People v. Uber Technologies, Inc.

• Court: California Court of Appeal
• Decided: October 22, 2020
• Docket: A160701
• Status: Published

Opinion

Filed 10/22/20

CERTIFIED FOR PUBLICATION

IN THE COURT OF APPEAL OF THE STATE OF CALIFORNIA

FIRST APPELLATE DISTRICT

DIVISION FOUR

THE PEOPLE, Plaintiff and Respondent, A160701, A160706 v. (City & County of San Francisco UBER TECHNOLOGIES, INC., et al., Super. Ct. No. CGC-20-584402) Defendants and Appellants.

To secure a preliminary injunction, the plaintiff must establish “that he is likely to succeed on the merits, that he is likely to suffer irreparable harm in the absence of preliminary relief, that the balance of equities tips in his favor,” and, particularly where public harm is implicated, “that an injunction is in the public interest.” (Winter v. Natural Resources Defense Council, Inc. (2008) 555 U.S. 7, 20 (Winter).) But it must always be kept in mind that interim injunctive relief is rooted in principles of equity and is fundamentally committed to the sound discretion of the trial court. “Flexibility is a hallmark of equity jurisdiction. ‘The essence of equity jurisdiction has been the power of the Chancellor to do equity and to mould each decree to the necessities of the particular case. Flexibility rather than rigidity has distinguished it.’ [Citation.] Consistent with equity’s character, courts do not insist that litigants uniformly show a particular, predetermined quantum of probable success or injury before awarding equitable relief. Instead, courts have

1 evaluated claims for equitable relief on a ‘sliding scale,’ sometimes awarding relief based on a lower likelihood of harm when the likelihood of success is very high.” (Id. at p. 51 (dis. opn. of Ginsburg, J.).) This reminder that the foundation of interim injunctive relief lies in equity comes from Justice Ruth Bader Ginsburg, who was renowned for her expertise in procedure long before she became the national icon known as RBG. What Justice Ginsburg says in Winter, though put forward on a point of federal law in dissent—a dissent that would have affirmed as within a trial judge’s considered discretion the issuance of a preliminary injunction in favor of a private party against an alleged violation of a federal statute by the Navy—happens to capture the essence of California law on the same point. (Butt v. State of California (1992) 4 Cal.4th 668, 678 (Butt) [a trial court’s decision to issue preliminary injunctive relief “must be guided by a ‘mix’ of the potential-merit and interim-harm factors; the greater the plaintiff’s showing on one, the less must be shown on the other to support an injunction”].) Justice Ginsburg’s cogent explanation of the governing standard as one that rests on a “sliding scale” calculus expresses a principle that will ultimately drive our analysis of this case. We have before us a civil enforcement action brought by the People1 against defendants Uber Technologies, Inc. and Lyft, Inc. (Uber and Lyft). Compared to Winter, the roles of the parties are reversed: It is the government that seeks interim injunctive relief against private parties. The core allegation in the case is that Uber and Lyft improperly misclassify drivers using their ride-hailing platforms as independent contractors rather

1 The Attorney General of California, joined by city attorneys of the cities of Los Angeles, San Diego, and San Francisco, brought this action on behalf of the People. We shall refer to plaintiffs collectively as the People.

2 than employees, thus depriving them of a host of benefits to which employees are entitled. This misclassification, it is alleged, also gives defendants an unfair advantage against competitor companies, while costing the public significant sums in lost tax revenues and increased social-safety-net expenditures that are foisted on the state because drivers must go without employment benefits. Mindful that—absent legal error—our role in reviewing a decision to issue interim injunctive relief is a limited one, we address here whether the trial court abused its discretion in granting a preliminary injunction that restrains Uber and Lyft from classifying their drivers as independent contractors. Seeing no legal error, we conclude the trial court acted within its discretion and accordingly affirm the order as issued.2

2 We have read and considered amicus curiae briefs: (1) in support of defendants, from (a) Bay Area Council, Earth Sparks, Internet Association, Silicon Valley Leadership Group, and TechNet; (b) Chamber of Commerce of the USA, California Chamber of Commerce, National Retail Federation, and HR Policy Association; (c) Communities-of-Color Organizations; (d) Independent Women’s Law Center (“IWLC”); (e) Independent Drivers and Independent Drivers’ Association; (f) Mothers Against Drunk Driving and California State Sheriffs’ Association; (2) in support of the People, from (a) Gig Workers Rising, Mobile Workers’ Alliance, Rideshare Drivers United, and We Drive Progress; (b) International Brotherhood of Teamsters, Service Employees International Union California State Council, State Building and Construction Trades Council of California, Transport Workers Union, United Food and Commercial Workers Union Western States Council, Unite Here, and California Labor Federation; (c) National Employment Law Project, ACLU of Northern California, Asian Americans Advancing Justice–Asian Law Caucus, Bet Tzedek Legal Services, California Employment Lawyers’ Association, The Center for Workers’ Rights, Centro Legal de la Raza, Council on American Islamic Relations, California Chapter, La Raza Centro Legal Workers’ Rights Program, Legal Aid at Work, Lawyers’ Committee for Civil Rights of the San Francisco Bay Area, Women’s Employment Rights Clinic of Golden Gate University School of Law, and Worksafe, Inc.

3 I. BACKGROUND A. Legal Framework—Assembly Bill 5 In 2019, the Legislature enacted Assembly Bill 5 (AB 5), which codified the decision of our high court in Dynamex Operations West, Inc. v. Superior Court (2018) 4 Cal.5th 903 (Dynamex). (See Stats. 2019, ch. 296, § 1.) As currently found in Labor Code section 2775,3 AB 5 provides in pertinent part: “For purposes of this code and the Unemployment Insurance Code, and for the purposes of wage orders of the Industrial Welfare Commission, a person providing labor or services for remuneration shall be considered an employee rather than an independent contractor unless the hiring entity demonstrates that all of the following conditions are satisfied: [¶] (A) The person is free from the control and direction of the hiring entity in connection with the performance of the work, both under the contract for the performance of the work and in fact. [¶] (B) The person performs work that is outside the usual course of the hiring entity’s business. [¶] (C) The person is customarily engaged in an independently established trade, occupation, or business of the same nature as that involved in the work performed.” (§ 2275, subd. (b)(1).)

(collectively “National Employment Law Project et al.”); (d) Public Rights Project, 15 Civil Rights Gender Justice, and Worker Rights Organizations; and (3) the parties’ respective responses to these amicus briefs. We appreciate the efforts of amici to provide us with valuable perspectives. 3 All undesignated statutory references are to the Labor Code. AB 5 was originally codified as section 2750.3, effective January 1, 2020. (Stats. 2019, ch. 296, § 2.) Effective September 4, 2020, section 2750.3 was repealed and the statutory provisions pertinent to this dispute were transferred with no substantive changes to section 2775, subdivision (b). (Stats. 2020, ch. 38, §§ 1–2.) For the sake of simplicity, we refer to the statutory scheme at issue as AB 5.

4 This standard for distinguishing employees from independent contractors is known as the “ABC” test. (Dynamex, at p. 916.) Centrally at issue in Dynamex was whether, for purposes of class certification in class action litigation, it was possible to determine on a classwide basis whether drivers who delivered packages for a “nationwide same-day courier and delivery service” offering “on-demand, same-day pickup and delivery services to the public generally” as well as to “a number of large business customers” were employees or independent contractors. (Dynamex, supra, 4 Cal.5th at p. 917.) On-demand drivers were paid “either a percentage of the delivery fee paid by the customer on a per delivery basis or a flat fee basis per item delivered” (ibid.), were required to make deliveries in their own vehicles (ibid.) and were obligated to pay all costs of operating those vehicles (ibid.). But they had the flexibility to set their own schedules, subject to requirements that they notify Dynamex when they intended to work and that, while working, they wear Dynamex uniforms and display its trade dress. (Id. at p. 918.) They were not required to accept delivery assignments; they were “generally free to choose the sequence in which they w[ould] make deliveries and the routes they w[ould] take”; and “when they [were] not making pickups or deliveries for Dynamex, drivers [were] permitted to make deliveries for another delivery company, including the driver’s own personal delivery business.” (Ibid.) The case arose in the wage and hour context, with the original named plaintiff’s complaint alleging that Dynamex “improperly failed to comply with the requirements imposed by the Labor Code and wage orders for employees with respect to” its drivers. (Dynamex, supra, 4 Cal.5th at p. 919.) The plaintiff driver asserted claims for unfair competition and for violation of wage and hour protections in the Labor Code. (Ibid.) All of these claims were

5 based on misclassification. Until 2004, Dynamex had treated its unscheduled, on-demand drivers as employees, but abruptly in 2004, to save business costs, it recharacterized them as independent contractors. (Id. at p. 917.) A key procedural premise underlying all of the claims asserted by the putative class plaintiff was that the alleged “Labor Code violations based on Dynamex’s failure to pay overtime compensation, to properly provide itemized wage statements, and to compensate the drivers for business expenses” were amenable to treatment on a classwide basis. (Id. at pp. 919– 920.) In what can fairly be described as a landmark opinion, our Supreme Court unanimously held that this issue was amenable to proof on a classwide basis. In so holding, the court carefully traced the state of the law governing “whether an individual worker should properly be classified as an employee or, instead, as an independent contractor” (Dynamex, supra, 4 Cal.5th at p. 912), historically a vexed question in federal and state law and one that the court acknowledged “has considerable significance for workers, businesses, and the public generally.” (Ibid.) Taking up a legal issue that had been left open in Ayala v. Antelope Valley Newspapers, Inc. (2014) 59 Cal.4th 522, 531 (Ayala), the court addressed whether the “suffer or permit” to work definition announced in Martinez v. Combs (2010) 49 Cal.4th 35 should apply for purposes of class certification. (Dynamex, at pp. 941–942, 943–944.) Martinez adopted a broad, pro-worker test as set forth in the “suffer or permit to work” definition of the terms “employer” and “employee” for purposes of California wage orders. (Dynamex, at pp. 935–939.) An alternative approach, urged by defendant Dynamex as “the only appropriate standard under California law for distinguishing employees and independent contractors” (id. at p. 915), was to construe those terms under the common

6 law test enunciated in S.G. Borello & Sons, Inc. v. Department of Industrial Relations (1989) 48 Cal.3d 341 (Borello) for purposes of the Workers’ Compensation Act.4 (See Dynamex, at pp. 929–935 [discussing Borello]; Dynamex, at pp. 941–942 [posing issue of whether to apply Martinez or Borello test in class-certification context].) Under the six-factor, highly fact- bound Borello standard, “ ‘the significance of any one factor and its role in the overall calculus may vary from case to case depending on the nature of the work and the evidence.’ ” (Dynamex, at p. 941, fn. 15, quoting Ayala, at p. 539, citing Borello, at p. 354.) The court chose the suffer or permit to work definition and affirmed an order of class certification on that basis. (Dynamex, supra, 4 Cal.5th at pp. 941–950, 965–967.) Tracing the origin of this “exceptionally broad” standard (id. at p. 952) to federal wage and hour legislation sponsored in 1937 by then-Senator Hugo Black (id. at p. 951)—who described it as “ ‘the broadest definition’ that has been devised for extending the coverage of a statute or regulation to the widest class of workers that reasonably fall within the reach of a social welfare statute” (ibid.)—the court explained that the “suffer or permit to work standard in California wage orders finds its justification in the fundamental purposes and necessity of the minimum wage and maximum hour legislation in which the standard has traditionally

4 “The Workers’ Compensation Act . . . extends only to injuries suffered by an ‘employee,’ which arise out of and in the course of his ‘employment.’ (§§ 3600, 3700; see Cal. Const., art. XIV, § 4 (former art. XX, § 21).) ‘Employee[s]’ include most persons ‘in the service of an employer under any . . . contract of hire’ (§ 3351), but do not include independent contractors. The Act defines an independent contractor as ‘any person who renders service for a specified recompense for a specified result, under the control of his principal as to the result of his work only and not as to the means by which such result is accomplished.’ (§ 3353.)” (Borello, supra, 48 Cal.3d at p. 349.)

7 been embodied.” (Id. at p. 952.) The court’s summary of these purposes bears emphasis. “Wage and hour statutes and wage orders were adopted in recognition of the fact that individual workers generally possess less bargaining power than a hiring business and that workers’ fundamental need to earn income for their families’ survival may lead them to accept work for substandard wages or working conditions,” the court explained. (Ibid.) It explained, further, that “[t]he basic objective of wage and hour legislation and wage orders is to ensure that such workers are provided at least the minimal wages and working conditions that are necessary to enable them to obtain a subsistence standard of living and to protect the workers’ health and welfare.” (Ibid.) And, it summed up, “[t]hese critically important objectives support a very broad definition of the workers who fall within the reach of the wage orders.” (Ibid.) All parties in this case acknowledge that AB 5 codified the holding in Dynamex, thus putting a legislative imprimatur on what our Supreme Court held there.5 But it is also significant to note that the Legislature went

5 Structurally, AB 5 enacted a statutory scheme that codifies the ABC test as the general rule in testing for employee versus independent contractor status (§ 2775), subject to a series of statutory exemptions (see § 2776 [certain business-to-business contracting relationships], § 2777 [certain referral agencies and service providers], § 2778 [certain contracts for professional services], § 2779 [individuals acting as sole proprietors or other business entities performing work pursuant to contract at the location of a single-engagement event]; § 2780 [certain occupations in connection with sound recordings or musical compositions]; § 2781 [certain contractors and subcontractors in the construction industry]; § 2782 [certain data aggregators]; § 2783 [certain persons and entities in insurance and financial services industries, professional health care services, professional licensees, salespersons, commercial fishers, newspaper distributors, and others]; and § 2784 [motor clubs]). “If a hiring entity can demonstrate compliance with all of [the] conditions set forth in any one of Sections 2776 to 2784, inclusive,

8 beyond Dynamex in some critically important respects. First, it expressly conferred on the Attorney General, district attorneys, and certain city attorneys and prosecutors the power to seek injunctive relief against those who misclassify employees as independent contractors. (§ 2786.) Second, while the Dynamex court repeatedly emphasized that the controversy before it—and implicitly its holding—was limited to the wage and hour context (Dynamex, supra, 4 Cal.5th at pp. 941–942, 948), the Legislature made clear that it was broadly adopting the Dynamex holding for purposes of all benefits to which employees are entitled under the Unemployment Insurance Code, the Labor Code, and all applicable wage orders. (§ 2775, subd. (b)(1).) The plaintiff public officers in this case—the Attorney General, and the city attorneys of Los Angeles, San Francisco, and San Diego—have taken full advantage of their enforcement power under AB 5, suing here for injunctive relief against these defendants’ misclassification of ride-share drivers, a practice they allege has deprived the drivers of minimum wages,6 overtime

then Section 2775 and the holding in Dynamex do not apply to that entity, and instead the determination of an individual’s employment status as an employee or independent contractor shall be governed by Borello.” (§ 2785, subd. (d), italics added.) Notably, the statutory scheme also contemplates potential non-statutory exemptions. (§ 2775, subd. (b)(3) [“If a court of law rules that the three-part test in paragraph (1) cannot be applied to a particular context based on grounds other than an express exception to employment status as provided under paragraph (2), then the determination of employee or independent contractor status in that context shall instead be governed by the California Supreme Court’s decision in [Borello]”].)

