Severance Under Duress: The Quiet Economics of Power
The document arrived while I was still processing the words "your position has been eliminated." A severance agreement. Pages of legal language drafted by attorneys I would never meet, reviewed by executives who would never read it, presented to me by an HR representative whose role was to facilitate my signature, not to ensure I understood what I was signing.
I had seven days to decide. Seven days to find an attorney, review the agreement, assess my options, and make a binding legal decision about whether to release all claims—discrimination, retaliation, wrongful termination, everything—in exchange for a payment that would cover approximately two months of expenses. Seven days during which my income had disappeared, my health insurance was counting down, and the financial pressure of immediate unemployment was operating as its own kind of persuasion.
This is how severance works in corporate America. It looks like a mutual agreement between two parties. It functions as a transaction in which one party has all the leverage and the other has a ticking clock.
The Architecture of Agreement
Severance agreements are legal documents, and they're designed by the party that can afford to design them. The company has legal counsel on retainer—attorneys who specialize in employment law, who draft these agreements regularly, who understand exactly what each clause accomplishes and what the company gains from each provision. The employee has whatever legal resources they can access on short notice while simultaneously managing the financial and emotional crisis of sudden unemployment.
The asymmetry is baked into every element of the process. The company determines the severance amount. The company sets the signing window. The company drafts the terms. The company decides what you're releasing, what you're agreeing to, and what consequences you'll face for non-compliance. The employee's role in this "negotiation" is to accept, reject, or attempt to modify terms from a position of zero leverage. The Economic Policy Institute has documented what this looks like in practice: workers at the point of termination have effectively no bargaining power because the employer controls the only thing the worker needs—money to survive the transition.
The seven-day window PPG initially offered is instructive. Seven days is long enough to create the appearance of informed consent—you had time to review it, didn't you?—and short enough to prevent genuine informed decision-making. Finding an employment attorney in seven days, while managing the emotional shock of termination and the immediate financial consequences of lost income, is not a meaningful opportunity to evaluate your options. It's a pressure mechanism dressed in the language of reasonableness.
I negotiated an extension to thirty days, which improved the situation marginally. But the extension itself required advocacy—required knowing that extensions were possible, having the composure to request one, and having the leverage (or the appearance of leverage) to make the request stick. Many terminated employees don't know they can ask. Many lack the emotional bandwidth to push back on anything during the first weeks of unemployment. The system's default setting—sign quickly or lose the offer—serves the company's interest in rapid resolution while exploiting the worker's interest in immediate financial relief.
The information asymmetry compounds the time pressure. The company knows what claims you might have. The company knows the legal landscape. The company has already assessed its exposure and calibrated the severance offer accordingly—just enough to make signing seem rational, not enough to reflect the actual value of what you're releasing. You, the employee, are working with incomplete information under time pressure in a state of emotional distress, trying to evaluate a legal document that was specifically designed by the other party to protect the other party's interests. The word "agreement" implies two parties reaching a shared understanding. What actually happens is one party presenting terms and the other party deciding whether they can afford to say no.
What makes this particularly insidious is the framing. Severance is presented as a benefit—something the company provides out of goodwill, a generous gesture during a difficult transition. The reality is more transactional. The company is purchasing something specific: the release of legal claims and the silence of the departing employee. The severance payment isn't charity. It's the price of ensuring that whatever happened during your employment stays buried under a non-disparagement clause and a comprehensive release of claims.
What You're Actually Signing
The release of claims is the centerpiece of most severance agreements, and its scope is deliberately comprehensive. You're not just agreeing that you won't sue for wrongful termination. You're releasing claims you might not even know you have. Discrimination claims. Retaliation claims. Claims under state and federal employment law. Claims that might become apparent only in retrospect, when you have the emotional distance and information access to understand what actually happened and why.
This is particularly significant when the termination follows a pattern that raises legal questions—as mine did. Disclosing ADHD in February 2024, being terminated in April 2024, and then being asked to release all discrimination and retaliation claims in exchange for a few weeks of salary creates a specific kind of bind. You're signing away the legal remedies that exist precisely for the situation you're in, at the moment when you're least equipped to evaluate whether you should.
