LIFE-CYCLE JUSTICE: ACCOMMODATING JUST CAUSE AND EMPLOYMENT AT WILL
NAME: Stewart J. Schwab *
* Professor of Law, Cornell University. B.A. 1975, Swarthmore College; J.D. 1980, Ph.D. (Economics) 1981,
University of Michigan. - Ed. For wonderful discussions and comments I thank Greg Alexander, Lynn Baker,
Chris Bruce, Jim Brudney, John Donohue, Ron Ehrenberg, Bob Hillman, Bob Hutchens, Susan Koniak, Saul
Levmore, Tom Rowe, Theodore St. Antoine, Bob Scott, John Siliciano, Kathy Stone, Steve Thel, Steve Willborn
and workshop participants at the University of Virginia Law School Faculty Workshop, the Labor Economics
Workshop at Cornell University's Industrial and Labor Relations School, and the annual meetings of the Amercian
Law and Economics Association and the Canadian Law and Economics Association. I thank Mike Holden for his
fine research assistance.
... "You're fired!" Most American workers hearing these words have no legal recourse under traditional employment law.
... A career--employment relationship faces two types of opportunism: opportunistic firings by an unfettered employer
and shirking by employees with job security. ... Inside productivity is above outside productivity at all stages, even before
much training, to emphasize the efficiency--wage story that harder work results in greater productivity throughout. ...
Such a firing clearly violates the good--faith notion that an employer must give the employee the benefit of the bargained--
for pension. ... In Bay v. Times Mirror Magazines, Inc., a fifty--four--year--old publisher who made nearly $200,000 per
year brought an ADEA claim after being fired from his position at Field and Stream. ... Unless the parties have clearly
agreed to at--will dismissals throughout the life cycle, such a firing smacks of an employer opportunistically firing an
employee who has committed the best years of his life and should reap, based on the norms of seniority, the benefits of
a career commitment to the employer. ... Under either approach, then, a life--cycle default rule is optimal. ... The life--
cycle framework that courts have developed provides the parties in a career employment relationship a legal structure that
checks opportunistic behavior. ...
"You're fired!" Most American workers hearing these words have no legal recourse under traditional employment
law. Unless they have specific protection in a contract - and few nonunion workers do - or evidence that the firing violates
a specific antidiscrimination statute, workers are treated at will. Under this traditional view, an employer can fire them for
a good reason, a bad reason, or no reason at all. n1
Yet, as any employer who has recently tried to fire a worker will complain, the traditional doctrine is now full of
holes. In the past decade or two, courts increasingly have become receptive to worker claims of unjust dismissal. Still,
the protections remain patchwork and the trend uneven. No state court has proclaimed a general right of employees
against arbitrary dismissal, although the Montana court has come close. n2 Indeed, the Supreme Courts of California
and Michigan, long considered to be among the leading innovators in this area, have recently issued major decisions
signaling hesitation with ever--expanding protections for employees. n3 On the legislative side, only [*9] Montana,
reacting to its extreme court decisions, n4 has a statute requiring employers to have good cause before firing a worker.
n5 However, the Commissioners on Uniform State Laws have recently passed a Model Employment Termination Act that
would finally abolish the at--will doctrine for most employees, n6 and we can expect further debate over employment at
will as legislatures confront the model act.
Snapshot examinations of current unjust dismissal law have led employment law commentators to see only chaos, a set
of cases with little internal coherence or rationale. A view of the decisions in their historical sweep causes commentators
to see evolution: in the beginning was employment at will, n7 now chaos exists, the natural ending point will be
just cause. Most commentators applaud the trend, urge its completion, and bemoan any hesitation or backsliding by the
courts. n8 Their continual refrain is that the United States lags behind [*10] the rest of the world on this issue. n9 At
the other extreme, some conservatives, most prominently Richard Epstein, hearken back to the heyday of employment at
will as the ideal state of employment law. n10 Such observers lament every move away from a strict presumption that all
employment contracts can be terminated at will by either party.
The fundamental problem with these perspectives is that they criticize current law but they do not understand or
explain it, except in the crudest way. The just--cause boosters simply applaud every pro--employee decision and decry the
others as backsliding responses to conservative political pressures. The at--will zealots simply cheer and boo the other
way. Neither perspective appreciates the apparent vacillation of current courts, which erode the at--will presumption
without rejecting it.
