MIT Sloan—Ideas Made to Matter: A new barrier to diverse hiring in tech

Apr 13, 2021 | News

Credit: Natanael Ginting / Shutterstock

 

READ ARTICLE ORIGINALLY PUBLISHED AT MITSLOAN.MIT.EDU

 

When life insurance was introduced to the United States in the early 1800s, people were horrified at the idea of putting a dollar price on a human life.

So insurance companies reframed it. No longer was life insurance a way to collect on someone’s death. Instead it offered a comfort to grieving families knowing they’d be financially supported after the loss of a loved one.

Life insurance is just one commodity that started as a taboo but was eventually adopted by U.S. consumers as an appropriate, and in some cases necessary, service.

Consider compensating organ donors: Even this is no longer always viewed as a “repugnant market” — an area of commerce that features transactions people find distasteful.

But one repugnant market that continues to be a challenge is companies’ use of recruiting and labor-matching platforms to help them hire a racially diverse workforce, writes Summer Jackson, a PhD candidate in MIT Sloan’s economic sociology program. For example, a platform may use artificial intelligence to develop a more diverse pool of candidates to interview or to strip resumes of information that identifies a candidate’s race.

In a 20-month study of a “fast-growing technology company” she calls ShopCo, Jackson found that managers looking to diversify the company’s tech workers chose not to work with recruiting and labor-matching platforms they perceived as exploiting and objectifying non-white candidates. But in doing so, they may have unintentionally sabotaged their efforts at diverse hiring.

Jackson’s findings are detailed in the forthcoming paper “(Not) paying for diversity: Repugnance and failure to choose labor market platforms that facilitate hiring racial minorities into technical positions.” She will join the faculty of Harvard Business School after graduation this spring.

When hiring generally (and when the candidates were most commonly white), managers at ShopCo used platforms that advertised benefits like how quickly they could pull together a large quantity of candidates using salary and career growth discussions. However, when filling technical roles with racially diverse candidates, ShopCo’s leaders worried that platforms promoting things like speed and compensation were exploiting those diverse candidates, while also raising what Jackson called the “specter of affirmative action” — that is, claims that affirmative action unfairly promotes racial minorities.

So instead, ShopCo worked with platforms that emphasized individuality, ethics, and authenticity, among other attributes. The result: it took longer for ShopCo to hire. And because the candidates it ultimately recruited through the latter set of platforms were more junior, they were less skilled and more likely to be laid off when the pandemic hit.

 

‘Well-intentioned’ but lacking diversity

Jackson spent September 2018 to April 2020 embedded with ShopCo, a “well-intentioned” company that was “constantly hiring technical talent, with many software engineering roles always open and accepting inbound applications.”

But only 8% of its technical employees were racial minorities, and across the whole company, only 18% of its employees were racial minorities. After a 2018 employee diversity survey revealed comments asking what the company was doing to increase its diversity, ShopCo leadership reevaluated the company’s hiring practices for technical roles.

In her dissertation presentation, Jackson said that ShopCo leaders wanted to hire more diverse technical employees. But they worried that using a recruiting platform in which most of the candidates were racial minorities could make the company look like it was objectifying and exploiting those candidates — using an algorithm to “dehumanize” them into an assigned demographic, to the financial benefit of the recruiter.

And though they did not want to have a homogenous team, ShopCo leaders didn’t want to come across as “cherry-picking candidates to satisfy diversity,” Jackson writes.

During her time with ShopCo, Jackson observed 13 hiring platforms and recruiters as they presented their recruiting processes to ShopCo’s hiring managers and leadership. These companies fell into one of two categories: the majority of hiring candidates were white, or the majority of hiring candidates were racial minorities. And within those two categories, these companies either used what Jackson calls a market-exchange relational package or a developmental relational package for recruiting candidates.

Market-exchange relational packages emphasized time, quantity, efficiency, and opportunity. Recruiting companies that used this type of relational package touted their “data-driven” hiring process, use of keyword searches to quickly bring in large numbers of candidate resumes and profiles, and practice of discussing salaries early to cull candidates who didn’t match a company’s pay scale.

