4200 words, 25 minutes’ reading time

For part 1, go here. For part 2, go here. For part 3, go here. For part 4, go here.

5. Pettit on employment-at-will
I said above that I agree in a broad way with Pettit’s critique of employment-at will. Let me put it this way: I agree that employment-at-will, at least as currently practiced in the American labor market, is a highly problematic institution, one that frequently exemplifies domination for just the reasons Pettit gives. But while this may sound like substantial-enough agreement, I think it conceals some subtle but significant disagreement. In this post, I want to work through some of the agreement and the disagreement.

Though Pettit doesn’t put things quite this way, I think we can probably agree that two things make employment-at-will problematic. One is its asymmetric character. The other are the stakes involved when it’s invoked and exercised.

5.1 Asymmetric power    As a rule, employers tend to have more power than employees or would-be employees. For this reason, though employment-at-will has a formally equal or reciprocal character about it, this turns out to be highly misleading. It’s technically true that in employment-at-will, both parties have the same right to exit the employment relationship at will, without cause or notice. But given asymmetries of bargaining power, this formal truth tends to be overturned by material realities. For one thing, if A is more powerful than B, and B is operating under sufficiently adverse conditions, nothing prevents A from inducing B to sign a contract which gives A at-will powers but denies them to B. In other words, nothing about the employment-at-will doctrine requires the at-will power to be a matter of even formal reciprocity. The doctrine is after all fully consistent with unilateral waiver of the at-will power.

Even if formally reciprocal, the at-will arrangement need not be materially so. Take a case in which both parties, formally speaking, operate at-will. Even so, employers regularly demand that employees sign at-will agreements formally granting reciprocal at-will rights, but then stipulating that failure to give notice of termination on the employee’s part will result in the employer’s withholding pay, benefits, or favorable references to the erstwhile employee. Thus while no breach of contract is alleged, and no explicit limitation on the at-will power is asserted (just the contrary), abrupt quitting by the employee is de facto penalized as though it was a breach of contract. Meanwhile the employer pays lip service to a formally reciprocal at-will contract. Indeed, in a sense the employee is worse off in cases of this sort than in cases where a breach of contract is alleged, because in this case, the loss of expected benefits is automatic: no allegation of breach is required, much less a legal proceeding designed to prove breach. Breach is de facto assumed; the penalty follows immediately.

But even in cases where both sides have at-will powers free and clear (without any further contractual stipulations or penalties reserved for one party’s exit from the relationship), having it often benefits employers more than employees. Take a situation in which employers have more of a stake in firing employees than employees have in quitting. In such cases, it doesn’t matter how reciprocal the right is; the value of its possession ends up being non-symmetric. If A has a stake in easy terminations, but B wants job security, it doesn’t matter that both A and B have at-will powers. Having them gives A what he wants while denying it to B.

5.2 Asymmetric stakes.  The point can be driven home by reflecting on a somewhat (but not entirely) contrived example, which I’ll dub “Lost Health Coverage.”(1)

Lost Health Coverage. Suppose that A enjoys substantial asymmetric bargaining power relative to B, and employs B to perform a defined, determinate task. Suppose that as part of B’s compensation package, A offers B coverage for health insurance. Suppose that health insurance is otherwise hard to get, and that as far as the demand for labor is concerned, it’s a seller’s market: it’s easy to fire workers and if need be replace them, but hard for a given worker to find a new job once fired. Now suppose that despite B’s doing excellent work, A feels an inexplicable rancor for B. Suppose further that B falls seriously ill. A decides, purely out of rancor, and in the knowledge of B’s illness, to fire B without notice or cause. B loses his health insurance, and in the process, goes bankrupt. Having done so, B is unable to pay for crucial health care, and prematurely loses his life. In this case, an asymmetric power relation has ended up having the ultimate in asymmetric stakes for one party, and at-will employment has played the lion’s share in making it happen.

It seems to me that three things about this case have to be granted at the outset. First, however contrived it may seem, the case can’t justifiably be dismissed as impossible or improbable. It’s neither. None of the conditions is individually or jointly impossible. And none is entirely implausible, either.

