 years – during which he tried to find alternative explanations. At the end, in 2007, he did publish his results – in a relatively obscure journal (Putnam, 2007).

As another example, consider this quote:

“The high volume of papers investigating the link between BD and firms’ (financial) performance has also spurred a number of meta-analyses on the topic–commonly confirming the positive connection between the two (Pletzer et al., 2015; Post and Byron, 2015; Anastasia et al., 2020).”
(Makkonen, 2022, p. 945)

Note that the author, Teemu Makkonen, quotes three meta-analytic studies which he writes are confirming the positive connection between “board diversity” (BD) and firms’ (financial) performance.

I have already described the first two of those two meta-analyses above. To summarize them here again, Post & Byron (2015) found a “very weak” or “negligible” positive correlation between board gender diversity and accounting returns, and no correlation between board gender diversity and market performance; Pletzer et al. (2015) found no relationship between board gender diversity and company financial performance.

The third meta-analysis, Anastasia et al. (2022), investigated studies made in a number of different countries. They also found only “very weak” correlations between board gender diversity and firm performance, even when differences between countries’ gender equality levels were taken into consideration.

Was this simply sloppiness by Makkonen or was it a sign of bias? I can’t tell. But it is clear that these three meta-analytic studies hardly confirm the “positive connection” between board gender diversity and firms’ (financial) performance – as stated by Makkonen.

All this warrants caution and skepticism not only towards the reports of McKinsey and other consulting companies, but also towards scientific results. However, one has to acknowledge that, in spite of general left-leaning biases, scientific research has a lot more controls and tests built in than research done by consulting companies. Scientific results reflect reality probably better than bombastic reports done by consulting companies.

Of course, scientific studies can come up with conflicting results as well – as shown above. Somebody with an ideological motivation can often find studies which support his ideology. But, if one is interested in the truth more than in supporting one’s ideology, the best strategy probably is to rely on information from meta-analytic studies which combine the results of many individual scientific studies.

What to do?

The “business case for diversity” seems like the Holy Grail of wokeness. Immense efforts have been exerted to find it. But, just as in the case of the mythical Holy Grail, the existence of a “business case for diversity” is highly doubtful, in spite of the hype created by consulting companies like McKinsey and others. If it exists, the diversity-performance relationship might be non-linear, and thus much more complex than previously thought; it seems very much dependent on moderating factors; and, it might be dependent on the way diversity and performance are defined and measured. Stating that “diversity leads to better performance” is highly misleading.

Instead of focusing on “diversity”, for increased productivity and excellence one should strive to identify particular combinations of features in teams. For example, a high diversity of personality types in a team will lead to conflict if some of those personality types are incompatible. Striving blindly for “diversity” in such a case is not just irrelevant but detrimental for team performance. Instead, one should try to include compatible personality types in the team, even though this might decrease this type of diversity. Also, in some teams an imbalance of personality types might be needed: for example, it is possible that the most beneficial combination of personality types is which contains a small percentage of bold, creative “leader” types, while all the rest are less creative, less bold, but hard working types who are happy to implement creative ideas.

Similar considerations should be taken for other types of diversity – e.g. age, gender or race – too. In some cases, increased performance and excellence might mean having no diversity at all. As an example, though both males and females are drafted into the Israeli army, there don’t seem to be any females serving in combat roles: there is zero gender diversity in such roles. This is amply demonstrated by the photos of Israeli soldiers fallen in the current conflict in Gaza and Lebanon, presented daily in Israeli newspapers (like The Times of Israel). With lives on the line, the Israeli army probably cares much less about “diversity” than about maximal efficiency.

Of course, companies might have other reasons to implement “diversity”, too. For example, the law might require employing “marginalized” minorities. Also, given the current “woke” mind set in large parts of the society, there might be societal pressures on the company coming from the media, etc., to implement diversity and DEI measures. Some CEO’s, influenced by the Human Resources department of their company – the source of many of the “woke” developments inside businesses – are maybe even convinced that they are doing the right thing by jumping on the DEI bandwagon. But it should be clear that these reasons have nothing to do with the “business case for diversity”. As described above, there is little current scientific evidence that diversity of the traditional kinds – race, gender, age, etc. – has beneficial effects for businesses. Also, in a number of cases the saying “Go woke go broke” became reality – the US companies Bud Light, Disney, Target and Gillette come to mind. These companies didn’t “go broke” but they lost millions of dollars when they tried to appeal to the “woke” mindset by pushing DEI objectives too hard.

