ORDER DENYING PLAINTIFFS' MOTION FOR PRELIMINARY APPROVAL

O'Connor, Docket No. 518

Yucesoy, Docket No. 206

EDWARD M. CHEN, United States District Judge.

I. INTRODUCTION

Plaintiffs brought the instant class action and putative class action against Defendant Uber Technologies, Inc., alleging that Uber misclassifies drivers as independent contractors rather than employees. O'Connor v. Uber Techs., Inc., Case No. 13-cv-3826-EMC, Docket No. 330 (Second Amended Complaint) (SAC) at ¶ 3; Yucesoy v. Uber Techs., Inc., Case No. 15-cv-262-EMC, Docket No. 198 (Fourth Amended Complaint) (FAC) at ¶ 2. Following three years of contentious litigation, the parties entered into a Settlement Agreement shortly before the O'Connor trial was to begin. O'Connor, Docket No. 518; Yucesoy, Docket No. 206.1

Plaintiffs' motions for preliminary approval came on for hearing before the Court on June 2, 2016. The Court has also reviewed the parties' briefing and supplemental briefing, as well as the many objections challenging the adequacy of the Settlement Agreement. It also invited and considered the comments of the California Labor and Workforce Development Agency (LWDA). While recognizing sizeable settlement sum and policy changes proposed by the Settlement Agreement and the significant risk that drivers face in pursuing this litigation, for the reasons explained below, the Court concludes that the Settlement as a whole is not fair, adequate, and reasonable and therefore DENIES Plaintiffs' motion for preliminary approval.

II. BACKGROUND

A. Procedural History

The Settlement Agreement at issue covers two lawsuits pending before this Court. O'Connor v. Uber Technologies, Inc. was brought on behalf of all individuals who worked as Uber drivers in California.2 Docket No. 330 (O'Connor Second Amended Complaint) (SAC) at ¶ 1. O'Connor alleged that Uber misclassified its drivers as independent contractors rather than employees. As employees, drivers would be entitled to the protections of the California Labor Code, including section 2802, which requires that employees be reimbursed for expenses such as gas and use of their vehicle. Id. at ¶¶ 3, 23. Plaintiffs also contend that although Uber advertised to customers that gratuity was included in the fare and that there was no need to tip drivers, drivers did not receive the total proceeds of any such gratuity. Id. at ¶¶ 1, 20. By failing to remit the full gratuity to drivers as required by California Labor Code section 351, Plaintiffs alleged that Uber violated California's Unfair Competition Law prohibition on unlawful business practices, and they sought to recover the portion of the gratuities that Uber withheld. Id. at ¶ 34. These claims, too, are predicated on drivers being employees rather than independent contractors.

Uber has argued that because it exercises minimal control over how drivers set their own hours and work schedule, its drivers cannot be considered employees. Plaintiffs, on the other hand, contend that Uber in fact exercised considerable control and supervision over the methods and means of its drivers' provision of transportation services, making drivers employees. See id. at ¶ 21.

Over the course of contentious litigation, the Court has adjudicated a motion to dismiss, for judgment on the pleadings, and for summary judgment, as well as numerous motions regarding class certification, arbitration, and stays. In its order denying Uber's motion for summary judgment, the Court applied California's two-step process for determining whether a worker is an employee or independent contractor. Docket No. 251 (March 11, 2015 Summary Judgment Ord.) at 6. First, it found that drivers provide a service to Uber because Uber is ultimately a transportation company, albeit a technologically sophisticated one. Id. at 10-11. The fact that Uber's drivers render a service to Uber created a rebuttable presumption of employment status. Id. at 15. Second, the Court applied California's Borello multi-factor test, focusing in particular on the most significant factor the putative employer's "`right to control work details.'" Id. at 6 (quoting S.G. Borello & Sons, Inc. v. Dep't of Indust. Relations, 48 Cal.3d 341, 350, 256 Cal.Rptr. 543, 769 P.2d 399 (1989)). It concluded that the ultimate determination of employment status had to be decided by a jury because there were disputes over material questions of fact, such as whether Uber has the right to significantly control the "manner and means" of drivers' transportation services. Id. at 20-25. The Court also found that a jury could reasonably find that the Borello test's secondary factors point in opposing directions, such that the test did not yield an unambiguous result. Id. at 25-27.

Following its denial of Uber's summary judgment motion, the Court certified the following class:

Docket No. 342 (September 1, 2015 Class Certification Ord.) at 7. The primary effect of this order was to limit the class to individuals who did not sign the 2014 arbitration agreements, as the Court found that individualized inquiries would be needed to determine whether there was procedural unconscionability with respect to those later contracts. Id. at 60-63. The September 1, 2015 certified class included approximately 8,000 drivers out of the estimated 160,000 California drivers. See Docket No. 519 (April 21, 2016 Liss-Riordan Dec.) at ¶ 15; September 1, 2015 Class Certification Ord. at 5. Further, the Court only permitted the class to pursue the gratuities claim; it did not certify the class to seek the expense reimbursement claim. See September 1, 2015 Class Certification Ord. at 66-67. Although Uber sought interlocutory review of this order, the Ninth Circuit denied Uber's petition for permission to appeal. Docket No. 389.

The parties then filed extensive supplemental briefing concerning whether the class could be expanded to include other California drivers who signed the later arbitration agreements, as well as whether a class could be certified as to the claim for expense reimbursement under California Labor Code section 2802. See Docket Nos. 359, 365. During a hearing on an unrelated motion, Plaintiffs' counsel for the first time argued that, as to the arbitration agreements, no unconscionability analysis was necessary because the arbitration clauses included a non-severable waiver of claims brought under California's Private Attorneys General Act of 2004 (PAGA), which was invalid as a matter of public policy. See Docket No. 379 (Nov. 4, 2015 Trans.) at 34:9-35:15.

After further briefing on this matter and oral arguments, the Court certified the following subclass of drivers:

Docket No. 395 (December 9, 2015 Class Certification Ord.) at 32. This expanded the certified class to over 240,000 drivers. See April 21, 2016 Liss-Riordan Dec., Exh. 1. Both the September 1, 2015 class and December 9, 2015 subclass were certified to pursue the expense reimbursement claim, as well as the gratuities claim. In certifying the December 9, 2015 subclass, the Court relied on Iskanian v. CLS Transportation Los Angeles, LLC, 59 Cal.4th 348, 173 Cal.Rptr.3d 289, 327 P.3d 129 (2014) and Sakkab v. Luxottica Retail North America, Inc., 803 F.3d 425 (9th Cir.2015), which had held that a waiver of PAGA claims was void as a matter of public policy. December 9, 2015 Class Certification Ord. at 24. The Court concluded that this PAGA waiver could not be severed from the remainder of the arbitration agreement, rendering the entire arbitration agreement void. Id. Because it relied on public policy rather than unconscionability, the Court did not engage in the procedural unconscionability analysis that defeated class certification in the original certification motion. Id.3

Two days after the Court found that the arbitration agreement was invalid as a matter of public policy, Uber issued a new arbitration agreement (hereafter, the December 2015 Agreement) to all Uber drivers, including members of the certified subclass. Plaintiffs in O'Connor, Yucesoy, and In re Uber FCRA Litigation filed separate motions to enjoin the December 2015 Agreement, arguing that it was an unauthorized communication designed to undermine or discourage participation in these and other pending cases against Uber. See Docket No. 435 (Rule 23(d) Ord.) at 2. The Court granted the motions pursuant to its Rule 23(d) "power to regulate the notice and opt-out processes and to impose limitations when a party engages in behavior that threatens the fairness of the litigation." Wang v. Chinese Daily News, Inc., 623 F.3d 743, 756 (9th Cir.2010), judgment vacated on other grounds, 565 U.S. 801, 132 S.Ct. 74, 181 L.Ed.2d 1 (2011). The Court explained that its rulings on Uber's motion for summary judgment and Plaintiffs' motions for class certification, including the voiding of the arbitration clauses, created a "legal landscape [that] has become materially more complicated for the drivers." Rule 23(d) Ord. at 3-4. While the Court explicitly declined to rule on whether the December 2015 Agreement was enforceable, it did conclude that this increasingly complex legal landscape required more robust notice to drivers, and ordered that the December 2015 Agreement could not be enforced without a revised cover letter and arbitration notice with a simplified opt-out option. Id. at 6-7. Uber moved to stay the Court's order pending appeal, arguing that the Rule 23(d) Order was unwarranted and burdened its First Amendment rights. Docket No. 439 at 2-3. This Court denied the motion; following an appeal of the order, the Ninth Circuit also denied to motion to stay. Docket No. 454; Case No. 16-15000, Docket No. 10.

