Microsoft Word - 42.2 (6.20.22).docx 42.2 COLUMBIA JOURNAL OF GENDER AND LAW 51 “LEAST FAVORED NATION”: PREGNANCY DISCRIMINATION DISPARATE IMPACT CLAIMS POST-YOUNG EMMA LUX* Abstract This Article analyzes disparate impact claims under the Pregnancy Discrimination Act (PDA) in light of the Supreme Court’s decision in Young v. United Parcel Service, Inc. In Young, the Court interpreted the PDA to provide plaintiffs who bring pregnancy-related disparate treatment claims pursuant to Title VII with additional protections that plaintiffs who bring non-pregnancy-related claims under Title VII do not receive. The Young Court reasoned that this interpretation flowed from the PDA because Congress intended the Act to modify Title VII and wanted to ensure that federal courts would not prematurely dismiss pregnancy discrimination claims, as they historically had done. This Article argues that such reasoning not only provides additional protections for plaintiffs who bring disparate treatment claims, but also furnishes similar safeguards for plaintiffs who bring disparate impact claims. The Article concludes by noting, however, that federal courts have not yet extended such special protection to plaintiffs with pregnancy- related disparate impact claims. The result is that courts still often prematurely dismiss their claims, despite the PDA’s purpose and Young’s reasoning. INTRODUCTION In Young v. United Parcel Service, Inc., the Supreme Court held that the Pregnancy Discrimination Act provides additional protection to plaintiffs bringing intentional pregnancy discrimination claims—protection that exceeds what plaintiffs © 2022 Lux. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits the user to copy, distribute, and transmit the work provided that the original author(s) and source are credited. * Associate at O’Melveny & Myers, LLP. J.D., Georgetown University Law Center, 2021. The author would like to thank Professor Irv Gornstein for making this Article possible. She is also grateful to the Columbia Journal of Gender and Law editors for their thoughtful feedback. 52 COLUMBIA JOURNAL OF GENDER AND LAW 42.2 with Title VII claims outside of the pregnancy discrimination context receive.1 Prior to Young, many plaintiffs encountered roadblocks at the prima facie stage of pregnancy-related Title VII disparate treatment claims,2 despite the fact that Congress had passed the Pregnancy Discrimination Act (PDA) nearly four decades earlier in an effort to ease the burden of proof placed on plaintiffs who attempt to raise an inference of intentional discrimination.3 Title VII disparate treatment claims require a showing that an employer intentionally discriminated against a member of a protected class.4 Plaintiffs who furnish direct or circumstantial evidence of intentional discrimination meet their prima facie burden, which shifts the burden to the employer to proffer a legitimate reason for the policy.5 One way to make the requisite showing of intent at the prima facie stage is to demonstrate that a similarly situated comparator was treated differently, raising an inference of discrimination.6 However, finding a person who is similar to a pregnant person in their ability or inability to work is “difficult, if not impossible.”7 Some commentators suggested that Young would make disparate treatment pregnancy discrimination cases less onerous at the prima facie stage.8 Post- 1 575 U.S. 206, 229 (2015). 2 See, e.g., Joanna Grossman & Gillian Thomas, Making Pregnancy Work: Overcoming the Pregnancy Discrimination Act’s Capacity-Based Model, 21 YALE J.L. & FEMINISM 15, 34–35 (2009) (“Although showing favorable treatment of a comparator is not the only way to prove disparate treatment under Title VII, courts tend to treat it as such in the pregnancy context.”). 3 See, e.g., Joanna Grossman & Gillian Thomas, Making Sure Pregnancy Works: Accommodation Claims After Young v. United Parcel Service, Inc., 14 HARV. L & POL’Y REV. 319, 321–22 (2020) (describing how the PDA was intended to overturn Gilbert’s finding that the plaintiff had not established a prima facie burden of discrimination). 4 See, e.g., Ricci v. DeStefano, 557 U.S. 557, 577 (2009). 5 E.g., Young, 575 U.S. at 228 (citing McDonnell Douglas Corp. v. Green, 411 U.S. 792 (1973)). 6 See Grossman & Thomas, supra note 2, at 34. 7 Id. (explaining that the unique burdens of pregnancy and lack of adequate comparators often hinder pregnancy-related disparate treatment claims). 8 Grossman & Thomas, supra note 3, at 331 (explaining that “the Young-modified prima facie analysis has resulted in more plaintiff-friendly rulings”). 42.2 COLUMBIA JOURNAL OF GENDER AND LAW 53 Young, however, the majority of pregnancy discrimination plaintiffs continue to encounter early dismissals when pursuing a disparate treatment theory.9 In light of the continued difficulties associated with challenging pregnancy- related discrimination under a disparate treatment theory of Title VII post-Young, some commentators have expressed hope that the “disparate impact framework . . . offers untapped potential for PDA litigants challenging failures to accommodate.”10 In addition to prohibiting disparate treatment, Title VII also prohibits policies with a disparate impact on protected classes, i.e., “practices that are not intended to discriminate but in fact have a disproportionately adverse effect on minorities.”11 To be sure, disparate impact cases under the PDA historically have not been the most successful.12 Before Young, many courts reasoned that plaintiffs who brought disparate impact claims sought impermissible “preferential” action, and that their claims therefore failed under Title VII as amended by the PDA.13 Plaintiffs pursuing pregnancy-related disparate impact claims have also struggled to amass statistical evidence of impact, “particularly in fields historically dominated by men.”14 9 See, e.g., DINA BAKST, ELIZABETH GEDMARK & SARAH BRAFMAN, A BETTER BALANCE: THE WORK & FAM. LEGAL CTR., LONG OVERDUE: IT IS TIME FOR THE FEDERAL PREGNANT WORKERS FAIRNESS ACT (2019), https://www.abetterbalance.org/wp-content/uploads/2019/05/Long-Overdue.pdf [https://perma.cc/76XW-BZBF]. 10 Grossman & Thomas, supra note 3, at 342. 11 Ricci v. DeStefano, 557 U.S. 557, 577 (2009). 12 See L. Camille Hébert, Disparate Impact and Pregnancy: Title VII’s Other Accommodation Requirement, 24 AM. UNIV. J. GENDER 107, 157 (describing cases where courts rejected disparate impact claims); Grossman & Thomas, supra note 3, at 342 (describing how “disparate impact claims under the PDA have enjoyed a checkered history, to be sure”). 13 See, e.g., Stout v. Baxter Healthcare Corp., 282 F.3d 856, 861 (5th Cir. 2002) (reasoning that a workplace policy did not disproportionately impact the plaintiff because “the PDA does not require preferential treatment of pregnant employees and does not require employers to treat pregnancy related absences more leniently than other absences”). 14 See Grossman & Thomas, supra note 2, at 44 (explaining that statistical evidence of discrimination against pregnancy is more difficult to collect in such fields, where “women may comprise a small portion of the workforce—or where the plaintiff literally might be the only female employee in the workplace”). 