I. Introduction

Among the plethora of pervasive and long-obscured women’s issues suddenly exposed by the #MeToo Movement, nondisclosure agreements (“NDAs”) have taken center stage. Indeed, it was the public revelation of Harvey Weinstein’s unchecked behavior and NDA protections that thrust the #MeToo Movement into the national limelight.1 Both the Weinstein2 and Stormy Daniels3 stories perfectly illustrate the reasons behind such intense public concern. Wealthy, sophisticated parties can bind financially vulnerable victims to private avenues of relief that are unlikely to remedy the harm caused and likely to perpetuate patterns of sexual harassment and abuse.4 To anyone unfamiliar with American law, the idea that courts enforce these agreements smacks of blatant injustice. Even many familiar with American law strongly believe that something ought to change.

The #MeToo Movement has made demand for reform more prevalent than ever. Empowered by their newfound public solidarity, some women have begun to breach the terms of their NDAs for the sake of the Movement and their own reparation.5 Legislators have scrambled to introduce new laws that either protect sexual harassment victims seeking to disclose information or prevent employers from hiding these stories from the public eye. As of February 2019, seven states have adopted bills that impose significant restrictions upon NDAs in the context of sexual harassment.6 In February 2018, the state attorneys general unanimously submitted a letter to Congress seeking a ban on arbitration agreements as pertained to workplace sexual harassment claims because of the “veil of secrecy” created by private arbitration and their related NDAs.7

As for the judiciary, Judge William Young of the District of Massachusetts cited #MeToo while radically rejecting a corporation’s “C” plea—a guilty plea where the prosecutor and the accused agree on punishment and avoid a jury—in favor of a public trial.8 He wrote:

Face it, if used in strong cases the “C” plea delegitimizes the central role of the trial judge. Any injustice rankles Americans, systemic injustice rankles them profoundly. Those of us who occupy the constitutional offices of the United States—in whatever branch we serve—must humbly acknowledge that there exists in America today a deep and pervasive sense of injustice.9

By 2017, #MeToo was already showing influence upon Judge Young’s decision-making in a case involving the systemic advantages granted to corporate defendants.10 It seems likely that other judges share his sentiment.11

It is surprising, then, that discussion on this topic has been mostly confined to popular media sources like the New Yorker12 and the Atlantic.13 Of course, legal scholars have denounced sexual harassment NDAs and proposed some legislative restrictions.14 However, none have investigated the substantive law that currently applies to NDAs in the context of sexual harassment, nor have they considered what the law implies for future legislation and whether it is likely to change. Part II of this Comment briefly reviews why the current legal regime fails to adequately protect future victims of workplace sexual harassment. Parts III and IV examine the judiciary’s present treatment of NDAs and the underlying rationale behind that treatment, respectively. Part V weighs the challenges to reform imposed by current case law against the advantages of new legislation. Ultimately, this Comment argues that the intersection between nondisclosure and workplace sexual harassment presents thorny legal issues that are better resolved by legislation than by judicial reform.

II.    The Need for Reform

The purpose of this Comment is not to determine what specific NDA regulations will serve the best interests of workplace sexual harassment victims. Other scholars have already taken that question to task, and there is increasing agreement that moderate—but not severe—restrictions should be imposed.15 On the one hand, NDAs are intuitively favorable for employers, since reputational damage and legal costs could easily outweigh the costs of settlement with the victim.16 On the other, NDAs may also provide many advantages for victims themselves, such as a greater likelihood of compensation from their employer, a quicker end to the painful ordeal, and greater privacy given lingering stigmas surrounding workplace victims.17 Therefore, laws that ban sexual harassment NDAs outright would deny victims a significant opportunity for confidential reparation. The debate continues, however, as to what these regulations should encompass and how far they should go to prevent such cases.18

Regardless, a majority of scholars have reached a powerful consensus: reform is necessary in order to prevent employers from crafting NDAs that do not amply compensate the victim or that enable perpetrators to continue patterns of sexual harassment and assault.19 While some settlement agreements boast six to seven-figure payoffs,20 others offer far less for similar harm;21 and even for victims who receive large payments, money cannot fix their trauma.22 Meanwhile, the stringent confidentiality of unfettered NDAs has enabled powerful individuals to avoid detection, both by other employees and the Equal Employment Opportunity Commission (the EEOC), thus opening the possibility for decades of abuse across multiple victims.23 Scholarly consensus and the plethora of victims’ jarring stories collectively illustrate the need for change.

III. Relevant Case Law

Before delving into the law that governs sexual harassment NDAs, it is important to note that provisions that obligate nondisclosure go by a few other names, such as “non-assistance” or “confidentiality” provisions. For the purposes of this Comment, an “NDA” refers to any agreement that prevents one party from disclosing some type of information to a third party. Moreover, NDAs are difficult to examine in a vacuum. This is to say: they rarely come about on their own. As the case law will show, the vast majority of NDAs (or at least the ones that generate litigation) occur as components of three sources: settlements, employment contracts, and arbitration agreements that arise from those contracts. Legal rules that take aim at these sources will necessarily have an impact on their nondisclosure provisions, and vice versa. This poses a challenge to this Comment’s analysis because there are some aspects of the three sources beyond standalone NDAs that are thematically similar to NDAs but outside of the scope of this Comment’s inquiry, e.g., arbitration procedures or for-cause termination. For the sake of simplicity, therefore, this Comment will discuss NDAs that emerge from different sources interchangeably. In the rare case where judge-made rules or legislation could have different effects on different forms of NDAs, this Comment will address those differences expressly.

A.    Unconscionability and Public Policy Doctrines

Judicial solutions to the threats posed by sexual harassment NDAs rely on two key components of contract law: the “unconscionability” and “public policy” doctrines. The former dictates that a judge may refuse to enforce a contract’s terms if she finds them to be “unconscionable” or otherwise modify them in order to prevent such a finding.24 While what constitutes “unconscionable” varies somewhat from state to state, the D.C. Circuit has advanced the most popular definition: the “absence of meaningful choice on the part of one of the parties together with contract terms which are unreasonably favorable to the other party.”25 Some jurisdictions hold that the “procedural” and “substantive” factors of this definition—absence of choice and one-sided terms respectively—must both be present for a court to render a term unconscionable.26 Within this group, some courts apply a “sliding scale:” the more procedurally unconscionable a term is, the less substantive unconscionability is required for a court to rescind that term, and vice versa.27 Other jurisdictions, in contrast, may render a contract term unconscionable even if the plaintiff only alleges procedural or substantive unconscionability.28

In order to determine whether a contract or term is procedurally unconscionable, a court considers factors such as disparate bargaining power between the parties, differences in business sophistication, representation by counsel, ambiguous or misleading contract language, and economic, social, or practical duress that may have compelled a party to execute the contract against her best interests.29 Contracts of adhesion—the often labyrinthine standard-form contracts typically provided by larger companies—may be procedurally unconscionable per se (though many courts must still find substantive unconscionability in the terms themselves).30 A term is substantively unconscionable where it is so “one-sided or overly harsh” or “exceedingly calloused” so as to be “shocking to the conscience.”31 Terms are not substantively unconscionable simply because they are much more favorable to one side; their one-sidedness ought to suggest that “the weaker party had no meaningful, no real alternative, or did not in fact assent or appear to assent to the unfair terms.”32 In summary, the unconscionability doctrine is justified by and applied according to the fundamental public policy principle that there could be no real freedom of contract if courts had to enforce contracts that boasted neither mutual interest nor mutual assent.

The public policy doctrine—in many ways the sibling of unconscionability—states that a term of an agreement is unenforceable if either (1) “legislation provides that it is unenforceable” or (2) “the interest in its enforcement is clearly outweighed in the circumstances by a public policy against the enforcement of such terms.”33 Courts consider a variety of factors when weighing enforcement interests, such as the parties’ justified expectations and the negative consequences of non-enforcement.34 As for weighing public policy interests, courts consider factors such as legislative precedent, the relation of the contract to the policy interest, and the seriousness of enforcement’s consequences as they pertain to public policy.35 Accordingly, there is a large variety of terms that a court may render void as against public policy. The most paradigmatic cases involve agreements to commit illegal acts,36 but the implicated action does not need to be explicitly illegal for a court to render it void.37

Exceptions and inconsistencies abound within and between jurisdictions, such that an attempt to map the public policy doctrine could quickly overwhelm the main focus of this Comment. Therefore, it must suffice to say that sexual harassment NDAs may be void as against public policy, but only in the specific circumstances that this Comment is about to address. Perhaps unconscionability plays a role in some of the forthcoming cases, but assuming most legal departments are savvy enough to avoid the “shocking”38 behavior necessary to render a contract unconscionable, it remains a secondary consideration for the purposes of this Comment’s investigation and analysis. The main problem that this Comment seeks to address emerges when the employer’s careful and sophisticated lawyers intervene in order to contain the damage that the perpetrator has already caused.39