Uber also moved to stay the case while it sought interlocutory review of the December 9, 2015 Class Certification Order, which this Court conditionally granted in part and denied in part. Docket No. 411. In that order, the Court ruled that the trial could proceed, but that it would not enter a final judgment as to the December 9, 2015 subclass if the appeals were still pending. Docket No. 429 at 9. The Ninth Circuit denied Uber's motion to stay the trial proceedings pending appeal. Case No. 15-17420, Docket Nos. 5, 14. However, on April 5, 2016, the Ninth Circuit granted Uber's petition for permission to appeal the December 9, 2015 Class Certification Order per Rule 23(f). Docket No. 512 (Case No. 15-80220, Docket No. 9). Uber immediately filed a new motion for this Court to stay the case, which remains pending.

In the meantime, Yucesoy was filed on behalf of all Massachusetts drivers, bringing similar claims for independent contractor misclassification, violation of the Massachusetts Tips law, tortious interference with contractual and/or advantageous relations, unjust enrichment/quantum meruit, breach of contract, and violation of the Massachusetts minimum wage and overtime law. Yucesoy, Docket No. 27 (First Amended Complaint). After the Court ruled on three motions to dismiss, the remaining claims in Yucesoy are: (1) independent contractor misclassification (and attendant failure to pay business expenses), (2) violation of the Massachusetts tips law, and (3) tortious interference with advantageous relations. Yucesoy, Docket No. 198 (Fourth Amended Complaint). No motion for class certification has been brought in the Yucesoy case.

B. Settlement Agreement

1. Monetary Terms

Shortly before the trial in O'Connor was scheduled to commence, the Yucesoy and O'Connor Plaintiffs entered into a Settlement Agreement with Uber. Under this Agreement, Uber has agreed to pay $84 million, plus an additional $16 million contingent on an initial public offering (IPO) reaching one-and-a-half times Uber's most recent valuation (i.e., about $93.75 billion). April 21, 2016 Liss-Riordan Dec., Exh. 6 (Settlement Agreement) at ¶¶ 58, 125. Of the $84 million, $300,000 will be used for class administration, a maximum of $73,000 will be allocated for enhancement payments to the named Plaintiffs and other Settlement Class members who contributed to the litigation, and $8.7 million will be treated as wages reported on IRS Form W-2. Settlement Agreement at ¶¶ 125, 129, 133. Plaintiffs' counsel is also permitted to seek a fee and expense award of up to 25% of the Settlement Fund ($21 to $25 million), although Plaintiffs' counsel has since informed the Court that she will reduce her fee request by $10 million. Settlement Agreement at ¶ 134; Docket No. 699 at 2. The $10 million reduction is made regardless of whether the $16 million contingency is triggered, thus resulting in an additional $10 million for distribution to the class. Docket No. 699 at 2 n.1.

The remaining Settlement Fund will be separated into two funds: approximately $5.5 to 6 million for the Massachusetts drivers, and $56 to 66.9 million for the California drivers.4 Settlement Agreement at ¶ 144. A driver must submit a claim form to receive a payment. Settlement Agreement at ¶ 138. The driver's payment is based on the number of miles driven for Uber. Settlement Agreement at ¶ 145. Drivers may also receive "double weight" for their mileage if they opted out of Uber's 2013 and 2014 arbitration agreements, and if they are members of the O'Connor certified class. Settlement Agreement at ¶ 144. If a driver both opted out and is a member of the O'Connor certified class, he or she will receive quadruple weight for his or her mileage. Docket No. 617 (May 20, 2016 Joint Supp. Briefing) at 4. Assuming a 100% claim rate, Plaintiffs' counsel estimates that California certified class members will receive an average distribution of $24 to $1,950, California non-certified class members will receive an average distribution of $10 to $836, and Massachusetts drivers will receive an average distribution of $12 to $979. April 21, 2016 Liss-Riordan Dec., Exh. 1. For example, of the 243,320 California drivers who were part of the certified class, if all drivers filed claims, the 122,297 drivers who drove between 0-750 miles will receive an average distribution of $24, while the 42,074 drivers who drove between 750-2,000 miles would receive an average distribution of $89. On the higher end of the scale, the 7,534 drivers who drove over 25,000 miles would receive an average distribution of $1,950. Of the 60,047 Massachusetts drivers, the 33,040 drivers who drove between 0-750 miles would receive an average of $12, the 9,258 drivers who drove between 750-2,000 miles would receive an average of $45, and the 1,489 drivers who drove over 25,000 miles would receive an average of $979. In sum, the vast majority of class members are slated to receive less than $100 each from the settlement.

The parties expect a 40% claim rate, which would increase the monetary amount paid to each claimant. See May 20, 2016 Joint Supp. Briefing at 58.

2. Non-Monetary Relief

In addition to the monetary payment, Uber has agreed to implement various forms of non-monetary relief. First, Uber has agreed to publish a comprehensive, written deactivation policy. Settlement Agreement at ¶ 135(a). Driver deactivation will only be allowed for sufficient cause, and low passenger acceptance rates will not be grounds for deactivation (although it would subject drivers to being logged out of the app for a limited period of time). Settlement Agreement at ¶ 135(a)(i); see also Driver Deactivation Policy — US ONLY, Uber, https://www.uber.com/legal/other/driver-deactivation-us-english/ (last visited August 3, 2016) (Uber Deactivation Policy). A driver may still be deactivated for having a high rate of cancellation, i.e., where the driver initially accepted the fare but then canceled it (in contrast to never accepting the fare to begin with). See Uber Deactivation Policy. Uber will also provide at least two advance warnings before a driver is deactivated for reasons other than safety issues, discrimination, fraud, or illegal conduct (each, an "excluded matter"). Settlement Agreement at ¶¶ 135(a)(iii)-(iv). If a driver is deactivated, Uber will provide the driver with an explanation for the deactivation. Settlement Agreement at ¶ 135(a)(v). A deactivated driver may appeal the decision to a Driver Appeal Panel, unless the deactivation resulted from certain circumstances such as low star ratings, criminal activity, physical altercations, or sexual misconduct. Settlement Agreement at ¶ 135(a)(vi). In addition, except for the excluded matters (e.g., safety issues, discrimination, fraud, or illegal conduct), drivers whose user accounts are deactivated will have the opportunity to take a "quality improvement course" and be reactivated upon completion of the course. Settlement Agreement at ¶ 135(a)(vii).

Second, Uber will provide more information regarding star ratings. Settlement Agreement at ¶ 135(d). Uber will also "consider" changes such as informing drivers how they rank against their peers, providing warnings when driver ratings go below a certain threshold, and warning drivers when their user accounts are at risk of deactivation for going below a certain threshold.

Third, the parties stipulate to the enforceability of the December 2015 Agreement. Settlement Agreement at ¶ 135(e). In exchange, Uber will pay for the filing and administrative arbitration fees in: (1) cases based on an alleged employment relationship between Uber and drivers, and (2) cases arising out of a final deactivation of a driver in the event of an excluded matter. The parties also agree to stipulate to vacating (retrospectively) this Court's Rule 23(d) Orders, and they agree that Uber has the option of voiding the Settlement Agreement should the Court not vacate these orders. Plaintiffs also agree to withdraw the charges filed with the National Labor Relations Board (NLRB) on behalf of John Billington and Catherine London, challenging the enforceability of the 2014 arbitration agreements as a violation of the National Labor Relations Act, and will not further cooperate with the NLRB's investigation unless compelled by subpoena or court order. Settlement Agreement at ¶ 33.

