Overcoming Loss Aversion to Boost Workplace Diversity

• A new study explores how managers often hire new employees who closely resemble those they are replacing, leading to “demographic stickiness” and hindering diversity.
• Demographic stickiness can slow diversification, but also help maintain diversity once achieved. One-time interventions to improve diversity might have lasting impacts on group composition.
• Furthermore, this article provides concrete strategies, such as blind recruitment, bias training, and setting diversity goals, to overcome biases and promote a more inclusive hiring process.

Introduction

Ever wondered why your office still looks the same despite ongoing diversity efforts? Part of the answer lies in the psychological concept of loss aversion.

This is the main finding of a new study by Edward H. Chang of Harvard University and Erika L. Kirgios of the University of Chicago. Published in Management Science, the authors conduct 2 large quantitative studies and 4 experiments to establish the existence and uncover the dynamics of loss aversion in the hiring process.

Read on to learn more about their research into the “demographic stickiness” of the workplace, and how we can promote more diverse and inclusive workplaces in general.

Background

Loss aversion is a psychological idea introduced by Kahneman and Tversky in 1979. It means people tend to feel losses more deeply than gains of the same size. For example, losing $5 feels worse than gaining $5 feels good.

This tendency can affect many decisions, including hiring new employees to replace those who leave. In hiring, loss aversion can show up in two main ways, leading to “demographic stickiness”—the habit of hiring replacements who look like the person who left.

1. Impact Aversion

First, loss aversion might cause managers to prefer keeping things the same or making as few changes as possible. When they need to replace someone, they might choose a new hire who is very similar to the previous employee. This behavior is called “impact aversion.” It happens because the risks of change seem scarier than the potential benefits.

Studies show that when managers can’t keep everything the same, they still try to minimize changes. This might lead them to hire people who share key traits with the person who left. Since race and gender are easy to notice, these are often the main factors considered. People tend to think that someone who looks similar will act similarly and fit into the group just like the old employee did.

2. Diversity Loss Aversion

Second, loss aversion can affect hiring through “diversity loss aversion.” Diversity, having different kinds of people in a group, is valuable. It can improve how a group works and looks to others. If managers see losing diversity as a big problem, they might be more likely to hire people from underrepresented groups to replace similar departing employees.

For example, if a non-White employee leaves, managers might focus on hiring another non-White person to avoid losing diversity. They want to keep the group’s diversity level and avoid negative reactions to becoming less diverse.

Comparing Impact Aversion and Diversity Loss Aversion

While both impact aversion and diversity loss aversion lead to similar hiring behaviors, they focus on different things:

• Impact aversion is about keeping the new hire as similar as possible to the old one, especially in terms of race and gender.
• Diversity loss aversion is about keeping the group’s overall diversity, which might lead to broader demographic considerations.

For instance, in a mostly White male environment, a manager focused on avoiding diversity loss might hire any non-White male, such as a woman or a person of color, to replace a non-White male who left. On the other hand, impact aversion would lead the manager to look for someone who closely matches the specific demographics of the departing employee, like hiring another Asian person to replace an Asian employee.

A Narrow Approach to Diversity

Both theories highlight a limited approach to diversity in hiring decisions. Instead of aiming for a certain level of diversity, managers may focus only on the demographic identity of the person who left. This narrow view can lead to decisions that maintain immediate demographic continuity rather than considering the group’s long-term diversity needs. For example, suppose a woman leaves a group of mostly men. In that case, the group might be more inclined to hire another woman simply to replace the lost demographic characteristic, rather than considering the overall diversity needs of the group.

Understanding how loss aversion affects hiring can help organizations create better strategies for building diverse and inclusive teams. By recognizing these tendencies, they can make more thoughtful and long-term decisions about diversity in their hiring practices.

Methods

This research investigates whether the demographic identity of outgoing members in organizational settings influences the selection of their replacements. The study spans two field settings and four controlled experiments, covering various contexts including U.S. federal judgeships and corporate board appointments.

Here’s a breakdown of the studies:

1. Analyzes U.S. federal judge appointments from 1945 to 2020.
2. Examines appointments to S&P 1500 corporate boards between 2014 and 2019.
3. Conducts an experiment manipulating the race of a departing group member.
4. Investigates the indirect effects of demographic consistency desires.
5. Compares responses to replacing a White man versus an unspecified individual.
6. Highlights the focus on immediate demographic consistency.