Current termination law does have an underlying coherence. We should recognize this coherence before we reject
the contemporary common law system for either the Model Act's futuristic scheme of arbitrating just cause or a return
to the heyday of employment at will. This paper attempts to articulate the coherence of current doctrine. Reacting to
the almost uniform polarization on this issue, I argue both positively and normatively for an intermediate position. The
current intermediate position of the common law balances two conflicting problems. A career--employment relationship
faces two types of opportunism: opportunistic firings by an unfettered employer and shirking by employees with job
security. An extreme legal rule can handle either problem alone, but only by ignoring the other. Thus, a legal presumption
of employment at will handles the shirking problem well but gives no protection against opportunistic firings. A just--
cause regime has the opposite virtue and flaw. The legal challenge is to find an intermediate rule that provides the optimal
check against both dangers.
The common law has groped towards such a rule by recognizing [*11] that the relative magnitudes of the two
problems vary over the life cycle of the worker. The danger of employer opportunism is greatest for late--career workers,
and it is also a problem for some beginning--career employees. By contrast, the greater problem at midcareer is shirking.
In response, the courts have begun to offer contract protections for workers at the beginning and end of the life cycle,
while maintaining a presumption of at--will employment for midcareer employees. In arguing for the wisdom of this
approach, I, like the courts, refrain from making a categorical statement that at--will or just--cause employment should
never - or should always - be the governing presumption.
Let me sketch at the outset the boundaries of my inquiry. Using an internal--labor--markets model of career
employment, this paper examines whether courts should presume, in the absence of any express contractual provision
about job security, that an employee is at will or protected by just cause. This paper therefore skirts the fascinating contract
issues involved in employee handbooks and other arguably express agreements about the standard of discharge. Nor will
it evaluate the appropriate scope of tort limitations on employment at will. n11 The goal of this article is to articulate
a coherent framework for understanding the default rules for employment termination. While most observers see chaos
here, I find a certain logic in the leading cases. The courts have been boldest when job protection is most appropriate, and
they have hesitated precisely when at will plays its most useful role.
An initial caveat about methodology is appropriate. When applying the model to the case law, I will not survey
every termination case, or even most or many of the thousands of cases. Rather, my method is to use "leading" cases -
those that may indicate trends in the law - and argue how these leading cases might fit together in a coherent structure.
Leading cases rarely typify the mass of cases, and counterexamples certainly exist. Leading cases are worth studying,
however, particularly in a rapidly evolving subject area, because they often give greater hints about the appropriate
normative structure of law than [*12] would an exhaustive study of the hidebound mass of cases. Nevertheless, because
of its reliance on leading cases, my inquiry must blend positive and normative analysis. Perhaps the question addressed
here should be phrased as whether the leading employment--termination cases can fit into a coherent structure that has
One final caveat before we begin. One could argue that we should scrap common law litigation over employment
terminations as not worth the cost of litigation. n12 Whatever rationality the current common law lines might draw, the
argument goes, its advantages are swamped by the costs of deciding on what side of the line a particular employee falls. To
my mind, this litigation--cost argument is the strongest argument for taking a polar position. By itself, however, it does not
point clearly toward either pole. Supporters of a rigid at--will presumption emphasize that fewer employees will challenge
terminations under their standard - thus saving on litigation costs. n13 At the other pole, proponents of just cause,
worried about the feasibility of courts reviewing every dismissal, typically link their proposals to arbitration rather than
litigation as the method of enforcement. n14 In any event, courts themselves are poor judges of the metadecision about
whether to have common law enforcement of employment terms. Reacting to the cases they see, courts have attempted to
create a coherent doctrine. Before rejecting their efforts on grounds of cost, one should appreciate the balance they are
trying to achieve. Until now, no one has attempted to find the method behind the current madness. While the intermediate
solution of the current common law inevitably will have detractors on both sides, much can be said on its behalf. The goal
of this article is to present that defense.
I. Employment as a Relational Contract
Workers have many types of jobs and many types of relationships with employers. Younger workers typically try
several jobs before beginning a long--term attachment to one employer. n15 This variety of [*13] relationships may
in itself counsel against a uniform legal approach to employment terminations. I will return to this issue at the end of
this article, n16 but most of my analysis will concentrate on the career employee. Termination of the career employee
provides the fact pattern for many leading cases in the erosion of employment at will.