54 COLUMBIA JOURNAL OF GENDER AND LAW 42.2 Nonetheless, Young offered reasons for optimism regarding the potential of the disparate impact theory for PDA plaintiffs challenging denials of light duty accommodations.15 While Young only concerned a disparate treatment claim, it contained indicia suggesting that employment policies like the one at issue, which only permitted light duty accommodations for employees injured on the job (and several other narrow classes of employees), may also disproportionately impact pregnant employees.16 During oral argument, Justice Breyer reflected that, had the plaintiff successfully brought a disparate impact claim in that case, it would have been a “beautiful vehicle” for her claim.17 Even Justice Kennedy, who dissented from the Court’s disparate treatment analysis, nonetheless underscored the continued availability of disparate impact claims under the PDA post-Young.18 This Article delves into disparate impact claims in the PDA context to determine whether and how Young affected the potential success of such claims. Part I explains the genesis of the PDA. Part II provides an overview of how the Young Court interpreted the PDA’s second clause, which states that “women affected by pregnancy, childbirth, or related medical conditions shall be treated the same for all employment-related purposes . . . as other persons not so affected but similar in their 15 Grossman & Thomas, supra note 3, at 342–44 (advocating for a disparate impact approach as a “way out of the comparator conundrum” post-Young). Light duty accommodations are common in manual labor fields and generally involve a reduction of the amount of weight an employee is expected to lift or carry, inter alia. See, e.g., Young, 575 U.S. at 224. 16 In Young, the light duty policy at issue permitted light duty only for individuals with on-the-job injuries, individuals who lacked a Department of Transportation certification, and individuals who qualified for accommodation under the ADA. Individuals whose disability arose off the job, like pregnant workers, could not obtain light duty. Young v. United Parcel Serv., Inc., 575 U.S. 206, 211– 12 (2015). 17 Young had attempted to amend her complaint to bring a disparate impact claim, but the district court denied her motion to amend the complaint, and the Fourth Circuit held that this was not an abuse of discretion. Young v. United Parcel Serv., Inc., 707 F.3d 437, 442 n.6 (4th Cir. 2013). In Justice Breyer’s view, Young’s theory of harm—that the policy at issue disproportionately impacted pregnant workers—fit more neatly into the disparate impact framework than into the disparate treatment framework, which requires an intent to discriminate. Transcript of Oral Argument at 9, Young v. United Parcel Serv., Inc., 575 U.S. 206 (2015) (No. 12-1226) (“[W]hat am I to do because I don’t know that you want to twist the disparate [treatment] claim out of shape when you have such a beautiful vehicle to bring a claim of [this] kind[.]”). 18 See Young, 575 U.S. at 252–53 (Kennedy, J., dissenting). 42.2 COLUMBIA JOURNAL OF GENDER AND LAW 55 ability or inability to work,” in the context of disparate treatment claims.19 Part III then explains why and how that same clause should theoretically modify the burden placed on pregnancy discrimination plaintiffs who bring disparate impact claims, as it does in the disparate treatment context. In Part IV, this Article examines the only post-Young federal appellate decision regarding PDA disparate impact claims, Legg v. Ulster County,20 and concludes that the Justices in Young may have been overly optimistic about the potential for successful disparate impact challenges under Title VII as amended by the PDA. I. The Pregnancy Discrimination Act Congress passed the PDA to overturn both the holding and reasoning in General Electric Co. v. Gilbert.21 Gilbert concerned the viability of an employment benefits plan that excluded pregnancy from coverage while providing benefits for conditions causing a similar inability to work.22 With respect to disparate treatment, the Gilbert majority reasoned that the policy’s exclusion of pregnancy was not sex discrimination, but rather discrimination against the identifiable condition of pregnancy.23 It held that, absent further evidence, the exclusion of pregnancy from a plan which covered other disabilities causing a similar inability to work did not alone raise an inference of intentional discrimination sufficient to satisfy the plaintiff’s prima facie burden.24 The Gilbert majority also analyzed the policy through the disparate impact framework.25 It held that the exclusion of pregnancy from a benefits plan, without 19 42 U.S.C. § 2000e(k). 20 Legg v. Ulster Cnty., 832 F. App’x 727 (2d Cir. 2020). 21 Newport News Shipbuilding & Dry Dock Co. v. EEOC, 462 U.S. 669, 678 (1983) (“When Congress amended Title VII in 1978 [with the PDA], it unambiguously expressed its disapproval of both the holding and the reasoning of the Court in the Gilbert decision.”). 22 Gen. Elec. Co. v. Gilbert, 429 U.S. 125, 128–29 (1976). 23 Id. at 136. 24 Id. (“We do not . . . infer that the exclusion of pregnancy disability benefits from [an employer’s] plan is a simple pretext for discriminating against women.”). 25 Id. at 138–39. 56 COLUMBIA JOURNAL OF GENDER AND LAW 42.2 more, did not satisfy the plaintiff’s prima facie burden of establishing a disparate impact, even though the plan covered other disabilities that, like pregnancy, may cause absences from work or necessitate accommodations.26 The Court held that, to establish a prima facie burden of disparate impact, a plaintiff would have had to produce additional proof that the “package [wa]s in fact worth more to men,” for instance, by showing that the policy caused comparatively more money to be distributed to men than women, or that it gave a benefit to men which women did not receive.27 With the PDA, Congress sought to overturn the reasoning and holding of Gilbert through two clauses, both of which appear in the Title VII definitions section.28 The first clause defines the term “because of sex” as used in Title VII to “include . . . because of or on the basis of pregnancy, childbirth, or related medical conditions.”29 This first PDA clause, the Supreme Court has found, was intended to supersede the Gilbert Court’s reasoning that discrimination against pregnancy was not discrimination against sex, but rather discrimination against the identifiable condition of pregnancy.30 The second clause states, “and women affected by pregnancy, childbirth, or related medical conditions shall be treated the same for all employment-related purposes . . . as other persons not so affected but similar in their ability or inability to work.”31 The Supreme Court has stated that this clause was “intended to overrule the holding in Gilbert and to illustrate how discrimination against pregnancy is to be 26 Id. at 138. 27 Id. 28 See 42 U.S.C. § 2000e(k). 29 Id. 30 See Cal. Fed. Sav. & Loan Ass’n v. Guerra, 479 U.S. 272, 284 (1987) (“[T]he first clause of the [PDA] reflects Congress’ disapproval of the reasoning in Gilbert [by] adding pregnancy to the definition of sex discrimination prohibited by Title VII.”). 31 42 U.S.C. § 2000e(k). 