Fourth, Uber will institute an internal process for drivers to raise concerns regarding the payment of specific fares in California and Massachusetts. Settlement Agreement at ¶ 135(f).

Fifth, Uber will collaborate with Plaintiffs regarding the creation and funding of a driver association as a means of "opening a dialogue between Uber and Drivers." Settlement Agreement at ¶ 135(g). The association's leaders are to be elected by drivers, and the leaders will have the opportunity to meet quarterly with Uber management to discuss driver concerns. Settlement Agreement at ¶¶ 135(g)(iv)-(v). The driver association will not be a union, and will have no right to bargain collectively with Uber. Settlement Agreement at ¶¶ 135(g)(ii)-(iii). The parties provided little detail on how the driver association will work in practice (in part due to the expected autonomy of each driver association), including what obligations Uber will have to fund the driver association.

Finally, Uber will make good-faith efforts to clarify its messaging to riders regarding tipping, i.e. that tips are not included in fares (except for UberTAXI), but that they are neither expected nor required. Settlement Agreement at ¶ 135(h); May 20, 2016 Joint Supp. Briefing at 18. Drivers will be permitted to put up signs requesting tips, although the parties disagree on whether this actually constitutes a change in policy. See May 20, 2016 Joint Supp. Briefing at 18 n.24 ("Uber expressly disputes Plaintiffs' claim that Uber's policy with respect to tipping signage will change as a result of this settlement").

3. Scope of the Class and Released Claims

The Settlement Agreement covers "all Drivers in California and Massachusetts who have used the Uber App at any time since August 16, 2009, up to and including the Preliminary Approval Date." Settlement Agreement at ¶ 103. Thus, the settlement class releasing claims will not only include the O'Connor certified class, but (1) all California drivers for Uber including California drivers who had been excluded by the class definition, i.e., drivers who drove for a third-party transportation company or who used fictitious or corporate names; and (2) all Massachusetts drivers.

Furthermore, although the O'Connor and Yucesoy cases were limited to claims based on expense reimbursement and the payment of tips, the Settlement Agreement contains an expansive release provision: it will require settlement class members — all drivers in California and Massachusetts — to release all claims based on or reasonably related to the employment misclassification claim, i.e., overtime, minimum wage, meal and rest breaks, and workers' compensation. Settlement Agreement at ¶ 105.5 The Settlement Agreement requires that the Plaintiffs file amended complaints expanding the causes of action to include all claims related to the alleged misclassification of drivers as independent contractors. See Settlement Agreement at ¶ 29 and Exhs. A (Proposed Yucesoy Fifth Amended Complaint), B (Proposed O'Connor Fifth Amended Complaint). As a result, the Settlement Agreement will cover claims that are brought in at least fifteen other lawsuits currently pending in federal and California state courts, effectively terminating those suits.6 Settlement Agreement at ¶ 28. It could also affect proceedings pending before various administrative bodies such as the NLRB. See NLRB v. Uber Techs., Inc., Board Case Nos. 20-CA-160717, 20-CA-160720 (filed September 24, 2015) (complainants required by Settlement Agreement to withdraw charges).7 The Settlement Agreement will also settle all civil penalties potentially due under PAGA, ending in all likelihood all currently pending PAGA litigation against Uber in the other lawsuits. Settlement Agreement at ¶ 105. Because PAGA is an action on behalf of the State, the PAGA settlement included in the Settlement Agreement would prohibit any other driver from bringing a PAGA claim (or obtaining relief through a PAGA representative suit) for the time period up to preliminary approval, even if that driver opts out of the Settlement Agreement. See Docket No. 732 (July 15, 2016 Joint Supp. Briefing) at 23; Docket No. 736 (Labor & Workforce Development Agency (LWDA) Resp.) at 2.

Finally, as noted above, the parties stipulate to the enforceability of the December 2015 Agreement, as well as to vacating the Court's December 2015 Rule 23(d) Order and January 19, 2016 Order regarding the arbitration agreement's notice provision and corrective cover letter (collectively, the Rule 23(d) Orders). Settlement Agreement at ¶ 135(e). The effect of such action, if agreed to by the Court, would be to retroactively strip drivers of the protections afforded by this Court's Rule 23(d) order. If the Court does not agree to vacate these orders, Uber is permitted to void the Settlement Agreement.

III. DISCUSSION

A. Standard of Review

Pursuant to Federal Rule of Civil Procedure 23(e), "[t]he claims, issues, or defenses of a certified class may be settled, voluntarily dismissed, or compromised only with the court's approval." As the Ninth Circuit has explained, "[t]he purpose of Rule 23(e) is to protect the unnamed members of the class from unjust or unfair settlements affecting their rights." In re Syncor ERISA Litig., 516 F.3d 1095, 1100 (9th Cir.2008). Accordingly, before a court approves a settlement, it must conclude that the settlement is "fundamentally fair, adequate, and reasonable." Hanlon v. Chrysler Corp., 150 F.3d 1011, 1026 (9th Cir.1998); see also Allen v. Bedolla, 787 F.3d 1218, 1222 (9th Cir.2015) (same). This inquiry:

Hanlon, 150 F.3d at 1026; see also Churchill Vill. L.L.C. v. Gen. Elec., 361 F.3d 566, 575 (9th Cir.2004) (same). "In determining whether the proposed settlement falls within the range of reasonableness, perhaps the most important factor to consider is plaintiffs' expected recovery balanced against the value of the settlement offer." Cotter v. Lyft, Inc., Case No. 13-cv-4065-VC, 176 F.Supp.3d 930, 935, 2016 WL 1394236, at *4 (N.D.Cal. Apr. 7, 2016) (internal quotation omitted); see also Procedural Guidance for Class Action Settlements, UNITED STATES DISTRICT COURT — NORTHERN DISTRICT OF CALIFORNIA, http://cand.uscourts.gov/ClassActionSettlement Guidance (last visited August 12, 2016). While it is not necessarily unusual or improper for a class action settlement agreement to release claims not originally brought by the plaintiff, the court must consider the strength and value of those claims in deciding whether to approve the settlement. See Cotter, 176 F.Supp.3d at 934-35, 2016 WL 1394236, at *4.

"[W]hether a settlement is fundamentally fair within the meaning of Rule 23(e) is different from the question whether the settlement is perfect in the estimation of the reviewing court." Lane v. Facebook, Inc., 696 F.3d 811, 819 (9th Cir.2012). However, "when ... the settlement takes place before formal class certification, settlement approval requires a `higher standard of fairness.'" Id.; see also Hanlon, 150 F.3d at 1026 ("settlement approval that takes place prior to formal class certification requires a higher standard of fairness [because t]he dangers of collusion between class counsel and the defendant, as well as the need for additional protections when the settlement is not negotiated by a court[-]designated class representative, weigh in favor of a more probing inquiry than may normally be required under Rule 23(e)"). This more "exacting review" is warranted "to ensure that class representatives and their counsel do not secure a disproportionate benefit at the expense of the unnamed plaintiffs who class counsel had a duty to represent." Lane, 696 F.3d at 819 (internal quotation omitted).

In this case, because the Settlement Agreement covers the claims of both certified class members and drivers who fall outside the class definition and thus have not been certified (for example, all Massachusetts drivers and the California drivers who drove for a third-party transportation company or under a corporate name), this Court must apply the more "exacting" standard in determining whether this settlement is fair, adequate, and reasonable. Further, with respect to all of the drivers, the parties propose to release all claims related to misclassification, including many which had not been brought in this case. This not only results in the addition of claims that Plaintiffs' counsel may not have fully investigated or reviewed, but would also eliminate at least fifteen other cases pending in California courts wherein the claims sought to be added and waived herein are being litigated. See Settlement Agreement at ¶ 28. Thus, exacting review is additionally warranted as newly added claims have not been subject to class certification. Moreover, as this Court has noted, the Court must be especially sensitive to the risk of collusion or a less than full adversarial process where claims pending in other lawsuits are released for minimal value, in order to induce the defendant to settle this case.8 See Docket No. 724 at 9.