A. The Specific Human--Capital Story
The key feature of the career employment relationship is that both sides are locked into it. The easiest explanation
for lock--in comes from a human--capital story that emphasizes "asset specificity." n17 Under the basic human--capital
model, workers become more productive as they learn the ways of the firm. Because the gains exceed the costs of training,
these firm--specific skills are worth learning. In contrast to general skills, however, these skills are not useful to other
The issue becomes whether the employer and employee can decide how to share the costs and benefits so that this
desirable training will occur. This issue can be resolved in a number of ways. One possibility is for the employer to pay
for the firm--specific training and to receive the benefits of the greater productivity by paying the worker throughout his
work life a wage equal to the wage he could get outside the firm. The problem is that the worker has no incentive to stay
with the firm because he earns no more than he could get elsewhere, and so the employer risks losing the employee before
it can recoup its training [*14] investment. Another possible resolution is for the worker to pay for the training, perhaps
by accepting a lower wage during the training period, and for the employer to pay the worker a higher wage once trained.
The mirror image of the earlier problem occurs in this situation: the firm has an incentive to underpay the worker once he
is trained, and the worker has no recourse - because the training is worthless elsewhere - as long as he is paid at least the
The best solution is for the employer and worker to share both the costs and benefits of firm--specific training. Figure
1 gives a stylizedFIGURE 1view of the situation. The outside wage is constant, equaling productivity based on general
skills. n19 In the training period at the firm, the worker accepts less than the outside wage, thereby paying for some of
the training, but the employer pays him more than his productivity during the training period, thereby paying for some
of the training. After training, the worker receives more than the outside wage, thereby reaping some of the benefits of
training, but the employer does not pay the worker for his full productivity, thereby allowing the employer to reap some of
the benefits of training. The worker will agree to this arrangement if the higher post--training wages, when appropriately
discounted, are at least as large as the lower training wages - [*15] that is, if A' is at least as large as A. The employer
will agree as long as the gains to it (B') are at least as large as its training costs (B). n20
In practice, these higher post--training wages take the form of seniority--based wages and late--vesting pensions,
which induce workers to stay with the firm after training. n21 Compared with the life cycle of otherwise similar
workers, the model predicts that workers who receive substantial on--the--job training will receive higher pay in later
years. Considerable empirical evidence supports this steep age--earnings profile of career employees and its relationship
to training early in the career. n22
A critical part of this simple human--capital story is the self--enforcing feature of the relationship. Because the parties
share the costs and benefits of training throughout the employee's work life, both parties want to continue the relationship.
The employer pays employees less than their full value later in their career. This protects employees from discharge
because a discharge would harm the employer as well. The late--career wage exceeds, however, the outside wage the
employee could receive, thereby discouraging the employee from quitting.
B. The Efficiency--Wage Story and the Potential for Opportunism
Gary Becker's human--capital theory n23 explained why wages rise with seniority, but puzzles arose that caused
commentators to question the theory that workers would receive less than their value late in their career. n24 One such
puzzle was mandatory retirement, which covered some thirty--five percent of the workforce before it was prohibited in
1986 by amendments to the Age Discrimination in Employment Act (ADEA). n25 Why would an employer, who is
paying older workers less than they are worth, demand that such workers retire? A related puzzle is the presence of
actuarially unfair pensions, which even after the 1986 ADEA amendments lawfully encourage older workers [*16] to
retire. The present value of many pensions declines if older workers keep working. n26 Again, why would an employer
establish pension plans that encourage retirement if, as the human--capital model suggests, the employer makes money
on older workers? A final puzzle stems from studies that suggest that workers' pay relative to others in their job grade
increases with seniority but their relative productivity does not. n27 This evidence conflicts with Becker's hypothesis that
productivity increases faster than wages. These puzzles suggest that employers want to end the relationship with late--
career workers at some point, probably because employers perceive their wages as exceeding current productivity.
