There are two paradigm shifts affecting businesses today. These two shifts are diversity, equity, and inclusion (DEI) issues and environmental, social, and governance (ESG) concerns. This article focuses on DEI and the issues surrounding how metrics are developed, measured, and reported.
Diversity, equity, and inclusion (DEI) efforts serve a societal good. Organizations are going to select from among a variety of options regarding which variables to measure and collect, and how to report their progress on DEI efforts. There is no right answer, no one-size fits all solution. But there are considerations that should be discussed and understood before decisions are made. There have been reports of “window dressing” an organization’s progress with diversity. Well-meaning and well-intentioned action should not be a cause of headline risk and negative publicity, and careful consideration of the measuring and reporting issues should mitigate those risks.
When it comes to DEI, even honest, accurate, and straightforward reporting can be misleading. It is important to understand the myriad possible interpretations of the reporting to avoid others from ascribing inaccurate motivations to it. Having an understanding of the potential for misleading reporting allows an organization to provide a fuller explanation of what it is trying to convey and avoid negative connotations.
Organizations that adopt DEI initiatives are attempting to do their part to increase the opportunities available to populations that face an unlevel playing field. How an organization measures and reports its efforts will provide a scorecard of how successful it has been. As organizations begin to consider the options available, it becomes obvious that there are numerous ways to ensure fairness.
If an organization reported that its five-member board of directors was diversified, composed of 20% female, 20% person of color, 20% person with a disability, and 20% LGBTQ+, what might an observer be picturing? Might a reasonable interpretation be one white man, one female, one person of color, one person with a disability, and one LGBTQ+ person? Or alternatively, would it be four white men and one female who is of color, with a disability, and LGBTQ+? Whichever the reality, the reporting is accurate, but potentially misleading. Just reporting the percentage breakdown in this case is not sufficient; a potential solution would be to include a metric of how many individuals these traits were dispersed over. In the case where each trait corresponded to a unique individual, the answer would be four (one person is female, a different person is of color, a different person has a disability, and a different person is LGBTQ+), as opposed to the extreme case, where the answer is one (a single person has all four traits).
How about a press release from an organization that reports that they have increased their senior female leadership fourfold? Impressive? Yes, but somewhat less so if they had 1,000 employees in senior leadership positions and they went from one female to five. The reporting is still accurate, but the underlying truth needs more detail to be properly understood in context. One alternative would be to include the overall ratios, beginning and ending, to provide context.
An organization utilizes the services of an Outsourced Chief Investment Officer (OCIO). An OCIO is a company that performs the duties of an investment office. In this case, the OCIO has discretion, meaning that it makes the investment decisions on behalf of the organization, which primarily includes the hiring of managers who do the actual investing of the organization’s funds. The OCIO is correctly described as meeting the organization’s diversity criteria. The organization reported that 100% of their funds were invested by firms meeting their standard for inclusion as a diverse organization. But in reality, none of the managers who do the actual investing met the organization’s requirements to be considered diverse. The organization utilized a thought process where it considered that, because the OCIO met the diversity standard, everything from that point forward met the diversity standard as well. Most would disagree with this line of thought.
This article will explore the thought process and procedures behind the decisions that need to be made to help an organization minimize headline risk due to inaccurate or misleading DEI measuring and reporting.
Defining the Focus
DEI is a broad topic that can comprise a multitude of measurable and reportable traits. There are two considerations regarding focus: first, the attributes the individual has and second, the role the individual has in the organization. An organization should consider what it wants to focus on, whether it be one, two, or more attributes. Some common areas for concentration are (in no particular order): gender, race, ethnicity, age, religion, sexual orientation, possession of a disability, veteran status, as well as others.
There are also sub-areas in each. For example, race can be subdivided in many ways: underrepresented populations, BIPOC (Black, Indigenous, and people of color), Latinx, Indigenous people, and others.
It would be quite difficult for an organization to prioritize all of these items simultaneously. In order to be manageable, a sample of the possible characteristics should be selected for the organization to concentrate on. Not all organizations will agree upon which characteristics are most imperative; just as some donors prioritize giving to the arts, and others medical research, there is no right or wrong.
