THE SHARING ECONOMY: DISRUPTING THE BUSINESS AND LEGAL LANDSCAPE

Panel 402 NAPABA Annual Conference Saturday, November 5, 2016 9:15 a.m.

1. Program Description

Tech companies are revolutionizing the economy by creating marketplaces that connect individuals who “share” their services with consumers who want those services. This “sharing economy” is changing the way Americans rent housing (Airbnb), commute (Lyft, Uber), and contract for personal services (Thumbtack, Taskrabbit). For every billion-dollar unicorn, there are hundreds more startups hoping to become the “next big thing,” and APAs play a prominent role in this tech boom.

As sharing economy companies disrupt traditional businesses, however, they face increasing regulatory and litigation challenges. Should on-demand workers be classified as independent contractors or employees? Should older regulations (e.g., rental laws, taxi ordinances) be applied to new technologies? What consumer and privacy protections can users expect with individuals offering their own services? Join us for a lively panel discussion with in-house counsel and law firm attorneys from the tech sector.

2. Panelists

Albert Giang Shareholder, Caldwell Leslie & Proctor, PC Albert Giang is a Shareholder at the litigation boutique Caldwell Leslie & Proctor. His practice focuses on technology companies and startups, from advising clients on cutting-edge regulatory issues to defending them in class actions and complex commercial disputes. He is the rare litigator with in-house counsel experience: he has served two secondments with the in-house legal department at Lyft, the groundbreaking peer-to-peer ridesharing company, where he advised on a broad range of regulatory, compliance, and litigation issues. Albert also specializes in appellate litigation, having represented clients in numerous cases in the United States Supreme Court, the United States Court of Appeals for the Ninth Circuit, and California appellate courts. He has been recognized as one of the “Most Influential Minority Lawyers” by the Los Angeles Business Journal, a “Rising Star” in the appellate field by Los Angeles Magazine, and one of the “Best Under 40” by the National Asian Pacific American Bar Association. Prior to joining Caldwell Leslie & Proctor, Albert served as a law clerk for the Honorable Richard A. Paez, United States Court of Appeals for the Ninth Circuit, graduated with distinction from Stanford Law School, and received his Bachelor of Arts degree magna cum laude from Amherst College.

Loni Mahanta Managing Counsel, Employment/Litigation at Lyft, Inc. Loni Mahanta is Senior Counsel, Employment/Litigation at Lyft, Inc., the groundbreaking peer-to-peer ridesharing company. In addition to overseeing employment issues at Lyft, she plays a leading role in the widely-followed disputes over whether Lyft and Uber’s drivers should be classified as independent contractors or employees. Prior to joining Lyft, Loni was an attorney at Folger Levin LLP and Crowell & Moring LLP, where she specialized in employment litigation. She obtained her B.A. in International Relations from Stanford University and graduated magna cum laude and Order of the Coif from Hastings College of the Law.

CALDWELL LESLIE & PROCTOR, PC / 725 S. FIGUEROA ST, 31ST FLOOR / LOS ANGELES, CA 90017 / 213 629 9040 / WWW.CALDWELL-LESLIE.COM

Jennifer Lam Senior Product Counsel, Airbnb Jennifer Lam is the Senior Product Counsel at Airbnb, where she plays a leading role in the company’s product and privacy work. She previously spent four years as in-house counsel at Zynga, where she was lead counsel on issues of privacy, product, promotions, and content. Prior to working at these innovative startups, Jennifer was an Associate at Kasdan, Simonds, Riley & Vaughan and graduated from Golden Gate University School of Law.

Steven Siger Managing Counsel, Thumbtack Steve Siger is Managing Counsel at Thumbtack, a groundbreaking company that connects businesses and customers with local service professionals. At Thumbtack, Steve is the primary legal advisor to the Product, Engineering, and Marketing teams, while also working on government affairs, trust and safety, and other legal issues. Previously, Steve served as Senior Counsel at Uber, where he led the team responsible for all regulatory and legal policy affairs for Uber in the western United States; as Chief of Staff and Acting Deputy Assistant Attorney General in the U.S. Justice Department’s Office of Legal Policy; as a Law Clerk to the Honorable Judith W. Rogers on the United States Court of Appeals for the District of Columbia Circuit; as an Associate at Latham & Watkins LLP; and on the advance staff of Obama for America. Steve graduated with distinction and pro bono distinction from Stanford Law School, and received his B.A. in Political Science magna cum laude and with distinction from Yale University.

3. Moderator

Albert Lin Partner, Ice Miller Albert Lin is a partner in the Columbus, Ohio office of Ice Miller, LLP. Lin’s practice focuses on government regulatory enforcement, complex litigation, and appeals. Prior to joining Ice Miller, Lin served as General Counsel to Ohio Attorney General Richard Cordray. As a member of the AG’s senior staff, Lin helped develop the office’s strategic plan and had primary responsibility for major financial litigation, criminal justice, gaming, and other matters. Prior to joining the AG’s office, Lin was an associate at Vorys, Sater, Seymour & Pease, LLP in Columbus, Ohio, where his practice focused on government regulatory matters and complex litigation. Lin also served as a law clerk to two federal judges: The Honorable R. Guy Cole, Jr. (6th Cir.) and The Honorable Ann Aldrich (N.D. Ohio).

4. Outline

a. Definition(s) of “Sharing Economy”

i. “Sharing economy is an umbrella term with a range of meanings, often used to describe economic and social activity involving online transactions. Originally growing out of the open-source community to refer to peer-to-peer based sharing of access to goods and services, the term is now sometimes used in a broader sense to describe any sales transactions that are done via online market places, even ones that are business to consumer (B2C), rather than peer-to-peer…. [A] means of describing a generally more democratized marketplace, even when it’s applied to a broader spectrum of services. Also known as shareconomy, collaborative consumption or peer economy, a common academic definition of the term refers to a hybrid market model (in between owning and gift giving) of peer-to-peer exchange. Such transactions are often facilitated via community-based online services.” [Wikipedia]

2

CALDWELL LESLIE & PROCTOR, PC / 725 S. FIGUEROA ST, 31ST FLOOR / LOS ANGELES, CA 90017 / 213 629 9040 / WWW.CALDWELL-LESLIE.COM

ii. “Increasingly, consumers and independent service providers are engaging in transactions facilitated by an Internet-based platform. The digital firms that provide the platforms are often collectively referred to as belonging to the ‘sharing’ or ‘collaborative’ economies, among other descriptors. [W]e narrow the focus and propose a definition of ‘digital matching firms’ that exhibit the following characteristics:

(1) They use information technology (IT systems), typically available via web-based platforms, such as mobile ‘apps’ on Internet- enabled devices, to facilitate peer-to-peer transactions.

(2) They rely on user-based rating systems for quality control, ensuring a level of trust between consumers and service providers who have not previously met.

(3) They offer the workers who provide services via digital matching platforms flexibility in deciding their typical working hours.

(4) To the extent that tools and assets are necessary to provide a service, digital matching firms rely on the workers using their own.” [U.S. Department of Commerce, Digital Matching Firms: A New Definition in the “Sharing Economy” Space (June 2016), available at http://www.esa.gov/sites/default/files/digital- matching-firms-new-definition-sharing-economy-space.pdf.]

iii. Span different industries: Lyft, Uber, Airbnb, Thumbtack

iv. Sometimes used interchangeably with: peer-to-peer economy, collaborative consumption, on-demand economy, gig economy

b. Economic impact

i. PARTICIPATION: According to a 2015 PWC report:

(1) 44% of US adults are familiar with the sharing economy.

(2) 18% of US adults have participated in the sharing economy as a consumer.

(3) 7% of the US population are providers in the sharing economy.

(4) The five key sharing sectors—travel, car sharing, finance, staffing, and music and video streaming—have global revenues of roughly $15 billion today, which are projected to increase to around $335 billion by 2025.

(5) “Airbnb averages 425,000 guests per night, totaling more than 155 million guest stays annually—nearly 22% more than Hilton Worldwide, which served 127 million guests in 2014. Five-year-old Uber operates in more than 250 cities worldwide and as of February 2015 was valued at $41.2 billion—a figure that exceeds the market capitalization of companies such as Delta Air Lines, American Airlines and United Continental.”

ii. PUBLIC OPINION: According to a 2015 PWC report:

(1) Among US adults familiar with the sharing economy, 86% agree it makes life more affordable, 83% agree it makes life more convenient and efficient, 72% agree they feel that the sharing economy experience is not consistent, and 69% 3

CALDWELL LESLIE & PROCTOR, PC / 725 S. FIGUEROA ST, 31ST FLOOR / LOS ANGELES, CA 90017 / 213 629 9040 / WWW.CALDWELL-LESLIE.COM

agree they will not trust sharing economy companies until they are recommended by someone they trust.

(2) Among consumer who have tried the sharing economy, 57% agree that “I am intrigued by companies in the sharing economy but have some concerns about them.”

(3) Among consumer who have tried the sharing economy, 64% say that peer regulation is more important than government regulation.

iii. FUTURE: According to a 2015 PWC report:

(1) Among US adults familiar with the sharing economy, 51% could see themselves being providers in the next two years, and 72% could see themselves being consumers in the next two years.

c. Legal challenges

i. Regulatory

(1) Airbnb case study

(2) “We initiate this proceeding to protect public safety and encourage innovators to use technology to improve the lives of Californians. The purpose of this Rulemaking is not to stifle innovation and the provision of new services that consumers want, but rather to assess public safety risks, and to ensure that the safety of the public is not compromised in the operation of these business models. The Commission invites all interested parties to participate in this proceeding to ensure that regulation is not a hindrance, but continues to be the safety net that the public can rely on for its protection.” [California Public Utilities Commission, Order Instituting Rulemaking (2012), Rulemaking 12-12- 011 (discussing transportation network companies)]

ii. Employment / Employee Classification

(1) Lyft case study

(2) Borello factors:

(a) “[T]he principal test of an employment relationship is whether the person to whom service is rendered has the right to control the manner and means of accomplishing the result desired.... [S]trong evidence in support of an employment relationship is the right to discharge at will, without cause.”

(b) Additional factors include: “(a) whether the one performing services is engaged in a distinct occupation or business; (b) the kind of occupation, with reference to whether, in the locality, the work is usually done under the direction of the principal or by a specialist without supervision; (c) the skill required in the particular occupation; (d) whether the principal or the worker supplies the instrumentalities, tools, and the place of work for the person doing the work; (e) the

4

CALDWELL LESLIE & PROCTOR, PC / 725 S. FIGUEROA ST, 31ST FLOOR / LOS ANGELES, CA 90017 / 213 629 9040 / WWW.CALDWELL-LESLIE.COM

length of time for which the services are to be performed; (f) the method of payment, whether by the time or by the job; (g) whether or not the work is a part of the regular business of the principal; and (h) whether or not the parties believe they are creating the relationship of employer-employee.” [S. G. Borello & Sons, Inc. v. Dep’t of Indus. Relations, 48 Cal.3d 341, 350–51 (1989)]

(3) “As should now be clear, the jury in this case will be handed a square peg and asked to choose between two round holes. The test the California courts have developed over the 20th Century for classifying workers isn’t very helpful in addressing this 21st Century problem. Some factors point in one direction, some point in the other, and some are ambiguous. Perhaps Lyft drivers who work more than a certain number of hours should be employees while the others should be independent contractors. Or perhaps Lyft drivers should be considered a new category of worker altogether, requiring a different set of protections. But absent legislative intervention, California’s outmoded test for classifying workers will apply in cases like this. And because the test provides nothing remotely close to a clear answer, it will often be for juries to decide. That is certainly true here.” [Cotter v. Lyft, Inc., 60 F. Supp. 3d 1067, 1081–82 (N.D. Cal. 2015)]

iii. Privacy/Consumer Protection

(1) Case study

(2) FTC Guidelines:

(a) Start with security.

(b) Control access to data sensibly.

(c) Require secure passwords and authentication.

(d) Store sensitive personal information securely and protect it during transmission.

(e) Segment your network and monitor who’s trying to get in and out.

(f) Secure remote access to your network.

(g) Apply sound security practices when developing new products.

(h) Make sure your service providers implement reasonable security measures.

(i) Put procedures in place to keep your security current and address vulnerabilities that may arise.

(j) Secure paper, physical media, and devices.

5

CALDWELL LESLIE & PROCTOR, PC / 725 S. FIGUEROA ST, 31ST FLOOR / LOS ANGELES, CA 90017 / 213 629 9040 / WWW.CALDWELL-LESLIE.COM

iv. Economic & Tort Relationship

(1) Doe v. Uber case study

(2) “In sum, the Court cannot determine – as Uber effectively argues –that as a matter of law sexual assault by Uber drivers is always outside the scope of employment, if the drivers are in fact ultimately found to be employees. The California Supreme Court has left this question open…. Plaintiffs may, or may not, ultimately prevail on their vicarious liability claims…. To the extent that these are close questions, the Court finds that they are more appropriately resolved at a later stage of the litigation.” [Doe 1 et al. v. Uber Technologies Inc., No. 3:15-cv-04670 (N.D. Cal.)]

5. Q&A

6. Resources

a. Sharing economy cases to watch

i. O’Connor v. Uber Techs., Inc., No. C-13-3826 EMC (N.D. Cal.): Misclassification, employment issues

ii. Airbnb Inc. v. City and County of San Francisco, No. 3:16-cv-03615 (N.D. Cal.): Regulatory issues, First Amendment, Communications Decency Act preemption

iii. Doe 1 et al. v. Uber Technologies Inc., No. 3:15-cv-04670 (N.D. Cal.): Tort law, vicarious liability issues, common carrier liability

b. Attachments

i. U.S. Department of Commerce, Digital Matching Firms: A New Definition in the “Sharing Economy” Space (June 2016), available at http://www.esa.gov/sites/default/files/digital- matching-firms-new-definition-sharing-economy-space.pdf.

ii. Molly Cohen & Arun Sundararajan, Self-Regulation and Innovation in the Peer-to-Peer Sharing Economy, 82 U. Chi. L. Rev. Dialogue 116 (2015), available at https://lawreview.uchicago.edu/page/self-regulation-and-innovation-peer-peer- sharing-economy.

iii. Vanessa Katz, Regulating the Sharing Economy, 30 Berkeley Tech. L.J. 1067 (2015), available at http://scholarship.law.berkeley.edu/cgi/viewcontent.cgi?article=2083&context=btlj.

iv. California Public Utilities Commission, Decision Adopting Rules and Regulations to Protect Public Safety While Allowing New Entrants to the Transportation Industry, Rulemaking 12-12-011 (September 19, 2013), available at http://docs.cpuc.ca.gov/PublishedDocs/Published/G000/M077/K112/77112285.PDF.

v. Federal Trade Commission, Start with Security – A Guide for Business (June 2015), available at https://www.ftc.gov/system/files/documents/plain-language/pdf0205- startwithsecurity.pdf.

vi. Cotter v. Lyft (N.D. Cal.) – Order on Motions for Summary Judgment 6

CALDWELL LESLIE & PROCTOR, PC / 725 S. FIGUEROA ST, 31ST FLOOR / LOS ANGELES, CA 90017 / 213 629 9040 / WWW.CALDWELL-LESLIE.COM

vii. Doe v. Uber (N.D. Cal.) – Order on Motion to Dismiss

viii. Doe v. Internet Brands (9th Cir.) – Order on Motion to Dismiss

7

CALDWELL LESLIE & PROCTOR, PC / 725 S. FIGUEROA ST, 31ST FLOOR / LOS ANGELES, CA 90017 / 213 629 9040 / WWW.CALDWELL-LESLIE.COM

U.S. Department of Commerce Digital Matching Firms:Economics A New Definition and in Statistics the “Sharing Administration Economy” Space Office of the Chief Economist

Digital Matching Firms: A New Definition in the “Sharing Economy” Space Executive Summary

Increasingly, consumers and independent service providers are engaging in transactions facilitated by an Internet-based platform. The digital firms that provide the platforms are often collectively referred to as belonging to the “sharing” or “collaborative” economies, among other descriptors. By However, in this paper, we narrow the focus and propose a definition of Rudy Telles Jr. “digital matching firms” that exhibit the following characteristics: 1. They use information technology (IT systems), typically available via web-based platforms, such as mobile “apps” on Internet- ESA Issue Brief enabled devices, to facilitate peer-to-peer transactions. #01-16 2. They rely on user-based rating systems for quality control,

ensuring a level of trust between consumers and service providers who have not previously met.

3. They offer the workers who provide services via digital matching platforms flexibility in deciding their typical working hours. Special thanks to 4. To the extent that tools and assets are necessary to provide a William Hawk and service, digital matching firms rely on the workers using their Jasmine Joung for own. conducting research on digital matching firms, In addition to defining these “digital matching services” the report offers as well as many an initial assessment of its size and scope based on publicly available data substantive suggestions on its largest firms, as well as an examination of its potential effect on (See back cover for consumers and service providers. The report closes with an overview of a full list of the benefits and challenges emerging from the growth of these firms. acknowledgements.)

June 3, 2016

Office of the Chief Economist  Economics and Statistics Administration  Page 1

Digital Matching Firms: A New Definition in the “Sharing Economy” Space Introduction

The Internet—particularly when accessed on smartphones and other mobile devices—is enabling sellers and buyers to conduct market transactions in ways that had not been possible in the past. What began as small and informal online exchanges of goods and services via message boards and rudimentary websites has, with the widespread adoption of fast, reliable mobile smartphones and access to GPS, evolved into a collection of firms that connect millions of consumers with other private citizens who can provide goods and services quickly and efficiently.

One can now open an app and quickly arrange and pay for a car ride; book lodging for the night in a private residence; or arrange for a local provider to clean a house, cook food, or even assemble one’s furniture or mount a television. Conversely, a person with free time and the right combination of skills and/or underutilized personal assets can use these same digital platforms to provide on-demand goods and services for profit, all on his or her own schedule, with low barriers of entry.

In the decade since the emergence of firms such as Uber, a transportation services company, and Airbnb, a platform for travel arrangements and reservation services, the number of people engaged in both obtaining and providing goods and services through digital matching platforms has grown considerably. A small number of digital matching firms are estimated to have valuations that rival many of the world’s largest firms.1

In this paper, we define “digital matching firms” as entities that provide online platforms (or marketplaces) that enable the matching of service providers with customers. By identifying a set of common characteristics that define these “digital matching firms,” we can explore what is new and unique about the phenomenon that is being called (among other names) “the sharing economy.” We then examine the size, scope, and growth of the digital matching firms, with the caveat that there is a relative dearth of public data available on these companies. Finally, we discuss the potential benefits and detriments of the growth in digital matching firms to both the buyers and providers of the services—that is, to consumers and workers. In this final context we also discuss some of the policy challenges that have emerged as some firms using this business model have rapidly expanded and begun to compete with existing firms in established markets.

Defining Characteristics of Digital Matching Firms

The companies that have pioneered this relatively new phenomenon have been classified by a number of names including the “sharing economy,” “e-lancing,” the “ICT-enabled economy,” among others. (See text box: A Plethora of Descriptors and Misnomers: Why We’re Not Describing “Sharing” or “Collaborative” Firms). In our examination of this subset of the broader digital economy, we wish to narrow our focus and define the “digital matching firms” as consisting of firms with business models that exhibit the following characteristics:

1 Uber is currently valued at approximately $62.5 billion (New York Times), while Airbnb has an estimated valuation of more than $25 billion. For comparison, Ford and Honda are worth approximately $60 billion, while GM has a market value of around $55 billion. The Hilton Hotel chain has a valuation of nearly $28 billion.

Office of the Chief Economist  Economics and Statistics Administration  Page 2

Digital Matching Firms: A New Definition in the “Sharing Economy” Space

1. Digital matching firms use information technology (IT systems), typically available via web-based platforms such as mobile “apps” on Internet-enabled devices, to facilitate peer-to-peer transactions.

Digital matching firms have introduced a variety of apps and other Internet platforms that provide a marketplace for secure, reliable, and efficient transactions between individuals. Digital matching platforms often allow individuals to access peer-to-peer services in real-time and also allow the digital matching firms to handle the financial transaction between the consumer and provider. For example, an Uber passenger pays for her ride using a credit card via the Uber app itself, with the firm paying the driver. This contrasts with a traditional taxi ride, during which the passenger pays the driver directly. In short, the worker providing the service has no role in collecting payment from the consumer.

2. Digital matching firms rely on user-based rating systems for quality control, ensuring a level of trust between consumers and service providers who have not previously met.2

In order to facilitate peer-to-peer transactions, digital matching firms all utilize some form of rating system to ensure a level of trust between individuals that are most often strangers. In addition to requiring public disclosure of aggregate ratings, service providers are often required to maintain a consumer feedback rating above a certain threshold in order to continue providing services via their platforms. Our definition only requires that a rating system is in place to evaluate service providers, but many rating systems are bilateral, giving service providers a sense of security about the integrity of the person to whom they are, for example, renting out an asset.

3. Individuals who provide services via digital matching platforms have flexibility in deciding their typical working hours.

Service providers for digital matching firms have the work flexibility of traditional freelance workers, which is why they are often referred to as “e-lancers.” Individuals only offer services when they choose, assuming they meet conditions that the digital service firm may set, such as: maintaining adequate user feedback ratings; having government-required licensing, training, and insurance; and having quality assets. As such, digital matching firm service providers, who are often not legally classified as “employees” of the digital matching firm, are often not required to be on call or work a specific amount by the digital matching firm in order to be eligible to provide services in the future.

2 It is worth noting that peer-to-peer financial services firms such as Lending Club or Funding Circle may or may not be considered digital matching firms, depending on how one interprets both the underlying peer-to-peer nature of the loans themselves and the necessity of a peer-to-peer rating system. Financial firms have robust rating systems, such as credit scores, that are independent from the peer-to-peer rating systems typical of a transportation, lodging, or peer-asset rental platform, but provide a mechanism for the trust necessary for consumers to use digital matching firms. In this paper, we will consider peer-to-peer lending companies as digital matching firms.

Office of the Chief Economist  Economics and Statistics Administration  Page 3

Digital Matching Firms: A New Definition in the “Sharing Economy” Space

4. To the extent that tools and assets are necessary to provide a service, digital matching firms rely on the workers using their own.

Unlike employees of traditional firms that use assets owned by the firm to perform services, these workers either own or have personal access to the assets that are used to provider services. Often, the assets must meet a set of criteria that the firm dictates. For example, Lyft requires that vehicles used by their service providers be from 2004 or later.3

Box 1. A Plethora of Descriptors and Misnomers: Why We’re Not Describing “Sharing” or “Collaborative” Firms

Digital matching firms have been referred to by many names and descriptors. Among the most ubiquitous labels for the collection of these firms in both media and academic reporting are the “sharing” and “collaborative” economies. However, terms such as “sharing” and “collaborative” do not adequately characterize what makes firms like Uber, Taskrabbit, and Airbnb innovative. As discussed at length in this report, these firms provide a platform for consumers and service providers to connect and complete a transaction safely and efficiently, using the capital assets of the service providers themselves, when such assets are required to provide the service.

Service providers using their own underutilized assets to provide a service are often characterized as “sharing” or “collaborating” with consumers, but this implies services being provided for free; the reality is that the so-called “sharing” or “collaboration” in these cases is not free. Service providers are simply using their assets to earn money. In digital matching firms, service providers bear the cost and risk of providing a service and, in many cases, use an asset they already own for another purpose, but they are not “sharing” their assets any more than a traditional taxi company is sharing its cars or a hotel is sharing its rooms. There are some true “sharing” economy platforms that help individuals provide their assets to others free of charge, such as Freecycle, which provides a place for people to give away their possessions. However, Freecycle is essentially free retail, and retail is not in the scope of our definition of digital matching services.

Similarly, a true “collaborative” economy consisting of individuals utilizing online platforms to provide services and/or produce products also exists, as in the environments that produced the UNIX operating system or R statistical software, for example. However, the types of activities conducted through the “collaborative” economy do not accurately describe the kinds of transactions conducted via digital matching platforms.

The use of a narrow set of conditions that define a digital matching firm is an intentional effort to separate firms using an innovative online business model from firms using more standard business

3 According to Lyft’s published requirements for Lyft vehicles, some states and cities require that service providers use a newer model vehicle, such as Seattle, where service providers must have 2006 or more recent vehicle models.

Office of the Chief Economist  Economics and Statistics Administration  Page 4

Digital Matching Firms: A New Definition in the “Sharing Economy” Space models that have a strong online component and from online activities that cannot be characterized as being provided by “firms.” (See text box: “Sharing” Firms that Are Not Digital Matching Firms).

Many digital matching firms share certain other attributes, but we did not consider them decisive or critical for defining digital matching firms because not all digital matching firms or the services provided by these firms have these attributes. For one, service providers utilizing digital matching platforms use capital assets that they already own to provide services, if such assets are required, but there are also providers that purchase or rent assets specifically to provide services via digital matching platforms.

For example, an individual may purchase a car via Uber’s partnership with a network of car dealers and lenders for the express purpose of providing transportation services through the Uber app, or lease a vehicle on a short-term basis for use with the Lyft or Uber platform; rental companies such as Breeze exist to provide vehicles for digital matching transportation firms. A provider may also purchase a condominium to rent out on Airbnb, or purchase a bicycle or tools to use for paid tasks such as delivering packages or assembling furniture via the Taskrabbit app.

While digital matching firms may set quality standards for their service providers, those standards may differ from those of their non-digital matching competitors. For instance, digital matching firm service providers may lack occupational licenses in the industries in which they provide services, and may be considered “amateurs.”4 However, digital matching firm service providers are not always amateurs, and some digital matching apps connect consumers with professionals.

Along with the wide variety of services offered are a range of pricing structures, as lodging digital matching platforms, such as Airbnb, and task platforms, such as Taskrabbit, allow service providers to set their own rates, relying on consumers and service providers to adjust prices themselves. Other firms, such as Uber and Lyft, set prices internally, and both consumers and service providers are reliant on the digital platform to determine cost of service.

4 For example, many localities, such as New York City, require that individuals complete a certification course and/or obtain a chauffeurs license before they can legally drive a taxi.

Office of the Chief Economist  Economics and Statistics Administration  Page 5

Digital Matching Firms: A New Definition in the “Sharing Economy” Space

Box 2. “Sharing” Firms that Are Not Digital Matching Firms5

Many companies that are commonly classified as members of the “sharing” and/or “collaborative” economy fall outside of the scope of our definition of digital matching firms. These include:

1. Firms that provide online classifieds such as Craigslist do, in fact, match consumers with goods and service providers, but lack rating systems and also do not process transactions via their own digital platform.

2. Companies that provide assets that are shared by multiple consumers on an ad-hoc basis, such as “bikesharing” and “carsharing” firms or movie rental kiosks. We exclude these firms because the assets provided are owned by the firm itself on a self-service basis. These firms operate more like rental services but without the need for a staff of retail salespeople.

3. Online retailers such as Amazon since a large portion of their sales consist of items warehoused by Amazon itself or provided via authorized, traditional third-party retailers. (In some cases, however, these large retailers have subsidiaries that provide services that fit our definition. Amazon Mechanical Turk, for example, connects consumers with freelancers who provide a service for a fee.)

4. Firms that facilitate the matching of a service without facilitating a monetary transaction. Examples of these platforms include couchsurfing, freecycle, Maine Tool Library, Neighborgoods. These firms are appropriately classified as part of the “sharing economy,” as the peer-to-peer transactions that take place via these apps involve the sharing or giving away of goods and services, often for altruistic purposes.

The Size and Scope of Digital Matching Firms

The apparent dramatic emergence of digital matching firms begs a number of questions about their size and scope, including what are the total revenues of all digital matching companies and how many people are engaged in providing services within this paradigm?

As the discussion below indicates, the evidence for definitive answers to these questions is limited. Though there are many high-profile privately held startups for which there are estimated large market values, much of the market analysis that has been done to date is broad in scope and speculative — there is little systematic information on the number and characteristics of individuals acting as providers on the digital matching platforms. Information about the customer experience and consumer surplus from using digital matching platforms versus more traditional business models is often anecdotal,

5 For a list of firms that do seem to qualify as digital matching firms, refer to the “Appendix: Examples of Digital Matching Firms” following the conclusion of this report.

Office of the Chief Economist  Economics and Statistics Administration  Page 6

Digital Matching Firms: A New Definition in the “Sharing Economy” Space although some solid economic research on specific markets—such as the price and market effects with respect to specific firms in specific cities—is beginning to emerge.

Estimates of the Size and Growth of Digital Matching Firms

Relatively little economic research on digital matching firms exists, and teasing out data about the size and value of these firms is difficult. Most digital matching companies are private and not subject to the disclosure requirements of publicly-held companies, limiting the availability of reliable data on factors such as yearly revenues. Therefore, estimates of the size and growth of digital matching firms tend to come from private-sector surveys of consumers and service providers.

PricewaterhouseCooper (PwC), consulting firm MBO Partners, investment research group PiperJaffrey, and the JPMorgan Chase Institute have each released reports that attempt to estimate the size and growth of the “sharing” and “collaborative” economies. These studies inevitably include many firms and industries that fall outside of the scope of our analysis. However, it’s worth examining the few studies that do exist, as many of the companies we classify as digital matching firms are represented in their analyses. Even using a broader definition than the one we propose, all of these studies suggest that the “sharing” economy comprises a relatively small portion of the overall economy.

In 2014, a study by PwC6 presented an estimate that five key sharing sectors—travel, car sharing, finance, staffing, and music and video streaming—had global revenues of about $15 billion in 2014 with the potential to increase to around $335 billion by 2025. In addition, PwC surveyed 1000 consumers in order “to comprehend consumer attitudes toward the sharing economy.” According to the PwC data, 8 percent of all adults have participated in some form of automotive sharing, and 1 percent have served as providers under this new model, “chauffeuring passengers around or loaning out their car by the hour, day or week”. The PwC study also suggests that service providers in the “sharing economy,” which they estimate to comprise 7 percent of the U.S. population, are made up of a wide variety of age and income groups though their estimates include firms that we would not consider digital matching firms.

The consulting firm MBO Partners produces an annual report titled the “State of Independence in America”7 that examines the U.S.’s “independent workforce, or those who work 15+ hours a week as an independent contractor.” In a proprietary supplement to this report titled “Independent Workers and the On-Demand Economy,” MBO estimates that 2.7 million Americans, or 9 percent of independent workers provide services through “on-demand economy platforms,” and that roughly 500,000 of the estimated 2.7 million U.S. on-demand independent workers are estimated to provide services for Uber, Lyft, and Airbnb, suggesting that service providers in the digital matching economy are concentrated in a small number of firms. Further the report found that, of those surveyed, independent workers in the on- demand economy reported lower earnings than independents not using these platforms and marketplaces. The data shows that 36 percent of independent workers using on-demand platforms reported earnings of $25,000 or less compared to 22 percent of independent workers not using such platforms. At the other end of the spectrum, only 17 percent of those providing services through on- demand platforms reported earning $75,000 or more; that compares to 28 percent of independent

6 PricewaterhouseCooper. “The Sharing Economy.” Consumer Intelligence Series. April 2015. 7 MBO Partners. “MBO Partners State of Independence in America 2015.”

Office of the Chief Economist  Economics and Statistics Administration  Page 7

Digital Matching Firms: A New Definition in the “Sharing Economy” Space workers that do not use on-demand platforms. However, like the PwC survey, MBO Partners included many companies in their report, such as Amazon and Ebay, which do not fit our criteria for digital matching firms.

The JPMorgan Chase & Co. Institute released a report titled “Paychecks, Paydays, and the Online Platform Economy: Big Data on Income Volatility,”8 that attempted to estimate the effect of what they call the “online platform economy” on income volatility. The authors used an anonymized sample of 1 million people who were customers of JPMorgan between October 2012 and September 2015, and a dataset of more than 260,000 individuals who have offered goods or services on an online platform. They estimate that more than 4 percent of adults, or approximately 10.3 million people, participated in the “online platform economy” over the three-year period of their study, and that 1 percent of adults earned income from an online platform in a given month. Moreover, the Institute estimated a 47-fold increase in the number of adults that earned income from online platforms over the course of the three- year period.