42.2 COLUMBIA JOURNAL OF GENDER AND LAW 57 remedied.”32 How exactly the clause should function in order to accomplish that goal, however, has “raise[d] several difficult questions of interpretation.”33 The following section provides an overview of how the Supreme Court, in Young v. United Parcel Service, Inc., interpreted the second clause with respect to disparate treatment claims. II. Young v. UPS: The PDA’s Second Clause in Disparate Treatment Cases The recent Supreme Court case, Young v. UPS, involved a disparate treatment challenge to UPS’s light duty policy.34 The Young plaintiff was a pregnant UPS worker who sought a light duty accommodation after her physician advised her that she should not lift more than twenty pounds while pregnant.35 UPS denied her accommodation request because its accommodations policy only permitted light duty for employees with on-the-job injuries, employees who lacked a Department of Transportation certification, and employees covered by the Americans with Disabilities Act.36 At trial, the plaintiff alleged that UPS intentionally discriminated against her because of pregnancy by using the McDonnell Douglas burden-shifting framework.37 The McDonnell Douglas framework provides a way for disparate treatment claimants to meet their prima facie burden where they lack direct evidence of an employer’s discriminatory intent.38 Under that framework, a plaintiff may raise an 32 Guerra, 479 U.S. at 285. 33 Young v. United Parcel Serv., Inc., 575 U.S. 206, 233 (2015) (Alito, J., concurring). 34 Id. at 212 (“This case requires us to consider the application of the second clause to a ‘disparate- treatment’ claim.”). 35 Id. at 211. 36 Id. at 211–12. 37 Id. at 211. The Young plaintiff also raised several other theories of intentional discrimination that are outside the scope of this Article. 38 E.g., Young v. United Parcel Serv., Inc., 575 U.S. 206, 228–29 (2015). 58 COLUMBIA JOURNAL OF GENDER AND LAW 42.2 inference of intentional discrimination sufficient to satisfy her prima facie burden by showing that “she belongs to the protected class, that she sought accommodations, that the employer did not accommodate her, and that the employer did accommodate others ‘similar in their ability or inability to work.’”39 A plaintiff who makes this showing shifts the burden to the employer to offer a legitimate reason for the policy.40 If the employer offers an apparently non-discriminatory reason for its action, “the plaintiff may in turn show that the employer’s proffered reasons are in fact pretext.”41 In Young, the lower courts found not only that the plaintiff failed to satisfy her prima facie burden under the McDonnell-Douglas framework, but that, even if she had, she also failed to show that the employer’s proffered reason for the policy was pretextual.42 As a result, outside of the pregnancy discrimination context, her disparate treatment claim would have failed.43 The Supreme Court, however, interpreted the second clause of the statute to provide Young and future PDA plaintiffs with a novel way to show intentional discrimination.44 Young held that, under the second clause of the PDA, a pregnancy discrimination plaintiff may continue to meet her prima facie burden pursuant to the previously articulated four-part standard in McDonnell Douglas.45 This shifts the burden to the employer to provide a legitimate non-discriminatory reason for the 39 Id. at 229. 40 Id. 41 Id. 42 See Young v. United Parcel Serv., Inc., No. DKC 08-2586, 2011 WL 665321, at *10–12 (D. Md. Feb. 14, 2011); Young v. United Parcel Serv. Inc., 707 F.3d 437, 442, 446 (4th Cir. 2013) (explaining that the district court found that Young failed to establish a prima facie case of sex discrimination because she could identify no similarly situated comparator who received more favorable treatment than she did). 43 See Young, 575 U.S. at 229. 44 Id. 45 Id. 42.2 COLUMBIA JOURNAL OF GENDER AND LAW 59 policy, as usual.46 The Young court added a gloss to the usual framework, though, emphasizing that this burden is not “onerous” and that a PDA claimant need not show that the employees who were accommodated were “similar in all but the protected ways.”47 After satisfying the prima facie burden, ordinarily a Title VII plaintiff would need to demonstrate that the employer’s proffered reason for the policy is pretextual, which the lower courts found that Young failed to do.48 However, the Supreme Court articulated a novel way for a PDA plaintiff alleging intentional discrimination to reach a jury on the question of intentional discrimination: by “providing sufficient evidence that the employer’s policies impose a significant burden on pregnant workers, and that the employer’s ‘legitimate, nondiscriminatory’ reasons are not sufficiently strong to justify the burden . . . .”49 Such a showing in the context of pregnancy discrimination claims may give “rise to an inference of intentional 46 Id. (“The employer may then seek to justify its refusal to accommodate the plaintiff by relying on ‘legitimate, nondiscriminatory’ reasons for denying her accommodation.”). 47 Id. at 228 (citing McDonnell Douglas Corp. v. Green, 411 U.S. 792, 802 (1973)). The lower court had reasoned that Young failed to meet her prima facie burden because, inter alia, “she could not identify a similarly situated comparator.” Young v. United Parcel Serv. Inc., 707 F.3d 437, 450 (4th Cir. 2013) (reasoning that pregnant workers are different than those workers who the policy covered— individuals disabled under the ADA, individuals who lacked DOT certification, and individuals injured on the job); cf. Grossman & Thomas, supra note 2, at 34 (explaining that the McDonnell Douglas framework in the pregnancy discrimination context “tends to invite the search for a similarly-situated comparator,” which makes it “difficult, if not impossible, to identify employees ‘similar in their ability or inability to work’”). The Supreme Court in Young, overturning the lower court’s holding, took pains to emphasize that the prima facie stage is not so “inflexible” and that a plaintiff can show similar employees obtained accommodations even though those employees may slightly differ in their ability or inability to work. Young, 575 U.S. at 228. 48 See Young v. United Parcel Serv., Inc., No. DKC 08-2586, 2011 WL 665321, at *10–12 (D. Md. Feb. 14, 2011); Young, 707 F.3d at 442, 446. 49 Young, 575 U.S. at 229. 60 COLUMBIA JOURNAL OF GENDER AND LAW 42.2 discrimination,”50 even though a showing of impact alone is usually insufficient to demonstrate pretext outside of the pregnancy context.51 The majority reasoned that this interpretation of the second clause is necessary in the PDA disparate treatment context to ensure that the purpose of the law is fulfilled.52 Without the PDA’s second clause, the Young Court reasoned, the first clause would only overturn the Gilbert majority’s reasoning regarding disparate treatment, but not its holding.53 For example, the Young Court could properly reason that the UPS policy’s exclusion of pregnancy was indeed because of sex.54 This approach would overturn the reasoning of Gilbert, and fulfill the instruction from the PDA’s first clause that sex discrimination includes discrimination “because of or on the basis of pregnancy, childbirth, or related medical conditions . . . .”55 Nevertheless, the Young Court could still hold that the challenged exclusion of pregnancy from eligibility for light duty—because pregnancy inherently arises off the job—does not raise an inference of intentional discrimination, without more, just as the Gilbert Court had reasoned with respect to a similar policy.56 In fact, the lower court in the Young case did just that, finding no evidence of intentional discrimination in UPS’s policy.57 Ultimately, the Young Court concluded that the PDA’s second clause must 50 Id. at 230. 51 Id. at 248 (Scalia, J., dissenting) (describing it as “topsy-turvy” that plaintiffs, in the pregnancy discrimination context, “can establish disparate treatment by showing that the effects of her employer’s policy fall more harshly on pregnant women than on others”). 52 Id. at 229. The Supreme Court also reasoned that the second clause must bear some substantive meaning, or else it would be superfluous. Id. at 226. 