6Sections 1182.12, 1182.13, 1194, and 1197, California Minimum Wage Order (MW-2019) (currently $13.00 per hour for employers with 26 or more employees), and I.W.C. Wage Order 9-2001, section 4; Los Angeles Minimum Wage Ordinance, Los Angeles Municipal Code, Chapter 18, Article 7, section 187.00 et seq. (currently $15.00 per hour); San Francisco Minimum Wage

9 wages,7 reimbursement for the necessary expenses of performing their work,8 meal and rest periods and premiums,9 wage statements,10 sick leave and health benefits,11 unemployment insurance and training fund contributions,12 disability insurance,13 and workers’ compensation benefits.14 They also allege injury to competitors who do provide employees these various benefits, and to the state because of defendants’ failure to pay their fair share of state and local payroll taxes and workers’ compensation insurance premiums. B. Defendants’ Businesses and Relationship with Drivers Both Uber and Lyft offer mobile phone applications (apps) that operate by matching those in need of a ride to ride-hailing drivers available to give them rides using their own vehicles. Defendants’ business models are

Ordinance, San Francisco Administrative Code, Chapter 12R (currently $15.59 per hour); City of San Diego Earned Sick Leave and Minimum Wage Ordinance, San Diego Municipal Code, Chapter 3, Article 9, Division 1 (currently $13.00 per hour). 7 Sections 510, 1194, and 1198 and I.W.C. Wage Order 9-2001, section 3(A). 8 Section 2802. 9 Sections 226.7 and 512 and I.W.C. Order 9-2001, sections 11 and 12. 10 Section 226. Section 246; Los Angeles Paid Sick Leave Ordinance, Los Angeles 11

Municipal Code, Chapter 18, Article 7, section 187.00 et seq.; San Francisco Paid Sick Leave Ordinance, San Francisco Administrative Code, Chapter 12W; San Francisco Health Care Security Ordinance, San Francisco Administrative Code, Chapter 14; San Francisco Paid Parental Leave Ordinance, San Francisco Police Code, Article 33H; San Diego Earned Sick Leave and Minimum Wage Ordinance, San Diego Municipal Code, Chapter 3, Article 9, Division 1. 12 Unemployment Insurance Code sections 976 and 976.6. 13 Unemployment Insurance Code sections 986, 2609, and 2652. 14 Sections 3207–3208 and 3700.

10 similar. Riders log into their accounts with Uber or Lyft through defendants’ apps and request a ride from one place to another. They are matched with nearby drivers who are available to give them a ride. The contracts between defendants and the drivers provide that the relationship between Lyft or Uber, on the one hand, and the drivers on the other, is not one of employment. Rather, the “Platform Access Agreement” for Uber’s “Rides” platform specifies that the parties’ relationship “is solely as independent business enterprises, each of whom operates a separate and distinct business enterprise that provides a service outside the usual course of business of the other.” Lyft’s “Terms of Service” provide that the driver and Lyft “are in a direct business relationship, and the relationship between the parties under this Agreement is solely that of independent contracting parties,” rather than an employment or agency relationship, and that Lyft does not control the drivers in their provision of rideshare services. Uber’s agreement with drivers recites that Uber has no right to direct or control the drivers; rather, under the agreement allowing drivers to use the Rides platform—i.e., the driver app and associated services—the drivers decide whether to use the app and whether to accept, decline, ignore, or cancel a ride request. Before accepting a ride request, drivers are given a prospective rider’s ratings, as well as information about the pickup location, requested destination, estimated trip duration, and estimated net fare, and riders may designate a preferred driver as a “favorite.” Although Uber provides navigation software, drivers may use any route they or their passengers choose on a ride. Uber does not limit the number of drivers who use its Rides platform, and it does not schedule them to drive at any particular time. Drivers need not accept any minimum number of rides to use the platform, and they may use any other platform or app in addition to

11 Uber’s. Uber does not interview drivers, collect resumes, or conduct reference checks. More than 311,000 drivers used Uber’s platform in California in 2019. Lyft similarly does not assign schedules, shifts, or driving areas to drivers. Drivers may use the app as much or as little as they want, including brief periods between other obligations, and they may log into other apps, such as Uber’s, while using Lyft’s app. They are free to accept or decline ride requests, and they may use a route of their or the passenger’s choosing. Lyft does not interview prospective drivers. Around 305,000 drivers used Lyft’s app in the year leading up to October 1, 2019. Each defendant ensures drivers meet certain standards before authorizing them to use the defendant’s platform and hold themselves out as Uber or Lyft drivers. Uber drivers are required to pass criminal background and driving record checks, and they must agree that their vehicles will be properly registered and maintained. Lyft requires its drivers to pass criminal background and driving record checks and to show that they are properly licensed and insured, that they have a right to drive their vehicles, and that their vehicles are in good operating condition and meet safety standards. Lyft limits the age and size of vehicles drivers may use; for instance, they must have four doors and at least five seats, and certain subcompact vehicles are ineligible. It requires vehicles to pass an inspection each year. Lyft and Uber both prohibit drivers using their apps from accepting street hails, bringing their friends along while providing rides, or receiving payment for rides in cash. Defendants offer incentives for drivers to drive at times when or in areas where there is higher demand. Defendants may monitor or collect information about drivers’ locations, communications with riders, and driving habits, such as speeding, braking,

12 and acceleration. Drivers and riders rate each other, and defendants may use low ratings to deactivate drivers. Defendants address riders’ complaints. Riders pay fares through Lyft’s and Uber’s apps, and Lyft and Uber deduct a fee for each ride and remit the remainder of the payment to the driver. Uber and Lyft set the base fare rates and time and distance rates. Uber maintains a bank account for the benefit of drivers, separate from its corporate accounts, into which fares and tips are paid, then transmits to the driver the fare and gratuity less any service fee. Lyft arranges payments through a payment processing service. Neither Uber nor Lyft compensates drivers for time they are logged on the apps but are not transporting passengers; they do not provide overtime premiums or paid rest periods; they do not reimburse drivers for the expenses necessary to do their work, such as vehicle maintenance, a mobile phone and data usage, or gasoline; and they do not provide workers’ compensation coverage or paid sick days. Uber and Lyft both encourage riders to obtain transportation through their apps. Uber’s internet site, for instance, tells potential riders they will receive “[a]lways the ride you want” and “a reliable ride in minutes,” and that they can “[r]equest a ride, hop in, and go.” Lyft’s web site advertises to potential riders “[t]he whole city. In the palm of your hand,” tells them to “[g]et a ride whenever you need one,” and represents that “[o]ur drivers are always nearby, so you can get picked up, on demand, in minutes.” Uber has recently made changes to its business practices that it contends are relevant to its relationship with the drivers. Drivers need not accept Uber’s base fare (or “surge” fare for busier times) but may set a multiplier to the base fare of their choosing, within limits set by Uber. They may also purchase “Drive Pass,” a subscription that entitles them to a

13 specified number of trip requests within a seven-day period, and Uber receives no additional service fee for those rides. Uber and Lyft each take the position that the drivers do not provide services to them and are not their employees, but instead are independent business people who pay for the use of their platforms to find opportunities to earn money. Lyft describes its business as “a multi-sided transportation platform that connects people who are looking for rides with drivers willing to provide them.” Its business, it asserts, is not providing rides, but “operating the software tools and a platform that connects riders and drivers.” Lyft describes both drivers and riders as users of its business. Similarly, Uber describes itself as a technology company that develops and maintains “multi-sided platforms,” or “digital marketplaces where providers or sellers of a good or service can connect with consumers of that good or service.” Its platforms, Uber asserts, “provide users (both the sellers and buyers) with various services, including matching and payment processing,” although it does not guarantee that all users will find a match. In addition to the rideshare platform at issue here (Rides), those services include food delivery (Uber Eats) and freight (Uber Freight). And its employees, Uber contends, are those who work on its technology and provide support services, not the drivers who use its products to find and receive compensation from passengers. C. Procedural History The People brought this action alleging that Uber and Lyft are transportation companies in the business of selling rides to customers and that their drivers are employees under Dynamex and AB 5. By misclassifying drivers as independent contractors, thus depriving them of the benefit of minimum wages, overtime pay, reimbursement for business expenses,

14 workers’ compensation, paid sick leave, disability insurance, and paid family leave, the People allege, defendants evade California’s workplace standards and safeguards and commit unfair business practices. The complaint asserts causes of action for injunctive relief, restitution, and penalties for violations of Business and Professions Code section 17200 and injunctive relief for violations of AB 5. Shortly after filing their complaint, on June 25, 2020, the People moved for a preliminary injunction enjoining defendants from continuing this practice, and prevailed. Granting the People’s motion, the trial court restrained Lyft and Uber, during the pendency of this action, from “classifying their Drivers as independent contractors in violation of [AB 5],” and from “violating any provisions of the Labor Code, the Unemployment Insurance Code, and the wage orders of the Industrial Welfare Commission with regard to their Drivers.” The court concluded the People had shown “a reasonable probability (indeed, an overwhelming likelihood)” of prevailing on the merits of their claim that Uber and Lyft were misclassifying their drivers as independent contractors in violation of AB 5; that substantial public harm would result in the absence of an injunction; and that the harm to defendants from erroneous entry of the injunction would not be grave or irreparable and would not outweigh the harm to drivers, businesses, and the general public in the absence of an injunction. The trial court issued its injunctive order on August 10, 2020, staying it for ten days to allow defendants to seek appellate relief. Defendants appealed and petitioned this court for a writ of supersedeas. On August 20, 2020, we granted the petition and stayed the order during the pendency of

15 this appeal subject to certain conditions, including an expedited briefing schedule, to which defendants agreed.15 II. DISCUSSION A. Legal Standards for Preliminary Injunction The decision whether to issue a preliminary injunction lies in the sound discretion of the trial court, which we do not disturb absent an abuse of discretion. (Hunt v. Superior Court (1999) 21 Cal.4th 984, 999 (Hunt); City of Corona v. AMG Outdoor Advertising, Inc. (2016) 244 Cal.App.4th 291, 298 (City of Corona).) On appeal we do not weigh conflicting evidence, but defer

15 The August 20 order reads in part as follows: “The petitions are granted and the preliminary injunction is stayed pending resolution of Lyft and Uber’s appeals, subject to the condition that, by 5:00 p.m. on August 25, 2020, Lyft and Uber shall both file written consents to the expedited procedures specified herein. If Lyft and Uber do not both file such written consents, the stay shall expire at 5:00 p.m. on August 25, 2020. The procedures are as follows: “1. Lyft’s and Uber’s appeals shall be consolidated . . . . “2. Lyft and Uber shall proceed with an appendix in lieu of a clerk’s transcript on appeal . . . . “3. Briefing shall proceed on . . . [a specified expedited] schedule . . . . Absent unforeseen extraordinary circumstances, there shall be no extensions. Oral argument shall be scheduled for October 13, 2020. “4. On or before September 4, 2020, each defendant shall submit a sworn statement from its chief executive officer confirming that it has developed implementation plans under which, if this court affirms the preliminary injunction and Proposition 22 on the November 2020 ballot fails to pass, the company will be prepared to comply with the preliminary injunction within no more than 30 days after issuance of the remittitur in the appeal. “5. Should Lyft or Uber fail to comply with these procedures, the People may apply to this court to vacate this stay. “Unless otherwise ordered, the stay will dissolve upon issuance of the remittitur in the appeal. (Cal. Rules of Court, rule 8.272.)” Both defendants consented to the conditions, and their chief executive officers have submitted declarations as required by paragraph 4 of the order.

16 to the trial court’s factual findings if they are supported by substantial evidence. (City of Corona, at pp. 298–299.) To the extent the trial court’s ruling rests on a legal issue, we review it de novo. (Huong Que, Inc. v. Luu (2007) 150 Cal.App.4th 400, 408.) The burden is on the party challenging the injunction to make a clear showing the trial court abused its discretion. (IT Corp. v. County of Imperial (1983) 35 Cal.3d 63, 69 (IT Corp.).) The trial court’s order on a request for a preliminary injunction “reflects nothing more than the superior court’s evaluation of the controversy on the record before it at the time of its ruling; it is not an adjudication of the ultimate merits of the dispute.” (People ex rel. Gallo v. Acuna (1997) 14 Cal.4th 1090, 1109 (Gallo); accord, Yee v. American National Ins. Co. (2015) 235 Cal.App.4th 453, 457–458.) The preliminary injunction is intended to “preserv[e] . . . the status quo until a final determination of the merits of the action.” (Continental Baking Co. v. Katz (1968) 68 Cal.2d 512, 528.) In general, when considering a request for a preliminary injunction, the trial court weighs two interrelated factors. The first is the likelihood the party seeking relief will prevail on the merits, and the second is the relative interim harm to the parties if the preliminary injunction is granted or denied. (Butt, supra, 4 Cal.4th at pp. 677–678; Hunt, supra, 21 Cal.4th at p. 999; IT Corp., supra, 35 Cal.3d at pp. 69–70.) The goal is to minimize the harm that an erroneous interim decision would cause. (IT Corp., at p. 73.) IT Corp. established a variation of this standard where a legislative enactment—there, a zoning ordinance—specifically provides for injunctive relief. (IT Corp., supra, 35 Cal.3d at pp. 66, 72–73.) The IT Corp. standard is highly pertinent here because AB 5 specifically authorizes the Attorney General or a city attorney or prosecutor to bring “an action for injunctive

17 relief to prevent the continued misclassification of employees as independent contractors.” (§ 2786; see former § 2750.3, subd. (j).) The injunction at issue in IT Corp. restrained a company from disposing of unauthorized wastes at a particular site, in violation of a zoning ordinance that specifically authorized injunctive relief. (IT Corp., supra, 35 Cal.3d at pp. 68–69.) In affirming the injunction, our high court concluded that the traditional balancing test should be adapted in such a circumstance. (Id. at p. 72.) The appropriate standard, the court explained, is as follows: “Where a governmental entity seeking to enjoin the alleged violation of an ordinance which specifically provides for injunctive relief establishes that it is reasonably probable it will prevail on the merits, a rebuttable presumption arises that the potential harm to the public outweighs the potential harm to the defendant. If the defendant shows that it would suffer grave or irreparable harm from the issuance of the preliminary injunction, the court must then examine the relative actual harms to the parties.” (Id. at p. 72, fn. omitted; see Water Replenishment Dist. of Southern California v. City of Cerritos (2013) 220 Cal.App.4th 1450, 1464 [since city did not establish it would suffer grave or irreparable harm from injunction, no need to weigh relative harms].) In carrying out this weighing, “an injunction should issue only if—after consideration of both (1) the degree of certainty of the outcome on the merits, and (2) the consequences to each of the parties of granting or denying interim relief—the trial court concludes that an injunction is proper.” (IT Corp., at p. 72.) Defendants argue the rule of IT Corp. is limited to prohibitory injunctions, and that it is inapplicable here because the injunction is mandatory, that is, it requires them to perform affirmative acts (such as changing their contractual relationship with their drivers or ceasing

