YLJO Essay

Losers’ Rules icon.pdf Print E-mail

Each year, the United States District Court for the District of Massachusetts holds an extraordinary panel. All active judges are present to answer questions from the bar. A lawyer’s question one year was particularly provocative: “Why are the federal courts so hostile to discrimination claims?” One judge after another insisted that there was no hostility. All they were doing when they dismissed employment discrimination cases was following the law—nothing more, nothing less.

I disagreed. Federal courts, I believed, were hostile to discrimination cases. Although the judges may have thought they were entirely unbiased, the outcomes of those cases told a different story. The law judges felt “compelled” to apply had become increasingly problematic. Changes in substantive discrimination law since the passage of the Civil Rights Act of 19641 were tantamount to a virtual repeal. This was so not because of Congress; it was because of judges.

Decades ago, law-and-society scholars offered an explanation for that phenomenon, evaluating the structural forces at work in law-reform litigation that lead to one-sided judicial outcomes. Focusing on employment discrimination claims, Marc Galanter argued that, because employers are “repeat players” whereas individual plaintiffs are not, the repeat players have every incentive to settle the strong cases and litigate the weak ones.2 Over time, strategic settlement practices produce judicial interpretations of rights that favor the repeat players’ interests.3 More recently, Catherine Albiston went further, identifying the specific opportunities for substantive rulemaking in this litigation—as in summary judgment and motions to dismiss—and how the “repeat players,” to use Galanter’s term, take advantage of them.4

In this Essay, drawing on my seventeen years on the federal bench, I attempt to provide a firsthand and more detailed account of employment discrimination law’s skewed evolution—the phenomenon I call “Losers’ Rules.” I begin with a discussion of the wholly one-sided legal doctrines that characterize discrimination law. In effect, today’s plaintiff stands to lose unless he or she can prove that the defendant had explicitly discriminatory policies in place or that the relevant actors were overtly biased. It is hard to imagine a higher bar or one less consistent with the legal standards developed after the passage of the Civil Rights Act, let alone with the way discrimination manifests itself in the twenty-first century. Although ideology may have something to do with these changes, and indeed the bench may be far less supportive of antidiscrimination laws than it was during the years following the laws’ passage, I explore another explanation. Asymmetric decisionmaking—where judges are encouraged to write detailed decisions when granting summary judgment and not to write when denying it—fundamentally changes the lens through which employment cases are viewed, in two respects. First, it encourages judges to see employment discrimination cases as trivial or frivolous, as decision after decision details why the plaintiff loses. And second, it leads to the development of decision heuristics—the Losers’ Rules—that serve to justify prodefendant outcomes and thereby exacerbate the one-sided development of the law.

I. The Skewed Evolution of Discrimination Law

Just as the social-psychological literature is exploding with studies about implicit race and gender bias—in organizational settings, in apparently neutral evaluative processes, and among decisionmakers of different races or genders—federal discrimination law lurches in the opposite direction, often ignoring or trivializing evidence of explicit bias.5 And just as empirical studies highlight the stubborn persistence of discrimination at all levels of jobs and in salaries,6 federal discrimination law assumes the opposite. In summary judgment decisions, judges search for explicitly discriminatory policies and rogue actors; failing to find them, they dismiss the cases.7 It is as if the bench is saying: “Discrimination is over. The market is bias-free. The law’s task is to find the aberrant individual who just did not get the memo.” The complex phenomenon that is discrimination can be reduced to a simple paradigm of the errant discriminator or the explicitly biased policy, a paradigm that rarely matches the reality of twenty-first-century life.8

Even without the contrary insights of social psychologists, this development is curious. Discrimination cases are about intent—in the language of the statute, whether an action was taken “because of” race or gender bias.9 Proof of intent is rarely direct. It is usually circumstantial, even multidetermined. In tort or contract cases, contests about intent require jury trials. Judges recognize that divining a person’s intent is messy and complex and that this issue usually involves a material dispute of fact for a jury to decide.10 Employment discrimination cases, in contrast, are typically resolved on summary judgment, although discriminatory intent may be more difficult to identify on a cold record than is the intent of a contract’s drafters or a putative tortfeasor’s state of mind.11 Why?