Investment research group PiperJaffray produced a report titled “Sharing Economy: An In-Depth Look At Its Evolution & Trajectory Across Industries”9 that estimated total “sharing” revenues from short-term person-to-person (P2P) home rentals, such as Airbnb, at 2 percent of the U.S. accommodations market, which includes hotels, hostels, bed and breakfasts, cruises and other short-term and P2P rentals. However, this report predicts that by 2025, P2P home rentals could represent as much as 10 percent of accommodation bookings, with revenue of $107 billion. In addition, Uber and other “ridesharing” companies are estimated to make up more than 5 percent of the $90 billion global taxi marketplace.

As noted earlier, the information available about the collective size of digital matching firms is sparse, the reports consist of small survey samples, and every study includes firms that that are outside the scope of our definition. For that reason, these estimates are not entirely applicable for our purposes. However, despite the inclusion of firms and industries that do not reflect the digital matching economy, these studies suggest that digital matching firms are quickly growing in size, yet remain a relatively small part of the greater U.S. economy.

The Largest Digital Matching Firms

Current estimates of the size and growth of the ”sharing” economy may not be appropriate for our purposes given the inclusion of firms that fall outside of our defined digital matching firm parameters. However, publicly released estimates of the size and growth of the largest individual digital matching firms are available and provide an insightful, if imperfect, glimpse into the rapid growth of the most successful firms.

8 JPMorgan Chase & Co. Institute. “Paychecks, Paydays, and the Online Platform Economy: Big Data on Income Volatility.” February 2016. 9 Olson, Michael J., Samuel J. Kemp. “Sharing Economy: An In-Depth Look At Its Evolution and Trajectory Across Industries.” PiperJaffray Investment Research. March 2015.

Office of the Chief Economist  Economics and Statistics Administration  Page 8

Digital Matching Firms: A New Definition in the “Sharing Economy” Space

Market intelligence group VB Profiles10 estimated that, worldwide, there are now 17 companies in the sharing or collaborative economies each worth more than $1 billion, with 60,000 employees and $15 billion in funding. As with the other studies discussed above, we would not consider many of these companies part of our more narrow definition of digital matching firms, but the list does include digital matching companies such as Uber, Lyft, and Airbnb, as well as other digital matching companies such as Chegg, which specializes in online textbook rentals.

 Uber—a privately held company—is the largest digital matching firm based on market valuation. The Wall Street Journal reported that its market value was estimated to be $62.5 billion in December 2015, up from $60 million in 2011.11 If this is accurate, its current valuation is higher than 80 percent of all S&P 500 companies. Reuters forecasts that Uber, from the 20 percent cut it takes from every ride, will generate approximately $2 billion in revenue worldwide in 2015.12

 Airbnb—also a privately held firm—is the second largest digital matching firm based on market valuation and the largest lodging accommodations provider among digital platform firms. The Wall Street Journal reported Airbnb’s estimated value at more than $25 billion, which is more than that of the Marriott hotel chain.13 The valuation increased from approximately $10 billion in April, 2014.

10 Koetsier, John. “The sharing economy has created 17 billion-dollar companies (and 10 unicorns).” VentureBeat. June 4 2015. 11 Isaac, Mike, Leslie Picker. “Uber Valuation Put at $62.5 Billion After a New Investment Round.” The New York Times. December 3, 2015. 12 Zhang, Shu, & Gerry Shih. “Uber seen reaching $10.8 billion in bookings in 2015: fundraising presentation.” Reuters. August 21, 2015. 13 Alba, Davey. “Airbnb Confirms $1.5 Billion Funding Round, Now Valued at $25.5 Billion.” Wired. December 7. 2015.

Office of the Chief Economist  Economics and Statistics Administration  Page 9

Digital Matching Firms: A New Definition in the “Sharing Economy” Space

Box 4. Disruption and Convergence

As with the introduction of e-commerce in the 1990s, Internet-based technologies in the form of digital matching apps have the potential to disrupt existing markets. The growth of the digital matching firms appears to have begun to cause some disruption in traditional industries such as transportation services and lodging, with the potential to do so in a variety of other industries.

Airbnb and other lodging-centric digital matching firms may have already had an effect on hotel revenues in some areas. According to a Boston University study titled “The Rise of the Sharing Economy: Estimating the Impact of Airbnb on the Hotel Industry,” each additional 10 percent increase in the size of Airbnb listings in Texas resulted in a 0.37 percent decrease in monthly hotel revenues. There is also some evidence that digital matching firms may have had an effect on the supply of long-term rentals in some areas, as some landlords in major cities, such as New York City, have chosen to operate homes as short-term rentals via Airbnb rather than lease them in a more traditional manner, decreasing the supply and potentially raising the price of rental properties within that market. To combat what they consider ad-hoc hotels, regulators in New York City have since proposed heavy penalties for property owners who violate the city’s ban on short-term rentals.14

In the transportation industry, the rise of firms such as Lyft and Uber likely have negatively affected the value of taxi medallions in New York City, as the price fell to roughly $805,000 in early 2015, down 23 percent from 2013’s peak of $1.05 million; corporate medallions, which may be owned in fleets, were down 28 percent from their peak.15 Taxi industry revenue has fallen considerably in a number of cities as well; in Seattle, taxi revenues dipped 28 percent in two years.16

At the same time, in an effort to compete with digital matching firms, traditional industries are beginning to incorporate digital matching technology into their services, in a process known as convergence, lowering their own costs and improving the consumer experience. For example, the Curb app for taxi services works much like Uber, connecting consumers with taxi drivers representing 90 taxi companies in 60 cities and allowing consumers to pay for rides via the app.

Along with incorporating technologies from digital matching firms into their business models, some regulatory hurdles are being modified to help make traditional firms more competitive with digital matching firms. For example, the New York City Taxi Commission has removed geography questions from the taxi license test in response to a decline in the number of driver applicants, while also acknowledging that reliable GPS technology has made rote knowledge of the New York City area less important for driver success and customer satisfaction.17

14 Gonzalez, Juan. “NYC Council to propose tough penalties for landlords who use sites like Airbnb, in effort to keep affordable housing.” New York Daily News, June 10, 2015. 15 Barro, Josh. “New York City Taxi Medallion Prices Keep Falling, Now Down About 25 Percent.” The New York Times, Jan. 7, 2015. 16 Samuelson, Rob. “Seattle taxi revenue dropping precipitously due to Uber and Lyft.” Seattle Sun Times. June 13, 2015. 17 Worland, Justin. “Cab Drivers No Longer Required to Learn N.Y.C.’s Streets.” Time. March 9, 2015

Office of the Chief Economist  Economics and Statistics Administration  Page 10

Digital Matching Firms: A New Definition in the “Sharing Economy” Space Benefits and Challenges Introduced by Digital Matching Firms

Benefits of Digital Matching Platforms

With the potentially rapid growth of digital matching firms, an important question is whether or not they benefit consumers, workers, and the overall economy. Given the developing nature of the sector, there is insufficient data to make any definitive judgments. However, given its inherent characteristics, digital matching firm technology has the potential to provide a number of benefits. This section explores the benefits often associated with the digital matching platforms.

1. Provides Lower Prices for Consumers Due to Reduced Transaction and Overhead Costs for the Service Provider: Transaction costs are the time, money, skill, and effort needed to facilitate a market transaction. Every day, consumers demand goods and services that could be provided by professionals and non-professionals in their communities. These market exchanges are often facilitated through firms, brokers, and sometimes government agencies. Digital matching platforms potentially reduce the costs of coordinating these transactions by connecting consumers with service providers directly and often in real-time, ostensibly cutting out the traditional firm and middlemen that would otherwise be needed to link them.

There is some evidence that these lower costs for the service provider has resulted in lower prices for consumers. For example, a Business Insider article reported that in 2014, an Uber ride was less expensive than a taxi in all but two of the 21 large cities studied, so long as surge pricing wasn’t activated.18 In addition, the previously discussed PwC survey found that, of those polled, 56 percent cited “better pricing” as the reason for their preference for “automotive sharing economy models.” The PiperJaffray report, which was also previously discussed, found that private accommodations available through digital platforms, such as Airbnb, are generally less expensive than hotels in cities throughout the world (see table).

18 The price of Uber services will rise during “peak” periods, when consumer demand for rides is highest. According to Uber’s FAQ, “At times of high demand, the number of drivers [Uber] can connect you with becomes limited. As a result, prices increase to encourage more drivers to become available.” These often include periods of inclement weather, holidays, and near areas in which special events are taking place. Given that Uber surge pricing is variable and that rates can climb by many multipliers of the base fare, it’s possible that consumers, on the whole, pay more for Uber.

Office of the Chief Economist  Economics and Statistics Administration  Page 11

Digital Matching Firms: A New Definition in the “Sharing Economy” Space

Table 1: Hotel vs. Private Rental Costs Throughout the World City Private Rental Index ($) Hotel Index ($) Difference (%) Singapore 68 202 67 Seoul 48 142 66 Rio de Janerio 102 257 60 Hong Kong 73 174 59 Barcelona 65 149 56 Zunich 87 199 56 New York 114 255 55 London 75 165 55 Istanbul 69 151 55 Budapest 45 99 54 Cannes 108 231 53 Palma de Mallorca 73 150 52 Sydney 94 195 52 Frankfurt 76 159 52 Melbourne 75 155 52 Los Angeles 94 192 51 Florence 76 152 50 Nice 89 174 49 Paris 98 183 47 Munich 87 159 46 Miami 115 214 46 Rome 87 155 44 Lisbon 63 113 44 Madrid 65 115 44 Milan 84 152 44 Seville 64 113 43 Prague 64 112 42 Berlin 65 112 41 Vienna 77 130 40 Venice 119 183 35 Brussels 93 136 32 Source: Study commissioned by W imdu and converted from EUR to USD by Piper Jaffray

2. Provides Flexible Employment Schedules and Additional Income for Workers: People who need extra income and/or can’t work traditional hours are often able to provide services via digital matching firms. Low barriers of entry and the utilization of ubiquitous, common capital assets, such as cars, bicycles, and extra bedrooms allow individuals to work during their “off” hours or while they’re otherwise unemployed19. For example, in a survey commissioned by Uber, 80 percent of their “driver-partners” were working full or part-time jobs just before they started driving on the Uber platform, and two-thirds of that group reported having a full-time job. In addition, of those

19 Evidence indicates that firms have encouraged individuals to purchase capital assets such as cars for use with the digital matching app. For example, Uber has a vehicle financing service that connects borrowers, including those with poor credit, with auto dealers. Individuals who purchase an asset to use specifically with a digital matching app in-fact may be losing the flexibility benefits that are ostensibly one of the draws of being a service provider for a digital matching firm, as they are now responsible for the payments and maintenance of that asset, and thus must work.

Office of the Chief Economist  Economics and Statistics Administration  Page 12

Digital Matching Firms: A New Definition in the “Sharing Economy” Space

surveyed, more than half had never previously worked as a driver, whether it be for a taxi, limo, or other for-hire transportation company, suggesting that the Uber platform provided an introduction into a new line of full or part-time work for the majority of its service providers.20

Further, the previously discussed JPMorgan Chase Institute study found that earnings from labor platforms helped to offset low or zero-income periods for workers with high levels of income volatility, notably when they were between jobs and when their income dipped. Although the number of people participating in what they call the “online platform economy” increased tremendously during the three-year period of their study, individuals mostly utilized online platforms as a secondary source of income, and their reliance on platforms for income remained stable over time in both the fraction of months that participants were active and the fraction of total income earned on platforms in active months.

Aside from providing employment opportunities for the unemployed and workers who require supplemental incomes, digital matching platforms also offer opportunities for non-traditional working populations, such as retired people and individuals with disabilities or health issues. Some companies are actively recruiting senior citizens; Uber, for example, announced a partnership with AARP's Life Reimagined, which would give members who sign up to be new drivers a bonus after they provide 10 rides through the service.

3. Leverages Excess Capacity: Digital matching firms provide a platform for service providers to take advantage of underutilized assets. Turo, for example, capitalizes on the existence of idle private vehicles by allowing users to rent out their cars to others when they’re not using them. Transportation services provided by Uber and Lyft capitalize on both underutilized cars that are theoretically sitting unused and drivers with both the time and desire to work. Rooms listed for rent on Airbnb or HomeAway are often guest rooms or in houses that are currently vacant due to vacation, travel, or other life events.

4. Potentially Stimulates New Consumption: By providing consumers access to services that were previously either unavailable or less convenient, digital matching firms may be able to access untapped markets and increase overall consumption. However, it’s possible that total consumption in the economy could actually decrease as consumers shift away from the more traditional economy. For example, if urban consumers begin using digital matching apps for their transportation needs to a large enough degree, they may hold off on purchasing a car, which could potentially decrease overall consumption in the economy. Reliable data examining the stimulative effect of digital matching firms is currently sparse.

5. Improves the Consumer Experience: The innovations introduced by digital matching firms could considerably lessen the inconvenient aspects of service transactions, increasing consumer welfare. For example, both Lyft and Uber allow consumers to pay for their services via their respective apps, removing the post-ride in-person transaction that is often required when they use traditional taxi

20 Hall, Jonathan V, & Alan Krueger. “An Analysis of the Labor Market for Uber’s Driver-Partners in the United States.”

Office of the Chief Economist  Economics and Statistics Administration  Page 13

Digital Matching Firms: A New Definition in the “Sharing Economy” Space

services. The apps also utilize GPS technology to allow consumers to track their driver so that they know in real-time when he or she will arrive.

6. Provides a Mechanism for Trust between Consumers and Individual Service Providers: Digital matching firms, via rating systems within their platforms, have provided the consumer an efficient mechanism through which they are willing to trust complete strangers to provide goods and services. Relying on crowdsourced information to establish trust between a consumer and a company is not new, as people have long used relatives, friends, co-workers, and neighbors to choose a company or specific service provider. However, robust public ratings systems may be a more efficient guidepost when deciding whether or not, for example, they will stay in a spare bedroom or have a stranger clean their house, mount their TV, or cook their food. Further, many digital matching firms are able and willing to ban service providers who fall below ratings thresholds, acting as an incentive for better service.

Challenges Introduced by the Digital Matching Platforms

Digital matching firms and their technologies have the potential to provide a number of benefits, but there are possible downsides to the emergence of these firms, most notably to service providers themselves. Partly because service providers are typically not classified as employees of the firm, risks are often shifted from the digital matching firm (that provides the platform) to the service provider (often an individual). There are also potential concerns about customer privacy and access to these services that need to be considered when evaluating the overall costs and benefits of these new services.

1. Potential Income Instability: Service providers in the digital matching economy are fully reliant on the digital matching platform’s ability to connect them with consumers, and they are not guaranteed to be matched. Thus, service providers in the digital matching economy are often unsure at any given time whether or not their services will be in demand. Also, in the case of digital matching firms that set rates themselves, service providers are unsure of the price until they begin providing those services, and the prices may change at any time.

2. Fewer Benefits and Protections for Service Providers: Since many digital matching firm service providers are classified as independent contractors, they are not eligible to receive many benefits, such as a minimum wage, overtime pay, health and life insurance benefits, collective bargaining rights, retirement and savings plans, protections from discrimination, and sick leave. Workers who sign up for their own benefits must additionally devote their own unpaid time to what is normally provided by human resource departments. In addition, service providers are often not compensated if, for instance, a client is running late or reneges on a service request.

3. Service Providers are Responsible for their Own Training: A digital matching company lacks the incentive to train its service providers lest they be classified as employees. For that reason, service providers either must already possess the knowledge and experience to provide a service or are forced to train themselves. Thus, for example, laborers providing handyman services via platforms such as Taskrabbit, must either already know how reliably to mount a TV, or teach themselves to do so.

4. Capital Investment and Maintenance Costs are the Responsibility of the Service Provider: Digital matching firms rely on service providers to use and maintain their own capital assets. If, for

Office of the Chief Economist  Economics and Statistics Administration  Page 14

Digital Matching Firms: A New Definition in the “Sharing Economy” Space

example, a service provider’s car breaks down or their tool malfunctions, he or she must cover replacement and repair costs. Further, when providing services through car transportation platforms such as Uber, service providers are responsible for fuel costs, depreciation, and insurance coverage.21

5. Consumer Privacy and Security: Digital matching firms by their very nature collect and have access to a substantial amount of consumer and service provider information, whether it be a consumer’s credit card information, home address, location, or travel history. As with all firms that conduct business via the digital economy, the safe handling and legal usage of such data by digital matching firms must be considered.

6. Access: In order to utilize digital matching platforms as either a consumer or a service provider, one must at least have access to the Internet, and also, in many cases, a smartphone. According to the Pew Research Center22, about two-thirds of American adults now own a smartphone, up from 35 percent in 2011. Although smartphone access has grown considerably, one-third of U.S. adults are effectively unable to utilize many digital matching applications without assistance, and many of those without smartphones are those with lower levels of educational attainment and those on the lower end of the socioeconomic spectrum. For instance, only 50 percent of Americans making less than $30,000/year own a smartphone, compared with 84 percent smartphone ownership among those making $75,000/year or more. Only 52 percent of Americans with a high school degree or less own a smartphone compared with 78 percent ownership among those with college degrees. Integrating Digital Matching Firms into the Regulatory Framework

When startups with innovative business models emerge, it may take time to figure out how they fit into the regulatory framework. For example, the rise of what are now considered traditional online retailers such as Amazon and Ebay brought with them a multitude of complicated policy issues that are still being debated, such as how, where, and when to tax purchases.23 If regulations are inequitable, this may lead to market distortions. The purpose of this section is to provide an overview of several of the issues that have emerged to-date as some of the prominent digital matching firms have increased their market share over the past few years. Although not a comprehensive list, these issues include:

1. Worker Classification: Currently, many digital matching firm service providers are classified as independent contractors, and not employees. As discussed earlier, in the United States this distinction carries with it differences in rules and regulations related to areas such as unemployment insurance, workers’ compensation, training, and health insurance coverage. The IRS has a list of 20 factors that “may be examined in determining whether an employer-employee relationship exists.” These factors include worker training and set hours of work. Government regulators are examining

21 https://www.uber.com/driver-jobs. Viewed on 3/24/2016. 22 Pew Research Center, “U.S. Smartphone Use in 2015.” April 1, 2015. 23For example, Amazon currently only collects sales taxes for transactions that take place in just over half of U.S. States.

Office of the Chief Economist  Economics and Statistics Administration  Page 15

Digital Matching Firms: A New Definition in the “Sharing Economy” Space

whether or not these service providers should be classified as employees, and some are ruling that they should be. California’s Labor Commission, for example recently ruled that Uber drivers should be classified as employees. The U.S. Department of Labor also recently reemphasized concern over companies claiming their workers as independent contractors when they should be employees, although the guidance did not specifically reference the “sharing” economy or any of the other commonly used designations.24 However, in anticipation of similar rulings, some companies, such as Shyp, have begun converting their independent contractors into employees.

Further, companies are required to withhold income taxes, pay unemployment taxes, and pay and withhold Social Security and Medicare taxes for workers classified as employees. As mentioned above, California regulators recently ruled25 that Uber drivers should be classified as employees, not independent contractors. If that ruling stands, Uber will be required to incur the administrative costs necessary to collect these employment-related taxes, as well as pay the Federal Unemployment tax (FUTA) and cover half of their employees’ social security and Medicare taxes26.

The sharp disparity between the way contractors and employees are regulated has led some27 to question whether a third worker classification should be enacted that covers workers who fall somewhere between independent contractors treated as self-employed businesspeople and traditional employees that are generally entitled to certain benefits and worker protections.

Our understanding of the extent of the worker classification issue is challenged because of the limited availability of data on this segment of the workforce, although the federal government is currently conducting several efforts to collect better data on the subject. (Box 5. Expanding the Collection and Availability of ‘Sharing’ Firm Data).

2. Taxation and Compliance: The applicability of hotel taxes to room and residence rentals via Airbnb and other lodging-specific digital matching firms have been raised in a number of localities. Initially, digital matching firms did not require that service providers pay lodging taxes that are typically required of hotels and other lodging establishments, potentially reducing government revenue and creating a competitive advantage for lodging-specific digital matching firms. Many localities, such as Santa Monica, California have banned the use of Airbnb-like services for short-term lodging unless the service provider obtains a business license and pays a hotel tax.28 In response, Airbnb has agreed to collect taxes in several cities, including the District of Columbia and Portland, Oregon in order to meet local tax collection responsibilities while not burdening potential service providers with the need to apply for licenses or collect taxes themselves.

24 https://www.dol.gov/whd/workers/misclassification/ai-2015_1.htm 25 http://www.reuters.com/article/us-uber-tech-drivers-lawsuit-idUSKCN0Y02E8 26 https://www.irs.gov/businesses/small-businesses-self-employed/understanding-employment-taxes 27 For example, The Hamilton Project at the Brookings Institute released a report titled “A Proposal for Modernizing Labor Laws for Twenty-First-Century Work: The “Independent Worker” that examined alternative employee classifications for gig economy workers. 28 Sam Sanders, ”Santa Monica Cracks Down On Airbnb, Bans 'Vacation Rentals' Under A Month,” NPR, May 13, 2015, http://www.npr.org/sections/thetwo-way/2015/05/13/406587575/santa-monica-cracks-down-on-airbnb- bans-vacation-rentals-under-a-month

Office of the Chief Economist  Economics and Statistics Administration  Page 16

Digital Matching Firms: A New Definition in the “Sharing Economy” Space

Box 5. Expanding the Collection and Availability of ‘Sharing’ Firm Data

Given the relative infancy of digital matching firms and the broader “sharing economy,” limited federal data is available about the size, scope, and growth of these firms or about the makeup of their employees and contractors. The lack of data makes it difficult for researchers and policymakers to study trends in this area. However, efforts are underway to expand the availability of data on these firms and the workers that provide services through them.

For example, the Department of Labor and the U.S. Census Bureau will reintroduce the Contingent Worker Supplement (CWS) as part of the Current Population Survey in 201729. The CWS was conducted 5 times from 1995 to 2005 in an attempt to measure more accurately the size of the contingent workforce30. With adjustments to reflect questions relevant to digital matching firms or to the ‘sharing economy,’ the CWS could be an important source of data on workers, both within the government and in the private sector, who participate in these parts of the economy but are not traditional employees.

Further, expanding the ability of federal statistical agencies to use limited Federal tax information holds promise for improving data on digital matching firms or the broader ”sharing economy.” Such data access potentially would enable the statistical agencies to measure income from sources such as payments made to a person who is not an employee, sources which are particularly relevant to these parts of the economy. Allowing statistical agencies access to this type of data would require a change in the tax code to expand the use of tax information for statistical purposes; for example, under current law, the Bureau of Economic Analysis only has access to Federal tax information of corporations (FTI) and the Bureau of Labor Statistics has no access to FTI for use in the statistics it produces. This barrier to accessing business tax information is a roadblock preventing measurement of the sharing economy’s financial size and employment scope.

3. Equal Access to Services for Individuals with Disabilities: A large number of businesses in the United States are included in the 12 categories that are considered “public accommodations” and are therefore covered by the American with Disabilities Act (ADA), including restaurants, hotels, movie theaters, schools, day care facilities, recreation facilities, taxi services and doctors’ offices. For instance, taxis services are required to ensure that a certain percentage of their fleets are equipped to transport passengers with disabilities; in the District of Columbia, each taxi and sedan company

29 Secretary Tom Perez, “Innovation and the Contingent Workforce,” U.S. Department of Labor blog, January 25, 2016. 30 According to the Department of Labor, contingent workers are “persons who do not expect their jobs to last or who reported that their jobs are temporary. They do not have an implicit or explicit contract for ongoing employment. Alternative employment arrangements include persons employed as independent contractors, on- call workers, temporary help agency workers, and workers provided by contract firms.”

Office of the Chief Economist  Economics and Statistics Administration  Page 17

Digital Matching Firms: A New Definition in the “Sharing Economy” Space

with 20 or more vehicles must dedicate a portion of its fleet to wheelchair accessible vehicles.31 Hotels and even bed and breakfast facilities must comply with ADA regulations regarding architectural barriers. Digital matching firms that provide transportation services may not be equipped to provide service to the disabled, and it is unclear whether or not they are required to do so under the ADA or related statutes and regulations. Also not clear is whether authorities require most rooms and houses listed on websites of lodging-specific digital matching firms to be ADA compliant.

4. Consumer Safety and Service Provider Certification: Traditional firms must often pass rigorous regulatory checks, such as health and safety inspections in hotels and in restaurants, to ensure that their services are safe for consumers. Service providers in certain traditional industries are also often subject to additional screening and certification requirements, such as ensuring the contractors they employ are licensed to conduct handyman services or have taxi licenses for their cabs. Digital matching firms may not meet these same consumer safety requirements. Airbnb and other lodging-specific digital matching firm service providers, for instance, are not subject to the health and sanitation inspections common among hotels and bed and breakfast facilities. Providers of handyman services through digital matching platforms may not have the required contractors’ licenses to do specific requested tasks. Firms such as Taskrabbit hedge against such issues by providing insurance coverage in the event of an accident, but their service providers potentially remain unlicensed.

Traditional service providers must also comply with federal, state and local environmental regulations to which digital matching service providers may not be subject. Hotels, for example, must adhere to a number of basic federal requirements under the Clean Air Act, Clean Water Act, and the Resource Conservation & Recovery Act and Toxic Substances Control Act, among others. Conclusion

In this paper, we proposed a definition for “digital matching firms” as firms that use Internet and smartphone-enabled apps to match service providers with consumers, help ensure trust and quality assurance via peer-rating services, and rely on flexible service providers who, when necessary, use their own assets. Notwithstanding the challenges of defining and measuring digital matching firms in the context of the greater economy, their rapid growth suggests that these firms are providing a unique and valuable platform to connect consumers and service providers. We found that many digital matching firms have grown considerably during the past five years, and although reliable public data about the size and scope of the digital matching economy as a whole is scarce, there are a number of digital matching firms that are reportedly valued in the billions of dollars, with Uber and Airbnb leading the pack at $62.5 billion and more than $25 billion, respectively. However, these firms remain a small part of the greater economy.

31 According to the DC Taxicab Commission Disability Advisory Committee, the portion of its fleet dedicated to wheelchair accessible vehicles must be at least 6 percent by December 31, 2014; at least 12 percent by December 31, 2016; and at least 20 percent by December 31, 2018.

Office of the Chief Economist  Economics and Statistics Administration  Page 18

Digital Matching Firms: A New Definition in the “Sharing Economy” Space

We discussed a number of potential benefits and challenges introduced by digital matching firms, with benefits including potentially lower transaction costs for services, flexible employment opportunities for service providers, the leveraging of excess capacity, improved customer experience, and the potential for stimulating new consumption in the economy. However, the introduction of digital matching firms is not without potential downsides. These detriments include potential income instability for service providers, the need for service providers to take care of their own asset maintenance costs, responsibility for obtaining the assets (such as a car or room) that they use to provide services, fewer worker benefits, and access issues for individuals who don’t have a readily available Internet source and/or smartphone.

Finally, we discussed some of the challenges that are emerging with the growth of this particular innovative business model. Like ecommerce firms in the 1990s, digital matching firms are promoting debate about how to capture the benefits of technology driven change without abandoning important aspects of the current industrial organization, such as workers’ rights, consumer safety, equal access, environmental protection, and privacy. . References

AARP. “Life Reimagined Announces Collaboration with Uber to Offer New Income Opportunities to Members.” July 30, 2015. Retrieved from http://www.aarp.org/about-aarp/press-center/info-07- 2015/lifereimagined-uber.html

Alba, Davey. “Airbnb Confirms $1.5 Billion Funding Round, Now Valued at $25.5 Billion.” Wired. December 7. 2015. Retrieved from http://www.wired.com/2015/12/airbnb-confirms-1-5-billion-funding-round-now- valued-at-25-5-billion/

Barro, Josh. “New York City Taxi Medallion Prices Keep Falling, Now Down About 25 Percent.” The New York Times, Jan. 7, 2015. Retrieved from http://www.nytimes.com/2015/01/08/upshot/new-york-city-taxi-medallion-prices- keep-falling-now-down-about-25-percent.html?_r=3&abt=0002&abg=0

Byers, John W., Davide Proserpio, &Georgios Zervas. “The Rise of the Sharing Economy: Estimating the Impact of Airbnb on the Hotel Industry.” 2015. Retrieved from http://people.bu.edu/zg/publications/airbnb.pdf

Gandel, Stephen. “Uber just beat Facebook's $50 billion record.” Fortune, July 31, 2015. Retrieved from http://fortune.com/2015/07/31/uber-valuation-funding-round/

GAO, “Contingent Workforce: Size, Characteristics, Earnings, and Benefits.” April 20, 2015. Retrieved from http://www.gao.gov/assets/670/669766.pdf

Gonzalez, Juan. “NYC Council to propose tough penalties for landlords who use sites like Airbnb, in effort to keep affordable housing.” New York Daily News, June 10, 2015. Retrieved from http://www.nydailynews.com/new-york/steep-penalties-coming-nyc- landlords-airbnb-article-1.2252541

Office of the Chief Economist  Economics and Statistics Administration  Page 19

Digital Matching Firms: A New Definition in the “Sharing Economy” Space

Hall, Jonathan V, & Alan Krueger. “An Analysis of the Labor Market for Uber’s Driver-Partners in the United States.” Working Papers (Princeton University. Industrial Relations Section); 587. January 2015. Retrieved from https://s3.amazonaws.com/uber-static/comms/PDF/Uber_Driver- Partners_Hall_Kreuger_2015.pdf

Houseman, Susan. “Measuring Nonstandard Employment in the United States.” Paper for the WIEGO meeting on “Measuring Informal Employment in Developed Countries.” October, 2008. Retrieved from http://wiego.org/sites/wiego.org/files/publications/files/Houseman_Measure_nonstandard_empl_US.p df

Isaac, Mike, Leslie Picker. “Uber Valuation Put at $62.5 Billion After a New Investment Round.” The New York Times. December 3, 2015. Retrieved from http://www.nytimes.com/2015/12/04/business/dealbook/uber-nears-investment-at-a-62-5-billion- valuation.html

JPMorgan Chase & Co. Institute. “Paychecks, Paydays, and the Online Platform Economy: Big Data on Income Volatility.” February 2016. Retrieved from https://www.jpmorganchase.com/corporate/institute/report-paychecks-paydays-and-the-online- platform-economy.htm

Kilgannon, Corey. “In New Exam for Cabbies, Knowledge of Streets Takes a Back Seat.” The New York Times. March 8, 2015. Retrieved from http://www.nytimes.com/2015/03/09/nyregion/the-best-route-once-sacred-cabby- wisdom-takes-a-back-seat.html

Koetsier, John. “The Sharing Economy has Created 17 Billion-Dollar Companies (and 10 Unicorns)” Venture Beat. June 4, 2015. Retrieved from http://venturebeat.com/2015/06/04/the-sharing-economy- has-created-17-billion-dollar-companies-and-10-unicorns/

Mcbride, Sarah, Dan Levine. “In California, Uber driver is employee, not contractor: agency.” Reuters. Jun 18, 2015. Retrieved from http://www.reuters.com/article/2015/06/18/us-uber-california- idUSKBN0OX1TE20150618#TIolVSIwD9yhTf4L.97

MBO Partners. “MBO Partners Highlights Key Characteristics of Independent Workers in the On-Demand Economy.” April 21, 2015 Retrieved from https://www.mbopartners.com/press-releases/characteristics-of-workers-on- demand-economy

MBO Partners. “MBO Partners State of Independence in America 2015.” 2015. Retrieved from https://www.mbopartners.com/state-of-independence

Olson, Michael J., Samuel J. Kemp. “Sharing Economy: An In-Depth Look At Its Evolution and Trajectory Across Industries.” PiperJaffray Investment Research. March 2015. Retrieved from http://collaborativeeconomy.com/wp/wp- content/uploads/2015/04/Sharing-Economy-An-In-Depth-Look-At-Its-Evolution-and- Trajectory-Across-Industries-.pdf

Office of the Chief Economist  Economics and Statistics Administration  Page 20

Digital Matching Firms: A New Definition in the “Sharing Economy” Space

Pew Research Center, “U.S. Smartphone Use in 2015.” April 1, 2015. Retrieved from http://www.pewinternet.org/files/2015/03/PI_Smartphones_0401151.pdf

PriceWaterhouseCooper. “The Sharing Economy.” Consumer Intelligence Series. April 2015. Retrieved from http://www.pwc.com/us/en/industry/entertainment- media/publications/consumer-intelligence-series/assets/pwc-cis-sharing-economy.pdf

Samuelson, Rob. “Seattle taxi revenue dropping precipitously due to Uber and Lyft.” Seattle Sun Times. June 13, 2015. Retrieved from http://seattle.suntimes.com/sea- news/7/79/144490/taxi-revenue-dropping/

Sanders, Sam. “Santa Monica Cracks Down On Airbnb, Bans 'Vacation Rentals' Under A Month.” NPR. May 13, 2015. Retrieved from http://www.npr.org/sections/thetwo- way/2015/05/13/406587575/santa-monica-cracks-down-on-airbnb-bans-vacation-rentals-under-a- month

Silverstein, Sara. “These Animated Charts Tell You Everything About Uber Prices In 21 Cities.” Business Insider. October 16, 2014. Retrieved from http://www.businessinsider.com/uber-vs-taxi-pricing-by-city- 2014-10

Sundararajan, Arun. “Peer-to-Peer Businesses and the Sharing (Collaborative) Economy: Overview, Economic Effects and Regulatory Issues.” Written testimony for the hearing titled, The Power of Connection: Peer-to-Peer Businesses, held by the Committee on Small Business of the United States House of Representatives, January 15, 2015. Retrieved from http://smallbusiness.house.gov/uploadedfiles/1-15- 2014_revised_sundararajan_testimony.pdf

Tangel, Andrew. “Trading Taxis for Uber, Drivers Riding a Boom.” The Wall Street Journal. July 31, 2015. Retrieved from http://www.wsj.com/articles/trading-taxis-for- uber-drivers-riding-a-boom-1438389363?mod=e2fb

Worland, Justin. “Cab Drivers No Longer Required to Learn N.Y.C.’s Streets.” Time. March 9, 2015 Retrieved from http://time.com/3737193/nyc-taxi-geography/

United States Department of Justice. Information and Technical Assistance on the Americans with Disabilities Act. Retrieved from http://www.ada.gov/ada_title_III.htm

Wallace, Alice. “Amazon to collect Colorado sales tax on purchases starting Feb. 1.” The Denver Post. January 15, 2016. Retrieved from http://www.denverpost.com/business/ci_29390906/amazon-collect- colorado-sales-tax-purchases-starting-feb

Weil, David. Wage and Hour Division, U.S. Department of Labor. (July 15, 2015). The Application of the Fair Labor Standards Act’s “Suffer or Permit” Standard in the Identification of Employees Who Are Misclassified as Independent Contractors. Retrieved from http://www.dol.gov/whd/workers/Misclassification/AI-2015_1.pdf

Office of the Chief Economist  Economics and Statistics Administration  Page 21

Digital Matching Firms: A New Definition in the “Sharing Economy” Space

Zhang, Shu, & Gerry Shih. “Uber seen reaching $10.8 billion in bookings in 2015: fundraising presentation.” Reuters. August 21, 2015. Retrieved from http://www.reuters.com/article/us-uber-tech- fundraising-idUSKCN0QQ0G320150821#6VSBdjilflUp3Q20.97

Office of the Chief Economist  Economics and Statistics Administration  Page 22

Digital Matching Firms: A New Definition in the “Sharing Economy” Space

Appendix: Examples of Digital Matching Firms

This list was compiled for the purpose of testing whether it was possible to develop a firm-based definition that captures the uses of a “new technology,” where the new technology is the on-demand or digital matching platform business model. The majority of this research was done in the latter half of 2015 using the methods described below the table. An effort has been made to ensure the firms in the list are still in operation and continue to meet the definition. However, given the dynamism of entrepreneurially activity in this area, the list should not be viewed as authoritative or as a comprehensive list of companies using the digital matching platform model.

