Greater Phila Chamber Commerce v. City of Philadelphia

PRECEDENTIAL UNITED STATES COURT OF APPEALS FOR THE THIRD CIRCUIT _____________ Nos. 18-2175 & 18-2176 _____________ GREATER PHILADELPHIA CHAMBER OF COMMERCE, Individually and on behalf of its members, Appellant in No. 18-2176 v. CITY OF PHILADELPHIA; PHILADELPHIA COMMISSION ON HUMAN RELATIONS, Appellants in No. 18-2175 ________________ On Appeal from the United States District Court for the Eastern District of Pennsylvania (D.C. No. 2-17-cv-01548) District Judge: Honorable Mitchell S. Goldberg ________________ Argued on March 15, 2019 Before: McKEE, ROTH, and FUENTES, Circuit Judges (Opinion filed: February 6, 2020) Benjamin H. Field Jane L. Istvan Nicole S. Morris Marcel S. Pratt, Esquire (Argued) City of Philadelphia Law Department 1515 Arch Street Philadelphia, PA 19102 Counsel for Appellants/Cross-Appellees Adam R. Pulver Scott L. Nelson Public Citizen Litigation Group 1600 20th Street NW Washington, DC 20009 Counsel for Amicus Public Citizen Inc. Maura Healey Elizabeth N. Dewar Genevieve Nadeau Erin K. Staab Office of Attorney General Massachusetts One Ashburton Place McCormack Building Boston, MA 02108 Counsel for Amicus Commonwealth of Massachusetts; District of Columbia; Commonwealth of Puerto Rico; Commonwealth of Virginia; State of Connecticut; State of Delaware; State of Illinois; State of New Jersey; State of New Jersey; State of New York; State of Vermont; State of Washington Zachary W. Carter Richard Dearing Devin Slack Eric Lee Jamison Davies New York City Law Department Room 6-178 100 Church Street New York, NY 10007 Counsel for Amicus City of New York; City of Berkeley; City of Columbus; City of Oakland; County of Santa Clara; City and County of San Francisco; City of Seattle; City of South Bend Terry L. Fromson Amal Bass Women’s Law Project 125 South 9th Street Suite 300 Philadelphia, PA 19107 Counsel for Amicus Womens Law Project; 2 36 Organizations Dedicated to Gender Wage Equity Richard A. Samp Cory L. Andrews Washington Legal Foundation 2009 Massachusetts Avenue, N.W. Washington, DC 20036 Counsel for Amicus Washington Legal Foundation Kellam M. Conover Miguel A. Estrada (Argued) Amir C. Tayrani Gibson, Dunn & Crutcher LLP 1050 Connecticut Avenue, N.W. Washington, D.C. 20036 Franco A. Corrado Marc J. Sonnenfeld Morgan, Lewis & Bockius 1701 Market Street Philadelphia, PA 19103 Kevin M. Siegel Suite 201 10000 Lincoln Drive East Marlton, NJ 08053 Counsel for Appellee/Cross-Appellant Michael L. Kichline Michael H. McGinley Dechert 2929 Arch Street 18th Floor, Cira Centre Philadelphia, PA 19104 Counsel for Amicus African American Chambers of Commerce of Pennsylvania., New Jersey and Delaware Robert L. Byer John E. Moriarty Robert M. Palumbos Andrew R. Sperl Duane Morris 3 30 South 17th Street Philadelphia, PA 19103 ________________ OPINION OF THE COURT ________________ Table of Contents I. BACKGROUND......................................................... 6 A. The Disparity And The Ordinance ....................... 8 B. Legislative Background ......................................... 9 1. Testimony Before the City Council ............. 10 a. Barbara Price .......................................... 10 b. Terry Fromson ........................................ 11 c. Marianne Bellesorte ............................... 14 d. Rue Landau ............................................. 15 2. Other Testimony Before the City Council .. 16 C. The Legal Challenge ............................................ 17 1. The Madden Affidavit .................................. 18 2. Declarations Filed by Chamber Members .. 23 D. The District Court Opinion ................................ 24 II. DISCUSSION .......................................................... 25 A. The Reliance Provision ........................................ 27 1. The District Court Correctly Concluded an Injunction as to the Reliance Provision Fails 4 Because the Provision does not Implicate Speech ............................................................ 27 2. None of the Chamber’s Arguments Call into Question the District Court’s Conclusions . 29 B. The Inquiry Provision ......................................... 32 1. The Legal Standard ...................................... 32 a. Commercial Speech ................................ 32 b. Intermediate Scrutiny Under Central Hudson is Appropriate ........................... 34 c. Strict Scrutiny is Inappropriate Here ... 35 2. The Inquiry Provision Satisfies Central Hudson Intermediate Scrutiny .................... 38 a. The Speech at Issue is not “Related to Illegal Activity” ........................................ 39 b. The City has a Substantial Interest in Closing the Wage Gap ............................ 41 c. The Inquiry Provision Directly Advances the City’s Interest in Pay Equity ........... 42 i. Caselaw Considering Whether a Legislature Relied on Substantial Evidence to Support a Speech Restriction Under Central Hudson Demonstrates that the City Presented Sufficient Evidence to Support the Ordinance ............... 45 ii. The Evidence Here is Stronger Than the Evidence Supporting the Restrictions in Florida Bar and King .............................................. 55 5 d. The Inquiry Provision is not More Extensive than Necessary ....................... 61 III. CONCLUSION ......................................................... 67 McKEE, Circuit Judge This appeal requires us to decide whether a Philadelphia Ordinance that prohibits employers from inquiring into a prospective employee’s wage history in setting or negotiating that employee’s wage violates the First Amendment. The district court held the Ordinance unconstitutional insofar as it prohibits that inquiry. However, the court upheld the provision of the Ordinance that prohibits reliance on wage history based on the court’s conclusion that such reliance did not implicate protected speech. For the reasons that below, we affirm the court’s order insofar as it upholds the Reliance Provision but reverse it insofar as it strikes down the Inquiry Provision. I. BACKGROUND In 2017, the City of Philadelphia enacted an ordinance to address the disparity in the pay of women and minorities that is often called the “pay gap.” The Ordinance contains two provisions: the “Inquiry Provision,” which prohibits an employer from asking about a prospective employee’s wage history, and the “Reliance Provision,” which prohibits an employer from relying on wage history at any point in the process of setting or negotiating a prospective employee’s wage. The Greater Philadelphia Chamber of Commerce filed this suit, individually and on behalf of some of its members, alleging that both provisions of the Ordinance infringe on the freedom of speech of the Chamber and its members. The Chamber concedes that the pay gap exists, and that the City has a substantial governmental interest in addressing it. However, the Chamber argues that the City passed the Ordinance “with only the barest of legislative records” and, therefore, did not present sufficient evidence to establish that the Ordinance would satisfy the City’s objective.1 1 Chamber Br. at 1. 6 Accordingly, the Chamber claims that the Ordinance cannot survive its First Amendment challenge under either strict or intermediate scrutiny. The district court agreed that the Inquiry Provision violated the First Amendment speech rights of employers and invalidated that part of the Ordinance. But the court concluded that the Reliance Provision withstood the Chamber’s First Amendment challenge because it did not impact speech. As we explain below, we conclude that the district court erred in holding that the Inquiry Provision was unconstitutional. We believe the court’s analysis of that provision applied a much higher standard than required. The Supreme Court has not demanded that the enacting authority achieve legislative certainty or produce empirical proof that the adopted legislation would achieve the stated interest even when applying strict scrutiny. Rather, the appropriate inquiry requires courts to determine whether the legislature “has drawn reasonable inferences based on substantial evidence.”2 The Supreme Court has even “permitted litigants to justify [analogous] speech restrictions by reference to studies and anecdotes pertaining to different locales altogether, or even, in a case applying strict scrutiny, to justify restrictions based solely on history, consensus, and ‘simple common sense.’”3 In short, the Supreme Court has upheld similar restrictions based on much less evidence than the City presented here. 2 Turner Broad. Sys., Inc. v. F.C.C., 520 U.S. 180, 195 (1997) (“Turner II”) (internal quotation marks and citation omitted). 3 Florida Bar v. Went For It, Inc., 515 U.S. 618, 628 (1995) (internal citations omitted). See also City Br. at 43-44 (citing Heffner v. Murphy, 745 F.3d 56, 92 (3d Cir. 2014), WV Ass’n of Club Owners & Fraternal Servs., Inc. v. Musgrave, 553 F.3d 292, 303–04 (4th Cir. 2009), Coyote Pub., Inc. v. Miller, 598 F.3d 592, 608 (9th Cir. 2010) (illustrating recent decisions reflecting the Supreme Court’s flexible approach to speech restrictions under intermediate scrutiny. As we explain below, we conclude that the Ordinance need only satisfy intermediate scrutiny.) 7 A. The Disparity And The Ordinance According to the 2015 census, women in Pennsylvania earned 79 cents for every dollar earned by similarly situated men.4 For women of color, the wage gap is even more profound. Black women earn 68 cents for every dollar paid to similarly situated men, and Latina women earn 56 cents for every dollar paid to similarly situated men.5 The gap begins for women as soon as they enter the workforce. In just the first year after college, full- time working women earn, on average, just 82% of what their male peers earn.6 Overall, women under the age of 35 earn 88- 91% of what their male peers earn.7 Rather than improve, as women gain experience in the work force the situation gets worse. Women aged 35 and over earn only 77-81% of what male peers earn.8 In response to this persistent wage disparity, the City of Philadelphia enacted the Ordinance at the center of this dispute. The Ordinance states: It is an unlawful employment practice for an employer . . . (i) To inquire about a prospective employee’s wage history, require disclosure of wage history, or condition employment or consideration for an interview or employment on disclosure of wage 4 See JA119–20 (discussing the City Council’s legislative findings supporting the Ordinance). 5 Id. 6 See, e.g., Christianne Corbett & Catherine Hill, Am. Ass’n of Univ. Women (“AAUW”), Graduating to a Pay Gap: The Earnings of Women and Men One Year After College Graduation, at 9 (Oct. 2012), https://www.aauw.org/files/2013/02/graduating-to-a-pay-gap- the-earnings-of-women-and-men-one-year-after-college- graduation.pdf. 7 U.S. Bureau of Labor Statistics, Highlights of Women’s Earnings in 2017, at 9 (Aug. 2018), https://www.bls.gov/opub/reports/womens- earnings/2017/pdf/home.pdf. 8 Id. 8 history, or retaliate against a prospective employee for failing to comply with any wage history inquiry. (ii) To rely on the wage history of a prospective employee from any current or former employer of the individual in determining the wages for such individual at any stage in the employment process, including the negotiation or drafting of any employment contract, unless such applicant knowingly and willingly disclosed his or her wage history to the employer, employment agency, employee or agent thereof. (c) For purposes of this Section 9-1131, “to inquire” shall mean to ask a job applicant in writing or otherwise. . . . 9 Employers who violate the Ordinance are subject to civil and criminal penalties, including compensatory damages, up to $2,000 in punitive damages per violation, and an additional $2,000 and 90 days’ incarceration for a repeat offense.10 B. Legislative Background The City seeks to justify the Ordinance by relying on the testimony of witnesses who testified before the City Council prior to the enactment of the Ordinance and an affidavit by Dr. Janice Madden that the City submitted to the district court in response to the Chamber’s constitutional challenge. Dr. Madden reviewed thousands of peer-reviewed research studies and concluded, among other things, that “there is wage discrimination in the labor market, suppressing the prior wages of women and minorities” and this “is consistent with the findings of thousands of research studies.”11 She concluded that “these scholarly studies show . . . that significant and substantial wage differentials by race and gender, which are 9 Phila. Code § 9-1131. 10 Id. §§ 9-1105(1)(c)–(d), 9-1121(2). 11 JA297. 9 not explained by credentials or qualification, persist.”12 The Chamber presented no evidence challenging any of Dr. Madden’s conclusions or the studies those conclusions were based on. 1. Testimony Before the City Council a. Barbara Price Barbara Price, the Public Policy Chair of the American Association of University Women, testified before the Council and submitted written testimony. She reiterated that “the [wage] gap still exists today at 80 cents nationally and 79 cents for Pennsylvania, which ranks [Pennsylvania] 27th as a state in the country.”13 She confirmed that “[t]he gap remains consistent across age groups, levels of education, and for full- time workers across a number of occupations.”14 She discussed research that showed, after accounting for choice of occupation, hours worked, economic sector, experience, grade point average, undergraduate institution, marital status and other factors, a significant gap between the earnings of men and women remained—beginning one year after graduation and widening in the years thereafter.15 For example, in 12 JA298 (emphasis added). 13 Council of the City of Philadelphia Committee on Law and Government: Hearing on Bill No. 160840, (Nov. 22, 2016) (hereinafter “Hearing Transcript” or “Hr’g Tr.”) at 70, available at http://legislation.phila.gov/transcripts/Public%20Hearings/la wngov/2016/lg112216.pdf; see also JA272-277. 14 JA273. 15 JA275 (citing AAUW, The Simple Truth About the Gender Pay Gap, at 17 (Fall 2014), https://fortunedotcom.files.wordpress.com/2014/12/the- simple-truth_fall.pdf (“After accounting for college major, occupation, economic sector, hours worked, months unemployed since graduation, GPA, type of undergraduate institution, institution selectivity, age, geographical region, and marital status, Graduating to a Pay Gap found that a 7 percent difference in the earnings of male and female college graduates one year after graduation was still unexplained.”). 10 Philadelphia, “[t]he single most common occupation for Latinas is that of maids, housekeepers, janitors or building cleaners where they make up 22 percent of the people employed in those jobs.”16 However, “Latinas who work full time in these occupations, year round, are paid just 58 cents for every dollar paid to White, non-Hispanic men in the same occupations.”17 She testified that the pay gap “costs a typical woman in Pennsylvania about $918,120 over the course of her career.”18 b. Terry Fromson Terry L. Fromson, the Managing Attorney for the Women’s Law Project, testified before the City Council that the practice of obtaining and using wage history to set pay is one contributor to the pay gap.19 She told the City Council that “a sizable wage gap exists between men and women in Pennsylvania, one that is substantially larger for women of color.”20 She testified that unequal pay “has persisted despite the existence of equal pay laws banning sex discrimination [in] wages for five decades.”21 She explained that discrepancy in pay continues, in part, because current laws targeting discrimination, such as “the Equal Pay Act[,] specifically[] allow[] employers to justify paying women less than men based on what is described as a factor other than sex.”22 Ms. Fromson explained that “many courts have interpreted prior wages as a factor other than sex, when in fact, it is typically not. It is not gender neutral.”23 She elaborated, “[a] woman’s prior pay may very well be based on a sex discriminatory assessment of her worth. It reflects historical market forces based on sex stereotypes and 16 JA273. 