Equitable Downsizing at AutoNow?
AutoNow, a 30,598-employee auto parts manufacturer in metro Detroit, has been acquired by private equity firm Tramonto Capital Partners. AutoNow's workforce is non-unionized. Strategy Director Zahra Nasser must reduce annual labor costs by $200 million through workforce reductions. Seven downsizing options have been proposed — each using a different criterion that makes no explicit reference to employee identity.
"Before any plan is finalized, it is required that you conduct a formal adverse impact analysis. A group is considered adversely impacted if it is meaningfully over-represented among those targeted for reduction relative to its share of the total workforce. This is both a legal obligation under Title VII and EEOC guidelines and a strategic necessity."
How to use this tool: Click any option to see its adverse impact profile. Expand details to read each VP's pitch from the case.
Select a Downsizing Option
Reading the charts: Positive values (maize) = group is over-represented among those laid off (harm). Negative values (blue) = under-represented (protection). Near zero (gray) = proportional.
Adverse Impact by Racialized Identity
Adverse Impact by Gender
Workforce Share vs. Downsized Pool
| Group | Workforce % | Downsized % | Impact (pp) | Status |
|---|
Status flags any group more than 2 percentage points off its workforce share — a descriptive marker, not the legal test. The EEOC 4/5 Rule card above asks a different question: whether any group's retention rate falls below 80% of the highest-retained group's rate. A group can be flagged here and the option can still pass the 4/5 Rule — both can be true at once.
All Options Compared — By Racialized Identity
All Options Compared — By Gender
Legal & Analytical Framework
What It Is: Adverse (Disparate) Impact
Adverse impact occurs when a facially neutral employment practice — one that does not explicitly reference race, gender, or other protected characteristics — nevertheless produces a disproportionate negative effect on a protected group. The concept is grounded in Title VII of the Civil Rights Act of 1964, which prohibits employment discrimination even when a policy appears neutral on its face. Under this framework, intent is not required — it is the outcome that matters.
How It's Assessed: Measurement & the EEOC 4/5ths Rule
The EEOC's Uniform Guidelines on Employee Selection Procedures (29 C.F.R. § 1607) require employers to evaluate whether their selection criteria produce demographic disparities. In this simulation, adverse impact is measured as the percentage-point difference between a group's share of the downsized pool and its share of the total workforce. A positive value means the group is over-represented among those laid off.
The primary screening tool is the 4/5ths (80%) rule: if the selection rate for any protected group falls below 80% of the rate for the highest-selected group, that is evidence of adverse impact. In layoff contexts, the relevant "selection rate" is the rate at which each group is retained. If any group's retention rate falls below 80% of the highest group's rate, the criterion is flagged.
How It's Enforced: Griggs v. Duke Power Co. (1971)
The legal standard for disparate impact was established in Griggs v. Duke Power Co. (1971), in which the Supreme Court struck down aptitude tests and diploma requirements that had the effect of excluding Black applicants from higher-paying positions — despite the absence of discriminatory intent. The Court held that Title VII "proscribes not only overt discrimination but also practices that are fair in form, but discriminatory in operation." An employer using a practice with disparate impact bears the burden of demonstrating that it is justified by business necessity. Consequences, not intentions, determine legality.
Going Further: Statistical vs. Practical Significance (optional)
Everything above tests practical significance — whether a gap is large enough, in percentage points, to matter under the 4/5ths rule. A separate question is statistical significance: whether an observed gap between two groups' selection rates is bigger than random chance would produce, given the size of each group. The two questions can point in different directions.
A standard test for this is a two-proportion z-test, comparing one group's selection rate to everyone else's. At AutoNow's scale — thousands of employees per group — even a modest percentage-point gap can be statistically undeniable. Under Option 5 (Geographic), Hispanic employees are downsized at 16.4%, versus 6.8% for everyone else; that gap produces a z-score of about 21.7, far beyond any conventional threshold for chance. Under Option 3 (LIFO), Black employees are downsized at 7.0%, versus 7.6% for everyone else — a gap in the opposite direction, and one small enough (z ≈ -1.6) that it would not be considered statistically significant at conventional thresholds.