Courts implementing Rule 23(e) have required a two-step process for the approval of class action settlements: the Court first determines whether class action settlement deserves preliminary approval and then, after notice is given to class members, whether final approval is warranted. In re High-Tech Emp. Antitrust Litig., Case No. 11-CV-2509-LHK, 2014 WL 3917126, at *3 (N.D.Cal. Aug. 8, 2014). As a general matter, "there is relatively scant appellate authority regarding the standard that a district court must apply in reviewing a settlement at the preliminary approval stage." Id. Some district courts "have stated that the relevant inquiry is whether the settlement `falls within the range of possible approval' or `within the range of reasonableness,'" looking at factors such as whether the settlement is the product of non-collusive negotiations, has no obvious deficiencies, does not improperly grant preferential treatment to class representatives or segments of the class, and falls within the range of possible approval. Id.; see also Harris v. Vector Mktng. Corp., Case No. C-08-5198-EMC, 2011 WL 1627973, at *7 (N.D.Cal. Apr. 29, 2011). Although the Ninth Circuit has not specified what standard should apply at the preliminary approval stage, "district courts often state or imply that scrutiny should be more lax." Cotter v. Lyft, Case No. 13-cv-4065-VC, 193 F.Supp.3d 1030, 1035-36, 2016 WL 3561742, at *3 (N.D.Cal. June 23, 2016).

More recently, in Cotter, Judge Chhabria questioned this "lax review," finding that:

Id. at 1037, at *4. "[B]y scrutinizing the agreement carefully at the initial stage and identifying any flaws that can be identified, the court allows the parties to decide how to respond to those flaws (whether by fixing them or opting not to settle) before they waste a great deal of time and money in the notice and opt-out process." Id.

The Court finds Judge Chhabria's view persuasive, particularly where, as here, the class includes nearly 400,000 individuals, and thus a great deal of expense would be incurred and substantial confusion could ensue were the Settlement Agreement preliminarily approved but ultimately disapproved on final review. And, as noted above, close review is particularly warranted where Plaintiffs seek to add new claims and drivers not previously certified, and the settlement would settle not only the instant case but claims brought in at least fifteen other pending lawsuits for relatively modest value. Further, the Settlement Agreement at issue has already been the subject of numerous objections challenging its adequacy even at the preliminary approval stage. As Judge Chhabria explained, it makes little sense to apply a lax standard at the preliminary approval stage to factors already known and amenable to analysis. Instead, the Court finds it more prudent to apply with full force the factors articulated by the Ninth Circuit in Hanlon and Churchill Village, while recognizing that some of these factors — such as the reaction of class members — are not currently known and cannot be assessed at the stage of preliminary approval and thus would have to await the stage of final approval.

B. Hanlon Approval Factors

1. Strength of the Plaintiffs' Case; the Risk, Expense, Complexity, and Likely Duration of Further Litigation; and the Risk of Maintaining Class Action Status

In considering the first three factors, the Court looks at the risks to both Plaintiffs and Uber in continuing this litigation.

a. Risks to Drivers

The most obvious risk to Plaintiffs is, of course, that the Ninth Circuit will uphold the validity of the arbitration provision contained in the 2013 and/or 2014 agreements, which this Court found was invalid as a matter of public policy in certifying the December 9, 2015 Subclass. See Ben Hancock, Uber ADR Pact May Get Green Light, THE RECORDER, June 16, 2016 ("A panel of federal appeals judges gave clear signs Thursday that it is ready to reverse a lower court decision finding the arbitration agreements that Uber Technologies Inc. circulated to its drivers were unenforceable"); Bonnie Eslinger, 9th Cir. Leans Toward Restoring Uber Arbitration Pacts, LAW360, June 16, 2016. This risk is heightened by the Ninth Circuit's decision to grant Uber's petition for permission to appeal the December 9, 2015 Class Certification Order. See Docket No. 512. A finding that one or both of the arbitration clauses is valid and enforceable would substantially change the scope and course of Plaintiffs' case, as it would likely require the vast majority of the class to go to arbitration on their non-PAGA claims, thus jeopardizing the scope and potential viability of the class action at bar. Plaintiffs face a considerable risk that they will not proceed as a class action in any court, or at least be limited to a class action greatly reduced in size. Even if the Ninth Circuit were to limit a finding of enforceability to the more recent contracts, and hold only the 2013 arbitration agreement not to be enforceable, this could substantially decrease the class from approximately 240,000 drivers to 8,000 drivers, dramatically lowering any class monetary recovery that Plaintiffs might obtain through the class action. See April 21, 2016 Liss-Riordan Dec. at ¶ 15. Requiring the drivers to arbitrate their claims individually will likely reduce by a substantial degree overall recovery for drivers, as typically only a fraction of individuals pursue arbitration.

In addition to this risk to maintaining class action status, as this Court has previously noted, Plaintiffs face risks on the merits of the case. The fundamental question of whether Uber drivers are employees or independent contractors is not a simple one. As this Court held in denying Uber's motions for summary judgment, there are factors under the Borello analysis that support each side's position. While the Court found that drivers performed a service for Uber and were therefore presumptively employees, it found that questions of fact existed as to the primary Borello inquiry of Uber's control over the drivers. March 11, 2015 Class Certification Ord. at 15, 25. For example, although drivers could choose their own days and hours or work, Uber controlled certain aspects of driver performance; for instance, Uber could terminate individuals with low acceptance rates. Id. at 21-25. Moreover, as to the secondary factors set forth in Borello, the Court found that these factors cut both ways. Id. at 25-26. Several factors pointed in favor of employment status, such as: (1) driving is an occupation that typically does not require close supervision, (2) driving does not require a special skill, and (3) the drivers were performing a regular and integral part of Uber's business. Id. at 26. On the other side, several factors favored independent contractor status, including: (1) the drivers' use of their own vehicles, (2) the ability of drivers to employ other drivers to drive on their own behalf, and (3) the drivers signing an agreement stating no employment relationship was created. Id. As a result, should the issue of employee versus independent contractor status proceed to trial, it would be up to the jury to make the ultimate determination, the outcome of which cannot be predicted with any certainty. See id. at 27 (finding that "[t]he application of the traditional test of employment — a test which evolved under an economic model very different from the new `sharing economy' — to Uber's business model creates significant challenges"); Cotter, 193 F.Supp.3d at 1037, 2016 WL 3561742, at *5 ("there is no straight answer to the question whether those drivers must be classified as employees or independent contractors under California law"); cf. Alatraqchi v. Uber Techs., Inc.,9 Case No. 13-cv-3156-JSC, Docket No. 9 at 6-9 (attaching California Labor Commissioner's August 1, 2012 decision, which found that a driver was an independent contractor) with Berwick v. Uber Techs., Inc., Case No. 11-46739 EK (June 3, 2015), available at http://digitalcommons.law.scu.edu/cgi/viewcontent.cgi?article=1988&context=historical (last visited August 5, 2016) (performing Borello analysis and concluding that a driver was Uber's employee).

Moreover, even if drivers were determined at trial to be employees, Uber challenges recovery on their claims. For instance, with respect to the expense reimbursement claim, Plaintiffs' counsel noted that Uber intended to argue that it already structured the fare to be "all-inclusive that takes into account things like expenses." April 21, 2016 Liss-Riordan Dec. at ¶ 31 (citing Gattuso v. Harte-Hanks Shoppers, Inc., 42 Cal.4th 554, 559, 67 Cal.Rptr.3d 468, 169 P.3d 889 (2007) ("We conclude that an employer may satisfy its statutory reimbursement obligation by paying employees enhanced compensation in the form of increases in base salary or increases in commission rates, or both, provided there is a means or method to apportion the enhanced compensation to determine what amount is being paid for labor performed and what amount is reimbursement for business expenses")).