B.    Voiding NDAs to Report Criminal Conduct

Given the obvious confidentiality of most sexual harassment NDAs and the often-steep penalties imposed for violating them, case law on the topic is quite limited.40 However, courts have provided enough for one to determine the extreme ends of whether these NDAs are unenforceable contracts as a matter of public policy. For one, it is clear that courts will not enforce NDAs that bar an individual “from providing relevant evidence regarding past allegations of sexual abuse in a domestic violence proceeding” or from “reporting future criminal behavior to a court.”41 Indeed, any NDA that “purport[s] to suppress information concerning the commission of felonies” is “illegal per se.”42 This also includes agreements that create a failure of one’s duty to disclose to a third party, involve a breach of contract with another party, or otherwise cause wrongful non-performance.43 However, “[t]he mere possibility that an employer could use a [nondisclosure] clause to hide illegal activity is . . . insufficient to void the clause on grounds of public policy.”44 Instead, courts “subject” NDA agreements to “implied exceptions for public policy purposes,” so that the agreements are enforceable unless they are used to prevent the disclosure of illegal activity to a court.45 As for the prevention of disclosure to non-governmental sources—even if there is potentially unlawful activity involved—such agreements are typically legitimate.46 This includes attempts to provide testimony in a third party’s civil claim against one’s employer, even though that disclosure would take place before a judge.47

In other words, if a victim has enough evidence that a felony has been committed by her harasser or abuser, she may report the felony, even when an NDA purports to bar her from such a report. However, as the #MeToo Movement has revealed, this is a big “if.” First, as damaging as sexual harassment can be, harassment alone rarely, if ever, qualifies as a felony under state law.48 Second, there are many disincentives for a victim to remedy her situation by means of a criminal case. For example, she may fear retaliation from other superiors at her corporation or affiliates of the accused, and criminal law does not necessarily supply her with any monetary remedy.49 Third, criminal law’s high standard of guilt—beyond a reasonable doubt—makes it incredibly difficult for a private individual to mount evidence sufficient to prevail at trial. This is especially true for sexual harassment and abuse, where perpetrators can often cover their tracks, and witnesses have personal incentives (including their own settlements) to keep quiet.50 Finally, criminal prosecution takes control of the case away from the claimant and places it in the hands of a prosecutor, whose central duty is to punish the perpetrator rather than ensure maximum recovery for the victim.51 In sum, reporting even expressly illegal activity is a difficult hurdle to overcome for an inherently nebulous transgression.

The EEOC works to ameliorate these barriers by institutionalizing some investigations;52 whistleblowers who violate their NDAs at the behest of the EEOC are protected.53 And yet, the EEOC cannot be everywhere at once, nor is every employee willing to take on a whistleblowing role, especially when employers compel internal disclosure first.54 Moreover, as mentioned above, the EEOC has little justification to launch a full investigation if it has not received any reports because the victims of a serial harasser are unanimously restrained by NDAs.55 Therefore, we must consider the options available outside of criminal law to individuals who seek to bypass their NDAs for the sake of obtaining a remedy from a civil suit or retribution by public disclosure.56

C.    NDAs and Self-Enforcement

EEOC v. Astra U.S.A., Inc.57 provides the leading case for unenforceability when criminal activity has yet to be established.58 In Astra, the EEOC brought a preliminary injunction against Astra USA, Inc., in order to prevent Astra from entering into or enforcing settlement agreements that prohibited employees from filing sexual harassment claims with the EEOC and assisting the EEOC in its investigation.59 Multiple Astra employees had communicated to the EEOC that they could not divulge any information because of such agreements.60 Indeed, only twenty-six out of ninety contacted employees even replied to the EEOC’s requests.61 In order to determine whether Astra’s NDAs were unenforceable as against public policy, the First Circuit applied a “bedrock” balancing test: “[A] promise is unenforceable if the interest in its enforcement is outweighed in the circumstances by a public policy harmed by enforcement of the agreement.”62 Specifically, the court weighed the impact of the NDAs on the EEOC’s enforcement of Title VII against the impact that outlawing those NDAs would have on private dispute resolution.63

According to this test, Astra quickly became an open-and-closed case. The EEOC is a public body empowered by Congress, and “it is crucial that the Commission’s ability to investigate charges of systemic discrimination not be impaired.”64 Furthermore, as the court found significant, the EEOC does not only benefit private parties; its investigations are designed to advance the public interest first and foremost.65 Even a “sprinkling” of prohibitions that “materially interfere[ ]” with communications between an employee and the Commission harms public interest.66 The court did recognize that public policy “strongly favors encouraging voluntary settlement,” but it ultimately found that unenforceability would not (1) create a “substantial disincentive” to settlement, (2) promote further litigation, or (3) disturb the finality of already-negotiated settlements.67 Read plainly, the First Circuit’s finding here is puzzling: confidentiality is a major incentive for corporations to privately resolve employee disputes. While arbitration has other advantages as well, public disclosure was clearly one of Astra, Inc.’s greatest concerns when it wrote the NDAs at issue.68 Therefore, it is unclear why the Astra court thought that nullifying a central provision of those agreements would not disincentivize settlement between Astra and its employees; if the nondisclosure provisions were unimportant, why would Astra fight for them in court?

If one were to read Astra as saying that any nondisclosure component of an arbitration agreement was non-substantial in terms of incentives for private settlement, one could expect public policy challenges against NDAs to be somewhat successful. However, the case law probably leans in the opposite direction, at least as it applies to private plaintiffs. For example, in Saini v. International Game Technology,69 the Nevada District Court applied Astra’s balancing test and determined that an employee’s confidentiality agreement was valid even though it would prevent employees from providing testimony that International Game Technology (IGT) had purposefully sold refurbished and defective gambling machines as new products.70 The court distinguished its case from Astra on multiple grounds. For one, the protection of trade secrets, which was the supposed purpose of IGT’s agreement, was more significant than the concealment of sexual harassment complaints.71 Second, the court found that “public policy [in uncovering the sale of defective products] is not as high a priority as enforcement of sexual harassment law by the EEOC, at least when, as here, the defect at issue is not a threat to the safety or economic well-being of the public at large.”72 Third, IGT’s terms were part of a “standard agreement;” they were not “specifically designed to stifle evidence of wrongdoing.”73 Finally, the court found it significant that the employee—instead of a regulatory agency—had moved to void the NDA and thereby “act as decisionmaker about what information IGT does and does not have a legitimate confidentiality interest in.”74 An employee’s belief that an NDA is concealing illegal activity is not enough.75

At first glance, Saini appears promising for individual victims attempting to void their employer’s NDAs. Indeed, the court asserted that the concealment of workplace sexual harassment in Astra was at least more significant than the dangers of concealing product defects. However, a closer reading of the opinion reveals that the court significantly limited its consideration of sexual harassment to the context of EEOC investigations. Namely, it specified its consideration of public policy concerns in Astra as the “enforcement of sexual harassment law by the EEOC.”76 It did not weigh a public interest in encouraging enforcement by private individuals—nor, as pertains more closely to #MeToo, the attempts of private individuals to seek better remedies in court or better retribution in the public eye. The court’s reservation is only made clearer by its final distinction that the plaintiff was a private individual and not a regulatory agency.77

The Saini court’s limitation of Astra’s findings to public agency enforcement is consistent with the language and underlying rationale of Astra itself. As noted above, the First Circuit considered the NDAs in question to be harmful to the EEOC’s own actions, not unjust to the actual victims involved. Moreover, it felt compelled to expressly recognize the otherwise intuitive notion that the EEOC’s actions benefit the public as a whole, not just “private” victims. That court did not, on the other hand, discuss whether private challenges to unfair NDAs can also similarly benefit the public, especially when harassment is too subtle for the EEOC to root out. Perhaps this explains why the First Circuit found that a preliminary injunction would not undermine prior settlements or discourage future ones: its holding really only applied to Astra’s relationship with the EEOC. Astra and Saini indicate there is a significant public interest in preventing workplace sexual harassment, but on the whole, that interest only overcomes the public’s interest in freedom of contract when it is the government who seeks to advance it.

1.    EEOC investigations

Surrounding case law supports the interpretation that government investigation / private dispute is the most significant predictor for determining whether a sexual harassment NDA is void as against public policy. As regards invalidated agreements, there are plenty of cases in which courts have refused to enforce agreements that are likely “to chill employees’ participation in legitimate investigations.”78 Given this Comment’s discussion so far, “legitimate” may very well be a stand-in for “government-led.” In EEOC v. International Profit Associates,79 the Northern District of Illinois invalidated a severance agreement that prohibited an employee from “disclosing anything relating to his employment . . . except as may be necessary in response to lawful process of any judicial or adjudicative authority or otherwise allowed by law.”80 Here, as in Astra, the court narrowly confined its ruling on the contract as “inapplicable to communications with the EEOC.”81 Such a limitation strongly invokes the “implied” public policy “exceptions” that allow reporting of criminal activity when the reporting party is bound by an NDA.82

In EEOC v. Morgan Stanley & Co.,83 the Southern District of New York invalidated a non-assistance agreement that required employees to notify their direct supervisor and the “Law” or “Compliance” departments before communicating with the EEOC.84 Even though the contract’s language only threatened the possibility of demotion or termination if an employee failed to notify the appropriate officers, the court found that this language was sufficient to frustrate the EEOC’s investigation.85 As in Astra, the agreement sported at least a “sprinkling” of prohibitions that “materially interfered” with the EEOC’s efforts.86 It seems reasonable that Morgan Stanley would want its Law and Compliance departments to stay well-informed about the EEOC’s interactions with its employees, so perhaps the court erred in denouncing the company’s prudence. Regardless, this court’s decision illustrates the judiciary’s eagerness to protect and advance the government’s own attempts to curb workplace sexual harassment and discrimination. Courts have therefore provided the EEOC with a great boon in enabling it to bypass corporate NDAs for the sake of its investigations. However, and again, the Commission does not have the resources to investigate behind every closed door.

