Can Data Drive Racial Equity?

The summer of 2020 saw an unprecedented number of corporate leaders publicly acknowledge and condemn institutional and structural racism, as part of a national reckoning on racial injustice. Issued in response to widespread protests against police violence, these statements called for an end to racial disparities across all systems, including health care, education, housing, criminal justice, and employment. This movement emerged at a moment when people of color were more likely to be working on the dangerous front lines of a pandemic, for low wages and with scant protective gear, while higher-wage earners sheltered at home.

Businesses claiming to support racial equity are, in reality, committing to change facts such as these: Black workers are more likely to work in low-wage industries with high turnover, earn less than their white counterparts, and report a median net worth of one-tenth that of whites. This fixed inequality holds even for Black college graduates, whose households report a lower net worth than those headed by white high school dropouts. The racial wealth gap is so entrenched that by some estimates, at the current growth rate, it would take Black families 228 years to amass the wealth that white families currently hold.1

Racial equity describes an alternate reality, in which race no longer predicts, in a statistical sense, how one fares. While most companies recognize that statements alone will not bring their workplaces — or the nation — closer to that reality, few are confident in their next steps. Fortunately, most employers already have what they need to guide their course: data. Over the past decade, data analytics has been used to improve the quality of products and services, improve efficiency in production and distribution, and fundamentally shape business models. Likewise, a different approach to analyzing your workforce data can help identify measurable and meaningful steps toward a more equitable workplace.

Why We Need a Data-Driven Racial Equity Framework

Racial equity strategies must be systemic, race-explicit, and outcome-oriented if they are to succeed.2 To better understand how well employers and other stakeholders in the world of workforce development are prepared to pursue such strategies, I conducted a research study that measured the degree to which participants had an accurate understanding of racial equity (as distinguished from diversity or inclusion), identified measurable equity outcomes, collected and disaggregated data to measure those outcomes, and implemented race-explicit policies with the goal of advancing equity.3

I found that while many employers expressed a strong desire to create equitable and inclusive workplaces, few had identified specific, measurable outcomes consistent with those goals. Even fewer had analyzed their policies and practices to determine whether they were helping or hindering their efforts. This makes it difficult, if not impossible, to measure progress.

Meanwhile, employers have access to an ever-increasing volume of workforce data sets and the technology needed to analyze them. However, our study revealed that most employers are not mining that data to better understand the equity impacts of their policies and practices, or to inform institutional change. One way for managers to craft more effective solutions is to adopt a data-driven framework that commits their organizations to do the following:

1. Collect, disaggregate, and analyze data related to race and ethnicity.
2. Identify racial disparities in workforce outcomes.
3. Name race explicitly when talking about disparities.
4. Investigate structural causes of racial disparities.
5. Develop strategies to eliminate policies, practices, and messages that reinforce differential outcomes by race.

Engaging in critical retention points like onboarding, evaluation, or promotion without applying this framework is likely to result in an avoidable loss of talent and revenue. Conversely, being able to pinpoint the locations of racial barriers makes it possible to eliminate them. While applying a framework like this requires work, it’s par for the course for any initiative that senior leadership has blessed. As one employer observed, “It doesn’t mean anything unless it is being measured, resourced, and supported. Until that happens, it’s all conversation.”

Collect and Analyze the Right Data to Identify Disparities

If your organization is large, it may already be collecting workforce data to comply with federal reporting requirements or to address issues episodically, such as a spike in attrition. These one-time statistical snapshots may aggregate the experiences of all workers, which can mask racial disparities. Disaggregating your data by race provides more valuable analytics.

While disaggregation may be simple, deciding what to do with the data can be far more complex. For example, equity in compensation and advancement signals to all employees that they have equal prospects for long-term growth, which is critical for retention. However, the ways in which race shapes those prospects and intersects with other policies such as seniority is complicated. Perhaps your recent hires are a more diverse cohort than your long-term employees, but your compensation policy values length of tenure over performance. Maybe your training programs are race-neutral but target female employees to remediate historic gender discrimination. Advanced analytics may flag these as racial equity issues; the most effective approach to resolving them remains a human decision.

Small businesses may not have large-scale data from which to glean patterns, but they can use qualitative data to shed light on the experiences of their employees, which can be equally powerful.4 Asking the right questions can reveal red flags — for example, that openings are largely communicated through word of mouth. While small businesses often rely on informal pipelines and personal connections, racialized networks — often resulting from residential and educational segregation — may fortify barriers to opportunity. This, too, is data, suggesting the need to identify new avenues of recruitment. Another option for small companies in search of more data is to pool data across their sector, revealing industry standards against which small businesses can benchmark.