18 operations in California) that will change the status quo. They rely upon long-established case law holding that preliminary mandatory relief is restricted to “ ‘extreme’ ” cases in which “ ‘the right thereto is clearly established . . . .’ ” (Board of Supervisors v. McMahon (1990) 219 Cal.App.3d 286, 295, quoting Hagen v. Beth (1897) 118 Cal. 330, 331; accord, Integrated Dynamic Solutions, Inc. v. VitaVet Labs, Inc. (2016) 6 Cal.App.5th 1178, 1184.) The trial court rejected this argument, concluding first, that an injunction that restrains a continued violation of state law is prohibitory in nature (see People ex rel. Brown v. iMergent, Inc. (2009) 170 Cal.App.4th 333, 342 [prohibition on continued violation of consumer protection laws]), and second, that the IT Corp. framework applies even where the injunction authorized by statute is mandatory. The parties continue to debate vigorously whether the injunction issued here is prohibitory because it merely restrains defendants from further violations of state law, or is mandatory because it requires defendants to take affirmative actions that will change the status quo in order to comply with the injunction. While the question whether the injunction is mandatory or prohibitory is complicated and not free from doubt on this record, we conclude it is ultimately academic here.16 Nothing in IT Corp. suggests its framework is

16 Professor John Leubsdorf, in an influential law review article on preliminary injunctions, cited on a different point by our Supreme Court in IT Corp., supra, 35 Cal.3d at p. 73, traces the historical roots of what he describes as a judicial aversion to interim mandatory injunctions and observes that “[o]ne might see” in this reluctance to grant such relief “a tendency to protect large businesses from judicial interference” because it allows these “[d]efendants [to] . . . argue that their activities ha[ve] become part of the status quo and that their profitability weighed against disruptive

19 limited to prohibitory injunctions. While there may be tension, there is no necessary conflict between applying this framework and recognizing the heightened scrutiny given to mandatory injunctions. (People ex rel. Herrera v. Stender (2012) 212 Cal.App.4th 614, 630 (Stender).) The People argue that any affirmative steps defendants must take to comply with the injunction are simply a matter of business choice in determining the proper way to bring themselves into compliance with law, while defendants claim that the pressure of the injunction, of necessity, will force a restructuring of their businesses. We need not pick sides in that debate. The burdens of coming into compliance may be—and shall be—taken into account under the IT Corp. framework in determining whether defendants face grave or irreparable harm, and if so, in the weighing of relative harms. We are guided by City of Corona, supra, 244 Cal.App.4th 291. The trial court there entered a preliminary injunction ordering the defendants to remove a billboard installed without a permit, pursuant to an ordinance that allowed abatement actions. (Id. at pp. 294, 296–297.) In reviewing the order, the appellate court acknowledged both the rule that a preliminary injunction that “mandates an affirmative act that changes the status quo . . . is scrutinized even more closely on appeal” and is granted only if the right is clearly established (id. at p. 299, citing Stender, supra, 212 Cal.App.4th at p. 630), and the rule that “a more deferential standard of review applies when the government is seeking to enjoin the violation of an ordinance” (City of Corona, at p. 299, citing IT Corp., supra, 35 Cal.3d at pp. 69–71, 73, and City of Claremont v. Kruse (2009) 177 Cal.App.4th 1153, 1166 (Kruse)).

interim relief.” (Leubsdorf, The Standard for Preliminary Injunctions (1978) 91 Harv. L. Rev. 525, 535, fn. 66.)

20 Also instructive is People ex rel. Feuer v. FXS Management, Inc. (2016) 2 Cal.App.5th 1154. An injunction there barred the defendants from operating a marijuana business or collective that was illegal under a city ordinance. (Id. at pp. 1157–1158.) Without discussing whether the injunction was mandatory or prohibitory, the appellate court applied the IT Corp. standard, presuming the existence of public harm because, in enacting a provision proscribing the activity at issue, the legislative body “ ‘already determined that such activity is contrary to the public interest.’ ” (Id. at p. 1162.) Defendants urge us to reject the IT Corp. standard, citing Stender, which considered an injunction requiring a lawyer and law firm to provide notice to clients that another lawyer from the firm had resigned with disciplinary charges pending, and to follow specified procedures in so doing (such as providing two copies of the notice within 30 days with a self- addressed stamped envelope, retaining returned copies of the notice, and including translations into three languages). (Stender, supra, 212 Cal.App.4th at pp. 619, 628–629.) The injunction was issued pursuant to Business and Professions Code sections 17203 and 17204, which authorize injunctive relief for unfair competition, based in part on violations of sections 6180 and 6180.1 of the same code, which require notice to clients in such a circumstance. (Stender, at pp. 621–622, 627.) In affirming the injunction, the court explained that a preliminary injunction that mandates an affirmative act or changes the status quo is scrutinized “ ‘ “even more closely” ’ ” for abuse of discretion, and is subject to stricter review on appeal. (Id. at p. 630.) The court did not apply, or even discuss, the IT Corp. framework for analyzing an alleged violation of an enactment that specifically authorizes injunctive relief. Stender does not persuade us that IT

21 Corp. is inapplicable here. First, a case is not authority for a proposition it does not consider. (People v. Ault (2004) 33 Cal.4th 1250, 1268, fn. 10.) In any case, the injunction in Stender required the defendants to carry out specific acts that went beyond the statutory notice requirements and beyond a mandate to cease violating the law. Here, as the trial court noted, AB 5 expressly authorizes injunctive relief to prevent misclassification of employees. (§ 2786.) Given this specific provision for injunctive relief, we believe it appropriate to apply the IT Corp. framework, which is premised on a recognition that, where the Legislature has specifically provided for injunctive relief, it has already determined that a violation of the statute will cause “significant public harm” and that “injunctive relief may be the most appropriate way to protect against that harm.” (IT Corp., supra, 35 Cal.3d at p. 70.) We therefore proceed to the first step of the IT Corp. analysis, which asks whether plaintiff has shown a reasonable probability it will prevail on the merits. (IT Corp., supra, 35 Cal.3d at p. 72.) B. Reasonable Probability of Prevailing on the Merits The crux of this lawsuit is whether, under the ABC test as adopted in Dynamex and codified in section 2775, ride-hailing drivers for Uber and Lyft are employees or independent contractors. This is a question of first impression in California. 1. The ABC Test and the “Hiring Entity” Issue Section 2775 establishes a presumption that one who “provid[es] labor or services for remuneration” is an employee. (§ 2775, subd. (b)(1).) This presumption may be rebutted if “the hiring entity” demonstrates that all three “ABC” conditions are satisfied. (§ 2775, subd. (b)(1)(A)–(C); Dynamex, supra, 4 Cal.5th at pp. 950–951, fn. 20, 956–958.)

22 As the California Supreme Court explained in Dynamex, the ABC test has been adopted by various jurisdictions. (Dynamex, supra, 4 Cal.5th at pp. 955–956 & fn. 23, citing Mass. Gen. Laws, ch. 149, § 148B; Del. Code Ann., tit. 19, §§ 3501(a)(7), 3503(c); N.J. Stat. Ann., § 43.21-19(i)(6)(A)–(C).) The question in Dynamex was whether workers should be treated as employees or as independent contractors for purposes of California wage orders. (Dynamex, at pp. 913–914.) After reviewing prior decisions, both in this state and in other jurisdictions, the court adopted a version of the ABC test that tracked that of Massachusetts. (Id. at p. 956, fn. 23.) This test places the burden on “the hiring entity to establish that the worker is an independent contractor who was not intended to be included within the wage order’s coverage,” and to do so by meeting all three factors in the ABC test. (Id. at p. 957.) Uber and Lyft argue the threshold question in an ABC analysis is whether they are “hiring entities.” Only if they are, they argue, does the court move on to consider whether the three ABC test factors are satisfied. They frame the “hiring entity” issue in this manner because, fundamentally, the case they make here rests on the theory that the drivers do not render services to them; rather, drivers are their customers, who render services to defendants’ other customers, the riders, using the two-sided platforms defendants developed. Pre-Dynamex out-of-state cases applying the ABC test describe a slightly different threshold inquiry: whether the worker “provided services” to the putative employer. Gallagher v. Cerebral Palsy of Massachusetts, Inc. (Mass.Ct.App. 2017) 86 N.E.3d 496, 499 (Gallagher), for instance, describes a “two-step inquiry”: first, whether the worker provided services to the putative employer and second, the three-part test allowing the putative employer to rebut the presumption of employment by proving the

23 person worked as an independent contractor. (See also Sebago v. Boston Cab Dispatch, Inc. (Mass. 2015) 28 N.E.3d 1139, 1147 (Sebago) [“The threshold question is whether the plaintiffs provided services to the defendants”].) We reject defendants’ invitation to import a threshold “hiring entity” inquiry into section 2775 by judicial construction. Drawing on the Massachusetts cases, they argue that the trial court committed legal error by applying the ABC test without first determining that they are “hiring entities” and therefore subject to section 2775. As the People point out in their respondent’s brief, however, the premise that Dynamex is limited by Massachusetts case law is mistaken. The Supreme Court went to great lengths to explain why the ABC test it adopted for purposes of California law derives from the suffer or permit to work definition embedded in the wage orders it was construing. As codified in section 2775, we think the phrase “hiring entity,” tracking the language of the Dynamex opinion, is intended to be expansive for reasons specific to California wage and hour laws and the longstanding social safety net objectives of those laws in this state. But even aside from defendants’ reliance on unpersuasive out-of-state authority, we reject their “hiring entity” argument on the merits because it rests on a false dichotomy. In defendants’ proffered mode of “hiring entity” analysis, we must first decide whether drivers’ services are rendered to riders, or to them, before applying the remainder of the ABC test. That, in our view, presents an artificial choice. What the argument masks is that drivers’ services may be rendered both to the hirer and to a third party, benefitting each one. In Dynamex itself, for example, the delivery services that drivers performed could have been characterized as having been carried out for the benefit of both the corporate dispatcher, Dynamex, and the shippers and recipients of packages. There was no suggestion in that case of

24 the need to address who received the drivers’ services before applying the ABC test. Nor is there any such need for ride-share drivers. Reading the term “hiring entity” in context, we think the phrase is used in Dynamex and in section 2775 for its neutrality, so that it covers both employment status and independent contractor status, and thus does not presuppose an answer one way or another. This construction contrasts with defendants’ insistence that the term “hiring entity” has talismanic significance as a threshold indicator of employment status. We note, further, that although we are construing California law as codified by AB 5 rather than applying the imported law of Massachusetts, we see no necessary inconsistency between this reading of the term and the Massachusetts cases cited by defendants, which ask at the outset whether a person rendered services to another, either as an employee or as an independent contractor. (Gallagher, supra, 86 N.E.3d at p. 499; Sebago, supra, 28 N.E.3d at p. 1147.) While the nature of the “hiring entity” is always central to the statutory analysis, that question in this case collapses into prong B of the ABC test, which looks to whether the drivers’ work is “outside the usual course” of defendants’ businesses. (§ 2775, subd. (b)(1)(B).)17 Most fundamentally, to make the determination of whether the party acquiring a worker’s service is a “hiring entity” an additional step in the ABC test—an analytical move that, in effect, creates a step zero and pretermits further analysis unless answered in the affirmative—is inconsistent with the holding in Dynamex: As our Supreme Court carefully delineated in that case, there are three steps to the ABC test, these steps may be considered in any

17The “hiring entity” status of a putative employer in a misclassification case may also be relevant to the inquiry under prong A, but since the trial court completed its analysis with prong B, and we conclude it was correct in doing so, we stop there as well. See Section II.B.2. post.

25 order, and the analysis is at an end if the putative employer fails at any step. (Dynamex, supra, 4 Cal.5th at pp. 956–963.) This is what makes the ABC test a streamlined analysis readily amenable to application as a matter of law, distinguishing it in most circumstances from the more elaborate Borello test. It is also what makes the ABC test a powerful enforcement tool, consistent with the spirit of the suffer or permit to work definition, especially where, as here, the government is the enforcer. Joined by a number of their supporting amici, defendants contend that, without an inquiry at the outset into whether they are “hiring entities,” there is the potential that the ABC test may be invoked and employment status will be found in myriad situations involving online marketplaces and routine commercial transactions. We view the handwringing over this prospect as overdrawn. Defendants are correct that there is a threshold test designed to prevent wholly inappropriate application of the ABC test, but it is not whether the putative employer is a “hiring entity.” The Legislature explicitly exempted numerous business sectors, professions, and commercial relationships from the scope of section 2775 (see ante, p. 8, fn. 5)—notably not including ride-sharing—and further recognized that there will be circumstances in which the ABC test “cannot be applied to a particular context based on grounds other than” one of the express statutory exemptions in the statutory scheme, thus triggering application of the Borello test instead of the more streamlined ABC test to determine employment status. (§ 2775, subd. (b)(3).) Defendants do not rely on a statutory exemption here. Nor have they invoked section 2775, subdivision (b)(3) on the ground that the ABC test cannot be applied in this “particular context.”

26 2. Application of the ABC Test a. Uber’s Showing Uber and Lyft both submitted expert and other evidence they contend show they provide services to the drivers, rather than employing them. Uber describes its Rides platform as a method for riders to connect with available drivers through its multi-sided platforms. Its “proprietary algorithm takes the inputs from riders and drivers, and uses that information to suggest optimum matches based on proximity,” after which Uber provides information to both driver and rider about how they have been rated on past rides, using its “bilateral rating system.” One of Uber’s expert witnesses, Dr. Terrence W. August, a business school professor whose expertise includes economic modeling, the economics of information systems, and operations management, described multi-sided platforms as “a type of business that facilitates transactions between two or more different groups such as purchasers and sellers,” who would not be able to find each other easily otherwise. Such platforms commonly provide services such as “ ‘[m]atching’ market participants on one side of the market to participants on the other side of the market in order to facilitate a transaction,” payment processing, collecting and processing information to support successful matching, providing sellers and purchasers with suggestions and information on pricing, and “[c]ommunicating and verifying the quality of market participants in order to facilitate more and better transactions.” Dr. August opined that Uber’s Rides platform is a two-sided market app that provides services in a fashion similar to other multi-sided platforms by matching drivers and riders, and that both riders and drivers “are customers of—rather than employees working on behalf of—the platform.”