2.    Private attempts to invalidate NDAs

The #MeToo Movement has made it clear: private individuals must also come forward with their own stories in order to effect meaningful social change.87 And yet, despite the judiciary’s eagerness to assist the EEOC, it has been less charitable with private actors who attempt to invalidate NDAs themselves. As the court proclaimed in Saini, it is not within the individual’s “prerogative” to determine for herself what type of information is “legitimately” confidential.88

For example, in Perricone v. Perricone,89 the Supreme Court of Connecticut enforced a divorce settlement NDA that prevented a celebrity-doctor’s ex-wife from divulging his purportedly condemnable behavior to popular public media sources.90 That court went a little further, interpreting Astra as a case purely concerned with communications to public enforcement agencies and thereby writing sexual harassment out of the script.91 It also limited voidable nondisclosure provisions to five main categories: terms that (1) restrict “the right to speak on matters of public concern” (referring to voting in public elections), (2) restrict communication with a public agency regarding civil rights law enforcement (Astra), (3) “require[ ] the suppression of criminal behavior,” (4) suppress information important to public health and safety, or (5) impose confidentiality for the benefit of a “public entity or official.”92 Taken literally, some of these categories should have helped those seeking to invalidate their NDAs. However, the Connecticut Supreme Court also urged that the public policy doctrine “should be applied with caution and only in cases plainly within the reasons on which that doctrine rests . . . .”93 Therefore, a term may be enforced even if it could potentially fall into one or more of these categories.94

Bandera v. City of Quincy95 provides an example of the judiciary’s hesitancy to apply the public policy doctrine in private disputes even when multiple Perricone categories may be implicated.96 In Bandera, the former director of Quincy’s Community Policing Commission attempted to invalidate a settlement agreement after she was allegedly coerced into it by her previous attorney.97 According to her trial testimony, for which the jury awarded her $135,000 dollars, Bandera had been excluded from meetings, ridiculed on the basis of her sex, and subjected to graphic stories of the sexual exploits of male officers.98 She was fired by the mayor of Quincy soon after bringing her complaint.99 Meanwhile, even the plain language of the settlement appeared suspicious. It demanded a general release of all claims and the imposition of non-admission and non-disclosure agreements concerning her claim in exchange for her then-attorney’s full compensation and a promise that the mayor would recommend Bandera for a permanent substitute teaching position at the city’s public school.100

The issues of unconscionability in this case are beyond the scope of this Comment.101 Nonetheless, the facts of Bandera raise various concerns along the lines provided by Perricone. The settlement and its NDA were unduly favorable to the city and its public officials, the potential criminal activities of the city and its officers were probably a matter of public concern, and the settlement was possibly designed around concealing those potentially criminal activities.102 Despite these potential red flags, the First Circuit found that the settlement’s provisions were not void as a matter of public policy, citing Astra but refraining from entering into a public policy balancing test.103

Whatever the import of the First Circuit’s refrain, Bandera compellingly illustrates the extra-legal difficulties facing private individuals who pursue litigation. Even if they are not barred by a boilerplate arbitration agreement, their employers typically have ample resources to compel settlement against the victim’s best interests and return the transgressions at issue to a state of nondisclosure. Any litigation is exhausting, resource-intensive, and full of uncertainty, especially for non-lawyers.104 This is doubly so in the case of a sexual harassment claim, which requires that the victim relive her harm at every court proceeding. The same does not apply to the well-informed and well-equipped employer. In Bandera, the First Circuit merely stayed the judgment in order for the district court to determine whether the settlement agreement had been properly formed.105 However, the case did not make it that far: Bandera, who then proceeded pro se, capitulated to the earlier settlement agreement, even though the jury’s award would have easily covered her legal fees and far exceeded her prospective income as a substitute teacher.106 Only Bandera’s psychological fatigue and lack of trustworthy legal resources could satisfactorily explain why she submitted to such a bad deal.

D.   Summary of Findings

This Comment has so far attempted to carve out a body of case law applicable to NDAs that purport to prevent victims of sexual harassment from divulging their stories to the public. Not every case that this Comment has examined is expressly linked by precedent. Nor does each case tackle the problem of sexual harassment NDAs head-on. Nonetheless, they all speak as one in providing the lay of the land: Courts are extremely hesitant to render an NDA unenforceable unless (1) the alleged harasser is under investigation by the EEOC or (2) the victim has strong evidence that the agreement was intended to prevent her from reporting a crime, of which there is also substantial proof.107 With this review of relevant case law complete, this Comment could conclude. However, in order to understand why courts have decided upon these doctrines and how they may change after #MeToo, one ought to take a closer look at the “bedrock” upon which Astra and its sister cases relied.

IV.   The Judiciary’s Underlying Rationale

The analysis above raises a crucial question: Why do courts draw the line for NDA invalidation between public and private enforcement, when such a line clearly leaves some harassment victims vulnerable to further harm? True, the foremost duty of the courts is to apply the law, not to make ad hoc moral judgments. However, in the case of the common law public policy doctrine, courts have and will continue to actively shape the legal regime that applies to sexual harassment NDAs. Their development of that regime can thus be evaluated for the purposes of determining its efficacy and the room it leaves for reform.

Therefore, the following sections of this Comment take a closer look at the underlying principles of the public policy doctrine and the specific factors that courts consider in weighing public policy in the sexual harassment context. Such an examination is an essential foundation for weighing the viability of judicial reform. Ultimately, this part of the Comment will find that even though the current judicial regime does not optimally advance the interests of sexual harassment victims, its approach is motivated by valid concerns for efficiency and accuracy in distributing justice.

A.    The Public Policy Doctrine’s Underlying Rationale

The line between public and private enforcement can be explained by the judiciary’s emphasis on the wrongdoer rather than the victim in public policy cases. Voiding a contract despite the universal public interest in freedom of contract is a significant deployment of a court’s authority. Were a court to do so in order to benefit an individual, it would raise major concerns about abusing its discretion, especially if that private benefit imposed a detriment upon the public overall (such as diminishing the binding power of private agreements). Therefore, as Dworkin and Callahan’s article, Buying Silence, elaborates, courts limit their application of the public policy doctrine to cases in which a bad actor’s behavior should not be enforced:

In general, “freedom of contract” is manifested by the courts’ disinclination to evaluate the substance of agreements between private parties. There are cases, however, in which this principle is superseded by societal interests, and a court may decline to enforce a contract term on public policy grounds. The introductory note to the Restatement (Second) of Contracts states that “reluctance to aid the promisee rather than solicitude for the promisor” provides the basis for such a determination. Refusals of enforcement on these grounds are designed to deter misconduct and to avoid using the courts as instrumentalities of questionable activity. 108

Voiding a term mainly to benefit a sympathetic individual risks injuring the public by undermining freedom of contract. However, punishing a bad actor by refusing to enforce his term advances public interest overall (whether because the actor was trying to get away with a crime or because he was clearly abusing contract law, among other reasons). Still, it may be difficult in practice for a court to draw the line between the two—that is, between an employer-favorable contract policy and a behavior so harmful to public policy that it must be rendered void. As suggested Part II, that inquiry becomes easier when other agencies have already addressed it themselves.

As Buying Silence points out, Astra provides a great example of the courts’ emphasis on refraining from assisting the promisee. If a government agency complains that a company is likely concealing unlawful action by means of NDAs, the court has good reason to take that agency at its word and void such agreements. It is not just that the agreements are wrongly concealing unflattering information; they represent a company’s unwillingness to submit itself to rule of law. Thus, even though the EEOC brought a civil injunction, the First Circuit’s refusal to enforce the NDAs was just a simple extension of the legal rationale that there is no “freedom of contract” interest when contracts are used to break the law.109

Nonetheless, Buying Silence does not address a crucial feature of both the Restatement and Astra’s decision: both expressly refuse to consider the “solicitude of the promisor” in determining whether an agreement is unenforceable for the sake of public policy. As already explained, Astra emphasized the EEOC’s function as “vindicat[ing] the public interest” rather than only serving private parties.110 The matter at hand was not a battle between remedying an employee’s sexual harassment and protecting a company’s right to enforce its agreements; it was a battle between a company’s (unnecessarily strict) private dealings and an agency’s enforcement responsibilities as laid down by Congress itself.111 What is more, the factors that the First Circuit weighed in its balancing test were even larger than the most basic stakes in the case. Namely, while the court considered individual impact on Astra for the sake of the injunction, as for its balancing test, it did not weigh the impact that voiding the NDAs would have on Astra’s relationships with its employees. Instead, it weighed “the impact [of its decision] on private dispute resolution” overall.112 In other words, the court did not weigh the impact of enforcement on the EEOC’s specific investigation against Astra; it weighed the impact that these types of agreements would have on all EEOC efforts if the court decided to enforce them.113 Nowhere in the opinion is there a speculation on how a ruling for one side or the other might affect the actual victims in the case.

B.    The Courts’ Current Approach Is at Odds with #MeToo’s Principles

The main thrusts of #MeToo and its sister movements114 urge us to focus on the exact same factors that courts leave out in public policy considerations, such as the harm to the victim, the power dynamics between victim and harasser that exacerbated the harm, and the remedy necessary not only to deter the harasser but more importantly to make the victim whole again.115 With these #MeToo factors in mind, the courts that this Comment examined could have inquired as to whether invalidating the NDA was necessary to ensure that the victim finds adequate relief. Under such a regime, companies could still have faith in the legitimacy of their employment contracts, so long as they provided their employees with generous arbitration terms and a thorough system for deterrence and accountability. However, the regime must be dealt with as it stands, not as how it should be.