53 Id. at 227–28 (citations omitted) (“The first clause of [the PDA] reflects Congress’ disapproval of the reasoning in Gilbert[.]”). 54 Cf. Gen. Elec. Co. v. Gilbert, 429 U.S. 125, 135 (1976) (reasoning that the exclusion of pregnancy from a disability benefits plan was not because of sex but rather discrimination because of an identifiable condition). 55 See 42 U.S.C. § 2000e(k). 56 See Young, 575 U.S. at 227–28. Cf. Gilbert, 429 U.S. at 129 (explaining that the plaintiff would have had to show further evidence of pretext to establish her prima facie burden). 57 See Young v. United Parcel Serv. Inc., 707 F.3d 437, 450 (4th Cir. 2013). 42.2 COLUMBIA JOURNAL OF GENDER AND LAW 61 provide substantive protection for PDA disparate treatment claimants to ensure that courts do not produce the same result as in Gilbert—the result that Congress intended the PDA to overturn.58 Similarly, the Young Court reasoned that the PDA’s goal of overturning Gilbert compelled the conclusion that an employer may not satisfy its burden to show a legitimate nondiscriminatory reason for the policy simply by claiming that “it is more expensive or less convenient to add pregnant women to the category of those” who the employee accommodates, reasoning that the Gilbert employer could have done just that, and Congress sought to avoid the Gilbert result.59 Young did not involve a disparate impact claim; the plaintiff there had attempted to raise such a claim in the district court, but it was dismissed on procedural grounds.60 As a result, though Young clarified the meaning of the PDA’s second clause in the disparate treatment context, it provided little explicit guidance regarding whether and how the PDA’s second clause bears on a disparate impact analysis.61 The following Part argues that the second clause should provide substantive protection to PDA claimants in the disparate impact context as well as to those in the disparate treatment context, to ensure that the Gilbert majority’s holding with respect to disparate impact is also overturned. III. The Function of the PDA’s Second Clause in Disparate Impact Cases This Part mirrors the reasoning of Young in the disparate impact context, examining the PDA’s text and purpose to interpret whether and how it modifies the ordinary Title VII framework. 58 Young, 575 U.S. at 230–31; § 2000e(k). 59 Young, 575 U.S. at 229. 60 Id. at 242. See Young, 707 F.3d at 442 n.6. 61 See Young, 575 U.S. at 242. See also Grossman & Thomas, supra note 3, at 343. 62 COLUMBIA JOURNAL OF GENDER AND LAW 42.2 As Part II explains, the PDA contains two relevant clauses.62 The first clause plays a fairly straightforward role in the disparate impact analysis. It defines “because of sex” to include “because of . . . pregnancy, childbirth or related medical conditions.”63 Title VII’s prohibition on disparate impact discrimination64 prohibits facially neutral employment practices that disproportionately impact protected individuals unless they are justified by a business necessity.65 As a result, the first clause of the PDA and Title VII’s disparate impact provision together establish the following rule: employment practices that are facially neutral but that disproportionately impact employees “because of . . . pregnancy . . .” are unlawful unless they are justified by a business necessity.66 How the PDA’s second clause fits into that disparate impact analysis is “less clear”67 from the text alone. That clause states “and women affected by pregnancy, childbirth, or related medical conditions shall be treated the same for all employment-related purposes . . . as other persons not so affected but similar in their ability or inability to work.”68 Does this text apply to disparate impact claims, as well as disparate treatment claims as Young found?69 And if it applies to disparate impact 62 See supra Part I. 63 § 2000e(k). 64 42 U.S.C. § 2000e-2(a)(2). 65 See Griggs v. Duke Power Co., 401 U.S. 424, 432 (1971). See also Grossman & Thomas, supra note 3, at 342 (describing a disparate impact cause of action under Title VII as amended by the PDA). 66 See Grossman & Thomas, supra note 3, at 342. Specifically, the first clause read together with Title VII’s prohibition on disparate impact discrimination says that it is an unlawful employment practice to “limit, segregate, or classify [employees] in any way which would deprive or tend to deprive any individual of employment opportunities or otherwise adversely affect his status as an employee, because of such individual’s . . . sex [which ‘include[s] but [is] not limited to, because of . . . pregnancy, childbirth, or related medical conditions’].” § 2000e-2(a)(2); § 2000e(k). 67 Young v. United Parcel Serv., Inc., 575 U.S. 206, 219 (2015). 68 See § 2000e(k). 69 Recall that Young found that the PDA’s second clause modified the ordinary disparate treatment test for pregnancy discrimination plaintiffs. See Young, 575 U.S. at 229. 42.2 COLUMBIA JOURNAL OF GENDER AND LAW 63 claims, how does the second clause operate to modify the ordinary disparate impact rule that flows from the first clause of the PDA and Title VII? This Part argues that the second clause does indeed apply to disparate impact claims, and that Congress intended for that clause to modify the ordinary disparate impact test for plaintiffs alleging pregnancy discrimination. Under the second clause, this Article proposes, PDA plaintiffs may show that an employment policy per se satisfies their prima facie burden to show a disparate impact when the policy disqualifies pregnant employees from obtaining a benefit that non-pregnant employees who are “similar in their ability or inability to work”70 may obtain, due to the cause of their conditions. The following section explains why the second clause should apply in PDA disparate impact claims under Young’s reasoning.71 It then demonstrates why this interpretation of the second clause represents the best reading of what Congress intended when it provided that, “women affected by pregnancy, childbirth, or related medical conditions shall be treated the same for all employment-related purposes . . . as other persons not so affected, but similar in their ability or inability to work.”72 A. The Second Clause’s Necessary Role in Disparate Impact Cases The PDA’s second clause should be interpreted to apply to disparate impact claims as well as disparate treatment claims.73 Though the text of the second clause is unclear regarding the scope of its application, it appears in the definitions section that modifies those Title VII provisions that prohibit both disparate treatment and disparate impact.74 There is no textual indication that the second clause should not, therefore, apply to both Title VII’s disparate treatment provision and to its disparate 70 See id. 71 See infra Section II.A. 72 § 2000e(k); see infra Section II.B. 73 See Young, 575 U.S. at 229. 74 The PDA’s two clauses modify the “subchapter” 42 U.S.C. § 2000e, which contains Title VII’s provisions prohibiting employment practices that discriminate intentionally or by causing an adverse impact. See 42 U.S.C. § 2000e-2(a)(1)–(2). It is significant to note that the original PDA also modified the subchapter where the Title VII prohibitions on disparate treatment and disparate impact discrimination then appeared. Pregnancy Discrimination Act, Pub. L. No. 95-555, 92 Stat. 2076. 