For obvious reasons, there are unlikely to be statistics on the frequency of cases exactly like Lost Health Coverage, but both bankruptcy and premature death due to denial of health insurance benefits take place with startling frequency in the United States.(2) And surely nothing prevents them from taking place due to termination of employment in situations like those in Lost Health Coverage. Though loss of benefits is mitigated to some degree by EMTALA, COBRA, Medicaid, safety net hospitals, and charity care, the existence of the first four contradict strict liberal strictures, and charity care covers only a portion of the relevant cases. COBRA is in any case often exorbitantly expensive, particularly so for someone deprived of an income through unemployment. And EMTALA only applies to a narrow range of emergency cases.

Take all but charity care out of the picture, and cases like Lost Health Coverage will tend to show up with alarming frequency. And taking these programs out of the picture is entirely justified if we’re trying to contrast strict liberalism with republicanism. The strictest forms of liberalism tend toward libertarianism, and the strictest forms of libertarianism rule out regulatory fixes like COBRA, social welfare programs like Medicaid and safety net hospitals, or legal mandates like EMTALA.(3)

Second, if we accept the facts Lost Health Coverage at face value, it seems obvious that A’s actions are deeply unjust. For A to fire B despite B’s doing excellent work is unjust: it flouts what B deserves. For A to fire B out of rancor (or envy) compounds the offense. For A to fire B in the knowledge that B is ill and depends on employment for health care, is likewise unjust: absent egregious wrongdoing by B, it expresses callous indifference to B’s suffering on top of the injustice done in failing to recognize B’s merit. Prima facie, and absent some normatively significant countervailing consideration, an impartial spectator with full knowledge of all the facts would be inclined to think that B’s bankruptcy and death ought not to have happened–and are the kind of thing that a just society should strive to make rare or impossible.(4)

Finally, the problematic character of the case is facilitated in an essential way by employment-at-will as a practice. Substitute a requirement of showing just cause for the at-will arrangement, and though you wouldn’t necessarily rule out such cases, it’s likely that you would make them much less probable.

So far, I think, Pettit and I would agree. Lost Health Coverage is possible, even probable; is unjust; and is facilitated in an essential way by employment-at-will. Were A forced to articulate his reasons for firing B in a public way, he would likely fail to do. If termination requires the articulation of a for-cause justification, B would likely not be fired. If B keeps his job, he keeps his health insurance. If he keeps his health insurance, ex hypothesi, he stays solvent and stays alive. Case closed. The problem, however, is that the case is closed as far as the case is concerned, but not as far as the practice or the doctrine. In other words, you can grant everything I’ve so far said, and still remain unconvinced that we should do away with employment-at-will as a matter of general policy. The problem here lies with the use of a single example to drive us to a general conclusion about a legal doctrine which has a multiplicity of different outcomes in different circumstances.(5) Lost Health Coverage is a compelling example, but that strength might also be its weakness.

Consider how I generated Lost Health Coverage in the first place. I deliberately set out to design a case that painted employment-at-will in the worst possible light. I had earlier suggested that employment-at-will is most problematic when it exemplifies two properties: strong asymmetry in the relative power of the parties, and high stakes for the weaker party. So for purposes of the example, I pictured a case that exemplified both properties in a deliberately tear-jerking way. In doing so, I baked a series of facts into the description of the example itself, all intended to heighten the problematic nature of the asymmetry and stakes involved. I then treated the example as though it was typical or modal of the broad range of cases. I ignored examples of a contrary nature, and omitted any causal account of how or why the situation took the form it did. The result is an example that’s possible and even plausible, but not necessarily representative of the normative complexity of the issue. It’s also an example that involves a sort of unexplained package deal: the various features of the example are treated as an irreducible whole; we’re then invited to treat the whole as a unified package rather than ask why the whole has the parts it has.

But this is a potentially misleading procedure. For one thing, the presentation of Lost Health Coverage ignores the many cases that have different, even radically different, outcomes and implications. Most obviously, it ignores cases where the asymmetries are less radical and the stakes lower. In such cases, it’s not obvious that the at-will arrangement yields an overall outcome that intuitively inclines us to the abolition of the practice.

