1. Acme Inc. has 40,000 white employees and 10,000 employees of color on payroll. The statistic would be 20%, if Acme were hiring at a constant rate by the same demographics.
2. However, suppose Acme hired the bulk of its employees during its growth phase 10 years ago. Acme's hiring back then was proportional, but the population has changed. Now only 60% of applicants are white, compared to 80% back then.
3. Acme lays off 5,000 staff (at random), and hires 1,000 (proportionally.) So they've laid off 4,000 white people and 1,000 people of color. And they've hired 400 people of color and 600 white people.
I'm too lazy to do the math but I think that works out as hiring a negative % of white people, even though it's just representative of demographic shifts.