64 COLUMBIA JOURNAL OF GENDER AND LAW 42.2 impact provision, given that both appear in the “subchapter” that is modified by the definitions section.75 Because the text provides little insight beyond the placement of provisions,76 however, this Article interprets the PDA in light of its purpose.77 As noted above, Congress passed the Pregnancy Discrimination Act with the explicit purpose of overturning the Supreme Court’s reasoning and holding in General Electric Company v. Gilbert. Recall that Gilbert assessed the viability of an employment benefits plan that, like the one in Young,78 excluded pregnancy despite covering other nonoccupational sicknesses and conditions that cause a similar inability to work.79 Young only addressed the Gilbert majority’s holding and reasoning in the disparate treatment context. For the same reasons as were expressed in Young, however, the second clause must provide substantive protections in the disparate impact analysis as well. If the second clause of the PDA does not apply to disparate impact claims, then Congress’s purpose to overturn the holding of Gilbert would be frustrated. Imagine, for instance, that Young had properly brought a disparate impact claim to challenge 75 § 2000e(k). 76 The Justices also acknowledged during oral argument in Young that the “and” that begins the second clause implies that the clause bears substantive protections, rather than mere definitional terms. Transcript of Oral Argument, supra note 17, at 57–58; see also Young, 575 U.S. at 226 (describing how the second clause is not merely “definitional”). 77 Cf. Young, 575 U.S. at 227 (interpreting the disparate treatment provision in light of the PDA’s purpose). 78 In Young, Justice Scalia argued in dissent that the Gilbert case involved different facts than in Young. Young, 575 U.S. at 246 (Scalia, J., dissenting). In his view, Young involved a neutral employment policy, whereas the one in Gilbert concerned an “otherwise comprehensive disability benefits plan that singled pregnancy out for disfavor.” Id. The Young Court, however, rejected Justice Scalia’s view, arguing that only the Gilbert dissents viewed the policy there as intentionally discriminatory. Id. at 227. The Young majority viewed the policy in Gilbert as akin to the policy in Young—both facially distinguished pregnancy on neutral grounds such as “accommodat[ing] only sicknesses or accidents.” Id. (explaining, though, that such neutral distinctions may raise an inference of intentional discrimination). For this reason, this Article will treat the policies in Young and in Gilbert as analogous—both “treat[ed] pregnancy less favorably than diseases or disabilities resulting in a similar inability to work.” Id. 79 Gen. Elec. Co. v. Gilbert, 429 U.S. 125, 128–29 (1976). 42.2 COLUMBIA JOURNAL OF GENDER AND LAW 65 the policy that denied light duty accommodations to pregnant workers while permitting such accommodations for non-pregnant workers similar in their ability to work. Just as the Gilbert court reasoned that a similar policy, without more, did not establish evidence of a disparate impact under the ordinary Title VII framework, a court could determine that the exclusion of pregnancy from light duty benefits does not establish an adverse impact absent further evidence.80 The second clause, which says that pregnant workers must be treated “the same” as non-pregnant workers, must be read to prevent such a result by furnishing substantive protections beyond the ordinary disparate impact framework. Otherwise, the clause would not only be superfluous; it would fail to overturn the Gilbert disparate impact holding.81 The following section elaborates on what the substantive protection that flows from the PDA’s second clause should look like in the disparate impact context. B. How the Second Clause Should Function in Pregnancy Discrimination Disparate Impact Claims Given that a clause is ordinarily read to have consistent usage throughout a statute,82 and given how the Supreme Court interpreted the PDA’s second clause in Young, it is not unreasonable to think that the Young Court’s interpretation of the second clause would be appropriate in the disparate impact context as well. Reading the second clause to have a consistent meaning in both the disparate treatment and disparate impact contexts,83 however, would not make sense. In Young, the Court 80 Id. at 138. 81 This Article relies on the text and purpose of the PDA to reach this conclusion. That the EEOC reached a similar conclusion, though—that the second clause provides additional substantive protection for pregnancy discrimination disparate impact plaintiffs—offers further support for this approach. See Enforcement Guidance on Pregnancy Discrimination and Related Issues, U.S. EQUAL EMPLOYMENT OPPORTUNITY COMM’N, https://www.eeoc.gov/laws/guidance/enforcement-guidance-pregnancy- discrimination-and-related-issues [https://perma.cc/7LWV-3WNN] [hereinafter Enforcement Guidance] (explaining what is “ordinarily” required to prove disparate impact, and then what is required for pregnancy discrimination plaintiffs, implying that the PDA substantively modifies the disparate impact claim); Gilbert, 429 U.S. at 156 (Brennan, J., dissenting) (internal citations omitted) (noting that EEOC interpretations “should receive great deference”). 82 See ANTONIN SCALIA & BRYAN GARDNER, READING LAW: THE INTERPRETATION OF LEGAL TEXTS 170 (2012); Young, 575 U.S. at 229. 83 Young, 575 U.S. at 229. 66 COLUMBIA JOURNAL OF GENDER AND LAW 42.2 viewed the PDA’s second clause as modifying the ordinary Title VII framework and providing that a PDA plaintiff may show that her employer’s proffered non- discriminatory reason for the challenged action is pretextual, inter alia, by producing evidence that the employer imposed a significant “burden” on pregnant women that the employer’s reasons are “not sufficiently strong” to justify.84 While that rule made sense in the disparate treatment context,85 Congress could not possibly have intended to define an adverse “impact” as a “burden”—such a rule would be conclusory and inadministrable. Rather, this Article argues that the PDA’s second clause should be interpreted to mean that a pregnancy discrimination plaintiff may establish her prima facie burden of disparate impact—not only with the usual method of “(1) identify[ing] a specific employment practice or policy; (2) demonstrat[ing] that a disparity [of impact] exists; and (3) establish[ing] a causal relationship between the two”86—but also by demonstrating that the employment practice categorically disqualifies pregnant employees from obtaining a benefit that non-pregnant employees who are “similar in their ability or inability to work”87 may obtain, because of the cause of their conditions. The latter showing would per se satisfy the plaintiff’s prima facie burden to show disparate impact; no additional statistical or evidentiary showing of impact would be necessary.88 A plaintiff who provides evidence of such a practice would establish a prima facie case of disparate impact, and the burden would shift to the employer to justify the impact with a business necessity.89 Importantly, for the same 84 Id. 85 Id. See supra Section II.A. 86 Chin v. Port Auth. of N.Y. & N.J., 685 F.3d 135, 151 (2d Cir. 2021). Accord Griggs v. Duke Power Co., 401 U.S. 424, 436 (1971) (describing that Title VII prohibits employment practices that cause a disparate impact unless the employer demonstrates that they are a business necessity); Enforcement Guidance, supra note 81 (describing the ordinary statistical showings necessary to establish a prima facie case of disparate impact). 87 42 U.S.C. § 2000e(k). 