This is not to say that there was no room in the doctrine for Astra to have considered the position of sexual harassment victims. The Restatement (Second) of Contracts provides four factors for a court to consider in determining whether a contract should be void as against public policy:

(a) the strength of that policy as manifested by legislation or judicial decisions;

(b) the likelihood that a refusal to enforce the term will further that policy;

(c) the seriousness of any misconduct involved and the extent to which it was deliberate; and

(d) the directness of the connection between that misconduct and the term.116

The First Circuit’s opinion explicitly involved considerations of factors (a) and (b), but factors (c) and (d) had no explicit role to play. If they did, as regards to (c), one might expect the court to have delved into EEOC’s investigation to see the degree of potential harassment at Astra. Alternatively, the court could have expressly assumed that (c) was satisfied because the possible activity raised to the level of gravity as to attract EEOC’s investigation in the first place. Either way, the stories of individual victims at Astra could have played a role in determining the degree of misconduct. As for (d), plenty in the contract suggested an intention to cover up misconduct. The court even noted (but again did not explicitly weigh) the concerning silence with which the EEOC’s investigation was met by numerous Astra employees.

There is evidence that other courts have externally considered factors (c) and (d). For example, in Saini, the District of Nevada court noted that IGT’s contracts were standard and not designed to cover up consumer fraud.117 Such a particular observation about the case at hand has little connection to the nationwide scope of Astra’s test, which Saini expressly followed. Furthermore, courts may focus on (c) and (d) in refusing to enforce NDAs that conceal illegal activity. However, this is likely because (a) and (b) are already easily satisfied: (a) Enabling crime is certainly against legal precedent and (b) allowing reports of that crime would surely help prevent it.118 This Comment will address whether the weighing of #MeToo factors is actually workable. For now, suffice it to show that the court in Astra and other courts have decided to orient their opinions towards broader questions of public interests rather than fact-intensive issues of individual harm.

From this—in addition to the language from other opinions that suggests likewise—one may gather that courts are more than comfortable recognizing the silent plight of workplace sexual harassment victims, but they have clearly delegated to the EEOC the question of how and when to protect those victims. In other words, the judiciary’s current approach refrains from determining how to empower those victims to protect themselves.119 One can only speculate as to why each court unanimously adhered to this particular construction of the public policy doctrine. On the one hand, they have a general obligation to follow precedent, but on the other, there was clearly room to branch out without totally diverging from the Restatement. Regardless, armed with the findings from Part III of this Comment, some speculation should prove productive.

C.    The Judiciary’s Current Approach Is Justified

As this Comment has argued, the best predictor for determining the validity of an NDA is whether the party that is seeking to invalidate the agreement is a government agency or a private individual. Drawing this line keeps the court’s determination simple, because, as in Astra, it makes satisfying the Restatement’s first two factors easy: (a) Congress clearly supports the EEOC’s Title VII investigations as a matter of public policy over a general desire to encourage freedom of contract, and (b) allowing companies to enforce their NDAs despite an EEOC investigation would severely frustrate the EEOC’s efforts and therefore Congress’s wishes. Again, such a ruling is merely an extension of the fundamental rationale that the public’s interest in law enforcement is prime.120 And if a court is concerned with factors (c) and (d)—even if they play no explicit part in the balancing test—the EEOC has done much of that consideration for them. From the court’s perspective: (c) the misconduct is serious because the EEOC is investigating it, and (d) there is likely a direct connection between the agreements and the misconduct, because the EEOC is asking the court to invalidate them.

Private attempts at invalidating NDAs would require a far more difficult deliberation before a court could justify a decision not to enforce the agreement. There is no universal piece of legislation proving that Congress or state legislatures desire private individuals to violate their NDAs in spite of the freedom of contract interest.121 Even if courts could comfortably determine that legislative bodies would want the victim to come forward in its case, it would then have to decide whether that coming-forward would actually advance, say, the public’s interest in having harassment-free workplaces. By the preliminary injunction stage, the court would have to not only estimate whether the victim’s case would be successful, but also whether its success would serve to deter future incidences of sexual harassment. It is far from clear that corporate liability for officer misconduct actually deters those officers from future misconduct.122

Meanwhile, if courts entertained invalidating an NDA at the sole behest of a private plaintiff, they could no longer rely on other agencies’ expert judgment—namely, the EEOC or an attorney general—as to whether alleged misconduct was serious and whether the agreements are connected to that misconduct. A court could compare the victim’s testimony and evidence with that of the company’s own (the dreaded “he-said / she-said” problem). However, as evinced by Bandera, courts are rightfully hesitant to conduct a mini-trial on such issues in order to determine whether or not the case can even proceed.123 Instead, it is simpler to adhere to basic contract principles in these cases rather than commit every NDA breach to intensive litigation before liability can even be established.124 Again, the hope is that the EEOC will protect these victims for whom arbitration or criminal law is insufficient.

V.    Statutory Reform and its Advantages

A.    Existing Case Law Does Not Provide for Effective Reform

A consideration of the many disadvantages posed by reforming the judiciary’s approach makes evident that legislative measures are the best option for improving the state of the law surrounding NDAs in the context of sexual harassment. As this Comment has already examined, the current case law fails to address many of the important interests and intricacies that the #MeToo Movement has brought to light. This is indubitably a major concern. First, however, there are the standard criticisms of case law as a platform for social change, both practical, such as its long time-horizon, and political, such as the needed cooperation from other branches of government.125 Second, there are problems facing judicial reform unique to the context of NDAs and sexual harassment. Sexual harassment is an extremely divisive issue, and not all courts are likely to share Judge Young’s pro-victim sentiments.126 Moreover, a balancing test that considers the victim’s harm in every attempt to invalidate an NDA is not a reasonably workable judicial regime.

As mentioned above, courts are hesitant to resolve complex disputes of fact before it is clear that a plaintiff actually has a valid claim.127 Determining whether an NDA is invalid because it is being used to cover up sexual harassment would require a court to figure out whether or not the victim was actually sexually harassed. Such a determination could take years, and a regime that requires it would enable complainants to unduly burden their employers and the courts. Moreover, even if the courts had the resources to reasonably allow for these private investigations in every case—and corporations had the resources to tolerate them without seriously affecting employment conditions—it is asking them to make some very difficult determinations in every instance.

Under this alternative regime, courts would have to consider:

(1) Was there sexual harassment for which the other party is responsible?

(2) Is an NDA preventing the disclosure of that harassment?

(3) What is the degree of harm?

(4) Are the standard remedies currently available to the plaintiff outside of this court sufficient to remedy that harm?

(5) If not, would invalidating the NDA make a better remedy available—in other words, is the NDA the reason for the victim’s lack of recourse?

(6) If invalidation would aid the victim, and recognizing that there is a default public interest in freedom of contract, would invalidation in this case create a precedent that would advance public interest overall?

Alternatively, courts could abandon the general public policy considerations of the current doctrine, but reinventing the wheel is dangerous when no better framework is apparent.

It is difficult to imagine how requiring these questions would proliferate an efficient legal system. It is even harder to imagine that judges, legal generalists by necessity, are equipped to answer these questions correctly in every case. For these reasons, the status quo is more justified than was first apparent. Namely, it is sound public policy in itself for a court to require that at least some of the questions surrounding a sexual harassment NDA are answered first by agency investigation and criminal prosecution. The tradeoff with this governmental approach, of course, is that it provides limited remedies to victims, even though it is most likely to focus on the most serious incidences of harassment and assault. And as some scholars convincingly argue, the current regulatory scheme may not do enough to affirmatively incentivize victims and witnesses to come forward.128 Nonetheless, because case law does not provide room for a compelling judicial alternative, we must look towards other avenues for reform.

B.    Advantages of Legislation

Legislation is the best avenue: it will better serve sexual harassment victims by preventing harmful NDAs from the outset rather than trying to strike them down retroactively in court. Some scholars have already put thought into what this legislation should look like. For example, in Targeting Repeat Offender NDAs,129 Ian Ayres suggests a “middle ground” that would allow for NDAs that could still protect victims who seek confidentiality but would prevent their validity when used to protect “serial offenders.”130 In contrast, this Comment does not seek to propose the “right” legislation. Instead, as with its consideration of case law, the Comment seeks to address legislation as it stands in order to determine whether legislatures can ameliorate the #MeToo concerns that have been raised throughout this inquiry.

1.    Examples of ongoing statutory reform

California and New York provide instructive case studies, not only because they are influential and commercially powerful states, but also because they are currently leading the way in passing new and ambitious #MeToo laws. Effective as of January 1, 2018, California’s newly amended civil procedure code states:

(a) Notwithstanding any other law, a provision within a settlement agreement that prevents the disclosure of factual information related to the action is prohibited in any civil action the factual foundation for which establishes a cause of action for civil damages for any of the following:

(1) An act that may be prosecuted as a felony sex offense. . . .

(4) An act of sexual assault, as defined in paragraphs (1) to (9), inclusive, of subdivision (e) of Section 15610.63 of the Welfare and Institutions Code, against an elder or dependent adult, as defined in Sections 15610.23 and 15610.27 of the Welfare and Institutions Code.

(b) Notwithstanding any other law, in a civil action described in paragraphs (1) to (4), inclusive, of subdivision (a), a court shall not enter, by stipulation or otherwise, an order that restricts the disclosure of information in a manner that conflicts with subdivision (a).131

Were Saini a case about sexual misconduct, this code’s language may have produced the opposite result. Specifically, the language prohibits the enforcement of any contract that purports to prevent a victim or crucial witness from coming forward in a civil claim. True, the code does not reach so far as sexual harassment, which is not a felony in California, but it would take some clever drafting to somehow specify that an NDA applies to disclosures about sexual harassment but not sexual assault. Bare minimum, such language would put employees and the court on notice that the corporation providing such agreements has sinister intent to conceal its misdeeds.