One should be aware that “diversity” is an ideological, political goal, having its origin partly in far-left political movements like Marcuse’s Critical Theory and its descendant, Critical Race Theory. The “business case for diversity” appeared post hoc, only after leftist ideology and politics started to push for diversity. There might not be any reason to search for a “business case for diversity” other than the fact that “diversity” is being forced onto companies, both by law and by social pressure.

Literature

Anastasia, R., Winoto, A., Wijayanto, D., Marbun, P.A. & Purba, H.M.T. (2020). Board Diversity and Firm Performance in All Industries : A Meta Analysis Study. Psychology and Education (2020) 57(9): 5005-5013.

Bermiss, S., Green, J., & Hand, J.R.M (2024). Does Greater Diversity in Executive Race/Ethnicity Reliably Predict Better Future Firm Financial Performance? Journal of Economics, Race & Policy 7(1): 45–60.

Ferrary, M. & Déo, S. (2022). Gender diversity and firm performance: when diversity at middle management and staff levels matter. The International Journal of Human Resource Management
2023, VOL. 34, NO. 14, 2797–2831. https://doi.org/10.1080/09585192.2022.2093121

Green, J. & Hand, J.R.M (2024). McKinsey’s Diversity Matters/Delivers/Wins Results Revisited. Econ Journal Watch 21(1), March 2024: 5–34.

Hanania, R. (2023). The origins of woke. New York : Broadside Books.

Herring, C. (2009). Does diversity pay? Race, gender and the business case for diversity. American
Sociological Review 74(2), 208–224.

Joecks, J., Pull, K. & Vetter, K. (2012). Gender Diversity in the Boardroom and Firm Performance: What Exactly Constitutes a ‘Critical Mass’? Available at SSRN: https://ssrn.com/abstract=2009234 or http://dx.doi.org/10.2139/ssrn.2009234

Makkonen, T. (2022). Board diversity and firm innovation: a meta-analysis. European Journal of Innovation Management, Vol. 25 No. 6, 2022 pp. 941-960. https://doi.org/10.1108/EJIM-09-2021-0474

McKinsey (2023). Diversity matters even more. The case for holistic impact. Report.

McKinsey (2020). Diversity wins. How inclusion matters. Report.

Meliá-Martí, E., Tormo-Carbó, G. & Fernández-Guadano, J. (2024). Board gender diversity and employee productivity. The moderating role of female leaders. European Research on Management and Business Economics, Volume 30, Issue 3, September–December 2024, 100257 https://doi.org/10.1016/j.iedeen.2024.100257

Pletzer, J.L., Nikolova, R., Kedzior, K.K. & Voelpel, S.C. (2015). Does Gender Matter? Female Representation on Corporate Boards and Firm Financial Performance – A Meta-Analysis. PLoS ONE 10(6): 1-20.

Post, C. & Byron, K. (2015). Women on boards and financial performance: A meta-analysis. Pre-print, accepted for publication in Academy of Management Journal, Vol. 58, No. 5.

Putnam, R.D. (2007). E Pluribus Unum: Diversity and Community in the Twenty-first Century The 2006 Johan Skytte Prize Lecture. Scandinavian Political Studies. Volume 30, Issue 2.137-174.

Reynolds, A. & Lewis, D. (2017). Teams Solve Problems Faster When They’re More Cognitively Diverse. https://hbr.org/2017/03/teams-solve-problems-faster-when-theyre-more-cognitively-diverse

Roberge, M-E. & van Dick, R. (2010). Recognizing the benefits of diversity: When and how does diversity increase group performance? Human Resource Management Review 20 (2010) 295–308.

Stojmenovska, D., Bol, T. & Leopold, T. (2019). Does Diversity Pay? A Replication of Herring (2009). American Sociological Review 1–11.

Wallrich, L., Opara, V., Wesołowska, M., Barnoth, D. & Yousefi, S. (2024). The relationship between team diversity and team performance: reconciling promise and reality through a comprehensive meta-analysis registered report. Pre-print, accepted for publication in: Journal of Business and Psychology, Volume 39, pages 1303–1354, (2024). https://link.springer.com/article/10.1007/s10869-024-09977-0

Williams, K. & O’Reilly, C. (1998). Forty years of diversity research: A review. In Staw B. M. & Cummings L.L. (Eds.), Research in Organizational Behavior (pp. 77–140). Greenwich, CT: JAI Press.