Plaintiffs' counsel also notes that there is a risk as to which IRS mileage reimbursement rate would apply. Plaintiffs have argued for use of the fixed rate while Uber would likely advocate for use of the variable rate, a potential reduction of 60%. Id. at ¶¶ 33-34. Further, as to Massachusetts drivers, Plaintiffs perceive an additional risk because Massachusetts does not have an express expense reimbursement statute, and thus recovery for expenses in Massachusetts may be less likely. Mot. at 23 n.25; May 20, 2016 Joint Supp. Briefing at 7 ("the law is not entirely established in Massachusetts as to whether employees may recover unreimbursed business expenses from their employers"). Similarly, Plaintiffs acknowledged risks to their tips claim, should a jury find that Uber's communications indicating that "tip is included" in a fare were too variable or not widespread enough to conclude that a tip was actually included. April 21, 2016 Liss-Riordan Dec. at ¶ 45.

There are additional risks to the claims the parties seek to add to this case and for which releases are sought. First, as to meal and rest breaks, California courts have found liability for failure to provide such breaks when the employer lacked a policy authorizing and permitting breaks. E.g., Benton v. Telecon Network Specialists, Inc., 220 Cal.App.4th 701, 725-26, 163 Cal.Rptr.3d 415 (2013). But Uber states that it has set up a system in which "drivers log in and out whenever they want, so that there can never be any circumstance in which a driver might feel pressure (even implicitly) not to take a break. In other words, Uber's entire system can be understood to constitute a policy of `permitting' or `authorizing' breaks whenever a driver wants." July 15, 2016 Joint Supp. Briefing at 11. There has been little argument that Uber drivers lack the freedom to choose their days and hours of work (although there has been evidence of the control Uber exercises over drivers when they do report to work), see March 11, 2015 Summary Judgment Ord. at 25, creating a potential risk that a jury would find that Uber had not deprived its drivers of meal and rest breaks.

Second, with respect to minimum wage and overtime, the primary question appears to be whether drivers would be entitled to compensation for time spent waiting to perform a task. See Docket No. 724 at 11. This Court previously dismissed with prejudice the minimum wage and overtime claims in Yucesoy, finding that Plaintiffs had failed to plead specific facts to support their claim that waiting time should be compensable. Yucesoy, Docket No. 194 at 10-11. For example, there Plaintiffs did not explain how often ride requests came in, how many requests they had to accept, and the magnitude of the risk of deactivation if requests were not accepted. Id. In so finding, the Court looked to the Ninth Circuit's test in Fair Labor Standards Act cases, the same test applied by California courts. Id.; Gomez v. Lincare, Inc., 173 Cal.App.4th 508, 523, 93 Cal.Rptr.3d 388 (2009) (looking to the Ninth Circuit's test to determine whether an employee was free to engage in personal activities while on call).10 While the Court does not conclude that drivers could not prevail on this claim were sufficient allegations pleaded and evidence presented, there are significant risks.11

Finally, regarding the workers' compensation claims, the Court notes that the settlement is not intended to release individual workers' compensation claims or personal injury claims against an employer who failed to carry workers' compensation insurance. July 15, 2016 Joint Supp. Briefing at 14. Rather, what is at issue is a claim in which a driver challenges Uber's failure to obtain workers' compensation insurance, and seeks to recover Uber's "savings" from not obtaining such insurance under California's Unfair Competition Law (UCL). See Docket No. 592 (Ghazi Obj.) at 1-2. There is a risk that such a claim could not be enforced through the UCL, as the California Supreme Court has generally found that the UCL cannot be used to recover money in which the plaintiff does not have an ownership interest. Korea Supply Co. v. Lockheed Martin Corp., 29 Cal.4th 1134, 1143-53, 131 Cal.Rptr.2d 29, 63 P.3d 937 (2003). While the State has brought a similar claim under the UCL, see People ex rel. Harris v. Pac. Anchor Transp., Inc., 59 Cal.4th 772, 775-76, 174 Cal.Rptr.3d 626, 329 P.3d 180 (2014), it is not clear that private individuals could recover under such a claim, in light of Korea Supply Co.12

b. Risks to Uber

While Plaintiffs thus face substantial risks both in their ability to maintain class certification and on the merits, the Court reiterates that Uber also faces substantial risks of losing on the misclassification question. As noted above, in its March 11, 2015 Summary Judgment Order, this Court held, as a matter of law, "Uber's drivers render service to Uber, and thus are Uber's presumptive employees." March 11, 2015 Summary Judgment Ord. at 15. Thus, the burden is on Uber to disprove an employment relationship, both in California and in Massachusetts. Id. at 16; see also Sebago v. Boston Cab Dispatch, Inc., 471 Mass. 321, 327, 28 N.E.3d 1139 (2015) ("an individual performing any service is presumed to be an employee[, and t]he purported employer may rebut the presumption of employment by establishing... three indicia of an independent contractor relationship") (internal quotations omitted). While Uber has emphasized that drivers are free to pick and choose when they work, this Court has found that this freedom does not preclude a finding of an employment relationship, as the more relevant inquiry is how much control Uber exercises over drivers while they are on duty for Uber. March 11, 2015 Summary Judgment Ord. at 25. Judge Chhabria recognized in Cotter that:

Cotter, 176 F.Supp.3d at 944, 2016 WL 1394236 at *11. And just as Judge Chhabria found that "if the jury reached a similar conclusion about Lyft drivers, the consequences for Lyft would be enormous," the Court concludes that the consequence for Uber of an adverse jury verdict would be substantial. Id.

Next, even if Uber were to prevail on its argument that the 2013 and/or 2014 arbitration agreements were enforceable, it would face substantial risks and costs absent settlement. First, PAGA claims13 cannot be compelled to arbitration. See Iskanian, 59 Cal.4th at 360, 173 Cal.Rptr.3d 289, 327 P.3d 129; Sakkab, 803 F.3d at 431-40.14 Thus, the employment classification question could still be decided by this or another court in the adjudication of a PAGA claim. See Docket No. 593 (Richardson Obj.).15 Should a court conclude that Uber drivers are, in fact, employees, both Plaintiffs and the LWDA conclude the statutory penalty against Uber would exceed $1 billion. See LWDA Resp. at 3. Further, such a PAGA judgment could have an influential and perhaps even binding effect on Uber in arbitration. In Arias v. Superior Court, 46 Cal.4th 969, 985, 95 Cal.Rptr.3d 588, 209 P.3d 923 (2009), the defendants argued that it would be unfair not to impose class action requirements for PAGA claims because "the defendant would be a party to every lawsuit while each of the various plaintiffs would be a party in only one lawsuit, [and therefore] the defendant would in later lawsuits be bound by any adverse determination of the common issues, while none of the plaintiffs would be similarly bound by prior determinations in the defendant's favor." The California Supreme Court concluded, however, there was no unfairness because a PAGA judgment would be "binding not only on the named employee plaintiff but also on government agencies and any aggrieved employee not a party to the proceeding." Id.

Second, if one or more drivers were not bound by arbitration (e.g., because like the class representatives in O'Connor, they opted out of arbitration), they would be free to litigate the merits of their claims, again raising the prospect that were Uber to lose on the merits of the classification question, that judgment would affect the outcome of arbitrations. Simply put, without the benefit of the release and waiver conferred by the Settlement Agreement, Uber still faces a substantial risk of litigation.

Further, even if the class were wholly or partially decertified and hundreds of thousands of drivers were remitted to arbitration, if even a fraction of the 380,000 drivers invoked arbitration, the mere transactional costs for Uber (in the absence of the broad release effected by the Settlement Agreement) of arbitrating thousands of cases would be substantial, not to mention the risks of findings of liability and imposition of damages (which for any driver could well be ten times greater than the award scheduled under the Settlement Agreement) by the arbitrators.