Beyond this, the example conceals the fact that employment-at-will might have positive outcomes. Imagine a case in which a terrible employee–e.g., one engaged in rampant sexual harassment, making threats of violence, or engaging in some other serious malfeasance–justifiably needs to be fired as soon as possible, without the employer’s having to waste time in giving notice and give a full explanation. If we know that the employee is guilty–if the evidence is good enough for that judgment–then it’s arguable that anything less than immediate termination is a form of delinquency.(6) Coming the other way around, we can imagine an employee who justifiably needs to leave his job immediately without being obliged to give notice or explain why. In cases like this, employment-at-will doesn’t look so bad, and doesn’t help Pettit’s case. All of this suggests that Lost Health Coverage might be possible and plausible without being modal.

Finally, it’s worth asking how cases of the generic type exemplified by Lost Health Coverage arise in the first place. The absence of such an account treats the etiology of the case as a kind of black box. But that in turn leaves open the possibility that the case has the package of features it has due to an etiology favorable to liberalism rather than republicanism. Suppose for instance that the explanation for the case, or for its most objectionable features, makes essential reference to past violations of liberal rights. In other words, suppose that the reason why employment-at-will is problematic is that a long series of past interferences into the economy has made it that way. Suppose that these interferences are precisely of the sort that liberalism rules out as violations of liberal freedom. And suppose, further, that had those interferences not taken place, employment-at-will would lack the objectionable features it’s taken to have in the example.

If so, the example doesn’t do the work the republican needs it to do. Rather, the case tends to show two things with jointly inconclusive implications:

• The adoption of a practice compatible with liberal freedom produces a highly objectionable outcome at t, and
• Violations of liberal freedom prior to t explain why the practice has the objectionable features that produce the objectionable outcome at t.

In other words, employment-at-will might produce bankruptcy and premature death in a given case right now (first bullet above), but past violations of liberal freedom might explain why employment and health care are so closely connected in the first place, and thereby explain why employment-at-will leads to bankruptcy and premature death (second bullet). We could, after all, imagine an economy in which the two things, employment and health care coverage, have far less bearing on one another than they do in my example. In such an economy, there might be no particular connection between losing your job and losing your health insurance, or even between being ill and needing health insurance. In that case, the example falls flat.

“Imagining” such an economy need not be a purely armchair exercise. It might arise, instead, by asking questions perfectly amenable to rigorous treatment by social science. Why, for instance, does the example assume that non-employment-based health insurance is hard to get, or that jobs are hard to find? Why does it assume that the individual’s health care bill is unmanageable enough to produce bankruptcy? And so on. If liberal political economy successfully explains the degree and counterfactual stability of the asymmetry between the parties, it might explain why at-will employment arrangements have stakes like Lost Health Coverage. And that’s something liberal economists have been trying to do since The Wealth of Nations. To the extent that they succeed, the power of Lost Health Coverage is significantly weakened, possibly defeated. It either becomes an anomaly swamped by contrary examples, or the unintended consequence of an illiberal, dirigiste approach to political economy.

A very specific historical trajectory explains why health care benefits in the contemporary United States are so closely connected to employment; why health care is so unaffordable on a self-pay basis; why insurance is so difficult to get except through an employer; and why employment can itself be so difficult to get. It’s misleading in this context to concoct a single contrived thought-experiment, and wield it as an all-purpose objection to the theory most adversely affected by it (or in favor of the theory most likely to benefit by it). However things end up, the history–the etiology–matters here. So it’s not clear we can adequately do normative theory by bracketing it. By bracketing it, we run the risk of adopting what Mill famously called a “one-eyed” point of view: clarity about one matter offset by fuzziness about others of equal importance.(7)

A republican critic of liberalism could of course stipulate that Lost Health Coverage typifies the world we inhabit, rule out further inquiries into the explanation behind it, and insist that we focus on this example for its own sake. But even if we accept this demand–and philosophers love making demands of this sort–a critic could with equal justification point out that if so, the lessons of Lost Health Coverage only apply to worlds in which the stipulated facts have the ontological power to crowd out any others. But we don’t live in that world. The world we actually inhabit is more complex than the stipulations of any given thought-experiment.

It’s at least possible that in the real world, employment-based health care arose through departures from liberal principles, and was entrenched through repeated departures. If so, a thought-experiment that conjoins employment-at-will with bankruptcy and premature death through the unavailability of employment-based health insurance is misleading and tendentious. Stipulate the reverse world, one in which health insurance is easily available outside of employment, and the would-be problem vanishes. And if one side can stipulate, so can the other. There are no asymmetries of bargaining power in philosophical argumentation.

