JUDICIAL FEDERALISM AND CAUSATION IN STATE EMPLOYMENT DISCRIMINATION STATUTES

JUDICIAL FEDERALISM AND CAUSATION IN STATE EMPLOYMENT DISCRIMINATION STATUTES

Most state and federal employment discrimination statutes prohibit employers from making certain decisions “because of” an employee’s protected characteristics or activities. Courts interpreting this language have developed a number of frameworks and standards to assess whether a plaintiff has demonstrated the causation required to make out a claim of employment discrimination. Two standards frequently invoked by courts are but-for causation and the less stringent motivating-factor standard. In the last decade, Supreme Court decisions involving the Age Discrimination in Employment Act and Title VII’s retaliation provisions have interpreted “because of” to require but-for causation. In so holding, the Court all but repudiated the motivating-factor standard for statutes that do not expressly embrace it.

In the wake of these decisions, how should state courts interpret “because of” in state discrimination statutes, and how much weight should be given to case law interpreting federal statutes? State courts confronting these questions have exhibited a range of responses, from absolute independence to total incorporation of federal standards. This Note surveys state courts’ various approaches and assesses them in the context of debates about the “new judicial federalism”—the practice of interpreting state constitutional provisions as more protective of individual rights than their federal analogues. When it comes to causation in employment discrimination statutes, this Note argues, the benefits of independent construction outweigh the costs. A presumption of independent construction better serves the values of state sovereignty, policy experimentation, and interjurisdictional dialogue that undergird our federal system.

Introduction

In June 2017, the Supreme Court of Iowa ordered a new trial in Tina Haskenhoff’s employment discrimination lawsuit against her former employer, Homeland Energy Solutions, LLC. 1 See Haskenhoff v. Homeland Energy Sols., LLC, 897 N.W.2d 553, 601 (Iowa 2017). Haskenhoff alleged sexual harassment and retaliation under the Iowa Civil Rights Act (ICRA). 2 Id. at 565. The trial court had instructed the jury that, to prevail on her retaliation claim, Haskenhoff needed to prove that her complaints about sexual harassment “played a part” in Homeland’s adverse employment decision. 3 See id. at 567–68. Notably, the court did not instruct the jury that the defendant could avoid liability by proving that it would have made the same decision absent the retaliatory motive—the so-called “same decision” defense common in motivating-factor frameworks. See id.; infra note 46 and accompanying text. The jury returned a verdict for Haskenhoff, which the high court vacated. 4 Haskenhoff, 897 N.W.2d at 561–62. The court held that while the “motivating factor” causation standard in the instruction may have been proper for status-based discrimination claims such as sex discrimination, retaliation claims required the higher “significant factor” standard. 5 Id. at 581–86. Had the court treated the claim as the U.S. Supreme Court treats retaliation claims under Title VII, it would have applied yet another standard: but-for causation. 6 See Univ. of Tex. Sw. Med. Ctr. v. Nassar, 133 S. Ct. 2517, 2533 (2013); see also infra section I.D.

The facts surrounding Haskenhoff’s employment and trial illustrate why these competing causation standards matter. Haskenhoff’s supervisor, Kevin Howes, engaged repeatedly in grossly inappropriate behavior, including commenting on Haskenhoff’s body, speculating out loud what it would be like to have sex with her, simulating sexual behavior in front of her, and suggesting that she was getting married for money. 7 Haskenhoff, 897 N.W.2d at 562–63. After this last incident, Haskenhoff called Howes expletives in front of her coworkers, went home early without permission, and then later sent Howes an email indicating her disgust. 8 Id. at 563–64. Howes prepared a written warning for Haskenhoff for leaving work early and indicated that he wanted to terminate Haskenhoff for insubordination. 9 Id. at 564–65. In the meantime, Haskenhoff complained of sexual harassment to the company’s management, which determined that Howes had made inappropriate comments in the workplace. 10 Id. A day after Howes interrupted a meeting to present Haskenhoff with a draft of a performance-improvement plan, Haskenhoff resigned. 11 Id. at 565. On the instruction that Haskenhoff need only prove that her complaints “played a part” in the adverse employment decision, the jury returned a verdict for Haskenhoff on the retaliation count. 12 Id. at 567–69. But one can see why it might be more difficult to prevail under a higher causation standard: She left work early without permission, openly used expletives to refer to her supervisor, and was under investigation for further evidence of insubordination. 13 See id. at 564. If this conduct turned out to be an independently sufficient reason to fire her, Haskenhoff would likely not prevail under a but-for causation standard.

The court in Haskenhoff tackled a question that now confronts many state courts: In the wake of a pair of recent Supreme Court decisions applying but-for causation to Title VII retaliation 14 See Univ. of Tex. Sw. Med. Ctr. v. Nassar, 133 S. Ct. 2517, 2533 (2013). and Age Discrimination in Employment Act (ADEA) 15 See Gross v. FBL Fin. Servs., Inc., 557 U.S. 167, 176 (2009). claims, how should state courts interpret similarly worded state statutes? 16 See infra Part II. Like  their  federal  analogues,  most  state  employment  discrimination  statutes  outlaw  making  certain  employment  decisions  “because  of”  an  individual’s  protected  characteristics  or  activities. 17 See Jerome Hunt, Ctr. for Am. Progress Action Fund, A State-by-State Examination of Nondiscrimination Laws and Policies 22–80 (2012), https://www.americanprogress.org/wp-content/uploads/issues/2012/06/pdf/state_nondiscrimination.pdf [https://perma.cc/A8RQ-UQ6E] (cataloguing state employment discrimination statutes). State  courts  looking  to  the  U.S.  Supreme  Court  for  persuasive  authority 18 Of course, Supreme Court cases cannot be binding authority on the construction of state statutes because state courts are the “ultimate expositors of state law.” Mullaney v. Wilbur, 421 U.S. 684, 691 (1975). now find two distinct interpre­tations of “because of” in the case law interpreting federal employment discrimination statutes. 19 Compare Price Waterhouse v. Hopkins, 490 U.S. 228, 240 (1989) (plurality opinion) (Brennan, J.) (rejecting the “but-for” interpretation of “because of” in Title VII), with Nassar, 133 S. Ct. at 2533 (interpreting “because of” as requiring but-for causation for the retaliation provisions of Title VII), and Gross, 557 U.S. at 176 (interpreting “because of” as requiring but-for causation for the ADEA).

State courts have exhibited a variety of responses, ranging from automatic application of the new federal standards 20 See, e.g., Hatheway v. Bd. of Regents of the Univ. of Idaho, 310 P.3d 315, 323 (Idaho 2013) (“[W]e apply the quantum of proof and standards promulgated in discrimination cases arising under the ADEA[, such as Gross].”). to completely independent construction of state statutes. 21 See, e.g., Smith v. Anchorage Sch. Dist., 240 P.3d 834, 842 (Alaska 2010) (“We decline to follow Gross . . . .”). Recently, a growing number of state appellate and supreme courts have applied the but-for requirement to state statutes, with varying degrees of attentiveness to the question of exactly how much deference state courts interpreting state statutes owe to the Supreme Court’s interpretations of analogous federal statutes. 22 See infra Part II.

This Note surveys and analyzes these state decisions in the context of debates about parallel construction and what has sometimes been called the “new” judicial federalism. 23 The term “new judicial federalism” refers to a trend among state courts of relying on state constitutions rather than the Federal Constitution to protect individual rights. See infra section III.A. The “new” judicial federalism is, by now, quite old. Although the discourse around the “new judicial federalism” focused initially on state constitutions, see infra section III.A, this Note uses the term in both the statutory and constitutional contexts. As in the case of state constitutions, state employment discrimination statutes protect individual employees’ rights in a way that may overlap with similarly worded federal statutes. For that reason, the statutory context raises many of the same considerations as the constitutional one. See generally Susan P. Fino, The Role of State Supreme Courts in the New Judicial Federalism (1987) (presenting an empirical study of “state supreme courts and their performance, especially in the area of the development of an independent and adequate body of state constitutional law”); Michael E. Solimine & James L. Walker, Respecting State Courts: The Inevitability of Judicial Federalism (1999) (studying the role of state courts in our bifurcated judicial system); Lawrence Friedman, The Constitutional Value of Dialogue and the New Judicial Federalism, 28 Hastings Const. L.Q. 93, 94–95 & n.5 (2000) (outlining the “lockstep,” “criteria,” and “primacy” approaches that state courts have followed in construing state constitutional provisions that parallel provisions of the federal Bill of Rights); Alex B. Long, “If the Train Should Jump the Track . . . ”: Divergent Interpretations of State and Federal Employment Discrimination Statutes, 40 Ga. L. Rev. 469, 480–505 (2006) (describing the rise of the new judicial federalism in constitutional law and noting similar dynamics in the employment discrimination context). Although state courts are not always explicit about all the considerations that enter into the decision of whether to follow the Supreme Court’s construction of a federal statute when construing a state analogue, the various arguments for either independent or parallel construction provide a framework for assessing these state-court decisions. 24 For an overview of the various arguments on both sides of the judicial federalism debate, see Long, supra note 23, at 505–39.

Part I of this Note provides an overview of current causation doctrine in federal employment discrimination law. Part II analyzes various state-court decisions addressing the question of whether to follow Gross v. FBL Financial Services, Inc. and University of Texas Southwestern Medical Center v. Nassar in applying but-for causation requirements to state employment discrimination statutes, including age discrimination and retaliation statutes analogous to those at issue in Gross and Nassar. From this survey, the Note concludes that, whether they decide to adhere to or depart from the GrossNassar standard, state courts rarely attend to the full range of policy considerations that underlie the judicial federalism debate and that ought to enter into their decisional calculus. Part III discusses the major arguments for and against parallel construction of similarly worded state and federal statutes. It concludes that the typical arguments for parallel construction—that it increases efficiency, better satisfies legislative preferences, and preserves the public’s perception of judicial legitimacy—carry little weight in deciding between competing causation standards. The benefits of independent construction, on the other hand—policy experimentation and fidelity to state policy preferences—are strong. Accordingly, this Note argues that state discrimination statutes should be interpreted independently of Gross and Nassar.