2. The Amount Offered in Settlement

a. Monetary Amount

Under the Settlement Agreement, Uber has agreed to make an $84 million guaranteed payment (including a $1 million payment for PAGA), as well as a $16 million payment contingent on the success of an IPO. As an initial matter, the Court cannot consider the $16 million contingent payment as part of the settlement amount because there is no information on the likelihood that this contingency will be triggered. At the June 2, 2016 hearing, Uber stated that the $16 million payment was "very likely" based on public reports about Uber's cash infusions and analysts' reports. See Docket No. 691 (June 2, 2016 Trans.) at 35:17-36:1. However, when the Court requested this public information, Uber stated that "it would not be proper for Uber to provide such information." July 15, 2016 Supp. Briefing at 25. While Plaintiffs provide a number of articles, none of the articles address the likelihood that an Uber IPO will yield an average valuation of at least 1.5 times Uber's most recent valuation within 365 days from the closing of the IPO. See Docket No. 733 (July 15, 2016 Liss-Riordan Dec. at 10 n.4). Thus, absent a showing that there is a realistic likelihood that the additional $16 Million will be realized, the Court will only consider the $84 million monetary amount in assessing the adequacy of the Settlement Agreement.

By comparison, Plaintiffs have estimated that the verdict value of the non-PAGA claims being released are:

            CLAIM:                       VALUE:

Expense Reimbursement (Mileage)16     $700 million

Expense Reimbursement (Phones)        $30 million

Tips                                  $122 million

Overtime                              $2.4 million

                          Total:      $854.4 million

[Editor's Note: The preceding image contains the reference for footnote16].

Mot. at 24; see also April 21, 2016 Liss-Riordan Dec. at ¶ 57. For all other non-PAGA claims, Plaintiffs' counsel attributes no value on the basis that "there would be a substantial risk of no recovery on this claim." As discussed above, the Court agrees with Plaintiffs that there were substantial risks as to the breaks claims, minimum wage and overtime claims, and workers' compensation claims, and it was therefore reasonable for Plaintiffs' counsel to assign no or little value to these claims when considering the overall full-verdict value. Compare with Cotter, 193 F.Supp.3d at 1038, 2016 WL 3561742, at *5 (finding that failure to assess value of a particular gratuity claim "does not automatically invalidate the settlement they reached. If the unconsidered claims are not particularly strong or valuable, such that they're not likely to have materially influenced the overall settlement, counsel's failure to consider the claims would not be a basis for rejecting the agreement").

After considering the information provided to the Court in response to the Court's order for supplemental briefing, the Court concludes that the parties' assessment of the value of all the non-PAGA claims is reasonably accurate. Thus, looking solely at the monetary relief, the settlement of $84 million constitutes about 10% of the full verdict value of the non-PAGA claims — i.e. a 90% discount off the verdict value of the non-PAGA claims. This substantial discount is well illustrated by the case of Ms. Berwick, who drove 6,468 miles and was awarded $3,878.08 in unreimbursed expenses by the California Labor Commissioner, would likely receive approximately $455 if she was a California class member, or $195 if she was a California non-class member. See Liss-Riordan Dec. Exh. 1 (average distributions); Berwick v. Uber Techs., Inc., Case No. 11-46739 EK.

b. Non-Monetary Relief

The Settlement Agreement is, of course, not limited to the $84 million payment, but includes a number of non-monetary relief that also provide benefits to the class. However, much of this non-monetary relief is not as valuable as the parties suggest, limiting their worth in considering the amount being offered in settlement.

First, Uber has agreed to "promulgate a comprehensive written deactivation policy," which will permit deactivation only for sufficient cause. Settlement Agreement at ¶ 135(a). Notably, a driver can no longer be deactivated because of low acceptance rates. Settlement Agreement at ¶ 135(a)(i); see also Uber Deactivation Policy. This is a significant change from prior contracts which seemed to allow Uber to fire its drivers for any reason at any time. See March 11, 2015 Summary Judgment Ord. at 20. But while Uber will no longer deactivate drivers because of low acceptance rates, it will still exercise substantial control over a driver's ability to accept or decline ride requests, as the deactivation policy still permits Uber to "temporarily... log[a driver] out of the app for a limited period of time" based on low acceptance rates. Uber Deactivation Policy; see also Travis Kalanick, Growing and Growing Up, UBER NEWSROOM (Apr. 21, 2016), https://newsroom.uber.com/growing-and-growing-up/ ("As part of the settlement, Uber has agreed not to deactivate drivers who regularly decline trips when they are logged into the app.... [W]here drivers do have low acceptance rates ... we will alert them to the issue. If things don't pick up, we may log them out of the app for a limited period of time").

Moreover, although Uber has added a Driver Appeal Panel to which a deactivated driver can appeal a deactivation decision, as well as agreed to pay for the arbitration costs of a challenge to a final deactivation decision in the event of an "excluded matter," both these avenues of review leave out an important reason for deactivation — low star ratings. Plaintiffs' counsel noted that star ratings are "a frequent reason for deactivation." June 2, 2016 Trans. at 84:12-13. While star ratings based on customer reviews might seem to be a relatively objective basis upon which Uber may deactivate a driver without an appeal process, but there may be value to such a process. For instance, Uber has expressed concern about bias and subjectivity on the part of passengers with respect to tipping; deactivations based on star ratings could likewise be subject to the same bias and subjectivity. See June 2, 2016 Trans. at 85:9-16.

Second, the Court is not convinced that the change to tipping policy will result in the "substantially increased income" that Plaintiffs' counsel promises. See May 20, 2016 Joint Supp. Briefing at 16. Plaintiffs' counsel suggests that if passengers tip 5% on average, Uber drivers would have earned an additional $125 million since 2009. Docket No. 611 (Plaintiffs' Resp. at 21). In support of this assertion, Plaintiffs cite Lyft's model, where customers have tipped drivers more than $85 million since the company's founding. May 20, 2016 Joint Supp. Briefing at 16. Plaintiffs' $125 million valuation suffers a number of flaws. As an initial matter, it relies on a 100% tipping rate, which is highly unlikely given that even Lyft, which includes an in-app tipping function (thus making it clear to riders that tipping is not already included in the fare), only has a 70% tipping rate. See Plaintiffs' Resp. at 21 n.18. By contrast, Uber has made it clear that it will not add an in-app tipping function, thus requiring riders to tip using cash (which many riders may not have readily on-hand, given Uber's emphasis on the cashless transaction).17 Importantly, while Uber has agreed to "clarify" its tipping policy to make clear that tips are not included in the fare, it has also actively discouraged tipping, arguing that it is inconsistent with its business model, drivers' interests, and a positive rider experience. May 20, 2016 Joint Supp. Briefing; see also Our Approach to Tipping, UBER UNDER THE HOOD (Apr. 28, 2016), https://medium.com/uber-under-the-hood/our-approach-to-tipping-aa0074c0fddc#.wb66dqmuq. In other words, Uber may be permitting tipping, but it is also telling riders not to tip, further decreasing the amount of tips that riders are likely to give. Given the lack of an in-app tipping function and Uber's active dissuasion of tipping, the value of this tipping policy (which Uber strenuously disputes is even a change, see May 20, 2016 Joint Supp. Briefing at 18) is, while not meaningless, not nearly as valuable as Plaintiffs suggest. The fact that no in-app tipping function will be included will also make it difficult, if not impossible, to measure the effectiveness of the new tipping policy.