27 Dr. August also explained that businesses that provide two-sided market apps are compensated by charging the seller and/or the purchaser, charging subscription fees, charging fixed fees per listing or sale, charging percentage fees, or a combination of those methods, and that Uber’s manner of charging riders and drivers is consistent with these approaches. The more users on one side of the platform—i.e., drivers or riders—the greater the value to the users on the other side of the platform. Other examples of two- sided market apps are StubHub, eBay, Angie’s List, EnergySage, and Teledoc. Another expert, Dr. Justin McCrary, a law professor and economist with expertise in economic modeling and econometric and statistical methods, opined that Uber’s matching service “is supported by advanced technology and technical employees,” and that Uber has “several distinguishing features that improve market efficiency and benefit both passengers and drivers”: there is a “relatively quick and easy enrollment process”; drivers may choose their own schedule and locations; drivers may simultaneously use other platforms’ apps to find leads; and Uber uses advanced technology to match driver supply with passenger demand efficiently. Dr. McCrary characterized Uber as a “network company” that “connects independent service providers and consumers, where the independent service provider is hired by the consumer to provide a one-time service.” He noted that many Uber drivers value having control over their own schedules, and pointed to evidence that more than half of Uber’s drivers work 12 hours or fewer per week, and less than 20 percent work more than 30 hours a week; that drivers’ schedules vary considerably from week to week; and that many drivers for Uber and other networks engage in other

28 types of work as well, including “traditional employment” or logging into more than one app (such as Uber and Lyft) at once. b. Lyft’s Showing Lyft’s Director of Data Science explained that its predecessor company, Zimride, began as an electronic message board for students to arrange carpools home from college. Lyft’s current platform was launched in 2012, and “allows users to arrange shorter distance rides on a peer-to-peer basis on demand.” Lyft’s business, he explained, is “operating the software tools and a platform that connects riders and drivers.” Lyft sees its drivers as users of its platform, and takes the position it does not receive services from them. Rather, it allows drivers to use their “spare time and unused seat capacity” to earn extra money. Lyft does not assign schedules or coverage areas, and drivers may switch between Lyft and any other platforms, including Uber. The value of the platform to users on one side of the platform—drivers or riders—increases as more users are added on the other side. Lyft’s technology matches drivers and riders by taking into account various factors, including distance between driver and rider, and it reduces the effort riders and drivers need to connect with each other. Drivers receive incentives to log in at times or places with higher demand, and riders may receive price discounts when the supply of drivers outstrips demand. Lyft offers payment-processing services that “reduce friction” between drivers and riders by requiring riders to have a payment method associated with their account and processing payments to drivers. Dr. Catherine Tucker, a business school professor who specializes in the economics of digital technology, testified on behalf of Lyft that a multi- sided platform acts as “a matchmaker or intermediary for distinct groups of

29 users who wish to interact in some way” and that participants are users, not employees, of the platform. Lyft’s business involves attracting groups of users to the platform and ensuring their interactions are positive. Lyft offers incentives to both riders and drivers, and the “symmetry” with which it treats both drivers and riders is inconsistent with an employment relationship with drivers. Dr. Tucker also testified that a flexible schedule is important to 91 percent of the drivers who use Lyft’s platform, and that most drive only in “short bursts,” with more than half of driving sessions lasting less than an hour and 84 percent less than three hours. Drivers earned more than $20 per hour on average in 2019 after taking into account the cost of fuel, maintenance, and depreciation per mile, and they would rarely have qualified for sick leave or overtime pay if they were employees. c. Analysis To prevail on their claim that the drivers are not their employees, defendants must establish that all three ABC factors apply. (Dynamex, supra, 4 Cal.5th at p. 963.) The trial court addressed only the second of these factors, that is, whether “[t]he person performs work that is outside the usual course of the hiring entity’s business” (§ 2775, subd. (b)(1)(B)), and concluded that the record before it showed defendants were in the business of transporting passengers for compensation. Our high court has explained this factor, prong B of the ABC test, as follows: “Workers whose roles are most clearly comparable to those of employees include individuals whose services are provided within the usual course of the business of the entity for which the work is performed and thus who would ordinarily be viewed by others as working in the hiring entity’s business and not as working, instead, in the worker’s own independent

30 business.” (Dynamex, supra, 4 Cal.5th at p. 959.) Thus, for example, an outside plumber repairing a leak in a retail store is not part of the store’s usual course of business, but a cake decorator regularly hired by a bakery to work on custom-designed cakes would be part of the bakery’s business. (Id. at pp. 959–960.) Although we need not conclude that Uber and Lyft’s position that they are not in the transportation business is frivolous, we find no abuse of discretion in the trial court’s conclusion that the People have shown a probability of prevailing on the merits based on prong B. In light of the overlap between the questions of (1) whether the drivers render services to defendants, and (2) the scope of defendants’ businesses, we look to cases considering both issues to assist us. A number of cases have considered contentions that ride-sharing companies such as Lyft and Uber are in the business solely of creating technological platforms, not of transporting passengers, and have dismissed them out of hand. In 2015—before our high court adopted the ABC test in Dynamex—the Northern District of California addressed whether Lyft should have paid the plaintiffs, former drivers, as employees rather than as independent contractors. (Cotter v. Lyft, Inc. (N.D.Cal. 2015) 60 F.Supp.3d 1067, 1070.) Lyft argued as a threshold matter that the drivers performed services not for Lyft but for the riders, while Lyft merely furnished the platform that allowed riders and drivers to connect. (Id. at p. 1078.) The court concluded, “[T]hat is obviously wrong. Lyft concerns itself with far more than simply connecting random users of its platform. It markets itself to customers as an on-demand ride service, and it actively seeks out those customers. [Citation.] It gives drivers detailed instructions about how to conduct themselves. Notably, Lyft’s own drivers’ guide and FAQs state that drivers are ‘driving for Lyft.’ [Citation.] Therefore, the argument that Lyft is

31 merely a platform, and that drivers perform no service for Lyft, is not a serious one.” (Ibid.; see also Rogers v. Lyft, Inc. (N.D.Cal. 2020) 452 F.Supp.3d 904, 911 [under AB 5, “Lyft drivers provide services that are squarely within the usual course of the company’s business, and Lyft’s argument to the contrary is frivolous”].) Similarly, the court in Cunningham v. Lyft, Inc. (D.Mass. 2020) 2020 U.S.Dist. Lexis 90333 (Cunningham), considering a request for a preliminary injunction enjoining Lyft from misclassifying its drivers as independent contractors, found the plaintiff drivers had shown a “substantial likelihood of success on the merits that, despite Lyft’s careful self-labeling, the realities of Lyft’s business—where riders pay Lyft for rides—encompasses the transportation of riders.” (Id. at pp. *2–*3, *28–*29.) The court went on to reject Lyft’s argument that drivers received a service from them, rather than providing one to them, and that its business was only connecting riders and drivers, noting, “Lyft ignores that the drivers are ‘provid[ing] transportation services to riders,’ and that service . . . is the service for which Lyft is being paid by riders.” (Id. at pp. *29–*30.) Uber has been similarly unsuccessful in making its pitch to the courts. In 2015, the Northern District of California rejected Uber’s argument that it was not a transportation company but a technology company, concluding this was “an unduly narrow frame,” and that “Uber does not simply sell software; it sells rides.” (O’Connor v. Uber Technologies, Inc. (N.D.Cal. 2015) 82 F.Supp.3d 1133, 1141.) The court went on, “Even more fundamentally, it is obvious drivers perform a service for Uber because Uber simply would not be a viable business entity without its drivers. [Citations.] Uber’s revenues do not depend on the distribution of its software, but on the generation of rides by its drivers.” (Id. at p. 1142, fn. omitted.) The court noted that Uber billed its riders directly for the entire amount of the fare charged—in which drivers

32 had no input—then paid the driver 80 percent of the fare. “Put simply, . . . Uber only makes money if its drivers actually transport passengers.” (Ibid.; see also Crawford v. Uber Technologies, Inc. (N.D.Cal. 2018) 2018 U.S.Dist. Lexis 33778, *11–*13 [concluding plaintiffs had “plausibly alleged that Uber is ‘primarily engaged in the business of transporting people,’ ” and noting, “[t]o say that Uber merely facilitates connections between ‘both sides of the two-sided ridesharing market’ obscures the fact that Uber arguably created a market for this type of transportation”]; Namisnak v. Uber Technologies, Inc. (N.D.Cal. 2020) 444 F.Supp.3d 1136, 1143 [“Uber’s claim that it is ‘not a transportation company’ strains credulity, given the company advertises itself as a ‘transportation system’ ”].) These authorities suggest the trial court did not abuse its discretion in finding plaintiff had met its burden to show a reasonable probability of success on the merits. Cases considering other companies involved in transporting passengers also provide useful insights. In this state, the court in Yellow Cab Cooperative, Inc. v. Workers’ Comp. Appeals Bd. (1991) 226 Cal.App.3d 1288 (Yellow Cab), considered whether an injured taxi driver was an employee of the company, Yellow Cab Cooperative (Yellow), that leased a cab to him, for purposes of a claim for workers’ compensation. (Id. at p. 1291.) The agreement between the driver and Yellow designated him as a lessee; he leased the cab for 10-hour shifts and paid a flat rate per shift, while Yellow provided telephone call service, radio service, and repair and maintenance service. (Id. at pp. 1291–1292.) The Workers’ Compensation Appeals Board concluded the driver was an employee, and the appellate court upheld the order. (Id. at p. 1291.) In doing so, it rejected Yellow’s position that the driver was not rendering a service to it when he was injured, stating, “Contrary to Yellow’s portrayal here, the essence of its enterprise was not

33 merely leasing vehicles. It did not simply collect rent, but cultivated the passenger market by soliciting riders, processing requests for service through a dispatching system, distinctively painting and marking the cabs, and concerning itself with various matters unrelated to the lessor-lessee relationship,” such as instructing drivers in service and courtesy, keeping the cabs clean, going on calls they were sent on, and being courteous and helpful to the public. (Id. at p. 1293.) Yellow’s enterprise, the court ruled, “consists of operating a fleet of cabs for public carriage. [Citations.] The drivers, as active instruments of that enterprise, provide an indispensable ‘service’ to Yellow; the enterprise could no more survive without them than it could without working cabs.” (Id. at pp. 1293–1294; accord, Linton v. DeSoto Cab Co., Inc. (2017) 15 Cal.App.5th 1208, 1221 (Linton).) We recognize that defendants’ business models are different from that traditionally associated with employment, particularly with regard to drivers’ freedom to work as many or as few hours as they wish, when and where they choose, and their ability to work on multiple apps at the same time. But some of the features of the delivery-driver model at issue in Dynamex are present here as well. Strip away the use of the internet as a mode of communication with drivers, and this case bears many similarities to that one. The dispositive issue there was not whether the defendant and its drivers followed what might be viewed as a traditional employment model, who may be said to receive the drivers’ services, or how payment was structured, but whether the mode in which the drivers were utilized met the elements of the ABC test. So too in this case. There is considerable evidence that the ride-share drivers involved here meet this test, despite the changes in the traditional workplace enabled by modern technology.

34 Most pertinent is the following. Uber and Lyft both solicit riders. (See Yellow Cab, supra, 226 Cal.App.3d at p. 1293.) They screen drivers and set standards for vehicles that can be used. Defendants track and collect information on drivers when they are using the apps, and they may use negative ratings to deactivate drivers. Riders request rides and pay for them through defendants’ apps, and the drivers’ portions are then remitted to them, either through a payment processing service or a dedicated bank account. (Ibid.; compare AC&C Dogs, LLC v. New Jersey Dept. of Labor (N.J.App.Div. 2000) 753 A.2d 737, 738, 740 [individuals who rented hot dog carts from company and paid rental charged based on number of hot dogs sold not employees of cart company; remuneration flowed to company, not from it]; see also Sebago, supra, 28 N.E.3d at p. 1149 [taxicab drivers provided services to radio associations where vouchers from associations’ corporate clients were submitted to drivers as payment, and associations gave drivers amount equal to fare and tip minus a “ ‘processing’ fee”]; Koza v. New Jersey Dept. of Labor (N.J.App.Div. 1998) 704 A.2d 1310, 1312 [for ABC test to apply, remuneration must flow from putative employer to alleged employee].) The remuneration here may reasonably be seen as flowing from riders to defendants, then from defendants to drivers, less any fee associated with the ride. With the possible exception of rides obtained using Uber’s Drive Pass subscriptions—which we discuss separately below—defendants’ revenues are directly connected to the fees that riders pay for each ride. (Compare Parks Cab Co. v. Annunzio (Ill. 1952) 107 N.E.2d 853, 854–855 [no employment relationship between taxicab drivers and cab company from which they leased licenses for flat fee where, apart from tort liability, “the company is not concerned with the operation of the cabs or the results of their

35 operation” and it has no control over operation of cabs]; accord, Metro East Cab Co. v. Doherty (Ill.Ct.App. 1999) 705 N.E.2d 947, 952.) These facts amply support the conclusion that, whether or not drivers purchase a service from defendants, they perform services for them in the usual course of defendants’ businesses. Defendants’ businesses depend on riders paying for rides. The drivers provide the services necessary for defendants’ businesses to prosper, riders pay for those services using defendants’ app, and defendants then remit the drivers’ share to them, either through a bank account in the case of Uber or a payment processing service in the case of Lyft. Arguing to the contrary, defendants reprise the theme that, under the contracts they have with drivers, drivers do not perform services for them but just the reverse—drivers are their customers. Not only is there substantial evidence in the record supporting the trial court’s rejection of that argument, but it was correct to do so as a legal matter as well under the rule that the parties’ characterization of their relationship is not dispositive because their “actions determine the relationship, not the labels they use.” (Linton, supra, 15 Cal.App.5th at p. 1217.) None of the cases relied upon by defendants persuades us otherwise. For instance, Lyft cites Ruggiero v. American United Life Ins. Co. (D.Mass. 2015) 137 F.Supp.3d 104, to argue that it provides matching services rather than transportation, but that case was persuasively distinguished in Cunningham. The plaintiff in Ruggiero sold insurance policies for a company that he contended was his employer because the sales were essential to its business. (Id. at p. 118.) Considering the relationship between a defendant who manufactures and administers insurance policies, and a plaintiff who offers the opportunity to buy those policies, the court explained: “Two categories can be discerned from the case law: on the one hand, there are