ShopCo found this relational package acceptable to use when engaging with white candidates, Jackson writes, but it was a “relational mismatch” for engaging with racial minority candidates because of ShopCo managers’ concerns about objectification, exploitation, and the specter of affirmative action. And some ShopCo leaders personally felt uncomfortable using this package. For example, one manager at ShopCo found the idea of “paying a membership fee to get [racial minority] candidates” distasteful.

Developmental relational packages used what Jackson calls a “triple-bottom-reward” system, which emphasized individuality, ethics, authenticity, and decisions that are good for profits, employees, and society. This type of recruiter emphasized the “development” of racial minority candidates, individual screening and outreach, and sourcing from diverse networks. One recruiting company told ShopCo leaders it built connections “with historically Black colleges and universities, and the National Black Society of Engineers.”

ShopCo found this relational package acceptable to use when engaging with both white and racial minority candidates.

But while the developmental relationship packages might have alleviated ShopCo’s worries about objectification, exploitation, and the specter of affirmative action, Jackson said, the decision to use this type of recruiting process had unintended consequences. This type of recruitment:

  • Ate up hiring time. It took more than 18 months for ShopCo’s technology department to increase its diversity from 8% to 10%.
  • Delayed access to skill sets. While the company did diversify its entry-level engineering roles thanks to the recruitment platforms that promised to expand the pool of underrepresented job candidates, diversity at the senior level didn’t happen as quickly, Jackson said. ShopCo also ran out of work for incoming entry-level roles, and had to turn its focus to hiring senior-level candidates.
  • Created instability in diversity and inclusion efforts. When the COVID-19 pandemic hit, ShopCo cut its workforce and laid off many junior employees, many of whom had been newly hired through recruiters and could have later been promoted to senior-level roles.

 

New questions and next steps

Jackson’s research illustrates a major challenge in diversity hiring goals: How can managers hire using criteria and methods they are comfortable with, while also ensuring they succeed in diversifying their technical workforce?

“Since Jackson is the first to highlight this problem, we do not know yet how to solve it,” saidan MIT Sloan professor of business administration, and advisor to Jackson.

Past research from Alexandra Kalev and Frank Dobbin showed that when employees are given diversity training, problems like unconscious bias can arise and be counterproductive to equity efforts, Kellogg said, especially among white men who might feel their jobs are at risk if they don’t follow the new rules. A way to address this, according to Kalev and Dobbin’s research, might be to engage managers in supporting diversity initiatives within their departments, versus training the employees themselves.

“One research study I’d like to see conducted is whether telling HR managers the results of Jackson’s study, rather than telling HR recruiters themselves, could help,” said Kellogg, who focuses on organizational and institutional change.

Particularly, Kellogg said she’d like to see HR managers suggest to their recruiters that they use market-exchange relational packages to engage with minority candidates the way they do with white candidates — and start on a smaller scale or pilot.

“If recruiters actually use these platforms, and get positive feedback from racial minority candidates while doing so, recruiters may become more comfortable choosing to implement the use of these platforms on a larger scale,” Kellogg said.

In an interview about her research, Jackson said choosing whether to use a market-exchange or developmental package depends on a company’s hiring needs. Hiring platforms with a market-exchange relational package tended to have more mid- to senior-level talent, versus the entry- or junior-level candidates on developmental relational platforms.

In ShopCo’s case, Jackson said, “they were selecting hiring platforms with a developmental relational package for the ‘right’ reasons. But an unforeseen consequence of these selection choices was that their hiring did not always match their actual hiring needs.”

Well-intentioned firms like ShopCo are likely to face repugnant market concerns when thinking about how to diversify their workforces, Jackson said, and firm leaders should remain concerned about employees’ experiences once they’re hired.

By doing that, Jackson said, “company leaders are going to be just as sensitive to whether there’s a revolving door or high turnover of racial minorities, or tokenism, racism, or isolation. And so thinking about these concerns while you’re hiring can make you a better employer to these candidates.”

X