Nowhere Else to Go

In its 2023 annual report, Amazon told the SEC that its ambition to be the best employer on Earth rested on several commitments, among them diversity and inclusion. In the 2024 filing, that sentence had been cut down to “We strive to be Earth’s best employer.” The rest was gone. Across the S&P 500, only about a third used the term “DEI” in their most recent annual report, down from roughly ninety percent a year earlier.

I have spent years drafting and reviewing that kind of language on the transactional side of my practice. Companies almost never edit it casually. It changes when someone inside the building decides the commitment now costs more than it returns.

What happened next is also public. In 2025, the employment rate for Black women fell 1.4 percentage points to 55.7 percent, one of the sharpest single-year declines in twenty-five years, while Black men and white women fell by no more than half a point. The steepest losses were among Black women holding bachelor’s degrees, whose employment rate dropped 3.5 points, more than in any other education category, including women without degrees.

The causal question is open and I am not going to close it. Much of that loss ran through the public sector, where federal layoffs account for a large share. The Economic Policy Institute, whose analysis this is, said it could not yet tell whether these numbers reflect general labor-market conditions or a specific retreat from equity commitments.

One figure resists the ambiguity. Labor force participation among college-educated Black women fell 2.3 points in the same year. These are not, for the most part, women who lost one position and took another. They are women who stopped looking.

There are two ways to read a number like that. The first is that a group of workers became discouraged, which is a fact about them. The second is that the market for their skills closed, which is a fact about the market. Lawyers on both sides try these cases on the first reading, and I think that is wrong.

Consider a woman fired unlawfully who cannot find comparable work. Her recovery turns on mitigation. Her former employer will argue that she did not try hard enough, and the answer is nearly always a log of applications and interviews, as though she had to prove her own diligence.

In most jurisdictions she does not. Failure to mitigate is an affirmative defense, which means the employer must prove that substantially similar positions were actually available to her. Practice has drifted from doctrine. The stronger answer is to hold the defense to the burden it already carries and make it name the jobs she should have taken.

That is provable, and the proof is public. The Bureau of Labor Statistics publishes employment counts by occupation and metropolitan area, which tells a jury how many positions of a given kind exist within driving distance. When that number is small, and the few employers holding those positions share executives and outside counsel, a two-year gap in a résumé stops being a character flaw. It becomes arithmetic.

Economists call this monopsony. A monopolist is the only seller in a market. A monopsonist is the only buyer, and in a labor market the buyers are employers. Labor markets are far more concentrated than the markets for the goods those workers make, because a product can ship anywhere and a worker has to commute.

No court has to adopt that theory for it to matter in a courtroom. Concentration is not a competing story about the case. It is evidence inside the case the plaintiff already has. Make the defense name the jobs it says were open to her, then show the jury that the same few employers control all of them. That is not background color; it is proof the jobs were never open.

I learned to look at things this way long before I had vocabulary for it. I grew up in the Aspen Place projects in Passaic, New Jersey, one of three children raised by a single mother on public assistance, and I attended public schools from kindergarten through twelfth grade. My mother was very good at working the choices in front of her. What I registered at the time, without words for it, was how few of those choices there were.

I spent years on this problem without recognizing it as the same one. My clients were artists and founders, and my work was helping them own what they made. During the race records era of the 1920s and 1930s, the Black musicians who made the records that became the foundation of American popular music were barred from holding copyright in their own work. They generated enormous value and captured almost none of it, and the reason was not that they negotiated poorly. There was no second label prepared to offer better terms.

I started my own firm in 2019, after nearly a decade at other people’s, and it now does two kinds of work that look unrelated. On one side, I help founders and creators establish ownership before someone with more leverage takes an interest in what they built. On the other, I represent employees, and only employees, after an institution has used its leverage on them. Both sides ask one question at different moments, which is who ends up owning what gets made.

The commitments came out of those filings quietly, in a single season, in the passive voice, the way expensive obligations usually get removed. The promises disappeared from the page, and the jobs closed for the people those promises were meant to protect. The people hired to make those commitments real did not go as quietly. A number of them are now my clients.