I. Causation in Federal Employment Discrimination Law

This Part provides a brief historical overview of the development of causation doctrine in federal employment discrimination law. This overview is crucial for understanding why different standards apply under different statutes, even when the statutes use equivalent language in describing the required causation. 25 See infra section I.D (describing the recent developments of Gross and Nassar). See generally Deborah A. Widiss, Undermining Congressional Overrides: The Hydra Problem in Statutory Interpretation, 90 Tex. L. Rev. 859, 863–64 (2012) (describing the “hydra problem” of statutory interpretation that results from complicated legislative history involving back-and-forth between courts and legislatures). Section I.A describes the classic burden-shifting framework for analyzing Title VII claims laid out by the Supreme Court in McDonnell Douglas Corp. v. Green 26 411 U.S. 792 (1973). and  notes  the  scholarly  and  judicial  debates  about  the  relationship  between  that  framework  and  causation  standards.  Section  I.B  describes  the  Court’s  application,  in  Price  Waterhouse  v.  Hopkins, 27 490 U.S. 228 (1989). of a motivating-factor analysis and the rejection of but-for causation by a plurality of the Court. Section I.C discusses the congressional adoption of the motivating-factor framework in the Civil Rights Act of 1991, which endorsed a version of this standard for the so-called “status-based discrimination” provisions of Title VII. Finally, section I.D discusses the recent and controversial Gross and Nassar cases, in which the Supreme Court held that the motivating-factor standard is unavailable to plaintiffs under the ADEA and the Title VII retaliation provision, in part because these statutes were not amended in 1991.

A. The Traditional Framework: The McDonnell Douglas Approach to Disparate Treatment Cases Under Title VII

Under Title VII of the Civil Rights Act of 1964, it is unlawful for an employer to discharge or refuse to hire an individual, or otherwise discriminate with respect to an individual’s terms of employment, “because of such individual’s race, color, religion, sex, or national origin.” 28 42 U.S.C. § 2000e-2(a)(1) (2012). While this provision indicates a clear congressional denunciation of employment discrimination on the basis of the listed factors, the precise contours of the prohibited conduct are less self-evident. Exactly what kinds of decisions on the part of employers can fairly be said to be “because of” these factors?

Intimately  tied  to  the  causation  question  is  the  issue  of  evidence.  If  the  plaintiff  in  an  employment  discrimination  suit  must  show  some  sort  of  causal  link  between  her protected status or activity and the adverse employment action she is challenging, what kinds of evidence does she need to bring to satisfy the causation requirement? 29 The Second Circuit illustrated one type of evidentiary difficulty when it distinguished between the admission “I fired him because he was too old” and the statement “You’re fired, old man.” The latter, though highly probative, nonetheless provides only indirect evidence of causation. See Tyler v. Bethlehem Steel Corp., 958 F.2d 1176, 1185 (2d Cir. 1992). Of course, as the Second Circuit noted, juries have always been allowed to make inferences from indirect evidence. See id.

In light of these difficulties, courts have long recognized the need for robust tools for determining liability in employment discrimination cases, developing various doctrinal frameworks for assessing discrimination claims in the absence of direct evidence of discriminatory intent. 30 See 1 Lex K. Larson & Arthur Larson, Employment Discrimination § 8.01 (2d ed. 2018) (“[The McDonnell Douglas] approach is useful in cases where the plaintiff does not have, at the initial stages of litigation, enough direct or circumstantial evidence showing that an adverse employment action was motivated by intentional discrimination.”). The most well-known of these is the burden-shifting framework laid out in McDonnell Douglas Corp. v. Green. 31 411 U.S. at 802–04. Under McDonnell Douglas, the plaintiff bears the initial burden of establishing a prima facie case of discrimination, which in McDonnell Douglas involved showing that: (1) the plaintiff belonged to a racial minority group; (2) he was qualified and applied for the position for which the employer was soliciting applicants; (3) he was rejected despite his qualifications; and (4) the position remained open after his rejection, and the employer continued to solicit applications from candidates with the individual’s qualifications. 32 Id. at 802.

Once the plaintiff establishes a prima facie case, the burden of production shifts to the defendant to articulate a “legitimate, nondiscrim­inatory reason” for the challenged employment action. 33 Id. By carrying this burden, the defendant rebuts the presumption created by the prima facie case, and the plaintiff must then ultimately prove that the proffered legitimate reason was pretextual and that the real reason for the employment action was discrimination. 34 Id. at 804. The burden of persuasion remains always with the plaintiff. 35 St. Mary’s Honor Ctr. v. Hicks, 509 U.S. 502, 511 (1993). There is some scholarly disagreement about the role of causation in the McDonnell Douglas framework. 36 Compare, e.g., Starceski v. Westinghouse Elec. Corp., 54 F.3d 1089, 1095 n.4 (3d Cir. 1995) (stating that the McDonnell Douglas framework requires plaintiffs to prove but-for causation), Robert A. Kearney, The High Price of Price Waterhouse: Dealing with Direct Evidence of Discrimination, 5 U. Pa. J. Lab. & Emp. L. 303, 310 (2003) (“[A] plaintiff relying on McDonnell Douglas would need evidence that her race was a ‘determining,’ or ‘but-for’ factor in the decision.” (footnote omitted)), and Michael J. Zimmer, The New Discrimination Law: Price Waterhouse Is Dead, Whither McDonnell Douglas?, 53 Emory L.J. 1887, 1930 (2004) (“[I]n McDonnell Douglas cases, the courts have typically required the plaintiff to prove that discriminatory motivation was the but-for or the determinative influence in the employer’s decision.”), with Martin J. Katz, Reclaiming McDonnell Douglas, 83 Notre Dame L. Rev. 109, 124–38 (2007) (arguing that the McDonnell Douglas framework does not prove or require but-for causation). Neither McDonnell Douglas nor any of the major Supreme Court cases elaborating on the McDonnell Douglas framework explicitly mention a particular causation standard. 37 See Hicks, 509 U.S. at 511; Tex. Dep’t of Cmty. Affairs v. Burdine, 450 U.S. 248, 252–53 (1981); McDonnell Douglas, 411 U.S. at 802. Nevertheless, the predominant view has equated McDonnell Douglas with but-for causation. 38 See Katz, supra note 36, at 116 (describing the “myth” that McDonnell Douglas proves or requires but-for causation as “universally held”).

B. Rejecting the But-For Requirement: Price Waterhouse and the Mixed-Motive Framework

A frequent criticism of McDonnell Douglas is that it presumes employers have a single rationale for all their decisions, ignoring the complex realities of human decisionmaking and workplace dynamics. 39 See, e.g., Dare v. Wal-Mart Stores, Inc., 267 F. Supp. 2d 987, 991–92 (D. Minn. 2003) (“In practice, few employment decisions are made solely on [the] basis of one rationale to the exclusion of all others. Instead, most employment decisions are the result of the interaction of various factors, legitimate and at times illegitimate, objective and subjective, rational and irrational.”). This perceived shortcoming of the framework was one of the driving forces behind the development of a subsequent important federal framework, commonly known as “mixed motive.” In Price Waterhouse, the Court was faced with a sex discrimination claim by a female employee who had been passed over for partnership in an accounting firm. 40 Price Waterhouse v. Hopkins, 490 U.S. 228, 231–32 (1989) (plurality opinion) (Brennan, J.). Although the defendant had persuaded the district court judge that it legitimately refused to promote the plaintiff to partnership because of the plaintiff’s poor interpersonal skills, Justice Brennan’s plurality opinion noted that there were “clear signs” that “some of the partners reacted negatively to Hopkins’ personality because she was a woman.” 41 Id. at 235–36.

Crucially, the plurality in Price Waterhouse opined on the meaning of “because of” in Title VII: “We take these words to mean that [the protected characteristic] must be irrelevant to employment decisions.” 42 Id. at 240. The plurality explicitly rejected Price Waterhouse’s argument that but-for causation should apply, arguing that Congress’s intent to prohibit employers from taking sex, race, religion, and national origin into account in making employment decisions “appears on the face of the statute.” 43 See id. at 239–40. Invoking “common sense,” Justice Brennan’s opinion argued that it would strain credulity to conclude that Congress, with its use of the simple “because of” language, intended to require Title VII plaintiffs to disentangle the legitimate and illegitimate motivations for an employment decision and prove that the illegitimate motivations rose to the level of but-for causation. 44 Id. at 241–42.

In addition to this common-sense or plain-meaning argument, the plurality drew on the statements of various legislators before the passage of the Act to conclude that the purpose of Title VII was to prohibit employers from taking the enumerated characteristics (race, color, sex, etc.) into account when making employment decisions and to promote hiring based on job qualifications. 45 Id. at 243–44. Because of this, the Court held that if a plaintiff showed that her gender played a motivating part in the challenged decision, as Hopkins had, then the defendant could only avoid liability by prevailing on a “same-decision” defense—that is, by proving that it would have made the same decision even had it not taken the protected characteristic into consideration. 46 Id. at 242, 258; id. at 259–60 (White, J., concurring in the judgment); id. at 261 (O’Connor, J., concurring in the judgment). In other words, the “same decision” defense means that the defendant bears the burden of disproving but-for causation. See William R. Corbett, Babbling About Employment Discrimination Law: Does the Master Builder Understand the Blueprint for the Great Tower?, 12 U. Pa. J. Bus. L. 683, 721 (2010) (“[T]he full mixed-motives analysis is still a but-for test; mixed-motives analysis simply bifurcates causation into two parts (motivating factor and same decision) and shifts the burden of persuasion to the defendant on the second part to disprove but-for causation.”). Although the difference between interpreting “because” to require the plaintiff to prove but-for causation and allowing the defendant the affirmative defense of disproving but-for causation may seem purely technical, the difficult evidentiary problems of employment discrimination cases make such burden shifting significant. See Gross v. FBL Fin. Servs., Inc., 557 U.S. 167, 190–91 (2009) (Breyer, J., dissenting) (arguing that it is difficult to prove but-for causation when considering the mental states that constitute motive and that employers are in a better position to prove how they would have acted in the hypothetical circumstance in which an impermissible motive was absent).

Under McDonnell Douglas, Hopkins likely would have lost because she would not have been able to show that the employer’s proffered reason, her poor interpersonal skills, was pretext. 47 See McDonnell Douglas Corp. v. Green, 411 U.S. 792, 804 (1973). Indeed, the district court judge in Hopkins’s case specifically found that Price Waterhouse had not invented its reservations about Hopkins’s interpersonal skills as pretext for discrimination. 48 Price Waterhouse, 490 U.S. at 236 (plurality opinion) (Brennan, J.). Moreover, the district judge intimated that Hopkins would not have satisfied a but-for causation requirement: “Because plaintiff had considerable problems dealing with staff and peers, the Court cannot say that she would have been elected to partnership if the Policy Board’s decision had not been tainted by sexually based evaluations.” 49 Hopkins v. Price Waterhouse, 618 F. Supp. 1109, 1120 (D.D.C. 1985). For that reason, the dissent would have directed judgment for Price Waterhouse. 50 Price Waterhouse, 490 U.S. at 295 (Kennedy, J., dissenting).

