Representation and belonging are not the same thing. Organisations that have learned this the hard way share a consistent pattern. Here is what it is, and what to do about it.
Increasing the diversity number without building the belonging culture produces a specific and costly outcome.
There is a pattern that appears in organisations that have invested seriously in diversity representation targets and have met them. The diversity number went up. Women, or people from non-dominant backgrounds, are now present at senior levels in numbers that reflect the organisation’s stated commitments. And yet, in the belonging surveys that run alongside the diversity tracking, the scores for the newly represented groups are declining rather than improving. They feel less that their voices are heard. They report higher levels of experience of being overlooked. They leave at higher rates than the dominant group at the twelve-month and twenty-four-month marks, creating a diversity revolving door that the recruitment metrics do not yet reflect. The organisation’s diversity number continues to look good because new diverse hires keep coming in. What is not visible is the rate at which the culture is chewing them up and spitting them out.
This pattern has a name. Researchers call it the DIVERSITY-BELONGING PARADOX: the finding, replicated across multiple studies, that increasing representation without investing in belonging culture can, under certain conditions, produce worse belonging outcomes for the newly represented groups than existed before the representation effort began. The mechanism is not mysterious. When diverse talent is added to a culture that has not been changed to include them, the culture does not absorb them. It assimilates them. It imposes its existing norms, behaviours, and implicit rules on people who did not grow up in the dominant group, and it evaluates their performance, their leadership style, and their cultural fit against a standard that was designed by and for people who are unlike them. The diversity number reflects who is in the building. The belonging culture reflects whether they can be themselves in it. These are not the same question, and organisations that treat representation as the primary success metric are, in many cases, answering the wrong one.
The Assimilation Tax
The term COVERING, developed by Kenji Yoshino in his research on identity management in professional settings, describes the work that people from non-dominant groups perform to manage the aspects of their identity that the dominant culture does not fully accommodate. Women who modify their communication style to match masculine norms of directness and authority. Leaders from non-metropolitan backgrounds who adopt particular registers of English or particular social behaviours to signal cultural fit with elite-educated peers. Professionals who do not acknowledge family responsibilities in work conversations because the culture treats such acknowledgement as a signal of reduced commitment. This covering is not a marginal phenomenon. Catalyst’s research on covering in the workplace finds that significant proportions of employees from every non-dominant group engage in covering behaviour, and that covering is associated with lower engagement, lower performance, and higher turnover intention.
The covering tax is a belonging tax. It is the cost that non-dominant group members pay to exist in a culture that has not been designed to include them. And it is a cost that most organisations are not measuring, because the instruments they are using, diversity representation metrics and general engagement surveys, are not designed to detect it. The representation metric shows the diverse hire is present. The engagement survey, which is typically designed for the dominant culture and asks questions whose norms are set by the dominant experience, may not capture the specific belonging deficit that covering produces. The organisation sees good numbers and is surprised when the diverse hire leaves at the two-year mark. What it is not seeing is the daily cost of performing a version of yourself that the culture can accept, which accumulates over months and years into a level of depletion that no career opportunity can fully compensate for.
In Indian corporate contexts, the covering tax takes specific forms that global frameworks do not fully capture. The language covering that regional professionals perform to approximate the dominant English register. The educational background covering that non-IIT/IIM graduates perform to signal credibility with pedigreed peers. The gender covering that women perform to be taken seriously in male-dominated leadership cultures. The caste covering that Dalit professionals perform in organisations that have invested in representation without acknowledging the cultural norms that reflect caste hierarchy. These forms of covering are not equivalent in their severity, but they share a common structure: the culture has room for the person’s skills while not having room for the person’s full identity. This is not belonging. It is conditional inclusion, and the conditions are set by the dominant group.