The timing deserves scrutiny. When the release of claims is presented simultaneously with the termination, the employee hasn't had time to understand what happened, let alone assess whether it constitutes actionable discrimination or retaliation. The full picture—the pattern of events, the organizational context, the documentation that might support or undermine a claim—only becomes clear with time and distance. But the signing window doesn't allow for time and distance. It demands a decision now, when understanding is minimal and vulnerability is maximum. The company gets the release before the employee fully comprehends what they're releasing.
The non-disparagement clause is the other critical provision, and its function extends beyond legal protection. Non-disparagement provisions don't just prevent you from saying false things about the company. They create a chilling effect on any public discussion of your experience. Even factual statements—accurate descriptions of what happened, honest accounts of workplace conditions—carry the risk of being characterized as "disparaging" by a company with the legal resources to make that characterization stick. The practical effect is silence: not because you've agreed that nothing wrong happened, but because the cost of being wrong about what counts as disparagement exceeds the benefit of speaking.
The NLRB has placed some limits on non-disparagement provisions in severance agreements, recognizing that overly broad clauses can infringe on workers' rights to discuss workplace conditions. But the NLRB's intervention addresses the legal boundary without addressing the power dynamic. A terminated employee deciding whether to discuss their experience is not weighing legal doctrine. They're weighing practical consequences: Can I afford to be sued? Can I risk the company enforcing this clause? Is the potential benefit of speaking worth the potential cost of the company's legal department deciding I've crossed a line?
For most people, the answer is no. The clause works not because it's legally unassailable but because it's practically intimidating. And the company knows this. That's the point.
The Mathematics of Dignity
I want to make the power asymmetry concrete, because abstraction lets institutions hide behind scale.
My severance was approximately $11,261. PPG's CEO received total compensation of $11,784,190 in the same fiscal year. That's a ratio of roughly 1 to 1,047. For every dollar the company offered me to release all claims and remain silent, its CEO received over a thousand dollars in compensation that included base salary, stock awards, and a bonus paid at 184% of target—during a year when the company's revenue declined 2.44%, earnings declined 12.13%, and the stock price declined 16%.
Let me make this more tangible. The CEO's compensation could pay 47,137 months of electric bills at $250 a month. That's 3,928 years of keeping the lights on. The top three executives' combined compensation of $19.2 million could heat 532 homes for a lifetime. My severance—the price of my silence, my legal claims, and my departure—would cover about 45 months of electric bills. Less than four years of a single utility.
These numbers aren't meant as an argument that executives shouldn't be paid well. They're meant to illustrate what "cost-cutting" actually looks like when you examine the distribution. PPG eliminated approximately 1,800 positions in its 2024 restructuring, recording $314 million in restructuring charges. The stated goal was $175 million in annual savings. The company framed this as "structural cost reduction"—the language of efficiency, of necessary optimization, of doing more with less.
But the company didn't cut costs. It transferred them. The expenses that disappeared from PPG's income statement—salaries, benefits, the overhead of employing 1,800 human beings—reappeared on 1,800 personal balance sheets as lost income, depleted savings, COBRA premiums, and the cascading financial consequences of sudden unemployment. The "savings" weren't savings. They were cost transfers from an entity that could absorb them to individuals who couldn't.
Meanwhile, during the same quarter that PPG recorded $314 million in restructuring charges—the quarter that eliminated those 1,800 positions—the company repurchased $250 million of its own stock. For the full year, PPG returned $1.4 billion to shareholders through dividends and buybacks. The money existed. It was simply directed toward people who already had the most rather than people who had just lost everything.