C. Congress Adopts Mixed-Motive: The Civil Rights Act of 1991

Two years after Price Waterhouse, Congress passed the Civil Rights Act of 1991. 51 Civil Rights Act of 1991, Pub. L. No. 102-166, 105 Stat. 1071 (codified as amended in scattered sections of 42 U.S.C.). Among other things, the Act codified an altered version of the mixed-motive method of proof established in Price Waterhouse. 52 See id. § 107, 105 Stat. at 1075; see also Tristin K. Green, Making Sense of the McDonnell Douglas Framework: Circumstantial Evidence and Proof of Disparate Treatment Under Title VII, 87 Calif. L. Rev. 983, 993–96 (1999) (noting that the Civil Rights Act of 1991 replaced the Price Waterhouse “mixed-motive” method of proof with a similar “motivating factor” method). As in Price Waterhouse, the amended Title VII stated that “an unlawful employment practice is established when the complaining party demon­strates that race, color, religion, sex, or national origin was a motivating factor for any employment practice, even though other factors also motivated the practice.” 53 42 U.S.C. § 2000e-2(m) (2012). However, unlike in Price Waterhouse, the 1991 amendments did not allow for the employer to prevail through the same-decision defense; rather, even if the employer could prove that it would have made the same decision absent the impermissible motive, this would not absolve the defendant of liability but merely limit the plaintiff to declaratory and injunctive relief, as well as attorneys’ fees and costs. 54 Id. § 2000e-5(g)(2)(B). For the five status-based discrimination claims listed in 42 U.S.C. § 2000e-2, then, the 1991 Act made Price Waterhouse obsolete. However, Congress did not make such an addition to various other federal employment discrimination statutes, including the Title VII antiretaliation provision 55 See id. § 2000e-3(a). and the ADEA. 56 See 29 U.S.C. § 623(a) (2012).

D. The Shifting Federal Tide: Gross, Nassar, and the Rise of But-For Causation

The Title VII amendments led to confusion about the reach of the Price Waterhouse decision. 57 Prior to the 1991 amendments, the application of Price Waterhouse was already a topic of considerable disagreement in the lower courts because of the fractured opinions. Most lower courts treated Justice O’Connor’s concurrence, requiring that the plaintiff show discrimination by “direct evidence” to be entitled to a mixed-motive burden shift, as controlling. See Green, supra note 52, at 992–93 (summarizing the different interpretations of Price Waterhouse’s holding among lower courts). Even this approach, however, was riddled with ambiguity, as some courts reasoned that “direct evidence” could not possibly mean “direct evidence,” given the extreme rarity of such evidence in discrimination cases. See, e.g., Tyler v. Bethlehem Steel Corp., 958 F.2d 1176, 1185 (2d Cir. 1992). Because Congress amended only the Title VII status-based provisions in 1991, where did that leave other federal statutes using the “because of” language to describe the required nexus between protected characteristics or activities and adverse employment actions?

The Supreme Court addressed this question head-on in Gross v. FBL Financial Services, Inc. 58 557 U.S. 167 (2009). Noting that Congress neglected to add a motivating-factor provision to the ADEA in 1991 when it added such a provision to Title VII, even though it amended the ADEA in other ways, the Court concluded that Title VII jurisprudence did not control its interpretation of the ADEA. 59 Id. at 174 –75. With Price Waterhouse out of the picture, the Court turned to the dictionary to ascertain the meaning of “because” and concluded that it meant but-for causation. 60 Id. at 176. Strangely, both parties and both lower courts seemed to assume that Price Waterhouse governed the interpretation of “because of” in the ADEA. In holding that a but-for requirement applied, the Court ignored the question posed in its own grant of certiorari. See Martin J. Katz, Gross Disunity, 114 Penn St. L. Rev. 857, 865–67 (2010) [hereinafter Katz, Gross Disunity]. It bears noting that, apart from the dictionary, which would presumably lead to the same interpretation of “because” no matter what statute it appears in, the only affirmative hint relied upon by the Court to discern the ADEA’s meaning was Congress’s failure to amend it in 1991. Though it is not clear precisely how much the holding depends on this fact, it is fair to say that the Gross Court placed less weight on the intent of the enacting legislature than the Price Waterhouse Court did. 61 Compare Gross, 557 U.S. at 174 (surveying no affirmative evidence of congressional intent, only the negative evidence of not amending the statute), with Price Waterhouse v. Hopkins, 490 U.S. 228, 243 (1989) (plurality opinion) (Brennan, J.) (citing various legislators’ statements as evidence of the congressional intent behind Title VII). Because the Court in Gross largely relied on the dictionary definition of “because,” it perhaps signaled that it would be willing to apply a similar reading to other similarly drafted statutes. 62 See Katz, Gross Disunity, supra note 60, at 859 & n.8 (arguing that the Court’s reasoning in Gross suggests that it is likely to apply the same standard to other statutes and noting cases in which lower courts have done so).

In 2013, in University of Texas Southwestern Medical Center v. Nassar, the Court did just that with respect to the Title VII antiretaliation provision. 63 See 133 S. Ct. 2517, 2533 (2013). Noting that Congress added the motivating-factor provision to the section of the statute dealing with status-based discrimination rather than a section dealing with all unlawful employment actions or with retaliation, the Court held again that but-for causation was the appropriate standard. 64 Id. at 2532. Citing Gross as persuasive authority for the meaning of “because of” and the significance of the 1991 amendments, the Court indicated its willingness to extend the Gross line of reasoning to other statutes. 65 Id. at 2527–28. And it went further by opening its analysis with the assertion that “[c]ausation in fact . . . is a standard requirement of any tort claim. This includes federal statutory claims of workplace discrimination.” 66 Id. at 2524–25 (citations omitted). Moreover, it all but repudiated Price Waterhouse by noting that “there is no reason to think that the different balance [of a lessened causation requirement and an affirmative defense] articulated by Price Waterhouse somehow survived [the 1991 Act’s] passage.” 67 Id. at 2534. After Nassar, it is doubtful that a motivating-factor analysis is available under any federal statute except those covered by the 1991 amendments, regardless of any arguments about congressional intent.

The Court’s decisions in Gross and Nassar have been subject to vigorous critique among scholars and plaintiff-side practitioners. 68 See, e.g., Brief of the Washington Lawyers Committee for Civil Rights & Urban Affairs et al. as Amici Curiae in Support of Respondent at 2, Nassar, 133 S. Ct. 2517 (No. 12-484), 2013 WL 1367758 (arguing that Gross was wrongly decided); Katz, Gross Disunity, supra note 60, at 857–58 (arguing that the Gross Court’s rejection of uniformity across statutes was “normatively problematic”); August T. Johannsen, Note, Mitigating the Impact of Title VII’s New Retaliation Standard: The Americans with Disabilities Act After University of Texas Southwestern Medical Center v. Nassar, 56 Wm. & Mary L. Rev. 301, 339 (2014) (“The Court in Nassar either misread or ignored the plain intent of Title VII by holding retaliation to a different, more stringent standard than status-based discrimination.”). One common criticism is that the Court committed precisely the error that the Price Waterhouse plurality warned against: Taking “because of” as a shorthand for but-for causation misunderstands the intent that the plurality found apparent on the face of Title VII—namely, that protected characteristics were to be irrelevant in employment decisions. 69 See, e.g., Johannsen, supra note 68, at 316 (critiquing Gross and Nassar along these lines). More pragmatically, the widespread perception was that Gross and Nassar made it significantly more difficult for plaintiffs to prevail in ADEA and Title VII retaliation suits. 70 See, e.g., Katherine Stark Todd, Note, But-For Nassar, There Would Not Be a Causation Conundrum in Title VII Retaliation Litigation: How University of Texas Southwest Medical Center v. Nassar Makes It Harder for Employees to Prevail, 21 Suffolk J. Trial & App. Advoc. 288, 311–12 (2016) (describing the standard in Gross and Nassar as “unrealistic . . . due to the nature of employment decisions where multiple causal factors come into play”). But see David Sherwyn & Michael Heise, The Gross Beast of Burden of Proof: Experimental Evidence on How the Burden of Proof Influences Employment Discrimination Outcomes, 42 Ariz. St. L.J. 901, 926–44 (2010) (presenting the results of an empirical study indicating that motivating-factor and McDonnell Douglas–based pretext jury instructions resulted in similar rates of favorable liability determinations but that plaintiffs who received motivating-factor instructions were much more likely to be awarded costs and attorneys’ fees); Blair Druhan Bullock, Judicial and Agency Enforcement of Nondiscrimination Laws 85–87 (May 2015) (unpublished Ph.D. dissertation, Vanderbilt University), https://law.vanderbilt.edu/phd/students/files/Bullock.pdf [https://perma.cc/F77Y-FKHS] (providing empirical evidence suggesting that the adoption of mixed-motive standards for certain claims increases the number of those claims that are brought and that frivolous claims largely account for the change). Those who agree with the Price Waterhouse plurality that the motivating-factor standard is the common-sense interpretation of “because of” in employment discrimination statutes have reason to lament the developments of Gross and Nassar: In both cases a 5-4 majority repudiated that interpretation and inserted in its place the tort concept of but-for causation. 71 See Sandra F. Sperino, Discrimination Law: The New Franken-Tort, 65 DePaul L. Rev. 721, 721 (2016) [hereinafter, Sperino, Franken-Tort] (“[T]he Court imported the concept of factual cause into discrimination law.” (emphasis added)).

That is not to say that plaintiffs seeking the benefit of the mixed-motive framework are out of luck for all claims except the status-based discrimination claims covered by the 1991 amendments. A client alert published by a defense-side firm shortly after the Nassar decision warned that, because subnational laws might still be analyzed using the motivating-factor standard, the Supreme Court’s holding in Nassar “ultimately may not reduce the number of retaliation and harassment claims lodged against employers. Rather, [Nassar] may serve only to push plaintiffs away from Title VII and towards any applicable state and local equivalents . . . .” 72 Julia E. Judish et al., Pillsbury Winthrop Shaw Pittman LLP, Impact of Supreme Court Pro-employer Title VII Decisions Blunted by State Laws 5 (2013), http://www.pillsburylaw.com/images/
content/4/3/v2/4318/Alert20130708LitigationImpactofSupremeCourtTitleVIIRulingsBlunte.pdf [http://perma.cc/LC7M-BXV8].
Of course, this strategy remains available only to the extent that state courts retain the motivating-factor standard in interpreting their state statutes after Gross and Nassar. In confronting this question, state courts—whether explicitly or implicitly—must contend with their role within our system of dual sovereignty. State courts may afford differing levels of deference to federal law, and their reasons for doing so may vary. How they approach these issues has profound implications for the federal system. It is to these state-court decisions that this Note now turns.