The Structural Cause
The diversity-belonging paradox is not a leadership failure in the conventional sense. Most organisations that have invested in diversity representation and are experiencing belonging deficits in their diverse talent populations are led by people who genuinely want both outcomes. The problem is structural. The organisation’s culture, the sum of its norms, behaviours, and implicit rules, is a system. Systems have homeostasis: they resist change and tend to return to their previous state when the input that produced the change is removed. Adding diverse hires to the system changes the composition of the system without changing the norms. The system then applies its existing norms to the new composition, and the result is either assimilation, where the new members adapt to the existing norms at significant personal cost, or exit, where the new members who cannot or will not assimilate leave.
The intervention required to break this cycle is not more representation investment. It is CULTURAL NORM REDESIGN, the deliberate examination and modification of the implicit rules that govern how the organisation works, specifically to identify and change the ones that disproportionately burden non-dominant group members. This is harder to design than a recruitment target. It requires facilitated conversations with multiple stakeholder groups about what the current culture actually requires of people from different backgrounds. It requires the courage of the dominant group members to examine their own normative assumptions rather than treating the culture as a neutral background. And it requires the patience of leadership to pursue a change that is measured in years, not quarters, and that is legible in belonging data and retention data rather than in the representation number that is easiest to report.
The leadership commitment required for this work is qualitatively different from the commitment to set a diversity target. Setting a diversity target is a goal-setting exercise. It is important, and in many organisations it has been what created the conditions for the representation progress that has occurred. But changing a culture to genuinely include the people who now populate it is not a goal-setting exercise. It is an organisational development exercise of significant complexity and duration, requiring the same quality of sustained attention and senior commitment that any complex transformation requires. The organisations that are most likely to produce genuine equity progress in the next five years are the ones that understand this distinction and have committed to the harder work it requires.
“If your diversity number is going up and your belonging data is not improving, the system is working exactly as designed. It is including people’s skills while excluding their identities. That is not diversity. That is a more equitable distribution of assimilation pressure.”
RK, Founder & Principal Consultant, ProventusHRResearch Reference
Yoshino, K. (2006). Covering: The Hidden Assault on Our Civil Rights. Random House. Catalyst (2023). Covering in the Workplace. McKinsey & Company (2022). Women in the Workplace. Lean In (2023). Women in the Workplace Annual Report.
The ProventusHR Perspective
ProventusHR’s approach to the diversity-belonging paradox begins with the diagnostic question that most organisations are not asking: which specific cultural norms are creating covering behaviour in which specific populations, and how are those norms encoded in the organisation’s processes, practices, and leadership behaviours? This question requires a qualitative diagnostic that goes beyond standard culture survey data. ProventusHR uses a combination of structured focus groups with non-dominant group members, facilitated leadership team conversations about cultural norm assumptions, and direct observation of key organisational processes to build a specific picture of where and how the culture is creating covering pressure.
The intervention that follows from this diagnostic is targeted at the specific norms and processes identified, rather than the general culture. This means working with the leadership team to make explicit the implicit rules that the culture has been running on, examining which of those rules serve the organisation’s performance goals and which serve only the comfort of the dominant group, and redesigning the ones that have been creating unnecessary barriers. This is not a painless process for the dominant group. It requires confronting the fact that many of the norms they experience as neutral are, in fact, structurally advantaging them over colleagues who are performing equivalent or superior work. The discomfort of this confrontation is the cost of genuine inclusion, and ProventusHR’s facilitation is designed to hold it productively rather than to manage it away.
For L&D leaders, the practical recommendation is direct: measure belonging separately from engagement, and measure it for each population rather than for the organisation overall. A belonging survey that produces an aggregate score is hiding the diversity of belonging experiences within it. A belonging survey that shows the distribution of belonging scores across gender, educational background, regional origin, and tenure tells you where the culture is working and where it is not, and it gives you the specific intervention targets that a general engagement score cannot. This disaggregated data is the starting point for addressing the representation-belonging gap. Without it, the organisation is, in a very real sense, flying blind on the dimension that matters most for retaining the talent it has invested in acquiring.
Every engagement begins with a discovery conversation. No templates, no generic programmes; only work that fits the context, the culture, and the people involved.
Start a Discovery Conversation