17 Id. 18 Id. 19 JA268; see also H’rg Tr. at 63-69. 20 H’rg Tr. at 65. 21 Id. at 65–66. 22 Id. at 66. 23 Id. 11 assumptions about the value of the equal work of one sex over the other.”24 Fromson also gave a detailed explanation of how wage history perpetuates and institutionalizes wage discrimination. “Wage policies challenged in recent years show how this happens.”25 One’s initial salary at a given employer is based in part upon the salary of the employee’s most recent job. “The wage gap data tells us that the woman’s salary is most likely less than the man who is equally situated to her.”26 Subsequent pay is then based on that starting salary “plus an increment that would be applied equally to the men and the women. . . . [E]very time a salary increase happens, an equal percentage of prior pay is applied. And so, women . . . remain paid less than men.”27 In other words, the initial discrepancy in pay is baked into all future pay increases, even in workplaces in which pay is increased at the same percentage for similarly situated men and women. Fromson told the Council: “[b]y specifically outlawing the practice of relying on prior wages to set a new employee’s pay, this [O]rdinance will provide clarity that will relieve women of having to gamble on whether a court will properly interpret this practice as unlawful” discrimination.28 It will therefore help ensure that wage growth and wage decisions are based on qualifications and job requirements “rather than a factor that likely reflects longstanding gender-based wage disparities in the labor market.”29 She also noted that the EEOC recognizes 24 Id. 25 Id. 26 Id. at 66–67 (emphasis added). This testimony does not distinguish between salary and wage discrepancy and the text of the Ordinance refers only to a wage discrepancy. However, we can discern no significant distinction for the purposes of our discussion and much of the testimony strongly suggests that the compensation gap that Fromson and others referred to for salaries is indistinguishable from the compensation gap in wages. 27 Id. at 67. 28 Id. 29 Id. 12 prior salaries of job candidates can reflect sex-based compensation discrimination.30 Fromson informed the Council that Massachusetts had approved a similar ban on inquiries into wage history and New York City had adopted an executive order to that effect insofar as public employees were concerned.31 She added that similar legislation was then pending in several jurisdictions including New Jersey, the District of Columbia, and New York City (to expand to all employers in the city and not just municipal employers).32 Fromson also suggested that the Council consider adding two clarifications to the proposed Ordinance based on provisions in the Massachusetts law and New York City executive order. The first provision would have prohibited employers from seeking an employee’s wage history from current or former employees. She noted that “while [the Ordinance] prohibits an employer from asking a job applicant for wage history, it does not bar inquiries directly to current and former employers.”33 The second suggested change was to allow reliance on wage history after an offer of employment and compensation had been made.34 She explained that without the second change, the proposed legislation allowed employers to consider wage history if the employee volunteered it before an offer was 30 Id. at 66. 31 Id. at 64. 32 Id. The state of New Jersey has since passed a similar wage history inquiry and reliance ban. See Act of July 25, 2019, Pub. L. No. 2019, c.199 (N.J. 2019). That act states: “it shall be an unlawful employment practice for any employer: (1) to screen a job applicant based on the applicant’s salary history, including, but not limited to, the applicant’s prior wages, salaries or benefits; or (2) to require that the applicant’s salary history satisfy any minimum or maximum criteria.” As Fromson’s testimony suggests, New Jersey’s recently enacted law appears to be part of an emerging trend that recognizes the extent to which reliance on wage history inevitably perpetuates historic wage disparity. 33 H’rg Tr. at 68. 34 Id. 13 made.35 Fromson suggested that this “place[d] applicants in an untenable position of having to choose between protecting what is biased information that may adversely affect their future wages or [] risk being denied a job.”36 In her opinion, that was “an inherently coercive situation for someone to be in.”37 Ultimately, these added provisions were not incorporated into the Ordinance. Councilman William Greenlee, Chairman of the Committee of Law and Government, noted that, in declining to include these added provisions, the Philadelphia Ordinance did not go as far as other proposed wage history bans around the country. He told the committee, “the Massachusetts law goes a little wider than we do. We’re trying to keep it real basic[;] . . . you’ll hear from a witness that thinks we don’t make it strong enough, but we’re trying to find that great balance that we always try to in legislation. . . .”38 We are trying “at this point [to] limit it to stopping the employer from asking, directly asking, the prospective employee what they make.”39 He also noted “for the record, as far as the Chamber of Commerce goes, the Boston Chamber of Commerce supported the Massachusetts law. . . . [T]hey obviously did not believe it was that injurious to businesses.”40 c. Marianne Bellesorte Marianne Bellesorte, Vice President of Advocacy at Pathways PA,41 analyzed wage gap data and concluded that the wage gap for women and men of color “are compounded” when these individuals are asked to share their wage history.42 She told the Council “[o]ne step in addressing wage inequality is ensuring 35 Id. 36 Id. at 68–69. 37 Id. at 69. 38 Id. at 15. 39 Id. 40 Id. at 80. 41 “PathWays PA works to end the cycle of poverty, homelessness, and abuse in the Philadelphia region.” Id. at 74. 42 JA276. 14 that a history of low salaries does not follow women into a new workplace.”43 She emphasized that “the wage gap is not just about women. It is also about people of color, men and women.”44 She explained how wage inquiries perpetuate discrimination for women and minorities: “Inequitable wages start right out of college, and they’re compounded when women and [minorities] apply for new jobs and are asked to share their pay history. Instead of starting their job on an equal footing, they enter with a lower salary because it was based on previous employment.”45 She continued: “Not surprisingly as women get older, the wage gap continues to grow and continues to affect women in retirement.”46 Bellesorte also explained: “By preventing potential employers [from] asking for salary history, Philadelphia’s workers gain the ability to earn what their work is actually worth. A woman who starts her career at the low end of a salary range will not be held to that standard for the rest of her work life.”47 d. Rue Landau Rue Landau, the Executive Director of the Philadelphia Commission on Human Relations, told the committee that, “as the agency charged with enforcing the Fair Practices Ordinance . . . the PCHR understand[s] that the wage gap is real.”48 According to Ms. Landau, “women working in Pennsylvania are paid only 79 cents for every dollar a man earns.49 In real numbers, median annual earnings in 43 H’rg Tr. at 74. 44 Id. at 75. 45 Id. 46 Id. 47 Id. at 76. 48 Id. at 8. 49 Jovida Hill, the Executive Director of Philadelphia’s Commission on Women, affirmed the testimony of Ms. Landau with nationwide empirical evidence. She testified that, “[a]ccording to calculations by the National Committee on Pay Equity, for a woman with a high school education, the difference [in pay arising from the pay gap] can amount to $700,000, $1.2 million for a woman with a college degree, 15 Pennsylvania are $51,212 for a man and $40,742 for a woman.”50 She testified “the practice of asking about an applicant’s wage history during the hiring process can perpetuate wage inequality, low wages, and poverty. . . . [A] jobseeker who has suffered from the wage gap can only be harmed when required to disclose her salary history.”51 She concluded “the PCHR strongly believes[] that taking out any obstacles that employers could use . . . to discriminate is a very important thing to do. This is one of these barriers.”52 2. Other Testimony Before City Council The Chamber did not present any witnesses in opposition to the Ordinance, but it did submit written testimony from Rob Wonderling, President and CEO of the Chamber. He wrote that the Ordinance “goes too far in dictating how employers can interact with potential hires.”53 Rather surprisingly, he submitted that employers use wage history to “have a better understanding of whether a candidate is worth pursuing based on previous compensation levels as well as the market value or salaries for comparable positions.”54 That of course is exactly why the City Council was considering the Ordinance. It was trying to cut the Gordian knot that continues to tie past discriminatory wages to future job opportunities and wages so that employers would not decide if a given employee was “worth pursuing based on previous compensation levels.” Wonderling also asserted that “[i]n speaking with our members . . . we hear that compensation decisions are based on a number of different factors such as market value, internal equity, funding limitations and competition. It is not made based on a candidate’s past salary history, gender or race.”55 As discussed and $2 million for women with advanced degrees.” H’rg Tr. at 12. 50 Id. 6. 51 Id. at 8. 52 Id. at 23. 53 JA124. 54 Id. 55 Id. This testimony is also surprising since, if salary history is not a factor in setting compensation levels, it is not at all 16 below, however, this claim was contradicted by the Chamber members’ own submissions to the district court in which they confirmed that they use wage history to set wages.56 The Chamber offered no testimony to refute the existence of the wage gap, the role of discrimination in the wage gap, or the conclusion that prohibiting inquiry into one’s wage history could help mitigate the wage gap. Based on this record, on December 8, 2016, after weighing the testimony and submissions, the City Council unanimously passed the Ordinance.57 C. The Legal Challenge In April 2017, the Chamber filed a Complaint and Motion for Preliminary Injunction alleging the Ordinance violated the First Amendment.58 The district court dismissed the original complaint for lack of standing. The Chamber addressed that deficiency in a subsequently filed Amended Complaint and refiled Motion for a Preliminary Injunction.59 The City responded to the Amended Complaint by submitting the affidavit of Dr. Janice Madden.60 clear how employers would be harmed or prejudiced by the Inquiry Provision. 56 See e.g., JA130 (Chamber Member Bittenbender: “Wage history information is essential to salary offers in positions where Bittenbender is unaware of the market wage.”). 57 JA283–89. On January 23, 2017, the bill was signed into law. JA122. 58 JA072-117. 59 JA74. The City has agreed that the Chamber and its members have standing to bring suit here. See JA250, ¶ 18; see also JA081-117. 60 See Affidavit of Janice F. Madden, Ph.D. JA291–306. Dr. Madden is a labor economist “with extensive experience in the analysis of labor markets and, in particular, gender and racial differentials in labor markets.” JA292. She attended the Wharton School after completing an M.A. and Ph.D. at Duke University and previously earned her B.A. in economics and mathematics at the University of Denver. She is a tenured faculty member of the University of Pennsylvania and teaches 17 1. The Madden Affidavit The City retained Dr. Madden, a highly respected labor economist, to summarize the research in each of the following areas: (1) the extent to which salaries of qualified job applicants have historically differed by race or gender; (2) the effect of starting salaries on the overall salary differentials of comparable qualified employees by race or gender– information that can be provided by an applicant’s salary history; and (3) whether there are “alternative sources of such information” to support the need for, and potential effectiveness of, the Ordinance. Her affidavit corroborated the testimony of the witnesses who had testified before the City Council.61 She concluded in her affidavit that “there is wage discrimination in the labor market[] suppressing the prior wages of women and minorities” and that this “is consistent with the findings of thousands of research studies.”62 Dr. Madden reviewed the research on pay differentials by race and gender for workers with equivalent skills and experience. She stated that “[h]undreds, possibly thousands, of scholarly studies over the years have decomposed the overall gender and undergraduate and graduate “courses dealing with economics, labor markets, and . . . relevant statistical methods.” Id. In addition, Dr. Madden has authored five books on economics and discrimination and has testified as an expert witness in over 45 cases in federal and state courts. JA292–93. 61 Although her affidavit was not before the Council when the Ordinance was passed, it was appropriately considered by the district court. As we have previously recognized, “[i]f a legislative body can produce in court whatever justification is required of it under the applicable constitutional doctrine, we perceive little to be gained by incurring the expense, effort, and delay involved in requiring it to reenact the legislative measure after parading its evidence through its legislative chamber.” Phillips v. Borough of Keyport, 107 F.3d 164, 178 (3d Cir. 1997). The district court was therefore correct in “consider[ing] post-enactment evidence offered in support of City Council’s decision.” Id. 62 JA297. 18 racial pay gaps into the proportion arising from gender and racial differences in experience, education, training, work hours, occupations and industries.”63 Madden concluded that “these scholarly studies show that the pay gap remains when comparing only men and women or minorities and non- minorities with the same education, experience, training, work hours, occupations and industries.”64 These studies found “significant and substantial wage differentials by race and gender, which are not explained by credentials or qualification, persist.”65 Dr. Madden reached several other conclusions based on her survey of the voluminous research supporting the need for the Ordinance, including that: • “Labor market researchers are in general agreement that women and/or members of racial and ethnic minorities have received and currently receive lower wages than comparably qualified and performing men and/or members of majority racial and ethnic groups.”66 • “Antidiscrimination laws, including the Civil Rights Act and the Equal Pay Act, have not eliminated the lower wages generally received by women and minority workers relative to men and majority workers of equivalent skill, ability, experience, and performance.”67 • “Starting salaries typically differ by race and gender for workers of equivalent skills and abilities.”68 63 JA298. 