Study 1: U.S. Federal Judges

• Source: Federal Judicial Center (FJC), the U.S. government’s judicial branch research and education agency.
• Period: Judges confirmed between 1945 and 2020.
• Sample: 2,163 judicial confirmations.
• Inclusion Criteria: Judges with a known predecessor to analyze demographic replacement patterns.
• Independent Variables: Gender and race of the predecessor judge.
• Dependent Variables: Gender of the new judge (binary: woman or not woman). Race of the new judge (binary: non-White or White).
• Control Variables: Political factors (e.g., nominating president’s party). Demographic makeup of the court’s jurisdiction. Number of women and racial minorities already on the court.
• Approach: Ordinary least squares (OLS) regressions with robust standard errors and clustering by president.
• Purpose: To determine if a judge’s demographic identity influences the selection of their replacement.

Study 2: S&P 1500 Corporate Board Directors

• Source: Institutional Shareholder Services (ISS).
• Period: Directors added to S&P 1500 boards from 2014 to 2019.
• Sample: 5,616 new director additions.
• Inclusion Criteria: Director additions following the departure of board members in the previous year.
• Independent Variables: Whether the board lost women or racial minorities in the preceding year.
• Dependent Variables: Gender of new directors (binary: woman or not woman). Race of new directors (binary: non-White or White).
• Control Variables: Firm-level variables (e.g., size, industry). Board size. Number of women and racial minorities already on the board.
• Approach: OLS regressions with robust standard errors and clustering by firm.
• Purpose: To assess demographic replacement patterns in corporate board settings.

Study 3: Experimental Manipulation of Race

• Source: Amazon’s Mechanical Turk.
• Sample: 600 participants (42.5% men).
• Design: Participants imagined managing a consulting firm’s group and had to replace either a departing White man or Black man.
• Task: Select a candidate from a pool with varying demographics but identical qualifications.
• Focus: Whether the race of the departing member influences the race of the replacement selected.

Study 4: Mediation Analysis for Gender

• Source: Amazon’s Mechanical Turk.
• Sample: 800 participants (47.2% men).
• Design: Participants chose replacements for a departing White woman or an unknown member.
• Task: Select a candidate from a pool with varying demographics but identical qualifications.
• Focus: Influence of known versus unknown predecessor gender on replacement selection.

Study 5: Comparison of White Man and Unspecified Departures

• Source: Amazon’s Mechanical Turk.
• Sample: 600 participants (52.8% men).
• Design: Participants replaced either a White man or an unspecified individual.
• Task: Select a candidate from a pool with varying demographics but identical qualifications.
• Focus: Test predictions of impact aversion versus diversity loss aversion.

Study 6: Highlighting Myopic Demographic Stickiness

• Source: Amazon’s Mechanical Turk.
• Sample: 900 participants (55.0% men).
• Design: Participants were either replacing a White man, White woman, or adding a new group member.
• Task: Select a candidate to join the group, considering consistent remaining group demographics.
• Focus: Myopic tendencies in maintaining demographic consistency.

In sum, the research employs a combination of historical data analysis and controlled experiments to reveal robust patterns and mechanisms driving these decisions.

Results

The results of the first two studies show clear patterns of “demographic stickiness,” where the gender and racial identity of the new appointees often mirror those of their predecessors.

Study 1

The study found that new judges’ gender was significantly influenced by the gender of their predecessors.

• When a man previously held a judicial seat, the new judge was a woman 16.15% of the time.
• In contrast, when a woman previously held the seat, a woman was chosen 36.60% of the time.
• The regression analysis confirmed this, showing a significant positive relationship (b = 0.089, p = 0.031) between the gender of the predecessor and the gender of the new judge.

Similarly, racial identity showed a significant pattern.

• When a White person held a seat, new judges were racial minorities 12.92% of the time, compared to 35.45% when the predecessor was a racial minority.
• The regression analysis indicated a significant positive effect (b = 0.117, p = 0.011) of the predecessor’s racial minority status on the likelihood of selecting a new racial minority judge.

The study also explored specific racial minority groups. It found that:

• A Black predecessor significantly increased the likelihood of selecting a Black new judge (b = 0.119, p = 0.003).
• Hispanic and Asian predecessors did not significantly influence the selection of Black new judges.
• Predecessors’ demographic characteristics did not significantly predict the selection of new judges from other racial minority groups.

Study 2

The patterns observed in judicial appointments were mirrored in corporate board appointments:

• When boards had lost at least one woman, new directors were women 35.31% of the time, compared to 25.98% when no women had left (p < 0.001).
• Similarly, new directors were racial minorities 24.58% of the time when a racial minority had left, compared to 16.85% otherwise (p < 0.001).

Study 3

The subsequent studies (Studies 3-6) further explored the concept of demographic stickiness in hiring decisions, focusing on controlled experimental settings and different organizational contexts.

Method: Participants were assigned to replace someone of a given demographic identity in a group. They chose new group members from a set of candidates.

• Participants were significantly more likely to select a Black man to replace a departing Black man (51.3%) than a departing White man (39.4%), (z = 2.94, p = 0.003).
• Participants were more likely to select a White man to replace a departing White man (17.5%) than a departing Black man (10.4%), (z = 2.52, p = 0.012).
• Selection rates for a White woman did not significantly differ across conditions (43.0% vs. 38.3%, z = 1.19, p = 0.232).