To explain the puzzles, economists have developed an efficiency--wage model. n28 The basic insight behind
efficiency--wage models is that workers often work harder when the job pays more. n29 High "efficiency wages" n30
increase worker effort by making the job more valuable to the worker. Because workers want to keep the valuable
job, they will work hard to avoid being dismissed. In effect, high wages increase the penalty for being dismissed - a
dismissed worker forgoes the large payout. Because workers labor harder than otherwise, the firm can afford the higher
compensation. An early version of the [*17] model n31 assumed that a firm employing this strategy n32 raised wages
by a constant amount throughout a worker's life cycle. One problem with the early, flat--wage version of the model is that
workers nearing retirement would care less about losing their high--wage job and would therefore be less deterred from
Later versions of efficiency--wage models suggest that compensation will rise over a worker's career to induce effort
throughout. n34 The implicit contract promises large payouts for senior workers, but it promises this reward only for
hard--working employees. n35 The firm recognizes that day--to--day monitoring of a worker's effort may be difficult, but
over a period of years the firm hopes to spot and weed out shirkers. The threat of being fired before the large payoff keeps
employees working hard. Because employees work harder than otherwise, the firm can afford the higher compensation.
Large law firms epitomize this model.
A related literature emphasizes that firms may conduct internal tournaments to induce high effort by junior and
midlevel management employees. n36 Tournaments are especially likely when firms cannot monitor actual effort or
output but can evaluate relative performance. A firm may (implicitly) tell an incoming class of workers that the best worker
will win the grand prize of C.E.O. A single prize may be insufficient inducement, however, so the firm may (implicitly)
offer several runner--up prizes of cushy vice--president jobs for those who try hard but fail. This model likewise suggests
an implicit agreement whereby firms pay late--career employees more than their current productivity. [*18]
Figure 2 grafts the efficiency--wage story to the human--capital model. As in the human--capital story, the outside
wage is flat n37 andFIGURE 2equal to outside productivity. The inside wage may be initially lower than the outside
wage, n38 but then rises. Inside productivity rises as job--specific training occurs. Inside productivity is above outside
productivity at all stages, even before much training, to emphasize the efficiency--wage story that harder work results in
greater productivity throughout. Figure 2 adds two other wrinkles that will become important later. First, because I will
specifically address the case law of employees who change employers, Figure 2 includes a preemployment "moving"
period. In this stage, an employee quits the outside job and incurs moving expenses, thus having a negative wage. Figure
2 also explicitly illustrates a postretirement pension. [*19]
The critical point of the expanded story is that the implicit contract is not always self--enforcing. As Figure 2
illustrates, after point T, firms pay late--career employees more than they currently produce. At this point, late--career
employees become vulnerable to opportunistic firing n39 because the general self--interest check on arbitrary firings
does not exist; firing such a worker does not hurt the employer but is instead in its immediate economic interest. One
can see a role for law in improving this situation. By policing against opportunism, the law can make the employment
relationship more secure for and valuable to both sides. To understand fully the role law can play, we must examine the
concept of opportunism more closely.
C. Contracting Problems in Career Employment
One solution to the problem of opportunism is for the parties entering into career employment to negotiate detailed
contracts, enforceable by courts, that specify appropriate behavior by both sides. Unfortunately, three contracting
challenges make detailed contracts an unsatisfactory solution. First, the parties cannot easily anticipate the future
contingencies, or states of the world, that will influence the relationship. n40 Will demand for the product stay strong?
Will the firm shift its focus from the employee's specialty to other areas that require more general skills? While all
predictions of the future are difficult, anticipating events twenty or thirty years in advance is an exceptional challenge for
employers and employees.
Second, even if parties can anticipate a future event, they may have difficulty specifying in detail the appropriate
contractual response, n41 particularly when one party has access to relevant information that the other cannot easily
observe. n42 A key element in the employment relationship is whether the employee is working hard, or, from the em
[*20] ployee's perspective, whether the employer is dismissing him for failure to work hard or for an unfair, opportunistic
reason. When monitoring is difficult, two alternatives emerge. First, the parties may decide not to make any part of the
contract contingent on difficult--to--monitor behavior. An at--will clause would accomplish this goal by making worker
efforts irrelevant to the permissibility of discharge. Alternatively, the parties may write a vague "best efforts" or "good
faith" clause for the contingency. n43 While this solution invites later court or arbitrator supervision over the meaning of
the terms, that supervision may be preferable to contractual language that straitjackets parties' future options.