64 Id. (emphasis added). 65 Id. 66 JA294. 67 Id. 68 Id. 19 • “The available evidence shows that when employers do not have access to salary history, they easily obtain information on past performance and skills of applicants and they select hires with this information as effectively as those using salary histories.”69 According to Dr. Madden, denying employers information about a perspective employee’s wage history does not deprive a perspective employer of information needed to make an informed employment decision, including determining an appropriate wage. Concomitantly, putting such wage history beyond the reach of new employers helps break the discriminatory chain linking an employee’s new salary to past salaries and any discriminatory judgments that may have influenced those past salaries.70 The studies cited in Dr. Madden’s affidavit included comprehensive reviews of scores of other studies. For example, she cites Stanley and Jarrell who performed a meta-regression analysis of fifty-five other studies and concluded that there is a “wide consensus that gender wage discrimination exists” and the “vast empirical economic literature, containing hundreds of studies, reveals that women are ‘underpaid’ disproportionate to their observed skills.”71 That study focused on determining the extent of the reported gaps. Dr. Madden also relied upon the research of Blau and Kahn, who found in their review of data from 1980 to 2010, “an unexplained gender wage gap in each year[’s data].”72 They explained that the “finding of such an unexplained gap is 69 JA295. 70 JA305-06. 71 T.D. Stanley & Stephen B. Jarrell, Gender Wage Discrimination Bias? A Meta- Regression Analysis, 33 J. Hum. Resources 947, 948 (Fall 1998) (hereinafter “Stanley & Jarrell”). 72 Francine D. Blau & Lawrence M. Kahn, The Gender Wage Gap: Extent, Trends, and Explanations, 31, NBER Working Paper No. 2193, National Bureau for Economic Research (2016), http://www.nber.org/papers/w21913 (published in 55 J. of Econ. Lit. 789 (2017)) (hereinafter “Blau & Kahn”). 20 fairly standard in the literature” and is “taken as an estimate of labor market discrimination.”73 Dr. Madden also cited Wilson and Rodgers who concluded “discrimination has consistently played a major role” in “the widening of racial wage gaps since 1979.”74 This study focused on the minority wage gap and the causes of the gap for specific minority sub-groups. It concludes, among other things: “Between the Great Recession of 2007–2009 and 2015, gaps among new-entrant women expanded more than among any other experience/gender group. The same factor that dominated prior to 2000—growing labor market discrimination—is the primary source of the erosion.”75 Additionally, “[a]mong black college graduates, growing discrimination was essentially the sole cause of the [wage] gap’s expansion, far outweighing the advantages black college graduates gained as a result of being slightly older (i.e., more experienced) than their white counterparts.”76 Her distilled conclusions of these studies were that “labor market discrimination continues to contribute to the wage gap;” “discriminatory wages persist;” and the “racial wage gap [is] increasing.”77 To eliminate the effect of variables other than race or gender such as: education, experience, training, occupation, and industry, which could explain the wage gap, Madden also cited the studies relied on by Blau and Kahn whose reviews focused on homogenous populations within the same industry. For example, they analyzed studies within a group of lawyers and MBAs that were able to control for very detailed characteristics, including, for example, grade point averages 73 Id. 74 Valerie Wilson & William M. Rodgers III, Black-White Wage Gaps Expand with Rising Wage Inequality, 4, Economic Policy Inst., (September 19, 2016) https://www.epi.org/files/pdf/101972.pdf (hereinafter “Wilson & Rodgers”). 75 Id. at 5. 76 Id. at 27. 77 JA298 n.3. 21 while in school.78 “The studies of lawyers and MBAs . . . find that, even if one accounts for variables related to family status, like work force interruption and fewer hours worked, unexplained gender earnings differences remain which are potentially due to discrimination.”79 Among lawyers, “men earned 11 percent more, controlling for an extensive list of worker qualifications and other factors, including grades while in law school, detailed work history data, and type and size of employer.”80 Among MBAs, “men earned nearly 7 percent more even accounting for work force interruptions, fewer hours worked, and gender differences in business school GPAs and finance courses taken.”81 Blau and Kahn also reviewed experimental studies that similarly concluded discrimination is a primary cause of the wage gap. The authors believed that experiments “provide[] particularly persuasive evidence of discrimination . . . [because] they offer estimates of the role of discrimination that are potentially less contaminated by unmeasured factors.”82 For example, the authors describe an experimental study that not only replicated the gender wage gap in otherwise identical candidates, but also showed that starting salaries for women in the study were set far lower than the (otherwise identical) male candidates. In the experiment, employers reviewed “the application materials of (fictitious) [applicants] who[m] they were told . . . applied for a science laboratory manager position.”83 Study participants “rated the male applicant as significantly more competent and suitable for the position than the (identical) female applicant. Participants also set a starting salary for male applicants that was almost $4,000 higher than the salary offered to female applicants and offered more career mentoring to the male applicants.”84 Blau and Kahn conclude this research “strongly suggests that discrimination cannot be 78 Blau & Kahn at 32. 79 Id. 80 Id. 81 Id. 82 Id. 83 Id. at 33. 84 Id. 22 discounted as contributing to the persistent gender wage gap.”85 In addition to synthesizing the conclusions reached in various studies, some of which are highlighted above, Dr. Madden’s affidavit relies on her “consulting experience with a wide range of employers over forty years.”86 That experience corroborates that “gender and racial pay gaps between otherwise equivalent workers largely arise from gender and racial differences in the salary set at hire.”87 2. Declarations Filed by Chamber Members Members of the Chamber filed declarations in support of their Motion for Preliminary Injunction. Those declarations asserted that a wage history ban would harm businesses because they use wage history as a factor in making salary offers and for other purposes.88 For example, Chamber Member Bittenbender stated that “[w]age history information is essential to salary offers in positions where [it] is unaware of the market wage.”89 Similarly, Comcast asserted it “frequently inquires” about an applicant’s “previous compensation and wage history,” among other things, to “understand the level of responsibility the applicant had,” evaluate the value the prior employer placed on the candidate, and “determine market wage for similar positions.”90 Similarly, the Children’s Hospital of Philadelphia submitted that it “relies on wage history in making a salary offer.”91 The Chamber and its members, however, presented no evidence that refuted or challenged the testimony before the City before passing the Ordinance. That evidence showed that prior wages of women and minorities is more indicative of compounded discrimination than an accurate assessment of the individual’s value to their prior employer. Thus, information obtained to 85 Id. at 50. 86 JA300. 87 Id. 88 See JA126–247. 89 JA130. 90 JA137–38. 91 JA147. 23 assess the applicant’s market value only perpetuates wage disparity. D. The District Court Opinion As we noted at the outset, the district court granted the Chamber’s motion for a preliminary injunction as to the Inquiry Provision. The court held that it likely violated the Chamber’s and its members’ free speech rights. However, it found that the Reliance Provision–which prohibits relying on an applicant’s wage history at any point in the process– regulated conduct rather than speech. Accordingly, the court refused to enjoin that provision.92 The court reasoned that the Reliance Provision, is “not subject . . . to First Amendment scrutiny” because the provision “does not ‘on its face, implicate the spoken or written word.’”93 Instead, “[t]o the extent the Reliance Provision is content- or speaker-based,” the court found the Reliance Provision “targets conduct and not speech.”94 Because the Chamber did not meet its burden of showing that the provision implicates speech, no First Amendment analysis was required.95 However, the court found that the Inquiry Provision did implicate speech and that it could not survive even the less stringent intermediate scrutiny required under the First Amendment.96 Thus the court did not discuss the actual level of scrutiny required to withstand the Chamber’s First Amendment challenge.97 Rather, the court held that the Ordinance was unconstitutional under the less stringent 92 See Chamber of Commerce, 319 F. Supp. 3d at 779. 93 Id. at 801, 803. 94 Id. at 803–04. 95 Id. at 804. 96 Id. at 785. 97 Id. (“[B]ecause I conclude infra that the Inquiry Provision does not pass muster under the Central Hudson framework, I need not determine whether the Central Hudson test has been broadened for content- or speaker-based restrictions. I will thus apply Central Hudson’s intermediate scrutiny to the Inquiry Provision.”). 24 standard of Central Hudson Gas & Electric Corp. v. Public Service Commission of New York.98 That decision rested on the court’s belief that the City had not presented substantial evidence to support a conclusion that the Inquiry Provision would help close the wage gap.99 The district court determined that the requirements for a preliminary injunction were met with respect to the Inquiry Provision because “the Chamber ha[d] alleged a real and actual deprivation of its and its members' First Amendment rights through declarations.”100 Accordingly, it found, “the City cannot claim a legitimate interest in enforcing an unconstitutional law” because “there is a significant public interest in upholding First Amendment principles.”101 This appeal and cross appeal followed. The Chamber argues that the district court erred in refusing to enjoin the Reliance Provision and that both provisions should have been reviewed under strict scrutiny. The City alleges the district court erred in enjoining the Inquiry Provision. II. DISCUSSION A preliminary injunction “is an extraordinary remedy, which should be granted only in limited circumstances.”102 As the 98 447 U.S. 557 (1980). 99 Chamber of Commerce, 319 F. Supp. 3d at 800 (“I conclude that there is insufficient evidence to establish the alleged harm of discriminatory wages being perpetuated in subsequent wages such that they contribute to a discriminatory wage gap.”). 100 Id. at 807 (citing e.g., Wonderling Decl. ¶¶ 16, 22 (“If the Ordinance is allowed to stand, it will harm the Chamber’s members named in the First Amended Complaint as well as other members within the Chamber’s broader membership by preventing them from making wage-history inquiries that they otherwise normally would make.”)). 101 Id. 807–08. 102 Instant Air Freight Co. v. C.F. Air Freight, Inc., 882 F.2d 797, 800 (3d Cir. 1989) (internal citations omitted). 25 district court explained, the moving party must establish four factors to get a preliminary injunction: (1) the likelihood that the plaintiff will prevail on the merits at final hearing; (2) the extent to which the plaintiff is being irreparably harmed by the conduct complained of; (3) the extent to which the defendant will suffer irreparable harm if the preliminary injunction is issued; and (4) [that] the public interest [weighs in favor of granting the injunction].103 Generally, the moving party must establish the first two factors and only if these “gateway factors” are established does the district court consider the remaining two factors.104 The court then determines “in its sound discretion if all four factors, taken together, balance in favor of granting the requested preliminary relief.”105 In First Amendment cases the initial burden is flipped. The government bears the burden of proving that the law is constitutional; thus, the plaintiff “must be deemed likely to prevail” if the government fails to show the constitutionality of the law.106 This is because “‘the burdens at the preliminary injunction stage track the burdens at trial,’” and the burden of proving the constitutionality of a law rests with the government.107 103 A.T.&T. Co. v. Winback & Conserve Program, Inc., 42 F.3d 1421, 1427 (3d Cir. 1994) (internal citations omitted) (quoting Merch. & Evans, Inc. v. Roosevelt Bldg. Prods., 963 F.2d 628, 632–33 (3d Cir. 1992)). 104 Reilly v. City of Harrisburg, 858 F.3d 173, 179 (3d Cir. 2017), as amended (June 26, 2017). 105 Id. 106 Id. at 180 (quoting Ashcroft v. ACLU, 542 U.S. 656, 666 (2004)). 107 Id. (quoting Gonzales v. O Centro Espirita Beneficente Uniao do Vegetal, 546 U.S. 418, 429 (2006)). 26 Therefore, in First Amendment cases, the moving party must first “mak[e] a colorable claim” that the law restricts some form of speech.108 The government must then “justify its restriction on speech under whatever level of scrutiny is appropriate (intermediate or strict) given the restriction in question.”109 If the government succeeds in showing constitutionality, “then the motion for a preliminary injunction fails because there is no likelihood of success on the merits.”110 If the government cannot establish that the law is constitutional, the challenger must still demonstrate irreparable harm, though that is generally presumed where the moving party’s freedom of speech right is being infringed.111 We review the grant or denial of a preliminary injunction for “an abuse of discretion, an error of law, or a clear mistake in the consideration of proof.”112 We review de novo the lower court’s conclusions of law but review its findings of fact for clear error.113 A. The Reliance Provision 1. The District Court Correctly Concluded that an Injunction as to the Reliance Provision Fails Because the Provision Does Not Implicate Speech As explained above, the Reliance Provision makes it illegal for employers to “rely on the wage history of a prospective employee from any current or former employer of the individual in determining the wages for such individual at any stage in the employment process, including the negotiation or 108 Id. at 180 n.5 (quoting Thalheimer v. City of San Diego, 645 F.3d 1109, 1116 (9th Cir. 2011)). 109 Id. 110 Id. 111 Id. 112 Doe by & through Doe v. Boyertown Area Sch. Dist., 897 F.3d 518, 526 (3d Cir. 2018) (citing Kos Pharm., Inc. v. Andrx Corp., 369 F.3d 700, 708 (3d Cir. 2004)). 