Alex Edmans, London Business School, CEPR, and ECGI; Caroline Flammer, Columbia University, NBER, and ECGI; Simon Glossner, Federal Reserve Board



Companies, investors, policy makers, and wider society are paying increased attention to diversity, equity, and inclusion (DEI) within firms. DEI initiatives have two motivations – to improve a company’s long-term financial performance, and contribute to societal goals. 

Under both motives, the relevant measures of DEI are holistic. New ideas, and thus superior financial performance, stem from cognitive rather than purely demographic diversity. Social outcomes stem from providing opportunities to underrepresented groups across all areas, such as demographic, socioeconomic, and educational. Moreover, both goals require not only diversity but also equity and inclusion. Hiring minorities to tick a box, but failing to ensure that they can thrive at work, will achieve neither the financial benefits of cognitive diversity nor the social outcomes of meaningful employment.

In practice, DEI initiatives focus almost exclusively on demographic diversity because it is easy to measure. Such initiatives have been justified by papers claiming a strong link between demographic diversity and financial returns, but they are deeply flawed. For example, numerous McKinsey studies claim a positive link between ethnic diversity and firm performance, but the results cannot be replicated even with their chosen performance measure (EBIT) and preferred methodology. Moreover, there is no link with other performance measures – gross margin, return on assets, return on equity, sales growth, or total shareholder return – or when using more established methodologies. The UK’s Financial Reporting Council published a study claiming that “Higher levels of gender diversity of FTSE 350 boards positively correlate with better future financial performance (as measured by EBITDA margin)”. Yet out of their 90 regressions linking diversity to the EBITDA margin, zero are significant.

Beyond demographic diversity

In a recent paper, we take a first step towards measuring the DEI of a company, employing proprietary data used by the Great Place to Work® (GPTW) to compile the list of the 100 Best Companies to Work For in America. Two thirds of the score that determines list inclusion stems from employee responses to the Trust IndexTM, a 58-question survey on various dimensions of employee satisfaction. Prior research by one of us found that the Best Companies enjoyed superior long-term shareholder returns over a 28-year period, suggesting that the dimensions captured by the list are financially material. However, this prior study only investigated list inclusion, since only this is public.

Via a confidentiality agreement with GPTW, we obtained the individual responses to all 58 questions from 2006 to 2021. We identified 13 questions that cover DEI, such as “This is a psychologically and emotionally healthy place to work”, “I can be myself around here”, and “Managers avoid playing favorites.” We aggregate employee responses to form a measure that we call DEI. Our measure is a ’grassroots’ indicator of actual DEI, as perceived by the responses of 250-5,000 employees in a company, in contrast to more superficial measures such as whether a company has a DEI policy.

Hitting the target, missing the point

We find that DEI is only weakly correlated with traditional measures of demographic diversity, such as the percentages of women and ethnic minorities in the board, senior management, CEO position, and wider workforce. Thus, DEI contains incremental information that would be missed by standard metrics that focus more narrowly on demographic diversity. This has implications for the substantial attention paid to diversity metrics – they omit a big piece of the picture. Companies can ’hit the target, but miss the point’ – improve diversity statistics without improving DEI.

The characteristics of high-DEI firms

We then study the ‘determinants’ of DEI – not in a causal sense, because both financial performance and firm characteristics are endogenous – but to understand what types of firms are associated with higher DEI. We find that DEI is positively associated with one- and three-year sales growth, positively associated with three-year but not one-year stock returns, negatively associated with leverage, and positively associated with dividends. This suggests that a strong financial position frees a company from having to focus on short-term pressures and instead allows it to address longer-term challenges such as DEI. Small and growth (low book-to-market) firms are associated with higher DEI, consistent with their greater ability to increase DEI, given management’s proximity to employees, and their greater incentives to do so, given the importance of human capital in such firms.

We then study the workplace policies and practices that are associated with DEI. DEI is positively correlated with the proportion of women in senior management, but unrelated to the presence of a female CEO and negatively linked to the percentage of women on the board. It is either unrelated or negatively related to ethnic diversity in senior management, at the CEO level, or in the boardroom. The insignificance of most demographic diversity variables suggests that an ’add diversity and stir’ approach is insufficient to improve DEI.

We also explore workplace policies, such as childcare, unpaid parental leave days, sabbaticals, and flexitime, to test whether DEI can be easily increased by implementing simple policies. All workplace policies are insignificantly associated with DEI, suggesting that improving DEI requires specific, targeted initiatives.

We then turn to the consequences of DEI, which we again do not interpret causally. We find that DEI is positively associated with seven out of eight measures of future profitability, such as return on assets, return on sales, profits divided by employees, and sales divided by employees. These results are after controlling for the percentages of female and minority employees; indeed, these variables are insignificant. In short, DEI is correlated with higher profits, but demographic diversity alone is not. We also find that DEI is positively associated with valuation measures, such as Tobin’s Q, suggesting that the market at least partially incorporates the value of DEI. However, DEI is positively linked to future earnings surprises, indicating that the market does not fully incorporate the performance benefits of DEI.

We also study future innovation performance. DEI is unrelated to either the number of future patents or patent citations. However, the granular nature of our data allows us to stratify the survey responses by job category. We find that DEI perceptions of professionals, a job category that includes R&D staff, are positively and significantly correlated with both innovation measures, but there is no positive link with the responses from executives, managers, and hourly workers. This is consistent with the fact that innovation is most likely to stem from professionals.

Beyond diversity, our paper highlights the dangers of a box-ticking approach to sustainability. Such an approach extends to other sustainability issues such as pay fairness, which is often reduced to the CEO-to-worker pay ratio; environmental performance, which is captured predominantly by carbon emissions; and human capital, which is typically measured by employee turnover or having certain workplace policies. Defenders of such metrics argue that partial information is always better than no information, even if it’s not the complete picture. However, such metrics may lead to companies, investors, and stakeholders focusing excessively on quantitative factors at the expense of the qualitative. This does not mean that these measures have no value, but that users should be aware of their incompleteness and look beyond the numbers.

You can read the full paper first published in May 2023.

This blog was also presented at the 2023 Academic Network Conference at PRI in Person, Tokyo. Watch here