Conclusion: The results confirmed demographic stickiness, with a significant tendency to replace departing group members with individuals sharing the same demographic identity.

Study 4

Method: Participants replaced group members of known or unknown demographic identities and indicated their hiring preferences.

• Participants selected a woman 57.2% of the time when replacing a departing White woman, compared to 43.0% when the identity was unknown, (z = 4.031, p < 0.001).
• Selection rates for White men were lower when replacing a White woman (9.2%) than when the identity was unknown (15.2%), (z = 2.588, p = 0.010).
• Selection rates for Black men were lower when replacing a White woman (33.5%) compared to the unknown identity condition (41.7%), (z = 2.408, p = 0.016).

Conclusion: Participants’ decisions were influenced by the desire to minimize changes, showing a significant demographic stickiness effect.

Study 5

Method: Participants guessed the identity of an unknown departing group member and made hiring decisions.

• 81.5% of participants guessed that the departing member was a White man.
• This assumption could explain why hiring decisions did not vary significantly when the identity was unknown compared to when replacing a White man.

Conclusion:
The study suggests that the assumption of a default identity (White man) influences hiring decisions, potentially masking demographic stickiness effects in certain conditions.

Study 6

Method: Participants replaced either a departing White man or a departing White woman and selected new group members.

• Participants selected a White woman 45.5% of the time when replacing a White woman, compared to 36.6% when replacing a White man, (z = 2.22, p = 0.026).
• The rates of selecting a White woman did not significantly differ between the departing White man and no one departing conditions (37.5% vs. 36.6%, z = 0.244, p = 0.807).

Conclusion: Participants showed a tendency to replace departing group members with candidates of the same demographic identity, reinforcing the concept of demographic stickiness.

Summary

The studies consistently found evidence of demographic stickiness in hiring decisions. Decision-makers tend to select candidates who resemble the departing group members in terms of gender and race. Furthermore, the effect is specific to demographic categories, with significant patterns observed for both gender and racial identity.

This has several important implications:

• Demographic stickiness can slow the diversification process in organizations. Loss-averse hiring managers may prefer to hire White or male workers for the sake of racial or gender continuity.
• However, once diversity is achieved, it is likely to be maintained due to this stickiness. Departing female or non-White employees are more likely to be replaced by another female or non-White candidate.
• The prevalence of this stickiness implies that one-time interventions to improve diversity might have lasting impacts on group composition.

What Can You Do?

Addressing the problem of female and non-White representation in employment requires active and sustained efforts. Here are some concrete steps you can take to make a difference:

1. Implement Blind Recruitment Processes: Remove personal information such as names, photos, and other identifiers from applications to focus solely on candidates’ qualifications and skills.
2. Set Diversity Goals: Establish clear, measurable diversity targets for hiring, promotion, and retention. Regularly review and adjust these goals to ensure progress.
3. Provide Bias Training: Offer training for all employees, especially hiring managers, to recognize and mitigate unconscious biases in the recruitment and selection processes.
4. Develop Inclusive Job Descriptions: Use inclusive language in job postings to attract a broader range of applicants. Avoid jargon and gender-coded language that might discourage diverse candidates from applying.
5. Expand Recruitment Channels: Partner with organizations, colleges, and job boards that focus on underrepresented groups. Attend diversity-focused job fairs and networking events.
6. Create Mentorship Programs: Establish mentorship and sponsorship programs to support the career development of female and non-White employees. This helps build a pipeline of diverse talent for future leadership roles.
7. Regularly Review Hiring Practices: Continuously assess and refine your recruitment and hiring practices to ensure they are fair and inclusive. Use data to identify areas of improvement and track progress over time.
8. Encourage Employee Resource Groups: Support the formation of employee resource groups (ERGs) for underrepresented employees. These groups can provide networking, professional development, and a sense of community.
9. Promote Flexible Work Arrangements: Offer flexible working hours, remote work options, and other policies that can make your workplace more attractive to a diverse range of candidates.
10. Hold Leadership Accountable: Ensure that company leaders are committed to diversity and inclusion. Include diversity metrics in performance evaluations and hold leaders accountable for meeting diversity goals.

By taking these steps, you can help create a more diverse, equitable, and inclusive workplace. This not only benefits your organization but also contributes to broader social change by challenging and dismantling systemic biases in employment practices.

What else can be done to improve female and non-White representation in employment, especially at the highest levels? Share your ideas below.

Stay informed and inspired! Subscribe to our newsletter, “Society This Week,” and get a free copy of our exclusive report, “Peaceful, Powerful: 225 Nonviolent Tactics for Impactful Activism!”