113 Id. 27 drafting of any employment contract.”114 The district court correctly concluded that this provision does not regulate speech. Accordingly, the court did not need to conduct a First Amendment analysis. As the court explained, the Reliance Provision does not “on its face, implicate the spoken or written word.”115 In arguing to the contrary, the Chamber claimed that the “Provision restricts [the] ability to communicate and/or convey a message.”116 The court found that here, unlike the situation in Wollschlaeger v. Governor of Florida117 and Holder v. Humanitarian Law Project,118 the conduct that the Reliance Provision regulates “is not executed through speech.”119 In Wollschlaeger, certain provisions of the Florida Firearms Owners’ Privacy Act (FOPA), prohibited medical professionals from, among other things, entering information about a patient’s gun ownership into medical records, or inquiring about gun ownership, and discriminating against a gun owner, unless the action was relevant to the patient’s care. In explaining why the case was not helpful, the district court correctly distinguished the “more specific actions” of “physical entry. . . into a patient log, making a written inquiry[] [and] asking a question” which “implicate[] speech on their face” from prohibiting reliance in the Ordinance.120 The district court also explained that not all of the provisions in Wollschlaeger were subject to First Amendment scrutiny. Like the Reliance Provision here, the Wollschlaeger court had concluded that the anti-discrimination provision of the FOPA did not “on its face, implicate the spoken or written word,” and therefore scrutiny under the First Amendment was not appropriate.121 114 Phila. Code. § 9-1131. 115 Chamber of Commerce, 319 F. Supp. 3d at 803. 116 Id. 117 848 F.3d 1293 (11th Cir. 2017). 118 561 U.S. 1 (2010). 119 Chamber of Commerce, 319 F. Supp. 3d at 803–04. 120 Id. at 803–04. 121 Id. at 803. 28 The statute at issue in Humanitarian Law Project banned providing “material support” to terrorist organizations.122 The Supreme Court found that the statute did implicate speech because it prohibited legal training and advice, which was support given “in the form of speech.”123 In rejecting the Chamber’s challenge to the Reliance Provision, the district court correctly concluded that “[h]ere, unlike in [Humanitarian Law Project], the conduct is not executed through speech. Reliance on wage history does not demand speech the way that providing legal advice necessarily does.”124 2. None of the Chamber’s Arguments Call into Question the District Court’s Conclusion The Chamber does not present any arguments before us that seriously challenge the district court’s reasoning or analysis of the Reliance Provision. The district court’s discussion of that provision is thorough, accurate, and persuasive. As the district court explained, the Reliance Provision does not restrain any expressive message. The Chamber argues that in “formulating a proposed salary,” a prospective employer is “communicating a message about how much that applicant’s labor is worth to the employer.”125 But the Reliance Provision does not restrict an employer from communicating an applicant’s worth. An employer may still discuss an applicant’s value based on his or her qualifications and abilities. The Ordinance simply attempts to prevent the employer from unknowingly incorporating past wage discrimination into the terms of an applicant’s job offer. The employer remains free to communicate its own valuation of the employee by making as many offers at whatever salary it 122 561 U.S. at 28. 123 Id. 124 Chamber, 319 F. Supp. 3d. at 804 (The “provisions [at issue] prohibited significantly more specific actions that implicated speech on their face[,] to the extent the Reliance Provision is content- or speaker-based, it targets conduct and not speech.”). 125 Chamber Br. at 29. 29 deems appropriate. The Ordinance merely attempts to ensure that any such offers are not unwittingly tethered to past discriminatory wage discrepancies. The Chamber also argues that because the Reliance Provision is “triggered” during the negotiation of a contract, it necessarily implicates speech.126 Consequently, the Chamber cites Valle Del Sol Inc. v. Whiting,127 and Centro de la Comunidad Hispana de Locust Valley v. Town of Oyster Bay,128 for the well settled proposition that negotiating the terms of an employment arrangement–either orally or in writing–is speech subject to the protections of the First Amendment. This argument relies upon a misreading of the Ordinance. The Reliance Provision is triggered not during negotiation but by the employer’s reliance on the employee’s wage history “at any stage in the employment drafting process.”129 The Chamber focuses on the phrase, “including the negotiating or drafting of the employment contract,” but that is merely one of the many “stage[s] of the employment process” during which the provision applies. It is not, as the Chamber argues, the conduct that makes the provision applicable. Moreover, even if the Chamber is correct that the Reliance Provision is “triggered” by negotiation, “it has never been deemed an abridgment of freedom of speech or press to make a course of conduct illegal merely because the conduct was in part initiated, evidenced, or carried out by means of language, 126 Chamber Br. at 29 (“[T]he conduct triggering coverage under the statute consists of communicating a message.”). 127 709 F.3d 808, 819 (9th Cir. 2013). 128 868 F.3d 104 (2d Cir. 2017). 129 Phila Code § 9-1131. “To rely on the wage history of a prospective employee from any current or former employer of the individual in determining the wages for such individual at any stage in the employment process, including the negotiation or drafting of any employment contract, unless such applicant knowingly and willingly disclosed his or her wage history to the employer, employment agency, employee or agent thereof.” Id. 30 either spoken, written, or printed.”130 As explained by the Supreme Court in National Institute of Family & Life Advocates v. Becerra,131 regulations that have an incidental impact on speech are not unconstitutional violations of the freedom of speech. The district court recognized that, to the extent that the Reliance Provision has an arguable effect on speech, it is incidental to the targeted reliance and does not place the provision under First Amendment scrutiny. Moreover, Valle Del Sol and Centro de la Comunidad Hispana de Locust Valley both dealt with ordinances that regulated day laborers’ abilities to advertise their availability for work.132 Advertising is prototypical speech that depends on spoken or written communication. Here, by contrast, the only activity being regulated by the Reliance Provision is the act of relying on wage history to set a salary. Under the Ordinance, the speech component of the negotiation process, i.e., the communication of a wage offer and any resulting discussion, is left intact. Other courts have reached similar results in analogous contexts.133 Accordingly, as the Chamber has not shown a likelihood of success on the merits of its constitutional challenge to this part of the Ordinance; the district court correctly refused to enjoin enforcement of the Reliance Provision. 130 Expressions Hair Design v. Schneiderman, 137 S. Ct. 1144, 1151 (2017) (quoting Rumsfeld v. Forum for Acad. & Inst’l Rights, Inc., 547 U.S. 47, 62 (2006)). During oral argument, the City Solicitor for the City of Philadelphia offered a very good analogy: An anti-discrimination Ordinance that prohibits hiring discrimination based on race does not implicate speech even though it may cause an establishment to remove a “Colored Applicants Only” sign. 131 138 S. Ct. 2361 (2018). 132 709 F.3d at 832; 868 F.3d at 113. 133 See, e.g., International Franchise Association, Inc. v. City of Seattle, 803 F.3d 389, 408 (9th Cir. 2015) (explaining that the minimum wage law at issue there was an “economic regulation that does not target speech or expressive conduct”). 31 B. The Inquiry Provision As discussed above, the Inquiry Provision of the Ordinance prohibits “ask[ing] a job applicant in writing or otherwise . . . about [the applicant’s] wage history, requir[ing] disclosure of wage history, or condition[ing] employment or consideration for an interview or employment on disclosure of wage history[.]”134 Unlike the Reliance Provision, the Inquiry Provision clearly regulates speech because it prevents employers from asking potential applicants specific questions. The district court was therefore correct in concluding that it was first necessary to determine the appropriate level of scrutiny to apply to that provision. 1. The Legal Standard The City argues that the speech at issue is commercial speech and therefore intermediate scrutiny under the test outlined in Central Hudson is appropriate. The Chamber argues that even if the speech at issue is commercial speech, we should apply strict scrutiny because the Inquiry Provision restricts expression based on content and speaker. We agree with the district court that the Inquiry Provision regulates commercial speech and that intermediate scrutiny under Central Hudson is the appropriate level of review. a. Commercial Speech The Supreme Court has described commercial speech as “expression related solely to the economic interests of the speaker and its audience.”135 A “proposal of possible employment . . . [is a] classic example[] of commercial speech.”136 Additionally, courts have recognized commercial 134 Phila. Code. § 9-1131. 135 Central Hudson, 447 U.S. at 561. 136 Pittsburgh Press Co. v. Pittsburgh Comm’n on Human Relations, 413 U.S. 376, 385 (1973) (“Each is no more than a proposal of possible employment. The advertisements are thus classic examples of commercial speech.”); see also Bigelow v. Virginia, 421 U.S. 809, 821 (1975) (finding the speech at issue “classic examples of commercial speech, for 32 speech in a range of employment-related contexts, including communications that advertise labor availability and terms of employment,137 as well as agreements “under which services will be exchanged for compensation.”138 We have recognized three factors that aid the inquiry into whether speech is commercial: “(1) is the speech an advertisement; (2) does the speech refer to a specific product or service; and (3) does the speaker have an economic motivation for the speech[?] . . . An affirmative answer to all three questions provides ‘strong support’ for the conclusion that the speech is commercial.”139 However, all three characteristics need not be present for a given expression to qualify as commercial speech.140 Expression pertaining to a possible offer of employment involves (1) an advertisement by the prospective employee to the employer; (2) the focus of the employee’s services for hire; and (3) by definition, an economic motive. The district court appreciated that the Inquiry Provision pertains only to communications between an employer and prospective employee and implicates no interests beyond the contract of employment. Because the speech occurs in the context of employment negotiations, the economic motive is clear. The regulated speech is part of a “proposal of possible employment.” Thus, the district court correctly concluded: [T]he Inquiry Provision prohibits Philadelphia- based employers from asking potential hires about their previous wage history. This inquiry occurs in the context of a job application or job interview, both of which propose a commercial each was no more than a proposal of possible employment”) (internal quotations omitted). 137 Valle Del Sol Inc, 709 F.3d at 818–19. 138 Nomi v. Regents for Univ. of Minn., 796 F. Supp. 412, 417 (D. Minn. 1992) vacated on other grounds, 5 F.3d 332 (8th Cir. 1993). 139 U.S. Healthcare, Inc. v. Blue Cross of Greater Phila., 898 F.2d 914, 933 (3d Cir. 1990) (citing Bolger v. Youngs Drug Prods. Corp., 463 U.S. 60, 66–67 (1983)). 140 Id. 33 transaction . . . [where] “all affected speech is either speech soliciting a commercial transaction or speech necessary to the consummation of a commercial transaction.”141 b. Intermediate Scrutiny under Central Hudson Is Appropriate “The Central Hudson analysis is commonly referred to as ‘intermediate scrutiny.’”142 Because commercial speech is “linked inextricably with the commercial arrangement it proposes, . . . the State’s interest in regulating the underlying transaction may give it a concomitant interest in the expression itself.”143 “The Constitution therefore accords a lesser protection to commercial speech than to other constitutionally guaranteed expression.”144 In Central Hudson, the Public Service Commission of New York City had attempted to address a fuel shortage in New York by promulgating an ordinance banning electricity-supply utilities from placing advertisements that promoted the use of electricity.145 A utility company challenged the ordinance arguing that it infringed on the company’s free speech rights because the ordinance banned speech based on the specific content of the speech and the identity of the speaker. In resolving the First Amendment issue, the Supreme Court “articulated a test for determining whether a particular commercial speech regulation is constitutionally 146 permissible[.]” Courts must determine whether: (1) the speech concerns lawful activity and is not misleading; (2) the asserted governmental interest is substantial; (3) the regulation 141 Chamber of Commerce, 319 F. Supp. 3d at 783 (citing Valle Del Sol, 709 F.3d at 818). 142 Retail Digital Network, LLC v. Prieto, 861 F.3d 839, 844 (9th Cir. 2017) (citing Florida Bar, 515 U.S. at 623). 143 Chamber of Commerce, 319 F. Supp. 3d at 784 (citing Edenfield v. Fane, 507 U.S. 761, 767 (1993)). 144 Central Hudson, 447 U.S. at 552–53. 145 Id. at 559. 146 Thompson v. W. States Med. Ctr., 535 U.S. 357, 367 (2002). 34 directly advances the governmental interest asserted; and (4) “whether it is not more extensive than is necessary to serve that interest.”147 As elaborated on below, under this test, the “fit” between the proposed restriction and the government’s interest need not be the least restrictive means. It need only be a “reasonable fit between the legislature’s ends and the means chosen to accomplish those ends.”148 c. Strict Scrutiny Is Inappropriate Here The Chamber argues that because the Ordinance only applies to employers and is focused squarely on content (wage history), strict scrutiny should have been applied.149 But as we described above, the Supreme Court has consistently applied intermediate scrutiny to commercial speech restrictions, even those that were content- and speaker-based, particularly when the challenged speech involves an offer of employment.150 We realize, of course, that it may be appropriate to apply strict scrutiny to a restriction on commercial speech that is viewpoint-based.151 If the regulation has the practical effect of 147 Central Hudson, 447 U.S. at 566. 148 Lorillard Tobacco Co. v. Reilly, 533 U.S. 525, 528 (2001). 149 Chamber Br. at 24. The Ordinance’s speech restrictions, the Chamber argues, are content-based due to their “appli[cation] to particular speech because of the topic discussed—namely, wage history.” Id. 150 See, e.g., Greater New Orleans Broad. Ass’n, Inc. v. United States, 527 U.S. 173, 176, 183–84 (1999) (applying intermediate scrutiny to prohibition on broadcast advertising of legal casino gambling); Rubin v. Coors Brewing Co., 514 U.S. 476, 478, 482, 488 (1995) (applying intermediate scrutiny to law prohibiting display of alcohol content on beer labels); Florida Bar, 515 U.S. at 620, 635 (1995) (applying intermediate scrutiny to prohibition on attorneys sending written solicitations to prospective clients relating to an accident or disaster). 