Is the explanation solely an ideological one? Is the cause a more conservative bench, and in particular a more conservative Supreme Court, that is far, far less supportive of antidiscrimination laws than it was in the past? That is surely part of it. But the outcomes I identify cut across the ideological spectrum, applying equally well to judges who would describe themselves as sympathetic to a discrimination plaintiff’s claims. Judges, as that District of Massachusetts panel reflected, feel that the law “compels” them to decide the cases as they do. But the “law” hardly compels anything in this context. Employment discrimination cases, after all, are factually complex, deal with state-of-mind issues, are typically proved circumstantially, and are rarely uncontested. The Supreme Court’s legal standards for summary judgment are so general that, for the most part, they provide a way to organize the record and frame the issues. They rarely mandate a result as would, for example, a statute of limitations or a failure to exhaust administrative remedies. Rather, the source of the law’s “compulsion” as the judges apparently experienced it was, at least in part, the phenomenon I call Losers’ Rules.

II. The Role of Asymmetric Decisionmaking

When the defendant successfully moves for summary judgment in a discrimination case, the case is over. Under Rule 56 of the Federal Rules of Civil Procedure, the judge must “state on the record the reasons for granting or denying the motion,”12 which means writing a decision. But when the plaintiff wins, the judge typically writes a single word of endorsement—“denied”—and the case moves on to trial. Of course, nothing prevents the judge from writing a formal decision, but given caseload pressures, few federal judges do.13 (During one case-management program in my district, the trainer, a senior judge, told the assembled judges, “If you write a decision, you have failed.” The message was clear: you would only write a decision when you absolutely had to.) Plaintiffs rarely move for summary judgment.14 They bear the burden of proving all elements of the claim, particularly intent, and must do so based on undisputed facts.15 Defendants need only show contested facts in their favor on one element of a plaintiff’s claim.16

The result of this practice—written decisions only when plaintiffs lose—is the evolution of a one-sided body of law. Decision after decision grants summary judgment to the defendant or, more recently, on the heels of the Supreme Court’s decisions in Twombly17 and Iqbal,18 dismisses the complaint. After the district court has described—cogently and persuasively, perhaps even for publication—why the plaintiff loses, the case may or may not be appealed. If it is not, it stands as yet another compelling account of a flawed discrimination claim. If it is appealed, the odds are good that the circuit court will affirm the district court’s pessimistic assessment of the plaintiff’s case.19

While the standard of review of summary judgment orders is de novo, appellate courts rarely reverse district courts’ decisions.20 It takes substantial work, not to mention a motivated decisionmaker, to dig into the voluminous summary judgment record and find a contested issue of fact. In my experience, few appellate court judges are so motivated in this area. On the contrary, they are even more affected than are district court judges by the skewed pool of cases they see—the selection effects of reviewing appeal after appeal of plaintiffs’ losses. They do not see the strong cases that settle. They may see appeals from successful plaintiffs’ verdicts, but those appeals are few and far between. What they mainly see is a litany of losing cases, each resolved on summary judgment for the defendant.

Although judges do not publish all the opinions they write, the ones they do publish exacerbate the asymmetry. The body of precedent detailing plaintiffs’ losses grows. Advocates seeking authority for their positions will necessarily find many more published opinions in which courts granted summary judgment for the employer than for the employee. And although one would expect that litigants would realize over time that their chances are slim and would then stop filing, the record proves otherwise. They continue to believe in the fairness of the system, notwithstanding the odds, flooding the courts with claims of all kinds—some frivolous, to be sure.21

But the problem is more than just the creation of one-sided precedent that other judges follow. The way judges view these cases fundamentally changes. If case after case recites the facts that do not amount to discrimination, it is no surprise that the decisionmakers have a hard time envisioning the facts that may well comprise discrimination. Worse, they may come to believe that most claims are trivial.