The areas and sub-areas should be unambiguous and mutually exclusive. If an organization chose to use three areas of race (BIPOC, Black, and Latinx) as its focus, someone who was Black could be categorized as BIPOC or Black, likewise someone who was Latinx. Having mutually exclusive categories is important to proper measuring and reporting.
For the ease of illustration, this article will use gender and underrepresented populations as the two areas of focus. This is arbitrary; any two could be utilized and it could be expanded to incorporate more. The considerations do not change because more characteristics are incorporated.
After determining the focus, an organization needs to consider the roles that individuals in the organization hold. Commonly, the roles utilized are ownership, leadership, and total staff. Leadership and total staff are denominated by head count, whereas ownership can be trickier—the number of owners is important, as well as the percentage of ownership.
Unit of Account
Determining the unit of account is a critical decision in measuring and reporting DEI efforts. When collecting and reporting data from your organization’s staff, the consideration is relatively simple; but when you are measuring and aggregating data from other organizations, such as the diversity of vendors or investment managers, it becomes more complicated. Do you count the individuals from across all your vendors or investment managers (individuals are the unit of account), or do you count the vendor or investment manager organizations who meet a certain threshold (e.g., the leadership must be composed of at least 50% of individuals who meet the standard chosen, making the organization the unit of account)?
An organization utilizes two firms for a certain purpose and is interested in measuring female ownership. Firm A has ten owners, including a female who owns 20%. Firm B has ten owners, and nine females are owners and together own 90%. If the measurement is based upon ownership percentage, then firm A has 20% female ownership and firm B has 90% female ownership, providing an average percentage female ownership of 55% [(20% + 90%)/2]. This might be interpreted to mean that all vendors are majority owned by females, when it is only one of the two firms that have female majority ownership.
The following more complete disclosure would help minimize misinterpretation:
Using the information from the previous example but changing the measurement from percentage ownership to head count, there are two firms with a combined ownership of twenty individuals; in this group, ten are female, indicating the 50% of the owners who are female. This is not misleading in this case because 50% of the firms are controlled by females, but a better presentation would be to expand the presentation from Example 4:
Changing the previous example so that each of the ten female owners (one in firm A, nine in firm B) each own 1% would provide that 50% of the owners are female (ten female owners out of twenty total owners), but the percentage owned by females is 1% in firm A and 9% in firm B; this represents a total ownership of 10% and an average ownership of 5%. This is clearly a different metric than stating the head count percentage of 50% of the ownership is female, even though both are accurate statements.
As in Example 5, a more complete presentation would be as follows:
Double counting is an issue when an individual possesses more than one of the characteristics being measured by an organization, such as in Example 1 when a single person could be female, of color, with a disability and LGBTQ+. Providing a single measurement can lead to the misinterpretation of how many individuals are represented. When an individual can check multiple boxes, each characteristic needs to be reported separately—and even then, it might still be misleading.
Using gender and underrepresented populations as the focus characteristics and leadership as the area, assume there was a firm with ten leaders and one of these leaders was female (not from an underrepresented group) and there was another female who was from an underrepresented group. One way of reporting this is to say that, of the ten leaders, 30% are women or from underrepresented groups. This would be accurate, but misleading; because the unit of account is head count, it would appear that three individuals meet the trait of gender or the trait of coming from an underrepresented group, when it is actually only two individuals. As in Example 1, stating how many individuals the 30% is dispersed over (in this case, two) would be useful in preventing misinterpretation of the data.
Another issue that needs to be decided is what happens when an individual represents multiple underrepresented groups. At a granular level, an individual can be 25% each of four underrep-resented groups. At the aggregate, the individual is 100% of the underrepresented category. An organization needs to develop a minimum threshold to determine the lowest percentage at which the individual can be said to be included in the statistic for the category. With larger populations aggregating fractional representation can lead misleading results. For example, if you have 1,000 employees where each is each 1% of a particular trait, would it be fair to say that 10 employees comprise that trait (1,000 × 1%)?
An organization has a focus on underrepresented populations. Specifically, they are measuring the percentage of their staff that are of color, Latinx, and white. They have an individual working for the organization that is 50% of color and 50% Latinx. This individual meets the criteria to be considered underrepresented, but it is less clear whether the individual should be counted as part of the subgroup of color or Latinx, both, or neither.