Finally, the parties have stipulated to the enforceability of the December 2015 Agreement and that this Court's Rule 23(d) Orders be vacated. Settlement Agreement at ¶ 135(e). Although included as part of the non-monetary relief, this portion of the Settlement is an additional benefit to Uber, not the class, and the Settlement Agreement is voidable at Uber's option should the Court not vacate its prior orders. By stipulating to the enforceability, drivers will be prevented from challenging the validity of the December 2015 Agreement, even in cases unrelated to employment misclassification such as whether arbitration would violate the National Labor Relations Act or on the ground that it denies a class member a contractual right to effectual relief, a claim that has been brought in Congdon v. Uber Technologies, Inc., Case No. 16-cv-2499-YGR. See July 15, 2015 Joint Supp. Briefing at 27.

Further, as expressed at the June 2, 2016 hearing, the Court is concerned about retroactively vacating its orders and the potential impact on the due process rights of drivers who may not have opted out of the December 2015 Agreement in reliance on those orders. See June 2, 2016 Trans. at 117:2-25. This is because the Court's Rule 23(d) Order specifically held that Uber could not enforce the December 2015 Agreement until it added more robust cover letter and notice provisions, and thus a driver was not required to exercise his or her opt-out right until Uber complied with Court's directives. Rule 23(d) Ord. at 6-7. Uber did not comply with the order. To retroactively revoke the protection that this Court imposed to protect the rights of drivers without affording drivers a right to now opt-out would be to put a driver in a worse position than if the Court had not issued the Rule 23(d) Orders at all. Such retroactive elimination of protection afforded by the Court could raise due process issues. Cf. Gen. Motors Corp. v. Romein, 503 U.S. 181, 191, 112 S.Ct. 1105, 117 L.Ed.2d 328 (1992) ("Retroactive legislation presents problems of unfairness that are more serious than those posed by prospective legislation, because it can deprive citizens of legitimate expectations and upset settled transactions"); Landgraf v. USI Film Prods., 511 U.S. 244, 266, 114 S.Ct. 1483, 128 L.Ed.2d 229 (1994) ("The Due Process Clause also protects the interests in fair notice and repose that may be compromised by retroactive legislation; a justification sufficient to validate a statute's prospective application under the Clause may not suffice to warrant its retroactive application") (internal quotation omitted). Further, vacating this Court's Rule 23(d) would effectively circumvent and nullify both this Court's and the Ninth Circuit's denial of Uber's motions to stay the Court's Rule 23(d) Orders on January 8 and January 13, 2016, respectively.

3. The Extent of Discovery Completed and the Stage of the Proceedings

The parties entered into the Settlement Agreement shortly before the O'Connor case was to go to trial, after having litigated two class certification motions and two summary judgment motions, as well as submitted their trial plans in preparation for trial. During this time, the parties conducted a significant amount of discovery, including a combined 326 Requests for Production, 216 interrogatories, 71 requests for production, and multiple depositions, including depositions of Uber's Senior Vice President of Operations, Ryan Graves, and five named plaintiffs. April 21, 2016 Liss-Riordan Dec. at ¶ 4. In addition, the parties exchanged discovery prior to settlement discussions. Id. at ¶ 25; see also July 15, 2016 Joint Supp. Briefing at 6.

4. The Experience and Views of Counsel

This Court previously found that Plaintiff's counsel is a "capable advocate" and a leading practitioner in the field of employment misclassification. September 1, 2015 Class Certification Ord. at 66. Plaintiff's counsel has strongly advocated for preliminary approval of the Settlement Agreement, particularly after considering the risks of this Court's arbitration orders being overturned, and even offered to reduce her fees by $10 million so that these funds could be distributed to the class, regardless of whether the $16 million contingent payment is triggered. Docket No. 699 at 2. However, as noted above, deference to the views of counsel must be tempered here where the Settlement Agreement at the eleventh hour folds in new claims and class members at the expense of litigation pending in other courts, while attributing almost no value to those claims, in order to induce Uber to settle the cases at bar.

5. The Presence of a Governmental Participant

In general, there has been no governmental participant in this case. However, as will be more fully discussed below, the California LWDA has, at the invitation of the Court, submitted a letter regarding the PAGA claim, in which it expresses serious reservation about the $1 million allocated to the newly added PAGA claim. See LWDA Resp. at 3.

6. The Reaction of the Class Members to the Proposed Settlement

Plaintiffs' counsel submits that since the announcement of the settlement, her firm has received feedback from more than 2,500 drivers; of these, 1,797 e-mails were from drivers wanting to confirm they were in the class or asking how to submit a claim. 71 class members expressed support for the Settlement Agreement, while 33 class members expressed negativity towards the Settlement Agreement. Docket No. 613-1 (Mason Dec.) at ¶¶ 8, 12, 13, 15. However, even at this preliminary state, this Court has received (and continues to receive) numerous objections, filed both by individuals and attorneys representing drivers in other California cases. E.g., Docket No. 529, 536-540, 546-548, 551-553, 556, 559, 561, 567, 569-571, 579, 581, 582, 584, 592, 594, 599, 601-604, 626, 652, 662, 688, 675, 690, 737. These objectives are in addition to five motions to intervene and one motion to disqualify Plaintiffs' counsel. Docket Nos. 588, 591, 627, 637, 644, 677.

7. Balancing the Factors

Balancing the Hanlon factors, the Court agrees with Plaintiffs' assessment that there is a substantial risk on the arbitration question in light of the Ninth Circuit's actions thus far, a risk that many of the objectors fail to appreciate. This risk would have the effect of substantially altering — if not effectively terminating — the class action in this Court, as well as in pending state court cases. These risks could well render a settlement providing for monetary relief reflecting a 90% discount off the verdict value along with limited non-monetary relief fair and adequate. Indeed, while at the low end of reasonable recovery,18 the Court would be inclined, after weighing the Hanlon factors, to find the consideration afforded by the settlement to be adequate for release of the non-PAGA claims.19 However, the parties' inclusion of waiver of PAGA claims as part of the settlement considerably alters the Court's assessment of the fairness and adequacy of the settlement as a whole.

C. Private Attorneys General Act (PAGA)

In 2003, California enacted the Private Attorneys General Act of 2004. Arias, 46 Cal.4th at 980, 95 Cal.Rptr.3d 588, 209 P.3d 923. As explained by the LWDA:

LWDA Resp. at 2 (quoting Cal. Lab. Code § 90.5(a)); see also Arias, 46 Cal.4th at 980, 95 Cal.Rptr.3d 588, 209 P.3d 923 (explaining that in passing PAGA, "[t]he Legislature declared that adequate financing of labor law enforcement was necessary to achieve maximum compliance with state labor laws, that staffing levels for labor law enforcement agencies had declined and were unlikely to keep pace with the future growth of the labor market, and that it was therefore in the public interest to allow aggrieved employees, acting as private attorney generals, to recover civil penalties for Labor Code violations"). The California Supreme Court has also recognized that "PAGA was clearly established for a public reason," such that a prohibition of a representative PAGA action would be contrary to public policy. Iskanian, 59 Cal.4th at 383, 173 Cal.Rptr.3d 289, 327 P.3d 129; see also id. at 383, 173 Cal.Rptr.3d 289, 327 P.3d 129 (explaining that a pre-dispute PAGA waiver "serves to disable one of the primary mechanisms for enforcing the Labor Code").

A plaintiff who brings a PAGA claim "does so as the proxy or agent of the state's labor law enforcement agencies." Arias, 46 Cal.4th at 986, 95 Cal.Rptr.3d 588, 209 P.3d 923. Because the "plaintiff represents the same legal right and interest as state labor law enforcement agencies," the California Supreme Court has found that "a judgment in an employee's action under the act binds not only that employee but also the state labor law enforcement agencies." Id. In short, because the employee's PAGA action acts as a "substitute" for a governmental action, the judgment binds all those who would be bound by an action brought by the government, including nonparty employees. Id. Thus, in a lawsuit which asserts a PAGA claims and seeks class certification for labor/wage claims, even class members who opt out of the class would be bound by an adverse PAGA judgment or settlement. For that reason, the LWDA rightly has stressed that:

LWDA Resp. at 2-3.