As discussed above, the connection between the NDA and the harm at bar is at least a tacit consideration during a public policy challenge to a contract provision.132 A court would be more comfortable voiding a contract as against public policy when its plain language suggests that the company aims to evade due accountability.133 Finally, and most importantly, this law remedies an essential problem with the judiciary’s current reliance on the EEOC and prosecutors to root out sexual assault: it ensures that victims can personally recover damages by means of a civil claim once the sexual misconduct has been established in a criminal proceeding.

New York has taken even greater steps, with an improved attention to the position of employee-victims. Its new legislation (Section 5003-b) says:

Notwithstanding any other law to the contrary, for any claim or cause of action, whether arising under common law, equity, or any provision of law, the factual foundation for which involves sexual harassment, in resolving, by agreed judgment, stipulation, decree, agreement to settle, assurance of discontinuance or otherwise, no employer, its officer or employee shall have the authority to include or agree to include in such resolution any term or condition that would prevent the disclosure of the underlying facts and circumstances to the claim or action unless the condition of confidentiality is the plaintiff’s preference. Any such term or condition must be provided to all parties, and the plaintiff shall have twenty-one days to consider such term or condition. If after twenty-one days such term or condition is the plaintiff’s preference, such preference shall be memorialized in an agreement signed by all parties. For a period of at least seven days following the execution of such agreement, the plaintiff may revoke the agreement, and the agreement shall not become effective or be enforceable until such revocation period has expired.134

Like California’s law, Section 5003–b prevents employers from independently establishing NDAs that would prevent their employees from disclosing information about sexual harassment in a civil suit. Inventively, under this statute, an employee can still agree for such a provision to be enforceable, but only after the required deliberation period and in her sole discretion. This certainly ameliorates general concerns about employee bargaining power in the face of corporate giants, while preserving an option for employees who desire confidentiality themselves.135 The language is somewhat unclear, but “underlying facts and circumstances to the claim” leaves open the possibility that the law prevents a company from restraining a victim from disclosing her story outside of court.

2.    Forecasting the impacts of reform

These statutes may be too recent to produce case law examples of their impact on victims’ attempts to invalidate NDAs.136 On the other hand, more than a year has passed since they both became effective.137 Silence itself could speak to their efficacy: a lack of new cases involving sexual harassment NDAs suggests a decrease in frequency for contractual cover-ups that provide insufficient solicitude and reparation for the victim. That is a victory for legislation in itself, even if #MeToo’s cultural shift has so far led the charge.138

Regardless, the future impact of these laws may be projected by imposing them on the cases examined above. For one, a statute in the style of California’s could have deterred the strongly one-sided settlement arrangement that featured prominently in Bandera.139 First, California only requires an act that “may” be prosecuted as a felony sex offense for its code to apply.140 Even if Bandera did not suffer from such an act—in California, the unwanted touching of an intimate area for the purpose of sexual gratification—the threat of such an allegation and the prolonged litigation required to validate it would have left the City of Quincy far more desperate to resolve the case.141 Second, Bandera’s claims would have served as an even stronger warning to the City of Quincy: if any of the City’s employee-perpetrators went further in their actions, the City would have to face a claimant whom they could not silence through settlement. California’s new law not only preserves disclosures to a court in a criminal case; by proactively preventing certain NDAs from being formed, it empowers a potential felony sex offense victim to share the information of her harm as she sees fit. Thus, the City of Quincy would have had stronger incentives to immediately eliminate any risks among its personnel in order to prevent public outrage after the victim’s disclosure, no matter the case’s actual outcome.

New York’s statute, which includes sexual harassment NDAs, would have prevented Bandera’s plight altogether.142 Because it only allows NDAs when they are the “plaintiff’s preference,” Quincy would likely have had to offer more money to Bandera in exchange for her silence. And even if Bandera mistakenly moved to accept the first deal due to bad counsel or ignorance of the law, the court could have invoked the New York code in preventing its signing or invalidating it retroactively. Indeed, the statute might have obliged the court to do so.

Finally, whatever the concrete impacts of these laws will be, they will likely influence how courts resolve attempts to void sexual harassment on public policy grounds. Returning to the Restatement, “the strength of that policy as manifested by legislation . . .” is one of the four main factors that courts consider in determining whether to enforce a challenged term.143 Therefore, a court whose state has a law that restricts sexual harassment NDAs can be more confident in refusing to enforce those agreements than a court whose state lacks them. Moreover, the more states that put such laws in place, the more likely a court will look beyond the laws of its own state when weighing the public policy interest in refusing to enforce an agreement. Were Congress to enact such a law, the effect would be all the more prolific.

C.    Counterarguments

The most immediate rebuttal to this Comment’s argument is that courts will have to change too—statutory reform alone will be insufficient. Indeed, it is challenging for legislatures to craft laws that provide answers for every incident. After all, sexual harassment law has not worked so far, and victims are likely to continue to slip through the cracks.144 However, arguing for an emphasis on legislative reform does not leave court reform by the wayside. In fact, legislative reform is probably the best chance at encouraging the courts to change their own doctrine. As this Comment just observed, the legislative scheme at hand is a primary consideration for courts in determining whether or not a contract is void as a matter of public policy.145 The more that Congress and state legislatures restrict sexual harassment NDAs generally, the more comfortable a court will be in stretching that law a little further to invalidate a contract in a special case. And as many law and economics experts have argued, the law itself has the capacity to alter cultural attitudes, including those of judges.146 In making the legal leap smaller for courts seeking to invalidate a sexual harassment NDA, we can expect those leaps to also become more common.

A critic might also argue that the positive influence of statutory reform upon the courts could be nullified—and some new laws preempted—by pre-existing legislation that favors private settlement, namely, the Federal Arbitration Act (FAA).147 In AT&T Mobility LLC v. Concepcion,148 the Supreme Court held that state laws and judge-made rules could be preempted by the Federal Arbitration Act when they “stand[ ] as an obstacle to the accomplishment and execution of the full purposes and objectives of Congress.”149 Specifically, the Court invalidated a Californian judge-made rule that rendered a consumer arbitration contract of adhesion unconscionable where disputes involve low damages and the weaker party alleges a scheme to defraud.150 It found that rule to be in conflict with the FAA, which “reflect[s] . . . a liberal federal policy favoring arbitration.”151 Hence, a skeptical court could invoke Concepcion in invalidating aggressive statutory reform by finding that an anti-NDA law is so restrictive that it has a “disproportionate impact” on Congress’s larger goal of encouraging (often confidential) arbitration.152 Alternatively, during a public policy inquiry, it could consider the FAA and Congress’s interest in encouraging arbitration as outweighing the influence of state laws that restrict certain sexual harassment NDAs.

First, however, the enforcement of arbitration at issue in Concepcion and the enforcement of sexual harassment NDAs are not wholly entwined issues. For one, arbitration does not always emerge from a harmful, prolific, and intentional act that is solely the fault of the dominant party. Instead, it is often an alternative approach for both parties in a mutual dispute. Moreover, the restriction of a specific contract term in a specific factual context is unlikely to have a material effect on the rate of arbitration overall. While confidentiality is one major advantage of arbitration, it is not an integral feature of it; courts can also impose confidentiality on their proceedings “for good cause.”153 And even if sexual harassment NDA restrictions were sufficient to be materially disruptive, employers still have plenty of incentives to prefer arbitration, including forum consistency, quicker procedure, et cetera.154

Second, lower courts have significantly limited the potential effect of Concepcion by upholding laws that have a far more significant impact on the enforcement of arbitration agreements than restrictions devoted solely to the sexual harassment context. For example, just after the Concepcion decision was handed down, the Massachusetts Supreme Judicial Court upheld an earlier decision that ruled that “an employment contract containing an agreement by the employee to limit or waive any of the rights or remedies conferred by [Massachusetts law] is enforceable only if such an agreement is stated in clear and unmistakable terms.”155 Most significantly, that court recognized an overriding public policy in preventing gender discrimination, even in light of the FAA and Massachusetts’ own pro-arbitration laws.156 It ultimately determined that restrictions on arbitration agreements still kept with the “generous spirit” of the FAA, so long as the parties were “free to agree” on arbitration.157 Such a ruling could save even New York’s strong anti-NDA statute, since that code does not prevent settlement or arbitration of harassment claims outright; the employer merely needs to offer more to the victim in order for her to agree to an NDA.158

Finally, the FAA itself states that arbitration agreements are enforceable “save upon such grounds as exist at law or in equity for the revocation of any contract.”159 The unconscionability doctrine was in place by the time the Act was passed in 1925, so it could not be construed to restrain the courts’ capacity to protect exploited victims.160 Given that the public policy doctrine stems from the same foundational principles as unconscionability, the logic that applies to the FAA and unconscionability should extend in turn.161

In summary, an invalidation of sexual harassment NDA laws via the FAA is unlikely due to their distance from the thrust of Concepcion and the current trend among lower courts of preserving judicial and legislative capacity to regulate employer-employee agreements.