Category Company Name Company Website

Art Rental Art.sy artsy.net Art Rental TurningArt turningart.com Art Rental Artsicle artsicle.com Bike Sharing Spinlister, Inc. (formerly Liquid) https://www.spinlister.com/ Car Sharing Turo, Inc. http://www.Turo.com/ Car Sharing Getaround, Inc. https://www.getaround.com/ Car Sharing SnappCar http://www.snappcar.com/ Car Sharing BMW Car Sharing, LLC (DriveNow) https://us.drive-now.com/ Care Dog Vacay, Inc. http://dogvacay.com/ Care A Place for Rover, Inc. http://www.Rover.com/ Care UrbanSitter, Inc. https://www.urbansitter.com/ Care Care.com, Inc. https://www.care.com/ Care Swifto https://swifto.com/ Care The Good Bear, Inc. (Doggybnb) http://doggybnb.com/ Care Zingy http://www.zingypet.com/ Clothing Swaps Dig N'Swap http://www.dignswap.com/ Delivery Postmates, Inc. https://postmates.com/ Delivery Food Lovers United Co. https://www.fluc.com/ Delivery Square, Inc. (Caviar) https://www.trycaviar.com/ Delivery DoorDash, Inc. https://www.doordash.com/ Dining Feastly, Inc. https://eatfeastly.com/ Dining EatWith Media Ltd. http://www.eatwith.com/ Dining Greased Watermelon LLC (LeftoverSwap) http://leftoverswap.com/ Dining SpoonRocket, Inc. https://www.spoonrocket.com/ Dining Munchery, Inc. https://munchery.com/ Dining Sprig, Inc. http://sprig.com/ Dining Gobble Errands TaskRabbit, Inc. (formerly RunMyErrand, Inc.) https://www.taskrabbit.com/ Errands JobRunners, LLC https://www.job-runners.com/ Errands Zaarly, Inc. https://www.zaarly.com/ Errands Dolly, Inc. https://getdolly.com/ Errands RedBeacon http://www.redbeacon.com/ Errands Expert Bids https://www.expertbids.com/ Errands Fancy Hands, Inc. https://www.fancyhands.com/ Errands Gorilly, Inc. http://www.gorilly.com/ Errands Alfred Club, Inc. https://www.helloalfred.com/ Errands NeighborFavor, Inc. https://favordelivery.com/ Errands Campus Bellhops, LLC https://getbellhops.com/

Office of the Chief Economist  Economics and Statistics Administration  Page 23

Digital Matching Firms: A New Definition in the “Sharing Economy” Space

Errands Shyp, Inc. http://www.shyp.com/ Errands Crowdflower http://www.crowdflower.com/ Fashion Tradesy, Inc. https://www.tradesy.com/ Fashion Le Tote, Inc. https://www.letote.com Fashion RocksBox, Inc. https://www.rocksbox.com/ Fashion Rent the Runway, Inc. https://www.renttherunway.com/ Fashion Bag Borrow or Steal bagborroworsteal.com Fashion Shopittome shopittome.com Fashion The Outnet theoutnet.com Funding Kickstarter, Inc. http://www.kickstarter.com Funding RocketHub, Inc. http://www.rockethub.com/ Funding IndieGoGo Inc. https://www.indiegogo.com/ Funding Prosper prosper.com Funding LendingTree, LLC (formerly Tree.com, Inc.) https://www.lendingtree.com/ Funding LendingClub Corporation https://www.lendingclub.com/ Funding Enterprise Den enterpriseden.com Funding Startsomegood startsomegood.com Funding Pozible pozible.com Gardens Servicevines General Online Rental AnyHire anyhire.com Goods Sharing Sugar Packet, Inc. (doing business as NeighborGoods) http://neighborgoods.net/ Goods Sharing HeyNeighbor, LLC http://www.heyneighborapp.com/ Goods Sharing 1000 Tools, Inc. https://www.1000tools.com/ Goods Sharing Boatbound, Inc. https://boatbound.co/ Goods Sharing Streetbank streetbank.com Goods Sharing Zi Group SA (Zilok) http://us.zilok.com/ Goods Sharing Sparkplug Marketplace, Inc. http://www.sparkplug.it/ Goods Sharing Friends With Things friendswiththings Goods Sharing Toolzdo toolzdo.com Goods Sharing RentStuff rentstuff.com Homesharing Couchsurfing International, Inc. https://www.couchsurfing.com/ Homesharing Airbnb, Inc. https://www.airbnb.com/ Homesharing FlipKey, Inc. https://www.flipkey.com/ Homesharing HomeAway, Inc. (formerly CEH Holdings) http://www.homeaway.com/ Homesharing Roomorama https://www.roomorama.com/ Homesharing Lifealike Limited (doing business as onefinestay) http://www.onefinestay.com/ Media and Entertainment Fon Wireless, Ltd. https://corp.fon.com/ Media and Entertainment SoundCloud Ltd. http://soundcloud.com/ Media and Entertainment Earbits, Inc. http://www.earbits.com/ Misc Services Wello wello.com Misc Services Nanny in the Clouds nannyintheclouds.com Parking Parkcirca www.parkcirca.com Parking Parking Panda Corp. (Parking Panda) https://www.parkingpanda.com/ Personal Services Rent a Friend rentafriend.com Personal Services Hire a Boston Wingwoman hireawingwoman.com Professional and Freelance Amazon.com, Inc. (Amazon Mechanical Turk) https://www.mturk.com Professional and Freelance Upwork Global, Inc. (formerly Elance-oDesk, Inc.) https://www.upwork.com/ Professional and Freelance Fiverr International Ltd. https://www.fiverr.com/ Professional and Freelance Thumbtack, Inc. https://www.thumbtack.com/ Professional and Freelance SpareHire, Inc. https://www.sparehire.com/ Professional and Freelance Websoft, Inc. (doing business as Guru.com) http://www.guru.com/ Professional and Freelance Wonolo, Inc. http://wonolo.com/ Professional and Freelance Gig Bureau, LLC (doing business as GigSalad) https://www.gigsalad.com/ Professional and Freelance Peers Benefit Corporation http://www.peers.org/ Professional and Freelance Turuly, Inc. (doing business as BlogMutt) https://www.blogmutt.com/ Professional and Freelance Gigwalk, Inc. http://www.gigwalk.com/ Professional and Freelance Creative Circle, LLC https://www.creativecircle.com/ Ridesharing Uber Technologies, Inc. http://www.uber.com Ridesharing Lyft, Inc. https://www.lyft.com/

Office of the Chief Economist  Economics and Statistics Administration  Page 24

Digital Matching Firms: A New Definition in the “Sharing Economy” Space

Ridesharing Sidecar Technologies, Inc. https://www.side.cr/ Ridesharing Tripda, Inc. https://www.tripda.com/ Ridesharing GoCarShare http://gocarshare.com/ Ridesharing Shuddle, Inc. https://shuddle.us/ Ridesharing AtoB LLC (Coride) https://www.coride.com/ Ridesharing Zimride https://zimride.com/ Ridesharing carma https://carmacarpool.com/ Ridesharing wingz https://wingz.me/ Ridesharing Nuride http://www.nuride.com/ Ridesharing Jayride http://us.jayride.com/ Taxi Sharing Taxi2 Taxi Sharing Weeels http://www.bandwagon.io/about#main Teaching CoachUp, Inc. https://www.coachup.com/ Teaching Chegg, Inc. (Chegg Tutors) https://www.chegg.com/tutors/ Teaching Skillshare, Inc. http://www.skillshare.com/ Teaching Udemy, Inc. https://www.udemy.com/ Teaching Service Scout, Inc. (doing business as TakeLessons) https://takelessons.com/ Teaching Myngle http://www.myngle.com/ Teaching Glovico http://www.glovico.org/ Teaching RiffRaff Community, Inc. http://www.riffraff.me/ Teaching Livemocha Textbook Rental Chegg Textbook Rental CampusBookRentals Textbook Rental BookRenter Toy Rental BabyPlays babyplays.com Unique Experiences Vayable, Inc. http://www.vayable.com/ As digital matching firms are a relatively new phenomenon and the companies have been commonly referred to as part of the “sharing” or “collaborative” economy, research into existing digital matching firms was conducted via basic online searches. The majority of these firms were found using Google search of the words “sharing economy,” or “collaborative economy,” which directed us to several news and journal articles written on the topic. We then researched the companies and compiled a list of those that have the characteristics of digital matching firms.

The articles used to identify these firms are cited below the table along with the names of companies mentioned..

1. Bloomberg Brief. “The Sharing Economy.” (2015). Available at: http://newsletters.briefs.bloomberg.com/document/4vz1acbgfrxz8uwan9/front

2. Botsman, Rachel. “The Sharing Economy Lacks a Shared Definition”. Fast Company and Inc. (2013). Available at: http://www.fastcoexist.com/3022028/the-sharing-economy-lacks-a- shared-definition

3. Fast Company. Available at: http://www.fastcompany.com/3042248/the-gig-economy-wont- last-because-its-being-sued-to-death

4. Federal Reserve Bank of Richmond. Available at: https://www.richmondfed.org/publications/research/econ_focus/2014/q4/cover_story

5. “Find Work.” Peers.org. (2015). Available at: http://www.peers.org/find-work/

Office of the Chief Economist  Economics and Statistics Administration  Page 25

Digital Matching Firms: A New Definition in the “Sharing Economy” Space

6. Forbes. Available at: http://www.forbes.com/sites/groupthink/2014/07/07/how-the-hotel- industry-got-blindsided-and-why-yours-could-be-next/

7. Geron, Tomio. “Airbnb and the Unstoppable Rise of the Share Economy”. Forbes. (2015). Available at: http://www.forbes.com/sites/tomiogeron/2013/01/23/airbnb-and-the- unstoppable-rise-of-the-share-economy/

8. Jolly, Jennifer. “Dog Needs a Walk? There’s an App for that”. The New York Times. (2015). Available at: http://well.blogs.nytimes.com/2015/07/07/dog-needs-a-walk-theres-an-app-for- that/?action=click&contentCollection=Your%20Money&module=MostEmailed&version=Full&re gion=Marginalia&src=me&pgtype=article&_r=0

9. LaBrecque, Sarah. “Eight of the Best Sharing Economy Companies”. The Guardian. (2014). Available at: http://www.theguardian.com/sustainable-business/eight-best-sharing-economy- companies

10. Scholz, Trebor. “Platform Cooperativism vs. the Sharing Economy”. Public Seminar. (2015). Available at: http://www.publicseminar.org/2015/04/platform-cooperativism-vs-the-sharing- economy/#.VaUvKvlVhBc

11. Sundararajan, Arun. “Peer-to-Peer Businesses and the Sharing (Collaborative Economy): Overview, Economic Effects, and Regulatory Issues.” (2014). Available at: http://smallbusiness.house.gov/uploadedfiles/1-15-2014_revised_sundararajan_testimony.pdf

12. Tanz, Jason. “How Airbnb and Lyft Finally Got Americans to Trust Each Other”. Wired: Business. (2014). Available at: http://www.wired.com/2014/04/trust-in-the-share-economy/

13. The Economist. “All Eyes on the Sharing Economy”. The Economist: Technology. (2013). Available at: http://www.economist.com/news/technology-quarterly/21572914-collaborative- consumption-technology-makes-it-easier-people-rent-items Acknowledgments

In addition to those whose work is cited, the author would like to thank the following persons who provided comments, suggestions, and other contributions to this report:

Department of Commerce Economics and Statistics Administration: Ellen Hughes- Cromwick, Chief Economist; Rob Rubinovitz, Deputy Chief Economist; Sabrina Montes, Team Lead and Economist; David Langdon, Team Lead and Economist; David Beede, Economist; Regina Powers, Economist; Sue Helper, former Chief Economist;

White House Council of Economic Advisers: Nirupama Rao; Robert Seamans; Martha Gimbel; Harris Eppsteiner; Sam Himel

Office of the Chief Economist  Economics and Statistics Administration  Page 26

Digital Matching Firms: A New Definition in the “Sharing Economy” Space

Department of Labor: Heidi Shierholz; Tanya Goldman; Jeff Vockrodt

Department of Treasury: Karen Dynan; Tara Watson; Ryan Nunn

White House National Economic Council: JJ Raynor

Any errors in the report are solely the authors’ responsibility. The author also wishes to express that the inclusion and/or discussion of any company is not to be characterized as an endorsement of neither the firm itself nor the services it provides.

Office of the Chief Economist  Economics and Statistics Administration  Page 27

SUNDARARAJAN_COHEN_FINAL (LMW).DOCX (DO NOT DELETE) 4/2/15 3:02 PM

Self-Regulation and Innovation in the Peer-to-Peer Sharing Economy Molly Cohen† & Arun Sundararajan††

INTRODUCTION A growing fraction of the world’s economy involves digitally enabled peer-to-peer exchange. This form of exchange has ex- panded dramatically in recent years, moving beyond simple re- tailing and free file exchange to personal, human-intensive serv- ices such as hosted accommodation, urban and city-to-city transportation, neighborhood logistics, peer-to-peer lending, eq- uity crowdfunding, and decentralized manufacturing. Such ex- change, while mediated by digital platforms like Uber, Airbnb, Lyft, Etsy, Funding Circle, and AngelList, frequently involves the peer-to-peer provision of familiar real-world services that are the traditional subjects of regulation. It blurs the line be- tween personal and professional in the provision of commercial services. Further, it often involves semianonymous transactions. Each of these factors creates a variety of regulatory challenges that could impede innovation, especially the grassroots innova- tion made possible by new opportunities for peer-to-peer ex- change. Additionally, these regulatory barriers may slow the growth of employment that involves individuals providing goods, services, labor, and capital through peer-to-peer platforms—a form of work that will in future years constitute a larger fraction of the economy than it does today. We argue that the resolution of these challenges must in- clude self-regulatory approaches. Self-regulation is not the same as deregulation or no regulation. Rather, it is the reallocation of regulatory responsibility to parties other than the government. We explain why platforms should not be viewed as entities to be regulated but rather as actors that are a key part of the regulatory

† National Center on Philanthropy and the Law Fellow, Office of the General Counsel, New York University. †† Professor and NEC Faculty Fellow, Leonard N. Stern School of Business, New York University. Jonah Blumstein, Valeriya Greene, and Eric Jacobson provided excellent research assistance. 116 SUNDARARAJAN_COHEN_FINAL (LMW).DOCX (DO NOT DELETE) 4/2/15 3:02 PM

2015] Self-Regulation and Innovation 117 framework in this arena. For nonintermediated peer-to-peer ex- change in the past, the primary solution to market failure was intervention by a government agency. But today, the existence of third-party platforms that mediate exchange fundamentally alters what the market is capable of providing on its own, and it creates a new institution capable of affecting what Michel Fou- cault referred to as the “conduct of conduct.”1 Nevertheless, be- cause the interests of digital, third-party platforms are not al- ways perfectly aligned with the broader interests of society, some governmental involvement or oversight is likely to remain useful. In this Essay, we describe different factors that may induce market failure and highlight the extent to which these are miti- gated by the existence of new digital platforms. We outline how self-regulation can form part of a broader innovation-enhancing solution, providing guidelines for sharing-economy regulation that draw from self-regulatory experiences in industries ranging from nuclear power and financial intermediation to chemical production and cotton supply. Demonstrated enforcement capa- bilities along with a clear perception of independence and le- gitimacy are essential. Leveraging reputational concerns can complement traditional regulation. We distinguish between those entities that, in a self-regulatory solution, are well suited to correct information asymmetries, and those that are well suited to address market failure from externalities. We conclude by noting some of the opportunities presented by delegated regu- lation through data—a form of regulation that has the potential to be superior to mandated marketplace transparency. We fre- quently use the peer-to-peer marketplace Airbnb and the context of short-term, hosted accommodation to illustrate our arguments.

I. BACKGROUND At the end of 2014, Airbnb’s CEO, Brian Chesky, announced that the company’s peer-to-peer marketplace listed over one mil- lion homes and was adding three thousand new listings per

1 Colin Gordon, Governmental Rationality: An Introduction, in Graham Burchell, Colin Gordon, and Peter Miller, eds, The Foucault Effect: Studies in Governmentality 1, 2–3 (Chicago 1991) (describing Foucault’s term “conduct of conduct” as a perspective of the self toward the structures of government, specifically vis-à-vis the conformance of personal conduct to broader social institutions).

SUNDARARAJAN_COHEN_FINAL (LMW).DOCX (DO NOT DELETE) 4/2/15 3:02 PM

118 The University of Chicago Law Review Dialogue [82:116 day.2 Airbnb is one of a number of new digital platforms that fa- cilitates widespread, peer-to-peer commerce. Uber, Lyft, and Sidecar allow drivers in hundreds of cities to offer taxicab-like point-to-point urban transportation that is requested through a mobile application. Paris-based BlaBlaCar and Munich-based Carpooling.com have each built city-to-city transportation net- works using online peer-to-peer marketplaces that connect indi- vidual drivers with potential passengers. Etsy, an online mar- ketplace centered on handmade and vintage items, has over one million sellers that power a decentralized, open-industrial- production-and-distribution system. Peer-to-peer lending plat- forms Lending Club and Funding Circle have mediated billions of dollars in loans among their members. These peer-to-peer, col- laborative, sharing-economy platforms3 are new market-firm hybrids that centralize certain activities associated with the provision of commercial services (like branding, trust, and pay- ments) while decentralizing others (like pricing, supply infra- structure, and service provision). Early forms of peer-to-peer accommodation, transportation, venture financing, lending, and labor provision may have emerged as a grassroots response to deficiencies in existing cor- porate alternatives.4 Today, however, platforms owned by shareholder corporations and funded by significant venture capi- tal drive the growth of the sharing economy.5 These platforms are

2 See Brian Chesky, Tweet by @bchesky (Twitter Dec 7, 2014, 12:09 PM), online at http://twitter.com/bchesky/status/541655860271783937 (visited Feb 22, 2015). 3 There is considerable debate in the popular press about what constitutes the sharing economy. For an overview of different kinds of peer-to-peer businesses, the tech- nological and economic drivers of their emergence, and their possible economic impacts, see Arun Sundararajan, Peer-to-Peer Businesses and the Sharing (Collaborative) Econ- omy: Overview, Economic Effects and Regulatory Issues, in The Power of Connection: Peer-to-Peer Businesses, Hearing before the House Committee on Small Business, 213th Cong, 2d Sess 20–28 (2014). See also Christopher Koopman, Matthew Mitchell, and Adam Thierer, The Sharing Economy and Consumer Protection Regulation: The Case for Policy Change *3–5 (Mercatus Research Dec 2014), online at http://mercatus.org/sites/ default/files/Koopman-Sharing-Economy.pdf (visited Feb 22, 2014). 4 For example, loans from relatives or friends are common alternatives to bank loans. Lending Club launched initially as a Facebook application aimed at using the trust created by social networks to attract young borrowers with credit histories that re- stricted these borrowers’ traditional financing alternatives. See Ming Kwan and Deepak Ramachandran, Trust and Online Reputation Systems, in Jennifer Golbeck, ed, Comput- ing with Social Trust 287, 301 (Springer 2009); Victoria Barret, Bank of Americans, Forbes 44, 44 (Dec 20, 2010). 5 As of January 31, 2014, Uber had received close to $5 billion in pre-IPO funding, Airbnb had received over $790 million, and Lyft had received over $330 million. Mike Isaac and Michael J. De La Merced, Ride-Hailing Service Lyft Is Said to Be in Talks to

SUNDARARAJAN_COHEN_FINAL (LMW).DOCX (DO NOT DELETE) 4/2/15 3:02 PM

2015] Self-Regulation and Innovation 119 frequently cast as the protagonists in conflicts between existing regulations and the commercial activity that the platforms en- able.6 Since the eventual commercial success of a platform de- pends on its participants’ ability to engage in exchange, the plat- forms have invested significant resources in addressing the misfit between existing regulation and the exchange that they facilitate.7 This leads to confusion between regulating the plat- forms themselves and regulating the exchange that their mar- ketplaces facilitate. We argue in this Essay that platforms should be viewed as part of the solution, rather than part of the problem, and they should be included as key actors in a self-regulatory regime. In order to guide the discovery of a new division of regulatory re- sponsibility among the government, these new market owners, and other societal stakeholders, we outline some sources of mar- ket failure commonly associated with the activities now con- ducted as part of the sharing economy, examine what self- regulatory organizations are and what factors make them effec- tive, and apply these insights to envision the shape that success- ful self-regulation in the sharing economy might take.

II. THE SHARING ECONOMY: REGULATORY ISSUES Generally, “regulation” refers to the use of legal instruments to implement social and economic policy objectives. Judge Rich- ard Posner defines economic regulation as follows: “Properly de- fined, the term refers to taxes and subsidies of all sorts as well as to explicit legislative and administrative controls over rates,

Raise $250 Million, Dealbook (NY Times Feb 12, 2015), online at http://dealbook.nytimes .com//2015/02/12/ride-hailing-service-lyft-is-said-to-be-in-talks-to-raise-250-million (visited Feb 22, 2015) (noting pre-IPO funding amounts for Uber and Lyft); Scott Austin, Airbnb’s Official Funding Haul: $475 Million, Digits Blog (Wall St J Aug 1, 2014), online at http://blogs.wsj.com/digits/2014/08/01/airbnbs-official-funding-haul-475-million (visited Feb 22, 2015) (discussing Airbnb’s funding totals). 6 For example, a summary of Airbnb’s operation in New York City provided by the New York State Office of the Attorney General (“NYAG”) states that, “[i]n late 2013, the [NYAG] launched an investigation of users of web platforms like Airbnb who run large- scale enterprises in violation of fire safety, zoning, tax, and other applicable laws.” New York State Office of the Attorney General, Airbnb in the City *4 (Oct 2014), online at http://www.ag.ny.gov/pdfs/Airbnb%20report.pdf (visited Feb 22, 2015). 7 See, for example, Rosalind S. Helderman, Uber Pressures Regulators by Mobiliz- ing Riders and Hiring Vast Lobbying Network (Wash Post Dec 13, 2014), online at http://www.washingtonpost.com/politics/uber-pressures-regulators-by-mobilizing-riders- and-hiring-vast-lobbying-network/2014/12/13/3f4395c6-7f2a-11e4-9f38-95a187e4c1f7 _story.html (visited Feb 22, 2015) (noting that as of December 2014, Uber had hired more than 161 lobbyists in at least 50 US cities and states to lobby on its behalf).

SUNDARARAJAN_COHEN_FINAL (LMW).DOCX (DO NOT DELETE) 4/2/15 3:02 PM

120 The University of Chicago Law Review Dialogue [82:116 entry, and other facets of economic activity.”8 When market practices lead to inefficient or inequitable outcomes (a situation often referred to as a “market failure”)—for example, due to asymmetric information, the problem of public goods, the threat of monopoly, or the existence of externalities that are not natu- rally internalized by market participants—regulation may be supplied as a corrective measure.9 In the absence of technologi- cal, self-regulatory, or governmental intervention, peer-to-peer exchange is susceptible to a variety of forms of market failure. We discuss three such species of market failure here.10

A. Information Asymmetry Most forms of peer-to-peer exchange are characterized by asymmetric information. For example, a passenger who enters a taxicab may not know the qualifications (or intentions) of its driver. A host knows more about the quality of her short-term accommodation than a potential guest does, and in turn, a guest knows more about her own reliability and level of cleanliness. A borrower knows more about her creditworthiness than a lender does. These and other forms of information asymmetry can lead to fewer transactions than are socially optimal—due to uncer- tainty about quality—in addition to a situation termed “adverse selection,” which occurs when the information asymmetry makes higher-quality traders less likely to participate.11 This asymme- try can also lead to moral hazard: because parties’ imperfect in- formation limits their ability to contract, one trading partner might display behavior that is less careful (for example, reckless

8 Richard A. Posner, Theories of Economic Regulation, 5 Bell J Econ 335, 335 (1974). 9 The “public interest theory” of regulation posits that regulation emerges in situa- tions of market failure as a response to public demand. The “capture theory” of regula- tion posits that, over time, governmental regulation is supplied to serve the interests of the regulated industries. See id at 335–36. See also George J. Stigler, The Theory of Eco- nomic Regulation, 2 Bell J Econ 3, 3 (1971) (outlining theories of regulation and arguing that regulation functions primarily for the benefit of the regulated industry). 10 Aside from these three types of market failure, there are a variety of other im- portant industry-specific and geography-specific regulatory issues. Such issues include commercial taxation, hospitality taxes, equal access, disaster preparedness and fire safety, affordable housing and rent control, and policies to combat discrimination. Ana- lyzing the economic, legal, and social aspects of each of these regulatory issues is beyond the scope of this Essay. 11 See George A. Akerlof, The Market for “Lemons”: Quality Uncertainty and the Market Mechanism, 84 Q J Econ 488, 493–94 (1970).

SUNDARARAJAN_COHEN_FINAL (LMW).DOCX (DO NOT DELETE) 4/2/15 3:02 PM

2015] Self-Regulation and Innovation 121 driving), of lower effort (for example, lower levels of cleanliness), or somehow riskier than otherwise would have been chosen. 12 Prior to the emergence of sharing-economy platforms, gov- ernmental intervention was a natural solution to market failure in traditional peer-to-peer businesses. For example, safety con- cerns about drivers and information asymmetries about the dis- tance or cost of a ride were alleviated in part through driver screening and metered fares by taxicab regulatory agencies.13 Safety and quality concerns could have led to inefficiently low levels of peer-to-peer exchange in other markets—for example, in the market for very short-term rentals of residential housing units. Third-party intermediaries like Airbnb, Uber, and Lyft now offer alternative solutions to this form of market failure, includ- ing the use of digital technologies that reduce information asymmetries.14 For example, Airbnb offers an online feedback system that allows guests to learn about the quality of hosts from prior guests and provides evidence of “social capital”15 via links to user profiles on platforms like Facebook and LinkedIn. It also uses technology to digitally verify the government IDs of its providers.16 But more important than the details of a specific implementation is that the mere existence of large, third-party intermediaries expands the set of alternatives available to miti- gate market failure. As noted before,17 the eventual commercial success of platforms is affected by the ability of their partici- pants to engage in exchange. Thus, platforms have a natural in- centive to alleviate exchange-deterring forms of information failure using all manner of resources that go well beyond provid- ing digital solutions. For example, Lyft, independent of any

12 See Bengt Holmström, Moral Hazard and Observability, 10 Bell J Econ 74, 74 (1979). 13 For a discussion of the emergence of average-price fares and the subsequent need for entry restrictions using a medallion system, see generally Edward C. Gallick and David E. Sisk, A Reconsideration of Taxi Regulation, 3 J L, Econ & Org 117 (1987). 14 For a broad discussion of the impact of Internet technologies on information asymmetry, see generally John C. Moorhouse, Consumer Protection Regulation and In- formation on the Internet, in Fred E. Foldvary and Daniel B Klein, eds, The Half-Life of Policy Rationales: How New Technology Affects Old Policy Issues 125 (NYU 2003). 15 A sociological term of art, “social capital” refers to “social connections and the attendant norms and trust.” Robert D. Putnam, Tuning In, Tuning Out: The Strange Disappearance of Social Capital in America, 28 PS: Polit Sci & Polit 664, 665 (1995). 16 See Airbnb, Trust at Airbnb, online at http://www.airbnb.com/trust (visited Feb 22, 2015). 17 See text accompanying note 7.