88 See Enforcement Guidance, supra note 81. 89 Grossman & Thomas, supra note 3, at 342 (internal citations omitted) (arguing that “the wholesale unavailability of [benefits for pregnancy-related conditions] poses a prima facie case of disparate impact—thereby shifting the burden to the employer to show that excluding pregnant workers from [the benefit] is a business necessity”). 42.2 COLUMBIA JOURNAL OF GENDER AND LAW 67 reasons as articulated in Young,90 this business necessity cannot merely constitute a “claim that it is more expensive or less convenient to add pregnant women to the category” of those “whom the employer accommodates."91 Under this reading, the second clause provides additional protection, “limited to the Pregnancy Discrimination Act context,”92 for plaintiffs seeking to establish a prima facie burden of disparate impact discrimination. It does not foreclose the ordinary avenues of proving a disparate impact under Title VII, for instance, if a pregnancy discrimination plaintiff does have the relevant statistical evidence as is often necessary to successfully bring a Title VII claim outside of the PDA context; it merely provides her with an additional protection, should she be unable to muster such evidence.93 This additional protection is necessary, as demonstrated above,94 to ensure that the “important congressional objective” of overturning the Gilbert majority’s entire holding—not just its holding with respect to disparate treatment—is carried out.95 The plaintiffs in Gilbert did not have statistical evidence of impact,96 but Congress intended for plaintiffs subject to a policy like the one in Gilbert to be able to demonstrate a prima facie case of disparate impact in the PDA context.97 Furthermore, though this Article relies on the text and purpose of the PDA’s second 90 See supra Section II.A (explaining how the Gilbert Court likely could have reasoned similarly, where the at-issue policy would cost more if pregnant women were covered). 91 Young v. United Parcel Serv., Inc., 575 U.S. 206, 229 (2015). 92 Id. 93 See id. 94 See supra Section II.A. 95 Cf. Young, 575 U.S. at 226. 96 See Gen. Elec. Co. v. Gilbert, 429 U.S. 125, 138 (1976). 97 See supra Section II.A. 68 COLUMBIA JOURNAL OF GENDER AND LAW 42.2 clause to reach this interpretation,98 the fact that the EEOC has similarly interpreted the second clause to provide PDA disparate impact plaintiffs with an additional way to demonstrate disparate impact without producing statistical evidence further demonstrates the wisdom of this approach.99 The strongest argument against the interpretation proposed here is that Congress did not intend for the PDA to enact radical change on the Title VII framework in the pregnancy discrimination context, but rather to restore the status quo prior to the Supreme Court decision in Gilbert.100 This legislative history implies that pregnancy discrimination plaintiffs should not be given additional protections above and beyond what ordinary Title VII plaintiffs have in disparate impact lawsuits—what Justice Scalia dubbed in Young, “most favored employee[]” status.101 This interpretation does not, however, elevate PDA plaintiffs to “most favored employee[]” status;102 it elevates PDA plaintiffs from “least favored” status in the workplace to that of equal stature, as Justice Ginsburg remarked during oral argument in Young.103 Title VII often under-protects plaintiffs suing for pregnancy 98 Cf. Gilbert, 575 U.S. at 162 (Stevens, J., dissenting) (reasoning that Title VII, by its very terms, prohibits the majority’s result in that case, without the need to look to administrative expertise from the EEOC). 99 Specifically, the EEOC’s Enforcement Guidance on Pregnancy Discrimination provides that: Proving disparate impact ordinarily requires a statistical showing that a specific employment practice has a discriminatory effect on workers in the protected group. However, statistical evidence might not be required if it could be shown that all or substantially all pregnant women would be negatively affected by the challenged policy. Enforcement Guidance, supra note 81; see also Gilbert, 575 U.S. at 156 (Brennan, J., dissenting) (internal citations omitted) (explaining that the EEOC’s interpretation is entitled to “great deference”). 100 See, e.g., H.R. REP. NO. 95-948, at 2 (1978). 101 Young v. United Parcel Serv., Inc., 575 U.S. 206, 243 (2015) (Scalia, J., dissenting) (“Prohibiting employers from making any distinctions between pregnant workers and others of similar ability would elevate pregnant workers to most favored employees.”). 102 Cf. id. 103 Transcript of Oral Argument, supra note 17, at 46. 42.2 COLUMBIA JOURNAL OF GENDER AND LAW 69 discrimination, for many reasons that commentators have identified.104 The PDA’s second clause is Congress’s response to “built-in headwinds” that pregnant workers already face in the workplace,105 and therefore actually serves Title VII’s guarantee of equality in the workplace. It is also true that the majority in Young rejected a per se rule in the disparate treatment context that is analogous to the per se approach advocated for here in the disparate impact realm.106 However, the Young Court’s refusal to read the second clause to mean what it says—that pregnant employees must be treated “the same as . . . [non-pregnant employees] similar in their ability or inability to work”—was wrong in the disparate treatment context, and would also be wrong in the disparate impact context.107 In passing the PDA, Congress agreed with the Gilbert dissent’s reasoning, that “pregnancy exclusions built into disability programs both financially burden women workers” and “exacerbat[e] women’s comparatively transient role in the labor force,” such that a narrow additional protection against the categorical exclusion of pregnancy from benefits was a necessary step toward equality.108 The Young Court’s refusal to interpret the PDA’s second clause to do what it says in the 104 See e.g., Grossman & Thomas, supra note 3, at 344 (underscoring how statistical evidence can be difficult to muster in male-dominated fields, for instance). 105 Griggs v. Duke Power Co., 401 U.S. 424, 432 (1971). See generally Hébert, supra note 12 (arguing that disparate impact claims are necessary to eradicate workplace discrimination involving neutral assumptions that employees do not become pregnant). 106 In Young, the plaintiff had argued for a similar rule as proposed here in the disparate treatment context—that a violation is shown if a policy does not treat pregnant employees “the same as other persons . . . similar in their ability or inability to work.” Young, 575 U.S. at 221. Justice Breyer, writing for the majority, rejected that rule in the disparate treatment context, instead adopting a split-the-baby approach that did not produce a violation upon such a showing but instead merely raised a sufficient inference of discrimination to reach a jury on the ultimate issue of whether intentional discrimination occurred. Id. 107 That the Court could not adopt a similar split-the-baby approach in the disparate impact context, where a violation is shown at the prima facie stage, further indicates the lack of wisdom of the Young Court’s approach. Cf. Young, 575 U.S. at 229. Though PDA disparate treatment claims are outside the scope of this Article, the reasoning suggests that the Young Court also should have established a per se approach under that theory to remain true to congressional intent. See H.R. REP. NO. 95-948, at 3 (1978) (explaining that the PDA “makes clear that distinctions based on pregnancy are per se violations of Title VII”). 