Consider Broader Societal and Systemic Factors

Despite the recent explosion in analytical software designed to detect payroll disparities, such tools may reveal little about the root causes of those inequities. For example, when it comes to gender, one HR analyst, having used the software, insisted that “when holding all other factors constant, there is actually very little salary inequity.” What does that mean? An equity framework challenges the assumption that in searching for disparities we should “hold all other factors constant.” Indeed, it is precisely within those “other factors” that structural discrimination can hide.

Historic de jure discrimination in housing, education, and employment formed the bedrock upon which today’s de facto segregation endures. If we strip from the analysis any information about an employee’s degree-granting institution, or length of corporate tenure, or rate of promotion, we may simply be ignoring — while wearing high-tech blinders — the very barriers we hope to remove.

Thus, an effective equity strategy must not only tackle the effects of systemic racism but also its root causes; it should seek to change systems, not individuals. A lack of diversity in your front-line workforce, for example, may reflect long-term disinvestment in public transportation and deep-rooted residential segregation. Just as employers have long advocated for public schools to equip students with future-focused skills, these data insights may propel you toward advocating for more equitable systems of transportation and community development.

Encourage Conversations About Race and Inequity

If we are to name race explicitly when talking about disparities, we must dramatically increase our capacity to have these conversations in the workplace. Our study found that employers are reluctant to speak about race and racism, often defaulting to race-neutral language like “economic opportunity” or “inclusion.” What our recent national reckoning has made plain, however, is that deleting race from the corporate vocabulary and creating color-blind policies that ignore race neither eliminate inequality nor foster inclusion. For most business leaders, developing the ability to facilitate or even productively participate in conversations about race will require a lifelong practice of learning and self-reflection.

Regardless of whether you feel comfortable talking about race, the demographics of your company are already telling a story. One hospitality employer explained, “Racial equity in the workplace is having leadership reflect the same diversity as first-level line associates. You show me an organization where everybody at the first line is Black or brown, and then you show me a leadership structure that is mostly white: There is a problem.” An inability to acknowledge such disparities bodes poorly for any leader’s efforts to eliminate them. One solution is to engage an external facilitator to help establish a common language, conduct an audit of existing policies and practices, and develop specific, measurable goals.

Once your organization has those goals in place, a data-driven framework can help establish the practice of racial equity as a managerial habit. For example, one manager described her first attempts to apply the framework, which for her meant “literally writing at the top of every meeting agenda, ‘How does this affect racial equity?’” At first it felt forced, she confessed, but over time this simple daily exercise helped create a culture where it was more “automatic” to ask, “What’s the racial equity and inclusion impact of that specific [fill in the blank] policy, tactic, data piece?”

In another workplace, the leadership team made a commitment to racial equity an explicit part of its hiring process. “It’s in the job description, and it’s in the interview. It’s in the questions we ask of references, and it’s a core criterion for annual performance evaluations. Every search committee includes multiple staff members across lines of difference to inform the evaluation of that candidate. It can be done,” said a manager in that organization. Over time, small but intentional changes such as these can measurably improve diversity, equity, and inclusion.

Remember That Data and Analytics Can Carry Bias

Proponents of the use of AI and machine learning tout their potential to eliminate racial bias in human decisions on hiring, promotions, salary rates, and training. Algorithms that consider experience and skills known to correlate with success can identify applicants who might otherwise be overlooked due to affinity bias or “like me” bias.5 In this way, large-scale data systems can help combat the kinds of implicit bias that can give rise to illegal discrimination. Similarly, the use of objective data to drive employment decisions can help mitigate wage disparities or occupational segregation long after an employee has been hired.

However, big data sets can also replicate racial bias by teaching machines to discriminate based on seemingly neutral proxies for race, such as educational institution, geographic location, and training history.6 As one manager cautioned, “One thing I struggle with is that the data we are collecting is affected by larger systems of racism.” While many employers have begun to integrate AI into HR decision-making, there is a valid concern that a failure to attend to bias in training data may result in technology tools that mirror rather than eliminate human biases. Applying an equity framework can help guard against using data, big or small, in ways that reproduce racialized barriers to opportunity.

Centuries of institutional and structural discrimination have created, and continue to replicate, devastating inequality in our communities and our workplaces. While public commitments to racial justice are encouraging, the real work of changing unjust policies, creating equitable opportunities, and fixing broken systems must continue. Employers have the tools to do so. As one executive explained, “It is amazing what one leader on the right side of equity can do to impact a business.”