SUNDARARAJAN_COHEN_FINAL (LMW).DOCX (DO NOT DELETE) 4/2/15 3:02 PM

122 The University of Chicago Law Review Dialogue [82:116 regulatory requirements, conducts in-person driver screenings that also include criminal background checks and an assessment of driving history.18 Similarly, as of July 2013, Airbnb employed three hundred people in its customer-service unit, fifty of whom were dedicated to promoting trust and safety.19

B. Externalities The choices of a buyer or provider in a peer-to-peer transac- tion may impose costs on (or result in benefits to) others, and these externalities often may not be naturally taken into ac- count (or internalized) when trading peers make choices. Some- times these externalities are negative. For example, an addi- tional taxicab on the road creates congestion and lengthens travel times for other drivers. A noisy Airbnb guest in an apartment building might impose costs on the other residents with his or her disturbing behavior. In other cases, externalities might be positive. For example, an increase in out-of-town visi- tors to a neighborhood induced by a high concentration of Airbnb hosts could benefit local restaurants. An increase in tourism caused by greater affordability and range of short-term accom- modation could benefit a variety of stakeholders in the hospital- ity and travel industries. Negative externalities lead to an oversupply of certain serv- ices; failure to internalize positive externalities could, correspond- ingly, lead to inefficiently low levels of market exchange. Some form of third-party regulatory intervention seems necessary in these situations, a point that we return to later in this Essay. 20

C. Blurring of Boundaries between the Personal and the Professional A distinguishing feature of commercial activity in the shar- ing economy is the way in which the provision of services often blurs lines between the personal and the professional. For ex- ample, most Airbnb hosts are not professional hoteliers.21 A large fraction of Lyft and Uber drivers are active on the platform

18 See Lyft, Safety, online at http://www.lyft.com/safety (visited Feb 22, 2015). 19 See Liz Gannes, After Home-Trashing Incident, Airbnb Builds an In-House En- forcer Team (All Things D July 16, 2013), online at http://allthingsd.com/20130716/after- home-trashing-incident-airbnb-builds-an-in-house-enforcer-team/ (visited Feb 22, 2015). 20 See Part IV. 21 In 2014, the NYAG classified a mere 6 percent of more than 25,000 Airbnb hosts as- sessed in New York as “commercial users.” NYAG, Airbnb in the City at *10 (cited in note 6).

SUNDARARAJAN_COHEN_FINAL (LMW).DOCX (DO NOT DELETE) 4/2/15 3:02 PM

2015] Self-Regulation and Innovation 123 fewer than fifteen hours per week.22 Only one in five sellers on Etsy considers their Etsy business a full-time job.23 Applying a regulatory regime developed for full-time or large-scale profes- sional providers to smaller, semiprofessional providers could create barriers to entry, stifling peer-to-peer exchange as well as the grassroots innovation that the sharing economy facilitates.24 Absent appropriate safeguards, there may also be new forms of market failure that occur as a consequence of the nonprofes- sional nature of supply and induce the emergence of new associa- tions or guilds that are part of a self-regulatory solution.

III. MAKING SELF-REGULATORY SOLUTIONS WORK We argue that the regulatory issues discussed in Part II may be best addressed by a self-regulatory approach that ac- tively involves sharing-economy platforms as well as existing nongovernmental stakeholders in both design and enforcement.

A. What Are SROs and Why Do They Emerge? Given its history and diversity, it is unsurprising that self- regulation (and self-regulatory organizations, or “SROs”) defies simple definition or categorization. Self-regulatory systems vary widely based on their levels of voluntariness, accountability, enforcement, and governmental intervention. For instance, Professor Julia Black distinguishes among four types of self- regulation.25 “Voluntary self-regulation” involves no direct gov- ernmental involvement or mandates.26 “Coerced self-regulation”

22 See Dan Levine and Sarah McBride, Uber, Lyft Face Crucial Courtroom Test over Driver Benefits (Reuters Jan 28, 2015), online at http://www.reuters.com/article/ 2015/01/28/us-uber-lyft-workers-idUSKBN0L11BN20150128 (visited Feb 22, 2015). 23 See Etsy, Redefining Entrepreneurship: Etsy Sellers’ Economic Impact *8 (Nov 7, 2013), online at http://extfiles.etsy.com/Press/reports/Etsy_RedefiningEntrepreneurship Report_2013.pdf (visited Feb 22, 2015). 24 For entities that facilitate grassroots innovation—often by nonprofessionals—the alternative paths to innovation are challenged by the fact that their approaches differ from traditional science, technology, and innovation. Thus, integration with mainstream institutions is a critical challenge. For further discussion of the interaction between grassroots innovators and established scientific and technological institutions, see generally Mariano Fressoli, et al, When Grassroots Innovation Movements Encounter Mainstream In- stitutions: Implications for Models of Inclusive Innovation, 4 Innov & Dev 277 (2014). 25 See Julia Black, Decentering Regulation: Understanding the Role of Regulation and Self-Regulation in a ‘Post-Regulatory’ World, 54 Current L Probs 103, 118 (2001). See also Saule T. Omarova, Wall Street as Community of Fate: Toward Financial Indus- try Self-Regulation, 159 U Pa L Rev 411, 424 (2011) (characterizing Black’s taxonomy as the framework “generally” adopted in the literature). 26 Black, 52 Current L Probs at 118 (cited in note 25).

SUNDARARAJAN_COHEN_FINAL (LMW).DOCX (DO NOT DELETE) 4/2/15 3:02 PM

124 The University of Chicago Law Review Dialogue [82:116 occurs when an industry formulates and imposes rules due to the threat of governmental regulation.27 Under a “sanctioned self-regulation” regime, the industry formulates rules subject to governmental approval.28 And, lastly, “mandated self-regulation” occurs when the government requires the industry to establish a self-regulatory framework.29 The overarching category of self- regulation may also include those situations in which the gov- ernment delegates to a third party the implementation of preex- isting federal law or regulation.30 Furthermore, self-regulatory regimes vary both by their methods and rigorousness of ac- countability and by their levels of formality.31 On the one hand, “partial self-regulation refers to situations in which private par- ties are responsible only for rulemaking, while enforcement is the domain of either public bodies or the market.32 On the other hand, “full self-regulation” occurs when the industry engages in both rulemaking and enforcement.33 Generally, SROs are privately run, typically with limited governmental involvement. Unlike trade organizations, which promote the well-being of an industry, SROs are meant to police an industry by formulating regimes of collective rulemaking in which entities come together to develop, monitor, and, at times, enforce standards to govern the behavior of members.34 Self-regulation is a naturally occurring phenomenon that has emerged repeatedly throughout the history of economic ac- tivity. For example, American Indian groups and the Maasai in Kenya both self-regulated access to and use of limited communal

27 Id. 28 Id. 29 Id. 30 See Douglas C. Michael, Federal Agency Use of Audited Self-Regulation as a Regulatory Technique, 47 Admin L Rev 171, 175–76 (1995). 31 Outside accountability may come from stakeholders, including consumer or community representatives, in “stakeholder self-regulation.” Black, 52 Current L Probs at 119 (cited in note 25) (quotation marks omitted). It may also come from third parties, such as NGOs, in “verified-self-regulation.” Id (quotation marks omitted). Or it could come as a close variant, such as “accredited self-regulation,” in which compliance is audited by a nongovernmental body like a standards council. Id (quotation marks omit- ted). See also Ian Bartle and Peter Vass, Self-Regulation within the Regulatory State: Towards a New Regulatory Paradigm?, 85 Pub Admin 885, 898–901 (2007) (discussing the intersection of self-regulation and accountability for the regulatory state). 32 Omarova, 159 U Pa L Rev at 424 (cited in note 25); Abraham L. Newman and David Bach, Self-Regulatory Trajectories in the Shadow of Public Power: Resolving Digi- tal Dilemmas in Europe and the U.S., 17 Governance 387, 390 (2004). 33 Omarova, 159 U Pa L Rev at 424 (cited in note 25); Newman and Bach, 17 Gov- ernance at 390 (cited in note 32). 34 See Omarova, 159 U Pa L Rev at 421 (cited in note 25).

SUNDARARAJAN_COHEN_FINAL (LMW).DOCX (DO NOT DELETE) 4/2/15 3:02 PM

2015] Self-Regulation and Innovation 125 goods like fish stock, prey, and grazing lands.35 Villages in Europe self-regulated access to and use of the commons, publish- ing communal farming-and-grazing regulations as early as the thirteenth century.36 Agreements among neighbors and groups of cultivators were formalized via bylaws that were quite com- prehensive in governing agricultural practices.37 More-formal merchant and craft guilds—early SROs—emerged in medieval times, imposing strict rules about their members’ wages, tools, technology, quality, and prices.38 These early examples of com- munal self-regulation of public goods are encouraging because they suggest that self-regulation is a natural part of economic development, and they challenge Professor Garrett Hardin’s thesis about the inevitable collapse of the commons.39 However, medieval occupational guilds may have also suffered from many of the issues of capture that Professor George Stigler argues are largely endemic across various forms of regulation40—for in- stance, the guilds may have enriched their members at society’s expense.41

B. Lessons from Contemporary SROs SROs continue to be widely prevalent in the modern world. They exist in many industries in which peer-to-peer transactions are common relative to firm-to-consumer transactions—for ex- ample, in medicine (the American Medical Association), real es- tate (the National Association of Realtors), and law (bar associa- tions). Often, modern SROs have significant enforcement and compliance capabilities, perhaps even quasi-judicial authority due to their ability to audit and penalize. We use examples of

35 See Donald Leal, Community-Run Fisheries: Avoiding the ‘Tragedy of the Com- mons’ (Property and Environment Research Center Sept 1996), online at http://perc.org/ articles/community-run-fisheries-0#sthash.64mLFFBf.dpuf (visited Feb 22, 2015); Rob- ert A. Blewett, Property Rights as a Cause of the Tragedy of the Commons: Institutional Change and the Pastoral Maasai of Kenya, 21 Eastern Econ J 477, 479 (1995). 36 See Susan Jane Buck Cox, No Tragedy on the Commons, 7 Envir Ethics 49, 55 (1985). 37 See id at 56. 38 See Alfred Kieser, Organizational, Institutional, and Societal Evolution: Medie- val Craft Guilds and the Genesis of Formal Organizations, 34 Admin Sci Q 540, 553 (1989). 39 See Cox, 7 Envir Ethics at 49 (cited in note 36). 40 See Stigler, 2 Bell J Econ at 3 (cited in note 9). 41 See Sheilagh Ogilvie, The Economics of Guilds, 28 J Econ Persp 169, 170 (2014) (arguing that the behavior of guilds can best be understood as being aimed at securing rents for guild members and legal privileges for the guild).

SUNDARARAJAN_COHEN_FINAL (LMW).DOCX (DO NOT DELETE) 4/2/15 3:02 PM

126 The University of Chicago Law Review Dialogue [82:116 contemporary SROs to highlight three factors that might lead to successful self-regulatory approaches for the sharing economy: credible enforcement mechanisms, a perception of legitimacy, and an emphasis on reputation.

1. Credible enforcement mechanisms. A primary factor that leads to a successful self-regulatory apparatus is the ability of an SRO to enforce its rules and regu- lations. Consider the example of the Institute of Nuclear Power Operations (INPO), an SRO that was established in 1979 follow- ing the nuclear accident at Three Mile Island,42 which has since successfully regulated nuclear-power operations in the United States.43 Its breadth of responsibility includes creating safety- risk-management standards, monitoring compliance, conducting routine evaluations of individual plants, investigating accidents, and providing technical assistance.44 The INPO did not gain instant respect; rather, it earned its reputation for fairness and gained credibility within the indus- try over time. In the 1980s, INPO inspectors found severe defi- ciencies at a Philadelphia nuclear plant. The INPO and the Nu- clear Regulatory Commission (NRC) worked closely together to rectify the flaws, and the ensuing regulatory criticism that they provided was so harsh that several top executives at the plant lost their jobs.45 This event established the INPO’s mettle and became a symbol of its power.46 The case of the INPO contrasts with the relative failure of Responsible Care, a set of self-regulatory principles promulgated by the International Council of Chemistry Associations, a chemical-industry SRO, in the late 1980s in the wake of the Bhopal Chemical Disaster.47 In the United States, the American Chemistry Council, a chemical-industry trade group, requires

42 See Institute of Nuclear Power Operations, About Us, online at http://www.inpo.info/AboutUs.htm (visited Feb 22, 2015). 43 The fact that nuclear power is subject to self-regulation should give pause to any critic who worries that public-safety stakes in the sharing economy are too high for a self-regulatory solution. 44 See INPO, About Us (cited in note 42); Omarova, 159 U Pa L Rev at 447–48 (cited in note 25). 45 See Joseph V. Rees, Hostages of Each Other: The Transformation of Nuclear Safety since Three Mile Island 110–18 (Chicago 1994). 46 See id. 47 See Andrew A. King and Michael J. Lennox, Industry Self-Regulation without Sanctions: The Chemical Industry’s Responsible Care Program, 43 Acad Mgmt J 698, 699 (2000).

SUNDARARAJAN_COHEN_FINAL (LMW).DOCX (DO NOT DELETE) 4/2/15 3:02 PM

2015] Self-Regulation and Innovation 127 that all its members adopt Responsible Care, but critics suggest that, since membership in the trade group is voluntary, this threat has little bite.48 An empirical study evaluating the effec- tiveness of this self-regulatory regime found no evidence that membership in Responsible Care influenced the rate of envi- ronmental improvement among members.49 In fact, members improved more slowly than nonmembers.50 The authors of the study concluded that this chemical-industry SRO has failed due to its inability to impose explicit sanctions.51

2. Perception of legitimacy. In addition to leading to effective policing of its members, the INPO’s enforcement mechanisms and early successes were valuable because they legitimized the SRO both inside and out- side the nuclear-power industry. In contrast, the Finance Indus- try Regulatory Authority (FINRA), a member-owned organiza- tion that regulates the activities of securities firms that transact with the public,52 suffers from a legitimacy problem. As a conse- quence, despite over eighty years of SRO oversight that have yielded tremendous financial innovation in the United States, the financial-services industry is rarely cited as a model for self- regulatory success. Part of FINRA’s difficulties may lie in its limited enforce- ment capabilities.53 However, the more serious challenge is that it is simply not perceived as being independent and sufficiently serious in its oversight of its members.54 As Professor Lawrence Lessig argues in regard to Congress in his recent book, the mere

48 See, for example, Aseem Prakash, Responsible Care: An Assessment, 39 Bus & Society 183, 197 (2000). 49 See King and Lennox, 43 Acad Mgmt J at 709 (cited in note 47). 50 See id. 51 See id at 713–14. 52 FINRA has oversight of trading in equities, corporate bonds, securities futures, and options, and it governs more than 4,000 securities firms with over 600,000 brokers. It is funded primarily through fees from member firms. See FINRA, FINRA: Our Story, online at http://www.finra.org/AboutFINRA (visited Feb 22, 2015). 53 For example, in the summer of 2014, Kara Stein, an SEC commissioner, ex- plained in a speech that FINRA’s enforcement actions were “too often financially insig- nificant for the wrongdoers.” Jean Eaglesham, Finra Weighs Tougher Stance (Wall St J June 19, 2014), online at http://www.wsj.com/articles/wall-street-watchdog-finra-under- pressure-to-toughen-sanctions-1403219509 (visited Feb 22, 2015). 54 See Jonathan Macey and Caroline Novogrod, Enforcing Self-Regulatory Organi- zation’s Penalties and the Nature of Self-Regulation, 40 Hofstra L Rev 963, 983 (2012).

SUNDARARAJAN_COHEN_FINAL (LMW).DOCX (DO NOT DELETE) 4/2/15 3:02 PM

128 The University of Chicago Law Review Dialogue [82:116 appearance of conflicts may be as harmful as conflicts themselves.55 Likewise, if an SRO appears untrustworthy or ineffective, gov- ernment agencies, citizens, and other stakeholders are likely to take steps that lower its relevance. A key lesson for the sharing economy is that, with self-regulation, perception and legitimacy go hand-in-hand, and it is essential that a clear image of objec- tivity and enforcement be created early, perhaps through some visible examples of enforcement and governmental partnering.

3. The power of reputation. Although the potential for digital reputation-and- monitoring systems to facilitate self-policing in the sharing economy is widely recognized,56 it has also become common for observers to question whether these reputation-based systems have significant regulatory power.57 However, many self- regulatory efforts that pre-date the digital economy rely exten- sively on the power of reputation to bolster compliance with indus- try rules. For example, the cotton industry, though not organized as an SRO, self-regulates via private law, with reputation play- ing a critical role. Merchant-to-mill and mill-to-mill cotton transactions are governed by rules adopted by various regional and national trade associations and exchanges,58 and these rules provide default contract provisions that cover issues such as per- formance, quality, payment, and damages.59 The associations also govern disputes—requiring arbitration and providing arbi- tration services—in which monetary sanctions may be imposed for noncompliance.60 New merchants attend a summer course at the Cotton Institute, learning trade rules and getting accultur- ated with the norms of the cotton industry.61

55 See Lawrence Lessig, Republic Lost: How Money Corrupts Congress—and a Plan to Stop It 226–48 (Twelve 2011). 56 See, for example, Arun Sundararajan, Why the Government Doesn’t Need To Regulate the Sharing Economy (Wired Oct 22, 2012), online at http://www.wired.com/ 2012/10/from-airbnb-to-coursera-why-the-government-shouldnt-regulate-the-sharing- economy (visited Feb 22, 2015). 57 See, for example, Tom Slee, Sharing Is Caring (Jacobin Mag Jan 24, 2014), on- line at https://www.jacobinmag.com/2014/01/sharing-and-caring (visited Feb 22, 2015). 58 See Lisa Bernstein, Private Commercial Law in the Cotton Industry: Creating Cooperation through Rules, Norms and Institutions, 99 Mich L Rev 1724, 1726–27 (2001). 59 See id at 1731–32. 60 See id at 1727–28. 61 See id at 1771–72.

SUNDARARAJAN_COHEN_FINAL (LMW).DOCX (DO NOT DELETE) 4/2/15 3:02 PM

2015] Self-Regulation and Innovation 129

Reputation is critical in the cotton industry because cotton grading is subjective, cotton prices are volatile, profit margins are low, and, as one person put it, “millions of dollars of business will be done on the basis of a thirty-second call.”62 The cotton in- dustry is thus able to enforce private law without significant noncompliance. Aside from deterring misconduct in the first place, reputational consequences can also reinforce the mone- tary sanctions available via arbitration: trading parties are less likely to unnecessarily challenge an industry arbitrator’s deci- sion in court.63

IV. SELF-REGULATORY SOLUTIONS FOR THE SHARING ECONOMY Peer-to-peer business facilitated by sharing-economy plat- forms has tremendous potential to expand grassroots entrepre- neurship and innovation by allowing society to tap into individ- ual abilities and aspirations that would otherwise not have been realized. Recent research results suggest that the economic ac- tivity on these platforms may benefit below-median-income con- sumers more than above-median-income consumers.64 Since the digital reach of sharing-economy platforms and the blurring of the personal and the professional dramatically increases the scale of peer-to-peer exchange, people may well conclude that anything other than delegated regulation imposes prohibitive governing costs on society. At the same time, new self-regulatory bodies need to be credible while being inclusive, policing misbe- havior without stifling experimentation and innovation. The preceding discussion highlights a number of key char- acteristics that could lead to successful self-regulation in the sharing economy. One characteristic of success that emerges quite clearly is the importance of being able to exert sufficient control to ensure compliance with whatever rules are developed. Put differently, for self-regulatory bodies to control their mem- bers, sanctions—including the ultimate punishment of expul- sion—must be costlier than the benefits of misbehavior.65 In some ways, this underscores the value of including the platforms themselves as enforcers of the self-regulatory solution. For

62 Bernstein, 99 Mich L Rev at 1746 (cited in note 58). 63 See id at 1740–41. 64 See generally Samuel P. Fraiberger and Arun Sundararajan, Peer-to-Peer Rental Markets in the Sharing Economy (NYU Stern School of Business Research Paper, Mar 6, 2015), online at http://ssrn.com/abstract=2574337 (visited Mar 24, 2015). 65 See Macey and Novogrod, 40 Hofstra L Rev at 976 (cited in note 54).

SUNDARARAJAN_COHEN_FINAL (LMW).DOCX (DO NOT DELETE) 4/2/15 3:02 PM

130 The University of Chicago Law Review Dialogue [82:116 example, Uber and Lyft have tremendous potential enforcement capabilities as regulatory entities: they control the channels for demand for their drivers, and as digital platforms, disconnecting a driver involves minimal transaction costs for the companies. The case of short-term accommodation is subtler. Granted, digital platforms like Airbnb are a primary channel for demand, and thus, are able to enforce trading limits and threaten plat- form expulsion with essentially no costs. However, since a ma- jority of Airbnb hosts rent out their primary residences only oc- casionally, the threat of expulsion from the platform may not be as significant a consequence. One might thus also involve a different set of entities in the self-regulatory solution for this sector: the increasingly ubiqui- tous co-op associations, condominium boards, and homeowners associations.66 These entities—which, for the sake of brevity, will all be referred to as “HOAs”—are entitled to regulate and collect dues from members, and they can control numerous aspects of the upkeep, maintenance, and living standards of the commu- nity.67 Given their broad-ranging authority, these organizations are quite likely to regulate short-term rentals. Courts have up- held bans on unit rentals, even if onerous and applied retroac- tively.68 Some HOAs do not ban but instead regulate unit rent- als, for instance by limiting the frequency or duration of rentals,69 or by requiring that owners pay a rental fee; provide names, ages, and license plates of the tenants; and perform background checks on tenants.70 The existence of these entities in the potential regulatory mix suggests an interesting potential division of regulatory re- sponsibility—delegate regulatory responsibility relating to in- formation asymmetry to platforms like Airbnb (whose interests are naturally aligned with the global aggregation of information

66 In 2002, the Community Association of America estimated that 50 percent of all new homes in major cities were a part of a community association. Renaud Le Goix and Chris J. Webster, Gated Communities, 2 Geography Compass 1189, 1190 (2008). 67 See Donna S. Bennett, Condominium Homeownership in the United States: A Selected Annotated Bibliography of Legal Sources, 103 L Library J 249, 268 (2011). 68 See, for example, Woodside Village Condominium Association, Inc v Jahren, 806 S2d 452, 460–62 (Fla 2002); Four Brothers Homes at Heartland Condominium II v Ger- bino, 691 NY2d 114, 114 (NY App 1999). 69 See Apple II Condominium Association v Worth Bank and Trust Co, 659 NE2d 93, 99 (Ill App 1995). 70 See Veronica Sanchez, Homeowner Fights HOA Rules (AZCentral.com July 21, 2009), online at http://archive.azcentral.com/12news/news/articles/2009/07/21/20090721HOAbattle .07212009-CR.html (visited Feb 22, 2015).

SUNDARARAJAN_COHEN_FINAL (LMW).DOCX (DO NOT DELETE) 4/2/15 3:02 PM

2015] Self-Regulation and Innovation 131 and the mitigation of adverse selection and moral hazard), and let HOAs play a key role in the regulation of local externalities, as the guest-noise and strangers-in-the-building externalities are typically local and primarily affect HOAs’ membership. Homeowners and renters have a continuous, high-bandwidth re- lationship with their HOA; these organizations are credible, can monitor compliance, and possess robust enforcement capabili- ties. Buildings and communities may then naturally differenti- ate into “Airbnb-friendly” and “Airbnb-free,” allowing future buyers and renters to self-select. We see reason to proceed cautiously with this division of re- sponsibility—after all, a growing fraction of residents supple- ment their income or cover their mortgage payments with short- term rental earnings, and switching residences involves signifi- cant transaction costs. Additionally, restricting economic activ- ity is one way to force individuals to internalize their negative externalities, but it is not the only one. Further, excessive re- strictions imposed by HOAs eliminate the individual benefits and positive spillovers associated with the economic activities that generate these negative externalities to begin with. The fragmented nature of HOAs is further cause for concern: if some uniformity in policy is desirable, their role may be better suited to enforcement rather than rule setting. A second insight from the modern-day SRO experience is that self-regulatory decisions may suffer from a lack of trans- parency or reviewability. This kind of opacity could limit the ability of public-interest groups to participate in the decision- making process or challenge the eventual outcome via the judi- cial system.71 Furthermore, without accountability to any out- side groups, there is no promise that organizations will protect or even consider the public interest in their decisionmaking.72 A self-regulatory solution for the sharing economy must therefore have some form of transparency and governmental oversight. One possible regime could involve a tripartite model in which third-party watchdogs evaluate SROs, and the level of govern- mental oversight and regulation is determined by a firm’s his- tory of compliance. We do not recommend mandating blanket marketplace transparency in the interest of promoting the self-regulatory

71 See Prakash, 39 Bus & Society at 184–85 (cited in note 48). 72 See Anthony Ogus, Rethinking Self-Regulation, 15 Oxford J Legal Stud 97, 98– 99 (1995).

SUNDARARAJAN_COHEN_FINAL (LMW).DOCX (DO NOT DELETE) 4/2/15 3:02 PM

132 The University of Chicago Law Review Dialogue [82:116 equivalent of “open government” because there are important consumer-privacy issues and trade secret costs that must be taken into account. Consider, instead, the fact that publicly traded corporations are, in some sense, “self-regulatory.” That is, they provide audited evidence (through their filings with the SEC), rather than being asked to provide raw operational data for a regulator to use in confirming compliance.73 Correspond- ingly, we might design a self-regulatory system for hospitality tax collection, for example. In such a system, rather than having each provider register with a government regulator, a platform like Airbnb would register as an SRO, use the data gathered from its operations to facilitate the collection and payment of taxes, and provide audited evidence of compliance without hav- ing to publish the detailed and sensitive underlying operational data that was used.

CONCLUSION The sharing economy promises tremendous decentralized innovation but needs a new regulatory framework in order to realize its potential. The approach that we propose is to utilize digital platforms as partners in the regulation of exchange, rather than view these platforms as adversaries or entities that require governmental regulation. This provides an attractive al- ternative to the cost of simply extending existing regulatory ap- proaches to the immense scale of digitally mediated peer-to-peer exchange. Self-regulation often emerges as a natural byproduct of eco- nomic exchange and has a long history of success. The experi- ence of self-regulation in a variety of modern industries reveals four factors essential to SRO success. First, an SRO must establish credibility early on through its performance. Second, self- regulatory actors must demonstrate strong enforcement capa- bilities. Third, SROs must be perceived as legitimate and inde- pendent. And finally, an SRO must take advantage of participants’ reputational concerns and social capital. Different market ineffi- ciencies for peer-to-peer transactions will require different enti- ties to act as partners in the self-regulatory solution. Furthermore,

73 See Miriam M. Weismann, Corporate Transparency or Congressional Window- Dressing? The Case against Sarbanes-Oxley as a Means to Avoid Another Corporate De- bacle: The Failed Attempt to Revive Meaningful Regulatory Oversight, 10 Stan J L Bus & Fin 98, 101–02 (2004).

SUNDARARAJAN_COHEN_FINAL (LMW).DOCX (DO NOT DELETE) 4/2/15 3:02 PM

2015] Self-Regulation and Innovation 133 governmental oversight and judiciously chosen transparency will increase the likelihood of SRO success. Making judicious choices about data transparency is truly critical. A progressive approach can have significant long-term benefits. For example, as the sharing-economy SROs (whether platforms themselves or third-party associations that emerge) establish a track record of credibility and enforcement and gain legitimacy as partners in regulation, they can then be called on to help invent self-regulatory solutions to societal issues that are especially difficult to address by centralized governmental inter- vention. One might imagine a variety of societal goals being achieved in part by the platforms applying machine-learning techniques to their data to detect patterns corresponding to, say, discriminatory practices, much like credit card issuers use automated systems to detect criminal fraud. This approach of regulatory delegation can yield far more expansive regulating- through-data alternatives than are feasible with complete transparency, and it suggests promising opportunities for self- regulation—ones that are appropriately reflective of the inter- esting meld of a decentralized marketplace and a centralized in- stitution that sharing-economy platforms represent.

Berkeley Technology Law Journal Volume 30 Article 18 Issue 4 Annual Review 2015

11-29-2015 Regulating the Sharing Economy Vanessa Katz

Follow this and additional works at: http://scholarship.law.berkeley.edu/btlj Part of the Law Commons

Recommended Citation Vanessa Katz, Regulating the Sharing Economy, 30 Berkeley Tech. L.J. 1067 (2015). Available at: http://scholarship.law.berkeley.edu/btlj/vol30/iss4/18

Link to publisher version (DOI) http://dx.doi.org/

This Article is brought to you for free and open access by the Law Journals and Related Materials at Berkeley Law Scholarship Repository. It has been accepted for inclusion in Berkeley Technology Law Journal by an authorized administrator of Berkeley Law Scholarship Repository. For more information, please contact [email protected].

REGULATING THE SHARING ECONOMY Vanessa Katz†

Internet commerce has enabled peer-to-peer (“P2P”) transactions on a larger scale than ever before.1 While the law slowly adapts to address P2P sales of goods, a wave of online marketplaces has emerged for P2P exchanges of services: short-term rentals of residential housing or office space,2 rentals of peer-owned assets,3 transportation network companies (“TNCs”) for short trips in personal vehicles or small planes,4 auction houses for temporary licenses,5 contracting services for daily chores and

© 2015 Vanessa Katz. † J.D. Candidate, 2016, University of California, Berkeley, School of Law. 1. See, e.g., David Gamage & Devin J. Heckman, A Better Way Forward for State Taxation of E-Commerce, 92 B.U. L. REV. 483, 484 (2012). 2. Some short-term rental platforms allow users to rent individual bedrooms or entire residences. E.g., AIRBNB, https://www.airbnb.com/; HOMEAWAY, http://www.homeaway.com/; ROOMORAMA, https://www.roomorama.com/. Others allow users to book a workspace, including individual desks and full conference rooms. E.g., PEERSPACE, http://www.peerspace.com/; ZIPCUBE, https://www.zipcube.com/ (all URLs last visited Feb. 26, 2015). 3. E.g., GETAROUND, https://www.getaround.com/ (cars); BOATBOUND, https://boatbound.co/ (boat rentals); CAMERALENDS, https://www.cameralends.com/ (cameras); BESCRAPPY, https://bescrappy.com/ (tools); LITERATOO, http://en.literatoo.com/ (books); STYLELEND, http://stylelend.com/ (high-fashion clothes) (all URLs last visited Feb. 26, 2015). 4. TNCs generally allow riders to contact black cars or other personal vehicles for short trips. LYFT, https://www.lyft.com/; SIDECAR, http://www.side.cr/. Some services have attempted to allow users to contract with small aircraft for flights. AIRPOOLER, https://airpooler.com/ (small aircraft); BLACKJET, https://www.blackjet.com/ (luxury planes) (all URLs last visited Feb. 26, 2015). P2P air transportation, however, has encountered both regulatory and financial setbacks. Scott Kirsner, ‘Uber for Private Jets’ Is Grounded, But CEO Says He Hopes to Raise More Money and Fly Again, BOSTON.COM (Dec. 19, 2013), http://www.boston.com/business/technology/innoeco/ 2013/12/uber_for_private_jets_is_groun.html; Jordan Valinksy, Sorry, But the FAA Has Decided Your ‘Uber for Planes’ Idea Can’t Fly, GIZMODO (Aug. 16, 2014, 5:00 PM), http://gizmodo.com/sorry-but-the-faa-has-decided-your-uber-for-planes-ide-1622734282/. 5. E.g., SWEETCH, http://www.getsweetch.com/ (last visited Feb. 26, 2015) (auction house for public parking spaces); BETRSPOT, http://www.betrspot.com/about (last visited Feb. 26, 2015) (auction house for spots in line and seats in public places). Reservation Hop originally allowed P2P auctions of restaurant reservations, but received so much negative attention that the founders created a new service for restaurant workers to exchange dropped shifts. Brian Mayer, ReservationHop Does a Hard Pivot: We Are Now OK Shift, BRIANMAYER.COM (Nov. 6, 2014, 8:00 AM), http://brianmayer.com/ 2014/11/reservationhop-does-a-hard-pivot-we-are-now-ok-shift/.