Statistics tell the story. A recent Federal Judicial Center report noted that roughly sixty percent of the summary judgment motions studied were granted in whole or in part, while more than seventy percent of such motions were granted in employment discrimination cases.22 From 1994 to 1995, “employers prevailed in approximately 86% of published appellate opinions.”23

III. One-Sided Heuristics: Losers’ Rules

In addition to contributing to one-sided outcomes, the asymmetry of the decisionmaking process distorts the evolution of substantive legal standards.24 Losers’ Rules evolve to justify the judicial analysis. These rules are heuristics, “simplistic, rule-like tests developed by the courts” to deal with otherwise complex cases in a more efficient manner.25 And they are particularly useful for organizing incomplete data,26 like those found in most summary judgment records. Although heuristics develop across many areas of the law, at many stages, the growing use of summary judgment in civil litigation in general,27 and its increased use in employment discrimination cases in particular, makes such “rule-like tests” all the more important. Thus, a pattern emerges. Courts create decision heuristics to enable them to quickly dispose of complex cases. They then write decisions employing the heuristics and publish their opinions. In short order, other courts rely on the heuristics, which become precedent, and the process is repeated over and over again.28

The problem with heuristics, however, is that they are subject to systematic errors in all directions.29 In the context of employment discrimination cases, false positives occur when a court finds that there may have been discrimination when there was not, and false negatives occur when a court finds no discrimination when there was.

When courts believe that most employment claims are without merit, as the decisional law suggests, they will be far more concerned with false positives than false negatives. One-sided heuristics—rules of thumb that oversimplify, dismiss, and often demean proof of discrimination—evolve. Sometimes their inculcation is explicit. At the start of my judicial career in 1994, the trainer teaching discrimination law to new judges announced, “Here’s how to get rid of civil rights cases,” and went on to recite a litany of Losers’ Rules. Indeed, he was right.

In Iqbal and Twombly, the Supreme Court—in effect—invited judges to use discrimination heuristics earlier in the litigation process than before, with far, far less information.30 Both cases involved a motion to dismiss the complaint under Rule 12(b)(6). The Iqbal Court encouraged judges to use their “common sense” and “judicial experience” to determine when claims are “plausible,”31 rather than to apply the far more objective notice pleading standards that predated these decisions.32 Under notice pleading, courts asked whether any set of facts could be proven consistent with the allegations in the complaint.33 Under Iqbal and Twombly, they are to determine whether alternative explanations for the events complained of are “more likely” than the allegations made by the plaintiff,34 a probabilistic determination in fact, despite the Court’s disclaimer.

The Court’s motivation could not have been clearer. It was concerned only about false positives—wrongful accusations of discrimination—not false negatives that leave meritorious claims of discrimination unredressed. It focused expressly on the transaction costs to defendants that such claims engender, not the impact on the plaintiffs whose claims are given short shrift.35 Its approach was compelled by the perception that case management, as the Court noted, had not been particularly effective in addressing the problem of insubstantial claims and had thus failed to control litigation expenses.36

But while it is one thing to be concerned about the limits of case management with respect to complex antitrust cases like Twombly, it is another to be concerned in connection with civil rights cases like Iqbal, where summary judgment has been wildly successful for employers and other defendants. Iqbal and Twombly have not yet produced wholesale dismissal of employment discrimination complaints,37 but, given employment discrimination heuristics and precedent in other fields,38 that is a fair prediction.

High on the list of heuristics that fundamentally distort the outcome of discrimination cases is what may be described as the “stray remarks” doctrine.39 This heuristic discounts explicitly discriminatory statements—ironically, just at a time when social psychologists are concerned about implicit bias. Consider the usual situation (which I noted in Diaz v. Jiten Hotel Management, Inc.):

  If a manager makes an ageist remark, it could well be a window on his soul, a reflection of his animus, or arguably, just a slip of the tongue somehow unrelated to his “true” feelings. If other managers were nearby, they could well have dismissed the overheard comment as an aberration, or it could have created a new norm of conduct for the company, an atmosphere of impunity. The point is that the inference to be given the remark should not be made by judges, particularly judges who have not heard the entire story.40  

The stray-remarks doctrine arose out of Justice O’Connor’s concurring opinion in Price Waterhouse v. Hopkins.41 The majority described a mixed-motive burden-shifting approach in Title VII cases.42 When a plaintiff alleges that an adverse employment action was taken for both a discriminatory reason and a legitimate reason, the plaintiff bears the burden of showing that the discriminatory reason was a motivating factor. But once the plaintiff does so, the burden of persuasion shifts to the employer to demonstrate that he or she would have made the same decision if the discriminatory reason was not considered. The concurring opinion argued for limiting the approach to cases where there was direct evidence of discrimination43 (a limitation that no longer applies44). Although Justice O’Connor noted that certain statements “by nondecisionmakers, or statements by decisionmakers unrelated to the decisional process itself” are not “direct evidence” of discrimination,45 she did not say that such remarks were never evidence of discrimination or never probative of discriminatory animus. Still, her statement birthed countless cases about stray remarks—a classic example of Losers’ Rules—that distort the original concept. Some courts have even suggested that stray remarks are not only insufficient in and of themselves to prove discrimination, but also may be inadmissible under Federal Rule of Evidence 403 because their unfair prejudicial impact substantially outweighs their probative value.46