Finally, having anticipated a future problem and specified the contractual response, a party may be unable to prove a
breach in court. Economists term this problem an unverifiable contract. n44 Unverifiability is particularly problematic
when the contractual language is vague - as it will often be in relational contracts. Both employer and employee might
know that the employee is not working as hard as "best efforts" require, but the employer cannot assemble sufficient
objective evidence to convince a court or arbitrator of this fact. If the parties cannot turn to outside enforcement, they
must develop self--enforcing mechanisms for any agreement to be effective. But, as we have seen, career--employment
contracts - particularly those following the efficiency--wage model - are not fully self--enforcing. n45
D. Potential for Opportunism
In the absence of enforceable, detailed contracts that regulate behavior, parties to a long--term relationship become
vulnerable to opportunism. They cannot easily leave the relationship because they would have to repeat the investments
or forgo their value. n46 The existence of "sunk costs" for one party creates a potential for opportunistic behavior by the
other side. A firm can pay workers less than they are worth or treat them more harshly than the initial agreement contem
[*21] plated, knowing they cannot easily move. The employees can produce less than their skills allow, knowing the
employer cannot easily replace them.
The law can sometimes help monitor opportunistic behavior, thereby increasing the parties' overall gains from the
relationship. If the law can enforce promises not to exploit the other side's vulnerability, the parties can more confidently
invest and the relationship will be more rewarding to both sides. The law has limits, however, because the contractual
language often will be general and vague, as we have seen. n47 More importantly, because both employer and employee
are investing, both can be exploited. A legal rule favoring one side would leave much opportunism by the other unchecked.
To curb opportunism adequately, courts must engage in difficult, case--by--case assessments or create more flexible
presumptions. As we will see, courts have responded by adopting presumptions that vary with an employee's life cycle.
1. Employer Vulnerability to Shirking
As the human--capital model indicates, employers make heavy investments in recruiting and training workers. n48
To ensure an adequate return on their investment, employers want workers to stay and produce for them after training.
Some scholars focus on employer recruitment and training costs in emphasizing employer vulnerability to employees
quitting, but this is not the trueproblem of opportunism. The basic human--capital model suggests that employers can
adopt delayed--payment schemes to discourage quitting, thus making the contract self--enforcing. Late--vesting pensions
and seniority--based wages can tell workers: "If you stick around, you will do well."
The greater risk to employers comes from employee shirking. The efficiency--wage model highlights the shirking
problem. Even if pensions and seniority wages discourage workers from quitting, an employer still faces problems when
workers stay. Workers often do not work as hard as they would under a fully specified and monitored contract. Economists
label this behavior shirking. n49 A fully specified op [*22] timal contract would designate an optimal level of effort.
Workers would agree to exert this effort because they prefer the higher wages that accompany it to an easier work life;
the employer would agree because the greater productivity is worth the higher wages. Once hired, however, an employee
may shirk from this optimal effort if employers have difficulty monitoring or replacing workers. Indeed, it is irrational for
workers to work up to "optimal" levels if they prefer coasting a little. Workers know that the employer will have to spend
money to catch shirkers and that, if it fires a shirker, the employer will have to recruit and train a replacement. As long as
workers perform better than a rookie would - considering the costs of monitoring, recruiting, and training - the firm must
accept less than optimal efforts from its workers.
Much of the debate over at--will employment addresses whether parties can write effective contracts to overcome
the shirking problem. To put it bluntly, the real question is whether the threat of firing for cause is sufficient to deter
substandard performance by workers. Proponents of at will emphasize the unverifiability of the performance standard in
many employment contracts. The employer may know the worker is shirking but cannot convince a court or arbitrator
that the conduct amounts to shirking. Oliver Williamson has emphasized the difficulty in distinguishing a consummate
performance from a perfunctory performance: "Consummate cooperation is an affirmative job attitude whereby gaps are
filled, initiative is taken, and judgment is exercised in an instrumental way. Perfunctory cooperation involves working to
rules and in other respects performing in a minimally acceptable way." n50 One problem employers have in documenting
a "perfunctory performance" is that particular instances of misconduct often seem trivial. Concluding that they add up
to a significant problem requires acknowledging that the whole problem exceeds the sum of the parts. The heart of the
employment-at-will argument is that proving cause under what is essentially an unverifiable agreement against shirking
places too great a burden on employers, preventing them from effectively using efficiency wages to deter shirking.