151 See Turner Broadcasting System, Inc. v. F.C.C. (“Turner I”), 512 U.S. 622, 658 (“Congress may not abridge the rights of some persons to engage in political expression in order to 35 promoting some messages or some speakers based on the content of the speech or the identity of the speaker, something more than intermediate scrutiny may be necessary to survive a First Amendment inquiry. “[S]peaker-based laws demand strict scrutiny when they reflect the Government’s preference for the substance of what the favored speakers have to say (or aversion to what the disfavored speakers have to say).”152 The Supreme Court addressed this in R.A.V. v. City of St. Paul, Minnesota.153 It explained that the rule that content-based speech restrictions are subject to strict scrutiny is “not absolute” and is inapplicable when the restriction does not “‘raise[] the specter that the Government may effectively drive certain ideas or viewpoints from the marketplace.’”154 Here, the Inquiry Provision precludes all employers from inquiring into wage history, without focusing on any particular viewpoint or favoring any particular employer or job. It also applies to all employees without regard to the employee’s prior salary or job title. It does limit the prospective employer’s speech, but only because that limitation prevents the tentacles of any past wage discrimination from attaching to an employee’s subsequent salary. This simply does not implicate the kind of viewpoint or speaker discrimination that the Chamber relies on in its attempt to distinguish Central Hudson and have us apply strict scrutiny. The Chamber points to Sorrell v. IMS Health, Inc,155 in support but Sorrell is unhelpful because the restriction there was viewpoint-based and “heightened scrutiny” was therefore necessary. In Sorrell, Vermont had passed a law restricting the sale, disclosure, and use of pharmacy records that revealed the prescribing practices of individual doctors.156 However, the enhance the relative voice of other segments of our society.”) (internal quotations omitted). 152 Id. 153 505 U.S. 377 (1992). 154 Id. at 387–88 (quoting Simon & Schuster, Inc. v. Members of N.Y. State Crime Victims Bd., 502 U.S. 105, 115 (1991)). 155 564 U.S. 552 (2011). 156 Id. at 557. 36 law contained exceptions that, for example, allowed entities engaging in “educational communications” to purchase the information, but barred disclosure when the recipients would use the information for marketing.157 Additionally, “Vermont could supply academic organizations with prescriber- identifying information to use in countering the messages of brand-name pharmaceutical manufacturers and in promoting the prescription of generic drugs,” but the law prevented pharmaceutical manufacturers from using the information for their own marketing purposes.158 Thus, the statute “disfavor[ed] marketing, i.e., speech with a particular content, as well as particular speakers, i.e., entities engaged in marketing on behalf of pharmaceutical manufacturers.”159 Strict scrutiny was therefore required. Moreover, even though the statute there was neither viewpoint neutral nor speaker neutral, it is not even clear that the Court applied strict scrutiny there. As the district court astutely recognized here, Sorrell merely stands for the proposition that some level of scrutiny above rational basis review applied. The district court explained: “Sorrell references a ‘heightened scrutiny,’ but it is just as likely that this is the same as intermediate scrutiny, which is stricter than rational basis scrutiny.”160 Moreover, after Sorrell, courts have continued to apply Central Hudson intermediate scrutiny to commercial speech restrictions and rejected the notion that Sorrell requires strict scrutiny in these cases just as the district court explained.161 That said, we need not resolve that issue here 157 Id. at 564. 158 Id. 159 Id. at 552. 160 Chamber of Commerce, 319 F. Supp. 3d at 784; see also Prieto, 861 F.3d at 847 (“There is nothing novel in Sorrell’s use of the term ‘heightened scrutiny’ to distinguish from rational basis review.”). 161 See, e.g., Prieto, 861 F.3d at 848–49 (rejecting notion that Sorrell’s reference to ‘heightened” scrutiny was intended to apply a standard to commercial speech cases that is greater than intermediate scrutiny); 1-800-411-Pain Referral Serv., LLC v. Otto, 744 F.3d 1045, 1055 (8th Cir. 2014) (“The upshot is that when a court determines commercial speech 37 because it is clear that the restrictions in the Ordinance are viewpoint neutral and do not merit strict scrutiny. Accordingly, we agree with the district court’s decision to subject the Ordinance only to intermediate scrutiny under Central Hudson. 2. The Inquiry Provision Satisfies Central Hudson Intermediate Scrutiny Under Central Hudson, speech “at least must concern lawful activity and not be misleading[]” to qualify for protection.162 If the speech concerns illegal activity or is misleading, then it is not subject to First Amendment protection at all and our inquiry ends.163 If the subject is not unlawful and the message not misleading, we must then determine whether the government has a substantial interest in the restriction. If it does, the challenged restriction must directly advance that interest.164 If it does directly advance the interest, the final prong of the Central Hudson inquiry requires us to decide if the restriction is nevertheless more extensive than necessary to serve the government’s substantial interest.165 The last two elements of the analysis are related because they “basically involve a consideration of the ‘fit’ between the legislature’s ends and the means chosen to accomplish those ends.”166 Determining whether the restriction is more extensive than restrictions are content- or speaker-based, it should then assess their constitutionality under Central Hudson.”). 162 Central Hudson, 447 U.S. at 566. 163 Id. 164 Id. 165 Id. 166 Posadas e Puerto Rico Assoc. v. Tourism Co. of Puerto Rico, 478 U.S. 328, 341 (1986). 38 necessary, is not to be confused with the “least restrictive alternative” inquiry required to survive strict scrutiny.167 a. The Speech at Issue Is Not “Related to Illegal Activity” The City has argued that inquiring about wage history is “related to illegal activity” because the Inquiry Provision prohibits acquiring information that cannot be legally used because of the restrictions in the Reliance Provision. In rejecting that argument, the district court explained that not all uses of wage history are illegal: “For example, acquisition of wage history is allowed in other contexts such as for gathering market information;” and, “the existence of a wage history is not in and of itself illegal.”168 The district court correctly concluded: “[s]imply because wage history could be relied upon in fashioning a salary in violation of the Reliance Provision does not render all other legal activity related to wage history illegal.”169 Accordingly, the Court held that the 167 See Lorillard Tobacco, 533 U.S. at 556 (“[I]t [is] clear that ‘the least restrictive means’ is not the standard; instead, the case law requires a reasonable ‘fit between the legislature’s ends and the means chosen to accomplish those ends, . . . a means narrowly tailored to achieve the desired objective’”). Under strict scrutiny the government faces a more difficult burden, it “must show that the ‘regulation is necessary to serve a compelling state interest,’” Burson v. Freeman, 504 U.S. 191, 198 (1992), and the regulation must be the least restrictive means of achieving the interest. McCullen v. Coakley, 573 U.S. 464, 478 (2014). 168 Chamber, 319 F. Supp. 3d at 786. 169 Id. (emphasis added); see also Dunagin v. City of Oxford, 718 F.2d 738, 743 (5th Cir. 1983) (en banc) (“The commercial speech doctrine would disappear if its protection ceased whenever the advertised product might be used illegally.”). 39 language the provision targets does not “concern unlawful activity.”170 We agree. The City relies in part upon Pittsburgh Press v. Human Relations Commission,171 in arguing that speech can be “related to unlawful activity” if only some of its uses are prohibited. In Pittsburgh Press, one of the provisions in a Pittsburgh Ordinance prohibited discrimination in employment and another prohibited “any notice or advertisement relating to ‘employment’ or membership which indicates any discrimination because of . . . sex.”172 The Pittsburgh Commission on Human Relations was in charge of implementing the Ordinance. The Commission concluded that Pittsburgh Press had violated the Ordinance through its practice of placing “help-wanted” advertisements in sex- specific columns (i.e., “Male Help Wanted,” “Female Help Wanted”). The final Commission Order, however, did not prohibit all sex-specific advertisements; it exempted certain jobs such as: “employment in domestic service,” and “jobs for which the Commission ha[d] certified a bona fide occupational exception,” and allowed exempted entities to advertise in a sex-specific manner.173 Pittsburgh Press sued, arguing that the Commission’s Order violated the First Amendment by restricting its editorial choices. The Supreme Court agreed, concluding that “[t]he advertisements, as embroidered by their placement, signaled that the advertisers were likely to show an illegal sex preference in their hiring decisions.”174 Accordingly, the Court 170 Chamber, 319 F. Supp. 3d at 787. Under the district court’s reasoning, on the other hand, a law that prohibited the advertising of the sale of cocaine, for example, would present a speech restriction that always and only related to illegal activity because there are no other legal uses/purposes behind the sale of cocaine. 171 413 U.S. 376 (1973). 172 Id. at 378. The Ordinance also prohibited “aid[ing] . . . in the doing of any act declared to be an unlawful employment practice under the Ordinance.” 173 Id. at 380. 174 Id. at 389. 40 found that “any First Amendment interest [that] might be served by [the advertisements] . . . [wa]s altogether absent when the commercial activity itself [wa]s illegal.”175 The City argues Pittsburgh Press is analogous because even though there were legal uses for sex-specific advertisements— i.e., the specific exemptions recognized by the Commission— the Court still concluded that sex-specific advertising was related to illegal activity and was therefore not protected by the First Amendment. Similarly, here, the City would have us decide that even though every inquiry into a prospective applicant’s wage history would not necessarily lead to a violation of law, reliance on that history would be illegal. Thus, the City urges us to hold that the Ordinance “concerns unlawful activity.” We, however, agree with the district court’s conclusion that commercial speech should not lose the protection of the First Amendment simply because a legislature has prohibited one of many uses of the regulated speech.176 As the district court reasoned, and as the Chamber argues, if the City’s position is upheld, a city could perform an easy end-run around First Amendment scrutiny by passing a speech restriction in conjunction with a law that made one use of the regulated speech illegal. The result would be that the prohibited speech would always “relate to unlawful activity” and therefore fail the first prong of the Central Hudson analysis. b. The City has a Substantial Interest in Closing the Wage Gap The Chamber does not dispute the district court’s conclusion that remedying wage discrimination and promoting wage equity is a substantial government interest, and we agree. Accordingly, we need not discuss the second prong of the Central Hudson inquiry. 175 Id. 176 Chamber of Commerce, 319 F. Supp. 3d at 787. 41 c. The Inquiry Provision Directly Advances the City’s Interest in Pay Equity The third prong of Central Hudson requires us to determine whether the Inquiry Provision directly “advances the Government’s interest in a direct and material way.”177 To survive that inquiry, the City must show that the “the harms it recites are real and that its restriction will in fact alleviate each of them to a material degree.”178 “[S]peculation or conjecture” cannot satisfy this burden.179 A court’s inquiry under this prong “is not a license to reweigh the evidence de novo, or to replace [legislators’] factual predictions with our own.”180 Rather, a court’s task is merely to determine whether the legislature has “drawn reasonable inferences based on substantial evidence.”181 This is the heart of the current dispute. The district court did not believe that the City produced sufficient evidence to establish that the Inquiry Provision would advance its substantial interest in mitigating the racial and gender-based pay gap. The court’s skepticism is summed up in the following passage from its opinion: While the conclusion that a discriminatory wage gap could be affected by prohibiting wage history inquiries was characterized by respected professionals as a logical, common sense outcome, more is needed. Like the Rubin case, the testimony in support of this theory is riddled with conclusory statements, amounting to “various tidbits” and “educated guesses.” Importantly, aside from Dr. Madden’s affidavit, the information relied upon by the City does not address the possibility that disparate wages could also be based on factors having nothing to 177 Florida Bar, 515 U.S. at 625 (internal citations omitted). 178 Id.at 626. 179 Id. 180 Turner I, 512 U.S. at 666. 181 Turner II, 520 U.S. at 195 (internal quotation marks and citation omitted). 42 do with discrimination, such as qualifications, experience, or any number of other factors.182 We disagree. It is clear to us that Dr. Madden’s affidavit would, by itself, satisfy the inquiry. However, that is not the point. Dr. Madden’s affidavit simply corroborated the testimony given to the City Council prior to it enacting the Ordinance with additional empirical evidence. The issue is the apparent failure by the district court to afford the testimony and studies presented to the City Council sufficient probative value given its equation of it with conclusory statements and educated guesses. The Supreme Court has “permitted litigants to justify speech restrictions by reference to studies and anecdotes pertaining to different locales altogether, or even, in a case applying strict scrutiny, to justify restrictions based solely on history, consensus, and ‘simple common sense.’”183 And it has often done so on records far less compelling than the record supporting the Inquiry Provision of this Ordinance. The Court has explained that “the quantum of empirical evidence [required]. . . var[ies] up or down with the novelty and plausibility of the justification raised.”184 And, especially relevant here, it has recognized that “[a] municipality considering an innovative solution may not have data that could demonstrate the efficacy of its proposal because the 182 Chamber of Commerce, 319 F. Supp. 3d at 797–98 (emphasis added) (internal quotation marks omitted). 183 Florida Bar, 515 U.S. at 628 (citing Burson v. Freeman, 504 U.S. 191, 211 (1992)) (Blackmun, J., plurality opinion)) (emphases added). See also City Br. at 43–44 (citing Heffner v. Murphy, 745 F.3d 56, 92 (3d Cir. 2014), WV Ass’n of Club Owners & Fraternal Servs., Inc. v. Musgrave, 553 F.3d 292, 303–04 (4th Cir. 2009), Coyote Pub., Inc. v. Miller, 598 F.3d 592, 608 (9th Cir. 2010) to show recent decisions reflecting the Supreme Court’s flexible approach to speech restrictions under intermediate scrutiny). 184 Nixon v. Shrink Mo. Gov’t PAC, 528 U.S. 377, 391 (2000). 