The accounting treatment tells its own story. Those 1,800 eliminated positions generated zero goodwill impairment. PPG carries approximately $5.7 billion in goodwill on its balance sheet—the premium paid for acquired companies, representing, in theory, the value of their people, relationships, institutional knowledge, and organizational capability. Eliminating 4% of the workforce didn't impair a single dollar of that goodwill. Under GAAP, firing people actually improves the discounted cash flow models that determine whether goodwill is impaired, because the projected savings increase future earnings. The accounting framework literally values the company as more capable after removing 1,800 people than it was with them.
I'll explore this paradox in more detail in Essay 17. But in the context of severance, it reveals something about how the system conceptualizes human contribution. The accounting says the institutional knowledge, the relationships, the capability those people carried—all of it was worth zero. Less than zero: the company is "better" without it. The severance agreement operationalizes that same logic at the individual level. Your claims, your story, your experience—all purchasable for a few thousand dollars. The system is internally consistent. It just happens to be internally monstrous.
A 10% reduction in the top three executives' compensation—approximately $1.9 million—could have preserved roughly 168 positions. Not all 1,800. But 168 people who would have kept their income, their insurance, their professional identity, and their dignity. The company chose not to make that trade. And the severance agreements that followed ensured that the people who bore the cost of that choice would remain silent about it.
Parallel Coercions
I was navigating the severance agreement during a period when another legal process was also demanding my attention: divorce proceedings. The parallel wasn't just unfortunate timing. It revealed something structural about how economic coercion operates when you're already vulnerable.
Both processes required legal expertise I couldn't easily afford. Both involved signing documents drafted by attorneys representing the other party's interests. Both created pressure to resolve quickly because delay carried its own costs. And both were operating on the assumption that I could make informed, rational, legally binding decisions while managing the emotional and financial fallout of losing my marriage and my job in the same period.
This is where class shows up in concrete ways. People with savings, with family financial support, with the ability to absorb months of unemployment without immediate crisis—those people can take the thirty days. They can hire an attorney. They can evaluate whether releasing all claims is worth the severance amount or whether the claims themselves might be worth pursuing. People living closer to the margin—people for whom two months of expenses is the difference between keeping the apartment and not—don't have that luxury. The "choice" to sign is constrained by circumstances that the agreement itself doesn't acknowledge.
The company knows this. The severance amount is calibrated to be meaningful enough that rejecting it feels reckless and small enough that accepting it feels like the only rational option. It's not designed to be fair. It's designed to be efficient—to resolve the separation quickly, cleanly, and with the company's legal exposure minimized. The human being on the other side of the agreement is a variable to be managed, not a person to be treated equitably.
I thought about this calibration when I ran the numbers later—when I had the distance to look at what I'd signed and what I'd released and what the company had paid its leadership during the same period. My severance was roughly two months of take-home pay. The CEO's bonus alone—just the bonus, not the base salary or the stock awards—was paid at 184% of target. During a year of declining revenue, declining earnings, and declining stock price. The message embedded in those numbers isn't subtle: executive underperformance is rewarded with above-target bonuses. Employee presence during the same period is rewarded with termination and a coerced agreement.
Financial decision-making research is clear on what happens when people make binding commitments under acute stress: they prioritize immediate relief over long-term interest. They accept unfavorable terms because the present pain of uncertainty exceeds the potential future benefit of negotiation. They sign. And the organizations that design severance agreements understand this research even if they don't cite it. The seven-day window, the comprehensive release, the non-disparagement clause—these aren't accidents of legal drafting. They're products of a system optimized to extract maximum concession from people at their point of minimum leverage.
What Humane Severance Would Look Like
None of what I'm describing is inevitable. Severance agreements could be structured differently—not as instruments of corporate protection but as genuine transitions that respect the dignity of the departing employee. The reforms I'm proposing aren't radical. They're the kind of basic fairness that any system claiming to treat people with dignity should already be practicing.
Longer review periods would be a start. Not seven days. Not even thirty. A review period proportional to the significance of what's being signed—sixty or ninety days—that gives employees genuine time to process the termination, consult with counsel, and make informed decisions. The argument that companies need rapid resolution doesn't survive contact with the reality that companies already tolerate far longer timelines for transactions of comparable legal significance. If a company can spend six months negotiating a vendor contract worth less than the aggregate cost of 1,800 severance agreements, it can give employees more than a week to evaluate whether they're surrendering claims worth pursuing.