VI.   Conclusion

Sexual harassment is one of the thorniest issues in modern discourse, and every strategy for reform will have its own set of advantages, disadvantages, and uncertainties. Nonetheless, as this Comment has argued, legislative reform is the most promising area of focus. Standing on its own, current case law and judicial procedure is too inflexible to incorporate all of victims’ best interests without severely clogging the courts—and hurting victims as employees in the long run. The legislative approach, in contrast, can both prevent harmful NDAs from being created in the first place and aid victims when they do seek to invalidate them in court. Moreover, legislation not only reflects a cultural change away from the acceptance of harmful sexual behavior, it has the possibility of spreading that change to judicial attitudes as well.

The Supreme Court’s decision in Concepcion may or may not serve as an outer limit for just how restrictive anti-NDA law can become. However, whether Concepcion applies to such laws is still uncertain, and the Court has let stand judicial embrace of some rules that would likely have a greater impact on arbitration overall. Finally, even if Concepcion will curtail some of the most extreme anti-NDA legislation, lesser restrictions can make material contributions to the plights of harassment victims. As this Comment observed in New York’s new code, tweaks to tertiary considerations such as a mandatory negotiation period can still provide a serious boon to underpowered plaintiffs, who typically do not have the resources to quickly understand the import of every legal matter.

Whether the reader accepts any of the preceding arguments, this Comment should serve as an accumulation and evaluation of key cases and statutes that will take center stage as sexual harassment NDA law continues to develop. This new “body” of law is in no way comprehensive, since the Comment had to draw on multiple categories of contract law in order to create a satisfactory picture of the legal landscape facing individuals who wish to breach their NDAs. But that legal landscape does exist, and even the most skeptical reader can appreciate the challenges it poses. The ultimate hope of this Comment is that if the reader takes away nothing else, she can agree on the need for reform.