43 solution would, by definition, not have been implemented previously.”185 This record contains a plethora of evidence that (1) the wage gap is substantial and real (indeed, the parties concede this point); (2) numerous experiments have been conducted, which controlled for such variables as education, work experience, academic achievement, etc. and still found a wage gap; (3) researchers over many years have attributed the gap, in substantial part, to discrimination; (4) existing civil rights laws have been inadequate to close the wage gap; and, critically, (5) witnesses who reviewed the data concluded that relying on wage history can perpetuate gender and race discrimination. Based on that substantial evidence, the City Council made a reasonable judgment that a wage history ban would further the City’s goal of closing the gap and ameliorating the discrimination inherent in the disparate wages. The district court believed that the evidence before the City didn’t account for variables other than gender and race. However, Barbara Price presented the Council with evidence to the contrary, and the studies of Blau and Khan summarized in the Madden affidavit isolated out the variables of gender and race, thereby ensuring they did not affect the results.186 This evidence showed that even after accounting for such variables as choice of occupation, hours worked, economic sector, experience, GPA, undergraduate institution, and marital status, there is a significant gap between the earnings of men and women beginning one year after graduation and widening in the years thereafter.187 The City merely “dr[ew] reasonable inferences based on substantial evidence[]’”188 that the Inquiry Provision would address the wage gap, and the district court erred when it “reweigh[ed] the evidence” and “replace[d] [the City’s] factual predictions with [its] own.”189 185 City of Los Angeles v. Alameda Books, Inc., 535 U.S. 425, 439–40 (2002). 186 Blau and Khan at 32; see also fn. 60, 75, supra. 187 See JA275. 188 Turner II, 520 U.S. at 181 (internal quotation marks and citation omitted). 189 Turner I, 512 U.S. at 666. 44 i. Caselaw Considering Whether a Legislature Relied on Substantial Evidence to Support a Speech Restriction Under Central Hudson Demonstrates that the City Presented Sufficient Evidence to Support the Ordinance. Our review of caselaw examining whether a legislature had sufficient evidence to support a challenged legislative enactment demonstrates that the Inquiry Provision is constitutional. In Burson v. Freeman,190 Central Hudson itself, and Tennessee Secondary School Athletic Association v. Brentwood Academy,191 the Supreme Court upheld laws restricting commercial speech even though they were supported by much less evidence than the City produced to demonstrate the need for the Inquiry Provision. In Burson, the Court considered whether a 100-foot bubble zone that prohibited political speech outside of polling places was constitutional.192 The Tennessee statute at issue implicated three fundamental First Amendment concerns because it regulated political speech, speech in a public forum, and the content of speech.193 The Court subjected the ordinance to strict scrutiny but still upheld it.194 We realize that the Court in Burson relied upon a “modified ‘burden of proof’” because the First Amendment right at issue there “threaten[ed] to interfere with the act of voting itself.”195 Nevertheless, the analysis in Burson provides helpful guidance in determining whether the City’s evidence was sufficient to survive the third prong of the Central Hudson inquiry. 190 504 U.S. 191 (1992). 191 551 U.S. 291 (2007). 192 Burson, 504 U.S. at 211. 193 Id. at 196. 194 Id. at 197, 211. 195 Id. at 208 n.11. 45 The Burson court explained that it “never has held a State to the burden of demonstrating empirically the objective effects” of a speech regulation.196 Accordingly, the Court relied on history, common sense, and one witness, noting that it would be “difficult for the states to put on witnesses who [could] testify as to . . . the exact effect” of the proposed law.197 In fact, rather than demand strict empirical evidence that the challenged restriction on speech advanced the underlying governmental interest, the Court’s analysis rested on the presumed logic of a 100-foot barrier around a polling place for the purpose of allowing voters fifteen seconds of uninterrupted contemplation before casting their ballots.198 There was no empirical evidence that voters needed fifteen seconds of uninterrupted contemplation to cast an informed ballot, nor was there any evidence that voters would use the fifteen seconds it took to traverse the 100-foot buffer zone for contemplation, as opposed to conversation, daydreaming, or reading a newspaper. The Supreme Court’s decision in Tennessee Secondary, is also informative. There, the Court accepted commonsense conclusions in the absence of empirical data in considering whether the enforcement of a rule governing interscholastic sports violated the First Amendment.199 The rule under review prohibited high school coaches from using “undue influence” when recruiting middle school students for athletic programs.200 Much like the record here, the evidence before the Court consisted of testimonial and documentary evidence, including letters sent by a school football coach to a group of unenrolled eighth-grade boys inviting them to participate in spring practice sessions.201 In upholding the sanction imposed on the coach’s speech, the Court noted that it “need[ed] no empirical data to credit [the agency’s] commonsense conclusion” that the speech at issue—an inquiry by a would- 196 Id. at 208 (quoting Munro v. Socialist Workers Party, 479 U.S. 189, 195 (1986)) (internal quotations omitted). 197 Id. at 208-211. 198 Id. 199 551 U.S. at 294. 200 Id. 201 Id. at 294-95. 46 be authority figure of a prospective team member—could exert the type of undue influence prohibited by the rule.202 Finally, in Central Hudson, the Court held that a prohibition on advertising by utilities was supported by substantial evidence.203 Rather than require strict empirical proof, the Court relied on the commonsense conclusion that “[t]here is an immediate connection between advertising and demand for electricity.”204 As we have summarized, the City did offer substantial evidence in the form of testimony and metanalysis of relevant research to support the need for the Inquiry Provision. Reasonable minds can debate whether the City’s evidence placed the need for, and potential effectiveness of, the Inquiry Provision beyond doubt. However, given the discussion in Burson, Tennessee Secondary, and Central Hudson, certainty of proof or empirical data is not required here. Rather, substantial evidence of the possibility that the speech restriction could favorably impact a concern that the state actor had a fundamental interest in addressing is sufficient. The City easily satisfied that standard. In concluding otherwise, the district court imposed too high a burden on the City. As noted earlier, all parties agree that there is a longstanding disparity in the pay of women and minorities compared to wages of White males. The district court readily accepted the existence of this pay gap.205 Moreover, the Chambers’ CEO 202 Id. at 300. 203 447 U.S. at 568. 204 Id., 447 U.S. at 569. Although the Court eventually found, under the fourth prong, that the law at issue was overbroad because it “suppresse[d] speech that in no way impairs the State’s interest in energy conservation,” id. at 570, under the third prong, the Court simply recognized the “immediate connection” that limiting advertising would have on demand for electricity. Id. at 569. 205 Chamber of Commerce, 319 F. Supp. 3d at 792 (“[P]ractically all of the . . . testimony amplifies a point that really is not in dispute – that there is a gender pay disparity.”) Although this excerpt from the district court opinion refers 47 stated that Chamber members relied on wage history “to have a better understanding of whether a candidate is worth pursuing based on previous compensation levels.”206 Nevertheless, the district court relied primarily on four cases in concluding that the City failed to meet its burden:207 Edenfield v. Fane,208 Rubin v. Coors Brewing Co.,209 Pitt News v. Pappert 210 and Wollschlaeger. However, the City’s proof here is much more robust than the records before those courts. In Edenfield and Rubin the restrictions on commercial speech were facially based on unsubstantiated fears supported by conclusory statements. In Edenfield, a Certified Public Accountant challenged a rule created by the Florida Board of Accountancy that prohibited CPAs from soliciting clients in- person.211 The Florida Board believed that in-person solicitation would lead to unethical conduct by CPAs.212 In striking down the restriction on commercial speech, the Court reasoned that the only evidence presented in support of the Florida Board’s position came from an affidavit by one of its former chairmen.213 He stated the solicitation ban was specifically only to the gender disparity, it is clear that the court also accepted the existence of a racial disparity. The court’s concern was not with the existence of these disparities, but with whether the City had established a sufficient “fit” between the Inquiry Provision and these disparities to support its conclusions that the Inquiry Provision was necessary to address the disparities. 206 JA124. 207 Chamber of Commerce, 319 F. Supp. at 794 (“Edenfield, Rubin, Pitt News, and Wollschlaeger instruct that some evidence is required for the legislature to conclude that the law at issue will directly advance the government’s substantial interest. Theories and unsupported opinions will not suffice to demonstrate that the asserted harms are real.”). 208 507 U.S. 761 (1993). 209 514 U.S. 476 (1995). 210 379 F.3d 96 (3d Cir. 2004). 210 Id. at 107–08. 211 Edenfield, 507 U.S. at 764. 212 Id. 213 Id. at 764, 771-72. 48 necessary to “prevent overreaching and vexatious conduct by the CPA.”214 His conclusion in his affidavit depended on the unsubstantiated theory that a CPA who solicits clients would be beholden to the client and thus willing to bend the rules.215 Consequently, the Court refused to credit his affidavit.216 In Rubin, the Federal Alcohol Administration Act prohibited beer labels from displaying alcohol content for fear of a “strength war” among brewers.217 The justification for the law was the purported “common-sense” conclusion that if the alcohol content were not advertised, customers would be less likely to buy the product based on the alcohol content.218 The Court found that the Act did not directly advance the stated purpose because the government’s regulatory scheme was “irrational.”219 Malt liquor, wine, and other alcohol sellers could and did label their bottles with the strength of the drink.220 The government, the Court noted, had relied on “anecdotal evidence and educated guesses” in contending that competition based on alcohol content was occurring and found that these “various tidbits” could not overcome the irrationality of the scheme.221 Thus, the very existence of a “strength war” was in doubt and no evidence was offered to establish that any such phenomena actually existed. Whereas, here, the wage gap that the Inquiry Provision seeks to address is a given, and the reasoned conclusions presented to the City Council were entitled to more credit than owed to the educated guesses before the Federal Alcohol Administration. The law we struck down in Pitt News, was similarly based solely upon “speculation and conjecture.” The law was premised on the assumption that prohibiting alcohol ads from appearing in university publications would “slacken the demand for alcohol by Pitt students” and help curb underage 214 Id. at 765. 215 Id. 216 Id. at 775-76. 217 514 U.S. at 478–79. 218 Id. at 487 219 Id. at 488. 220 Id. at 486–89. 221 Id. at 490. 49 drinking.222 We found that the legislature’s conclusion was “counterintuitive and unsupported by any evidence.”223 There was no evidence, for example, that the removal of the ads would make it harder to find places near campus to buy alcohol. Furthermore, not only were students able to see alcohol ads in many other publications and on television, more that 75% of the university population was of the legal drinking age.224 Finally, Wollschlaeger is similarly unpersuasive because of the tenuous reasoning supporting the restriction on commercial speech. There, the Court of Appeals for the Eleventh Circuit struck down a law that had been enacted based solely on a few anecdotes.225 Certain Florida laws prevented doctors from asking patients “whether they own firearms or have firearms in their homes, or from recording answers to such questions.”226 The legislature asserted that the law was necessary to protect gun-owning Floridians from the “private encumbrances” on their Second Amendment Rights that allegedly came from being subject to such questions by physicians.227 The legislature had relied on “six anecdotes and nothing more” to justify enacting the restrictions.228 In striking down the legislation, the court observed that while anecdotes can provide evidence, there was “no other evidence, empirical or otherwise” presented by the legislature, and the six anecdotes could not show that the harms were “real, [and] not merely conjectural,” such that the regulations “will in fact alleviate [the] harms in a direct and material way.”229 Thus the Wollschlaeger court required something more than anecdotal evidence and less than empirical evidence if the restriction was to survive the third prong of the Central Hudson inquiry. 222 379 F.3d at 107. 223 Id. (emphasis added). 224 Id. at 108. 225 848 F.3d at 1319. 226 Id. at 1303. 227 Id. at 1312. 228 Id. (emphasis added). 229 Id. (quoting Turner II, 512 U.S. 622 at 664). 50 The City’s proof of the nexus between its substantial interest in eliminating the real phenomenon of a racial and gender- based wage gap and the need for the limitations that are at the heart of the Inquiry Provision is in a different category than the cases we have just discussed. There is testimony here that the gender disparity in pay in Pennsylvania has existed for the past five decades despite the passage of laws over that period to remedy such discrimination.230 Terry Fromson explained how this wage gap is compounded through institutional discrimination and explained how other states have addressed this issue.231 Marianne Bellesorte researched the wage gap for women and men of color, and explained how the inequities began right out of college and continued to affect women, in particular, until retirement. Finally, Jovida Hill and Rue Landau provided empirical evidence that substantiated the distilled conclusions of Fromson and Bellesorte.232 This testimony is much more than “conclusory statements, . . . and ‘educated guesses[.]’”233 Moreover, Dr. Madden’s affidavit amplified this testimony by viewing it through the empirical lens of thousands of studies she summarized.234 There is therefore ample evidence to establish the fit between the Inquiry Provision and the societal evil it was intended to address. Our conclusion that the district court imposed too high a burden on the City’s proof is consistent with the en banc opinion of the Court of Appeals for the Ninth Circuit in Rizo v. Yovino.235 There, the en banc court held that an employer’s reliance on the plaintiff’s prior salaries to justify paying a female less than her male cohort’s salary was a violation of the 230 H’rg Tr. at 66. 231 Id. at 75. 232 Id. at 8-12. 233 Chamber of Commerce, 319 F. Supp. At 798. 234 JA297. 235 887 F.3d 453, 460–61 (9th Cir. 2018) (en banc) (holding that a female employee’s prior salary does not qualify as a “factor other than sex” under the federal Equal Pay Act that can justify paying her less than a male employee who performs substantially equal work), vacated on other grounds by Yovino v. Rizo, 139 S. Ct. 706 (2019). 