II. Interpreting State Employment Discrimination Statutes

This Part examines state-court interpretations of state employment discrimination statutes in the wake of Gross and Nassar. Although these courts interpret different statutes and differ considerably in the weight that they give to federal precedent in interpreting those statutes, a relatively circumscribed set of jurisprudential and policy concerns underlies the decision in each case to follow or depart from federal doctrine. These concerns track broader debates about judicial federalism, balancing the values of state sovereignty, interjurisdictional dialogue, and experimentation against considerations of uniformity, efficiency, and perceptions of judicial legitimacy. 73 See infra Part III. Nevertheless, as this Part will detail, these concerns rarely come to the surface. Rather, in deciding to follow or depart from federal law, state courts often give little explicit consideration to the weighty issues on both sides of the debate. This Part’s catalogue of different state-court approaches sets the stage for Part III, which assesses these approaches in light of the broader considerations at play.

Section II.A lays out the field by surveying the state and local statutory landscape, noting some common differences between subnational employment discrimination statutes and their national analogues. Section II.B analyzes the case law that has emerged from state courts, especially state courts of last resort, in the years since Gross and Nassar. It concludes that these state courts often give short shrift to the federalism concerns that lurk beneath their treatment of federal law.

A. The State Statutory Landscape

All fifty states have some form of employment discrimination statute. 74 See Hunt, supra note 17, at 22–80. While many of these statutes were modeled partly after Title VII, 75 See Long, supra note 23, at 524–25 & n.301 (noting the primary models from which state legislatures likely drew in enacting their own employment discrimination statutes, and surveying various state-court decisions that conclude that these state statutes were modeled after Title VII). none of them are completely identical to federal models in all significant details. 76 Sandra F. Sperino, Revitalizing State Employment Discrimination Law, 20 Geo. Mason L. Rev. 545, 557–58 (2013) [hereinafter Sperino, Revitalizing]. Unlike the federal model, most states have omnibus statutes that include all of the protected statuses under one statutory framework. 77 Id. at 560. Some state and local discrimination statutes contain directives to the courts to construe the statutes liberally, 78 See, e.g., Del. Code tit. 6, § 4501 (2019); Iowa Code § 216.18(1) (2019); Wash. Rev. Code § 49.60.020 (2018); W. Va. Code Ann. § 5-11-15 (LexisNexis 2019); Wis. Stat. § 111.31(3) (2019); N.Y.C., N.Y., Admin. Code § 8-130(a) (2019). language that is absent from the federal statutes. 79 Haskenhoff v. Homeland Energy Sols., LLC, 897 N.W.2d 553, 608 (Iowa 2017) (Appel, J., concurring in part and dissenting in part).

Although it is frequently said that most state employment discrim­ination statutes were modeled after federal statutes such as Title VII, 80 See, e.g., Long, supra note 23, at 524–25 (claiming that state employment discrimination statutes were “by and large” inspired by Title VII). around half of the states had fair employment statutes at the time of Title VII’s enactment. 81 Andrea Catania, State Employment Discrimination Remedies and Pendent Jurisdiction Under Title VII: Access to Federal Courts, 32 Am. U. L. Rev. 777, 783 n.24 (1983). These statutes anticipated many of the features of Title VII, complicating the narrative that the state statutes were “modeled” on federal ones. 82 See Haskenhoff, 897 N.W.2d at 608 (Appel, J., concurring in part and dissenting in part) (arguing that these preexisting state statutes may also have served as models for Title VII and later state statutes). Additionally, in 1966 the Uniform Law Commissioners drafted a Model Anti-discrimination Act. 83 Model Anti-discrimination Act (Nat’l Conference of Comm’rs on Unif. State Laws 1966). Although the Model Anti-discrimination Act uses the same “because of” language as Title VII to describe the practices that it prohibits, 84 See Carl A. Auerbach, The 1967 Amendments to the Minnesota State Act Against Discrimination and the Uniform Law Commissioners’ Model Anti-discrimination Act: A Comparative Analysis and Evaluation, 52 Minn. L. Rev. 231, 236 (1967) (“Section 302(a)(1) of the Model Act makes it unlawful for an employer to ‘fail’ as well as to ‘refuse’ to hire ‘an individual’ because of race, color, religion, sex, or national origin.”); see also 42 U.S.C. § 2000e-2(a) (2012) (“It shall be an unlawful employment practice for an employer . . . to discriminate against any individual . . . because of such individual’s race, color, religion, sex, or national origin . . . .”). at least one state supreme court has noted that portions of the state employment discrimination statute were modeled after the Model Act—rather than Title VII—in declining to follow federal Title VII doctrine. 85 See Carlson v. Indep. Sch. Dist. No. 623, 392 N.W.2d 216, 220 (Minn. 1986). Another state supreme court has noted that its state legislature had before it both Title VII and the Model Act when it enacted its sex discrimination statute, but the parties in the case agreed that no legislative history existed regarding the origin of the state provision. See Colo. Civil Rights Comm’n v. Travelers Ins. Co., 759 P.2d 1358, 1361 n.6 (Colo. 1988) (en banc).

In fact, these state and local statutes bear a variety of relationships to Title VII and federal employment discrimination policy. At one extreme of the spectrum, some state statutes explicitly endorse or indicate their subordination to the federal policy embodied in statutes like Title VII. For example, the Texas Commission on Human Rights Act notes that its purpose is to “provide for the execution of the policies of Title VII of the Civil Rights Act of 1964 and its subsequent amendments,” as well as those “embodied in Title I of the Americans with Disabilities Act of 1990 and its subsequent amendments.” 86 Tex. Labor Code § 21.001(1), (3) (2017). This kind of language seems to militate against independent construction, linking the state statute to federal policy. 87 The reality can be somewhat more complicated, however, when the state statute has itself been amended in ways not entirely concordant with the development of federal law. See infra notes 130–132 and accompanying text.

A unique case on the other extreme is the New York City Human Rights Law (NYCHRL), which reads:

a. The provisions of this title shall be construed liberally for the accomplishment of the uniquely broad and remedial purposes thereof, regardless of whether federal or New York state civil and human rights laws, including those laws with provisions worded comparably to provisions of this title, have been so construed.

b. Exceptions to and exemptions from the provisions of this title shall be construed narrowly in order to maximize deterrence of discriminatory conduct.

c. Cases that have correctly understood and analyzed the liberal construction requirement of subdivision a of this section and that have developed legal doctrines accordingly that reflect the broad and remedial purposes of this title include Albunio v. City of New York, 16 N.Y.3d 472 (2011), Bennett v. Health Management Systems, Inc., 92 A.D.3d 29 (1st Dep’t 2011), and the majority opinion in Williams v. New York City Housing Authority, 61 A.D.3d 62 (1st Dep’t 2009). 88 N.Y.C., N.Y., Admin. Code § 8-130 (2019).

The NYCHRL—and the case law that interprets it, now incorporated into the statute by amendment—presents perhaps the strongest legislative endorsement of independent construction of a subnational statute. One of the cases cited approvingly in the statute, Bennett v. Health Management Systems, Inc., held that because of the broad-construction provision of the NYCHRL, a defendant seeking summary judgment had to establish that no jury could find the defendant liable under either the traditional McDonnell Douglas analysis or a mixed-motive analysis encompassing the motivating-factor causation requirement. 89 936 N.Y.S.2d 112, 121 (App. Div. 2011). This holding applied even though the wording of the status-based discrimination provision of the NYCHRL is substantially similar to that found in Title VII. 90 Compare 42 U.S.C. § 2000e-2(a) (2012), with N.Y.C., N.Y., Admin. Code § 8-107(1)(a). The NYCHRL is an extreme example of why the text of state or local statutes might compel state courts to interpret state discrimination statutes in a manner that diverges from the federal case law, even though the statutory language that suggests causation—“because of”—is substantially the same.

In spite of the common differences between federal, state, and local statutes, state courts often look to federal employment discrimination cases for guidance on questions of statutory interpretation, including which causation standard the statute implies. One possible reason for this is that federal courts are often tasked with interpreting state discrimination statutes because plaintiffs bring claims under both federal and state laws in federal court. 91 See Sperino, Revitalizing, supra note 76, at 587 (“It is often the federal courts that are interpreting state law.”). Federal courts often use similar analytical frameworks for analyzing state and federal claims brought together. 92 See id. at 582 & n.255 (arguing that the common practice of using a single analysis for federal and state claims appended makes less sense given the current doctrinal fracture).

Moreover,  state  courts  confronted  with  the  issue  of  how  to  react  to  Gross  and  Nassar  often  face  similar  practical  constraints  on  their  ability  to  develop  interpretations  that  depart  from federal  case law. As longtime chief judge of the New York Court of Appeals Judith Kaye has noted, a “crucial distinction between state courts and federal courts interpreting statutes is the quantity of the legislative history that is available.” 93 Judith S. Kaye, State Courts at the Dawn of a New Century: Common Law Courts Reading Statutes and Constitutions, 70 N.Y.U. L. Rev. 1, 29 (1995). Chief Judge Kaye noted that, more often than federal courts, state courts have to infer legislative intent from the language of the statute itself. 94 Id. at 30. Understanding this range of policy concerns and practical constraints that influence state-court decisions is a crucial prerequisite to evaluating state-court responses to Gross and Nassar.

B. State Courts Decide the Causation Question

In light of the opposing pulls of state autonomy and deference to federal law, it is perhaps not surprising that many state courts have deferred deciding whether Gross and Nassar should apply to claims under state statutes. 95 See, e.g., Dussault v. RRE Coach Lantern Holdings, LLC, 86 A.3d 52, 61 n.5 (Me. 2014); Haddad v. Wal-Mart Stores, Inc., 914 N.E.2d 59, 77 n.27 (Mass. 2009); O’Brien v. Telcordia Techs., Inc., 20 A.3d 1154, 1163 (N.J. Super. Ct. App. Div. 2011) (“[W]e defer a decision on the thorny issue of the continued viability of the use of a Price Waterhouse mixed-motive analysis in light of the United States Supreme Court’s decision in Gross in an age discrimination case instituted pursuant to the [New Jersey Law Against Discrimination].”). With increasing frequency, however, state courts—including some state supreme courts—are deciding whether the but-for causation standard laid out in Gross and Nassar applies to the interpretation of various state discrimination statutes. This section surveys that landscape, which is only now beginning to take shape. Section II.B.1 discusses the states that have applied but-for causation or another requirement higher than the motivating-factor standard. Section II.B.2 discusses the states that have rejected the application of Gross and Nassar to state statutes and held that motivating-factor standards apply.

1. State Courts Applying a But-For or Other Heightened Causation Requirement. — In the wake of Gross and Nassar, state courts applying a heightened causation requirement have generally done so in one of three ways: (1) deciding the issue on independent state law precedent and noting the congruence with recent Supreme Court cases as dicta; (2) establishing congruence with federal law as a precedential principle that outweighs state law precedent on the correct standard; or (3) automatically applying the federal standards without justification.

a. Deciding the Issue on Independent State Law Precedent. — State courts might, of course, independently decide that the state law dictates the same result as the federal case law. Perhaps the clearest example of this is Asbury University v. Powell, in which the Supreme Court of Kentucky held that a but-for causation standard applied to a retaliation claim under the Kentucky Civil Rights Act (KCRA). 96 486 S.W.3d 246, 256 (Ky. 2016). The court explained that it was merely following state case law, which had always followed the but-for standard that the federal courts had only recently adopted in Nassar. 97 Id. By “happenstance,” the Nassar Court “merely restated the law as it has always been interpreted in Kentucky.” 98 Id.