36 cases in which the defendants equip the plaintiffs with the tools, resources, and opportunity to sell or provide the defendants’ products, often earning a commission or percentage of the sales, and essentially franchising their business; and on the other, there are cases in which the defendants merely give the plaintiffs a license or a product and leave the plaintiffs to their own devices to make a profit from it.” (Id. at p. 119.) On the facts before it, the court in Ruggiero found the plaintiff’s services were “ ‘merely incidental’ ” to the insurance company’s business. (Id. at p. 122.) Considering drivers for Lyft, the court in Cunningham noted that Lyft’s proceeds were directly dependent on the drivers’ services, and concluded there was a “substantial likelihood that on the merits this case will fall in [Ruggiero’s] first category” and that Lyft would thus be considered an employer. (Cunningham, supra, 2020 U.S.Dist. Lexis 90333 at pp. *30–*32.) Defendants also draw our attention to a number of cases from other states in which brokers who matched consumers and workers were not treated as the workers’ employers. (See, e.g., Trauma Nurses, Inc. v. Bd. Of Rev. (N.J.Super.Ct.App.Div. 1990) 576 A.2d 285, 286, 290 [service supplying hospitals with temporary nurses]; State Dept. of Employment, Training & Rehab., Employment Securities Div. v. Reliable Health Care Services Of South Nev., Inc. (Nev. 1999) 983 P.2d 414, 418 [health care worker temporary placement agency; “providing patient care and brokering workers are two distinct businesses”]; Q.D.-A., Inc. v. Ind. Dept. of Workforce Development (Ind. 2019) 114 N.E.3d 840, 843, 847–848 [business connected drivers with customers who needed too-large-to-tow vehicles driven to them]; but see O’Hare-Midway Limousine Service, Inc. v. Baker (Ill.Ct.App. 1992) 596 N.E.2d 795, 797–798 [business of furnishing chauffeur services constituted employment where limousine drivers paid percentage of commission to

37 company, establishing “financial interdependence, or a direct financial stake with the limousine company”].) None of these cases involves the continual coordination between worker and company at every stage of the work performed or the financial interdependence that is present here. The principal California case the defendants rely upon in support of the argument that their business models meet the requisites of prong B is Curry v. Equilon Enterprises, LLC (2018) 23 Cal.App.5th 289. The issue there was whether a manager at a service station was employed by Shell Oil Products US (Shell) as well as by the service station operator. Shell leased service stations to entities that operated the service stations; Shell owned the gasoline that was sold and received all revenue from fuel sales; and the operators retained all profits from the service stations’ convenience stores and carwash facilities. (Id. at pp. 292–293.) The manager was hired by the service station operator. (Id. at p. 295.) Even assuming that Dynamex extended beyond the independent contractor context to the joint employment context at issue in Curry—a point it questioned—the appellate court concluded there was no triable issue of fact as to prong B of the ABC test, whether the manager’s work was part of Shell’s usual course of business. (Id. at pp. 314–315.) The service station operator, not Shell, was responsible for all aspects of the employment relationship, including hiring and compensation, and controlled its employees’ daily work, and Shell did not acquiesce in the manager’s employment. (Id. at p. 311.) The operator, not Shell, operated the service stations. The court concluded that “Shell was not in the business of operating fueling stations—it was in the business of owning real estate and fuel.” (Id. at pp. 307, 315.) The case before us is readily distinguishable from Curry. This is not a situation in which a putative joint employer leases facilities to a worker’s direct employer and has

38 no involvement in the worker’s employment or compensation. Rather, defendants’ usual course of business involves the day-to-day task of matching riders and drivers each time a user requests a ride, arranging for riders’ payments to be processed, and retaining a portion of the proceeds from each ride. Curry does not assist defendants. Based on the breadth of the term “hiring entity” as well as the conspicuous absence of an express exemption for ride-sharing companies in the statutory scheme enacted by AB 5, we have little doubt the Legislature contemplated that those who drive for Uber and Lyft would be treated as employees under the ABC test. Indeed, as the trial court pointed out, Uber is currently—and so far, unsuccessfully—challenging the constitutionality of the measure in federal court, arguing that the legislation “irrationally targets gig economy companies and workers.” (Olson v. California (C.D.Cal. Feb. 10, 2020, No. CV 19-10956-DMG (RAOxO)) 2020 U.S.Dist. Lexis 34710, *13– *14.) At oral argument, counsel for Uber confirmed that his client does indeed take this view, though he was quick to add that the Legislature may have “targeted and missed.” While one might quibble with the word “target” given the breadth of the AB 5 statutory scheme, we appreciate the candor, because the legislative history does appear to show an awareness that the misclassification issues AB 5 sought to address are prevalent not just in traditional “brick and mortar” businesses, but in modern technology-driven companies as well.18

18 For example, one comment made by a member of the Assembly was that “[i]t is not the intent of AB 5 to distinguish between ‘platform’ and ‘brick and mortar’ businesses. Both types of business rely on individuals to perform work as part of the usual course of their business.” (Letter from Assembly Member Lorena Gonzalez to E. Dotson Williams, Chief Clerk of the Assembly (September 13, 2019).) Assemblymember Gonzalez sponsored AB 5. One

39 Pointing to steps it took in an apparent effort to adjust to the new legal standard following passage of AB 5, Uber contends that changes it made to its business practices since then—such as allowing drivers the option of setting a fare multiplier and allowing them to purchase Drive Passes rather than having Uber’s service fee withheld on a per-ride basis—now take it outside the statute’s ambit. And, Uber points out, riders’ payments are processed through a bank account maintained for the benefit of drivers, separate from Uber’s corporate accounts. We acknowledge these newly adopted business practices, but are not persuaded they make a difference to the analysis. The fare multipliers are within limits set by Uber, and the Drive Passes provide only a limited number of leads, including leads the

opponent of the measure argued to the Assembly Committee on Labor and Employment, “[I]ndependent contractor status has fostered the growth of the so-called ‘gig’ economy, with companies like Uber and Lyft, which enables thousands of college students, active duty military personnel and others to fill spare-time hours and generate income.” (Hearing Report, Ass. Comm. on Labor and Employment (April 3, 2019), at p. 6.) Others, supporting the bill, observed, “[A]n internet application, no matter how clever, cannot turn lead to gold.” (Hearing Report, Sen. Comm. On Labor, Public Employment and Retirement (July 10, 2019), p. 10.) At oral argument, counsel for Lyft correctly pointed out that as a general matter, statements of individual legislators, including bill sponsors, may not be relied upon in using legislative history to construe the meaning of ambiguous statutes. (Quintano v. Mercury Casualty Co. (1995) 11 Cal.4th 1049, 1062.) That is equally true for the views of supporters or opponents from outside the Legislature. (See Kaufman & Broad Communities, Inc. v. Performance Plastering, Inc. (2005) 133 Cal.App.4th 26, 37–39 [authoring legislator’s files, letters, press releases and other statements not communicated to the Legislature as a whole not properly cognizable in assessing legislative history].) Although we see no need to go beyond the plain text and structure of AB 5 in construing the statutory scheme, we do note that these statements by Assemblymember Gonzalez and others are consistent with the acknowledgment of Uber’s counsel that the Legislature “targeted” ride-sharing companies, even if its aim was not a rifle-shot.

40 driver rejects. And Uber has not shown that Drive Passes account for a significant portion of its business or that any drivers use them exclusively. Even if some drivers take advantage of these options, we agree with the trial court that these changes do not alter the basic fact that providing transportation is part of Uber’s usual course of business. While these details relating to how drivers are compensated might to a limited extent bear on whether the drivers are free from Uber’s direction and control or whether the drivers are engaged in an independently established trade—prongs A and C of the ABC test—they do not support Uber’s contention that the drivers’ work is outside the usual course of its business under prong B. (§ 2775, subd. (b)(1).) Quite to the contrary, according to the People, “Drive Pass . . . financially incentivizes the Driver to accept every dispatched ride [and thus] . . . is further evidence of why Drivers are within Uber’s usual course of business—to provide rides.” Another set of arguments arises from the fact that defendants are regulated by the Public Utilities Commission (PUC) as “transportation network companies” (TNC), defined as “an organization . . . that provides prearranged transportation services for compensation using an online- enabled application or platform to connect passengers with drivers using a personal vehicle.” (Pub. Util. Code, § 5431, subd. (c).) Defendants contend the trial court’s evaluation of the merits was based on a misapplication of the statutes governing TNC’s. As they point out, when identifying the nature of defendants’ businesses for purposes of prong B of the ABC test, the trial court first looked to provisions of the Public Utilities Code establishing TNC’s as a new category of charter party carriers, and defining charter carriers as “engaged in the transportation of persons” and TNC’s as “provid[ing] prearranged transportation services for compensation

41 . . . .” (Pub. Util. Code, §§ 5440, subd. (a), 5360, 5360.5, 5431, subd. (c).) These provisions, the court concluded, show that defendants are in the business of transporting passengers for compensation. Defendants disagree, contending the statutes and regulations governing TNC’s do not establish that drivers are their employees: Uber points out that the PUC has expressly disavowed any intention to “meddle into their business model” by requiring them to designate their drivers either employees or contractors. (Cal.P.U.C. Dec. No. 13-09-045 (Sept. 19, 2013) 2013 Cal.P.U.C. Lexis 504, *98–*99 (2013 PUC Decision).) Lyft, in turn, points out that, consistent with PUC regulations, its permit from the PUC recites that it is authorized to “facilitate rides between passengers and private drivers using their own personal vehicles” (italics added) and that a transportation network company is not permitted to own vehicles used in its operation. (See Cal.P.U.C. Res. No. TL- 19129 (Oct. 25, 2018) 2018 Cal.P.U.C. Lexis 535, *55–*56.) These provisions, Lyft contends, are inconsistent with a definition of its business as providing rides. The trial court’s contrary conclusion was a clear error of law, we are told. To the extent defendants argue that the statutory provisions and the PUC’s regulatory decisions did not decide the issue now before us, we agree, although, to be sure, as the trial court correctly recognized, these rulemaking decisions may be given some limited weight in determining what defendants’ businesses actually entail. (Yamaha Corp. of America v. State Bd. of Equalization (1998) 19 Cal.4th 1, 7–8 [significance of agency decisionmaking as it bears on judicial construction of statutes, “[d]epending on the context, . . . may be helpful, enlightening, even convincing,” and it “may sometimes be of little worth”]; see New Cingular Wireless PCS, LLC v. Public Utilities Com. (2016) 246 Cal.App.4th 784, 808–810.) Lyft is emphatic that the trial

42 court misread the Public Utilities Code as construed in the 2013 PUC Decision when the PUC created the regulatory category of TNC’s. But we do not see deference to the PUC’s regulatory decisionmaking as central to the trial court’s reasoning—nor is it central to ours—for the court went on to consider and reject on the merits defendants’ argument that they merely operated as multi-sided platforms rather than providing transportation services. Thus, we are not persuaded that the court’s limited reliance on the Public Utilities Code as construed by the PUC makes much difference here.19 On a related note, Lyft takes the position that statements in the PUC’s rulemaking orders show that defendants are not in the transportation business. To a large extent, Lyft takes these comments out of context. For instance, Lyft points to the PUC’s comment in the 2013 PUC Decision that Uber is “the means by which the transportation service is arranged.” (2013 PUC Decision, supra, 2013 Cal.P.U.C. Lexis 504, at p. *17, italics added.) But this statement was made in the course of rejecting the assertion that TNC’s “are nothing more than an application on smart phones, rather than part of the transportation industry.” (Ibid.) Similarly, the PUC’s statement that TNC permits are granted only to “companies utilizing smart phone technology applications to facilitate transportation of passengers in the driver’s personal vehicle” (id. p. *39, italics added) refers to the means used to connect drivers and passengers, not the nature of a TNC’s business. Lyft

19Going further afield, Lyft draws our attention to an unemployment insurance decision of the Wisconsin Labor and Industry Review Commission finding a Lyft driver was not its employee. (Ebenoe v. Lyft Inc. (Jan. 20, 2017) Hearing No. 16002409MD.) We do not find this decision useful, as it is grounded in statutory definitions of employee, transportation network company, and “participating driver” that do not mirror California law. (See Wis. Stat. §§ 108.02(12), 440.40(3) & (6); compare Lab. Code, § 2775, Pub. Util. Code, § 5431, subds. (a) & (c).)

43 asserts that in 2018, the PUC held that Uber was subject to regulation as a transportation charter party carrier (TCP) because it provides a technology platform for “ ‘ “independent” providers of transportation services to connect with riders.’ ” (Cal.P.U.C. Dec. No. 18-04-005 (April 26, 2018) 2018 Cal.P.U.C. Lexis 180, *35, italics added.) But the PUC there was merely summarizing Uber’s own position, and it went on to conclude that despite Uber’s claim that the drivers were “independent service providers,” “it is Uber . . . that is engaged in running the TCP operation.” (Id. at p. *37.) At oral argument, Lyft argued the PUC said in the same decision that Uber was simply a “catalyst” for transportation; but the PUC made this statement in the course of rejecting the argument that Uber did not “provide[] prearranged transportation services” and noting that it had previously rejected the claim that Uber was “simply a technology company engaged in the business of developing and licensing software.” (Id. at pp. *21–*23; see id. at p. *24 [“Rather than behaving as a passive technology company, Uber is actively involved in facilitating . . . transportation services”].) These decisions do not establish that providing transportation falls outside the scope of defendants’ businesses. Uber also points out that some of the practices discussed above—such as ensuring that drivers are properly licensed and insured, ensuring their vehicles are inspected, checking their background and driving history, suspending drivers who use intoxicating substances, carrying out driver training programs, and reporting the number of rides requested and accepted in each zip code—are required by either the governing statutes or the PUC. (See Pub. Util. Code, §§ 5444, 5445.2, 5445.3; 2013 PUC Decision, supra, 2013 Cal.P.U.C. Lexis 504, at pp. *41–*51; Cal.P.U.C. Dec. No. 16-04-041, (April 21, 2016) 2016 Cal.P.U.C. Lexis 208, *1–*5, *87–*93.) Uber argues

44 that its compliance with these public safety standards does not make the drivers its employees. (See Linton, supra, 15 Cal.App.5th at p. 1223 [“A putative employer does not exercise any degree of control merely by imposing requirements mandated by government regulation”]; but see Secci v. United Independent Taxi Drivers, Inc. (2017) 8 Cal.App.5th 846, 858–859 [rejecting argument that when taxi company exercises control over drivers to comply with public regulations, that activity cannot be considered in determining whether agency or employment relationship exists].) We agree that the fact defendants comply with legal requirements imposed on them is of lesser importance than their affirmative business choices, but on the other hand it does not follow that legally compelled practices are irrelevant to an assessment of the scope of their normal course of business. Viewing the conduct of defendants’ businesses as a whole, we conclude the trial court properly found—based on prong B alone—that there is more than a reasonable probability the People will prevail on the merits at trial. (City and County of San Francisco v. Evankovich (1977) 69 Cal.App.3d 41, 54 [“The substantial evidence rule applies to preliminary injunctions, as well as the additional rule requiring us, when weighing the question of a trial court’s exercise of discretion in granting a preliminary injunction, to view the facts most favorably to the court’s disposition”].) We emphasize that our conclusion here is not a final resolution of the merits; that is a matter ultimately to be determined after a full trial. (Gallo, supra, 14 Cal.4th at p. 1109.) At this stage, the trial court’s task was essentially a predictive one. Balanced against the relative harms of a preliminary injunction issuing— harms that we consider below—we think the court correctly gave the most weight to the People’s almost “inevitable success on the merits.” (Winter, supra, 555 U.S. at p. 53 (dis. opn. of Ginsburg, J.).) We assess the merits