51 Equal Pay Act.236 The court’s explanation was straightforward. “The question before us is . . . simple: can an employer justify a wage differential between male and female employees by relying on prior salary? . . . [T]he answer is clear: No.”237 There, the employer had argued that the plaintiff’s disparate salary was not barred by The Equal Pay Act because, in paying her a wage based on her prior salaries, the differential was based on a factor other than sex which is explicitly allowed under the Equal Pay Act.238 The court held that that consideration of salary history “allow[s] employers to capitalize on the persistence of the wage gap and perpetuate that gap ad infinitum.”239 Other courts have reached the same conclusion.240 Notwithstanding our recitation of the impressive record that supports this Ordinance, we think it important to emphasize that neither scores of empirical studies nor proof to scientific certainty is necessary to carry the City’s burden here. Even though we find the City’s evidence here more than sufficient to carry its burden under the third prong of Central Hudson, it is important not to lose sight of the fact that where a legislature presents an “innovative solution,” the Supreme Court has recognized that it “may not have data that could [conclusively] demonstrate the efficacy of its proposal because the solution would, by definition, not have been implemented 236 Id. at 456 (citing 29 U.S.C. § 206(d)(1)). 237 Id. 238 29 U.S.C. § 206(d)(1)(iv) allows “a differential based on any factor other than sex.” 239 Rizo, 887 F.3d at 456–57 (emphasis added). 240 See, e.g., Irby v. Bittick, 44 F.3d 949, 955 (11th Cir. 1995) (“if prior salary alone were a justification, the exception would swallow up the rule and inequality in pay among genders would be perpetuated”); Riser v. QEP Energy, 776 F.3d 1191, 1199 (10th Cir. 2015) (Equal Pay Act “precludes an employer from relying solely upon a prior salary to justify pay disparity”) (citation omitted); but see, e.g., Wernsing v. Dep’t of Human Servs., 427 F.3d 466, 468-70 (7th Cir. 2005) (holding that prior salary alone can justify wage disparities). 52 previously.”241 Nevertheless, the City did produce such evidence here and clearly carried its burden. However, as we held in King v. Governor of the State of New Jersey,242 and as we recount in detail below, legislatures are not “constitutionally required to wait for conclusive scientific evidence before acting to protect [their] citizens from serious threats of harm.”243 In Alameda Books, the City of Los Angeles enacted legislation that prohibited “more than one adult entertainment business” from inhabiting “the same building, structure or portion thereof.”244 The Court of Appeals for the Ninth Circuit invalidated this restriction on speech finding that “the city failed to present evidence upon which it could reasonably rely to demonstrate that its regulation of multiple-use establishments [wa]s “designed to serve” the city’s substantial interest in reducing crime.”245 The Supreme Court disagreed. The Court concluded that the City had presented sufficient evidence upon which to base the speech restriction. Justice O’Connor, joined by the Chief Justice, Justice Scalia and Justice Thomas, explained that the respondents “ask[ed] the city to demonstrate, not merely by appeal to common sense, but also with empirical data, that its ordinance will successfully lower crime.”246 But they concluded that “[o]ur cases have never required that municipalities make such a showing, certainly not without actual and convincing evidence from plaintiffs to the contrary.”247 241 City of Los Angeles v. Alameda Books, Inc., 535 U.S. 425, 439–40 (2002). 242 767 F.3d 216 (3d Cir. 2014), abrogated on other grounds by Nat'l Inst. of Family & Life Advocates v. Becerra, 138 S. Ct. 2361 (2018). 243 767 F.3d at 239. 244 Alameda Books, 535 U.S. at 429. 245 Id. at 433. 246 Id. at 439. 247 Id. The court noted, “Respondents’ claim assumes that the . . . study proves that all adult businesses, whether or not they are located near other adult businesses, generate crime. This is a plausible reading of the results from the 1977 study, but 53 Here, as in Alameda Books, the Plaintiff has offered no proof to counter the City’s conclusion about the need for, or effectiveness of, the Inquiry Provision. In fact, as we have explained, some of the Plaintiff’s proof substantiates the City’s position. A lack of contrary evidence lightens the legislature’s burden.248 The substantial legislative record here is simply not analogous to the “irrational,” “conclusory,” “speculative,” and purely anecdotal evidence presented in Edenfield, Rubin, Pitt News and Wollschlaeger. Nonetheless, the Chamber argues, that even though “conclusive scientific evidence of the Ordinance’s effect is not required, ‘substantial evidence’ means ‘some concrete evidence is required.’”249 In support, the Chamber cites to “a 106-page summary of [a] 2-year study,”250 relied upon in Florida Bar and the “empirical judgments” of “a number of well-known, reputable professional and scientific organizations,” from our decision in King.251 respondents do not demonstrate that it is a compelled reading. Nor do they provide evidence that refutes the city’s interpretation of the study, under which the city’s prohibition should on balance reduce crime.” Id. at 438. Accordingly, the Court concluded that the City had supported the law with sufficient evidence. 248 See Nixon v. Shrink Missouri Government PAC, 528 U.S. 377, 394 (2000) (“[t]here might, of course, be need for a more extensive evidentiary documentation if respondents had made any showing of their own to cast doubt on the apparent implications of [the government’s] evidence and the record here”). 249 Chamber Br. at 49. 250 Id. 251 767 F.3d at 238. 54 ii. The Evidence Here is Stronger Than the Evidence Supporting the Restrictions in Florida Bar and King The Chamber makes much of Florida Bar, and the district court cited it as demonstrative of the type of “extensive” record necessary to sustain a speech infringement.252 There, the Florida Bar Association enacted rules banning direct-mail solicitation of clients in the 30 days following an accident or disaster. Members of the Florida bar sued, claiming that the law infringed their right of commercial speech. In rejecting that challenge, the Court relied upon the Bar Association’s citation to a 106-page study purporting to show the harm that the Bar was attempting to mitigate.253 However, a closer look at the study reveals that it contained information that was less relevant, less methodologically sound, and much less informative than the evidence supporting the Inquiry Provision here. The Majority of the Court described the study as “contain[ing] . . . statistical and anecdotal [data] . . . supporting the Bar’s contentions” that the direct-mail solicitations in the wake of accidents “reflects poorly on the profession.”254 The Court accepted that evidence as sufficiently probative even though much of the data in the surveys did not address the specific issues the restriction was supposed to address. The Court pointed to a subset of the findings from the study: it cited one survey of Florida adults that “indicated . . . Floridians ‘have negative feelings about those attorneys who use direct mail advertising.’”255 It also provided a handful of statistics about Floridians’ views of lawyer advertising.256 However, only one question referred to the reputation of the legal community–the 252 Chamber, 319 F. Supp. at 796, 800 (“Unlike in Florida Bar, there are no comprehensive studies demonstrating the alleged harm.”). 253 Florida Bar, 515 U.S. at 627. 254 Id. at 626. 255 Id. at 626-27. 256 Id. 55 harm that the law was apparently aiming to remedy.257 It appeared from the responses to that question that direct mail solicitation in general, rather than solicitation in the 30 days following an accident, was what lowered the views of the legal profession (and did so in only one quarter of those surveyed).258 The primary evidence relied upon by the Court, did not squarely address the harm that the rule was enacted to remedy. Despite the fact that the study gave “few indications of the sample size or selection procedures employed” and even though “no copies of the actual surveys employed,” were presented to the Court, the Court held that the Bar adequately supported the law.259 In dissent, Justice Kennedy noted that the record: (1) contained no explanation of methodology, sampling or framework; (2) dealt primarily with television and phone book advertising, which were not at issue; and (3) only two pages of the more than 100 focused on direct-mail solicitation.260 He concluded by saying that the “few pages of self-serving and unsupported statements by the State” should have been clearly insufficient to “demonstrate that a regulation directly and materially advances the elimination of a real 257 Id. 258 The report did include “excerpts from complaints of direct-mail recipients,” id. at 627, some of whom complained about solicitation in the wake of an injury or accident, but the Bar presented no evidence that a solicitation ban only in the first 30 days after an accident would do anything to mitigate these complaints. Additionally, the comments included favorable statements about direct mail solicitation as well. Id. at 641. 259 Id. at 640 (Kennedy, J. dissenting). 260 Id. at 640–41 (Kennedy, J. dissenting) (“[N]o actual surveys, few indications of sample size or selection procedures, no explanations of methodology, and no discussion of excluded results [were presented]. . . . [N]o description of the statistical universe or scientific framework that permits any productive use of the information [was presented].”). 56 harm.”261 Yet the Court upheld the statute on the basis of this evidence.262 Even if we view the supporting evidence in Florida Bar in the most favorable light possible, we still conclude that the City has made a stronger evidentiary showing here. The studies presented by the City address the specific issue that the Ordinance was enacted to remedy—discriminatory wage gaps. And, unlike the study before the Court in Florida Bar, the studies the City relied upon are peer-reviewed research studies, many of which were meta-studies that summarized the findings of hundreds of other such studies.263 The studies support the City’s conclusion that the wage gap is not attributable to “legitimate” factors such as education, experience or qualifications.264 Moreover, researchers’ conclusions that discrimination is the likely cause of the gaps has been present in the academic literature for decades.265 The conclusion that the wage gap is most likely the result of discrimination is also consistent with voluminous unrebutted independent evidence of workplace discrimination.266 261 Id. 262 Id. at 641. 263 See, e.g., Stanley & Jarrell (meta-analysis of more than 50 studies investigating the wage gaps). 264 See, e.g., id. at 948 (concluding there is a “wide consensus that gender wage discrimination exists” and the “vast empirical economic literature, containing hundreds of studies, reveals that women are ‘underpaid’ disproportionate to their observed skills”). 265 See, e.g., Blau & Kahn at 32 (finding from 1980 to 2010, “an unexplained gender wage gap in each year[’s data],” and explaining that the “finding of such an unexplained gap is fairly standard in the literature” and is “taken as an estimate of labor market discrimination”). 266 See e.g., Arin N. Reeves, “Written in Black and White: Exploring Confirmation Bias in Racialized Perceptions of Writing Skills,” Nextions Yellow Paper Series (2014) (concluding from the results of a controlled experiment on law firm partners reviewing an identical memo from African- American Thomas Meyer and Caucasian Thomas Meyer that the greater number of negative comments and a .9 reduction 57 The Chamber also cites our decision in King v. Governor of the State of New Jersey267 as another case in which the legislature presented substantial evidence to support a law. The Chamber argues, unconvincingly, that the showing in King was more robust than the City’s evidentiary showing here. In King, we upheld a New Jersey law prohibiting sexual orientation change efforts (“SOCE”) therapy to persons under the age of 18 over a challenge by individuals and organizations providing such counseling.268 in score on a scale of 5 for African-American Thomas Meyer was the result of “commonly held racially-based perceptions about writing ability . . . unconsciously impact [law firm partners’] ability to objectively evaluate a lawyer’s writing. Most of the perceptions uncovered in research thus far indicate that commonly held perceptions are biased against African Americans and in favor of Caucasians.”). The results of this controlled experiment are consistent with others like it conducted in various fields designed to ensure that the only variable that could explain the more positive reaction to White employees was the perceived race or gender of the person they were being compared to. In a similar, well publicized experiment published under the title: “Are Emily and Greg More Employable Than Lakisha and Jamal? A Field Experiment on Labor Market Discrimination,” researchers Marianne Bertrand and Sendhil Mullainathan of the University of Chicago and MIT, found “large racial differences in callback rates. Applicants with White names need[ed] to send about 10 resumes to get one callback whereas applicants with African-American names need[ed] to send about 15.” The fictional White applicant therefore had a 50 percent greater probability of getting a call back than the fictional African-American applicant. Marianne Bertrand & Sendhil Mullainathan, Are Emily and Greg More Employable Than Lakisha and Jamal? A Field Experiment on Labor Market Discrimination, NBER Working Paper No. 9873, National Bureau for Economic Research (2003), https://www.nber.org/papers/w9873.pdf. 267 767 F.3d at 216. 268 Id. at 221. 58 The legislative record there “demonstrate[d] that over the last few decades a number of well-known, reputable professional and scientific organizations ha[d] publicly condemned the practice of SOCE.”269 And we, in reviewing that record, specifically noted that the American Psychological Association, the American Psychiatric Association, and the Pan American Health Organization “have warned of the ‘great’ or ‘serious’ health risks accompanying SOCE counseling, including depression, anxiety, self-destructive behavior, and suicidality.”270 We also noted, “[m]any such organizations have also concluded that there is no credible evidence that SOCE counseling is effective.”271 We also found, that “the bulk of empirical evidence regarding the . . . harmfulness of SOCE counseling currently falls short of the demanding standard imposed by the scientific community.”272 We recognized there was a “limited amount of methodologically sound research” on the counseling and that “the few early research investigations . . . refus[ed] to make a definitive statement about whether SOCE is safe or harmful . . . due to a lack of scientifically rigorous studies.”273 Nevertheless, we concluded the legislature was not “constitutionally require[d] to wait for conclusive scientific evidence before acting to protect its citizens from serious threats of harm.”274 Instead, we were convinced by the legislature’s “highly plausible” judgment that SOCE could be harmful to minors, and concluded that the statute “directly advanced” New Jersey’s stated interest.275 Here, the district court concluded that many of the studies cited by the City did not conclusively prove that discrimination is the sole cause of the wage gap. That level of certainty is not required. The City made a well-reasoned judgment based on the testimony presented to it and the unrefuted existence of the wage gap that banning wage history inquiries would prevent 269 Id. at 238. 