Employer-funded independent legal consultation would address the access inequity. Not the company's own counsel "explaining" the agreement—that's a conflict of interest, not a solution. Independent employment attorneys, funded by the employer, whose role is to represent the departing employee's interests in reviewing the terms. If the agreement is fair, independent review shouldn't threaten the company. If the company resists independent review, that resistance is informative.
Proportional severance—packages that reflect tenure, role, and the circumstances of departure rather than the minimum the company calculates it needs to spend to purchase silence—would acknowledge that the relationship being severed had value on both sides. The current model treats severance as an expense to be minimized. A humane model would treat it as an obligation to be honored. A person who worked at a company for four and a half years, built automation systems, improved processes, and contributed to the organization's capability has earned more than two months of pay and a gag order.
And non-disparagement provisions should be mutual and specific rather than unilateral and vague. If the company wants the employee's silence, the employee should receive the company's silence in return—no negative references, no internal commentary about the circumstances of departure, no unofficial signals to future employers. If the company wants to define disparagement, the definition should be narrow enough that factual statements about workplace conditions aren't chilled. The current model—where the company can say whatever it wants about why you left while you can say nothing about why you were pushed—is a one-sided gag order that serves institutional narrative control, not legitimate legal protection.
These reforms wouldn't eliminate the power asymmetry. Nothing short of structural economic change would do that. But they would reduce the asymmetry enough that the word "agreement" would mean something closer to what it implies: two parties, both informed, both with genuine options, reaching terms that reflect shared understanding rather than unilateral imposition.
The Purchase of Silence
Every severance agreement that includes a non-disparagement clause is, at its core, a transaction in which the company purchases the employee's story. Not just their legal claims. Their narrative. Their ability to say, publicly and without legal risk, "this is what happened to me."
For organizations that have nothing to hide, this purchase is unnecessary. If the termination was fair, the workplace conditions were appropriate, and the process was handled with integrity, the departing employee's account of their experience wouldn't threaten the company. The insistence on silence—built into the legal architecture of nearly every severance agreement—is itself an admission that the company anticipates accounts it would prefer not to address publicly.
I signed the agreement. Most people do. Not because the terms were fair—they weren't—but because the alternative was worse. Rejecting the severance would have meant losing the only financial bridge between employment and whatever came next. It would have meant entering an already difficult job search without the cushion, however thin, that the payment provided. It would have meant choosing principle over pragmatism at a moment when pragmatism was the only thing keeping the bills paid.
And this is the mechanism's elegance, from the company's perspective. It doesn't require malice. It doesn't require conscious exploitation. It just requires a system where one party designs the terms, sets the timeline, and controls the resources—and another party decides whether they can afford to object. The coercion is structural, not personal. The HR representative who presented the agreement was probably a decent person doing their job. The attorneys who drafted it were probably competent professionals serving their client. Nobody in the process needed to be cruel. The cruelty is in the architecture.
This is the quiet economics of power. The company doesn't need to threaten you. It doesn't need to coerce you in any way that would be legally recognizable as coercion. It just needs to structure the choice so that one option—silence and a check—is marginally less devastating than the alternative. The power isn't in the document. It's in the conditions that make the document feel like the only rational choice.
I think about the CEO's $11.78 million and my $11,261 and I don't feel bitterness. I feel clarity. The numbers tell a story about what the organization values, who it protects, and how it distributes the consequences of its decisions. The executives who oversaw a year of declining revenue, declining earnings, and declining stock price received bonuses at 175-184% of target. The employees who bore the consequences of those declines received severance agreements that purchased their silence for less than the cost of a used car.
The company didn't cut costs. It redistributed them—from the people who had the most to the people who had the least, and then it bought their silence so the redistribution would remain invisible.
That's not efficiency. That's extraction. And the severance agreement is the receipt.