Large sample sizes make statistical significance easy to reach and easy to over-read. A statistically significant gap is not automatically a legal violation — Option 5 clears the 4/5ths rule regardless of its z-score — and a gap that fails to reach significance is not automatically fair. Statistical significance, the 4/5ths rule, and the question this case asks are three separate tests. Clearing one says nothing about the other two.
The Stakes
Why every option harms someone different. AutoNow's workforce is not demographically uniform across departments, roles, sites, or performance tiers. Historical hiring, assignment, and evaluation patterns have sorted employees along identity lines. When any identity-neutral criterion is applied to this structurally unequal workforce, it inherits the demographic composition of whatever organizational feature it targets. The criterion is neutral; the structure is not; the outcome replicates the structure. Every option reaches the $200M target. Every option produces measurable adverse impact on at least one protected group — each time a different group. There is no safe answer.
Lawful does not mean equitable. Importantly, none of these options violate the EEOC's four-fifths rule. Every option is legally defensible. But legality is not the same as equity, and compliance is not the same as conscience. The data will tell you what your decision does — to whom, by how much, and through which mechanism. It will not tell you what to do. That requires knowing what you believe, what your organization stands for, and which harms you are willing to justify.
The human cost is not hypothetical. A $200 million headcount reduction affects thousands of real people — their families, their health, their communities. The research is sobering: job displacement increases annual mortality risk by approximately 15% over a 20-year period (Sullivan & von Wachter, 2009), with effects persisting 30 years even in countries with universal healthcare (Zellers et al., 2025). Targeted layoffs — where individuals are selected rather than a facility closing — produce distinct physiological harm: elevated diabetes and cardiac risk markers, and increased deaths from suicide, violence, and substance abuse (Michaud et al., 2016). By one estimate, roughly 1 additional death per 111 employees laid off can be attributed to displacement over a 20-year window.
This is the landscape, not an anomaly. Large-scale workforce reductions are not receding — they are accelerating, driven by automation, AI adoption, offshoring, and cost-of-capital pressures. The scenario on this page is not a thought experiment from the past. It is the decision that leaders across industries are being asked to make right now, often with fewer good options than they would like and more at stake than a spreadsheet can capture.
The toll extends to those who decide. The harm of downsizing is not borne only by those who lose their jobs. The leaders who design, approve, and implement reductions carry a psychological and moral weight that persists long after the decision is made. Naming that reality — the inevitability of harm, the absence of a clean outcome, and the burden on those responsible — is not weakness. It is the starting point for making the decision well.
? What if Zahra flipped a coin?
What would happen if AutoNow selected employees for layoff at random — no criterion, no algorithm, no discretion? In expectation, a coin flip produces zero adverse impact on every group. Can your preferred option justify the disproportionate harm it causes relative to a process that causes none?
Works Cited
EEOC. (1978). Uniform Guidelines on Employee Selection Procedures. 29 C.F.R. § 1607. https://www.law.cornell.edu/cfr/text/29/part-1607
Griggs v. Duke Power Co., 401 U.S. 424 (1971). https://supreme.justia.com/cases/federal/us/401/424/
Kalev, A. (2014). How you downsize is who you downsize: Biased formalization, accountability, and managerial diversity. American Sociological Review, 79(1), 109–135. https://doi.org/10.1177/0003122413518553
Michaud, P.-C., Crimmins, E. M., & Hurd, M. D. (2016). The effect of job loss on health: Evidence from biomarkers. Labour Economics, 41, 194–203. https://doi.org/10.1016/j.labeco.2016.05.014
Sullivan, D., & von Wachter, T. (2009). Job displacement and mortality: An analysis using administrative data. Quarterly Journal of Economics, 124(3), 1265–1306. https://doi.org/10.1162/qjec.2009.124.3.1265
Zellers, S., Azzi, E., Latvala, A., Kaprio, J., & Maczulskij, T. (2025). Causally-informative analyses of the effect of job displacement on all-cause and specific-cause mortality. Social Science & Medicine, 369, 117711. https://doi.org/10.1016/j.socscimed.2025.117711