One might wonder, however, to what extent a state court applying but-for causation on state law grounds is influenced by the gravitational pull of the recent Supreme Court decisions. 99 Cf. Scott Dodson, The Gravitational Force of Federal Law, 164 U. Pa. L. Rev. 703, 706 (2016) (“[S]omething more than independent parallel conduct is afoot: federal law exerts a widespread gravitational pull on state actors.”). These cases provide an aura of legitimacy to state-court decisions and may help tip the balance in favor of a heightened causation requirement when the weight of state law authority is not necessarily clear.

Nassar may have played this role in the majority opinion in Haskenhoff. 100 See supra notes 1–12 and accompanying text. Noting that “the retaliation provision of the ICRA mirrors almost exactly the retaliation provision of Title VII” 101 Haskenhoff v. Homeland Energy Sols., LLC, 897 N.W.2d 553, 584 (Iowa 2017) (emphasis omitted). and that the Iowa Supreme Court has “consistently employed federal analysis when interpreting the ICRA,” 102 Id. (internal quotation marks omitted) (quoting Estate of Harris v. Papa John’s Pizza, 679 N.W.2d 673, 678 (Iowa 2004)). the  court  supported  its  rationale  for  a  higher  causation  standard  with  reference  to  the  Supreme  Court’s  decision  in  Nassar. 103 Id.; see also Univ. of Tex. Sw. Med. Ctr. v. Nassar, 133 S. Ct. 2517, 2534 (2013). Although the Iowa majority rejected allegations of blind adherence to federal law, it invoked the “long-standing practice of looking to federal decisions to interpret the same or equivalent statutory language.” 104 Haskenhoff, 897 N.W.2d at 585. Nevertheless, the court explicitly relied on state law precedent to hold that the proper standard is “significant factor” and declined to apply the but-for standard of Nassar itself. 105 Id. at 585–86; see also supra notes 5–6 and accompanying text.

A  vigorous  dissent  disagreed,  arguing  that  “the  Nassar  case  has  no  bearing  in  the  interpretation  of  the  ICRA”  because  Nassar  relied  upon  the  legislative  history  of  Title  VII,  which  was  fundamentally different  from  the  legislative  history  behind  the  ICRA. 106 Haskenhoff, 897 N.W.2d at 635–37 (Appel, J., concurring in part and dissenting in part). Noting the similar causation language in the status-based discrimination and retaliation provisions of the ICRA, the dissent saw no reason to interpret the two provisions as requiring different standards, especially given the ICRA’s mandate to “broadly construe [its] provisions,” a directive absent from Title VII. 107 Id. at 634 –37. The dissent stressed that “the ICRA is not simply a knockoff of the Federal Civil Rights Act” 108 Id. at 606. and critiqued the majority’s “mirror” theory by noting that both the ICRA and Title VII drew upon state statutes that preceded Title VII. 109 Id. The dissent urged that interpreting the ICRA required “serious, provision-by-provision analysis, recognizing similarities [with Title VII] when they appear, but also honoring the differences.” 110 Id. at 607.

Thus, although by the explicit terms of its decision the Haskenhoff majority was following state law precedent, 111 See id. at 585 (majority opinion). it gave the Nassar decision considerable persuasive weight, citing Nassar at length for its policy concern about the proliferation of unfounded claims. 112 See id. at 584. Additionally, the majority, contrary to its assertion that it was deciding based on state precedent, noted that the enactment of the ICRA after Title VII rendered “look[ing] to federal decisions for guidance” appropriate. 113 Id. at 585–86. Thus, even when the state courts purport to apply heightened causation requirements based on state precedent, the Gross and Nassar decisions may gravitationally pull the court’s analysis in the direction of a certain result.

b. Establishing Congruence with Federal Law as a Binding Precedential Principle. — Perhaps the most common reason that state courts follow Gross and Nassar is that the case law has established some principle or practice of following analogous federal decisions. Recently, the Supreme Court of Florida authorized a new set of standard jury instructions in civil cases. 114 See In re: Standard Jury Instructions in Civil Cases—Report No. 16-01 (Standard Jury Instructions), 214 So. 3d 552, 555–57 (Fla. 2017). Although the main text of the jury instruction simply uses the “because of” language from the statute, 115 Fla. Stat. § 760.10(1) (2018). the instruction also reads: “(If necessary, clarify the causation standard further.).” 116 Standard Jury Instructions, 214 So. 3d at 555. In the “Notes on Use” for the discrimination jury instruction, the court noted that decisions construing the ADEA apply to the Florida Civil Rights Act’s age discrimination provision. 117 Id. at 556. The practice of looking to the ADEA was so entrenched in Florida law that the court felt the need to cite only a state appellate court to justify relying on Gross. 118 See id.

Similarly, the Court of Special Appeals of Maryland has applied a presumption of federal–state congruity. In White v. Parker, the court opined, “We see no reason why we would not adopt the but-for causation standard established in Nassar for [Maryland Fair Employment Practices Act] retaliation claims.” 119 No. 2171, 2017 WL 727794, at *11 (Md. Ct. Spec. App. Feb. 24, 2017). The court wrote that absent specific language or policy considerations that support an independent construction, it would presumptively interpret the Maryland Fair Employment Practices Act according to federal standards. 120 Id. Similar cases include Wholf v. Tremco Inc., 26 N.E.3d 902, 908 (Ohio Ct. App. 2015) (adopting the reasoning of Nassar because the Ohio statute was “modeled after Title VII”), and Goree v. United Parcel Serv., Inc., 490 S.W.3d 413, 439 (Tenn. Ct. App. 2015) (“We believe that the Tennessee Supreme Court would adopt the reasoning of Nassar and require but for causation for a retaliation claim pursuant to the [Tennessee Human Rights Act] . . . .”). This presumption is quite similar to the approach advocated by at least one proponent of parallel construction. 121 See infra notes 164–166 and accompanying text.

c. Automatic Application of the Federal Standards. — In some instances, state courts, even state high courts, cite Supreme Court precedent as if it is binding authority on the interpretation of state statutes, without any further inquiry into why the court should follow the Supreme Court’s reasoning. Occasionally, these cases note that their interpretations of the state statutes are guided by federal law, but they go on to cite federal decisions as if binding. Such cases may arise when the principle of following federal law is so entrenched that courts feel free to cite federal case law without acknowledging the distinctness of the state and federal regimes.

For example, in Robinson v. Board of Supervisors for the University of Louisiana System, the Louisiana Supreme Court cited a Louisiana state appellate court for the proposition that Gross applies to age discrimination claims under Louisiana’s Age Discrimination in Employment Act. 122 225 So. 3d 424, 431 (La. 2017) (citing Eastin v. Entergy Corp., 42 So. 3d 1163, 1182 (La. Ct. App. 2010)). Although the court noted that “Louisiana courts have traditionally looked to federal case law for guidance” in interpreting the LADEA, 123 Id. (internal quotation marks omitted) (quoting LaBove v. Raftery, 802 So. 2d 566, 573 (La. 2001)). it conducted minimal independent analysis in applying Gross directly to the state statute. 124 See id.

In 2013, the Idaho Supreme Court automatically applied the but-for requirement of Gross to an age discrimination claim under the Idaho Human Rights Act. 125 See Hatheway v. Bd. of Regents of the Univ. of Idaho, 310 P.3d 315, 323 (Idaho 2013). The court noted that “[f]ederal law guides this Court’s interpretation of the [Idaho Human Rights Act].” Id. at 322. The Idaho Supreme Court has previously treated the “guid[ance]” of federal law as binding. See, e.g., Ostrander v. Farm Bureau Mut. Ins. Co. of Idaho, 851 P.2d 946, 949 (Idaho 1993) (finding that the limitation of Title VII’s protections to employees applied to the Idaho Human Rights Act because of the “guid[ance]” of federal law). Similarly,  in  Villiger  v.  Caterpillar,  Inc.,  an  Illinois  appellate  court  merely  asserted  that  but-for  causation  was  the  requirement  in  age  discrimination  cases,  with  a citation  to  Gross. 126 See No. 3-12-0739, 2013 WL 2298474, at *3 (Ill. App. Ct. May 23, 2013). For another state decision that falls into this category, see Termonia v. Brandywine Sch. Dist., No. N10C-12-174 ALR, 2014 WL 1760317, at *4–5 (Del. Super. Ct. Apr. 16, 2014), aff’d, 108 A.3d 1226 (Del. 2015).

Given that state courts are the ultimate authorities in the construction of state statutes, 127 See supra note 18. these cases may constitute an abdication of the state courts’ role. In the case of state high courts, such as the Idaho Supreme Court, the role of interpreting the state statute is even more important because federal courts may certify questions of state law to the state’s highest court. 128 See generally Jonathan Remy Nash, Examining the Power of Federal Courts to Certify Questions of State Law, 88 Cornell L. Rev. 1672, 1674 (2003) (“Most state high courts now offer federal courts faced with questions of state law, as well as similarly situated state courts, the opportunity to ‘certify’ those questions to the state high court.”). Following federal standards lockstep would effectively collapse the analogous federal and state provisions into one.

2. State Courts Applying a Motivating-Factor or Substantial-Motivating-Factor Standard. — Other state courts have more vigorously asserted their prerogative to construe state statutes independently from case law interpreting federal analogues. Although these decisions are rarely as comprehensive as the Haskenhoff dissent, they often turn on the same factors identified by that opinion as weighing in favor of independent construction of a statutory provision. 129 See supra notes 106–110 and accompanying text. Because these decisions typically identify particular features of the state statutes at hand that help justify independent construction, they are less susceptible to categorization than cases which establish lockstep-like principles. Sometimes relying on the textual differences between an omnibus statute and the fragmented federal statutory scheme, sometimes relying on a provision instructing state courts to construe the state statute liberally, and sometimes simply disagreeing with the reasoning of Gross and Nassar, state courts have applied the motivating-factor standard with varying degrees of sensitivity to the efficiency and legitimacy considerations that should enter into a court’s decisionmaking calculus.