The organization should adopt a minimum percentage to use for inclusion. Although a minimum of 51% would indicate that this is their dominant make-up, that would also mean two characteristics of 50% each would not be captured. It is possible to adopt a procedure where an individual with a 50/50 division between subgroups is placed into one category or the other, but not both nor neither. This may be left to the organization to decide, or perhaps the individual chooses which they most closely identify with, as long as the process is transparent.
Changing the percentages from Example 8 to 25% of color, 25% Latinx, 25% Indigenous, and 25% white yields a similar dilemma—if you need above 50% for individuals to be counted as being from the underrepresented population that standard has been met, but in which sub-class should they be counted? In percentage terms, they have an equal percentage of all race classes, but should there be a minimum threshold percentage necessary to be counted? Using the 50% threshold from before would not allow inclusion in any subgroup computation. A different rule, perhaps that the sub-group with the highest percentage is used, would aid in classification in some situations, but not this one. Perhaps the best solution is to allow the organization to select which subclass to place the individual into. A statement could accompany the reporting along the lines of: “In cases where individuals have multiple fractional representation, they are placed in the subgroup that has the highest percentage. Where the highest percentage is shared by more than one subgroup, the individual is placed in the subgroup selected by the organization.” It would be useful to specify the criteria of how the organization makes the placement decision.
Inter-organization or Intra-organization
There is research that suggests that the more diverse a group of decision makers is, the better decisions they make. Let’s say an entity has two investment managers, and the characteristic that diversity is upon based on is ownership. Investment manager 1 is 100% owned by white males, and investment manager 2 is 100% owned by females of color. The entity could say that it has 50% gender diversity and 50% racial diversity across the ownership in their manager pool. Within each investment manager, however, there is no diversity, because as each investment manager is owned by individuals categorized similarly according to the framework being used.
Does the better decision making come from diversity within each of the investment managers or across the investment managers? This is not commonly discussed, but it is important if better decision making is used a rationale for hiring diverse managers. Does the diversity derive from the manager being diverse from other managers, or does it derive from the diversity of ownership within the manager’s ranks? This is not to say one is better than the other, just that an organization needs to be clear about what it is doing and why it is doing it. It would be perfectly appropriate to hire an entity that has no diversity within itself if this nondiverse population meets the definition of the diversity the organization is trying to achieve.
Size and Scope
Any time an organization utilizes multiple entities within a category, there is the issue of size and scope when measuring and reporting the results of DEI efforts.
Assume an entity hires two law firms to represent them. One firm is not diverse, the other is fully diverse. The organization can report that it has achieved a level of diversity, as 50% of the law firms meet its definition of being diverse.
Taking the two law firms from Example 10, the amount of business (billings) each firm receives is now included. If the nondiverse firm received 80% of the work, then how should that be reported? These are not issues easily resolved. Perhaps the best approach is to be fully transparent and provide multiple metrics to capture the nuance and complexity of measurement and reporting.
An organization has ten investment managers currently investing funds on its behalf. One manager does not meet the standard to be considered diverse (having the majority of ownership percentage held by an underrepresented population), whereas the other nine do. The non-diverse manager holds 90% of the funds, whereas the other nine share the remaining 10%. What is the proper way to report the diversity of the entity’s investment managers?
An Ongoing Mission
Organizations that have embarked on a mission to increase the representation of owner, employees, firms, entities, and companies in areas where these groups have been historically underrepresented deserve credit. There are numerous ways to define, measure, and report these efforts; without careful consideration, this can lead to misinterpretation and, worse, claims of misleading reporting.
Every organization should be able to clearly present what they have done. It is the responsibility of the reporting party to ensure that they have thought through the different interpretations of what they report and have taken steps to minimize the potential for misinterpretation. The organization should also not hesitate to adjust their policies, procedures, and processes when difficulties arise. Transparency is the key to fair, clear, concise, and straightforward reporting.
Author’s note: The preferred language when discussing diversity, equity, and inclusion is fluid. In order to avoid unintended offense, the “Inclusive Language Guidelines” of the American Psychological Association (https://www.apa.org/about/apa/equity-diversity-inclusion/language-guidelines) was used.