45 similarly on this record. At trial, defendants will face a presumption against them on all three prongs of the ABC test; an adverse decision on any one of the three prongs will result in an employment relationship being found; and, taking a generous view for defendants, their chances of prevailing on prong B alone may be characterized as daunting. Compared to the six-factor, fact-bound Borello test for independent contractor status—which can be very difficult for plaintiffs to meet at an early stage of litigation, short of a full-blown trial—the Dynamex court “create[d] a simpler, clearer test for determining whether the worker is an employee or an independent contractor,” one that “presumes a worker hired by an entity is an employee and places the burden on the hirer to establish that the worker is an independent contractor.” (Dynamex, supra, 4 Cal.5th at p. 951, fn. 20.) In applying this test, it is important to bear in mind the procedural posture of the order under review in Dynamex. The case arose on review of a class certification order, on a limited record far less robust than would have been the case upon review of a decision on a trial record. Our Supreme Court affirmed the class certification order as a matter of law, holding that the simplicity of the ABC test made it possible to decide at that early stage in the litigation that common questions were sufficiently predominant to warrant class treatment. The procedural posture here differs, but we take Dynamex as instructive in determining that the ABC test may be applied, and may be applied with confidence, in the context of a motion for interim injunctive relief. C. Grave Harm to Defendants and Balance of Relative Harms The next steps in the IT Corp. framework require the trial court to determine whether the defendant has shown it would suffer grave or irreparable harm from the issuance of the preliminary injunction and, if so, to

46 balance the relative actual harms to the parties, while taking into account the degree of certainty of the outcome on the merits. (IT Corp., supra, 35 Cal.3d at p. 72.) The Supreme Court has explained, “At this stage of the analysis, no hard and fast rule dictates which consideration must be accorded greater weight by the trial court. For example, if it appears fairly clear that the plaintiff will prevail on the merits, a trial court might legitimately decide that an injunction should issue even though the plaintiff is unable to prevail in a balancing of the probable harms. On the other hand, the harm which the defendant might suffer if an injunction were issued may so outweigh that which the plaintiff might suffer in the absence of an injunction that the injunction should be denied even though the plaintiff appears likely to prevail on the merits.” (Id. at pp. 72–73.) The goal is to minimize the harm that would be caused by an erroneous interim decision. (Id. at p. 73.) 1. Evidence of Harm if Injunction Granted or Denied Defendants submitted extensive evidence of the harm they claim would be caused by an erroneous preliminary injunction. Lyft’s director of data science, Christopher Sholley, testified that an injunction would require Lyft to change its business model. Currently, when drivers are logged into the driver app but not on a ride or on their way to pick up a rider, they may use their time as they wish, including running personal errands or using other platforms. If Lyft were required to compensate drivers for this time, it would need to find ways to control drivers’ time, for instance by having them work in scheduled shifts, at designated times and places, or for multiple hours at a time, in order to direct drivers’ work to times and places with the most demand for rides. Lyft might also need to prohibit drivers from using other platforms, such as Uber, while logged into the app or from unilaterally rejecting or cancelling rides.

47 Sholley testified that as a result of these changes, Lyft would probably need to reduce the number of drivers who use its platform, rather than allowing an unlimited number of drivers to use its app. Converting drivers to employee status would impose additional costs on Lyft and might force it to respond by increasing prices, costing Lyft goodwill with riders. Areas with lower demand, such as outlying suburbs or smaller towns, might have fewer drivers available. Drivers whose relationship with Lyft was severed as a result of these events might look for other opportunities and not return to Lyft if the injunction were later lifted. Lyft also submitted expert testimony that it would incur significant costs in converting its system to treat drivers as employees, including substantial changes to its “organizational structure, hiring processes, software tools and management systems, and company culture.” For instance, Lyft would have to spend extensive time ensuring each driver filled out the necessary paperwork and verifying their eligibility to work, and it would need employees to supervise the drivers, additional human resources support staff, additional accounting staff, and new recruiting staff. It would need to develop expanded infrastructure technology to run its payroll system for an influx of new employees. Uber also submitted evidence it would incur substantial costs if it were required to treat drivers as employees. Uber would incur unrecoverable costs such as hiring additional corporate staff to recruit and manage the expanded work force. An injunction would give Uber an economic incentive to reduce the number of drivers, enforce a fixed work schedule, and limit the number of hours each employee could drive in order to reduce overtime costs. A reduced number of drivers would mean fewer rides available during busy times, longer waiting times for rides, and higher prices. And Uber argues its drivers

48 would, if reclassified, lose access to certain federal benefits for self-employed workers during the current COVID-19 pandemic. Both Lyft and Uber submitted declarations by drivers, who variously attested that they use the Lyft and Uber apps to make money on the side, at their own convenience, and that they value or need the ability to set their own work hours. Some use the apps between other work commitments or when they have free time. Some are unable to work a regular schedule, either because of health conditions of their own or their responsibilities for children or ailing family members. Some use the apps to make extra money to pay unexpected bills.20 The People also submitted declarations from Uber and Lyft drivers, some of whom drive long hours but are not paid for overtime work or for rest breaks. They do not receive sick leave or health insurance coverage. Some drivers stated they had difficulty obtaining unemployment benefits during the COVID-19 pandemic because defendants did not report their earnings. Drivers pay for their own vehicles, insurance, gas, inspections (in the case of Uber), and cell phone service. One driver testified that because of the low

20 See, e.g., Declaration of Jesus Sauceda in Support of Uber’s Opposition to Motion for Preliminary Injunction (July 20, 2020) [“I use the Uber app because of the flexibility to drive when and where I want. . . . My independence is important to me. I don’t want anything to change”]; Declaration of Gerhard Kleindl in Support of Lyft, Inc.’s Opposition to Plaintiff’s Motion for Preliminary Injunction (July 14, 2020) [“I can’t imagine having this level of flexibility as an employee, and I’m grateful that this arrangement allowed me to help provide for me and my wife, including while I was still recovering from surgery”]; Declaration of Michael Delfino in Support of Lyft, Inc.’s Opposition to Plaintiff’s Motion for Preliminary Injunction (February 27, 2020) [“I operate my driving business around my work schedule at the steel company. . . . I adjust when I drive based on my school schedule and study needs as well”].

49 rates paid by Uber and Lyft and her responsibility for expenses such as gas, insurance, and maintaining her vehicle, it was “hard to count on driving for Uber or Lyft to make ends meet.”21 2. The Trial Court’s Findings The trial court found substantial public harm would result in the absence of an injunction, looking first to the Supreme Court’s decision in Dynamex, which explained the significance of how a worker is classified: A worker who is properly classified as an employee obtains the protection of applicable labor laws and regulations, including payment of Social Security and payroll taxes, unemployment insurance taxes, workers’ compensation insurance, and enactments governing wages, hours, and working conditions. An independent contractor, on the other hand, gains none of the numerous labor law benefits, and the public may be required to assume additional financial burdens. (Dynamex, supra, 4 Cal.5th at pp. 912–913.) By misclassifying employees as independent contractors, a business may obtain a competitive advantage over others that classify their workers properly; moreover, misclassification “is a very serious problem, depriving federal and

21 Declaration of Linda Salamone in Support of the People’s Motion for Preliminary Injunction (June 11, 2020); see also, e.g., Declaration of Michael Dominguez in Support of the People’s Motion for Preliminary Injunction (June 17, 2020) [“I have been unemployed ever since I stopped driving for Uber in March. . . . [¶] Not receiving unemployment benefits has been extremely hard on me. I’m 59 years old and I’ve had to go down to Social Services to get on food stamps. I can’t pay my rent, my insurance, or the registration for my vehicle. I haven’t been able to pay my utility bills either.”]; Declaration of Jose Funes in Support of the People’s Motion for Preliminary Injunction (June 16, 2020) [“Uber’s control over my ability to make money as a driver has had severe economic consequences to me and my family. . . . [¶] Not only do I have insufficient money to rent my own place, I have no money for health insurance and therefore cannot go to the doctor when I’m sick. Sometimes I even struggle to buy food.”].

50 state governments of billions of dollars in tax revenue and millions of workers of the labor law protections to which they are entitled.” (Id. at p. 913.) In enacting AB 5, the Legislature declared its intent to “ensure workers who are currently exploited by being misclassified as independent contractors instead of recognized as employees have the basic rights and protections they deserve under the law, including a minimum wage, workers’ compensation . . . , unemployment insurance, paid sick leave, and paid family leave.” (Stats. 2019, ch. 296, § 1(e).) The Legislature stated it intended to “restore[] these important protections to potentially several million workers who have been denied these basic workplace rights that all employees are entitled to under the law.” (Ibid.) The trial court also noted the declarations of individual drivers attesting to the “precariousness of their financial existence, which is directly attributable to Defendants’ refusal to classify and treat them as employees entitled to protection under California law.” The court concluded defendants had not shown they would suffer grave or irreparable harm from the issuance of a preliminary injunction: although implementation would require defendants to change the nature of their business practices in significant ways, those costs were fundamentally financial. The court recognized the injunction’s adverse effect on some drivers who desired the flexibility of defendants’ current business model, but noted first, that those drivers who worked for only a small number of hours a week would suffer correspondingly minor effects, and second, that during the current pandemic, many drivers were working less or not at all, further reducing the interim consequences of an injunction. And, even assuming defendants’ showing amounted to grave or irreparable harm, the court concluded it did not outweigh the harm in the absence of an injunction.

51 3. Analysis As a threshold matter, the parties dispute whether Uber and Lyft have demonstrated grave or irreparable harm. Defendants argue this type of harm is found in the burden of restructuring their businesses; the loss of goodwill they will suffer from terminating their contractual relationships with many drivers, a prospect they claim is inevitable; the lost income of drivers who are not hired as employees or who will not be able to work on a fixed schedule; and the community’s loss of transportation options. (See American Trucking Ass’ns v. City of Los Angeles (9th Cir. 2009) 559 F.3d 1046, 1058 [irreparable harm in incurring large costs in restructuring business and losing customer goodwill].) They feature declarations—from Ron Hamilton, a human resources and business operations expert for Lyft, and from Brad Rosenthal, director of strategic operational initiatives for Uber—attesting to the need to add extensive internal management systems, including vast human resources and related information technology services, to support an employee workforce many times the size they have now. These systems, defendants point out, cannot be added overnight. The People counter, correctly, that a party suffers no grave or irreparable harm by being prohibited from violating the law (see People ex rel. Reisig v. Acuna (2010) 182 Cal.App.4th 866, 882 [no harm from restrictions on activities that constitute public nuisance]) and that defendants’ financial burdens do not rise to the level of irreparable harm (see IT Corp. supra, 35 Cal.3d at p. 75 [although party would suffer substantial loss of waste disposal and transportation revenues, no showing of grave or irreparable injury because it could still process wastes]). Moreover, the People contend, again correctly, nothing in the preliminary injunction prevents defendants from allowing drivers to maintain their flexibility rather

52 than assigning rigid shifts. (See Dynamex, supra, 4 Cal.5th at p. 961, fn. 28 [business may allow workers to set own hours and to accept or decline a particular assignment while treating them as employees for purposes of wage order]; Cunningham, supra, 2020 U.S. Dist. Lexis 90333, at p. *33 [describing as “red herring” argument that classifying drivers as employees was inconsistent with flexible schedules].) To the extent defendants base their claim of harm on the plea that the necessary changes cannot be made “on the flick of a switch,” the trial court correctly observed that defendants have had more than two years—since Dynamex was decided—to make the necessary adjustments. The facts in Dynamex, though they arose in a low-tech setting compared to what we have here, bear a number of similarities to those in this case. One could not reasonably read that opinion as of April 2018 and not come away with an expectation that, without legislative relief, the foundation of defendants’ ride- sharing business model, to the extent it was based on treating drivers as independent contractors, would highly likely have to change. The passage of AB 5 in October 2019, obviously, should have heightened the importance of urgent contingency planning for an employment-based model. And when the trial court ordered that change in August 2020, we gave defendants an additional reprieve, putting a stay in place during the pendency of this appeal, subject to the submission of sworn statements from their chief executive officers confirming that implementation plans have been made so the companies will be able to comply if we affirm the injunction and if the governing law is not changed by a proposition on the upcoming November 3 ballot. Given the time that has elapsed since Dynamex was decided, the idea that it was unreasonable for the trial court to expect rapid compliance is untenable.

53 Nonetheless, for purposes of our analysis—bearing in mind both the evidence of disruption to defendants’ businesses and the fact that we must consider the potential harm caused by an erroneous interim decision (IT Corp., supra, 35 Cal.3d at p. 73)—we shall assume that if the injunction were ultimately determined to have been wrongly entered, the harm to defendants could fairly be considered grave or irreparable. Even with this assumption, however, the trial court did not abuse its discretion in concluding the balance of harms favored the issuance of a preliminary injunction. In the end, it matters—and it matters in a profoundly important way to the bottom-line discretionary calculus—that the Legislature specifically authorized the government to seek injunctive relief as a means of enforcing AB 5, and that IT Corp. gives the government the benefit of a presumption when it champions the public interest in an enforcement action invoking that authority. Uber and Lyft disagree, contending that the trial court effectively made the IT Corp. presumption irrebuttable. They are incorrect. At the last step of the IT Corp. analytical framework, the court treated neither side’s showing as conclusive, leaving it free to strike the appropriate balance in its considered discretion. What defendants overlook is that, in the final analysis under IT Corp., “if it appears fairly clear that the plaintiff will prevail on the merits”—as it does in this case—“a trial court might legitimately decide that an injunction should issue even though the plaintiff is unable to prevail in a balancing of the probable harms.” (IT Corp., supra, 35 Cal.3d at pp. 72–73, italics added.) We do not underestimate the difficulty of the trial court’s task at this stage of the IT Corp. analysis, but so long as it properly understood its discretion as a legal matter, as we believe it did, we must defer to its exercise of discretion so long as the choice it made was within the permissible range of options before

54 it. (See Cahill v. San Diego Gas & Electric Co. (2011) 194 Cal.App.4th 939, 957 (Cahill) [“there is no abuse of discretion requiring reversal if there exists a reasonable or fairly debatable justification under the law for the trial court’s decision or, alternatively stated, if that decision falls within the permissible range of options set by the applicable legal criteria”].) The order under review here meets that test. On this record, there were compelling policy arguments favoring both sides of the choice the court faced. Some of the competing arguments now advanced in favor of, and against, the order it ultimately entered even come from within the same constituencies of third parties potentially affected.22 In the end, the trial court had a reserve of discretionary power under IT Corp. to choose between the contending

22 Compare Amicus Curiae Brief of Communities-of-Color Organizations, at pp. 12–15 (contending injunction will harm drivers of color by depriving them of income-earning opportunities and will harm communities of color by depriving them of a critically needed mode of transport in areas lacking public transportation options, in ways exacerbating harms caused by COVID-19 pandemic) with Amicus Curiae Brief of National Employment Law Project et al., at pp. 23–31 (defendants “do not offer ‘opportunities’ to marginalized workers and communities of color [and t]heir misclassification model deepens the desperation of workers who have been excluded from stable employment, with Black and Latino workers made to bear the brunt”); and compare Amicus Brief of IWLC, at pp. 4–13 (“[t]he gig economy, and the independent contractor model upon which it relies, is critical for working women” because it provides employment opportunities for “tens of thousands of California workers, many of them women, at precisely a time where flexible work arrangements, and the ability to earn a living in the manner one chooses, is more critical than ever”) with People’s Response to Amici Supporting Uber Technologies, Inc. and Lyft, Inc., at p. 22 (“[i]n arguments strongly reminiscent of those advanced by Uber and Lyft today,” the early twentieth-century garment industry argued that “women workers who knit part-time, were paid on a piece-rate basis, and were not treated as employees” similarly “fell outside of the protections of employment law because the[y] . . . could knit at home, at times of their own choosing, and often worked for more than one person”).