108 Gen. Elec. Co. v. Gilbert, 429 U.S. 125, 158 (1976) (Brennan, J., dissenting). 70 COLUMBIA JOURNAL OF GENDER AND LAW 42.2 context of disparate treatment is also particularly troubling, given the federal courts’ history of under-protecting pregnancy discrimination plaintiffs despite the intentions of Congress.109 For the foregoing reasons, the second clause should be interpreted to modify the ordinary disparate impact framework and lessen the evidentiary burdens for pregnancy discrimination plaintiffs at the prima facie stage. The following Part, however, illustrates that courts have not yet interpreted the disparate impact framework broadly in this manner. IV. How the Second Clause Functions in Practice: The Second Circuit’s Misapprehension of the PDA’s Second Clause in Legg v. Ulster County The Second Circuit’s recent decision in Legg v. Ulster County demonstrates why the second clause is necessary to modify the disparate impact test as well as the disparate treatment case.110 Legg is the first post-Young federal appellate decision involving a pregnancy discrimination disparate impact claim.111 The Legg court did not apply this Article’s interpretation of the second clause—it did not engage in any analysis whatsoever of how the second clause must function in order to overrule Gilbert. And the Legg holding mirrored the Gilbert holding, finding that an employment practice does not establish a prima facie case of disparate impact discrimination even where there is evidence that the employment policy categorically disqualifies pregnant employees from obtaining a benefit that non- pregnant employees who are “similar in their ability or inability to work”112 may obtain because of the cause of their conditions. In Legg, the Second Circuit applied a disparate impact theory of discrimination as it considered whether a correctional facility policy violated Title VII as amended 109 See, e.g., H.R. REP. NO. 95-948, at 2 (1978) (explaining that the Gilbert majority interpreted Title VII incorrectly, and that the “dissenting Justices correctly interpreted the Act”); see also Young, 575 U.S. at 227 (describing how Justice Scalia “reasoned in part just as” the Gilbert majority did). 110 832 Fed. App’x 727 (2d Cir. 2020). 111 Grossman & Thomas, supra note 3, at 343; Legg, 832 Fed. App’x at 732. 112 Gilbert, 429 U.S. at 138–39; 42 U.S.C. § 2000e(k). 42.2 COLUMBIA JOURNAL OF GENDER AND LAW 71 by the PDA.113 The policy at issue provided that an “employee injured in the line of duty and prohibited from performing full duty status assignments by an attending physician may be placed on line of duty injury leave.”114 A pregnant correctional officer, Ann Marie Legg, sought to be placed on “line of duty injury leave.”115 She produced a doctor’s note that stated, “Ann Marie is able to work at this time but shouldn’t have direct contact with inmates.”116 Soon after, Legg’s supervisor told her that she did not qualify for line of duty injury leave under the policy because “[e]mployees are afforded light duty assignments at the Sheriff’s discretion for work- related injuries/illnesses only.”117 Instead, her supervisor gave Legg “the option of being re-evaluated by [her] attending physician and returning to work full capacity” or utilizing accrued sick time and filing for disability benefits.118 The same day that Legg received this response, she provided her supervisor with a revised evaluation from her physician stating that Legg was “able to work with no restrictions.”119 Legg continued to work; she “again requested a transfer to a position without inmate contact in October 2008” but was denied.120 Having sustained “a serious health- related complication after witnessing an inmate fight” that occurred while she was seven-and-a-half months pregnant, Legg was finally forced to resign until the conclusion of her pregnancy, leaving her unemployed for three-and-a-half months.121 The district court held that Legg failed to establish a prima facie case that the policy had caused a disparate impact because of pregnancy. The court reasoned that the policy could hypothetically treat pregnant corrections officers differently from 113 Legg, 832 Fed. App’x at 731. 114 Id. at 729. 115 Id. at 729–30. 116 Id. at 730. 117 Id. 118 Id. 119 Id. 120 Id. at 732. 121 Id. at 703, 732. 72 COLUMBIA JOURNAL OF GENDER AND LAW 42.2 others similar in their ability or inability to work.122 However, the court nonetheless determined that there was no proof of an adverse impact because the record contained no evidence that Legg or other pregnant correctional officers were unable to work full duty; Legg’s second doctor’s note provided that she could work full duty, and two other officers worked full duty throughout their pregnancies.123 The Second Circuit Court of Appeals affirmed, finding that the district court’s holding was not clearly erroneous.124 The court reasoned that, because Legg did not provide additional evidence beyond her original, ultimately retracted, physician’s note, it was within the district court’s discretion to determine that Legg did not meet her evidentiary burden to show that she was adversely impacted compared to others similar in their “inability to work.”125 The PDA does not establish any “assumptions about a pregnant woman’s ability or inability to work,” the Second Circuit reasoned.126 In order to make out a prima facie case of adverse impact, Legg would have had to produce some “quantum” of evidence that pregnant correctional officers were unable to work full duty in order to satisfy her prima facie burden of adverse impact, something which the court explained she had not done.127 To support this conclusion, the Second Circuit cited two cases in which plaintiffs produced sufficient evidence to show impact, both of which were from outside the pregnancy discrimination context.128 It did not explicitly consider how the PDA’s second clause 122 Legg v. Ulster Cnty., No. 1:09-CV-550, 2017 WL 3207754, at *7 (N.D.N.Y. July 27, 2017). 123 Id. at *8 (citing cases outside the pregnancy discrimination context to find that the plaintiff did not meet her prima facie burden). 124 Legg, 832 Fed. App’x at 731. 125 Id. (concluding that, “to establish a prima facie case, a plaintiff pursuing a disparate impact claim under the Act must demonstrate, at a minimum, that some pregnant women are ‘similar in their ability or inability to work’ to non-pregnant comparators receiving the accommodations sought by the plaintiff”). 126 Id. 127 Id. at 731 (reasoning that “[s]imply proving that all pregnant women will automatically be denied accommodations under the County's Policy is insufficient because it is tantamount to requesting that the district court ‘assume that pregnant women are inherently incapable of working full-duty’”). 