Some may argue that commentators overstate this shirking problem because the employee's desire for a good
reputation deters shirking. Even if shirking is possible in a just--cause world, this counterargument runs, benefits accrue
to employees with a reputation for hard work. Not only may the incumbent employer reward hard work with promotions
and pay raises, but employees with good repu [*23] tations are most attractive to outside employers. Nevertheless, this
reputation argument, in both its inside and outside reputation forms, ignores several important facts. First, an individual
worker with a good inside reputation may not reap major rewards. As we saw, employers often establish pay and promotion
ladders that do not depend on current individual productivity in order to discourage quits with promises of big paydays
in the future. n51 In such internal labor markets, pay scales attach to jobs rather than workers, and seniority rather
than merit often determines who gets the jobs. n52 Promoting individual workers simply because of individual hard
work may not be worth the disruptions in the general progression system. In these internal labor markets, then, unusually
good effort may not be rewarded even though unusually bad effort is punished by firing. n53 Second, an employee
may find it hard to acquire a good outside reputation if his skills are firm--specific. An academic whose publications are
useful to many potential employers can obtain an outside reputation. Indeed, some might argue that the major inside job
of the academic is to acquire an outside reputation. An engineer working on a classified defense project finds it more
difficult - and therefore has less incentive - to obtain an outside reputation for hard work. The reluctance of employers to
give candid references, itself a response to defamation law, exacerbates the difficulty for employees seeking to establish
A second response to the shirking argument involves a quick comparative law lesson. In the rest of the industrialized
world, at--will employment is unknown, n54 yet workers manage to work hard without the threat of firing. n55 As Jack
Beermann and Joseph Singer lament, why [*24] does our society assume it can trust employers not to abuse the power
of arbitrary firings while it refuses to trust employees protected by just cause? n56 Of course, one can overdramatize the
comparative lesson. Industrial tribunals in Europe, having found a dismissal to be unjust, usually award modest severance
pay that rarely exceeds six months duration. n57 Further, commentators of "Eurosclerosis" would caution against using
Europe as a model for productive labor markets. n58
Ultimately, the verifiability problem involves a question of degree, and the problem is greater for some jobs than
for others. To the degree that clear contracts against shirking are difficult to write, monitor, and enforce, opportunistic
behavior by employees will remain a threat.
2. Employee Vulnerability
As both the human--capital and efficiency--wage models emphasize, employees invest heavily as they pursue a career
with a single employer. First, they obtain training that is more useful for their own employer than it would be elsewhere -
what economists term job--specific human capital. n59 Second, they join the company's career path. This path, as we
have seen, ties pay, promotions, and benefits to sen [*25] iority and generally forbids lateral entry. n60 A major cost
of pursuing a career with one firm is that one forgoes other ladders and must start over at the bottom if one leaves the
firm. Additionally, as they plan for a lifetime with an employer, workers put down roots, establish networks of friends
in the workplace and the community, buy homes within commuting distance of the job, and build emotional ties to the
Losing these investments, roots, and ties can be devastating. Many studies document how even impersonal plant
closings lead to increases in "cardiovascular deaths, suicides, mental breakdowns, alcoholism, ulcers, diabetes, spouse
and child abuse, impaired social relationships, and various other diseases and abnormal conditions." n61 Being singled
out and fired may be even more devastating. n62
Because of these tremendous costs, no employee wants to lose his job involuntarily. Further, these investments, roots,
and ties are sunk costs that trap the worker in his current firm, inhibiting him from departing voluntarily. Even if the
career does not proceed as anticipated, the employee is reluctant to quit because the job remains preferable to alternative
jobs. Such trapped workers are vulnerable to opportunism. The employer might pay them less than the implicit contract
requires or work them harder, knowing they cannot easily quit.
By itself, this potential for opportunism does not justify a just--cause standard. Employers want such exploited
workers to stay, not to leave. Only when conditions become so intolerable that the employee prefers to quit for another
job might termination law come into play. In the economist's framework, this situation occurs when the employer has
appropriated all the gains from the relationship, making the career no longer better than alternative jobs. Lawyers label
these intolerable conditions a constructive discharge.