There are striking examples of courts dismissing extraordinary statements of bias as stray remarks. In Shorter v. ICG Holdings, Inc.,47 a supervisor’s derogatory racial slur about the plaintiff was treated as a stray remark.48 The court wholly discounted that comment even though it was directed at the plaintiff and was made by the decisionmaker in her case, finding, incredibly to be sure, that nothing linked them to the decision to terminate her.49 They constituted “only” the speaker’s “personal opinion.”50 Even if the court in Shorter had found that those statements were ambiguous—itself a shocking conclusion—questions of ambiguity are not remotely suitable for a judge to resolve on summary judgment.51

More employment heuristics—more Losers’ Rules—have evolved, the net effect of which has been to substantially lighten the employer’s burden of proof and make summary judgment in his or her favor increasingly likely. There is the “honest belief” doctrine—which starts from the notion that the employer’s belief was wrong, not based on the record, or foolish, but was “honest,”52 or at least “honestly described.”53 This approach takes even a pretextual reason offered by an employer out of the discrimination calculus. Typically, when an employer’s reason for an adverse employment decision is not supported by the record, the plaintiff can argue that it suggests he is covering up his unlawful bias. Under this doctrine, however baseless the employer’s explanation may be, that element is trumped by the employer’s “honest belief,” and a judge, not a jury, may make such a determination on summary judgment.

There are also doctrines that require the comparator in the discrimination case to be “nearly identical” to the plaintiff.54 In a discriminatory discharge case, for example, the plaintiff has to show that the misconduct for which he or she was fired was virtually identical to the misconduct of another employee outside the protected class.55 It is not enough to show that that the nonminority employee’s conduct for which he escaped discipline was similar, comparable, or, indeed, more serious.56

There are “rules” that turn the employment laws on their head, in which the court refuses to second-guess an employer’s decision and defers to its business judgment.57 The employment laws necessarily require second-guessing an employer when the plaintiff makes out a prima facie case, or, at the very minimum, the laws permit a jury to do so.

There are also Losers’ Rules that are based on flawed sociological judgments about how discrimination occurs, and in what setting. Some courts have concluded that it is unreasonable to believe that the same decisionmaker who hired the plaintiff would later discriminate against him or her.58 That conclusion assumes that bias will be reflected in every decision of an employer—that it cannot be repressed at, for example, the hiring stage and surface as the individual starts work.59 Once the decisionmaker hires someone in the protected class, his or her later dealings with that employee are somehow immunized.60


The list of “rules” provides a blueprint for the judge to grant the defendant summary judgment or to dismiss the complaint. Courts recite these “rules” in case after case, without regard to context, without examination, like the child’s game of telephone. A complex social phenomenon—discrimination—is disaggregated, or “sliced and diced,”61 into unrecognizable and sometimes unintelligible boxes, determined by the judge, not the jury, with predictable results. And, as did the judges responding to the questioner in Massachusetts, the judge here truly believes that he or she is “just following the rules” in dismissing the claim.

What to do about it? First, the problem has to be named: judges have made rules that have effectively gutted Title VII. These rules are not required by the statute, its legislative history, or the purposes of the Act. Second, the problem has to be addressed directly. Congress could, for example, amend Title VII to make its prohibitions more explicit. Alternatively, judicial-education programs can train judges not on how to “get rid” of these cases, but rather on how to analyze the merits in a way that privileges jury decisionmaking and reminds the decisionmakers what the law was designed to reform in the first place. And finally, to address the asymmetry, courts must write decisions if only to show what counts as discrimination, and not simply what does not.62


Nancy Gertner is a professor at Harvard Law School and a former United States District Judge for the District of Massachusetts.


Preferred citation: Nancy Gertner, Losers’ Rules, 122 Yale L.J. Online 109 (2012), http://yalelawjournal.org/2012/10/16/gertner.html.