128 Id. The Second Circuit distinguished Legg’s evidentiary showing from the plaintiff’s showing in Bradley v. Pizzaco, Inc., 939 F.2d 610, 612–13 (8th Cir. 1991), where the record contained medical evidence that African-American men were more likely to have a condition preventing them from shaving in a race-based disparate impact challenge to a no-beard policy. The Legg court also 42.2 COLUMBIA JOURNAL OF GENDER AND LAW 73 may modify the ordinary Title VII analysis for pregnancy discrimination plaintiffs specifically, nor did it address how plaintiffs should produce such statistical evidence in male-dominated fields where such evidence may be difficult to “adduce[].”129 Under the approach advocated for by this Article,130 the Second Circuit should have found that the district court clearly erred in holding that Legg did not satisfy her prima facie burden of establishing disparate impact. The correctional facility’s policy that only employees injured in the line of duty “may be placed on line of duty injury leave”131 categorically disqualifies pregnant employees from a benefit for which non-pregnant employees “similar in their ability or inability to work”132 are eligible. Non-pregnant employees who are injured on the job are eligible to be considered for injury leave upon submission of a physician’s note purporting that they are unable to work full duty.133 By contrast, pregnant employees are completely ineligible for line of duty injury leave because of the cause of their condition, which inherently arises off the job; a submission of a doctor’s note by a pregnant employee does not trigger the same consideration for line of duty injury leave as it does for a non-pregnant employee. Rather, it triggers an additional burden—produce a revised note or take unpaid leave.134 This categorical disqualification from a benefits program, the PDA’s second clause provides, should per se establish a pregnancy discrimination plaintiff’s burden to show disparate impact.135 While a plaintiff distinguished the record evidence from that in Lynch v. Freeman, 817 F.2d 380, 384–85 (6th Cir. 1987), where women challenging bathroom conditions at a job site demonstrated that they were more likely to suffer urinary tract infections as a result of the policy. 129 Legg, 832 Fed. App’x at 732. 130 See supra Part II. 131 Legg, 832 Fed. App’x at 729. 132 42 U.S.C. § 2000e(k); see also Grossman & Thomas, supra note 3, at 343 (internal citations omitted) (arguing that the “wholesale unavailability of light duty in the physically taxing and dangerous job of correctional officer poses a prima facie case of disparate impact—thereby shifting the burden to the employer to show that excluding pregnant workers from the light duty scheme is a business necessity”). 133 Legg, 832 Fed. App’x at 729. 134 Id. 135 See supra Part II; Grossman & Thomas, supra note 3, at 343. 74 COLUMBIA JOURNAL OF GENDER AND LAW 42.2 outside of the pregnancy discrimination context may need to produce additional statistical evidence of an impact to shift the burden to the employer,136 the second clause provides additional protection to address the “headwinds”137 of “pregnancy exclusions built into disability programs.”138 The Second Circuit misapprehended the second clause not only by requiring some “quantum” of additional evidence beyond the categorical exclusion of pregnant employees, but also by suggesting in dicta that the relevant evidence to show disparate impact would be proof that a substantial number of pregnant workers are unable to work full duty.139 Interpreting the second clause in this manner, the Second Circuit found that Legg’s claim necessarily failed due to a lack of record evidence that she was unable to work full duty, because her revised doctor’s note was subject to two reasonable interpretations.140 However, the second clause should not be read to limit PDA plaintiffs to showing impact on one’s ability to work; Legg was adversely impacted by the policy even if she technically could work full duty, just as the Gilbert plaintiffs were adversely impacted by the policy there even if they were technically able to continue working throughout their pregnancies without taking disability leave.141 In both cases, the impact occurred not because of an inability to work without accommodations—it arose because pregnant workers were categorically excluded from obtaining a benefit that non-pregnant employees similar in their ability or inability to work could receive. This exclusion, Congress provided in the PDA’s second clause, causes an inherent prima facie adverse impact by “financially burden[ing] women workers” and “exacerbat[ing] women’s 136 See Enforcement Guidance, supra note 81. 137 Griggs v. Duke Power Co., 401 U.S. 424, 432 (1971). 138 Gen. Elec. Co. v. Gilbert, 429 U.S. 125, 158 (1976) (Brennan, J., dissenting). 139 Legg, 832 Fed. App’x at 731. 140 Id. 141 Cf. Gilbert, 429 U.S. at 132 (explaining that most pregnancies caused an inability to work, but finding that pregnancies are often subject to different levels of disability). Congress intended to overrule Gilbert. See supra Section II.B. But if evidence that some pregnant employees could keep working without an accommodation defeats every pregnant employee’s claim, the case would not be overturned. See Gilbert, 429 U.S. at 132. 42.2 COLUMBIA JOURNAL OF GENDER AND LAW 75 comparatively transient role in the labor force,”142 which is exactly what happened to Ann Marie Legg when her employer’s policy caused her to take unpaid leave for the last three-and-a-half months of her pregnancy.143 As a result, the Second Circuit was incorrect to subject PDA plaintiffs to the same statistical burdens as plaintiffs outside the pregnancy discrimination context, and should not have relied on non-pregnancy-related cases to deny Legg’s claim.144 Moreover, it also misinterpreted the second clause when it explained in dicta that showing an adverse impact would require proof that pregnant employees are unable to work full duty.145 The evidence of a categorical exclusion that Legg produced should have been sufficient proof of impact,146 but instead the Second Circuit mimicked the Gilbert majority and held that it was not. CONCLUSION This Article argues that the PDA’s second clause should theoretically provide substantive protection to pregnancy discrimination plaintiffs alleging that a denial of an accommodation caused a disparate impact.147 The Second Circuit’s decision in Legg, however, demonstrates that federal courts are not interpreting the second clause to provide such protections.148 The continued barriers that pregnancy discrimination plaintiffs face when suing under a disparate impact theory of discrimination, in addition to those barriers present in the disparate treatment 142 See 42 U.S.C. § 2000e(k); Gilbert, 429 U.S. at 158 (Brennan, J., dissenting). Accord H.R. REP. NO. 95-948, at 2 (1978). 143 Legg, 832 Fed. App’x at 703, 732. 144 Legg, 832 Fed. App’x at 732. This reasoning also flouts the Young Court’s holding that the burden of making out a prima facie Title VII discrimination claim must not be “onerous.” See Young v. United Parcel Serv., Inc., 575 U.S. 206, 228 (2015). 145 Legg, 832 Fed. App’x at 732. 146 Grossman & Thomas, supra note 3, at 343; supra Section II.B. 147 See supra Part II. 148 See supra Part III. 76 COLUMBIA JOURNAL OF GENDER AND LAW 42.2 context,149 demonstrate the need for new comprehensive federal pregnancy discrimination legislation.150 Ideally, this new legislation would be less ambiguous regarding the scope and type of protection its provisions provide, so that it would be more effective than the PDA’s second clause has been in promoting workplace equality.151 149 See generally Grossman & Thomas, supra note 3. 150 See Bakst et al., supra note 9 (arguing that Congress should pass the Pregnant Workers Fairness Act, which is currently pending in the Senate, to ensure that pregnant employees obtain equal stature in workplaces). The lack of constitutional protections for pregnancy-related discrimination further underscores the need for strong statutory protections. See Geduldig v. Aiello, 417 U.S. 484, 497 (1974). 151 Compare 42 U.S.C § 2000e(k), which the Supreme Court has called “less [than] clear.” Young v. United Parcel Serv., Inc., 575 U.S. 206, 219 (2015).