270 Id. 271 Id. 272 Id. at 239. 273 Id. (emphasis added) (citation omitted). 274 Id. 275 Id. at 239. 59 further perpetuation of gender and race discrimination in this context. Moreover, we won’t ignore the fact that the very nature of discrimination in employment is such that showing discrimination by negative inference is often necessary. As the Supreme Court has recognized in the context of gender discrimination in the workplace, “[a]s should be apparent, the entire purpose of the McDonnell Douglas prima facie case is to compensate for the fact that direct evidence of intentional discrimination is hard to come by.”276 We, too, have previously recognized, “the instances in which employers . . . openly [discriminate against] employees appear to be declining. Regrettably, however, this in no way suggests that discrimination based upon an individual’s race, gender, or age is near an end.”277 “It has become easier to coat various forms of discrimination with the appearance of propriety, or to ascribe some other less odious intention to what is in reality discriminatory behavior.”278 Accordingly, demonstrating discrimination by controlling for legitimate factors like education, training, experience, age, skills, and other factors that could otherwise “legitimately” explain wage gaps, and through experimental evidence, are essential means of showing discrimination. Because “direct evidence . . . is hard to come by,” negative inferences can be persuasive evidence of discrimination, especially where they are entirely unrebutted.279 276 Price Waterhouse v. Hopkins, 490 U.S. 228, 271 (1989) superseded by statute as stated in Burrage v. U.S., 571 U.S. 204, 214 n. 4 (2014) (permitting a showing that discrimination was a “motivating” or “substantial factor to shift the burden of persuasion to the employer, which was made moot after Congress amended the statute to remove but- for causality). 277 Aman v. Cort Furniture, 85 F.3d 1074, 1081 (3d Cir. 1996). 278 Id. at 1082 (emphasis added). 279 Price Waterhouse, 490 U.S. at 272. 60 As some of the studies on subliminal or implicit bias which we have discussed establish, bias is often not even something that that the actor is aware of.280 This makes it exceedingly difficult to address such issues as wage disparity because simply educating employers about the pay gap will not deter an employer who is not even aware of the fact that s/he is setting a discriminatory salary. Indeed, without challenging the existence of the pay gap, the CEO for the Chamber without pause admitted that Chamber members gain a “better understanding of whether a candidate is worth pursuing based on previous compensation.” Consequently, as the Court of Appeals for the Ninth circuit recognized in in Rizo v. Yovino, and as the Managing Attorney for the Women’s Law Project, Ms. Fromson, testified here, criteria that may at first appear to be race and gender neutral (such as wage history) may be proxies for race or gender. d. The Inquiry Provision is Not More Extensive Than Necessary “The last step of the Central Hudson analysis complements the third step, asking whether the speech restriction is not more extensive than necessary to serve the interests that support it.”281 However, “‘the least restrictive means’ is not the standard; instead, the case law requires a reasonable ‘fit between the legislature’s ends and the means chosen to accomplish those ends, . . . a means narrowly tailored to achieve the desired objective.’”282 The “scope” of the law must be “in proportion to the interest served.”283 The Court does not 280 For a thorough discussion of the prevalence and impact of such subliminal bias, see Mahzarin R. Banaji and Anthony G. Greenwald, Blind Spot: Hidden Biases of Good People (2013). 281 Lorillard Tobacco, 533 U.S. at 556 (internal quotations omitted). 282 Id. (citing Florida Bar, 515 U.S. at 632). 283 Board of Trustees of State Univ. of New York v. Fox, 492 U.S. 469, 479–80 (1989) (citing In re R.M.J., 455 U.S. 191, 203 (1982)); see also Matal v. Tam, 137 S. Ct. 1744, 1764 (2017) (“[t]he regulatory technique may extend only as far as 61 “impose upon [regulators] the burden of demonstrating that . . . the manner of restriction is absolutely the least severe that will achieve the desired end.”284 Instead, the legislation must provide “a fit that is not necessarily perfect, but reasonable;” one that “represents not necessarily the single best disposition” but a “proportion[ate]” one.285 Understandably, the district court here did not reach the third or fourth prong of the Central Hudson inquiry because it found that the Ordinance failed under prong two. However, the last two prongs “are not entirely discrete.”286 These two prongs “have been considered, somewhat in tandem, [as courts must] determine if there is a sufficient ‘fit between the [regulator’s] ends and the means chosen to accomplish those ends[.]’”287 Given our conclusion that the City has satisfied its burden of establishing the relationship between the legislative objective of mitigating the wage-gap and the remedy afforded by the Inquiry Provision, we will address the fourth prong. The Inquiry Provision is narrowly tailored. It only prohibits employers from inquiring about a single topic, while leaving employers free to ask a wide range of other questions, including qualifications, work history, skills and any other job- related questions relevant to performance or fit with the the interest it serves.”) (citing Central Hudson, 477 U.S. at 565). 284 Board of Trustees, 492 U.S. at 480. 285 Id. The Court does not invalidate a commercial speech restriction “that went only marginally beyond what would adequately have served the governmental interest,” rather “almost all of the restrictions disallowed under Central Hudson’s fourth prong have been substantially excessive . . . .” Id. at 479. 286 Greater New Orleans Broadcasting Ass’n, Inc. v. United States, 527 U.S. 173, 183 (1999). 287 Bad Frog Brewery, Inc. v. New York State Liquor Auth., 134 F.3d 87, 98 (2d Cir. 1998) (quoting Puerto Rico Assocs. 478 U.S. at 341). 62 company.288 Additionally, the provision does not prohibit employers from obtaining market salary information from other sources. The Ordinance simply seeks to insulate any discriminatory impact of prior salary levels on subsequent wages. The Ordinance is thus more narrowly tailored than similar wage history Ordinances that have been passed since 2017.289 As enacted, it simply prohibits employers from inquiring about wage history at a specific point in time—after a prospective employee has applied for a job and before s/he is hired and a wage is set—when the City has determined that the risk is greatest for conduct that perpetuates discrimination. Moreover, applicants can voluntarily provide salary history if they feel it is in their best interest.290 The Chamber argues that the Ordinance is not sufficiently tailored because it indisputably “regulate[s] speech that poses no danger to the asserted [governmental] interest.”291 According to the Chamber, the Ordinance does not achieve its interest “when it is applied to White male job applicants, whose salaries the City acknowledges are not tainted by past 288 We caution, however, that, as the discussion of the Court of Appeals for the Ninth Circuit in Rizo makes clear, some questions may raise the specter of a wage inquiry, even though not expressed in so many words. 289 See e.g., H’rg Tr., at 15–16 (“Actually, the Massachusetts law goes a little wider than we do. We’re trying to keep it real basic, and I think you’ll hear from a witness that thinks we don’t make it strong enough, but we’re trying to find that great balance that we always try to in legislation and at least at this point limit it to stopping the employer from asking, directly asking, the prospective employee what they make. Massachusetts law goes a little farther as far as how far the employer can inquire, and we’re not ready to go there yet and we think that could have added more controversy to the bill. . . . [W]e want to try to keep it real basic as far as the inquiry of past wages.”). 290 This, of course, does not suggest that an employer can goad or cajole an employee into disclosing prior wages or salary. 291 Central Hudson, 447 U.S. at 565. 63 discrimination.”292 At oral argument, the Chamber even went so far as to argue that the Ordinance should therefore not apply to White men. The suggestion was offered in all seriousness, and it shows the difficulty of, and very limited avenues for, addressing this persistent problem. Counsel for the Chamber actually suggested that the City set up a system in which employers are free to ask salary histories of White male job applicants but are precluded from doing the same for women and minorities. Aside from the clear equal protection implications, the suggestion for such a carve-out fails to understand the nature of the wage gap. As amici point out, a system that perpetuates higher salaries for men based on their higher salary histories is no better than one that perpetuates lower salaries for women and minorities based on their lower salary histories.293 Indeed, it is the very same system. Asking White men their prior salary and allowing it to impact an offer of employment would ensure that the historic salary advantage enjoyed by White males would continue. Employers operating under such a scheme would unwittingly be helping White males to continue to enjoy salary advantages on new jobs because they would be carried over from their prior jobs. More importantly, even were we to credit the Chamber’s suggestion, we would nevertheless not be free to ignore the Supreme Court’s decision in Florida Bar where the Court considered and rejected a similar overbreadth argument. The Respondents in Florida Bar argued that the ban on communications to all accident victims within 30 days of an accident was overbroad because it did not distinguish between those whom the provision was aiming to protect—injury victims who were especially vulnerable—and “those accident victims who are ready, willing and able to utilize a lawyer’s advice.”294 Rather than require the Bar Association to “draw[] difficult lines,” the Court concluded that the blanket “ban applicable to all post-accident or disaster solicitations for a brief 30–day period” was sufficiently narrowly tailored.295 292 Chamber Br. at 55. 293 See Br. of Amicus City of NY et al. 294 515 U.S. at 632. 295 Id. at 633. 64 Thus, even if we were to credit the Chamber’s argument that the law is overbroad, it would not prevent the Inquiry Provision from surviving intermediate scrutiny. The Supreme Court has refused to invalidate restrictions on commercial speech “that [go] only marginally beyond what would adequately have served the governmental interest.”296 We have no trouble concluding that the City has demonstrated a “proportionate” fit between its substantial interest and its legislative attempt to advance that interest. The Chamber also argues that “underinclusiveness plagues the Ordinance. Despite the City’s assumption that the wage history of female and minority applicants is ‘tainted’ by past discrimination, the Ordinance permits employers to base a salary offer on wage-history information that an applicant voluntarily discloses.”297 However, underinclusiveness is only important to our inquiry if it “raises serious doubts about whether the government is in fact pursuing the interest it invokes, rather than disfavoring a particular speaker or viewpoint.”298 There is no suggestion of such insincerity here. Moreover, the alleged underinclusiveness is more of a strength than an infirmity. It allows a female or minority who may have historically been paid above the normal salary levels because of extraordinary qualifications to inform a potential employer of that salary history rather than remain silent and risk forfeiting the higher salary that s/he may well deserve. 296 Fox, 492 U.S. at 479. 297 Chamber Br. at 57-58. 298 Brown v. Entm’t Merchs. Ass’n, 564 U.S. 786, 801–02 (2011). See also R.A.V., 505 U.S. at 387 (“[T]he First Amendment imposes not an ‘underinclusiveness’ limitation but a ‘content discrimination’ limitation upon a State’s prohibition of proscribable speech. There is no problem whatever, for example, with a State’s prohibiting obscenity (and other forms of proscribable expression) only in certain media or markets, for although that prohibition would be ‘underinclusive,’ it would not discriminate on the basis of content.”). 65 Even when this is not the case, Central Hudson does not require that the Ordinance “redress the harm completely.”299 The City may choose to regulate only “part” of the speech that causes harm.300 Here, as part of the regulatory scheme, the City has chosen to allow owners of their own prior salary data to remain in control of that information and thereby allow the employee to decide whether s/he wants to disclose it. The Chamber also suggests that more rigorous enforcement of current antidiscrimination laws is an alternative that the City must attempt before passing an Ordinance such as this. Intermediate scrutiny, however, does not require that the City adopt such regulatory measures only as a last alternative or that the City demonstrate that the legislation is the least restrictive response.301 Moreover, it is clear on this record and from some of the cases we have discussed (see Rizo) that the wage gap has survived other remedial measures, including the Equal Pay Act.302 The testimony supporting the Inquiry Provision establishes that, despite the presence of antidiscrimination laws, that “[t]he gender wage gap has narrowed by less than one-half a penny per year in the United States since 1963.”303 299 Mariani v. United States, 212 F.3d 761, 774 (3d Cir. 2000). 300 Id. 301 See Fox, 492 U.S. at 476–78. 302 The Chamber also cited to our recent decision in Bank of Hope in its Rule 28(j) letter to argue that the City was required to attempt a host of other alternatives before implementing the Ordinance. Central Hudson scrutiny does not require the City to adopt the least restrictive means to achieve its goal. Moreover, in Bank of Hope v. Miye Choni, 938 F.3d 389 (3d Cir. 2019). We concluded that there “neither the magistrate judge nor the district court considered a single alternative.” Id. at 396. In contrast, the City here, considered and appropriately rejected a number of alternatives, including the patently deficient alternatives suggested by the Chamber, such as simply enforcing current antidiscrimination laws, which have been insufficient to meaningfully close the wage gap. 303 § 9-1131(1); see also JA299-300 (summarizing testimony before the City regarding existing laws that have insufficiently closed the pay gap). 66 The City enacted the Inquiry Provision in an attempt to address this persistent problem and the record is clearly sufficient to withstand this First Amendment challenge to it. III. CONCLUSION For the foregoing reasons, we will affirm the district court’s denial of a preliminary injunction as to the Reliance Provision and we will vacate the district court’s grant of a preliminary injunction as to the Inquiry Provision and remand with directions to the district court to deny the preliminary injunction as to the Inquiry Provision. 67