In the most straightforward cases, state legislatures have enacted statutory language either explicitly or implicitly endorsing the motivating-factor standard—much in the same way that Congress did in 1991. In 2011, for instance, a Texas appellate court noted that the law was unsettled as to whether Gross applied to age discrimination claims under the Texas Commission on Human Rights Act (TCHRA). 130 See Hernandez v. Grey Wolf Drilling, L.P., 350 S.W.3d 281, 285 (Tex. Ct. App. 2011). Because the relevant provision of the TCHRA, like the 1991 Title VII amendments, contained the motivating-factor language, the court held that Gross did not apply. 131 See id. This was in spite of the court’s acknowledgement that, “[b]ecause the TCHRA’s stated purpose is to ‘provide for the execution of the policies of Title VII of the Civil Rights Act of 1964 and its subsequent amendments,’ Texas courts apply analogous federal case law when interpreting the [TCHRA].” 132 Id. at 283 (quoting Tex. Lab. Code § 21.001(1) (2006)). Interestingly, although the TCHRA provides for the execution of the policies embodied in both Title VII and the Americans with Disabilities Act of 1990 (ADA), it makes no analogous mention of the ADEA. The court in Hernandez did not seem to think this discrepancy worth noting. See id. A less explicit but no less certain endorsement of the motivating-factor standard appears in the NYCHRL, which gives legislative imprimatur to a decision holding that the motivating-factor standard applies under the NYCHRL. 133 See supra notes 88–90 and accompanying text.

Less straightforward are cases in which state courts focus on differences between state statutes and their federal analogues to justify departing from federal precedent. In the unreported decision of West Virginia American Water Co. v. Nagy, the Supreme Court of West Virginia upheld a mixed-motive jury instruction in a state law age discrimination case, emphasizing that the West Virginia Human Rights Act encompasses all forms of discrimination and disparate treatment within a single omnibus statutory scheme. 134 See No. 101229, 2011 WL 8583425, at *20 (W. Va. June 15, 2011). Because this was one of the main considerations that the court discussed in rejecting the Gross rule, it is likely that the critics of the new judicial federalism would find the “bootstrap reasoning” in West Virginia American Water Co. to be inadequate, given the prudential reasons for parallel construction. 135 See infra notes 161–162 and accompanying text (describing the possibility that state courts will rely on trivial differences between state and federal statutes to disguise what amounts to mere differences of opinion with the Supreme Court). The inclusion of age discrimination in the same statutory scheme as the other status-based discrimination claims may, to some critics, be a flimsy pretext for what is essentially a policy-based disagreement with the Gross opinion.

Courts may be somewhat less susceptible to these criticisms in states whose employment discrimination statutes contain broad-construction provisions. Faced with a race and disability discrimination claim under the Alaska Human Rights Act, the Alaska Supreme Court in 2010 noted that “while we look to federal discrimination law jurisprudence generally,” Alaska’s employment discrimination law was intended to be construed more broadly than its federal analogue. 136 Smith v. Anchorage Sch. Dist., 240 P.3d 834, 842 (Alaska 2010). Further noting that, unlike in the federal scheme, the Alaska statute included age discrimination and other discrimination claims in the same sentence, the court declined to apply Gross to the claims at issue. 137 Id. A Connecticut appellate court reached a similar result based not on a broad-construction provision but on a line of state case law. See Bissonnette v. Highland Park Mkt., Inc., No. HHDCV106014088S, 2014 WL 815872, at *4 (Conn. Super. Ct. Jan. 28, 2014) (“[O]ur fair employment practices statutes were enacted to eliminate discrimination in employment. They are remedial and receive a liberal construction.” (alteration in original) (internal quotation marks omitted) (quoting Vollemans v. Town of Wallingford, 928 A.2d 586, 605 (Conn. App. Ct. 2007))).

Another case in which a state court distinguished the state statute at hand from federal analogues based on differing provisions, including a liberal-construction provision, came down in California in 2013. 138 See Harris v. City of Santa Monica, 294 P.3d 49, 55 (Cal. 2013). The  California  Supreme  Court  rejected  the  argument  that  Gross  controlled  the  interpretation  of  “because  of”  in  the  sex  discrimination  provision  of  California’s  Fair  Employment  and  Housing  Act (FEHA). 139 See id. (“Our precedent has recognized, however, that ‘but for’ causation is not the only possible meaning of the phrase ‘because of’ in the context of an antidiscrimination statute.”); see also Cal. Gov’t Code § 12940(a) (2019). Drawing on the FEHA’s stated purposes, as well as the provision that the statute should be “construed liberally for the accomplishment of [its] purposes,” the court applied on state law grounds a standard similar to the one that Congress enacted for status-based discrimination in 1991. 140 See Harris, 294 P.3d at 60. Specifically, if the plaintiff proves by a preponderance of the evidence that discrimination was a “substantial factor motivating” an adverse employment decision, then the employer will be held liable under the FEHA; nevertheless, if the employer proves by a preponderance of the evidence that it would have made the same decision absent the impermissible motive, then the plaintiff may not receive damages, back pay, or an order of reinstatement but is limited to declaratory and injunctive relief, as well as reasonable attorneys’ fees and costs. 141 Id. at 72. Crucially, the court applied this interpretation to all of section 12940(a) of the FEHA, which encompasses not just the status-based discrimination covered by Title VII but also discrimination on the basis of age, sexual orientation, physical disability, mental disability, gender identity, and a host of other protected characteristics. 142 Id.; see also Cal. Gov’t Code § 12940(a).

In construing the FEHA, the California Supreme Court noted that its goal was to “give effect to the Legislature’s intent” and that it had to “begin with the statutory text” to discern intent. 143 Harris, 294 P.3d at 54. The  plain  meaning  of  the  statute,  however,  was  unavailing  because,  “[l]inguistically,  the  phrase  ‘because  of’  is  susceptible  to  many  possible  meanings.” 144 Id. Furthermore,  the  FEHA’s  legislative  history  did  not  justify  the  same  kind  of  negative  inference  that  the  Supreme  Court  had  made  in  Gross. Thus, the court turned its attention to the enacting legislature’s statements about the FEHA’s aims. 145 Id. at 59. The court relied heavily on the FEHA’s stated purpose of preventing and deterring unlawful discrimination in rejecting but-for causation. 146 See id. at 64–65 (“When discrimination has been shown to be a substantial factor motivating an employment action, a declaration of its illegality serves to prevent that discriminatory practice from becoming a ‘but for’ cause of some other employment action going forward.”). Nevertheless, cautious about the imposition of liability for “mere thoughts or passing statements unrelated to the disputed employment decision,” the court declined to apply the motivating-factor standard, instead applying a higher “substantial motivating factor” standard. 147 Id. at 66. The court suggested that, in rejecting the mere “motivating factor” standard in favor of the “substantial motivating factor” test, it was in accord with Justice O’Connor’s concurrence in Price Waterhouse. See id.; see also supra notes 46, 57 and accompanying text. Because this holding applied to all the protected character­istics listed in section 12940(a), the result is that California, as compared to federal law, imposes a higher causation standard for the status-based discrimination claims covered by the 1991 Title VII amendments but a lower standard for age discrimination. 148 See Harris, 294 P.3d at 72.

From the foregoing survey, it is clear that state courts are not engaged in any sort of unified discourse on the relative merits of parallel and independent construction. Whether it is because they feel the inexorable gravitational pull of federal law, they are bound by stare decisis to decide the question in a certain way, or they simply do not think the issue worth discussing, state courts tend to give little attention to federalism considerations when deciding to follow or depart from federal law. But systematic application of these principles to the specific problems of interpretation that these courts face may help clarify how courts should weigh the competing concerns of autonomy and deference in this context.

III. Why Federalism Considerations Favor Independent Construction

This Part assesses the range of state-court approaches discussed in Part II in light of scholarly debates about the merits and drawbacks of parallel construction. It argues that while the benefits of parallel construction are weak when it comes to causation standards in employment discrimination, the countervailing benefits of independent construction are abundantly present. These include fidelity to state policy preferences and interjurisdictional dialogue. Section III.A provides background on the discourse of “new judicial federalism” and argues that these debates provide a useful frame for the systematic analysis of the problem at hand. Although the new judicial federalism is most commonly associated with the interpretation of state constitutional provisions independent of their federal constitutional analogues, this section builds on existing scholarly work to argue that a similar set of concerns underlies the employment discrim­ination context. Section III.B argues that the standard legitimacy and efficiency concerns counseling in favor of parallel construction are relatively absent in the context of causation in employment discrimination statutes. It also explores differences between statutory and constitutional law, positing that, although there may be less room for productive vertical (between state and federal) dialogue on this particular question of statutory interpretation than on questions of fundamental constitutional rights, there is still the possibility of horizontal dialogue. Ultimately, this section argues that the benefits of state independence here outweigh the costs. Concerns about efficiency, legislative preference, and legitimacy should not discourage state courts from fulfilling their independent function within our system of dual sovereignty.

A. The “New” Judicial Federalism

State supreme courts are the final, authoritative voices in interpreting state statutes and constitutions. 149 E.g., Friedman, supra note 23, at 100. As one scholar has put it, “[I]n a situation where a state supreme court interprets a state constitutional provision—even one textually indistinguishable from the federal provision—the [U.S.] Supreme Court, far from being final, has nothing at all to say.” 150 Charles Fried, Reflections on Crime and Punishment, 30 Suffolk U. L. Rev. 681, 710–11 (1997). Indeed, Justice Brennan famously urged state courts to be open to the possibility of interpreting state constitutions to provide stronger protections for individual rights than the U.S. Constitution. 151 See William J. Brennan, Jr., State Constitutions and the Protection of Individual Rights, 90 Harv. L. Rev. 489, 491 (1977) (“[S]tate courts cannot rest when they have afforded their citizens the full protections of the federal Constitution.”). Commentators sometimes use the term “new judicial federalism” to refer to the practice in which state high courts rule on questions of individual rights by relying on state constitutional provisions as adequate and independent grounds of decision. 152 See, e.g., John Kincaid, Foreword: The New Federalism Context of the New Judicial Federalism, 26 Rutgers L.J. 913, 913 n.1 (1995). Some scholars also use the phrase to refer to a similar dynamic in the statutory context. 153 See Long, supra note 23, at 483.

Not all commentators celebrate this independent interpretation of state law, however. Some insist that, when it comes to state constitutional provisions, state courts should defer to the Supreme Court’s interpretation of analogous federal constitutional provisions. 154 See G. Alan Tarr, Understanding State Constitutions 175 (1998). Critics of the new judicial federalism often invoke concerns about efficiency, 155 See Long, supra note 23, at 531–36 (advancing efficiency arguments for parallel construction). legislative preference, 156 See id. at 524 – 31 (arguing that state legislatures generally prefer parallel construction). and legitimacy 157 See Friedman, supra note 23, at 96 (summarizing the legitimacy-based criticisms of the new judicial federalism); Long, supra note 23, at 538–39 (arguing that parallel construction bolsters the credibility of state courts). to advocate for parallel construc­tion. Parallel construction bolsters efficiency, the argument goes, because courts deciding cases arising under parallel federal and state constitutional and statutory provisions need only learn and apply one set of rules. As to legislative preference, the hypothesis is that legislatures generally prefer parallel construction with federal statutes. 158 See Long, supra note 23, at 538–39. If true, this hypothesis also raises a legislative- efficiency rationale, as erring on the side of parallel construction conserves the resources needed for the legislative monitoring and abrogation of undesirable judicial rules. 159 See id. at 531–36.