1068 BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 30:385 professional services,6 and lending.7 These businesses have experienced rapid growth over the past five years and are taking root worldwide.8 Commentators have dubbed this phenomenon the “sharing economy.”9 However, the sharing economy has raised difficult legal questions. P2P service marketplaces create uncertainty for participants and third parties because these services do not fall neatly into traditional legal categories.10 Businesses in the sharing economy have caused confusion under insurance, tax, employment, and civil rights statutes.11 Commentators have called the

6. E.g., TASKRABBIT, https://www.taskrabbit.com/ (chores, including assembling furniture and moving); FANCYHANDS, https://www.fancyhands.com/ (administrative tasks); HOMEJOY, https://www.homejoy.com/ (maid service); WASHIO, http://www.getwashio.com/ (laundry and dry cleaning); ZAARLY, https://www.zaarly.com/ (home services including plumbing and gardening) (all URLs last visited Feb. 26, 2015). 7. E.g., PROSPER, https://www.prosper.com/ (last visited Feb. 26, 2015); LENDING CLUB, https://www.lendingclub.com/ (last visited Feb. 26, 2015). This Note does not address the unique regulatory issues facing P2P lenders, which differ substantially from other sharing platforms. For background information on the regulation of P2P lenders, see generally Paul Slattery, Note, Square Pegs in a Round Hole: SEC Regulation of Online Peer-to-Peer Lending and the CFPB Alternative, 30 YALE J. ON REG. 233 (2013); Eric C. Chaffee & Geoffrey C. Rapp, Regulating Online Peer-to-Peer Lending in the Aftermath of Dodd-Frank: In Search of an Evolving Regulatory Regime for an Evolving Industry, 69 WASH. & LEE L. REV. 485 (2012). 8. The Rise of the Sharing Economy, ECONOMIST (Mar. 9, 2013), http://www.economist.com/node/21573104/. 9. There is no standard definition of the sharing economy. The use of the term “sharing” traces back to a seminal article by Yochai Benkler. Yochai Benkler, Sharing Nicely: On Shareable Goods and the Emergence of Sharing as a Modality of Economic Production, 114 YALE L.J. 273, 334 (2004); see also RACHEL BOTSMAN & ROO ROGERS, WHAT’S MINE IS YOURS: THE RISE OF COLLABORATIVE CONSUMPTION 58 (2010) (referencing Benkler’s article in describing the origins of what Botsman terms “collaborative consumption”). Many have argued that “sharing” is not an accurate description of P2P service models, in part because users exchange money. E.g., Brad Tuttle, Can We Stop Pretending the Sharing Economy Is All About Sharing?, TIME (June 30, 2014), http://time.com/money/2933937/sharing-economy-airbnb-uber-monkeyparking/. Other authors have framed the sharing economy as a new form of sustainable consumption and see the sharing economy as a societal shift to an access model rather than an ownership model. See generally, e.g., BOTSMAN & ROGERS, supra; LISA GANSKY, THE MESH: WHY THE FUTURE OF BUSINESS IS SHARING (2010); JANELLE ORSI, PRACTICING LAW IN THE SHARING ECONOMY: HELPING PEOPLE BUILD COOPERATIVES, SOCIAL ENTERPRISE, AND LOCAL SUSTAINABLE ECONOMIES (2013). 10. See, e.g., Molly Cohen & Corey Zehngebot, What’s Old Becomes New: Regulating the Sharing Economy, 58 BOSTON B.J. 6 (2014). 11. See infra Section IV.B.

2015] REGULATING THE SHARING ECONOMY 1069 sharing economy a “disruptive innovation”12—a product, service, or business model that repackages old technology to create a new market and thereby “disrupts” incumbent firms.13 Professor Nathan Cortez’s term “regulatory disruption”—an “innovation[] that disrupt[s] existing regulatory schemes”—aptly describes the sharing economy.14 For example, RelayRides is a P2P car rental service. Car owners list their personal vehicle on the RelayRides website, and users can then search available vehicles and schedule a rental. RelayRides provides insurance coverage of $1 million per accident to all registered car owners and drivers on the site. However, in 2012, a RelayRides rental driver caused a collision, killing himself and injuring four passengers in the other vehicle, resulting in estimated combined damages of $1.3 million.15 While the parties ultimately reached a settlement,16 the incident raised awareness of the risks of participating in the sharing economy and the need for regulatory protections. This Note surveys regulatory issues surrounding the sharing economy and assesses current proposed solutions to challenges the sharing economy creates. The analysis will largely focus on short-term rental platforms and TNCs because these are the largest sectors of the sharing economy.17 Part I explores what makes businesses in the sharing economy different from traditional service providers. Part II identifies existing laws for traditional service providers, and who has a stake in the outcome of efforts to reform those laws. Part III explains how regulators have attempted to adapt existing laws to the sharing economy. Finally, Part IV discusses what risks regulators should address moving forward. Part V concludes.

12. Robert T. Shannon, Disruptive Innovation Demands Delicate Regulation, LAW360 (Nov. 12, 2014), http://www.law360.com/articles/595081/disruptive- innovation-demands-delicate-regulation. 13. CLAYTON M. CHRISTENSEN, THE INNOVATOR’S DILEMMA: WHEN NEW TECHNOLOGIES CAUSE GREAT FIRMS TO FAIL 23–24 (1997); Nathan Cortez, Regulating Disruptive Innovation, 29 BERK. TECH. L.J. 175, 177 n.2 (2014). 14. Cortez, supra note 13, at 177. 15. Ron Lieber, Fatal Collision Makes Car-Sharing Worries No Longer Theoretical, N.Y. TIMES (Apr. 13, 2012), http://www.nytimes.com/2012/04/14/your- money/relayrides-accident-raises-questions-on-liabilities-of-car-sharing.html. 16. M.P. McQueen, Beware the Liability of Sharing Your Car With Strangers, INVESTOPEDIA (Oct. 11, 2013), http://www.investopedia.com/articles/ personal-finance/101013/beware-liability-sharing-your-car-strangers.asp. 17. The Sharing Economy: Boom and Backlash, ECONOMIST (Apr. 26, 2014), http://www.economist.com/news/business/21601254-consumers-and-investors- are-delighted-startups-offering-spare-rooms-or-rides-across-town/ (“Along with transport, the market most affected by sharing has been accommodation . . . .”).

1070 BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 30:385

I. DEFINING THE SHARING ECONOMY What makes the sharing economy unique is at the heart of an ongoing regulatory debate.18 Many argue that there is nothing new about the sharing economy. These critiques often analogize sharing services to traditional service providers, and portray the sharing model as a thinly veiled ruse for avoiding regulation.19 Others contend that the sharing economy disrupts existing regulatory schemes precisely because the business model is so new; for that reason, they argue that sharing companies require special regulatory treatment.20 In practice, the sharing model encompasses mundane features of existing online businesses as well as novel implementations of those features. This Note defines the sharing business model as (A) an online intermediary that (B) acts as a market for P2P services and (C) facilitates exchanges by lowering transaction costs.

A. SHARING PLATFORMS ACT AS ONLINE INTERMEDIARIES First, most sharing platforms operate through either a web portal or a mobile application (“app”). But unlike websites that act as online

18. Decision Adopting Rules and Regulations to Protect Public Safety While Allowing New Entrants to the Transportation Industry, Decision 13-09-045 (Cal. P.U.C. 2013), available at http://docs.cpuc.ca.gov/PublishedDocs/Published/ G000/M077/K192/77192335.PDF [hereinafter Cal. P.U.C. Decision]. 19. See, e.g., Tom Slee, Sharing and Caring, JACOBIN (Jan. 24, 2014), https://www.jacobinmag.com/2014/01/sharing-and-caring/ (arguing that the sharing economy obscures profit-driven behavior by tapping into long-standing leftist values such as “decentralization, sustainability, community-level connectedness, and opposition to hierarchical and rigid regulatory regimes”); Jon Evans, When Old-Economy Jobs Become New-Economy Gigs, TECHCRUNCH (Sept. 6, 2014), http://techcrunch.com/2014/09/06/serfing-the-wave/ (comparing “microentrepreneurs” in the sharing economy to serfs); Andrew Leonard, “Sharing economy” shams: Deception at the Core of the Internet’s Hottest Businesses, SALON (Mar. 14, 2014, 4:43 AM), http://www.salon.com/2014/03/14/sharing_economy_shams_deception_at_the_core_of_ the_internets_hottest_businesses/ (arguing that the sharing economy represents an exploitive movement by companies to profit by avoiding regulation while claiming to promote broader social welfare). 20. See, e.g., The Power of Connection: Peer-to-Peer Businesses Before the H. Comm. on Small Business, 113th Cong. 35 (2014) (written testimony of Philip Aeurswald) (“The question is . . . whether there is anything fundamentally new about the sorts of peer-to- peer businesses that have been proliferating in the past five years. The answer is yes.”); see also Roberta A. Kaplan, Regulation and the Sharing Economy, N.Y. L.J. (2014) (arguing that “innovative technologies [like sharing platforms] present new and fundamentally different circumstances that were unforeseen at the time of the original regulatory enactment”); CHRISTOPHER KOOPMAN ET AL., THE SHARING ECONOMY AND CONSUMER PROTECTION REGULATION: THE CASE FOR POLICY CHANGE 16–19 (2014) (arguing that the sharing economy is a self-regulating market and does not require government intervention).

2015] REGULATING THE SHARING ECONOMY 1071 storefronts, sharing platforms are not direct service providers.21 Instead, the platform allows “users” (purchasers of services) to connect and transact with “providers” (sellers of services). DogVacay, for example, allows dogsitters to connect with dog owners through an online web portal. The DogVacay platform itself does not provide dogsitting services or employ dog sitters.22 This functionality is not new; passive platforms for hosting user-provider transactions predate the sharing economy. In fact, the founders of DogVacay originally ran a small dogsitting service through Craigslist.23 The critical distinction between sharing platforms and other online services does not turn solely on the character of the services or parties involved, but rather on the degree of control the platform exercises over each transaction.24 Sharing platforms exercise control over transactions by directing the form and content of listings, issuing minimum quality standards for providers, providing an electronic payment system, and charging a transaction fee for each exchange.25 DogVacay charges fifteen

21. For example, Fetch! Pet Care, is a traditional pet-services provider with an online storefront. Users directly contract with Fetch for pet services. In other words, Fetch acts as a provider, rather than as a platform. FETCH! PET CARE, http://www.fetchpetcare.com/ (last visited Feb. 2, 2014). 22. How It Works, DOGVACAY, http://dogvacay.com/how-it-works/ (last visited Feb. 26, 2015). DogVacay requires that pet-sitters receive payment through the platform. Host Frequently Asked Questions, DOGVACAY, http://dogvacay.com/faq-host (last visited Feb. 26, 2015) (“all payments must go through DogVacay.com”). 23. Erin Griffith, Exclusive: DogVacay Raises $25 Million, FORTUNE (Nov. 10, 2014, 10:16 AM), http://fortune.com/2014/11/10/dogvacay-raises-25-million/. Likewise, Couchsurfing has acted as a passive, user-provider platform for short-term rentals since 2004. About Us, COUCHSURFING, http://about.couchsurfing.com/about/ (last visited Feb. 26, 2015). Car sharing programs have existed since the early 70s, and some online car sharing networks have operated for over a decade. Susan Shaheen et. al, Carsharing in Europe and North America: Past, Present and Future, 52 TRANSP. QUART. 35 (1998); About Us, CITY CARSHARE, https://citycarshare.org/about-us/ (last visited Feb. 26, 2015). 24. The degree of control exercised is critical for establishing indirect liability. See infra Subsection III.C.2. The issue of control has surfaced in many regulatory and judicial decisions regarding sharing platforms. E.g., Cal. P.U.C. Decision, supra note 18, at 72– 75 (“We reject Uber’s assertion that TNCs are nothing more than an application on smart phones, rather than part of the transportation industry. Uber . . . performs essentially the same function as a limousine . . . .”); Bos. Cab Dispatch, Inc. v. Uber Techs., Inc., No. 13-10769-NMG, 2014 WL 1338148, at *6 (D. Mass. Mar. 27, 2014) (“[T]here is sufficient evidence that Uber exercises control over (or is ‘in charge of’) vehicles-for-hire that compete with plaintiffs in the private transportation business.”). 25. Courts have considered similar issues in the context of personal jurisdiction over internet sellers. See, e.g., Zippo Mfg. Co. v. Zippo Dot Com, Inc., 952 F. Supp. 1119, 1124 (W.D. Pa. 1997) (establishing a sliding scale test where “the likelihood that personal jurisdiction can be constitutionally exercised is directly proportionate to the

1072 BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 30:385 percent of each booking for its services,26 and transaction fees can range from three to twenty percent.27 Sharing companies fall somewhere along a spectrum between purely passive message boards and direct service providers. Message boards reserve the right to remove listings, but provide minimal guidance on the form and content of listings. These platforms have no employment relationship with providers, and no financial stake in any particular transaction.28 Direct service providers, by contrast, exert complete control over the form and content of listings, and directly benefit from every purchase. Determining whether a platform acts as a passive intermediary or exerts sufficient control over providers to establish liability is a highly fact-dependent question.29 Sharing platforms generally seek to minimize their own liability by characterizing their services as close equivalents to message boards.30 Many critics, however, argue that sharing platforms in practice operate like direct service providers.31 Some have decried sharing platforms for using deceptive tactics to evade regulation.32 Though this nature and quality of commercial activity that an entity conducts over the Internet.”). Unlike sales of goods through ecommerce, the provision of services to an injured party through a sharing platform would typically constitute a sufficient connection to establish personal jurisdiction. 26. Host Frequently Asked Questions, DOGVACAY, http://dogvacay.com/faq-host/ (last visited Feb. 26, 2015). 27. What Are Host Service Fees?, AIRBNB, https://www.airbnb.com/ help/article/63/ (last visited Feb. 26, 2015) (charging a three percent host service fee per booking, in addition to a separate guest service fee); Ellen Huet, Uber Now Taking Its Biggest UberX Commission Ever—25 Percent, FORBES (Sept. 22, 2014), http://www.forbes.com/sites/ellenhuet/2014/09/22/uber-now-taking-its-biggest-uberx- commission-ever-25-percent/. 28. See, e.g., Terms of Use, CRAIGSLIST, http://www.craigslist.org/ about/terms.of.use/ (last visited Feb. 26, 2015) (requiring a fee for certain categories of posts, but having no stake in any transaction resulting from a post); Terms of Use, ADOOS, http://support.adoos.com/answer/6/ (last visited Feb. 26, 2015). 29. See infra Section III.C. 30. E.g., United States and Canada: Terms & Conditions, HOMEJOY, https://www.homejoy.com/terms/us/en/ (last visited Feb. 2, 2015) (“THE COMPANY DOES NOT PROVIDE CLEANING SERVICES, AND THE COMPANY IS NOT A CLEANING SERVICE PROVIDER.”); Terms of Service, GETAROUND, https://www.getaround.com/terms/ (last visited Feb. 2, 2015) (“GETAROUND WILL NOT BE A PARTY TO OR IN ANY WAY MONITOR ANY TRANSACTION BETWEEN YOU AND THIRD-PARTY PROVIDERS OF PRODUCTS OR SERVICES.”). 31. E.g., Letter from Mark Saber, Central Penn Taxi Ass’n, to Penn. Pub. Util. Comm’n (Aug. 28, 2014), available at http://www.puc.state.pa.us/transport/ motor/EnBanc/Testimony-CPATA_082814.pdf (“A transportation network company is nothing more than a call or demand carrier in disguise.”). 32. See, e.g., Avi Asher-Schapiro, Against Sharing, JACOBIN (Sept. 19, 2014),

2015] REGULATING THE SHARING ECONOMY 1073 debate remains unresolved in many jurisdictions, the intermediary status of sharing platforms affects both platform liability and local regulatory authority over platforms.33 B. SHARING PLATFORMS ARE MARKETS FOR PEER-TO-PEER SERVICES Second, sharing platforms allow informal, small-scale actors to exchange services online. Catering to peer-to-peer transactions, rather than to business-to-consumer transactions, serves two purposes. It allows providers to profit from underused personal assets,34 or to market their skills on a freelance basis.35 For example, instead of leaving a car in the garage during a family vacation, P2P car rental platforms allow car owners to earn money by leasing their personal vehicles. Per-task cleaning services like Homejoy give providers exceptional flexibility over their working hours.36 In addition, informal providers may not be subject to the same regulations as larger businesses.37 However, sharing platforms do not necessarily prohibit professional or skilled providers. In fact, some platforms like Zaarly, a per-task contractor service for home repairs and cleaning, require that all providers have proper accreditations.38 Although sharing platforms generally attempt to exclude businesses and large-scale operators from participating on their sites,39 data from short-term rental

https://www.jacobinmag.com/2014/09/against-sharing/ (criticizing TNCs and the sharing economy generally for depressing wages and creating unfair working conditions). 33. See infra Section III.C. 34. See, e.g., BOTSMAN & ROGERS, supra note 9, at 83 (claiming that sharing platforms exploit the “idling capacity” of underused resources). 35. See, e.g., The New ‘Sharing Economy’ Can Enrich Micro-Entrepreneurs But at What Cost?, PBS (Oct. 10, 2014), http://www.pbs.org/newshour/bb/ sharing-economy-enrich-micro-entrepreneurs-promote-unregulated-big-business/ (discussing the impact of “microentrepreneurs” on existing businesses). 36. Start Cleaning with Homejoy, HOMEJOY, https://www.homejoy.com/ apply/ (last visited Feb. 26, 2015) (“You choose when and where you want to work within Homejoy’s availability. It’s entirely up to you!”). 37. See infra Sections II.B, III.B. 38. How Zaarly Works, ZAARLY, https://www.zaarly.com/howzaarlyworks/ (last visited Feb. 26, 2015). 39. E.g., Host Frequently Asked Questions, DOGVACAY, http://dogvacay.com/faq- host/ (last visited Feb. 26, 2015) (limiting dog-sitters to hosting three dogs per day); Discover the Difference, VRBO, http://www.vrbo.com/global/advantages.htm (last visited Feb. 26, 2015) (differentiating short-term rentals from hotels through a comparison chart). In many cases, sharing platforms restrict access to individuals rather than businesses, but do not necessarily restrict the hours each individual works per week. E.g., Start Cleaning with Homejoy, HOMEJOY, https://www.homejoy.com/apply/ (last visited Feb. 26, 2015) (seeking applications from individual cleaners).

1074 BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 30:385 platforms and anecdotal evidence from TNCs suggests that a certain percentage of providers transact at substantially higher volumes than most others.40 Sharing platforms offer a wide range of services, from boat rentals to electrical repair. Unlike ecommerce platforms for sales of goods, exchanges of services generally require users and providers to interact in person. Thus, sharing platforms raise distinct legal issues from ecommerce platforms. Contracting with strangers online involves many risks, and therefore much higher transaction costs than similar exchanges with larger businesses.41

40. The New York Attorney General’s recent investigation into Airbnb suggests a stark divide between commercial and casual hosts. Eric T. Schneiderman, N.Y. STATE ATTORNEY GENERAL, AIRBNB IN THE CITY (2014), available at http://www.ag.ny.gov/pdfs/Airbnb%20report.pdf [hereinafter N.Y. Attorney General Report]. While most hosts list casually, the small minority of full-time commercial hosts are responsible for the vast majority of bookings. Id. at 10–11. Airbnb, however, has emphasized that most hosts use the service on an occasional basis, and that Airbnb hosts often rely on the service to help pay for rent rather than treating it as a business. Hotels vs. Regular New Yorkers, AIRBNB, http://publicpolicy.airbnb.com/hotels-vs-regular-new- yorkers/ (last visited Feb. 26, 2015). Some have even argued that scaling up an Airbnb business is not profitable outside a few high-density locations. Nin-Hai Tseng, Profiting $13K a Year on an Airbnb Rental? Maybe., FORTUNE (Nov. 7, 2013, 7:46 PM), http://fortune.com/2013/11/07/profiting-13k-a-year-on-an-airbnb-rental-maybe/. Commentators have suggested a similar division for TNC drivers. Full-time drivers likely work between forty and seventy hours per week. Part-time drivers likely work thirty hours or less per week. Johana Bhuiyan, What Uber Drivers Really Make (According to Their Pay Stubs), BUZZFEED (Nov. 19, 2014), http://www.buzzfeed.com/johanabhuiyan/what- uber-drivers-really-make-according-to-their-pay-stubs; see also Jack Flanagan, The Trick Uber Drivers Use to Boost Their Salaries, DAILYDOT (June 20, 2014), http://www.dailydot.com/business/uber-driver-actual-salary-earnings-wage-90000/ (noting that “smart” part-time drivers can earn more per hour than full-time drivers). Many full-time drivers are former cab drivers, and some may continue to work for both a traditional dispatch service as well as multiple TNCs. Harry Campbell, Do Drivers Make More Money With Lyft or Uber?, THE RIDESHARE GUY (Aug. 13, 2014), http://therideshareguy.com/do-drivers-make-more-money-with-lyft-or-uber/; Megan Burks, Taxi-Turned-Uber Drivers Get a New Lease on Life, KPBS.ORG (Aug. 15, 2014), http://www.kpbs.org/news/2014/aug/15/san-diegos-taxi-turned-uber-drivers-get-new- lease-/. 41. Interacting with strangers is, in fact, necessary in most offline business transactions. For a discussion of why users perceive that peer-to-peer transactions involve greater risks than traditional services, see, for example, Jason Tanz, How Airbnb and Lyft Finally Got Americans to Trust Each Other, WIRED (Apr. 23, 2014), http://www.wired.com/2014/04/trust-in-the-share-economy/.

2015] REGULATING THE SHARING ECONOMY 1075

C. SHARING PLATFORMS REDUCE TRANSACTION COSTS To reduce transaction costs for users and providers, sharing platforms rely on a variety of features to decrease the information cost of determining whether a provider is trustworthy. Many sharing platforms incorporate a social networking feature—they allow users to (1) create individual profiles, (2) articulate connections to other users’ profiles, and (3) search their connections.42 This feature allows sharing platforms to track user and provider account information and activity,43 and create “reputation systems”—a review and rating tool for screening bad actors.44 In addition, sharing platforms lower transaction costs by standardizing exchanges and providing quality or safety guarantees. Measures to standardize exchanges include user-friendly and uniform interfaces, location-based searches, price comparison, account activity and payment records, and scheduling and booking tools.45 Quality or safety guarantees might consist of insurance policies, customer service hotlines, cancellation and refund policies, reputations systems, provider training, and background checks.46 Embedding these tools and services into sharing platforms allows users and providers to transact at higher volumes, and encourages casual participants to join the market.47 Sharing platforms thus facilitate exchanges that might otherwise never occur due to high

42. See PAUL LAMBERT, SOCIAL NETWORKING: LAW, RIGHTS AND POLICY 13– 15 (2014). Examples of social network services are friend websites (e.g., Facebook), blogs (e.g., Tumblr), and message boards (e.g., Reddit). To some degree, users can accomplish the same transactions through social networks as through sharing platforms. For example, a dogowner could find a dogsitter by posting a status update or group message on Facebook. The dogowner’s Facebook connections would then respond to the dogowner through Facebook. However, this does not make sharing platforms the functional equivalent of social networks. On a social network, the user searches for providers within her circle of connections, rather than through a specialized market for services. And sharing platforms provide additional features that facilitate exchanges between parties. 43. For a discussion of privacy and account tracking on sharing platforms, see infra Section III.C.1. 44. E.g., Erica Swallow, The Rise of the Reputation Economy, FORBES (Oct. 9, 2013), http://www.forbes.com/sites/ericaswallow/2013/10/09/reputation-economy/. For a discussion of reputation systems, see infra Section IV.B.2. 45. Many sharing platforms also enable nonmonetary exchanges. See, e.g., 99GAMERS, https://99gamers.com/ (last visited Feb. 26, 2015). 46. See, e.g., Tanz, supra note 41. 47. It is important to note that, unlike an online storefront or social network, sharing platforms function as two-sided markets. In short, sharing platforms coordinate two constituencies—providers and consumers—and the value of the platform for each constituency depends on the size of the other, also known as an indirect network effect. David S. Evans & Michael Noel, Defining Antitrust Markets When Firms Operate Two- Sided Platforms, 2005 COLUM. BUS. L. REV. 667 (2005).

1076 BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 30:385 transaction costs.48 In part because they encourage high volumes of P2P transactions, these platforms have also raised new regulatory concerns.

II. THE EXISTING REGULATORY LANDSCAPE This Part describes existing regulatory frameworks that might affect sharing platforms and explains why these regulations matter. Many P2P services are closely analogous to traditional service providers. For that reason, sharing platforms often seem to fall under the purview of laws governing traditional service providers.49 The following discussion will address two features of these regulatory frameworks. First, modern statutory schemes have roots in common law special relationships, and accordingly impose limits on the freedom of contract between users and providers on sharing platforms. Second, state and local regulatory agencies create complex, overlapping requirements for providers and platforms.50 However, sharing platforms differ substantially from traditional service providers, and the sharing economy does not neatly fit into existing regulatory frameworks.51 Many providers and platforms cannot comply with local regulations. Others face uncertainty as a result of gaps or ambiguities in the law.52 To address these problems, sharing platforms have advocated strongly for new rules tailored specifically for P2P service providers.53 This Part concludes by discussing the various interests at stake in the ongoing debate over how to regulate the sharing economy.

48. Kaplan, supra note 20 (“Sharing apps and websites act like ‘virtual matchmakers’ by facilitating relationships that otherwise might be too costly or burdensome to arrange.”). 49. E.g., Shannon, supra note 12 (describing the “crux” of a suit by the Chicago taxi industry against the city “over enforcement of its existing public chauffeur rules” as a “claim by the taxi industry that [TNCs] have avoided regulations imposed on taxis, even though both are engaged in the same basic activity”). 50. Infra Sections II.A–B. 51. See infra Section III.B. 52. For a discussion of why sharing platforms elude categorization under current rules, see Emily Badger, Why We Can’t Figure Out How to Regulate Airbnb, WASH. POST (Apr. 23, 2014), http://www.washingtonpost.com/blogs/ wonkblog/wp/2014/04/23/why-we-cant-figure-out-how-to-regulate-airbnb/. 53. E.g., Donna Tam, Airbnb, Lyft Partner with New Share-Economy Advocacy Group, CNET (July 31, 2013, 12:50 PM), http://www.cnet.com/news/ airbnb-lyft-partner-with-new-share-economy-advocacy-group/ (describing a partnership between several platforms and advocacy group Peers, an organization providing support to sharing platform users and providers); Carolyn Said, Uber, Lyft, Airbnb Harness Users to Lobby Lawmakers for Them, SFGATE (Jan. 11, 2015, 10:18 AM), http://www.sfgate.com/business/article/Uber-Lyft-Airbnb-harness-users-to-lobby-60055 62.php. Several TNCs have hired prominent lobbyists to advocate for responsible

2015] REGULATING THE SHARING ECONOMY 1077

A. COMMON LAW SPECIAL RELATIONSHIPS PRINCIPLES INFORM MODERN RULES Many sharing platforms offer services with clear common law analogues: TNCs provide transportation services analogous to common carriers,54 P2P rentals are governed by product liability,55 per-task contractors are affected by various tort law doctrines—particularly premises liability,56 and short-term rentals lie at the nexus of landlord- tenant and innkeeper law.57 In general, these bodies of law reflect a tension between freedom of contract and the strong public interest inherent in certain transactions.58 Based on certain “special relationships” between service providers and consumers, various common law doctrines establish heightened duties for the service provider.59 In the context of this Note, special relationships include landlord-tenant, common carrier-passenger, innkeeper-guest, and lessor-lessee—a close parallel to manufacturer- consumer under products liability. This Section explores three assumptions that justify the common law treatment of “special relationships,” and how these assumptions continue to inform modern regulatory frameworks.

regulation. James R. Hood, Lyft Hires Its First Lobbyists to Clear the Way for Its Ride- Sharing Service, CONSUMER AFFAIRS (Apr. 17, 2014), http://www.consumeraffairs.com/news/lyft-hires-its-first-lobbyists-to-clear-the-way-for- its-ride-sharing-service-041714.html; Rosalind S. Helderman, Uber Pressures Regulators by Mobilizing Riders and Hiring Vast Lobbying Network, WASH. POST (Dec. 13, 2014), http://www.washingtonpost.com/politics/uber-pressures-regulators-by-mobilizing-riders- and-hiring-vast-lobbying-network/2014/12/13/3f4395c6-7f2a-11e4-9f38-95a187e4c1f7_ story.html. In addition, Airbnb has funded several economic studies to demonstrate its positive impact on communities and influence lawmakers. Emily Badger, Why Airbnb Wants You to Know How Much Its Users Are Spending, WASH. POST (Mar. 25, 2014), http://www.washingtonpost.com/blogs/wonkblog/wp/2014/03/25/why-airbnb-wants- you-to-know-how-much-its-users-are-spending/. 54. See Shannon, supra note 12. 55. See KOOPMAN, supra note 20, at 18 (suggesting generally that “torts and product liability law, and other legal remedies exist when things go wrong” on sharing platforms). 56. For a discussion of premises liability in the context of short-term rentals, see Ron Lieber, The Insurance Market Mystifies an Airbnb Host, N.Y. TIMES (Dec. 19, 2014), http://www.nytimes.com/2014/12/20/your-money/the-insurance-market-mystifies-an- airbnb-host.html. Similar concerns apply for per-task contractors, but the liable party in that case is likely the user who invites a contractor to her property. See infra Section III.C. 57. See infra text accompanying notes 77–80. 58. See, e.g., Joseph William Singer, No Right to Exclude: Public Accommodations and Private Property, 90 NW. U. L. REV. 1283, 1304–31 (1996) (discussing the origins of duties under public accommodations law and whether these duties applied to retail stores). 59. E.g., id.

1078 BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 30:385

First, common law special relationships assume that certain transactions pose unique health, safety, and financial concerns for consumers.60 Landlord-tenant law illustrates this heightened public interest in private transactions. In the mid-twentieth century, many jurisdictions adopted the implied warranty of habitability, “a duty on landlords to provide residential tenants with a habitable dwelling unit.”61 This wave of reform reflected growing acknowledgement of the importance of minimum housing quality standards for public health and safety,62 and strongly influenced modern landlord-tenant statutes. Most states have codified the implied warranty of habitability.63 In addition, many states have enacted some form of the Uniform Residential Landlord and Tenant Act to establish remedies and protections against retaliation for tenants.64 Today, most jurisdictions prohibit landlords from taking self-help measures to evict tenants, and embed at least some protections for tenants in summary eviction proceedings.65 Related to the unique health and safety risks of certain transactions, common law special relationships assume that certain service providers occupy a disproportionately strong bargaining position in relation to

60. See id. at 1292–94 (describing the prevailing view among scholars that “necessity required special obligations to protect travelers from hardship when they had no place to sleep at night and were vulnerable to bandits on the highways”); Escola v. Coca Cola Bottling Co. of Fresno, 24 Cal. 2d 453, 441 (1944) (Traynor, J., concurring) (noting the inevitable occurrence of accidents resulting from defective products, and arguing that “[a]gainst such a risk there should be general and constant protection and the manufacturer is best situated to afford such protection”). 61. Roger A. Cunningham, The New Implied and Statutory Warranties of Habitability in Residential Leases: From Contract to Status, 16 URB. L. ANN. 3, 74 (1979). One of the most widely cited cases on point, Javins v. First Nat’l Realty Corp., 428 F.2d 1071 (D.C. Cir. 1970), described the modern lease agreement as a contract for “a well[-]known package of goods and services—a package which includes not merely walls and ceilings, but also adequate heat, light and ventilation, serviceable plumbing facilities, secure windows and doors, proper sanitation, and proper maintenance.” Id. at 1074. 62. David A. Super, The Rise and Fall of the Implied Warranty of Habitability, 99 CALIF. L. REV. 389 (2011), 401–03; Gerald Korngold, Whatever Happened to Landlord- Tenant Law, 77 NEB. L. REV. 703, 706 (1998); Mary Ann Glendon, The Transformation of American Landlord-Tenant Law, 23 B.C. L. REV. 503, 518–19 (1982). 63. Super, supra note 62, at 394 (asserting that the implied warranty of habitability was adopted by “almost every state’s legislature or courts”). 64. Gary Goldman, Uniform Commercial Landlord and Tenant Act—A Proposal to Reform “Law Out of Context”, 19 T.M. COOLEY L. REV. 175, 185–91 (2002); Olin L. Browder, The Taming of a Duty—The Tort Liability of Landlords, 81 MICH. L. REV. 99, 113–16 (1982). 65. The Eviction Process, CA. DEP’T CONSUMER AFFAIRS, http://www.dca.ca.gov/publications/landlordbook/evictions.shtml (last visited Feb. 26, 2014); Browder, supra note 64, at 111–12.