  • 1See Samantha Schmidt, #MeToo: Harvey Weinstein Case Moves Thousands to Tell Their Own Stories of Abuse, Break Silence, Wash. Post (Oct. 16, 2017), [] (detailing the mass proliferation of #MeToo tweets after the Weinstein story broke).
  • 2See Ronan Farrow, Harvey Weinstein’s Secret Settlements, New Yorker (Nov. 21, 2017), [].
  • 3See Nina Burleigh, Stormy Daniels Defies Trump to Join Chorus of Women Violating Nondisclosure Agreements about Sex, Abuse and Harassment, Newsweek (Mar. 8, 2018), [].
  • 4For a quick overview of the argument, see Ian Ayres, Targeting Repeat Offender NDAs, 71 Stan. L. Rev. Online 76, 79–83 (2018).
  • 5See Burleigh, supra note 3.
  • 6Antone Melton-Meaux, Sex Harassment Settlements: A New Scarlet Letter for Employers?, 75 Bench & B. Minn. 18, 19 (2018); see also NAVEX Global, The Sexual Harassment Legislation Watch List (State by State), JD Supra (Sept. 24, 2018), [].
  • 7See Melton-Meaux, supra note 6, at 19.
  • 8See Fed. R. Civ. P. 11(c)(1)(C).
  • 9United States v. Aegerion Pharm., Inc., 280 F. Supp. 3d 217, 228 (D. Mass. 2017) (citing Sophie Gilbert, The Movement of #MeToo, Atlantic (Oct. 16, 2017), []) (other citations omitted).
  • 10Id.
  • 11Cf. Linda L. Berger, Bridget J. Crawford & Kathryn M. Stanchi, Using Feminist Theory to Advance Equal Justice under Law, 17 Nev. L.J. 539, 539–40 (2017) (noting that judges were among the audience of the U.S. Feminist Judgments Project: Writing the Law conference, where experts discussed how “judges applying feminist perspectives could bring about change in the development of the law . . . .”).
  • 12See, e.g., Farrow, supra note 2.
  • 13See, e.g., Hiba Hafiz, How Legal Agreements Can Silence Victims of Sexual Assault, Atlantic (Oct. 18, 2017), [].
  • 14See generally, Ayres, supra note 4; Vicki Schultz, Open Statement on Sexual Harassment from Employment Discrimination Law Scholars, 71 Stan. L. Rev. Online 17, 46 (2018).
  • 15See, e.g., Ayres, supra note 3; Vasundhara Prasad, If Anyone Is Listening, #MeToo: Breaking the Culture of Silence around Abuse through Regulating Non-Disclosure Agreements and Secret Settlements, 59 B.C. L. Rev. 2507, 2515–16 (2018) (collecting articles).
  • 16Saul Levmore & Frank Fagan, Semi-Confidential Settlements in Civil, Criminal, and Sexual Assault Cases, 103 Cornell L. Rev. 311, 314 (2018).
  • 17See Annie Hill, Nondisclosure Agreements: Sexual Harassment and the Contract of Silence, Gender Pol’y Rep. (Nov. 14, 2017), [ 7VVW-LXWV]; Elizabeth Grace, Confidentiality of Settlements in Sexual Abuse Cases—Necessary Evil or Positive?, Lerners Pers. Inj. Group (Jan. 31, 2013), [].
  • 18See id.
  • 19In addition to the many contributions in this volume, see Schultz, supra note 14 (writing on behalf of Rachel Arnow-Richman, Ian Ayres, Susan Bisom-Rapp, Tristin Green, Rebecca Lee, Ann McGinley, Angela Onwuachi-Willig, Nicole Porter, and Brian Soucek); see also, Ayres, supra note 4; Prasad, supra note 14; Margaret Ryznar, #MeToo & Tax, 75 Wash. & Lee L. Rev. Online 53, 53–57 (2018); Ramit Mizrahi, Sexual Harassment Law after #MeToo: Looking to California as a Model, 128 Yale L. J. Forum 121, 140–41 (2018); Minna J. Kotkin, Invisible Settlements, Invisible Discrimination, 84 N.C. L. Rev. 927, 976–77 (2006).
  • 20See Farrow, supra note 2.
  • 21This Comment will directly address one such settlement in Part III. See Bandera v. City of Quincy, 344 F.3d 47, 49 (1st Cir. 2003) (refusing to invalidate a settlement agreement in which the employer offered to recommend the victim for a substitute teaching position in exchange for dropping her case).
  • 22See Farrow, supra note 2 (detailing the many psychological harms that the settlement process itself caused upon one of Weinstein’s victims, Ambra Battilana Gutierrez).
  • 23See Ryznar, supra note 19, at 54–57.
  • 24Restatement (Second) of Contracts § 208 (“If a contract or term thereof is unconscionable at the time the contract is made a court may refuse to enforce the contract, or may enforce the remainder of the contract without the unconscionable term, or may so limit the application of any unconscionable term as to avoid any unconscionable result.”); see also Hume v. United States, 132 U.S. 406, 411 (1889).
  • 25Williams v. Walker-Thomas Furniture Co., 350 F.2d 445, 449 (D.C. Cir. 1965); see also Samuel Williston & Richard A. Lord, A Treatise on the Law of Contracts § 18:9 (4th ed. 1993 & Update 2018) (collecting cases).
  • 26See Bynum v. Maplebear Inc., 160 F. Supp. 3d 527 (E.D.N.Y. 2016); Summers v. Crestview Apartments, 236 P.3d 586, 590 (Mont. 2010).
  • 27See Grayiel v. Appalachian Energy Partners 2001-D, LLP, 736 S.E.2d 91, 102 (W. Va. 2012); Ferguson v. Countrywide Credit Industries, Inc., 298 F.3d 778, 783 (9th Cir. 2002).
  • 28See Gandee v. LDL Freedom Enter., Inc., 293 P.3d 1197, 1201 (Wash. 2013) (rendering contract terms void because they were substantively unconscionable, without a consideration of procedural unconscionability); In re Poly-America, L.P., 262 S.W.3d 337, 360 (Tex. 2008).
  • 29See Tedesco v. Home Sav. Bancorp, Inc., 407 P.3d 289, 298 (Mont. 2017).
  • 30See Day v. CTA, Inc., 324 P.3d 1205, 1209 (Mont. 2014); see also Grayiel, 736 S.E.2d at 104.
  • 31Gandee, 293 P.3d at 1999.
  • 32Quicken Loans, Inc. v. Brown, 737 S.E.2d 640, 657 (W. Va. 2012).
  • 33Restatement (Second) of Contracts § 178; see also Williston, supra note 25 at §§ 12:1–3 (collecting cases).
  • 34Restatement (Second) of Contracts § 178(2).
  • 35Id. at § 178(3). This Comment will examine these factors more fully below.
  • 36See, e.g., Kaiser Steel Corp. v. Mullins, 455 U.S. 72, 77 (1982) (rendering unenforceable terms of a collective bargaining agreement that violated federal antitrust and labor laws).
  • 37See, e.g., Restatement (Second) of Contracts § 192 (promises to commit torts are also unenforceable).
  • 38See Gandee, 293 P.3d at 1999.
  • 39Granted, some lawyers themselves can be coarse and intimidating, but professional counsels would stay clear of inducing the level of duress necessary for unconscionability.
  • 40This was even more true in 1998, when Terry Dworkin and Elletta Callahan first tackled confidentiality more generally. See Terry Dworkin & Elletta Callahan, Buying Silence, 36 Am. Bus. L.J. 151, 153 (Fall, 1998). Dworkin and Callahan’s work represents the most extensive discussion on the topic to date. New case law since their date of publication and our particular focus on the sexual harassment context distinguishes this Comment from their earlier work.
  • 41Lana C. v. Cameron P., 108 P.3d 896, 902 (Alaska 2005).
  • 42Bowman v. Parma Bd. of Educ., 542 N.E.2d 663, 667 (Ohio Ct. App. 1988). Accord Perricone v. Perricone, 972 A.2d 666, 689 (Conn. 2009); Unami v. Roshan, 659 S.E.2d 724, 727 (Ct. App. Ga. 2008) (citing 7 Williston on Contracts § 16:14 (4th ed.)).
  • 43See Unami, 659 S.E.2d at 727 (citing Lachman v. Sperry–Sun Well Surveying, 457 F.2d 850, 852–54 (10th Cir. 1972); S. R. & P. Imp. Co. v. Am. Union Bank, 204 N.Y.S. 755, 755–56 (N.Y. Sup. Ct. 1924); Eggleston v. Pantages, 175 P. 34, 36–37 (Wash. 1918)).
  • 44Cooper Tire & Rubber Co. v. Farese, 423 F.3d 446, 457 (5th Cir. 2005).
  • 45Katz v. South Burlington Sch. Dist., 970 A.2d 1226, 1229 (Vt. 2009) (citing Camp v. Eichelkraut, 539 S.E.2d 588, 598 (2000)).
  • 46See id.; see also Saini v. Int’l Game Tech., 434 F. Supp. 2d 913, 922 (D. Nev. 2006).
  • 47See Saini, 434 F. Supp. 2d at 922; Uniroyal Goodrich Tire Co. v. Hudson, No. 95-1130, 1996 WL 520789, at *10 (6th Cir. Sept. 12, 1996).
  • 48See, e.g., Conn. Gen. Stat. Ann. § 53a-73a (2013) (limiting unwelcomed sexual contact with a non-minor, even if the perpetrator is in a position of power over the victim, to a class A misdemeanor); Del. Code Ann. Tit. 11, §§ 763–773 (2018) (limiting sexual contact with non-minors to a misdemeanor offense); D.C. Code Ann. § 22-3005 (2019) (requiring that a sexual contact felony be brought about by reasonable threat of death, bodily injury, or kidnapping); 720 Ill. Comp. Stat. Ann. 5/11-1.60 (2019) (imposing a requirement for sexual abuse felonies that the perpetrator used or threatened to use a dangerous weapon or caused bodily harm, among others); Minn. Stat. § 609.345 (2019) (requiring a finding that the perpetrator used force or the threat of force in order to charge him with felony).
  • 49See Schultz, supra note 14, at 38–39; Nicole Porter, Ending Harassment by Starting with Retaliation, 71 Stan. L. Rev. Online 49, 50 (2018).
  • 50See Rachel Arnow-Richman, Of Power and Process: Handling Harassers in an At-Will World, 128 Yale L. J. F. 85, 93–94 (June 18, 2018).
  • 51See Lissa Griffin & Ellen Yaroshefsky, Ministers of Justice and Mass Incarceration, 30 Geo. J. Legal Ethics 301, 312 (Spring, 2017) (“It is well accepted . . . that the prosecutor is a fiduciary who represents the sovereign and must make decisions in the public interest, for society at large—not any individual client.”) (citing Bruce A. Green, Why Should Prosecutors “Seek Justice”?, 26 Fordham Urb. L. J. 607, 633–34 (1999)).
  • 52The EEOC has endorsed the estimate that one in four women have experienced sexual harassment in the workplace, but that many women do not report it. See Equal Emp. Opportunity Comm’n, Women in the American Workforce (2015). Even when they do, in any given year, the EEOC resolves roughly 50 percent of their charges under the category of “No Reasonable Cause,” and 20.9 percent under “Administrative Closure” (another form of charge failure). Equal Emp’ Opportunity Comm’n, Charges Alleging Sex-Based Harassment (Charges Filed with EEOC) FY 2010–FY 2018 (2018). Only 22.8 percent of charges resolve with positive outcomes for the victim (as defined by the EEOC, not the victim), so there is no way of knowing whether the victim has truly received adequate reparation. See id. The EEOC does not provide data for the number of NDAs it has challenged in court. However, given the meager case law, that number is likely very low compared to the roughly 8,000 charges it resolves each year. See id. Indeed, about 9 percent of EEOC charges are resolved via settlement, so filings with the EEOC may be creating more sexual harassment NDAs. Id.
  • 53See, e.g., Sheets v. Teddy’s Frosted Foods, Inc., 427 A.2d 385 (Conn. 1980); Shores v. Senior Manor Nursing Center, Inc., 518 N.E.2d 471 (Ill. App. Ct. 1988); Boyle v. Vista Eyewear, Inc., 700 S.W.2d 859 (Mo. Ct. App. 1985); McQuary v. Bel Air Convalescent Home, Inc., 684 P.2d 21 (Or. Ct. App. 1984); Harless v. First Nat’l Bank in Fairmont, 246 S.E.2d 270 (W. Va. 1978).
  • 54See Dworkin & Callahan, supra note 40, at 190.
  • 55See Ryznar, supra note 19, at 54–57.
  • 56See Arnow-Richman, supra note 50, at 90–92 (discussing the inadequacy of antidiscrimination law to cover all forms of sexually motivated harm); Tristin K. Green, Was Sexual Harassment Law a Mistake? The Stories We Tell, 128 Yale L.J. F. 152 (June 18, 2018).
  • 5794 F.3d 738 (1st Cir. 1996).
  • 58See generally id.
  • 59Id. at 740–41.
  • 60Id. at 741.
  • 61Id.
  • 62Id. at 744 (quoting Town of Newton v. Rumery, 480 U.S. 386, 392 (1987)).
  • 63Id. at 744.
  • 64Id. (quoting EEOC v. Shell Oil Co., 466 U.S. 54, 69 (1984)).
  • 65Id.
  • 66Id. (citing EEOC v. Cosmair, Inc., L’Oreal Hair Care Div., 821 F.2d 1085, 1090 (1987)). It is important to note for later how the court reserves this interference only for communications with the EEOC, not any other party. It does not even suggest that nondisclosure in the sexual harassment setting could be harmful to public interest per se.
