Workforce Reduction Equity

The Opportunity

Organizations that stress-test the selection criteria behind a workforce reduction before executing it reduce legal exposure, protect hard-won representation gains, and avoid the kind of headline risk that follows when layoff patterns become public. A brief review before a reduction costs far less than the reputational and legal consequences of one that goes wrong.

The Business Case

In the 2022 and 2023 tech layoffs, women made up 33 percent of FAANG employees but 44 percent of those laid off.¹ Women in tech are 65 percent more likely to be laid off than men.² In federal workforce reductions in 2025, women comprised between 52 and 64 percent of those cut in the five most heavily targeted departments, despite representing 46 percent of the total workforce.³ These patterns create legal exposure and can erase years of representation progress in a single quarter.

The practices that produce disproportionate impact often look neutral, and it rarely happens through a rule that names gender. Last-in-first-out policies affect women more because women were historically excluded from hiring longer and carry shorter average tenure. Cuts to support functions affect women more because support functions are where women are concentrated. Eliminating roles designated non-essential affects women more because women's work was historically undervalued. The problem was never an explicit gender rule. It was a neutral-sounding rule with a predictable gender result that nobody checked in advance.

That is why this standard is built around the criteria, not the headcount. Where a review surfaces a disparity, the response is never to swap individuals by gender. It is to re-examine the criteria themselves. Auditing the rule is lawful, defensible, and gets at the actual source of the disparity. Adjusting individual outcomes by gender is neither, and it exposes the very people it is meant to protect.

The AI Dimension

AI-driven workforce changes belong in this pillar because they carry the same risk at greater scale and faster speed. In the United States, 79 percent of employed women hold positions categorized as high risk for automation, compared to 58 percent of men.⁴ The ILO found that 29 percent of female-dominated occupations are exposed to generative AI, compared to 16 percent of male-dominated ones.⁵ Of the 6.1 million workers most likely to be displaced by AI with the fewest pathways to adapt, 86 percent are women.⁶

The roles at highest risk, administrative assistants, legal secretaries, medical secretaries, payroll clerks, and receptionists, are between 89 and 96 percent female.⁷ Organizations making AI-driven workforce changes without reviewing the gender impact of which functions they chose to automate, without offering upskilling pathways, and without accounting for the institutional knowledge being lost are making an equity decision by default rather than by design.

What a Review Looks Like

A pre-reduction adverse-impact review, ideally conducted under legal privilege, asks three straightforward questions before any notice goes out.

First, do the selection criteria fall harder on women than on the workforce overall? If so, the criteria, not the individuals, get re-examined.

Second, is each criterion genuinely job-related and applied consistently? A standard that cannot be justified on the merits is removed or corrected before the reduction proceeds.

Third, are upskilling, transition support, and severance distributed equitably? The off-ramp deserves the same scrutiny as the decision itself.

These questions do not require a dedicated analyst or a lengthy process. They require someone to look at the criteria before the notices go out rather than after.

A Note on Company Size

A small company letting two people go does not need a formal audit. It needs one honest question, asked before the decision is final: are the criteria we are using fair, job-related, and applied the same way to everyone, and is the result a coincidence or a pattern? Asking it early, and writing down the answer, costs nothing. It is also the single best protection both for the people affected and for the company itself.

What Good Looks Like

Good represents accessible baseline practices. Better reflects more intentional investment. Best describes what the most forward-thinking companies are doing right now.

Questions Worth Asking

• Are the criteria we use to decide who stays genuinely job-related, and are they applied the same way to everyone?
• In our last workforce reduction, did anyone review how our selection criteria fell across the workforce before the notices went out?
• Are we planning any AI-driven role eliminations, and have we looked at the gender distribution of the roles at risk?
• Do employees in roles with high automation exposure know that, and do they have access to upskilling resources?
• Are our severance and transition packages designed with the same care for all affected employees?