2015] REGULATING THE SHARING ECONOMY 1079 consumers.66 For example, both innkeepers and common carriers have a duty of non-discrimination—absent extenuating circumstances, each must accept all guests or passengers.67 Denying innkeepers and common carriers the right to refuse service represents a substantial restraint on freedom of contract. However, the particular vulnerability of travelers has historically justified this restraint. Travelers are unfamiliar with local pricing, have limited access to competitor services, and are likely subject to time or circumstantial constraints on their ability to negotiate (e.g., the need to find lodging before sunset, or arrange transportation during a snowstorm).68 Modern statutory regimes likewise assume that guests and passengers require special protections, and most jurisdictions today have codified the duty of non-discrimination for hotels and taxis.69 In the case of taxis, local ordinances additionally specify metering requirements and set rates.70 Finally, common law special relationships typically assume that the service provider is the least-cost avoider—the party who can adopt precautions against a given risk at the lowest cost.71 These relationships therefore allocate greater liability to the service provider rather than the consumer. For example, in some jurisdictions, a commercial lessor qualifies as a “seller or distributor” under products liability law.72 Lessors may incur liability for product defects or for failure to warn the consumer of dangers related to use of the product.73 Courts have typically justified these heightened duties based on the relative positions of the seller and the consumer: the seller has superior knowledge of the condition of a product, and consumers cannot reasonably inspect or protect against the risks

66. See, e.g., Robert M. Hardaway, Taxi and Limousines: The Last Bastion of Economic Regulation, 21 HAMLINE J. PUB. L. & POL’Y 319, 357–58 (2000). 67. E.g., Singer, supra note 58, at 1304–31; JOHN E.H. SHERRY, THE LAWS OF INNKEEPERS: FOR HOTELS, MOTELS, RESTAURANTS, AND CLUBS 33–34 (1981). 68. SHERRY, supra note 67, at 126; Hardaway, supra note 66, at 357–58; Edmund W. Kitch, Marc Isaacson & Daniel Kasper, The Regulation of Taxicabs in Chicago, 14 J. L. & ECON. 285, 306 (1971). 69. See JACK P. JEFFERIES & BANKS BROWN, UNDERSTANDING HOSPITALITY LAW 35–37 (2001). 70. Lee A. Harris, Taxicab Economics: The Freedom to Contract for a Ride, 1 GEO. J.L. & PUB. POL’Y 195, 201–05 (2002). 71. Richard A. Posner, Guido Calabresi’s The Costs of Accidents: A Reassessment, 64 MD. L. REV. 12, 16 (2005) (describing and critiquing Guido Calabresi’s claims regarding liability for the “cheapest cost avoider”). 72. RESTATEMENT (THIRD) OF TORTS § 20(b) (1998). 73. RESTATEMENT (THIRD) OF TORTS § 1 (1998).

1080 BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 30:385 associated with many modern manufacturing processes.74 Just as sellers and lessors represent the least-cost avoider, so do hotels, taxis, landlords, and other service providers with respect to the average guest, passenger, tenant, or consumer. As discussed below, sharing platforms disrupt some of the fundamental assumptions underlying common law “special relationships.”75 Unlike two-party relationships between service providers and consumers, sharing platforms create three-party relationships between the platform, provider, and user. These new relationships require a different allocation of liability and statutory protection. Consequently, the influence of common law special relationships on modern statutory frameworks often creates problems for sharing platforms. Many jurisdictions, for instance, prohibit leases of fewer than thirty days.76 These ordinances prevent landlords from operating residential properties as hotels, but also bar most short-term rental platforms from operating.77 However, guests who stay for over thirty days become entitled to statutory protections for tenants.78 If those guests refuse to leave the property at the end of the rental period, hosts must pursue formal eviction proceedings.79 By contrast, many jurisdictions allow hotels to take self-help measures to evict transient guests.80 This example demonstrates how common law special relationships can affect the rights and remedies of sharing platform users under existing regulatory regimes.

B. REGULATORY AUTHORITY IS DIFFUSE, LOCALIZED, AND SPECIALIZED Beyond the influence of common law “special relationships,” modern statutory frameworks create many other challenges for sharing platforms. First, regulatory authority governing traditional service providers is highly

74. See Price v. Shell Oil Co., 466 P.2d 722, 726 (Cal. 1970) (“[W]e think it makes good sense to impose on the lessors of chattels the same liability for physical harm which has been imposed on the manufacturers and retailers. The former, like the latter, are able to bear the cost of compensating for injuries . . . .”). 75. See infra Section III.C. 76. Charles Gottlieb, Residential Short-Term Rentals: Should Local Governments Regulate the “Industry”?, 65 PLAN. & ENVTL. L. 4 (2013). 77. Id. 78. Chuck Thompson, ‘Professional Scammers’ Refuse to Leave Airbnb Host’s House, CNN (July 24, 2014), http://www.cnn.com/2014/07/23/travel/airbnb-squatters/. 79. Id. 80. Who Is A “Landlord” and Who Is A “Tenant”, CAL. DEP’T CONSUMER AFFAIRS, http://www.dca.ca.gov/publications/landlordbook/whois.shtml (last visited Feb. 26, 2015).

2015] REGULATING THE SHARING ECONOMY 1081 localized. Sharing platforms must navigate rules promulgated by multiple agencies at both state and local levels. Second, the relevant local codes are complex and, as discussed in the previous Section, impose heightened compliance burdens on service providers. Sharing platforms must contend with multiple regulatory agencies at both the state and local levels. Short-term rentals are typically regulated at the city level.81 Yet state laws can also affect short-term rental platforms. Notably, Florida has prohibited local legislatures from passing new restrictions on short-term rentals since 2010.82 TNCs, by contrast, are often subject to regulation by state public utility commissions and state legislatures.83 However, in many jurisdictions cities retain concurrent or separate regulatory authority over TNCs. The Pennsylvania Public Utilities Commission, for instance, has granted certain TNCs a two-year experimental license, but the city of Philadelphia exercises independent authority over for-hire vehicles and continues to ban TNCs.84 Moreover, the local and state rules governing traditional service providers are highly complex, and thus compliance may be costly for providers. For short-term rentals, relevant local ordinances include residential zoning restrictions, health and fire codes, transient occupancy taxes, and licensing and permitting regimes.85 These ordinances may also affect per-task contractors operating on residential properties, such as DogVacay.86 Taxis are subject to detailed rules governing metering rates, required coverage areas, vehicle inspections, driver background checks, and licensing regimes.87 Enforcement of these rules also takes place at the local

81. See Gottlieb, supra note 76. 82. FLA. STAT. § 509.032(7)(b) (2014); see also Gottlieb, supra note 76. 83. Niraj Chokshi, Colorado Passes Nation’s First Law Regulating UberX, Lyft, WASH. POST (June 6, 2014), http://www.washingtonpost.com/blogs/govbeat/wp/ 2014/06/06/colorado-passes-nations-first-law-regulating-uberx-lyft/. 84. Paul Nussbaum, PUC Approves UberX For State, Not Philadelphia, PHILLY (Nov. 15, 2014), http://articles.philly.com/2014-11-15/business/56394938_1_ride-sharing- service-uberx-uber-service. 85. See Gottlieb, supra note 76. 86. E.g., KENNELS, CTY. SAN DIEGO, PLANNING & DEV. SERVS. (2014), available at http://www.sandiegocounty.gov/pds/zoning/formfields/PDS-359.pdf (prescribing zoning regulations for kennels, and defining kennels as any lot housing “seven or more dogs, cats, or similar small animals”). 87. Harris, supra note 70, at 201–12.

1082 BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 30:385 level,88 and agencies tasked with investigating violations of city zoning or health codes may have limited resources. Highly localized and specialized regulatory frameworks pose a challenge for sharing platforms, particularly those that operate nationwide. To clarify ambiguities in local regulations or contest regulations that prohibit sharing services, platforms must advocate for reform on a city-by- city basis.89 Even as cities and states begin to pass legislation tailored to sharing platforms, each jurisdiction takes a different approach. Sharing platforms must therefore adapt to a wide range of regulatory solutions. However, many other interest groups also have an interest in the outcome of regulatory action governing sharing platforms. C. MANY INTEREST GROUPS HAVE A STAKE IN THE REGULATION OF SHARING PLATFORMS Most jurisdictions have taken an “experimental” approach to regulating sharing platforms,90 and a variety of interest groups continue to shape the rulemaking process.91 The growth of the sharing economy has created substantial economic opportunities for users, providers, and platforms. The New York Attorney General’s Office predicts that in 2014 Airbnb alone will conduct $282 million of short-term rental transactions within New York.92 Moreover, these platforms have acquired broad user bases.93

88. For example, the Mayor’s Office of Special Enforcement enforces rules against illegal hotels in New York City. Mayor’s Office of Special Enforcement, N.Y. CITY, http://www.nyc.gov/html/cjc/html/quality/mose.shtml (last visited Feb. 26, 2015). 89. See, e.g., Rosalind S. Helderman, Uber Pressures Regulators by Mobilizing Riders and Hiring Vast Lobbying Network, WASH. POST (Dec. 13, 2014), http://www.washingtonpost.com/politics/uber-pressures-regulators-by-mobilizing- riders-and-hiring-vast-lobbying-network/2014/12/13/3f4395c6-7f2a-11e4-9f38- 95a187e4c1f7_story.html. 90. E.g., Kim Lyons, PUC Grants Uber License to Operate Experimental Service, PITTSBURGH POST-GAZETTE (Nov. 13, 2014, 11:43 PM), http://www.post gazette.com/business/2014/11/13/UbergetslicenseapprovalfromPennsylvaniaPUC/stories /201411130305; Allen Wallace, Uber Granted Temporary License to Operate in South Carolina, COLA DAILY (Jan. 30, 2015), http://coladaily.com/2015/01/30/uber-granted- temporary-license-to-operate-in-south-carolina/. 91. See, e.g., Polly Mosendz, Face-Off: NYC Lawmakers Grill Airbnb on Illegal Hotels, NEWSWEEK (Jan. 21, 2015), http://www.newsweek.com/face-nyc-lawmakers-grill- airbnb-illegal-hotels-301060/ (describing the various interest groups participating in a recent New York City Council meeting to debate the merits of changing the city’s short- term rental laws). 92. These figures reflect the overall revenue of hosts; as noted in Part I, the platform itself takes only a six to twelve percent transaction fee. Airbnb’s direct revenue from transaction fees is approximately $30 million. N.Y. Attorney General Report, supra note 40, at 7.

2015] REGULATING THE SHARING ECONOMY 1083

According to Airbnb, total listings on the site grew from 120,000 to 300,000 just in 2012.94 TNCs have experienced similarly rapid growth, though some evidence suggests that user signup rates have flattened in primary markets.95 Uber claims that it adds 50,000 new drivers per month.96 These new stakeholders have a strong financial interest in creating rules that allow sharing platforms to continue operating. However, sharing platforms also generate negative externalities for third parties. Short-term rentals can change the character of a building or even a community,97 and many residents object to finding strangers routinely entering the common areas of an apartment complex.98 Likewise, TNC drivers’ accidents can injure not only passengers but also pedestrians and other drivers. TNCs also affect other entities like airports, which strictly control the traffic of for-hire vehicles through licenses.99 Moreover, the success of sharing platforms threatens traditional service providers. Short-term rental services often serve as close substitutes for hotels, particularly low-end hospitality services,100 and TNCs serve as direct substitutes for taxis.101

93. The Shared Economy’s Growth Spurt, CHRISTIAN SCI. MONITOR (Jan. 2, 2014), http://www.csmonitor.com/Commentary/the-monitors-view/2014/0102/The- shared-economy-s-growth-spurt. 94. Sonya Chudgar, Airbnb’s Annual Report: We’re Still Growing, INC. (Feb. 8, 2013), http://www.inc.com/sonya-chudgar/airbnb-annual-report-skirts-legal-issues.html. 95. Chris Nicholson, Study: Uber Pulls Ahead of Lyft in Riders and Revenue With 12x Lead in U.S., FUTURE ADVISOR (Sept. 11, 2014), http://blog.futureadvisor.com/ study-uber-pulls-ahead-of-lyft-in-riders-and-revenue-with-12x-lead-in-u-s/. 96. Douglas MacMillan, Uber’s Army of Drivers Growing 50,000 a Month, WALL ST. J. (Sept. 8, 2014, 4:35 PM), http://blogs.wsj.com/digits/2014/09/08/ubers-army-of- drivers-growing-50000-a-month/. 97. Matt Stevens & Martha Groves, Malibu to Crack Down on Short-Term Rentals Via Airbnb, Other Websites, L.A. TIMES (May 27, 2014, 8:09 PM), http://www.latimes.com/local/la-me-malibu-renting-20140528-story.html (noting that many Malibu residents have complained that short-term rentals change the character of neighborhoods). 98. N.Y. Attorney General Report, supra note 40. 99. Ellen Huet, California Threatens To Shut Down Uber, Lyft, Sidecar Over Airport Rides, FORBES (June 12, 2014), http://www.forbes.com/sites/ ellenhuet/2014/06/12/california-threatens-to-shut-down-uber-lyft-sidecar-over-airport- rides/. The overall impact of TNCs on the environment and traffic congestion remains unclear. 100. Gergios Zervas & Davide Proserpio, The Rise of the Sharing Economy: Estimating the Impact of Airbnb on the Hotel Industry (Bos. Univ. Sch. Mgmt., Working Paper No. 2013-16, 2013), available at http://cs-people.bu.edu/dproserp/papers/ airbnb.pdf. 101. A recent survey by MKM Partners suggests that up to twenty-four percent of Uber users “use it as a replacement for rental companies all of the time.” Jack Hough,

1084 BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 30:385

Regulating the sharing economy therefore requires a balancing of interests. State and local governments must ensure that new rules adequately protect both participants in the sharing economy and third parties. In addition, local governments have a strong public interest in the regulation of sharing platforms. Sharing services can generate substantial tax revenue, but can also disrupt the market for affordable housing102 and create serious consumer safety hazards.103 Many jurisdictions are currently attempting to find regulatory solutions to address the sharing economy.

III. CHALLENGES FOR REGULATORS The following discussion describes how regulators have approached the sharing economy so far, and the enforcement challenges that regulators face under any approach. This Part then explores what makes the sharing economy a regulatory disruption, and argues that existing rules tailored for two-party relationships cannot address the needs of three- party relationships between platforms, providers, and users. Nor do existing models of intermediary liability provide workable solutions for structuring sharing platform liability.

A. REGULATORS HAVE STRUGGLED TO ADAPT EXISTING REGULATIONS TO SHARING PLATFORMS Given that existing regulatory frameworks are highly localized,104 state and municipal legislatures have taken a wide range of approaches in addressing sharing platforms. Sharing platforms often operate legally even without legislative reform, including many per-task contractors, P2P rentals, and some P2P auction houses. Other sharing platforms operate illegally—or arguably illegally105—and regulators have struggled to deal with these platforms. In particular, short-term rentals of residential housing and TNCs frequently violate local ordinances. Some jurisdictions have attempted to ban these platforms, others have granted “experimental”

Uber, Airbnb a Danger to Car Rentals, Hotels?, BARRON’S (Sept. 25, 2014), http://online.barrons.com/articles/SB52784017629588234037504580174550733121636. 102. See, e.g., Jennifer Peltz, NYC Debates: Does Airbnb Help or Hurt Affordable Housing?, ABC NEWS (Jan. 20, 2015, 7:02 PM), http://abcnews.go.com/ Health/wireStory/lawmakers-airbnb-affects-nyc-housing-economy-28339911/. 103. Ben Popken, States Warn of Rideshare Risks for Passengers, NBC NEWS (May 28, 2014, 1:52 PM), http://www.nbcnews.com/business/consumer/states-warn-rideshare- risks-passengers-n116736/. 104. See supra Section II.B. 105. Katie Lobosco, Uber Agrees to Temporarily Suspend Service in Portland, CNN MONEY (Dec. 18, 2014, 7:13 PM), http://money.cnn.com/2014/ 12/18/technology/uber-portland/.

2015] REGULATING THE SHARING ECONOMY 1085 licenses to certain platforms before issuing formal rules, and an increasing number of cities have passed legislation for short-term rentals and TNCs.106 Generally, bans on sharing platforms have met with limited success, and the outcomes under experimental licenses and early legislation are difficult to predict. And under each of these approaches, regulators face similar enforcement challenges. 1. Bans Absent regulatory reform, many sharing platforms violate local ordinances. Short-term residential rentals—often defined as rentals for fewer than thirty days—are explicitly prohibited in many jurisdictions.107 In some cities, certain categories of residential housing may host short- term rentals (e.g., single-family homes) while others may not (e.g., multiple dwelling units or apartment complexes),108 or certain residential

106. The Sharing Economy: Boom and Backlash, ECONOMIST (Apr. 26, 2014), http://www.economist.com/news/business/21601254-consumers-and-investors-are- delighted-startups-offering-spare-rooms-or-rides-across-town/: Corey Owens, head of public policy at Uber, says that in his industry local authorities can be put into three “buckets”. In the first are those with strict rules that they intend to keep, such as Austin, London, New York and Philadelphia. In the second are places where the future is ambiguous: here he puts Baltimore, Brussels, Paris and Washington, DC. Authorities in the third bucket have recognised [sic] that the world is changing. California, where the taxi regulator adopted new rules for ride-sharers last year, is the “primary example.” As noted in this Section, many of these cities have moved into different “buckets” since Owens made this statement. 107. For example, until October 2014, San Francisco prohibited rentals for less than thirty days. The city technically allowed residents to obtain a special permit for short- term rentals, but the expense and difficulty of obtaining such a permit made this option impractical for the average host. Carolyn Said, S.F. Cracks Down on Airbnb Rentals, SFGATE (Apr. 7, 2014, 2:26 PM), http://www.sfgate.com/bayarea/article/S-F-cracks- down-on-Airbnb-rentals-5381237.php; Doug Sovern, SF Law Prohibits Short-Term Rentals Popularized By Airbnb, CBS SF BAY AREA (Aug. 4, 2011, 9:19 AM), http://sanfrancisco.cbslocal.com/2011/08/04/sf-law-prohibits-short-term-rentals- popularized-by-airbnb/. 108. The New York Multiple Dwelling code provides that certain categories of “multiple dwelling” units “shall only be used for permanent residence purposes.” The statute defines the relevant category of “class A” multiple dwelling units to include “tenements, flat houses, maisonette apartments, apartment houses, apartment hotels, bachelor apartments, studio apartments, duplex apartments, kitchenette apartments, garden-type maisonette dwelling projects, and all other multiple dwellings except class B multiple dwellings.” N.Y. Mult. Dwell. Law § 4(8)(a) (McKinney 2011). The statute further defines residential use as “occupancy of a dwelling unit by the same natural person or family for thirty consecutive days or more . . . .” Id. State Senator Liz Krueger has clarified that the statute does not prohibit multiple dwelling residents from renting single

1086 BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 30:385 zones may grant limited licenses while other residential zones do not.109 Similarly, TNCs arguably fall under the definition of for-hire vehicles under many local taxi ordinances.110 Since TNCs generally do not (and often cannot feasibly) comply with taxi ordinances, some courts have ruled that TNCs are illegal taxi operations.111 P2P auction houses for parking spaces may also violate local ordinances.112 Rather than revising local codes to accommodate sharing platforms, some jurisdictions chose to enforce existing laws to prohibit these P2P services. Many cities and states responded to TNCs by issuing cease-and- desist orders, fining platforms, and seeking injunctions.113 Several cities have also engaged in aggressive efforts to shut down short-term rental hosts,114 and some have revised local ordinances to expressly prohibit particular categories of short-term residential rentals.115 New York City and San Francisco have launched highly visible campaigns against illegal

rooms for less than thirty days, so long as a permanent resident also occupies the unit for the full duration of the short-term rental. Liz Krueger, Answers for New Yorkers Concerned or Confused About the Illegal Hotel Law, N.Y. SENATE (May 27, 2014), http://www.nysenate.gov/report/answers-new-yorkers-concerned-or-confused-about- illegal-hotel-law/. 109. Bianca Barragan, The Few Places in Los Angeles Where Airbnbs Might Be Legal, CURBED L.A. (Mar. 24, 2014), http://la.curbed.com/archives/2014/03/ the_few_places_in_los_angeles_where_airbnbs_might_be_legal.php. 110. Broward County in Florida has recently clarified that TNC drivers count as taxis under local ordinances. Brittany Wallman, Uber Drivers Cited 250 Times, Fined $35,000, SUN SENTINEL (Jan. 27, 2015, 4:47 PM), http://www.sun-sentinel.com/local/broward/ fl-uber-showdown-broward-20150127-story.html. 111. E.g., Eric M. Johnson, Ridesharing Firm Uber Suspends Operations in Nevada, REUTERS (Nov. 27, 2014), http://www.reuters.com/article/2014/11/27/us-usa-nevada- ridesharing-idUSKCN0JB18J20141127. 112. Laura Entis, San Francisco Says Enough Monkey Business: Tells Parking Spot App to Shut Down, ENTREPRENEUR (July 11, 2014), http://www.entrepreneur.com/article/235575/. 113. E.g., Sean Doogan, Judge Sides with City in Decision Allowing Uber to Offer Only Free Ride Sharing, ALASKA DISPATCH NEWS (Oct. 13, 2014), http://www.adn.com/article/20141013/judge-sides-city-decision-allowing-uber-offer- only-free-ride-sharing/; Josephine Woolington, City Threatens to Fine Uber Drivers, REGISTER GUARD (Jan. 17, 2015), http://registerguard.com/rg/news/local/32661556- 75/eugene-warns-uber-drivers-of-potential-fines.html.csp. 114. David Streitfeld, Airbnb Listings Mostly Illegal, New York State Contends, N.Y. TIMES (Oct. 15, 2014), http://www.nytimes.com/2014/10/16/business/airbnb-listings- mostly-illegal-state-contends.html. 115. Katherine Sayre, New Orleans City Council Toughens Ban on Unlicensed Short- Term Vacation Rentals, NOLA (July 10, 2014, 12:23 AM), http://www.nola.com/business/index.ssf/2014/07/new_orleans_city_council_tough.html.

2015] REGULATING THE SHARING ECONOMY 1087 hotel operators.116 In addition, the San Francisco Attorney General issued multiple cease-and-desist orders to P2P auction houses for parking spots.117 These efforts have met with limited success. Short-term rental hosts continue to operate illegally in many cities. Some jurisdictions like New Orleans simply have not acted to enforce formal bans on short-term rentals.118 Even in New York City, enforcement campaigns against short- term rentals have encountered serious obstacles. Likewise, TNCs have often ignored cease-and-desist orders.119 In many cases, TNCs have negotiated with cities to gain temporary authorization rather than discontinuing operations, and in most cases TNCs have obtained temporary licenses.120 2. Legalizing Sharing Platforms: Revising Existing Rules or Creating New Rules Increasingly, states and cities have attempted to legalize or partially legalize TNCs and short-term rental platforms. Most jurisdictions have approached short-term rentals by revising existing ordinances, though some have also created new permitting regimes. Short-term rental provisions vary widely. Common restrictions include geographic caps,

116. Streitfeld, supra note 114; Katy Steinmetz, San Francisco Cracks Down on Airbnb ‘Abuses’, TIME (Apr. 15, 2014), http://time.com/63810/san-francisco-airbnb-crackdown/ (discussing efforts to pass an ordinance clarifying the legality of some short-term rentals, and to encourage enforcement efforts against illegal rentals); Guy Marzorati, Enforcement Issues Remain as San Francisco’s Airbnb Law Takes Effect, KQED NEWS (Feb. 2, 2015), http://ww2.kqed.org/news/2015/02/02/Enforcement-Issues-San-Francisco-Legalize- Airbnb-Rentals (discussing opposition to the short-term rental ordinance, and enforcement challenges related to the ordinance). 117. Entis, supra note 112. 118. Robert McClendon, In a Push for Short-Term Rental Legalization, Some New Orleans Owners Ask ‘Why Bother?’, NOLA (Nov. 6, 2014, 7:38 AM), http://www.nola.com/politics/index.ssf/2014/11/new_orleans_short-term_rental.html. 119. Jeremy Allen, Uber, Lyft ignore Ann Arbor’s cease and desist demand; no tickets issued following city’s orders, MLIVE.COM (July 7, 2014), http://www.mlive.com/ business/ann-arbor/index.ssf/2014/07/uber_lyft_ignore_ann_arbors_ce.html; Ashley Dejean, Uber And Lyft Ignore Virginia’s Demand To Cease And Desist, WAMU 88.5 AM. UNIV. RADIO (June 6, 2014), http://wamu.org/news/14/06/06/uber_and_ lyft_refuse_virginias_demand_to_cease_and_desist; Amos Maki, Uber, Lyft Operating Despite Cease-and-Desist Notices, MEMPHIS DAILY NEWS (July 17, 2014) http://www.memphisdailynews.com/news/2014/jul/17/uber-lyft-operating-despite-city- notices/. 120. See text accompanying note 90; see also Mike Isaac, Uber Suspends Portland Operations in Deal with City Regulators, N.Y. TIMES (Dec. 18, 2014), http://bits.blogs.nytimes.com/2014/12/18/uber-suspends-portland-operations-in-deal-with- city-regulators/.

1088 BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 30:385 maximum durations, maximum rentals per year, occupancy limits, and exclusions for multiple dwelling residences.121 Some jurisdictions have created affirmative duties for short-term rental hosts, such as ensuring adequate parking for guests, providing notice of short-term rentals to neighbors, or liability for nuisances caused by guests.122 In jurisdictions with permitting regimes, hosts typically pay an application fee and undergo a safety inspection.123 Many jurisdictions have also decided that transient occupancy taxes apply to short-term rental hosts.124 By contrast, most cities and states have approached TNCs by creating a new legal category distinct from for-hire vehicles like taxis and limousines. In some states, the public utilities commission has issued early guidance for TNCs.125 Other states and cities have already passed legislation governing TNCs, including Colorado, the city of Chicago, and the District of Columbia.126 Some common features of early rules for TNCs include: minimum insurance requirements, regular inspections, driver background checks, duties not to discriminate against disabled riders, vehicle registration and licensing, requirements for distinctive trade dress to mark registered vehicles, and limitations on surge pricing—a practice of increasing fares during peak hours to encourage drivers to work during those hours.127 In addition, most states and cities have expressly excluded TNCs from common carrier status, relieving TNCs of the duty of non-discrimination.128 Many new rules for TNCs will not take effect until early 2015 or later, and it is too early to judge the success of these regulatory efforts.

121. Gottlieb, supra note 76. 122. Id. 123. See Vacation Rental Licensing, AUSTIN, http://www.austintexas.gov/str/ (last visited Feb. 26, 2015). 124. See, e.g., Portland, Or., Planning and Zoning § 33.207.070. 125. Cal. P.U.C. Decision, supra note 18, at 72–75. 126. COLO. REV. STAT. §§ 40-10.1-601–40-10.1-608 (2014); CHI., ILL., CODE § 9-115-100 (2014); Vehicle-for-Hire Innovation Amendment Act of 2014, 61 D.C. Reg. 012430 (Dec. 5, 2014). 127. Douglas MacMillan, Uber Laws: A Primer on Ridesharing Regulations, WALL ST. J. (Jan. 29, 2015, 10:37 PM), http://blogs.wsj.com/digits/2015/01/29/ uber-laws-a-primer-on-ridesharing-regulations/. 128. E.g., COLO REV. STAT. § 40-10.1-603 (“A transportation network company is not subject to the [public utility] commission’s rate, entry, operational, or common carrier requirements . . . .”), available at http://cdn.colorado.gov/ cs/Satellite/DORA-PUC/CBON/DORA/1251655091163. Maryland is a notable exception, and has ruled that TNCs are common carriers. Md. P.U.C. Order No. 86528 (Aug. 6, 2014).

2015] REGULATING THE SHARING ECONOMY 1089

Moreover, some jurisdictions have issued experimental licenses to TNCs, but have yet to issue formal rules.129 3. Enforcement Whether state and local governments choose to ban or legalize sharing platforms, regulatory agencies must develop strategies for enforcing new or existing laws. Policing sharing platforms often proves more difficult than enacting legislation. Regulators have two main options for enforcement: target providers or target platforms. In some cases, private causes of action may also contribute to enforcement strategies. a) Targets: Providers Versus Platforms To target providers, regulators must decide how to identify bad actors and what remedy to seek. Regulators have identified illegal conduct by soliciting user information from platforms,130 through inspections or policing efforts,131 and through complaint hotlines.132 After regulators uncover illegal conduct, they impose fines133 or they seek injunctive relief to either disband commercial actors or remove accounts from platforms.134

129. See test accompanying note 90; see also Mike Isaac, Uber Suspends Portland Operations in Deal with City Regulators, N.Y. TIMES (Dec. 18, 2014), http://bits.blogs.nytimes.com/2014/12/18/uber-suspends-portland-operations-in-deal-with- city-regulators/. 130. See Airbnb, Inc. v. Schneiderman, 989 N.Y.S.2d 786 (N.Y. Sup. Ct. 2014) (quashing as overbroad a subpoena from the New York Attorney General requiring that Airbnb provide personal information of hosts relevant to identifying illegal listings); Columbus v. Lyft Inc., No. 2014 EVH 060145, 2014 WL 4805037 (Sept. 19, 2014) (granting in part and denying in part a public records request from the City of Columbus seeking Lyft’s records of drivers’ names and personal information). 131. E.g., Marc D. Anderson, Fairhope to start cracking down on short-term vacation rentals, AL.COM (Aug. 11, 2014) http://www.al.com/news/mobile/ index.ssf/2014/08/fairhope_to_start_cracking_dow.html. 132. E.g., Marc Saltzberg, VENICE NEIGHBORHOOD COUNCIL, SHORT TERM VACATION RENTAL REPORT (2013), available at http://www.venicenc.org/wp -content/uploads/2012/05/Short-Term-Vacation-Rental-Report.pdf (“The City's Department of Building and Safety (DBS) is charged with enforcing the ordinances which ban vacation rentals. DBS is complaint driven - that is, they don't enforce unless there is a complaint.”). 133. See Ron Lieber, A $2,400 Fine for an Airbnb Host, N.Y. TIMES (May 21, 2013), http://bucks.blogs.nytimes.com/2013/05/21/a-2400-fine-for-an-airbnb-host/ (reporting on Nigel Warren, an Airbnb user in New York who received $2,400 in fines, and questioning whether Airbnb has a duty to warn or block users in areas where Airbnb hosting is illegal). 134. San Francisco City Attorney Dennis Herrera filed suit in April 2014 against two landlords for evicting tenants under the Ellis Act to convert apartments into short-term rental listings. Bob Egelko, S.F. city attorney sues 2 landlords over short-term rentals,

1090 BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 30:385

Many cities have addressed short-term rental platforms by policing providers. The New York Attorney General’s office has aggressively pursued hosts with illegal listings.135 In August, the Attorney General requested that Airbnb reveal the identities of 124 accounts.136 San Francisco has similarly pursued actions against landlords who attempted to convert apartments into vacation rental units by evicting long-term tenants.137 But shutting down individual hosts requires immense resources, and new offenders can enter the market with ease. On the other hand, targeting platforms presents less of a logistical challenge. Regulators have generally targeted TNCs, car sharing services, and auction houses at the platform level. Because it is not always clear which local or state agency has the authority to regulate platforms, regulators have sometimes faced an initial hurdle when establishing the correct channel for enforcement.138 Regulators can impose fines for specific instances of noncompliance with existing statutes.139 However, as discussed above, public enforcement efforts have limited traction because of limited government resources, and because platforms often continue to operate in defiance of cease-and-desist orders or repeated fines. b) Public Versus Private Enforcement The previous discussion has assumed that regulatory agencies will primarily enforce legislation governing sharing platforms. The focus on

SFGATE (Apr. 24, 2014, 6:52 AM), http://www.sfgate.com/bayarea/article/S-F-city- attorney-sues-2-landlords-over-5425826.php. 135. Polly Mosendz, Face-Off: NYC Lawmakers Grill Airbnb on Illegal Hotels, NEWSWEEK (Jan. 21, 2015, 1:26 PM), http://www.newsweek.com/face-nyc-lawmakers- grill-airbnb-illegal-hotels-301060. 136. Tom Huddleston, Jr., Airbnb to reveal names of 124 hosts to New York’s Attorney General, FORTUNE (Aug. 22, 2014), http://fortune.com/2014/08/22/ airbnb-to-reveal-124-hosts-identities-to-new-yorks-attorney-general/. 137. Bob Egelko, S.F. City Attorney Sues 2 Landlords Over Short-Term Rentals, SFGATE (Apr. 24, 2014, 6:52 AM), http://www.sfgate.com/bayarea/article/ SFcityattorneysues2landlordsover5425826.php. 138. During the administrative review period in 2013, Uber challenged the California Public Utilities Commission’s authority over the service “because it is a technology company that does not provide transportation services.” Patrick Hoge, Uber, taxis appeal PUC ruling on private drivers for hire, S.F. BUS. TIMES (Oct. 24, 2013, 9:53 AM), http://www.bizjournals.com/sanfrancisco/ blog/2013/10/uber-lyft-sidecar-taxicab-puc.html. 139. The Pennsylvania Public Utility Commission first conducted undercover investigations of Uber and Lyft’s services, then threatened the TNCs with fines— $130,000 for Lyft and $95,000 for Uber. Yazhou Sun, Uber and Lyft Face First-Ever Fines in the United States, ABC NEWS (June 11, 2014), http://abcnews.go.com/US/uber- lyft-face-fines-united-states/story?id=24089617.