Independent construction raises a potential legitimacy concern if it is likely that divergent state-court opinions will be seen as capricious or less justifiable than opinions that follow federal case law in a “lockstep” fashion, a concern amplified in states where judiciaries are elected and therefore subject to more immediate political pressures than their federal counterparts. 160 See id. at 538–39. The “new” judicial federalism has been frequently criticized for its “result-oriented cast.” For instance, Justice Brennan has been accused of advocating for state constitutionalism precisely to make up for the erosion of the Warren Court’s individual rights jurisprudence under Chief Justice Burger. 161 See Friedman, supra note 23, at 93–94 & n.3 (summarizing these criticisms). Critics inveigh against the “bootstrap reasoning” whereby state courts, disagreeing with the U.S. Supreme Court’s interpretation of a federal provision, focus on flimsy or trivial differences between the federal and state provisions to justify reaching a different result under state law. 162 See, e.g., James A. Gardner, State Constitutional Rights as Resistance to National Power: Toward a Functional Theory of State Constitutions, 91 Geo. L.J. 1003, 1063 (2003). Such departures from federal law may “call into question unnecessarily the wisdom or moral authority of the Supreme Court.” 163 Long, supra note 23, at 507.

In light of these perceived shortcomings of the new judicial federalism, Professor Alex B. Long proposed—years before the Gross and Nassar decisions—that state courts adopt a substantive canon of construction favoring parallel construction of similarly worded state and federal statutes unless courts can “identify some meaningful difference between state and federal law, some fundamental change in approach at the federal level, or some outright error on the part of the federal courts.” 164 Id. at 556. This approach, Long argues, rejects “[b]lind adherence to federal precedent” 165 Id. while also avoiding the harms to legislative efficiency, legislative preference, and judicial credibility that accrue from irresponsible independent construction. 166 See id. at 556–57.

Of  course,  under  this  canon,  it  is  still  incumbent  on  state  courts  to  determine  whether  a  federal  case  represents  a  “fundamental  change  in  approach”  or  an  “outright  error.” 167 Id. at 556. To the extent that one believes, as Long does, that departing from Supreme Court doctrine may “call into question” the Court’s “wisdom or moral authority,” see supra note 163 and accompanying text, it is unclear why accusing the Court of “outright error” would be any less troublesome. In any event, even the critics do not completely absolve state courts of the responsibility of evaluating federal decisions and determining whether to follow them; however, they envision for state supreme courts a much more modest and deferential role in the federal system than do the proponents of the new judicial federalism.

Between the two extremes of this debate—a focus on state sovereignty on the one hand and an emphasis on federal supremacy on the other—is a more recent body of writing on federalism that justifies the independence of state courts, not because states are sovereigns that are entitled to operate in blithe indifference to the national conversation, but because a proper allocation of power to state governments produces benefits for the federal system as a whole. 168 See, e.g., Robert A. Schapiro, Polyphonic Federalism: Toward the Protection of Fundamental Rights 7 (2009) (theorizing a “polyphonic federalism” divorced from traditional notions of state sovereignty and focused on “the organizational benefits of multiple agents of power”); Friedman, supra note 23, at 97 (arguing that “dialogue” between state courts and the Supreme Court buoys the legitimacy of the Supreme Court); James A. Gardner & Jim Rossi, Dual Enforcement of Constitutional Norms, in New Frontiers of State Constitutional Law: Dual Enforcement of Norms 1, 10 (James A. Gardner & Jim Rossi eds., 2010) [hereinafter New Frontiers of State Constitutional Law] (collecting essays exploring a dialogic model of the development of constitutional norms because of communication and interaction between state and federal actors). This body of thought is distinct from—but shares some broad themes with—what Professor Heather K. Gerken has termed the “nationalist school of federalism.” See generally Heather K. Gerken, Federalism as the New Nationalism: An Overview, 123 Yale L.J. 1889, 1890 (2014) (describing the “nationalist school of federalism,” which “offers a descriptive and normative account that is deeply nationalist in character”). The nationalist school offers a vision of and justification for federalism that is “detached from the notion that state autonomy matters above all else [and is] attentive to the rise of national power and the importance of national politics.” Id. at 1889–90.

One of the most frequently commented-upon benefits is that of dialogue between federal and state actors about fundamental questions of governance. 169 See, e.g., Schapiro, supra note 168, at 133–38 (balancing judicial federalism’s benefits of plurality, dialogue, and redundancy against the values of uniformity, finality, and hierarchical accountability); Friedman, supra note 23, at 97–98 (“[State courts’ engagement in dialogue about constitutional values] can inform interpretive debates among judges, scholars, and citizens about the meaning of constitutional text, and thereby balance the interpretational judgment of the Supreme Court.”); Lawrence G. Sager, Cool Federalism and the Life Cycle of Moral Progress, in New Frontiers of State Constitutional Law, supra note 168, at 15, 16–17 (speculating that moral progress in the United States frequently occurs as the result of the “interplay between state and federal sensibilities and consensus”). As Justice Brandeis famously wrote, “a single courageous State may . . . serve as a laboratory; and try novel social and economic experiments without risk to the rest of the country.” 170 New State Ice Co. v. Liebmann, 285 U.S. 262, 311 (1932) (Brandeis, J., dissenting). In at least some visions of federalism, this dialogue does not lead to unending plurality and uncertainty but rather helps to forge a national consensus. 171 See Sager, supra note 169, at 18–19 (arguing that the moral progress that comes from judicial federalism may yield consensus because states on the “moral frontier” articulate the heretofore unimaginable, which other states may then follow). But the dialogue need not lead to a grand telos in order to be beneficial: The accumulation of knowledge may lead to better legal rules even without the prospect of ultimate consensus. Lockstep approaches to parallel statutes may impoverish this process of knowledge accumulation, leading to stagnation. 172 See Dodson, supra note 99, at 746.

Moreover, state-court independence likely results in greater fidelity to the policy preferences of state electorates. As Professor Scott Dodson has argued, “state courts eager to track nonpreemptive federal law may misinterpret state law, resulting in further distance between the preferences of state citizens and the laws that govern them.” 173 Id. at 747. Greater variation among states may well result in interpretations that are, in the aggregate, better anchored in the preferences of each state’s citizenry. 174 See id. Thus, independent interpretations may in fact improve perceptions of judicial legitimacy.

Just as importantly, as Professor Dodson argues, when states give into the gravitational force of federal law, they “lend[ ] credence to the position that states are just not as good at being sovereign as the federal government is.” 175 Id. at 748. Dodson’s article examines the “gravitational pull” of nonbinding federal law on not only the state courts but also state rulemakers and legislators. See id. at 706. This degrades the role of the states within our system of dual sovereignty. 176 See id. at 751. Moreover, states’ incorporation of federal law may be self-reinforcing: “If states routinely mimic federal law rather than innovate, then all eyes will train on the federal leader[,]” creating a feedback loop that only entrenches state actors’ habit of following. 177 See id. at 752.

B. Motivating-Factor Standards Do Not Raise the Typical Concerns About the New Judicial Federalism

As noted above, critics of the new judicial federalism sometimes invoke notions of legitimacy, legislative preference, and legislative efficiency to argue for parallel interpretation. 178 See supra notes 155–163 and accompanying text. Moreover, although it is possible that states and localities may arrive at different causation standards as a matter of policy, the preeminence of Title VII and the other federal discrimination statutes may suggest that employment discrimination is, to paraphrase Chief Justice Rehnquist, “truly national” as opposed to “truly local.” 179 United States v. Lopez, 514 U.S. 549, 568 (1995). Unless there is reason to think that a more lenient causation standard is an appropriate expression of the state’s particular custom or policy preference, which is unlikely given the scarcity of legislative history, 180 See supra notes 93–94 and accompanying text. it might seem sensible to adopt Professor Long’s presumption in favor of parallel construction. 181 See supra notes 164–166 and accompanying text. Adopting such a presumption would also avoid the complication of presenting two sets of jury instructions and asking juries to consider two different causation standards simultaneously, which some commentators have worried would overwhelm “even the most competent juror.” 182 Nico Piscopo, Note, A Pragmatic Approach to Age Discrimination Claims in Connecticut: Aligning with the Federal Standard, 16 Conn. Pub. Int. L.J. 281, 293–94 (2017).

However, these concerns fail to justify such a renunciation of the state courts’ obligation to interpret state statutes within a federal system characterized by overlapping statutory protections. As a threshold matter, Title VII’s text itself limits its preemptive effect to state law “which purports to require or permit the doing of any act which would be an unlawful employment practice” under federal law. 183 42 U.S.C. § 2000e-7 (2012). Therefore, as a matter of statutory scheme, states are empowered to put in place higher protections than the federal law. In other words, Title VII itself seems to envision setting a federal floor for employee rights rather than a ceiling. This alone, however, does not justify independent construction: While Title VII clearly indicates that state legislatures can impose higher requirements, as the TCHRA and NYCHRL have, 184 See supra notes 88–89, 131 and accompanying text. it does not necessarily suggest what limitations are appropriate for state courts operating within a national system. 185 See Gardner & Rossi, supra note 168, at 3 (“To constrain [the discretion of state courts to construe state constitutions independently], state courts require a stock of disciplining jurisprudential concepts. The modern history of American subnational constitutional interpretation is a story about the search for such concepts.”).