55 positions, with inevitable trade-offs entailed either way. Uber argues that “equity demands that courts ‘take into account the public interest’ when assessing the propriety of injunctive relief.” “Indeed,” it points out, “this is the most fundamental requirement in exercising equitable authority.” We agree. As Uber observes, “Courts sitting in equity have a ‘duty to arrive at a just solution’ (Barquis v. Merchants Collection Assn. (1972) 7 Cal.3d 94, 112),” but that is precisely what the trial court did here in discharging its duty to decide at the final step of its IT Corp. analysis. Taking a slightly different tack focused on what they claim is the meager evidentiary showing of actual irreparable harm made by the People, defendants charge that the trial court improperly relied on general statements in Dynamex and AB 5, rather than evidence, in reaching its result. (See Herb Reed Enterprises v. Florida Entment. Mgmt. (9th Cir. 2013) 736 F.3d 1239, 1250 [trial court relied on “platitudes rather than evidence” in enjoining trademark infringement]; see also People v. Pacific Land Research Co. (1977) 20 Cal.3d 10, 21 [party seeking injunction must make showing by admissible evidence].) They contend the Legislature has not made—and could not properly make—findings applicable to this particular case (see Communist Party v. Peek (1942) 20 Cal.2d 536, 548 [“it is not the function of the Legislature to determine whether a statute declaring a general policy has been violated in a particular case”]; accord, Mack v. State Board of Education (1964) 224 Cal.App.2d 370, 374–375), and they argue that the policy rationale behind AB 5 is insufficient to show that anyone in this case is suffering injury. We are unpersuaded by this line of argument. The trial court’s application of the balancing test at the final step of its IT Corp. analysis was grounded firmly in record evidence. It is true that the People cited various

56 articles and studies that were never offered into evidence or made the subject of a timely request for judicial notice.23 But reading the record as a whole, we are satisfied that, even assuming that all of the improper citations to such material were disregarded, there was still ample admissible evidence— including a great deal of undisputed evidence offered by Uber and Lyft themselves—from which the court could infer that irreparable harm to drivers on a broad scale is ongoing, and that immediate, pendente lite relief is warranted, despite the competing considerations put forward by defendants. (People v. Pacific Land Research Co., supra, 20 Cal.3d at p. 22 [agreeing that principal affidavits submitted by party seeking injunction against violation of Subdivided Lands Act should have been excluded on evidentiary grounds, but affirming grant of preliminary injunction because “strong circumstantial evidence” remained that was “sufficient to support an inference” of illegal lot subdivision].)

23 These materials, cited in the People’s brief in support of the preliminary injunction, included Benner et al., On-Demand and On the Edge: Ride-Hailing & Delivery Workers in San Francisco (May 5, 2020) U.C. Santa Cruz [finding majority of ride-hailing and delivery drivers rely on platform work as primary source of income and many struggle financially] [as of June 22, 2020]; Mishel, Uber and the Labor Market (May 15, 2018) Economic Policy Institute [finding that, after deducting fees and expenses, Uber driver compensation averages $11.77 an hour] [as of June 22, 2020]; Miller et al., Paid Sick Days and Health: Cost Savings from Reduced Emergency Department Visits (Nov. 14, 2011) Inst. for Women’s Policy Research, at p. 7 [finding workers without paid sick days are more likely to delay needed medical care, which can turn minor health problems into more serious and costly ones]; Jacobs et al., What Would Uber and Lyft Owe to the State Unemployment Insurance Fund? (May 7, 2020) UC Berkeley Labor Center [finding Uber and Lyft would have paid $413 million into state Unemployment Insurance Fund between 2014 and 2019 had they treated workers as employees] [as of June 18, 2020].

57 It is, for example, undisputed that neither Uber nor Lyft offers any of the benefits the People allege they have illegally withheld from hundreds of thousands of ride-share drivers who utilize their apps. While some facts at the margins are sharply contested (such as the average hourly wage drivers earn, and the percentage of drivers who drive casually for only a few hours per week versus the percentage who drive more or less full time), the dispute here boils down to what inferences should be drawn from largely undisputed facts. As counsel for the People pointed out at oral argument, defendants’ detailed showing of their hundreds of thousands of drivers statewide, the size and scale of their respective operations, and the ripple effects on various third parties that they insist will flow from the trial court’s injunction— supported by arguments couched in terms that sound very much like defendants are saying that they are “too big to be enjoined”—ultimately cut against them by confirming the extent of the harm being inflicted by virtue of their undisputed failure to provide the benefits of employment to many thousands of ride-share drivers across the state. Defendants portray the record as if it were wholly one-sided; as if they, and they alone, came forward with evidence going to the issue of irreparable harm; and as if the trial court had no defensible choice but to deny interim relief in the face of their showing. We do not read the record that way. The People were not required to counter the array of impressively credentialed experts marshalled by defendants with experts of their own. They submitted 10 declarations from individuals who drive for Uber or Lyft attesting to the hardships they are currently suffering. The trial court was entitled to credit these declarations and draw reasonable inferences from them in light of defendants’ own evidence of how all drivers sign standard form contracts and all drivers are treated in standardized ways, much as a court would do in a

58 class action when deciding whether common issues of fact and law predominate. We cannot improve upon what the trial court said about the harm to drivers it inferred from the evidence before it. “[T]hese harms are not mere abstractions; they represent real harms to real working people”— consisting for instance of receiving low pay for long hours, having no overtime pay, breaks, health insurance, or sick leave, and being forced to pay business expenses. Defendants invited the court to draw a different set of inferences, to look at them as mere purveyors of the software on which their platforms operate, and to rely on their 32 driver declarations as more accurately describing the interests and preferences of most drivers than the People’s 10 driver declarations. Defendants, on the strength of their driver declarations, insist that most Uber and Lyft drivers do not wish to be employees and are not interested in employment benefits. But the differences in the various driver perspectives offered on this issue have limited bearing here. The governing ABC test is not decided by plebiscite. And if there is a segment of drivers—even a large one—who do not need, wish to have, or even understand they are entitled to employment benefits, that does not strip others of rights the People seek to ensure may be claimed by all. What matters for substantial evidence purposes, at this stage, on appeal, is that there is competent evidence in the record supporting the People’s showing of irreparable harm. (City and County of San Francisco v. Evankovich, supra, 69 Cal.App.3d at p. 54 [principal declaration submitted by union defendants in opposition to preliminary injunction, controverting facts in plaintiff city’s verified amended complaint, did not require reversal of injunction since, “viewing the facts most favorably to the court’s disposition . . . , the court did not abuse its discretion, despite the apparent conflict”].)

59 Arguing that the record is bereft of any actual evidence of irreparable harm, despite the driver declarations the People submitted, Uber relies on Sampson v. Murray (1974) 415 U.S. 61, for the proposition that “temporary loss of income, ultimately to be recovered, does not usually constitute irreparable injury.” (Id. at p. 90.) That case involved a terminated civil service probationary employee who claimed wrongful discharge and was granted reinstatement by preliminary injunction during the pendency of her lawsuit. (Id. at pp. 62–63.) The High Court reversed on the ground the plaintiff’s wage loss and difficulties finding new employment under the cloud of a termination did not constitute irreparable harm. (Id. at pp. 88–89.) We see the circumstances here quite differently, starting with the fact that this is a government enforcement action, not an individual employment case. In Sampson, not only was the plaintiff’s loss measurable in money, but backpay was “the usual, if not the exclusive” form of available relief for the statutory violation alleged there. (Id. at p. 91.) In a case brought by a private plaintiff seeking individual relief for loss of employment benefits, the general principle that money damages will supply a legally adequate remedy may often carry the day, as it did in Sampson, but here too we must bear in mind that this is an enforcement action by government plaintiffs invoking the public interest to forestall the need for a multiplicity of individual actions, under a statutory scheme that specifically authorizes them to seek injunctive relief. Similarly contending that lost employment income is not irreparable harm, Lyft offers a version of this argument that is somewhat more blunt than the one Uber advances. “Injunctions do not issue to order the payment of money,” Lyft contends. (Friedman v. Friedman (1993) 20 Cal.App.4th 876, 890 [“monetary loss does not constitute irreparable harm” unless the amounts are unrecoverable].) According to Lyft, “[c]ase after case has

60 specifically held that wage protections, expense reimbursement, and other similar alleged harms—however important they may be—are not irreparable because they can be remedied through later monetary relief.” (See, e.g., Lucas v. Bechtel Corp. (9th Cir. 1986) 800 F.2d 839, 847 [for nonpayment of “hourly wages in addition to travel and subsistence pay, the injury involves only monetary harm” and is “ ‘not usually sufficient to establish irreparable harm’ ”].) “Usually,” in private actions, as we note above, that is true, but generalized as Lyft would have it, we think this reading of California law tends to undervalue the importance of statutory wage and hour and other related protections extended to employees. California courts have long recognized, for example, that “ ‘wages are not ordinary debts . . . and that, because of the economic position of the average worker . . . it is essential to the public welfare that he receive his pay when it is due.’ ” (Smith v. Superior Court (2006) 39 Cal.4th 77, 82, quoting In re Trombley (1948) 31 Cal.2d 801, 809–810.) Unlike situations in which records are maintained and from which damages can be calculated, there is no comparable way to measure the failure to pay minimum wages accurately, much less overtime wages, or to provide meal and rest breaks, when no records exist for the time that drivers are not transporting passengers. By the same token, how can the failure to provide wage statements, sick leave and health benefits, unemployment insurance and training fund contributions, disability insurance, and workers’ compensation benefits, much less the impact on competitors impacted by defendants’ failure to comply with the law, be measured? They cannot be, and that is why interim injunctive relief at the request of the government was appropriate in this case. When violation of statutory workplace protections takes place on a massive scale, as alleged in this case, it causes

61 public harm over and above the private financial interest of any given individual. This is particularly true nowadays, with the diminished efficacy of private enforcement of workplace remedies due to the widespread adoption and ready enforceability of contractual arbitration clauses.24 In sum, our assessment is as follows. The trial court found that rectifying the various forms of irreparable harm shown by the People more strongly serves the public interest than protecting Uber, Lyft, their shareholders, and all of those who have come to rely on the advantages of online ride-sharing delivered by a business model that does not provide employment benefits to drivers. Under IT Corp. the court determined that the balance of interim harms tips in favor of the People under the sliding scale analysis that must inform any equitable decision of the kind presented here. In striking that balance, the court relied on more than abstract expressions of policy, untethered to the facts before it. And it properly considered the harm shown by the record, in light both of those policies and of its determination that the People showed a reasonable probability—indeed, an “overwhelming likelihood”—of prevailing at trial. Accordingly, we conclude that the trial court correctly applied the law and that the choice it made to grant preliminary injunctive relief was “within the permissible range

24 Notably, by way of counterattack to the People’s motion for a preliminary injunction, both defendants moved to compel arbitration of the People’s claim for restitution, arguing that although none of the government plaintiffs who brought this suit is a party to an arbitration agreement with them, the People are acting on behalf of drivers who are bound by such agreements. And as a companion to that motion, defendants moved for an indefinite stay of this action pending completion of their proposed arbitration with the People, various ongoing individual arbitrations with drivers, and other related litigation. The trial court denied the stay motions and deferred ruling on the motions to compel arbitration.

62 of options set by the applicable legal criteria.” (Cahill, supra, 194 Cal.App.4th at p. 957.) There was no abuse of discretion here. D. The Injunction Is Not Vague or Overbroad In fashioning a remedy, a court should “strive for the least disruptive remedy adequate to its legitimate task” and tailor it to the harm at issue. (Butt, supra, 4 Cal.4th at pp. 695–696.) And an injunction against legitimate business activities “should go no further than is absolutely necessary to protect the lawful rights of the parties seeking such injunction.” (People v. Mason (1981) 124 Cal.App.3d 348, 354.) Defendants contend the injunction in this case violates these principles. According to Lyft, the injunction was improper without a showing that all of its drivers are suffering irreparable harm. Lyft relies upon O’Connell v. Superior Court (2006) 141 Cal.App.4th 1452 (O’Connell), where this division found overbroad an injunction, based on an equal protection claim, restraining the California Board of Education from denying diplomas to high school students who had not passed both portions of an exit exam, because the injunction “affected every high school in the state regardless of circumstances” and regardless of how many students were “actually educationally disadvantaged.” (Id. at pp. 1479–1480.) Lyft also points to Stormans, Inc. v. Selecky (9th Cir. 2009) 586 F.3d 1109 (Stormans), in which a trial court issued an injunction restraining enforcement of regulations requiring pharmacists to dispense Plan B emergency contraceptives without limiting the injunction to the plaintiffs before the court, who asserted religious objections. (Id. at p. 1118.) Without such limitation, the Court of Appeals concluded, the injunction was “fatally overbroad because it is not limited to the only type of refusal that may be protected by the First Amendment—one based on religious belief.” (Id. at p. 1141.)