Defenders of at will contend that employer self--interest protects productive employees from discharge. n63 An
employer hurts itself by arbitrarily terminating a productive worker or by causing him to quit [*26] because it wastes the
recruiting and training investment in the employee. To avoid its own sunk--cost losses, an employer wants to keep good
workers and fire only workers who fall below the standard of new entrants. Figure 1 illustrates this self--enforcing feature.
After the training period, workers' productivity exceeds their wages, and the employer would lose this gain by firing such
While employers can make mistakes, the self--enforcing feature should minimize firing of productive workers.
Employees do not need the grand and expensive apparatus of the law for further protection, claim at will's defenders.
Indeed, its very expense harms employees as well as employers, for wages will inevitably fall as terminations become
Opponents of at--will employment remain skeptical. A major concern is that an employer is not a monolith but rather
a hierarchy of high--level managers and low--level supervisors. Often low--level supervisors make the decision to fire, and
the factors influencing their decisions are often not perfectly aligned with the profit--maximizing interest of shareholders.
Thus, while shareholders may not want employees to be fired arbitrarily, supervisors might. Personality conflicts and
power trips may lead supervisors to fire valuable and productive employees.
Again, one can overstate the dangers of front--line supervisors running amok. The firm has incentives voluntarily to
reduce supervisor mistakes so long as the gains in employee satisfaction outweigh the costs of supervising the supervisors.
Just--cause advocates cannot make their point simply by showing that agency costs exist. They must show further that
employers will not take cost--effective steps to ensure that they treat their employees fairly.
One check on such opportunism - emphasized in the efficiency--wage literature - is the employer's concern for
its reputation. If word gets out that an employer routinely fires older workers, it will be harder for the employer to
recruit entrants into career jobs. Perhaps more damaging than its outside reputation is its inside reputation with fellow
employees when older, productive workers are fired. n65 This loss of collegiality may encourage other workers to quit.
Problematically [*27] for the employer, the most productive workers likely have the greatest opportunities for moving
In many situations, reputation is unlikely to check fully the employer's incentive to fire late--career workers. Young
job entrants cannot easily assess an employer's reputation for how it handles senior workers. n66 Great problems arise
in passing on knowledge of a firm's opportunistic firings between generations of workers. n67 These problems are
particularly acute in small or new firms, n68 where much of the workforce works. n69 Finally, a reputation for harsh
personnel policies may not greatly harm declining firms that are not hiring many new workers. n70
Because reputation is not a full check on opportunism, firms must compensate workers for the risk that the delayed
bonanza may not accrue. Early--career wages, or the late--career bonuses and pensions, must be higher than they would
have to be were reputations more secure. Court scrutiny of opportunistic firings may offer another method of policing
long--term contracts. Such third--party scrutiny may allow employers to offer efficiency--wage contracts at lower overall
cost. The danger, of course, is that court intervention will diminish the employer's flexibility in firing workers whose
shirking a court can [*28] not verify. The question is whether court intervention can be limited to opportunistic firings,
rather than to a broader supervision against unfair firings in general.
II. An (Inconclusive) Appeal to Empiricism
We see, then, that both the employer and employee are vulnerable to opportunism. Which is the greater problem -
shirking or unjust firings? Or, more precisely, what legal rule can best reduce the problems? The choice is highly
political. Those sympathetic to workers will feel their exploitation is the greater problem and favor just--cause. Others
who instinctively rely on private markets may see a greater need to curb worker shirking and so favor at will.
Some scholars have looked at actual practices to help resolve the debate. n71 It is useful to sketch these appeals to
empiricism before rejecting them as inconclusive and moving to the life--cycle theory of career employment.