The typical concerns about efficiency, legislative preference, and judicial legitimacy, though they may justify going with the federal tide on some issues of statutory interpretation, have comparatively little purchase on the question of which causation standard the word “because” expresses. The legislative-preference and legislative-efficiency justifications for parallel construction rest on strong, and not necessarily defensible, presumptions about how enacting legislatures want their statutes interpreted: “The initial decision to borrow federal law . . . can reasonably be viewed as establishing a general preference in favor of future parallel construction of the state statutes, provided that those constructions are reasonable and generally consistent with the progression of federal law existing at the time of borrowing.” 186 Long, supra note 23, at 528. An equally strong presumption, however, might hold in the opposite direction: State legislatures would have little reason to enact separate discrimination statutes if they expected or wanted their citizens’ rights under the statutes to be exactly duplicative of their rights under federal analogues. 187 The “lockstep” approach has been frequently criticized on grounds similar to these. Cf. James A. Gardner, Why Federalism and Constitutional Positivism Don’t Mix, in New Frontiers of State Constitutional Law, supra note 168, at 39, 55 (“[Criticizing unreflective lockstep analysis] is warranted because, in a genuinely federal system, the state polity has some degree of independent agency in the formulation of its constitutional rules of self-governance and courts cannot assume . . . that the state polity has chosen entirely to forgo any exercise of that agency.”). The same could be said in the case of parallel state and federal statutes. There is just as good a reason to presume that the legislative preference would be for judicial rules that are foreseeable at the time of enactment. 188 Cf. Dodson, supra note 99, at 711 (“[W]hile state rulemakers may have intended that state rules be interpreted in light of then-existing federal precedent, it is far more tenuous to infer that the state rulemakers intended for post-adoption federal precedent to be indicative of the state rule’s meaning.”). Here, given the weight of commentary describing the causation standards applied under Gross and Nassar as “new” or surprising, one may conclude that the but-for requirement was not foreseeable at the time of enactment of many state discrimination statutes. 189 See, e.g., Lawrence D. Rosenthal, A Lack of “Motivation,” or Sound Legal Reasoning?: Why Most Courts Are Not Applying Either Price Waterhouse’s or the 1991 Civil Rights Act’s Motivating-Factor Analysis to Title VII Retaliation Claims in a Post-Gross World (but Should), 64 Ala. L. Rev. 1067, 1089 (2013) (noting that the motivating-factor question was not before the Court in Gross); Michael J. Zimmer, Title VII’s Last Hurrah: Can Discrimination Be Plausibly Pled?, 2014 U. Chi. Legal F. 19, 62 (“The second, surprising step [in the Court’s reasoning in Gross] was to reject the burden-shifting approach the Supreme Court had found to apply to Title VII cases in its 1989 decision in Price Waterhouse.”); Todd, supra note 70, at 290 (“[Many people] may be confused as to why the standards of causation [for discrimination and retaliation after Nassar] differ where retaliation has previously been considered a form of discrimination itself.”). The point here is not that one presumption regarding legislative preference is more defensible than the other; the point is that appeals to legislative preference do not necessarily support parallel construction.

And,  of  course,  explicit  evidence  of  the  enacting  legislatures’  preferences  for  whether  and  how  the  interpretation  of  “because”  would  track  federal  law  is  likely  nonexistent. 190 See supra notes 93–94 and accompanying text (discussing the relative lack of state legislative history). It is worth recalling that, faced with a similar absence of evidence as to Congress’s intent, the plurality in Price Waterhouse invoked “common sense” in finding that the congressional intent was to prohibit consideration of the protected characteristics or activities in employment decisions, necessitating a motivating-factor standard. 191 See supra notes 42–44 and accompanying text. The much scarcer legislative history in the state context makes this sort of reasoning even more critical. 192 See supra notes 93–94. This is not to say that the motivating-factor standard is the only commonsensical interpretation of “because” in discrimination statutes; rather, the commonsensical appeal of the Price Waterhouse plurality’s interpretation could prevent courts from having to follow highly speculative paths about how legislatures foresaw the development of the interpretation of their enacted language.

Indeed, apart from state-specific considerations, there may be other good reasons for state courts to consider not following Gross and Nassar. Recent empirical research has suggested that the but-for test is far more restrictive than layperson understandings of “because of.” 193 See James A. Macleod, Ordinary Causation: A Study in Experimental Statutory Interpretation, 94 Ind. L.J. (forthcoming 2019). A  state  court  might  also  consider  adopting  the  motivating-factor  standard  as  what  Professor  Einer  Elhauge  calls  a  “preference-eliciting  default  rule.” 194 See Einer Elhauge, Preference-Eliciting Statutory Default Rules, 102 Colum. L. Rev. 2162, 2165 (2002) [hereinafter Elhauge, Preference-Eliciting Rules] (arguing that some rules of statutory construction should and do operate as default rules whose purpose is to induce legislative specification). Elhauge’s general theory of statutory canons is that they are meant to maximize the satisfaction of “enactable preferences,” or the legal rules that would be enacted if they were on the legislative agenda. 195 See Einer Elhauge, Statutory Default Rules: How to Interpret Unclear Legislation 7–8 (2008) [hereinafter Elhauge, Statutory Default Rules]. One might think, then, that canons would tend to favor the powerful because legislative preferences more often track the preferences of the powerful. Elhauge’s insight, however, is that many canons that favor the politically powerless (for example, the rule of lenity, the canon favoring Indian tribes, and others) are preference-eliciting default rules. 196 Id. at 168–87; Elhauge, Preference-Eliciting Rules, supra note 194, at 2165–66. These canons are more likely to elicit legislative reactions that help us understand which preferences are enactable (because the canons would result in rules unfavorable to the politically powerful, who have the resources to influence the legislative agenda). 197 See Elhauge, Statutory Default Rules, supra note 195, at 168–87. This analysis may be especially attractive in the state context, in which legislative preferences are less clear (due in part to the paucity of legislative history), as well as in the context of discrimination causation standards, about which it seems unlikely that enacting legislatures have very coherent preferences. Therefore, a lower causation might be preferable for eliciting more frequent legislative reactions.

Additionally, the “common sense”–style reasoning of Price Waterhouse may be even more amenable to the institutional strengths of state courts as compared to federal courts. When courts apply causation standards to discrimination statutes, it is technically a question of statutory interpretation. But because of the indeterminacy of the statutory language to be interpreted, it involves courts acting more in their traditional common law capacity than many modern questions of statutory interpretation do. 198 Cf. C. Scott Hemphill, Less Restrictive Alternatives in Antitrust and Constitutional Law 2 (2015), http://scholarship.law.berkeley.edu/cgi/viewcontent.cgi?article=1031&context=law_econ [https://perma.cc/MCK2-LRGZ] (“[The Sherman Act’s] glittering generalities leave courts with a lot of work to do.”). See generally Guido Calabresi, A Common Law for the Age of Statutes (1985) (arguing that the primacy of statutes in modern America has upset the judicial–legislative balance and advocating for a more robust role for common law reasoning). Indeed, the enacting legislatures may have intended to delegate the authority to develop causation standards to the courts. 199                                                                                See generally Margaret H. Lemos, The Other Delegate: Judicially Administered Statutes and the Nondelegation Doctrine, 81 S. Cal. L. Rev. 405 (2008) (arguing that courts exercise lawmaking functions when filling statutory gaps). This is consistent with courts and commentators’ sense that causation in employment discrimination is more “tort-like” than in the average statutory question. 200 See, e.g., Univ. of Tex. Sw. Med. Ctr. v. Nassar, 133 S. Ct. 2517, 2524–25 (2013) (treating employment discrimination as a species of tort); Sperino, Franken-Tort, supra note 71, at 721 (“In a series of cases, the U.S. Supreme Court imported tort law into federal discrimination law.”). The more “tort-like” a statutory claim is, arguably, the less pressing the concerns about legislative preference and efficiency (because the enacting legislatures presumably knew that standards would need to be developed and elected to leave that task to the courts) and judicial legitimacy (because state courts are more in the business of common law reasoning than federal courts). 201 See Kaye, supra note 93, at 6 (“That state courts—not federal courts—are the keepers of the common law has long been American orthodoxy.”).

Although there is potentially some loss to judicial efficiency that results from the divergence of state and federal standards, this is no more than the general inefficiency involved in federal courts adjudicating state law claims that are outside their expertise. In fact, because federal courts must apply the motivating-factor analysis for Title VII status-based claims by statute, 202 See supra section I.C. they will be well versed and will not have to learn any separate state causation tests. As to juror confusion, jurors must already understand two causation standards when plaintiffs bring federal age and sex discrimination claims together, or when they bring race discrimination and retaliation claims together. There is no compelling reason to believe that having to consider the same type of discrimination under two different standards should be any more confusing for jurors than applying different standards to different types of discrimination.

Finally, as suggested above, too much state following of federal law poses legitimacy concerns at least as grave as those raised by too little following. 203 See supra notes 173–176 and accompanying text. Parallel construction means that state courts presumptively follow “even abrupt and counterintuitive changes in federal law.” 204 See Dodson, supra note 99, at 705. If done uncritically, following federal law could misinterpret state law 205 See supra note 173  and accompanying text. and degrade states as quasi-independent sovereigns in our federal system. 206 See supra note 176 and accompanying text.

As to the dialogic values that some federalism theorists espouse, it is possible that parallel state and federal statutes provide fewer opportunities for productive dialogue than parallel constitutional provisions. Unlike the establishment of a new fundamental right under a state or federal constitution, or a novel or unexpected statutory interpretation expanding the previously understood reach of the statute, 207 See, e.g., Hively v. Ivy Tech Cmty. Coll., 853 F.3d 339, 341 (7th Cir. 2017) (holding that discrimination on the basis of sexual orientation is sex discrimination under Title VII). the applicable causation standard in various types of discrimination suits does not provide lower federal courts any room to engage in dialogue: They are bound by straightforward pronouncements by the Supreme Court and, if called upon to interpret the analogous state statutes, will simply follow the state supreme court. There is good reason to believe that disagreement among the state supreme courts as to the proper causation standard will have little direct effect on the interpretation of federal law.

In any event, as noted above, variation can be valuable even if it does not lead to a future consensus. 208 See supra note 172 and accompanying text. Interjurisdictional dialogue accumulates knowledge and may lead to better legal rules. Independent of any questions about the preferences of the enacting legislatures, variation may lead to rules that are better suited to local conditions and public preferences. This is all the more true because of the closeness of the question at the Supreme Court. Because a bare minority of Justices in Price Waterhouse interpreted “because” to mean motivating-factor causation 209 See supra section I.B. and a bare majority of Justices in Gross and Nassar found it to mean but-for causation, 210 See supra section I.D. it seems likely that some state electorates prefer the Price Waterhouse approach. Moreover, variation may improve accountability by disallowing state judges from hiding behind federal case law to avoid political backlash. 211 See Dodson, supra note 99, at 740. See generally David E. Pozen, The Irony of Judicial Elections, 108 Colum. L. Rev. 265 (2008). These benefits of independent construction outweigh the at-best-equivocal arguments in favor of parallel construction.

Conclusion

As federal employment discrimination and retaliation laws continue to be interpreted in more plaintiff-unfriendly ways, the role of state courts and state statutes in the broader employment discrimination project grows potentially larger. When the Supreme Court all but repudiated Price Waterhouse—which made a fair attempt to effectuate the congressional purpose regarding causation in employment discrimination litigation—in favor of a narrow textualist approach, it likely reduced the number of viable discrimination and retaliation claims in federal court. State courts interpreting analogous provisions in state antidiscrimination law face a dilemma as to how closely to adhere to federal doctrine. Ultimately, concerns about judicial federalism should not restrain courts from exercising their independent judgment, lest they cede their authority in our federal scheme to vindicate the rights of their state’s citizens.