2015] REGULATING THE SHARING ECONOMY 1091 action by regulatory agencies stems in part from the strong public interest involved in some sharing platforms—particularly TNCs, short-term home and car rentals, P2P lending platforms, and P2P auction houses. Yet in some cases private causes of action offer a viable alternative to regulatory action. In the context of short-term rental platforms, landlords may constitute the first line of enforcement against hosts. Most landlords include a sublet clause in both commercial and residential leases.140 A landlord can succeed in a summary eviction proceeding by demonstrating that the tenant illegally hosted short-term guests in violation of the lease.141 Drivers, competitors, and riders have also filed private suits against TNCs. Drivers have claimed that Uber and Lyft misclassify drivers as independent contractors.142 Riders, on the other hand, have filed both personal injury suits against TNCs143 and class actions based on violations of the Americans with Disabilities Act (“ADA”)144 and unfair business

140. E.g., Ga. Dep’t of Cmty. Affairs, Basic Rental Agreement or Residential Lease, available at http://www.dca.ga.gov/housing/specialneeds/programs/documents/C- 2SampleLEASE.pdf; Coy Davidson, The Importance of Assignment and Subletting Clauses in Commercial Leases, TENANT ADVISOR (Sept. 6, 2009), http://www.coydavidson.com/leasing-tips/the-importance-of-assignment-and-subletting- clauses-in-commercial-leases/. 141. Kevin Davis, Guest Wrong, 100 A.B.A.J. 19, 19–20 (2014). 142. See Cotter v. Lyft, Inc., No. 13-cv-04065-VC, 2014 WL 3884416 (N.D. Cal. Aug. 7, 2014). The IRS, state tax bureaus, and other state agencies each apply similar but varied tests for determining whether a worker qualifies as an employee rather than an independent contractor. These tests generally focus on the degree of control the “employer” exercises over the conduct of the employee. In fact, most states allow taxi dispatch services to hire drivers as independent contractors, depending on the nature of the contractual relationship. UNREGULATED WORK IN THE TAXI INDUSTRY IN NEW YORK CITY, NELP.ORG (2007), available at http://www.nelp.org/page//EJP/ Unregulated_Work_Taxi.pdf; INFORMATION SHEET: TAXICAB INDUSTRY, EMP’T DEV. DEP’T OF CAL. (2009), available at http://www.edd.ca.gov/pdf_pub_ctr/ de231tc.pdf; Industry Specific Help for Worker Classification, OREGON.GOV, http://www.oregon.gov/IC/pages/07-industries.aspx#taxicab (noting that some dispatch services hire drivers as independent contractors, and these services must meet state criteria including “(1) that [drivers] be free from the right of another to control how services are performed and, (2) that the worker operates an independently established business”). Taxicab drivers have filed claims against dispatchers for misclassification of drivers as independent contractors in numerous states, with mixed results. 143. See infra Section III.B; see also Mazaheri v. Doe, No. CIV-14-225-M, 2014 WL 2155049 (W.D. Okla. May 22, 2014). 144. McPhail v. Lyft, Inc., No. A-14-CA-829-LY, 2014 WL 6680393 (W.D. Tex. Nov. 25, 2014); Salovitz v. Uber Techs., Inc., No. A-14-CV-823-LY, 2014 WL 5318031 (W.D. Tex. Oct. 16, 2014).

1092 BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 30:385 practices.145 Likewise, taxi dispatch services in Boston and Chicago have pursued claims against Uber under the Lanham Act and state unfair competition laws.146 Private actions against sharing platforms can supplement government enforcement efforts. These actions may also help to shape the ongoing legislative efforts to address sharing platforms. Although state and local governments have taken action to regulate the sharing economy, many problems remain unresolved. Even sharing platforms that already operate legally may raise novel insurance or employment issues warranting regulatory attention. The following Section explores why sharing platforms disrupt existing regulatory frameworks.

B. LEGAL GRAY AREAS: SHARING PLATFORMS DISRUPT TRADITIONAL LEGAL CATEGORIES Sharing platforms often operate in legal gray areas, and the regulatory vacuum surrounding these services has raised complex legal questions.147 Sharing companies erode the distinction between commercial and personal, public and private, and the definition of goods and services.148 For example, TNC drivers operate personal vehicles for a commercial purpose. Similarly, short-term rental hosts temporarily offer private, residential housing to the public. P2P auction houses and per-task contractor services allow users to purchase temporary licenses, such as a position in line for a movie, a table in a coffee shop, or a public parking spot. Blurring these lines creates difficult legal issues because these categories affect determinations of liability in almost every field of law, from tax149 to civil rights.150

145. Ehret v. Uber Techs., Inc., No. C-14-0113 EMC, 2014 WL 4640170 (N.D. Cal. Sept. 17, 2014). 146. E.g., Bos. Cab Dispatch, Inc. v. Uber Techs., Inc., No. 13-10769-NMG, 2014 WL 1338148 (D. Mass. Mar. 27, 2014); Yellow Grp. LLC v. Uber Techs. Inc., No. 12 C 7967, 2014 WL 3396055, at *5 (E.D. Ill. July 10, 2014). 147. Cohen & Zehngebot, supra note 10 (“To grossly generalize, the law tends to prefer binary divisions: public and private, business and personal, donation and sale, consumer and provider, and, most saliently, my property and yours. In the sharing economy, many companies blur these boundaries, resulting in a legal gray area.”). 148. See Badger, supra note 52 (arguing that the real “tough call” for regulators is how to regulate “quasi-professionals,” providers who fall in between traditional regulatory categories). 149. See Thomas A. Dickerson & Sylvia O. Hinds-Radix, Taxing Internet Transactions: Airbnb and the Sharing Economy, 86 N.Y. ST. B.J. 49 (2014) (“While Expedia, Priceline and Hotwire are best defined as retailers or resellers and, as such, can be controlled and taxed accordingly, it is much more difficult to find a comparable taxing analogue for the Internet-sharing economy.”).

2015] REGULATING THE SHARING ECONOMY 1093

1. Insurance Insurance liability for TNCs illustrates the gray area between personal and commercial activity. Most personal insurance plans do not cover commercial activity, such as charging fares for rides. Thus TNC drivers cannot rely on personal insurance policies for accidents that occur while they have the TNC app on.151 Some states now require TNCs to purchase insurance to cover this “gap” in the TNC driver’s personal insurance policy.152 Insurers break down TNC liability into three “periods.” The first period starts when the TNC driver turns on the app and ends when the TNC driver accepts a match. The second period covers the time between the match and when the driver picks up the rider. The third is the time between rider pick up and drop off.153 Many jurisdictions have adopted an “app on” rule, where TNCs must provide insurance to drivers from the moment the driver turns on the TNC app.154 Although TNCs have not always provided primary coverage during periods two and three,155 most TNCs now voluntarily purchase insurance

150. For example, federal civil rights statutes prohibit discrimination based on gender in “places of accommodation,” and in advertisements of housing rentals. Title II of the Civil Rights Act, 42 U.S.C. § 2000a (2012); Fair Housing Act, 42 U.S.C § 3604 (2012). The Fair Housing Act, however, makes an exception for advertisements expressing a preference for a particular gender. Memorandum from Roberta Acthenberg, Assistant Sec’y for Fair Housing and Equal Opportunity, U.S. Dep’t Hous. & Urb. Dev. 2 (Jan. 9, 1995), available at http://www.hud.gov/offices/fheo/disabilities/sect804achtenberg.pdf. Whether short-term rental hosts can take advantage of this exception is unclear. 151. See, e.g., Ron Lieber, The Question of Coverage for Ride Service Drivers, N.Y. TIMES (Sept. 5, 2014), http://www.nytimes.com/2014/09/06/your-money/auto- insurance/offloading-the-risk-in-renting-a-car-ride.html. 152. E.g., CAL. PUB. UTIL. CODE § 5433(c) (West 2015). California state assemblywoman Susan Bonilla introduced the current insurance mandate for TNCs, bill number AB 2293, in response to Sofia Liu’s death. Ron Lieber, The Question of Coverage for Ride Service Drivers, N.Y. TIMES (Sept. 5, 2014), http://www.nytimes.com/2014/09/06/your-money/auto-insurance/offloading-the-risk- in-renting-a-car-ride.html; Josh Eidelson, Is California About to Disrupt Uber?, BLOOMBERG BUSINESSWEEK (June 26, 2014), http://www.businessweek.com/ articles/2014-06-26/is-california-about-to-disrupt-uber. 153. See Letter from Dave Jones, Ins. Comm’r, Cal. Dep’t Ins., to Michael R. Peevey, President, Cal. Pub. Utils. Comm’n 2 (Apr. 7, 2014), available at http://www.insurance.ca.gov/0400-news/multimedia/0030VideoHearings/upload/CDI- CPUC20140407.pdf. The California insurance mandate for TNCs instead divides TNC liability into two periods. CAL. PUB. UTIL. CODE § 5433. 154. Don Jergler, Uber, Lyft, Sidecar Toe-to-Toe With Insurers State-by-State, INS. JOURNAL (June 27, 2014), http://www.insurancejournal.com/news/national/ 2014/06/27/332942.htm. 155. Prior to 2014, Lyft, Uber, and Sidecar only purchased backup policies in case a driver’s personal insurance declined to cover an accident. Most personal insurance

1094 BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 30:385 plans for these periods without a statutorily imposed insurance mandate.156 However, TNCs have vigorously contested statutory insurance requirements for the second period. TNCs have argued that providing commercial insurance during the second period would be economically wasteful and incentivize drivers to commit insurance fraud. Commercial coverage is more expensive than personal insurance coverage because of the increased liability risks involved. However, drivers have an incentive to leave the TNC app on while doing routine chores or other personal trips to benefit from higher coverage.157 Regulators have argued that TNCs benefit from drivers who leave the app on even when those drivers do not currently have a passenger request, because the driver is then “available” to prospective passengers.158

policies, however, have an exception for commercial ride sharing. By requiring drivers to first report claims to their personal insurers, this TNC policy created a perverse incentive for drivers to lie to their insurance providers. See Lieber, supra note 151. 156. See Jergler, supra note 154. 157. Sidecar and other TNCs argued during the California Department of Insurance investigatory hearing that mandating full coverage during period one creates a moral hazard for drivers, who may be tempted to leave the TNC app on to take advantage of the higher insurance coverage. See Don Jergler, TNCs, Insurers Square off at California Hearing, INS. JOURNAL (Mar. 21, 2014), http://www.insurancejournal.com/ news/west/2014/03/21/324004.htm. TNC driver support forums suggest that this may be a realistic concern; some drivers may literally leave the app on twenty-four hours a day. Uber Jax, Just Curious as to who leaves their App on 24/7, UBERPEOPLE.NET (Oct. 3, 2014), http://uberpeople.net/threads/just-curious-as-to-who-leaves-their-app-on-24- 7.4537/ (“I never turn my driver App off! I leave it on 24/7. . . . I just go about my daily life and when I get a ping I drop what I’m doing and take care of business.”). Insurance Commissioner Jones aptly summarized the “moral hazard” argument of TNCs: “(1) drivers may be running personal errands; (2) drivers may have multiple applications open at the same time; (3) drivers with low limits on their personal automobile insurance policy will turn on the application in the event of an accident to secure more robust coverage; and (4) drivers start to look more like employees or independent contractors if the TNC covers this period.” Letter from Dave Jones, Ins. Comm’r, Cal. Dep’t Ins., to Michael R. Peevey, President, Cal. Pub. Utils. Comm’n 2 (Apr. 7, 2014), available at http://www.insurance.ca.gov/0400-news/multimedia/0030VideoHearings/ upload/CDI-CPUC20140407.pdf. 158. California Insurance Commissioner Dave Jones argued in his letter to California Public Utilities Commission: “Even if a driver is running errands during Period 1, if that driver has the app open, the TNC benefits from the driver showing availability to provide rides to customers.” Letter from Dave Jones, Ins. Comm’r, Cal. Dep’t Ins., to Michael R. Peevey, President, Cal. Pub. Utils. Comm’n 2 (Apr. 7, 2014), available at http://www.insurance.ca.gov/0400news/ multimedia/0030VideoHearings/upload/CDI-CPUC20140407.pdf.

2015] REGULATING THE SHARING ECONOMY 1095

2. Tax Blurring the line between commercial and personal activity complicates many legal areas beyond insurance, such as whether a TNC driver may deduct the cost of car repairs as a business expense for tax purposes. The IRS permits deductions based on the percentage of a driver’s yearly mileage attributable to “business” use.159 This seems to cover the second and third insurance periods for TNCs, but it is unclear whether the app-on rule for insurance liability also establishes that driving during the first insurance period is a “business” purpose when calculating income tax deductions. 3. Lessor Liability P2P rental platforms may also complicate the distinction between high-volume and casual commercial activity. For example, in many jurisdictions, product liability depends on the commercial character of the seller.160 Some jurisdictions similarly create an exception to limit “casual lessor” liability for manufacturing defects.161 This would seem to encompass services like Equipify or Spinlister, which allow providers to rent their personal sporting equipment to users, including skis and mountain bikes.162 Yet providers may transact at vastly different volumes on these platforms.163 It is not yet clear at what point a casual lessor of snowboards on a sharing platform becomes a commercial lessor in various

159. Travel, Entertainment, Gift, and Car Expenses: For Use in Preparing 2013 Returns, IRS 16 (2014), available at http://www.irs.gov/pub/irs-pdf/p463.pdf. 160. Both the second and third restatements limit liability for “occasional” or “noncommercial” sellers. RESTATEMENT (SECOND) OF TORTS § 420A cmt. f (1965); RESTATEMENT (THIRD) OF TORTS: PROD. LIAB. § 1 cmt. c (1998). 161. E.g., 1 MICHAEL WEINBERGER, NEW YORK PRODUCTS LIABILITY § 5:7 (2d ed. 2014) (“[A] commercial lessor of a product will be held strictly liable for personal injuries, just as a product manufacturer would . . . . However, casual or occasional lessors of products are not subject to claims of strict liability. At most, the duty of a casual or occasional lessor is to warn the person to whom the product is supplied of known defects that are not obvious or readily discernible.”) 162. EQUIPIFY, http://equipify.me/; SPINLISTER, https://www.spinlister.com/. Similar concerns may also arise for casual rentals of office space, in the context of premises liability. Services like Peerspace allow business to occasionally rent out office space; these social work environments are popular with freelancers and remote workers, and even small businesses hosting occasional retreats. PEERSPACE, http://www.peerspace.com/. The duty of care for hosts on these platforms may vary by jurisdiction. See generally, e.g., Rowland v. Christian, 69 Cal. 2d 108 (1968) (abandoning the categories of trespassers, licensees, and invitees in favor of a reasonableness standard). PeerSpace’s terms of service indicate that hosts and guests are expected to provide their own insurance. Terms of Use, PEERSPACE, http://www.peerspace.com/terms/. 163. See infra Part I.B.

1096 BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 30:385 jurisdictions. These platforms often provide form contracts to providers and users, and in some cases require the parties to sign a contract before processing a transaction.164 However, contracting does not completely resolve the issue as jurisdictions vary widely in their enforcement of exculpatory clauses.165 4. Civil Rights Statutes Likewise, TNC platforms’ process for matching riders and drivers may establish duties under civil rights statutes. When drivers have the TNC app on, they count as “available” to all users for the purpose of calculating wait-time and surge pricing rates—a dynamic pricing scheme that increases fares to encourage drivers to work during peak hours.166 Once a user requests a ride, however, TNC drivers may decline to accept the match; in making this decision, the driver typically has access to the user’s profile, including her name and potentially her photograph.167 Given this mechanism, it is unclear whether the “public” nature of TNCs subjects them to obligations under civil rights statutes, including the ADA.168 Recent lawsuits have raised the issue of Uber’s obligations under the ADA. For example, the National Federation of the Blind alleges that drivers discriminate against blind riders by refusing service to riders with service animals or mishandling those animals.169 Two other class action

164. Spinlister requires that renters and listers sign a form contract. Ride Rental Agreement, SPINLISTER, available at https://www.spinlister.com/safety/. 165. See C. Connor Crook, Validity and Enforceability of Liability Waiver on Ski Lift Tickets, 28 CAMPBELL L. REV. 107, 120–21 (2005) (“Courts are generally reluctant to enforce exculpatory clauses, especially those that include the negligence of the party attempting to enforce the clause. However, . . . courts can take very nuanced approaches . . . .”). Spinlister includes a sweeping exculpatory clause in its rental agreement. Ride Rental Agreement, supra note 164, at § 3.2 (including both “as is” clause and an assumption of risk clause—“EVEN IF SUCH LOSS OR DAMAGE IS DUE TO ANY NEGLIGENCE OF LISTER, SPINLISTER, [or] THEIR AGENTS”). 166. What Is Surge Pricing and How Does It Work?, UBER, https://support. uber.com/hc/en-us/articles/201836656-What-is-surge-pricing-and-how-does-it-work-/. 167. Lyft allows drivers to see a photograph of potential riders before accepting a request, while Uber by design does not. Jenna Wortham, Ubering While Black, MATTER (Oct. 23, 2014), https://medium.com/matter/ubering-while-black-146db581b9db/ (discussing the concern that TNC drivers racially discriminate against riders based on profile pictures). 168. 42 U.S.C. §§ 12131–65 (2012); AM. ASS’N OF PEOPLE WITH DISABILITIES, EQUITY IN TRANSPORTATION FOR PEOPLE WITH DISABILITIES (2012), available at http://www.aapd.com/resources/publications/transportation-disabilities.pdf. 169. Complaint, Nat’l Fed. Blind Cal. v. Uber Techs., Inc., No. 3:14-cv-04086 (N.D. Cal. Sept. 9, 2014), available at http://dralegal.org/sites/dralegal.org/ files/content/0001_complaint_accessible.pdf.

2015] REGULATING THE SHARING ECONOMY 1097 suits have claimed that Uber discriminates against disabled riders by failing to provide sufficient numbers of accessible vehicles.170 Uber has responded by issuing statements confirming its policy of deactivating drivers who discriminate, but denying any legal duty to do so under the ADA.171 Some public utilities commissions have expressly imposed a duty not to discriminate on TNCs, but these mandates do not specify exact benchmarks for compliance.172 Many states have not yet clarified a TNC’s duty to monitor driver behavior, nor do they impose guidelines for what percentage of a TNC’s drivers must provide wheelchair accessible vehicles.173 5. Public Performance and Social Host Liability for Short-Term Rentals The distinction between public and private activity plays a significant role in many other areas of the law. For example, providing video rental services in a hotel room qualifies as a “public performance” under copyright infringement law.174 If a short-term rental host charges guests an extra fee for the use of her home theater and DVD collection, does this also count as a public performance? In some cases, the private nature of the short-term rental might also burden the host with liability. Should a host be liable under a social host ordinance if an underage guest drinks a bottle of wine from her cabinet without permission?175 In this case and

170. McPhail v. Lyft, Inc., No. A-14-CA-829-LY, 2014 WL 6680393 (W.D. Tex. No. 25, 2014); Salovitz v. Uber Techs., Inc., No. A-14-CV-823-LY, 2014 WL 5318031 (W.D. Tex. Oct. 16, 2014). 171. Bob Egelko, Blind Riders Sue Uber, Allege Discrimination, SFGATE (Sept. 11, 2014, 8:12 AM), http://www.sfgate.com/bayarea/article/Uber-accused-in-suit-of- refusing-customers-with-5746412.php. 172. The Colorado statute prohibits discrimination against protected categories such as disability. It provides that if a driver is unable to transport a disabled passenger, “the driver shall refer the rider to another driver or transportation service provider” with an accessible vehicle. It does not, however, require TNCs to supply any number of wheelchair-accessible vehicles. Nor are TNCs liable for driver discrimination unless the TNC has received a complaint in writing and “failed to reasonably address the alleged violation.” COLO. REV. STAT. § 40-10.1-605 (2014). 173. See Cal. P.U.C. Decision, supra note 18, at 72–75. 174. See On Command Video Corp. v. Columbia Pictures Indus., 777 F. Supp. 787 (N.D. Cal. 1991) (“Hotel guests watching a video movie in their room through On Command’s system are not watching it in a ‘public place’ but they are nonetheless members of ‘the public.’”). 175. See SOCIAL HOST ORDINANCES IN WISCONSIN, WISCONSIN ALCOHOL POLICY PROJECT 1 (2012) (“Effective social host ordinances cover a wide range of locations and situations, noting the adult ‘hosts’ must take reasonable steps to prevent

1098 BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 30:385 many others, operating in a gray area not only creates confusion for regulators, but exposes providers on sharing platforms to liability on multiple fronts. 6. Sales of Temporary Licenses While many sharing companies like TNCs simply provide an old service through a new channel,176 some sharing companies challenge the definitions of “goods” and “services.” In particular, some sharing companies create new marketplaces for temporary licenses; these platforms act as auction houses for a temporary right to occupy a position. The two primary examples are auction platforms for public parking spaces and for seating in coffee houses or restaurants.177 Neither sellers nor purchasers “own” or “provide” the temporary license; the municipal government is responsible for maintaining public parking and restaurants are responsible for providing seating to customers. Many have criticized these platforms as parasitic, and claim that these services interfere with third-party rights. Bloggers have even coined the term “jerk tech” to describe these platforms.178 Others contend that auction houses optimize efficient use of scarce resources and prevent congestion; users who value a parking spot most can pay for that privilege, and price-sensitive users will accept less conveniently located spots.179 Because markets for temporary licenses can

underage drinking. Adults who allow or facilitate an underage drinking party may be ticketed even when they are not present during the party.”). 176. In some cases, sharing companies offer a package of products and services. For example, dining-experience sites like Feastly allow users to contract with chefs for single meals at the user or chef’s home. The chef provides a meal (a product) but the user also pays for the overall experience (a service). Yet the same reasoning applies to restaurants, and the line between products and services has always been blurry in some sectors. 177. SWEETCH, http://www.getsweetch.com/ (an auction house for parking spots); RESERVATION HOP, https://reservationhop.com/ (an auction house for restaurant reservations); BETRSPOT; http://www.betrspot.com/about (auction house for spots in line or seating: “We all jockey for space with others. BetrSpot is now THE ADVANTAGE you have to get that spot someone may be leaving before others do.”). 178. E.g., Josh Constantine, Stop the JerkTech, TECHCRUNCH (July 3, 2014), http://techcrunch.com/2014/07/03/go-disrupt-yourself/; Tekla Perry, Drawing the Line Between “Peer-to-Peer” and “Jerk” Technology, IEEE SPECTRUM (July 18, 2014, 1:59 PM), http://spectrum.ieee.org/view-from-the-valley/at-work/start-ups/drawing-the-line- between-peertopeer-and-jerk-technology/. 179. See Brian Ng, The Race to Reinvent Parking: Are Monkey Parking, Sweetch and ParkModo Unethical, Illegal, or the Next AirBnB?, STRIKINGLY, http://blog. strikingly.com/strikingly_stories/reinvent-parking-monkey-parking-sweetch-parkmodo/ (last visited Oct. 27, 2014) (discussing an interview with Paolo Dobrowolny, cofounder of Monkey Parking, on the parking auction business model).

2015] REGULATING THE SHARING ECONOMY 1099 implicate public goods and disrupt legitimate business operations, these platforms also pose unique regulatory challenges.180 Current regulatory efforts have not yet addressed many of the legal issues raised by sharing platforms. As the sharing economy continues to grow, regulators will need to confront these gray areas. A particular area of concern is platform and provider liability for user and third-party injuries resulting from transactions on the platform.

C. INTERMEDIARY LIABILITY: TWO-PARTY RULES NO LONGER FIT THREE-PARTY RELATIONSHIPS For the most part, transactions on sharing platforms do not introduce new risks but new parties. The same accidents can occur in hotels as apartments, taxis as TNC cars, and ski rental shops as P2P ski rental services. For that reason, regulators have argued that providers on sharing platforms should meet the same requirements as traditional service providers. But the balance of power has shifted between the parties on either side of those accidents. Providers do not have the same “special relationship” to users that traditionally justified a higher regulatory burden on service providers. Nor can regulators hold platforms indirectly liable for all accidents resulting from transactions on the platform. To provide clear guidance to courts and enforcement agencies, regulators must decide how to divide liability between providers and platforms. 1. Provider Liability The sharing economy disrupts many of the assumptions that traditionally justified heightened duties for providers: the particular vulnerability of consumers, the superior bargaining position of providers in relation to consumers, and the treatment of providers as least-cost avoiders. Providers on sharing platforms are typically casual or informal market participants. They operate on a smaller scale than traditional service providers (a single bedroom rather than a hotel), use personal resources (a personal vehicle or home) rather than commercial assets (a licensed taxi or commercial building), and are likely less sophisticated than

180. San Francisco Dennis Herrera issued a cease-and-desist order to Monkey Parking in June 2014. Monkey Parking initially ignored the order, but later “temporarily disabled” the app within San Francisco. Laura Entis, Monkey Business: Tells Parking Spot App to Shut Down, ENTREPRENEUR (July 11, 2014), http://www.entrepreneur.com/ article/235575/. It has since expanded its operations to Santa Monica, California. Gene Maddaus, Kicked Out of San Francisco, MonkeyParking App Plans a Fresh Start in Santa Monica, L.A. WEEKLY (Sept. 18, 2014), http://www.entrepreneur.com/ article/235575/.

1100 BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 30:385 experienced professionals working for a traditional service provider (a casual cyclist renting her own bicycle versus a professional mechanic or salesman at a bicycle rental business). Thus, providers may be unable to shoulder the same regulatory burdens that apply to traditional service providers. Although sharing services pose the same health and safety risks for users as traditional service providers, these transactions put providers at risk as well. In many cases, providers are just as vulnerable as users. For example, TNC riders have reported incidents of assault and battery,181 sexual assault,182 and reckless driving while using the platform.183 Yet TNC drivers face many of the same risks as riders,184 and generally receive minimal if any training on how to handle difficult or dangerous riders.185 Likewise, short-term rental hosts may encounter serious risks of personal injury or property damage by inviting strangers into their homes. Moreover, sharing platforms standardize transactions such that providers no longer clearly occupy a superior bargaining position. In many

181. Ellen Huet, Uber Rider Might Lose An Eye From Driver’s Hammer Attack. Could Uber Be Held Liable?, FORBES (Sept. 30, 2014), http://www.forbes.com/sites/ellenhuet/2014/09/30/uber-driver-hammer-attack-liability/. 182. Hal Eisner, Rideshare Driver Accused of Kidnapping, Sexually Assaulting Passenger, MY FOX LA, (June 3, 2014), http://www.myfoxla.com/ story/25684061/rideshare-driver-in-custody-accused-of-kidnapping-woman/. 183. Ben Popken, States Warn of Rideshare Risks for Passengers, NBC NEWS (May 28, 2014), http://www.nbcnews.com/business/consumer/states-warn-rideshare-risks- passengers-n116736/. 184. OSHA reports that taxi and for-hire drivers are “over 20 times more likely to be murdered on the job than other workers.” PREVENTING VIOLENCE AGAINST TAXI AND FOR-HIRE DRIVERS, OSHA (2010), available at https://www.osha.gov/Publications/taxi-driver-violence-factsheet.pdf. TNCs eliminate many of the conditions that make taxis high risk, such as transacting electronically rather than in cash and requiring passengers to register for accounts with a valid credit card. But like taxis, TNCs often serve inebriated passengers, and unlike taxis, TNCs drive personal vehicles that do not have physical controls like barriers between the driver and passenger, or silent alarms. LISA RAYLE, ET AL., APP-BASED, ON-DEMAND RIDE SERVICES: COMPARING TAXI AND RIDESOURCING TRIPS AND USER CHARACTERISTICS IN SAN FRANCISCO, UNIV. CAL. TRANSP. CTR., UNIV. CAL. BERKELEY (2014) (finding in a survey that nineteen percent of users listed “[d]idn’t want to drive after drinking” as one of the top two reasons to request a ride). Moreover, while TNCs claim to conduct background checks on drivers, no such checks are in place for riders. See, e.g., LYFT, https://www.lyft.com/ (last visited Feb. 26, 2015) (“With just one tap, get matched with a friendly, background-checked driver.”). 185. Emily Huet, Uber Skimps On Driver Training, Then Charges Drivers $65 For Basic Driver Skills Course, FORBES (Oct. 8, 2014), http://www.forbes.com/sites/ ellenhuet/2014/10/08/uber-skimps-on-driver-training-then-charges-drivers-65-for- basic-driver-skills-course/.