A. The Pervasiveness of At--Will Employment
Certainly, at--will employment is the legal norm for the vast majority of American workers. As a long--repeated
practice by many players negotiating many contracts, at--will employment is a plausible candidate for being the optimal
arrangement, in the sense that the overall gains, which primarily accrue to the employer, outweigh the costs of potentially
arbitrary firings, which are borne primarily by employees. Both sides can share this net gain by increasing the wage to
compensate workers for the risk of firing. Richard Epstein uses this empirical observation to bolster his argument for the
desirability of at--will employment: "The survival of the contract at will, and the frequency of its use in private markets,
might well be taken as a sign of its suitability for employment relations." n72
The empirical claim that most private employees are at will is rarely disputed. n73 One needs some faith in the
efficiency of labor markets, however, to infer that this proves that at--will employment is effi [*29] cient, in the sense
that the costs to employers from a switch to just cause would outweigh the gains to workers. Unequal--bargaining--power
theorists are unlikely to have such faith.
The noted institutional economist John Commons put it well long ago in disputing the claim that the employer and
the at--will employee have equal freedom to quit the relationship:
If the corporation has 10,000 employees it loses only one ten thousandth part of its working force if it chooses to not--
employ the man, and cannot find an alternative man. But the man loses 100 per cent of his job if he chooses to not--work
and cannot find an alternative employer. n74
Commons's observation envisions an employer with market power, perhaps a monopsonist or even a company town.
Even a monopsonist, however, would not insist on at--will employment if it harmed employees more than it helped the
monopsonist. A monopsonist can increase its profits by offering cost--effective fringe benefits because employees would
agree to work for a lower wage. n75 If employers had unlimited bargaining power, an employer would only pay workers
a subsistence wage. n76 This does not accord with the reality of the labor market for most American workers.
In any event, the monopsony model is inapt. Most employees today have alternative job options, n77 at least at
the beginning of a relationship, which limits the monopsony power of employers. If at will presents a serious danger
of exploitation and beginning employees see the danger, they will require higher wages before they accept the danger.
Refusing to concede that workers will receive benefits for which [*30] they are willing to pay, unequal--bargaining--power
enthusiasts shift their argument away from the existence of a monopoly to the absence of real bargaining. Even if workers
have several entry--level job options, the argument runs, workers have no real choice because every employer offers at--will
employment on a take--it--or--leave--it basis. Duncan Kennedy, hardly an unabashed enthusiast of the economics vision,
provides the best response to this argument. n78 As Kennedy realizes, one cannot test whether buyers as a class influence
compensation packages by studying whether individuals dicker over terms. n79 The analogous argument would be that
consumers have no influence over breakfast cereals because they cannot bargain with the grocery store manager over the
terms of sale. n80
Of course, as I have been emphasizing, both employer and employee gain more from the long--term relationship than
they would by starting over with others. This relationship surplus must be divided between them. There is no clear theory
or evidence about how the parties should divide the surplus. n81 Importantly, though, even an employer who captures
most of the surplus has an incentive to promise just--cause terminations if that is the efficient promise. The fact that we
virtually never see employers make such promises is some indication that at will is the efficient relationship, especially
when at will is supported by a plausible model of its efficiency.
B. The Unionized Just--Cause Standard as Counterfact
Rather than resort to inequality of bargaining power to refute the [*31] significance of at will's pervasiveness, a
weightier rebuttal n82 emphasizes that the just--cause standard always governs terminations among unionized workers.
n83 Free--market proponents might attempt a quick dismissal of this fact by saying that unions distort the market; one
cannot expect the unionized market to reflect efficiency. n84 But this response is too hasty.
The competitive market may reflect the preferences of marginal workers, but it does not necessarily reflect the wishes
of inframarginal workers. n85 In a competitive labor market, employers learn how to structure their employment package
from the ease with which they fill positions. They rely, in other words, on signals from workers they are trying to recruit,
or workers who are deciding whether to quit. If these workers do not reflect the interests of inframarginal workers, the
competitive market will not reach an efficient result.
The inframarginal workers - the career employees upon whom we have been focusing - are locked in. They have
made significant investments in the workplace and cannot easily leave. In a competitive market that relies on exit and
entry to register preferences, locked--in employees will not be included in the calculus. This is not a problem unless their
preferences differ in some dimension from the preferences of marginal workers less tied to the firm. It is quite plausible,
however, that the average worker has significantly greater preference for job security than does the marginal worker. The
marginal worker is typically young and in the process of trying out several jobs before [*32] finally settling on a firm
with which to establish a career. n86 This marginal worker may not recognize or value fully the firm's reputation for job security and may not realize that he may one day be locked into a career with this firm. n87