63 These cases do not assist defendants. Stormans is a First Amendment case. Under conventional constitutional overbreadth principles, narrow tailoring is compelled where a court-ordered restraint indiscriminately trenches on constitutionally protected conduct while restraining constitutionally unprotected conduct. There is no such issue in this case. Nor is this a situation similar to the one in O’Connell, where clearly legal conduct and protections for people who are not entitled to such protection were brought within the scope of an umbrella injunctive order that, self- evidently, could have been crafted in a more targeted fashion. The court held that the plaintiffs in that case failed to bear their burden of showing that the injunction they sought was appropriately fitted to the harm they alleged. (O’Connell, supra, 141 Cal.App.4th at p. 1481.) That is not what happened here. The People proposed, and the court adopted, an injunction “ ‘mould[ed] . . . to the necessities of the particular case’ ” (Winter, supra, 555 U.S. at p. 51 (dis. opn. of Ginsburg, J.)), and the necessities of the case called for a broadly framed injunction. By proposing a form of injunction fitted to the scale of the ongoing violations of law shown by the evidence—a proposal evidently put forward on the assumption the trial court would draw inferences from the record favoring them as movant—the People carried their initial burden of requesting appropriately tailored relief. But it is important to bear in mind that the court did not simply rubber-stamp the relief the People sought. To address the issue of potential overbreadth, the court issued an order prior to the August 6, 2020 hearing on the preliminary injunction, putting the following question to the parties: “How could an injunction, if granted, be framed so as to minimize the claimed harm to Defendants’ businesses and participating drivers?” In response, defendants stood mute. They devoted all

64 of their energies to arguing, instead, that a preliminary injunction should not issue at all. Indeed, after this suit was brought, the first public show of seriousness by these defendants about the likely need to make significant changes to their mode of doing business came in response to this court’s order directing each of them to submit a declaration from its chief executive officer, as a condition of our issuance of a stay pending appeal, attesting to the fact that implementation plans exist should we affirm the trial court’s injunctive order. Defendants are in no position now to claim it was an abuse of discretion for the trial court to overlook the possibility of a more narrowly tailored order, when they declined to cooperate in identifying what such an order might look like.25 The strategic posture they assumed on this issue no doubt carried high risks, but those were the risks they chose. Here, on appeal, defendants continue to say nothing specific about how an injunction pending trial might be framed to minimize interim harm to them or others. Even assuming defendants preserved their ability to advance overbreadth objections on appeal despite their default on the issue below, we

25 It is not difficult to imagine, for example, an alternative proposed form of injunction that, at least on an interim basis until trial, would have covered only a subset of self-identified drivers who are users of either Uber’s or Lyft’s ride-sharing app, who rely on that app as a sole source of income, who do not drive for any ride-sharing competitor, and who drive some minimum threshold number of hours per week, just as the class defined in Dynamex “consisted only of individual Dynamex drivers who had returned complete and timely questionnaires and who personally performed delivery services for Dynamex but did not employ other drivers or perform delivery services for another delivery company or for the driver’s own delivery business.” (Dynamex, supra, 4 Cal.5th at p. 920.) We do not know and cannot speculate whether such an alternative proposal would have been acceptable to the People, would have been feasible in defendants’ eyes, or, most importantly, would have been adequate to the task of preventing the interim harm the trial court sought to prevent.

65 see nothing in the record that compels a conclusion that only a subset of the drivers who use defendants’ apps are entitled to the protections of employment status under Dynamex and AB 5. The closest either defendant comes to articulating a specific basis for crafting the injunction more narrowly is Lyft’s suggestion that some drivers, even if ultimately found to be employees, will not be harmed by continuation of their current business practices and will in fact suffer harm if Lyft elects to sever its contractual relationship with them. But because Lyft fails to tie this argument about potential driver harm to a suggested narrower framing of the injunction, we conclude that it supplies no basis to find that the trial court abused its discretion by not limiting the injunction to only some of its drivers. (See IT Corp., supra, 35 Cal.3d at p. 69 [preliminary injunction rests in sound discretion of trial court]; see also Kruse, supra, 177 Cal.App.4th at p. 1180 [scope of injunction is within trial court’s discretion].) Uber puts a slightly different twist on this issue, arguing that the injunction goes further than permissible because it does not inform Uber what additional changes to its business practices would suffice to allow it to treat its drivers as independent contractors in a way consistent with AB 5, thus exposing it to a “contempt trap.” And, both Uber and Lyft contend, the second part of the injunction—restraining them from “violating any provisions of the Labor Code, the Unemployment Insurance Code, and the wage orders of the Industrial Welfare Commission with regard to their Drivers”—is an impermissible “obey the law” injunction that likewise exposes them unfairly to contempt because it gives no guidance on how to comply. In support of this argument, Uber cites several federal authorities. (U.S. v. Dixon (1993) 509 U.S. 688, 695 (Dixon); National Labor Relations Board v. Express Pub. Co. (1941) 312 U.S. 426, 435 (N.L.R.B.); Del Webb Communities,

66 Inc. v. Partington (9th Cir. 2011) 652 F.3d 1145, 1150; Hughey v. JMS Development Corp. (11th Cir. 1996) 78 F.3d 1523, 1531–1532.) Those cases are readily distinguishable. In Dixon, supra, 509 U.S. 688, the High Court merely noted a general rule at common law that injunctions “would not issue to forbid infringement of criminal or civil laws, in the absence of some separate injury to private interest” (id. at p. 695) as background to its analysis of a novel Double Jeopardy challenge to criminal prosecutions based on conduct that had previously led to criminal contempt proceedings. In N.L.R.B., supra, 312 U.S. 426, the Court held that the NLRB exceeded its authority because, after finding that a company had violated its duty under the National Labor Relations Act (the Act) to bargain in good faith with a union, the Board ordered the company “not to violate ‘in any manner’ the duties imposed on the employer by the statute.” (Id. at p. 432.) The high court rejected the Board’s contention that, because the employer had violated one provision of the Act, the Board was “not only free to restrain violations like those . . . committed, but any other unfair labor practices of any kind which likewise infringe any of the rights enumerated in [the Act], however unrelated those practices may be to the acts of respondent which alone emerged in course of the hearing.” (Id. at pp. 432–433.) Here, by obvious contrast, the injunction includes no restraints on the commission of unlawful acts “dissociated from those which a defendant has committed.” (Id. at p. 436.) The federal Court of Appeals decisions Uber cites, meanwhile, involved injunctive orders—or overbroad portions of such orders—that left the enjoined parties so wholly in the dark as to what was prohibited that the challenged orders violated Federal Rule of Civil Procedure 65(d)(1), which requires that an injunction “state its terms specifically and describe in reasonable detail . . . the act or acts restrained or required.” (Del Webb

67 Communities, Inc. v. Partington, supra, 652 F.3d at p. 1150; Hughey v. JMS Dev. Corp., supra, 78 F.3d at pp. 1531–1532.) In any event, we have the benefit of California case law on this point— case law that provides more specific guidance than the federal precedent on which Uber relies. At bottom, the governing test rests on the due process principle of fair notice. We ask whether the directive at issue is set forth “ ‘in terms so vague that men of common intelligence must necessarily guess at its meaning and differ as to its application.’ ” (In re Berry (1968) 68 Cal.2d 137, 156.) And we do not pursue the inquiry in the abstract. To be considered unconstitutionally vague, an injunction must suffer from vagueness in all its applications (Gallo, supra, 14 Cal.4th at p. 1116), since “[a] contextual application of otherwise unqualified legal language may supply the clue to a law’s meaning, giving facially standardless language a constitutionally sufficient concreteness.” (Ibid.) Taken in context, the conduct sought to be restrained—continued misclassification of drivers—is specifically identified on the record presented here. We do not demand the detail of an engineer’s instruction manual, only that the injunction provide “ ‘reasonable specificity.’ ” (Gallo, supra, 14 Cal.4th at p. 1117.) This injunction passes that test. It is specific to the Labor Code provisions, Unemployment Insurance Code provisions, and Wage Orders with which these defendants must comply, and it is directed to the drivers who are the subject of this action. If, in altering its policies, Uber is unsure what it will take either to convert its drivers to employment, or to modify its business sufficiently to make its drivers genuinely “independent” and properly susceptible to classification as independent contractors, it can always petition the court for modification of the injunction. (Code Civ. Proc., § 533; see Donovan v. Sureway Cleaners (9th Cir. 1981) 656 F.2d 1368, 1373

68 [enforcing injunction under the Fair Labor Standards Act against an employer who failed to seek clarification and, after making “superficial” changes to its business model, continued to misclassify its workers].) As for overbreadth, City of Redlands v. County of San Bernardino (2002) 96 Cal.App.4th 398, is the most instructive case. At issue there was an injunction issued in a writ proceeding under the California Environmental Quality Act (Pub. Res. Code., § 21000 et seq.; CEQA) enjoining the County of San Bernardino from readopting an invalid amendment to its general plan, “ ‘or any similar amendment(s),’ ” without first “ ‘preparing and considering an [environmental impact report] and fully complying with [CEQA].’ ” (Id. at p. 415.) The court rejected the county’s arguments that the injunction went beyond what was at issue in the lawsuit and that it was an impermissible “obey the law” injunction. (Id. at pp. 415–416.) “While a court may not issue a broad injunction to simply obey the law, thereby subjecting a person to contempt proceedings for committing at any time in the future some new violation unrelated to the original allegations, the court is entitled to restrain the person from committing similar or related unlawful activity.” (Id. at p. 416.) The injunction before us does no more than that. Defendants insist it does, describing the injunction here as “radical” and “unprecedented.” But these adjectives perhaps say more about the reach of modern technology and the scale of today’s technology-driven commerce than they do about the order itself. Although the business context may be relatively new, we conclude that the injunction was properly issued in accordance with enduring principles of equity. It is broad in scope, no doubt, but so too is the scale of the alleged violations.

69 III. DISPOSITION The August 10, 2020 order is affirmed. The stay issued on August 20, 2020, shall expire 30 days after issuance of the remittitur in this appeal.

70 _________________________ STREETER, J.

WE CONCUR:

_________________________ POLLAK, P. J.

_________________________ BROWN, J.

People v. Uber Technologies, Inc. et al. (A160701, A160706)

71 Trial Court: City & County of San Francisco

Trial Judge: Hon. Ethan P. Schulman

Counsel: Keker, Van Nest & Peters, Christa M. Anderson, Rachael E. Meny, R. James Slaughter; Munger, Tolles & Olson, Rohit K. Singla, Miriam Kim, Justin P. Raphael, Jeffrey Y. Wu for Defendant and Appellant Lyft, Inc.

Gibson, Dunn & Crutcher, Theodore J. Boutrous, Jr., Theane Evangelis, Blaine H. Evanson, Heather L. Richardson for Defendant and Appellant Uber Technologies, Inc.

Crowell & Moring, A. Marisa Chun, Kayvan Ghaffari, Alice Hall-Partyka for Bay Area Council, Earth Sparks, Internet Association, Silicon Valley Leadership Group, and Technet as Amici Curiae on behalf of Defendants and Appellants Lyft, Inc. and Uber Technologies, Inc.

Horvitz & Levy, Jeremy B. Rosen, Felix Shafir, Steven S. Fleischman for Chamber of Commerce of the United States of America, California Chamber of Commerce, National Retail Federation, and HR Policy Association as Amici Curiae on behalf of Defendants and Appellants Lyft, Inc. and Uber Technologies, Inc.

Willenken, Amelia L. B. Sargent, Kenneth M. Trujillo-Jamison for California Asian Pacific Chamber of Commerce, California Hispanic Chambers of Commerce, California State National Action Network, CA-NAACP State Conference, Los Angeles Metropolitan Churches, Los Angeles Urban League, National Action Network Sacramento Chapter Inc., National Asian American Coalition, National Black Chamber of Commerce, National Diversity Coalition, National Hispanic Council on Aging, National Newspaper Publishers Association, and Southern Christian Leadership Conference of Southern California (“Communities-of-Color Organizations”) as Amici Curiae on behalf of Defendants and Appellants Lyft, Inc. and Uber Technologies, Inc.

72 Durie Tangri, Benjamin B. Au, Raghav R. Krishnapriyan for Denise Alvarado, Tony Do, Mimi Fan, Victoria Broussard, Jim Pyatt, Daniel Farris, Robert Prather, Howard Tanner, and Independent Drivers Association of California as Amici Curiae on behalf of Defendants and Appellants Lyft, Inc. and Uber Technologies, Inc.

Independent Women’s Law Center, Jennifer C. Braceras; Littler Mendelson, Bruce J. Sarchet, Michael J. Lotito for Independent Women’s Law Center as Amicus Curiae on behalf of Defendants and Appellants Lyft, Inc. and Uber Technologies, Inc.

Eimer Stahl, Robert Dunn, John D. Tripoli for Mothers Against Drunk Driving and the California State Sheriffs’ Association as Amici Curiae on behalf of Defendants and Appellants Lyft, Inc. and Uber Technologies, Inc.

Xavier Becerra, Attorney General, Michael L. Newman, Senior Assistant Attorney General, Satoshi Yanai, Supervising Deputy Attorney General, Minsu D. Longiaru, Marisa Hernández-Stern, Mana Barari, and R. Erandi Zamora-Graziano, Deputy Attorneys General; Michael N. Feuer, City Attorney (Los Angeles), Michael Bostrom, Managing Assistant City Attorney; Mara W. Elliott, City Attorney (San Diego), Mark Ankcorn, Chief Deputy City Attorney, Kevin B. King and Marni Von Wilpert, Deputy City Attorneys; Dennis J. Herrera, City Attorney (San Francisco), Ronald P. Flynn, Chief Deputy City Attorney, Yvonne R. Meré, Chief of Complex and Affirmative Litigation, Molly J. Alarcon, Sara J. Eisenberg, and Matthew D. Goldberg, Deputy City Attorneys, for Plaintiff and Respondent.

Partnership for Working Families, Reynaldo Fuentes; Law Office of Beth A. Ross and Beth A. Ross for Gig Workers Rising, Mobile Workers Alliance, Rideshare Drivers United, and We Drive Progress as Amici Curiae on behalf of Plaintiff and Respondent.

73 Altshuler Berzon, Stacey Leyton, Andrew Kushner; Weinberg, Roger & Weinberg, David A. Rosenfeld; Law Office of Beth A. Ross and Beth A. Ross for International Brotherhood of Teamsters, Service Employees International Union California State Council, State Building and Construction Trades Council of California, Transport Workers Union, United Food and Commercial Workers Union Western States Council, Unite Here, and California Labor Federation as Amici Curiae on behalf of Plaintiff and Respondent.

National Employment Law Project, Nayantara Mehta, Brian Chen; Legal Aid at Work, George Warner for National Employment Law Project, ACLU of Northern California, Asian Americans Advancing Justice–Asian Law Caucus, Bet Tzedek Legal Services, California Employment Lawyers’ Association, the Center for Workers’ Rights, Centro Legal de la Raza, Council on American-Islamic Relations–California Chapter, La Raza Centro Legal’s Workers’ Rights Program, Legal Aid at Work, the Lawyers’ Committee for Civil Rights of the San Francisco Bay Area, the Women’s Employment Rights Clinic of Golden Gate University School of Law, and Worksafe, Inc. as Amici Curiae on behalf of Plaintiff and Respondent.

Public Rights Project, Jill E. Habig, Jonathan B. Miller, Lijia Gong, Sophia Tonnu for Public Rights Project, A Better Balance, Center for Popular Democracy, ChangeLab Solutions, Equal Justice Society, Equal Rights Advocates, National Center for Law and Economic Justice, National Center for Lesbian Rights, National Partnership for Women and Families, National Women’s Law Center, One Fair Wage, Open Markets Institute, People’s Parity Project, Public Counsel, Towards Justice, and Women’s Law Project as Amici Curiae on behalf of Plaintiff and Respondent.