Given that, I would put my disagreements with Pettit on employment-at-will this way:

1. Pettit’s view condemns employment-at-will not just “as currently expressed in the American labor market,” but as such, regardless of context. By contrast, I would say that context matters.
2. Pettit’s conception of employment-at-will doesn’t just suggest that it frequently exemplifies domination but that it does so as a matter of conceptual necessity. I prefer the weaker view.
3. Pettit doesn’t just condemn employment-at-will, but wants to use the power of the State to abolish it. I’m sympathetic to that view, but reluctant to adopt it. There are just too many complexities here to fast forward to that conclusion.
4. Pettit’s condemnation of employment-at-will implies that we can condemn instances of it without worrying about how it got the problematic features it’s now come to have. Contrary to Pettit, it seems to me essential to consider the etiology in question.

In short, while I agree with much about Pettit’s critique of employment-at-will, it seems to me that the issue is more complicated than he makes it.

Notes

1. The example is modeled on an actual case that arose when I was co-chair of the Department of Philosophy and Religious Studies at Felician University in the 2013-14 AY. I had originally posted a summary of it here, but had trouble with the formatting. It was in any case excessively long for an endnote, so I’ve deleted it for now, and will at some point post it in a series of appendices to this series.
2. See Xan et al, “Associations of Medical Debt with Health Status, Premature Death, and Mortality in the US,” JAMA Network Open (March 2024).
3. Strictly speaking, charity care is itself a matter of government mandate, and involves liberal interference: “Among other government regulations, federal law requires that nonprofit hospitals—which account for nearly three-fifths (58%) of community hospitals—provide some level of charity care as a condition of receiving tax-exempt status, and many state governments require all or a subset of hospitals to extend eligibility for charity care to certain groups of patients,” Zachary Levinson et al, “Hospital Charity Care: How It Works, and Why It Matters,” KFF (Nov. 2022).
4. Cf. Richard Epstein, “In Defense of the Contract at Will,” University of Chicago Law Review 51:2 (1984), p. 972. Though Epstein complains about the problematic use of hypothetical examples by critics of employment-at-will, and about the totalitarian implications of attempts to regulate the practice, he fails to acknowledge even the possibility of cases like the one in the text, and a fortiori makes no effort to deal with them. “The strength of contract at will,” he claims, “should not be judged by the occasional cases in which it is said to produce unfortunate results, but rather by the vast run of cases where it provides a sensible private response to the many and varied problems in labor contracting” (p. 982). In fact, Epstein provides no evidence to show or even suggest that the “unfortunate” cases are “occasional,” or that the “vast run of cases” have positive outcomes specifically as a result of adherence to the doctrine. Virtually every empirical claim he makes is either an armchair pronouncement detached from empirical evidence, or a generalization licensed by the unwarranted (and extremely implausible) assumption that the cases prominent in the legal literature must of necessity be representative of conditions in the labor market. Obviously, the more disadvantaged the victim, the less likely to be able to afford legal representation; the less able to afford legal representation, the less likely it is that her case will appear in the legal literature.
5. Cf. Richard Epstein, Simple Rules for a Complex World,p. 157, who makes this point. Though I grant the general point Epstein makes, it seems to me that he himself violates it fairly often in practice.
6. Things become murkier the further we drift from full-fledged knowledge. Ironically, I’ve devised this example from cases I know from my time at Felician University that exemplify the exact reverse of the situation described in the text: people were fired at-will from false accusations of malfeasance where the “evidence” was entirely (and transparently) cooked. In both of the cases I have in mind, male professors were falsely accused of sexual harassment, and were fired on grounds of the (confabulated) “danger” they presented to female students. There was in fact no such evidence, and rumors aside, no bona fide evidence was ever presented to anyone. In both cases, I know with certainty that the charges were fraudulent, and that both terminations were made in egregious bad faith. That said, the examples of politicians like Donald Trump, Robert Menendez, and Eric Adams et al all suggest the plausibility of the example in the text.
7. Mill describes Bentham as “one-eyed” in section 2 of “Bentham”: “For our own part, we have a large tolerance for one-eyed men, provided their one eye is a penetrating one: if they saw more, they probably would not see so keenly, nor so eagerly pursue one course of inquiry.”