  • 67See id. (citing Carson v. Am. Brands, Inc., 450 U.S. 79, 88 n.14 (1981)).
  • 68See Jean R. Sternlight, Creeping Mandatory Arbitration: Is It Just?, 57 Stan. L. Rev. 1631, 1638 (2005).
  • 69434 F. Supp. 2d 913 (D. Nev. 2006).
  • 70See id. at 923.
  • 71See id. at 923.
  • 72Id. at 921.
  • 73Id. This finding is somewhat questionable if ITG was in fact intentionally short-changing its customers.
  • 74Id. at 922.
  • 75Id.
  • 76Id. at 921 (emphasis added).
  • 77Id. at 922.
  • 78See, e.g., EEOC v. Int’l. Profit Assocs., No. 01 C 4427, 2003 U.S. Dist. LEXIS 6761, at *3–4 (N.D. Ill. Apr. 21, 2003); EEOC v. Shell Oil Co., 466 U.S. 54, 69 (1984).
  • 79No. 01 C 4427, 2003 U.S. Dist. LEXIS 6761 (N.D. Ill. Apr. 21, 2003).
  • 80Id. at *2.
  • 81Id. at *6.
  • 82Katz v. South Burlington Sch. Dist., 970 A.2d 1226, 1229 (Vt. 2009).
  • 8301 Civ. 8421 (RMB) (RLE), 2002 U.S. Dist. LEXIS 17484 (S.D.N.Y. Sept. 20, 2002).
  • 84See id. at *5–6.
  • 85See id.
  • 86EEOC v. Astra U.S.A., Inc., 94 F.3d 738, 744 (1st Cir. 1996).
  • 87See, e.g., Rhitu Chatterjee, A New Survey Finds 81 Percent of Women Have Experienced Sexual Harassment, Nat’l Pub. Radio, Inc. (Feb. 21 2018), [].
  • 88Saini v. Int’l Game Tech., 434 F. Supp. 2d 913, 922 (D. Nev. 2006). See also Dworking & Callahan, supra note 40, at 166–68 (discussing other cases supporting the proposition that “the courts seem reluctant to excuse an employee from a confidentiality agreement where breach is motivated by personal gain, even in situations where disclosing information might advance public safety”).
  • 89972 A.2d 666 (Conn. 2009).
  • 90See id. at 689.
  • 91See id. at 688.
  • 92See id. at 688 (citing Leonard v. Clark, 12 F.3d 885, 891 (5th Cir. 1993); EEOC v. Astra U.S.A., Inc., 94 F.3d 738, 744 (1st Cir. 1996); Bowman v. Parma Bd. of Educ., 542 N.E.2d 663, 664 (Ohio 1988); Pansy v. Stroudsburg, 23 F.3d 772, 787–88 (3rd Cir. 1994)).
  • 93Id. at 687 (citation omitted).
  • 94See also Katz v. South Burlington School Dist., 970 A.2d 1226, 1228 (Vt. 2009).
  • 95344 F.3d 47 (1st Cir. 2003).
  • 96See id. at 50.
  • 97See id. at 49.
  • 98See id.
  • 99See id.
  • 100See id.
  • 101Even if the harassing employee exerted undue influence on the victim, the NDA is most often between the victim and the employer. Employers, as opposed to individual harassers, often have significant incentives and sufficient legal knowledge to prevent the extra risks that would emerge from soliciting an NDA via undue influence. Therefore, while the unconscionability doctrine is thematically relevant to sexual harassment, courts rarely apply it when the other party of an NDA is an employer (assuming the employer is not the individual harasser).
  • 102The public officers involved had likely violated Massachusetts law. See Mass. Gen. Laws Ann. ch. 151B § 1.18 (2014).
  • 103Bandera, 344 F.3d at 55.
  • 104See also Arnow-Richman, supra note 50, at 95–96; Schultz, supra note 14, at 46.
  • 105Bandera, 344 F.3d at 55.
  • 106The agreement did not even guarantee a position at the school. Mayor Sheets merely promised to recommend Bandera for a substitute position if she could not secure a full time position herself. See id. at 49. Quincy Public Schools currently pays its substitute teachers $85 per day, or roughly $15,000 per year if the substitute works every day of the year. Quincy Pub. Sch. Dist. #172: Internal Postings: Substitute Tchr., (Oct. 30, 2018), [].
  • 107As laid out in section B, such cases exist in the context of other crimes, but there has been no case in which a private plaintiff has succeeded in invalidating an NDA by a showing of clear evidence that she was sexually harassed by the other side of the agreement.
  • 108Dworkin & Callahan, supra note 40, at 162 (citing Restatement (Second) of Contracts, introductory note (1981)).
  • 109See, e.g., Bowman v. Parma Bd. of Educ., 542 N.E.2d 663, 667 (Ohio Ct. App. 1988). See also, Katz v. S. Burlington Sch. Dist., 970 A.2d 1226, 1228 (Vt. 2009).
  • 110EEOC v. Astra U.S.A., Inc., 94 F.3d 738, 744 (1st Cir. 1996).
  • 111See id.
  • 112Id.
  • 113See id.
  • 114For example, #TimesUp has also gained significant traction in the past two years since Weinstein’s behavior was revealed. See About Time’s Up, Time’s Up, [] (last visited Jan. 28, 2019).
  • 115See, e.g., Arnow-Richman, supra note 50, at 101–02; Green, supra note 56, at 166–68.
  • 116§ 178(3)(a)–(d) (1981). Of course, this language was available to the First Circuit by 1996. Dworking and Callahan discuss this in Buying Silence but only as pertains to whistleblowers. See Dworking & Callahan, supra note 40, at 179–90.
  • 117Saini v. Int’l Game Tech., 434 F. Supp. 2d 913, 921 (D. Nev. 2006).
  • 118See, e.g., Katz v. S. Burlington Sch. Dist., 970 A.2d 1226, 1228 (2009).
  • 119See Saini, 434 F. Supp. 2d at 921.
  • 120Bowman v. Parma Bd. of Educ., 542 N.E.2d 663, 667 (Ohio Ct. App. 1988). Accord Perricone v. Perricone, 972 A.2d 666, 689 (Conn. 2009); Unami v. Roshan, 659 S.E.2d 724, 727 (Ct. App. Ga. 2008) (citing 7 Williston on Contracts § 16:14 (4th ed.)).
  • 121As we will see, this may change in the future.
  • 122Harvey Weinstein is too easy an example. See also Arnow-Richman, supra note 50, at 95–96.
  • 123Bandera v. City of Quincy, 344 F.3d 47, 52 (1st Cir. 2003) (“But by the same token the district court cannot summarily deny enforcement simply because material facts are in dispute: the task is to resolve the dispute. And, in this instance, it is unlikely—though perhaps not impossible—that the matter could be resolved without an evidentiary hearing.”).
  • 124Efficiency is an important consideration when courts resolve issues of first impression or demands for reform. For example, the Supreme Court has many times invoked the federal judiciary’s limited case capacity in defining the limitation of “Federal Question Jurisdiction”—the ability of federal courts to hear cases involving federal issues. See, e.g., Grable & Sons Metal Prod. v. Darue Eng’g & Mfg., 545 U.S. 308, 318–19 (2005) (“For if the federal labeling standard without a federal cause of action could get a state claim into federal court, so could any other federal standard without a federal cause of action. And that would [mean] a tremendous number of cases.”); see also 28 U.S.C. § 1331 (2012). Letting too many cases into federal court would not only create an imbalance between the state and federal systems, it would pose a serious threat to the efficiency of federal courts. See also Merrell Dow Pharm. Inc. v. Thompson, 478 U.S. 804, 811–12 (1986) (noting the “increased [complexity and] volume of federal litigation” as “considerations that should inform” efficiency decisions in the interpretation of Federal Question). Promptness and accuracy are both essential components in delivering justice, so an inefficient system threatens to become an unjust one. See Fed. R. Civ. P. 1 advisory committee’s note to 1993 amendment (“The purpose of this revision, adding the words ‘and administered’ to the second sentence, is to recognize the affirmative duty of the court to exercise the authority conferred by these rules to ensure that civil litigation is resolved not only fairly, but also without undue cost or delay. As officers of the court, attorneys share this responsibility with the judge to whom the case is assigned.”).
  • 125See generally, Gerald N. Rosenberg, The Hollow Hope: Can Courts Bring About Social Change? (2nd ed. 2008).
  • 126See United States v. Aegerion Pharm., Inc., 280 F. Supp. 3d 217, 228 (D. Mass. 2017); Megan Keller, Poll: Over 40 Percent Believe #MeToo Movement Has Gone Too Far, Hill (Oct. 31, 2018) [].
  • 127Bandera, 344 F.3d at 52.
  • 128See Schultz, supra note 14, at 39 (citing Nicole Buonocore Porter, Ending Harassment by Starting with Retaliation, 71 Stand L. Rev. Online 49 (2018)); Deborah L. Brake, Retaliation in an EEO World, 89 Ind. L.J. 115, 136–39 (2014)).
  • 129Ayres, supra note 4, at 79–83.
  • 130See id. at 78–79.
  • 131Cal. Civ. Proc. Code § 1002(a)–(b) (2019).
  • 132See Restatement (Second) of Contracts, § 178(3)(a)–(d) (1981); Saini v. Int’l Game Tech., 434 F. Supp. 2d 913, 921 (D. Nev. 2006).
  • 133See id.
  • 134N.Y. C.P.L.R. § 5003–b (McKinney 2018).
  • 135See generally Arnow-Richman, supra note 50; Ayres, supra note 4.
  • 136As of September 13, 2019, no case has cited New York’s statute. As for California’s statute, no cases have emerged, though it is cited by a series of complaints against USA Gymnastics and Michigan State University for their alleged cover up of Dr. Lawrence Nassar’s sexual violations. See, e.g., Complaint at 202, Davis v. Mich. State Univ., No. 1:17-cv-00029, 2018 WL 4329266 (W.D. Mich. Sep. 10, 2018). As one can tell from the complaint’s jurisdiction, the California statute is not central to the plaintiff’s claim.
  • 137California and New York’s laws became effective on January 1, 2018, and July 11, 2018, respectively.
  • 138See Kenworthey Bilz & Janice Nadler, Law, Moral Attitudes, and Behavioral Change, in The Oxford Handbook of Behavioral Economics and the Law 242–67 (Eyal Zamir & Doron Teichman eds., 2014) (concluding that law can also play an important role in changing “moral attitudes,” especially in the sexual harassment setting).
  • 139Bandera v. City of Quincy, 344 F.3d 47, 49 (1st Cir. 2003).
  • 140Cal. Civ. Proc. Code § 1002(a) (2019).
  • 141See Cal. Penal Code § 243.4 (2019).
  • 142N.Y. C.P.L.R. § 5003–b (McKinney 2018).
  • 143Restatement (Second) of Contracts, § 178(3)(a).
  • 144See generally Green, supra note 56.
  • 145See Restatement (Second) of Contracts, § 178(3)(a).
  • 146See Bilz & Nadler, supra note 138, at 243 (suggesting that even where judges fail to follow the public trend, public outrage against their decisions as motivated by new laws encourages them towards legal reform) (citing Cass R. Sunstein, Social Norms and Social Roles, 96 Colum. L. Rev. 903–68 (1996)).
  • 147Federal Arbitration Act, 9 U.S.C. §§ 1–16 (2012).
  • 148563 U.S. 333 (2011).
  • 149Id. at 352 (quoting Hines v. Davidowitz 312 U.S. 52, 67 (1941)).
  • 150Id. at 338.
  • 151Id. at 339 (quoting Moses H. Cone Memorial Hospital v. Mercury Constr. Corp., 460 U.S. 1, 24 (1983)).
  • 152Id. at 342.
  • 153See, e.g., Fed. R. Civ. P. 5.2, 26.
  • 154See Sternlight, supra note 68, at 1638.
  • 155Warfield v. Beth Israel Deaconess Medical Center, Inc. 910 N.E.2d 317, 326 (Mass. 2009); accord Joule, Inc., v. Simmons, 944 N.E.2d 143, 149 (Mass. 2011); see also Harris v. Bingham McCutchen LLP, 154 Cal.Rptr.3d 843, 849 (Ct. App. Cal. 2013) (following Massachusetts’s approach in determining that the FAA did not preempt the court’s refusal to enforce a certain term).
  • 156See Warfield, 910 N.E.2d at 325–26.
  • 157Id. at 327.
  • 158See N.Y. C.P.L.R. § 5003–b (McKinney 2018).
  • 1599 U.S.C. § 2.
  • 160See Hume v. United States, 132 U.S. 406, 411 (1889) (embracing the unconscionability doctrine).
  • 161Restatement (Second) of Contracts, § 208 (stating that the two doctrines overlap). According to Frank P. Darr, the unconscionability doctrine is a form of public policy, since it enables courts to dictate to society what proper contract behavior ought to be. See Frank P. Darr, Unconscionability and Price Fairness, 30 Hous. L. Rev. 1819, 1849 (1994).