Here, Plaintiffs seek formally to add the PAGA claim to the suit and settle it for $1 million, despite having previously argued that the PAGA claim could result in penalties over $1 billion. April 21, 2016 Liss-Riordan Dec. at ¶ 82. As noted above, the LWDA also concludes that the verdict value of the PAGA claim in this case exceeds $1 billion. See LWDA Resp. at 3 ("LWDA believes it is accurate to estimate the potential PAGA penalty exposure as in excess of $1 billion"). This $1 billion amount makes up more than half of the total verdict value of the case. Plaintiffs propose settling the PAGA claim for 0.1% of its estimated full worth. The Court is cognizant that even if a verdict were rendered for PAGA plaintiff(s), a penalty of $1 billion could well be reduced, as a court may reduce the penalty when "to do otherwise would result in an award that is unjust, arbitrary and oppressive, or confiscatory." Cal. Lab. Code § 2699(e)(2).20 Nonetheless, as the LWDA concludes, there is "no rationale for allocating $1 million of the proposed settlement fund to the PAGA claim ... other than that this is a `round' number and a large figure in comparison to other PAGA settlements." Id. The parties have failed to demonstrate how the Hanlon factors or any other coherent analysis justifies settling the PAGA claim for such a relatively meager value.

It is important to note that where plaintiffs bring a PAGA representative claim, they take on a special responsibility to their fellow aggrieved workers who are effectively bound by any judgment. See Iskanian, 59 Cal.4th at 381, 173 Cal.Rptr.3d 289, 327 P.3d 129 ("When a government agency is authorized to bring an action on behalf of an individual or in the public interest, and a private person lacks an independent legal right to bring the action, a person who is not a party but who is represented by the agency is bound by the judgment as through the person were a party"). Such a plaintiff also owes responsibility to the public at large; they act, as the statute's name suggests, as a private attorney general, and 75% of the penalties go to the LWDA "for enforcement of labor laws ... and for education of employers and employees about their rights and responsibilities under this code." Cal. Lab. Code § 2699(i). This duty imposed upon the PAGA representative is especially significant given that PAGA does not require class action procedures, such as notice and opt-out rights. The Court must be cognizant of the risk that despite this responsibility, there may be a temptation to include a PAGA claim in a lawsuit to be used merely as a bargaining chip, wherein the rights of individuals who may not even be members of the class and the public may be waived for little additional consideration in order to induce the employer to agree to a settlement with the class.

This is not to suggest that the PAGA claim must comprise a disproportionate amount of class settlements which include a PAGA claim, with the majority of funds going to the State simply because the PAGA penalty has the potential to be larger than the actual claims. Such a requirement would come at the expense of the workers, who might otherwise benefit from a larger non-PAGA settlement. Rather, in reviewing a settlement that includes both a Rule 23 class and a PAGA claim, the Court must closely examine both aspects of the settlement. While a proposed settlement must be viewed as a whole, see In re Online DVD-Rental Antitrust Litig., 779 F.3d 934, 944 (9th Cir.2015), the Court must evaluate the adequacy of compensation to the class as well as the adequacy of the settlement in view of the purposes and policies of PAGA. In doing so, the court may apply a sliding scale. For example, if the settlement for the Rule 23 class is robust, the purposes of PAGA may be concurrently fulfilled. By providing fair compensation to the class members as employees and substantial monetary relief, a settlement not only vindicates the rights of the class members as employees, but may have a deterrent effect upon the defendant employer and other employers, an objective of PAGA. Likewise, if the settlement resolves the important question of the status of workers as employees entitled to the protection of the Labor Code or contained substantial injunctive relief, this would support PAGA's interest in "augmenting the state's enforcement capabilities, encouraging compliance with Labor Code provisions, and deterring noncompliance." LWDA Resp. at 2. But where, as here, the compensation to the class amounts is relatively modest when compared to the verdict value, the non-monetary relief is of limited benefit to the class, and the settlement does nothing to clarify the status of drivers as employees versus independent contractors, the settlement of the non-PAGA claims does not substantially vindicate PAGA. In these circumstances, the adequacy of settlement as a whole turns in large part on whether the PAGA aspect of the settlement can stand on its own.

Here, the Court cannot find that the PAGA settlement is fair and adequate in view of the purposes and policies of the statute. Plaintiffs propose settling PAGA for only 0.1% of the potential verdict value, a reduction that the LWDA has found has no rational basis.21 This 99.9% reduction does not adequately reflect the parties' respective risks, particularly when the PAGA claim would not be subject to the same arbitration risk that this Court has found justifies in part the 90% discount in the verdict value of the non-PAGA claims. Instead, the risks at issue rest primarily on the merits of drivers' labor codes claims and the discretionary reduction of statutory penalties, not on the risk of compelled arbitration. However, as discussed above, those risks are not limited to Plaintiffs; Uber also takes on a significant risk that should a representative PAGA claim be litigated and adjudicated, it could lose on this question (especially given that this Court has found a presumption of employee status, see March 11, 2015 Summary Judgment Ord. at 15), and such an adverse judgment would carry not only a direct monetary penalty, but potentially could affect other litigation including arbitrations. Instead of adequately considering these risks to Uber and the full value of the PAGA claim, in settling the PAGA claim herein, Plaintiffs appear to treat the PAGA claim simply as a bargaining chip in obtaining a global settlement for Uber's benefit, even though the PAGA claim alone is worth more than half of the full verdict value of all claims being released. The Court shares the LWDA's view that there is "no rationale for allocating $1 million of the proposed settlement fund to the PAGA claim ... other than that this is a `round' number and a large figure in comparison to other PAGA settlements," LWDA Resp. at 3. Given the sweeping consequences of the proposed PAGA waiver, viewed in the context of a relatively modest settlement of the non-PAGA claims, the Settlement Agreement is not as a whole is fair, adequate and reasonable.

Even if the PAGA claim were not separately scrutinized, viewing all the claims combined (PAGA and non-PAGA), the Settlement Agreement yields less than 5% of the total verdict value of all claims being released. Although the litigation risks to the plaintiffs are substantial, absent the sweeping releases conferred by the Settlement Agreement, Uber faces significant risks and costs, regardless of the outcome of pending interlocutory appeals. The settlement as a whole as currently structured is not fair, adequate, and reasonable.

IV. CONCLUSION

The Court therefore DENIES Plaintiffs' motion for preliminary approval.

Although the Court denies Plaintiffs' motions for preliminary approval and thus refuses to vacate its Rule 23(d) orders pursuant to the parties' stipulation, it will TERMINATE its Rule 23(d) orders. The legal landscape no longer requires the protection afforded by the Orders. See Docket No. 522; Case No. 14-5200, In re Uber FCRA Litig., Docket No. 175. Thus, Uber is permitted to issue the December 2015 Agreement to new drivers without satisfying the enhanced notice provisions required by the Court. Uber may also re-issue the December 2015 Agreement to current drivers, with the exception of the certified O'Connor class and claims (which, according to Uber, they did not intend the December 2015 Agreement to affect to begin with, see Docket No. 408, Exh. C; Docket No. 410 at 4; Docket No. 428 at 38:24-39:7). The Court will not, however, retroactively vacate its Rule 23(d) orders, and thus it will not deem the December 2015 Agreement effective as to drivers who did not timely opt out of the arbitration agreement during the pendency of the Rule 23(d) orders; the Court does not rule on the enforceability of the December 2015 Agreement.

Because the Court denies Plaintiffs' motion for preliminary approval, Uber's motion to stay filed on April 6, 2016 remains pending. Docket No. 506. The parties are ordered to meet and confer to discuss how they wish to proceed with that motion, as well as the general status of this case in light of the Court's ruling and the pendency of the appeals pending in the Ninth Circuit. A joint status report will be due on September 8, 2016, and a Status Conference will be held at 10:30 a.m. on September 15, 2016.

This order disposes of O'Connor, Docket No. 518 and Yucesoy, Docket No. 206.

IT IS SO ORDERED.