2015] REGULATING THE SHARING ECONOMY 1101 cases, the terms of the platform govern all aspects of the transaction, leaving both providers and users only the freedom to accept or reject a potential match.186 In other cases, the platform provides a default contract or baseline set of terms, but allows users and providers to negotiate privately.187 It is unclear how often providers actually negotiate with users. Even where these negotiations result in material changes to the contract, both parties occupy relatively even footing. Users can propose and modify terms by bargaining with providers, and can easily comparison-shop.188 While providers may occupy a superior position in certain markets with limited supply and high user demand, this does not guarantee that providers will have the upper hand across the platform.189 Finally, on sharing platforms, providers do not necessarily represent the least-cost avoider. Most providers do not transact at sufficient volumes to absorb the cost of accidents.190 Personal insurance policies may not cover property damage during commercial transactions, and providers may be unable to purchase or afford commercial insurance.191 Since providers

186. Sharing services that operate on this model include TNCs, per-task contractor services, and to some degree, peer-to-peer lending services. 187. Examples include short-term rental platforms, equipment-rental platforms, and auction houses for goods or licenses. 188. Jasper Ribbers, Three Tips For Airbnb Hosts Negotiating With Guests, THE TRAVELING DUTCHMAN (July 29, 2014), http://www.thetravelingdutchman.com/three- tips-airbnb-hosts-negotiating-guests/ (last visited Feb. 26, 2015). 189. Studies of Airbnb use patterns in San Francisco show that hosts may earn more in tourist hotspots. On the other hand, potential suppliers may quickly enter the market in response to high prices from excess demand, and the most expensive listings may simply be higher value properties (luxury apartments) rather than run-of-the-mill spare rooms. See Carolyn Said, Window into Airbnb’s hidden impact on S.F., SAN FRANCISCO CHRONICLE, http://www.sfchronicle.com/business/item/window-into-airbnb-s-hidden- impact-on-s-f-30110.php (last visited Oct. 27, 2014). 190. Most short-term rental platforms advise both users and guests to purchase travel or vacation rental insurance, and only offer limited coverage for property damage—not personal liability—through the platform. E.g., How do I protect my property from potential damage?, VRBO, http://help.vrbo.com/articles/en_US/Article/How-do-I-protect-my- property-from-potential-damage?. San Francisco recently passed legislation requiring platforms to indemnify guests and hosts for up to $500,000. DAVID CHIU, SHORT- TERM RENTAL REGULATION SUMMARY (2014), available at http://www.sfbos.org/Modules/ShowDocument.aspx?documentid=50472. Purchasing appropriate vacation rental insurance, however, may be challenging for hosts who only list single bedrooms. See Insurance on a rental vacation rental, COMMUNITY FROM HOMEAWAY, https://community.homeaway.com/thread/1488 (last visited Oct. 27, 2014). 191. See Lieber, supra note 56. Peers, an advocacy group for sharing economy platforms, has introduced a new insurance product to help address the needs of users and providers. See also Ellen Huet, Peers Launches Home Liability And Car-Replacement Insurance For Airbnb, Uber, Lyft Workers, FORBES (Dec. 4, 2014),

1102 BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 30:385 generally list their personal assets on sharing platforms, the financial toll of unreimbursed property damage may place providers in a highly vulnerable position. With respect to accidents involving users, this may not only prevent providers from guarding against crushing liability, but also leave injured users without adequate remedies. Nor can providers afford or even feasibly adopt the security measures available to small businesses. For example, TNC drivers cannot reasonably install security barriers in their personal vehicles. Short-term rental hosts may not have control over common areas within apartment complexes.192 In certain cases where providers may face personal security risks as a result of user behavior, users may be the least-cost avoider. Per-task contractor services like TaskRabbit are a perfect example. The provider generally travels to the user’s home or workplace and performs services at the user’s direction. The user likely has superior knowledge of concealed dangers on the premises as well as any equipment or goods at issue. Thus, if a user hires a per-task contractor to mow her lawn, the user is in the best position to warn the contractor of pitfalls hidden in her yard.193 Given the relative vulnerability of casual providers, it is inappropriate to impose the same duties for providers on sharing platforms as traditional service providers. Regulators may therefore be tempted to establish indirect liability for platforms for accidents related to transactions on the platform.

http://www.forbes.com/sites/ellenhuet/2014/12/04/peers-home-liability-car- replacement-insurance/. 192. By contrast, courts have imposed duties of care on both landlords and hotels for tenant or guest safety in common areas. The applicable standard of care often requires that the owner of the premises have notice of relevant risk, and the capacity to adopt protective measures. See Kline v. 1500 Mass. Ave. Apartment Corp., 439 F.2d 477 (D.C. Cir. 1970). 193. Homeowner’s insurance would cover this accident regardless of whether the contractor had any affiliation with a per-task contractor service. Filipowicz v. Diletto, 796 A.2d 296 (N.J. Super. Ct. App. Div. 2002) (finding that a garage sale shopper was an invitee on the homeowner’s premises, therefore the homeowner owed a duty of care with respect to maintenance of his yard); Who cares if the guy mowing my lawn isn't insured, injuries are his problem, not mine. Right?, SAFETY INSURANCE GROUP INC., https://www.safetyinsurance.com/resource_center/homeowners/lawnmowing.html (last visited Oct. 27, 2014). However, contractors on TaskRabbit accept a variety of tasks, including deliveries and event staffing services; thus, while the user’s duty of care may be clear cut in some circumstances, this may not always be the case. Nor does TaskRabbit require that users carry homeowner’s insurance.

2015] REGULATING THE SHARING ECONOMY 1103

2. Platform Liability In many cases, the sharing platform represents either the least-cost avoider or the easiest target for enforcement agencies. But sharing platforms act—or purport to act—as neutral intermediaries that facilitate transactions between users and providers.194 Therefore, regulators can only hold platforms accountable for user and provider activity through indirect liability.195 This Section describes three models of intermediary liability for online service providers, and how these models might apply to sharing platforms. Whether platforms are truly neutral intermediaries is not always clear. In some cases, platforms have no formal contractual relationship with providers. For example, P2P rental services and short-term rental platforms do not have an employment relationship with providers. In other cases, platforms hire providers as independent contractors, and therefore are not liable for a contractor’s torts.196 For example, in January 2014 an Uber driver struck and killed a four-year old girl, Sofia Liu. The driver did not have a passenger or a passenger request at the time of the accident, but did have the Uber app on and was distracted by the GPS on his phone. Liu’s family later filed suit against Uber for wrongful death. Uber replied that it does not provide a transportation service, and instead “allows riders to connect with, and request transportation services from, a range of independent transportation providers.”197 Therefore Uber contended that it should not be liable for the driver’s negligence.

194. Anthony Ha, SideCar’s Sunil Paul On Working With (And Battling) Regulators, TECHCRUNCH (May 12, 2013), http://techcrunch.com/2013/05/12/sidecar-sunil-paul- backstage/ (citing an interview where Sunil Paul, Sidecar cofounder, described the company as an “information provider and matching service”); Sarah Buhr, Brian Chesky Talks About Just How Different The Hotel Business Is From Airbnb, TECHCRUNCH (Sept. 9, 2014), http://techcrunch.com/2014/09/09/brian-chesky-hotels-and-airbnb-are-the- same-but-different/ (quoting Brian Chesky, cofounder of Airbnb, claiming that the service is not a hotel: “We had to ask ourselves a couple of years ago we’re not in the business of homes, we’re not in the business of space, we’re in the business of trips”). 195. This is true regardless of whether sharing companies properly classify providers as independent contractors or employees. Cf. Doug Lichtman & Eric Posner, Holding Internet Service Providers Accountable, 14 SUP. CT. ECON. REV. 221 (2006) (proposing indirect liability for internet service providers by drawing a comparison to “the common law principle of vicarious liability [that] obligates employers to monitor, train, and otherwise exercise control over the behavior of their employees”). 196. DAN B. DOBBS ET AL, THE LAW OF TORTS (2d ed. 2011) §§ 431–32. 197. Answer and Affirmative Defenses at 6, Liu v. Uber Techs., Inc., No. CGC-14- 536979 (Cal. Super. Ct. Jan. 27, 2014).

1104 BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 30:385

Commentators have questioned whether sharing platforms misclassify employees as independent contractors,198 and several suits have challenged TNCs on this basis.199 However, determining whether an employer has misclassified an employee is a fact-dependent analysis.200 Taxi drivers, for instance, are typically considered independent contractors, but in some cases courts have found that taxi drivers qualify as employees.201 In addition, different states apply different legal standards. Providers on some sharing platforms may qualify as independent contractors while others do not, and some jurisdictions may reach conflicting decisions. Assuming that platforms do act as online intermediaries, regulators may still hold platforms liable through indirect liability. Scholars disagree on whether internet intermediary liability is desirable, and if so, under what circumstances.202 This Section describes three models for intermediary liability that apply to other online services: (1) the absolute immunity of the Communications Decency Act (“CDA”), (2) the qualified immunity or safe harbors of the Digital Millennium Copyright Right Act (“DMCA”), and (3) an active duty to monitor, comparable to the voluntary enforcement practices of payment intermediaries.

198. E.g., Kevin Roose, Does Silicon Valley Have a Contract-Worker Problem?, N.Y. MAG. (Sept. 18, 2014, 8 :42 AM), http://nymag.com/daily/ intelligencer/2014/09/silicon-valleys-contract-worker-problem.html. 199. Cotter v. Lyft, Inc., No. 13-cv-04065-VC, 2014 WL 3884416 (N.D. Cal. Aug. 7, 2014). 200. DAN B. DOBBS, ET AL, THE LAW OF TORTS (2d ed. 2011) §§ 431–32. 201. NAT’L EMP. LAW PROJECT, UNREGULATED WORK IN THE TAXI INDUSTRY IN NEW YORK CITY (2007), available at http://www.nelp.org/page/- /EJP/Unregulated_Work_Taxi.pdf; EMP’T DEV. DEP’T OF CAL., INFORMATION SHEET: TAXICAB INDUSTRY (2009), available at http://www.edd.ca.gov/ pdf_pub_ctr/de231tc.pdf; Industry Specific Help for Worker Classification, Oregon.gov, http://www.oregon.gov/IC/pages/07-industries.aspx#taxicab (noting that some dispatch services hire drivers as independent contractors, and these services must meet state criteria including “(1) that [drivers] be free from the right of another to control how services are performed and, (2) that the worker operates an independently established business”). 202. Lichtman and Posner have argued that the desirability of indirect intermediary liability for ISPs depends on two considerations: (1) control—whether the intermediary “is in a good position to detect or deter” user and provider misconduct, and (2) activity— whether “liability would serve to encourage [the intermediary] to internalize some significant externality unavoidably associated with its activities.” Lichtman & Posner, supra note 195, at 230. Likewise, Ronald Mann and Seth Belzley have criticized ISP immunity, calling for indirect intermediary liability in cases where ISPs are least-cost avoiders. Ronald J. Mann & Seth R. Belzley, The Promise of Internet Intermediary Liability, 47 WM. & MARY L. REV. 239 (2005). Mark MacCarthy argues that intermediary liability should depend not only on a cost-benefit analysis, but also equitable considerations. What Payment Intermediaries Are Doing About Online Liability And Why It Matters, 25 BERKELEY TECH. L.J. 1037 (2010).

2015] REGULATING THE SHARING ECONOMY 1105

At one end of the spectrum, the CDA provides absolute immunity to “interactive computer service provider[s],” meaning “any information service, system, or access software provider that provides or enables computer access by multiple users to a computer server.” More specifically, § 230 provides that “[n]o provider or user of an interactive computer service shall be treated as the publisher or speaker of any information provided by another information content provider.”203 Courts have broadly applied § 230 to absolve internet intermediaries from liability for user- generated content, even if the intermediary is aware that the content is illegal and declines to remove it.204 Thus, recent decisions have held that dating services are not responsible for user-created fraudulent dating profiles,205 eBay is not responsible for defects in products sold by users,206 and StubHub is not responsible for user violations of anti-scalping laws.207 Intermediaries may, nevertheless, incur liability for directing or shaping user generated content “in some way [that] specifically encourages the development of what is offensive about the content.”208 However, § 230 contains a significant exception for intellectual property claims.209 To fill this gap, the DMCA provides limited safe harbors to internet service providers (“ISPs”) for transmitting, caching, storing, or indexing copyright infringing content at the direction of users.210 The conditions of these safe harbors vary. With respect to “[i]nformation residing on systems or networks at [the] direction of users,” ISPs can only benefit from the DMCA safe harbors by implementing a “notice and take-down” system (a system for removing infringing content upon the copyright owner’s request) and a system to identify and terminate the accounts of “repeat infringers.”211 Beyond these requirements, DMCA § 512(c) has two exceptions that prevent ISPs from taking advantage of this safe harbor: an ISP can be liable for vicarious

203. 47 U.S.C. § 230(c)(1) (2012). 204. Catherine R. Gellis, 2012 State of the Law Regarding Internet Intermediary Liability for User-Generated Content, 68 BUS. LAW. 289, 289–90 (2012); Mark A. Lemley, Rationalizing Internet Safe Harbors, 3 (2007) (unpublished draft). 205. Christianne Carafano v. Metrosplash.com, Inc., 339 F.3d 1119 (9th Cir. 2003). 206. Inman v. Technicolor USA, Inc., No. 11-666, 2011 WL 5829024 (W.D. Pa. Nov. 18, 2011). 207. Hill v. StubHub, Inc., No. COA11-685, 2012 WL 696223, at *13 (N.C. Ct. App. Mar. 6, 2012). 208. Fair Hous. Council of San Fernando Valley v. Roommates.com, LLC, 666 F.3d 1216 (9th Cir. 2012). 209. 47 U.S.C. § 230(e)(2); Lemley, supra note 204, at 4. 210. 17 U.S.C. § 512; Lemley, supra note 204, at 4. 211. 17 U.S.C. § 512(c)–(d); Gellis, supra note 204, at 293–94.

1106 BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 30:385 infringement where the ISP receives a “direct financial benefit attributable to the infringing activity” or where the ISP is aware—or should be aware—of the infringing activity.212 At the other end of the spectrum, legislators could impose an active duty to monitor.213 Though this approach is rare, Congress has proposed legislation that would codify affirmative duties for payment intermediaries to screen out merchants trafficking in child pornography or controlled substances, selling tobacco to minors, and violating online gambling prohibitions.214 Payment intermediaries like Visa and MasterCard currently collaborate with law enforcement agencies voluntarily through a “two-part program,” by instituting “due diligence requirements” to keep offending merchants out of the payment system and a “monitoring program to detect and expel” offending merchants.215 Regulators could adapt one of these models to establish liability for sharing platforms. For example, regulators could grant platforms immunity for harms that arise from transactions between users and providers. On the other hand, regulators might also create a notice-based system where platforms only become liable for the actions of providers when the platform has previously received notice of the provider’s poor conduct. For instance, a platform might not be liable where a driver assaults a passenger. However, if users have previously submitted complaints about a particular driver’s behavior to the platform, then the platform might become liable for future injuries caused by that driver. Alternatively, regulators could require platforms to take affirmative steps to protect users from dangerous drivers. It is unclear which of these three approaches to intermediary liability would appropriately allocate liability between sharing companies, users,

212. 17 U.S.C. § 512(c)–(d); Gellis, supra note 204, at 293–94. 213. In 2011, Congress nearly passed legislation that would have largely displaced the DMCA, adding private causes of action to enjoin illegal streaming services, and increasing penalties for infringing streaming services. By proxy, this legislation would have imposed an active duty to monitor on ISPs. Stop Online Piracy Act, H.R. 3261, 112th Cong. (2011); Preventing Real Online Threats to Economic Creativity and Theft of Intellectual Property Act of 2011, S. 968, 112th Cong. (2011); see also Chris Civil, When The Net Went Dark: SOPA, PROTECT IP and the Birth of an Internet Movement, BERKELEY TECH. L.J.: THE BOLT (Feb. 14, 2012), http://btlj.org/2012/02/14/when- the-net-went-dark-sopa-protect-ip-and-the-birth-of-an-interent-movement/. 214. MacCarthy, supra note 202, at 1059–98. 215. Id. at 1075–76, 1079–80, 1090–95. Courts have also declined to impose affirmative duties on payment intermediaries to monitor transactions where websites may contain copyright infringing content. E.g., Perfect 10, Inc. v. Visa Intern. Serv. Ass’n, 494 F.3d 788, 807 (2007).

2015] REGULATING THE SHARING ECONOMY 1107 and providers. Unlike internet service providers, sharing companies are not passive; most platforms exercise at least some control over provider and user transactions, and almost all platforms have a financial stake in transactions between users and providers.216 This suggests that immunity for platforms is not appropriate. In addition, a notice-based system of liability comparable to the DMCA § 512(c) safe harbor may not sufficiently safeguard users. Yet imposing an affirmative duty to monitor user and provider compliance with local regulations is extremely costly. Even the comparatively easier task of screening offending merchants from payment systems has proven controversial.217 Regulators in various jurisdictions have begun to experiment with solutions to these problems. The following discussion highlights and critiques some of these rules, and proposes general principles for future legislation.

IV. PROPOSED SOLUTIONS As previously discussed, sharing platforms are markets for P2P services. These markets allow informal participants to operate at a small scale by minimizing transaction costs. Sharing services are often closely analogous to traditional service providers.218 For the most part, this means that regulators must address old risks that arise from transactions between new parties.219 P2P service markets also create some unique risks for users and providers, however. This Part addresses regulatory solutions to old and new risks in turn.

A. OLD RISKS, NEW PARTIES: REGULATORS SHOULD CONSIDER TIERED REGULATIONS, LIMITED INTERMEDIARY LIABILITY, AND DUTIES TO THIRD PARTIES The proper allocation of liability between users and providers on sharing platforms is highly contextual because both the least-cost avoider and the easiest target for enforcement agencies vary between and within

216. Sharing platforms impose a wide range of transaction fees. See supra note 27. Interchange fees might serve as an indirect measure of the degree of control a platform exercises over transactions. 217. See MacCarthy, supra note 202, at 1083–87 (arguing that imposing statutory duties to monitor for child pornography, controlled substances, and gambling is unnecessary, and even voluntary monitoring programs may not be cost effective); Mann & Belzley, supra note 202, at 260 n.59 (claiming that voluntary agreements between intermediaries and enforcers develop in the shadow of the law). 218. Supra Sections I.B–II.A. 219. Supra Section III.C.

1108 BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 30:385 sharing service models. Since sharing platforms often operate on a large scale over multiple jurisdictions, the least-cost avoider for compliance may often be users and providers, who are most familiar with local rules and their own property. Moreover, casual and high volume providers may require different treatment. Yet in other situations platforms may be best positioned to address user or provider misconduct. Regulators will likely need to address sharing platform liability not just on a case-by-case basis, but also on an issue-by-issue basis. This Note therefore proposes that regulators (1) adopt tiered regulatory schemes for providers, (2) design nuanced frameworks for intermediary liability, and (3) impose duties on platforms and providers to third parties. 1. Tiered Regulation for Providers Because previous justifications for provider liability often do not apply in the sharing context, regulators should consider relaxed regulatory requirements for casual providers.220 However, many providers operate at sufficiently high volumes to merit stricter standards. The Chicago ordinance governing TNCs provides a working example of tiered regulation. The city offers Class A licenses to drivers whose average service operation is twenty hours or less per week, and Class B licenses to drivers who work in excess of twenty hours per week. Class A licenses cost less and require less stringent yearly inspections.221 Attempting to block high volume users may needlessly inhibit beneficial activity on the platform. For example, the San Francisco short- term rental ordinance requires that hosts occupy the residence “for no less than 275 days out of the calendar year in which [it] is rented.”222 The legislature intended this provision to prevent the “widespread conversion of residential housing to short-term rentals” and ensure that listed properties “remain truly residential in use.”223 However, this provision may bar desirable uses of residential properties. Potential hosts who might not meet the 275-day requirement include professors on sabbatical,

220. The Portland ordinance on short-term rentals, for example, differentiates between Type A rentals, “where no more than [two] bedrooms are rented,” and Type B rentals, “where [three] or more bedrooms are rented.” PORTLAND, OR., PLANNING AND ZONING § 33.207.020 (2015). 221. CHI., ILL., CODE § 9-115-030 (2014). 222. S.F., CAL., ADMIN. CODE § 41A.5(g)(1)(A) (2014). The Portland ordinance on short-term rentals imposes a similar requirement on Type A short-term rentals. PORTLAND, OR., PLANNING AND ZONING § 33.207.040(A)(1) (2015) (“a resident must occupy the dwelling unit for at least 270 days during each calendar year”). 223. S.F., CAL., ADMIN. CODE § 1(c)(1)-(2).

2015] REGULATING THE SHARING ECONOMY 1109 salespeople who regularly travel for business, and those who own a second home in the city. It also permits high volume uses of rooms that might otherwise serve as residential housing. An apartment tenant can rent a second bedroom year round, so long as she co-occupies the residence for at least 275 days. Rooms like this might otherwise serve as long-term housing.224 Rather than imposing caps on use, regulators should impose higher taxes and stricter regulations on high volume providers.225 Tiered regulation properly allocates risk to repeat players who benefit most from the platform. For instance, providers who transact at volumes comparable to bed and breakfasts should incur similar liability to a traditional service provider. And TNC drivers who operate like full-time taxis should comply with safety standards comparable to taxis. In addition, tiered regulation levels the playing field between sharing platforms and traditional service providers, and discourages abuse of the platform. In a tiered scheme, regulators can combat “hotelization” by raising the cost for high volume listings of short-term apartments to the point where long-term residential leases become more profitable. 2. Limited Intermediary Liability for Platforms Given the public interest in consumer and worker welfare, absolute immunity for sharing platforms rarely makes sense. Notice-based systems may also fail to adequately protect consumers, because platforms only become liable after a user has interacted with a dangerous provider. Yet imposing affirmative duties on internet intermediaries is an equally extreme solution. Regulators should therefore consider addressing platform liability on an issue-by-issue basis. Regulators should establish affirmative duties where platforms exercise sufficient control over the relevant aspect of the transaction, and that transaction poses a serious risk of injury for users or providers. For example, several ordinances on TNCs restrict the maximum number of hours that a driver can operate per day, but some TNC platforms do not directly restrict drivers’ hours.226 If a TNC driver leaves the app on twenty-

224. The same criticisms apply to the New York Multiple Dwelling Statute, which requires that the host co-occupy the residence for the duration of the rental. N.Y. Mult. Dwell. Law § 4(8)(a) (McKinney 2011). 225. Contra Badger, supra note 52. 226. E.g, COLO. REV. STAT. § 40-10.1-605(1)(e) (2014). Messages on TNC driver forums suggest that it is not uncommon for drivers to work twelve-hour shifts or longer. E.g., SoCal_Uber, Your longest continuous hours of Uber driving?, UBERPEOPLE.NET,

1110 BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 30:385 four hours a day and recklessly accepts fares without adequate rest, should the platform have a duty to terminate that driver? The platform exercises sufficient control such that it represents both the least-cost avoider and easiest enforcement target. Platforms can monitor the number of hours drivers keep their apps on, and they have the ability to terminate drivers for poor conduct. Moreover, enforcement officials would find it relatively difficult to uncover individual driver misconduct, but could easily determine platform negligence by obtaining the platform’s records on driver app usage. Other potential affirmative duties include insurance mandates,227 background checks for providers,228 and routine inspections for vehicles or rental property.229 In addition, regulators could take a sliding scale approach: the greater the risks for users and providers, the less control required to establish an affirmative duty. Even where an affirmative duty to monitor is an appropriate solution, enforcement agencies may find it preferable to request that TNCs take voluntary measures to police provider activity, rather than adopting formal legislation.230 Where the platform does not exercise sufficient control over user or provider behavior, regulators should instead consider a notice-based system or a limited duty to warn. Under a notice-based system, platforms only become liable if they fail to take action after receiving a complaint from a user or provider. For example, some statutes have established notice-based liability provisions for TNC platforms where drivers discriminate against disabled riders231 or violate a zero-tolerance policy for drugs or alcohol.232 Beyond background checks for providers, the platform has few additional opportunities to monitor driver behavior. Notice-based systems therefore shield platforms from excessive liability without immunizing the platform. Drivers and platforms remain liable for injuries http://uberpeople.net/threads/your-longest-continuous-hours-of-uber-driving.180/ (last visited Feb. 26, 2015). 227. CAL. PUB. UTIL. CODE §§ 5431–37 (West 2015). 228. D.C. CODE § 20j-2 (2015). 229. COLO. REV. STAT. § 40-10.1-605(1)(g). 230. See MacCarthy, supra note 202, at 1083–87; Mann & Belzley, supra note 202, at 260 n.59. 231. The Colorado TNC law enacts a notice-based system for violations of the duty not to discriminate. COLO. REV. STAT. § 40-10.1-605(7)(a). The Cal. P.U.C. has also imposed a notice-based system for a zero-tolerance policy. CAL. P.U.C., BASIC INFORMATION FOR TRANSPORTATION NETWORK COMPANIES AND APPLICANTS (2013), available at http://www.cpuc.ca.gov/NR/rdonlyres/1788F1F1-EA38-4B68- B221-4116994F2252/0/TNC_App_Instrctns.pdf. 232. D.C. CODE § 20j-1(9) (2015).

2015] REGULATING THE SHARING ECONOMY 1111 that occur as a result of dangerous conduct. However, so long as a platform subsequently takes action to remedy the problem, the platform does not incur liability for violating civil rights statutes or fines related to its zero-tolerance policy. Other cases that merit a notice-based system include violations of local zoning laws and use of the platform for illegal purposes like prostitution or drug trafficking.233 In other situations, regulators might consider establishing a limited duty to warn. For example, the San Francisco ordinance on short-term rentals requires that platforms inform hosts of the local requirements for compliance with the housing and tax code.234 Airbnb has also voluntarily provided similar notice of the New York multiple-dwelling statute to potential hosts within the city.235 Regulators should consider imposing similar duties where providers represent the least-cost avoider, but platforms can easily convey the relevant information during the hiring or sign-up process. 3. Duties to Third Parties Beyond risks for users and providers, sharing platforms may create negative externalities for third parties.236 Short-term rentals, for example, affect residents of multiple-dwelling buildings, neighborhoods, and communities as a whole.237 Regulators in some jurisdictions have created ordinances to address noise complaints and other nuisances, and some have established complaint hotlines for residents to report bad behavior.238 These enforcement efforts, however, may prove costly for local

233. E.g., Dana Sauchelli & Bruce Golding, Hookers Turning Airbnb Apartments Into Brothels, N.Y. POST (July 31, 2014), http://nypost.com/2014/04/14/hookers-using- airbnb-to-use-apartments-for-sex-sessions/; Michael Arrington, Another Airbnb Victim Tells His Story: “There Were Meth Pipes Everywhere,” TECHCRUNCH (July 31, 2011), http://techcrunch.com/2011/07/31/another-airbnb-victim-tells-his-story-there-were- meth-pipes-everywhere/. 234. S.F., CAL., ADMIN. CODE § 25(A) (2014). 235. Ron Lieber, A $2,400 Fine for an Airbnb Host, N.Y. TIMES (May 21, 2013, 2:22 PM), http://bucks.blogs.nytimes.com/2013/05/21/a-2400-fine-for-an-airbnb-host/. 236. E.g., N.Y. Attorney General Report, supra note 40, at 38–39 (detailing complaints from neighbors of Airbnb hosts); see also supra text accompanying notes 196– 197 (discussing TNC liability for driver accidents involving third parties). 237. Jessica Pressler,“The Dumbest Person in Your Building Is Passing Out Keys to Your Front Door!”, N.Y. MAG. (Sept. 23, 2014), http://nymag.com/news/ features/airbnb-in-new-york-debate-2014-9/. 238. GRAND COUNTY, COLO., ZONING REGULATIONS, SECTION XIV SUPPLEMENTAL REGULATIONS § 14.8(2) (2014); Vacation Rentals, CITY OF PALM SPRINGS, http://www.ci.palm-springs.ca.us/index.aspx?page=1098 (last visited Feb. 26, 2015).

1112 BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 30:385 administrative agencies. Regulators should shift this burden to platforms and providers—those who most directly benefit from sharing platform activity—by requiring (1) that platforms adopt third-party dispute resolution systems and (2) that providers notify third parties of their activities. First, platforms should provide a complaint hotline and dispute resolution procedure for third parties who seek to remove listings from the platform or report poor conduct.239 For short-term rental platforms, affected third parties might include landlords who want to prevent tenants from posting listings in violation of a lease, or neighbors of hosts who seek a remedy against a guest for property damage.240 For TNC or P2P car rental platforms, third parties might report dangerous conduct by TNC drivers, or submit complaints about TNC driver congestion in high traffic areas like airports. These complaints could create a duty to take action under notice-based schemes for platform liability. In addition, platforms could provide limited remedies such as reimbursement for small damages claims, or in some cases removal of an offending listing or profile. Any dispute resolution forum should fairly account for the legitimate interests of users and providers as well as third parties. For instance, before removing a short-term rental listing, platforms might require that landlords demonstrate ownership of the relevant property and an enforceable provision in the lease prohibiting subletting. In some cases, regulators might also consider requiring providers to give notice of their activities to third parties. Maui requires that short- term rental hosts provide notice to neighbors within 500 feet of the listed property.241 Jurisdictions could also create a standardized notice form that requires hosts to provide a contact number for the host, platform, and local government complaint hotline.

239. Some platforms already run call centers for user and community complaints. Spencer Peterson, Airbnb Creates a Call Center That’s Not a Total Nightmare, CURBED (Dec. 17, 2014), http://curbed.com/archives/2014/12/17/airbnbs-portland-office- design.php; Contact Us!, DOGVACAY, http://dogvacay.com/contact/ (last visited Feb. 26, 2015). Others have complaint email services directed toward users and providers. ROOMORAMA, https://www.roomorama.com/ (last visited Feb. 26, 2015); Help Center, LYFT, https://www.lyft.com/help/contact/passenger (last visited Feb. 26, 2015). 240. E.g., Joe Coscarelli, Airbnb Poster-Child Was Evicted for Airbnb-ing a Converted Barn She Didn’t Own, N.Y. MAG. (Oct. 10, 2014), http://nymag.com/daily/intelligencer/2014/10/airbnb-poster-child-shell-evicted-for- airbnbing.html. 241. MAUI, HI., COUNTY CODE § 19.65.050(B) (2014), available at http://www.co.maui.hi.us/DocumentCenter/View/91217.

2015] REGULATING THE SHARING ECONOMY 1113

Regulatory agencies should address negative externalities generated by the sharing economy.242 Yet regulators should balance the costs of regulation against the benefits of innovation. The solutions detailed above—tiered regulation for providers, limited intermediary liability, and duties to third parties—can help regulators to allocate protections and burdens more flexibly.243 B. NEW PARTIES, NEW RISKS: REGULATORS MUST ADDRESS BIG DATA, REPUTATION SYSTEMS, AND COMPETITION Although sharing platforms largely pose the same risks as equivalent traditional service providers, these platforms also create new consumer protection issues related to user privacy and reputation systems. Regulators should address both of these issues. In crafting new rules, regulators should also consider whether certain provisions overly restrict competition. 1. Big Data: Regulators Should Anticipate Privacy Concerns TNCs and other sharing companies collect vast quantities of sensitive user data.244 Rider and driver accounts on TNCs, for instance, include personal information like credit card numbers, as well as logs of prior trips.245 Airbnb likewise maintains a “Verified ID” system, which requires

242. Adopting this role would address the concerns of many critics of the sharing economy. See, e.g., Dean Baker, Don’t Buy the ‘Sharing Economy’ Hype: Airbnb and Uber Are Facilitating Rip-Offs, GUARDIAN (May 27, 2014), http://www.theguardian.com/commentisfree/2014/may/27/airbnb-uber-taxes-regulation (arguing that the sharing economy imposes costs on society at large, and arguing in favor of regulation to create a “level playing field”). 243. For a thorough discussion arguing in favor of balanced regulation, see DEBBIE WOSSKOW, UNLOCKING THE SHARING ECONOMY: AN INDEPENDENT REVIEW (2014). Wosskow’s report, prepared for the United Kingdom Minister of State for Business, Enterprise and Energy, provides detailed suggestions for reforms that would allow communities and government agencies to take advantage of the sharing economy. 244. Granting access to this data in response to a search warrant could have serious implications for criminal investigations. See Andrea Lance, Comment, Back to the Future of Your Privacy Rights, 95 MASS. L. REV. 214 (2013) (discussing the Court’s decision in United State v. Jones, 132 S. Ct. 945 (2012), regarding the installation of a GPS tracking device on a suspect’s vehicle). The significance of this issue applies not only to average criminal investigations, but potentially civil discovery orders. In some cases the public interest in allowing disclosure of rider logs may outweigh the rider’s privacy interest. For example, consulting firm Hamilton Places Strategies calculated based on congressional campaign filings that Members of Congress relied on Uber for sixty one percent of rides during the 2014 election cycle. Tom Kise, Uber: Congress’ New Private Driver, HPS (Nov. 11, 2014), http://www.hamiltonplacestrategies.com/news/uber-congress-new- private-driver/. 245. Uber has faced intense criticism for its internal use of rider log data. See, e.g., Johana Bhuiyan & Charlie Warzel, “God View”: Uber Investigates Its Top New York

1114 BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 30:385 that users provide a copy of a “government-issued ID, such as [a] driver’s license or passport.”246 And the potential for misuse extends beyond disclosure of individual account activity logs. Mining aggregated data can reveal information far more sensitive than that in any single account.247 In many ways, the accumulation of data through sharing platforms parallels the same concerns and disputes faced by any online service provider. One might argue that short-term rental logs are comparable to online databases for hotel guest records, that per-task contractor services grant only as much access to homes as traditional cleaning or repair services, and that TNC rider logs are comparable to GPS logs from personal vehicles.248 Enforcement agencies could therefore address these issues through existing privacy and data security laws. However, sharing platforms create unique issues regarding platform disclosure of user information to providers and vice versa. For example, TNCs create dummy numbers to allow drivers and riders to communicate before and after a ride. Drivers can only continue to reach the rider through the dummy number for up to half an hour following the ride, but riders can continue to reach the driver indefinitely. In some cases, the dummy number exposes the driver to harassment from disgruntled or

Executive For Privacy Violations, BUZZFEED (Nov. 18, 2014, 8:27 PM), http://www.buzzfeed.com/johanabhuiyan/uber-is-investigating-its-top-new-york- executive-for-privacy/; Craig Timberg et al., Uber executive stirs up privacy controversy, WASH. POST (Nov. 18, 2014), http://www.washingtonpost.com/business/ technology/uber-executive-stirs-up-privacy-controversy/2014/11/18/d0607836-6f61- 11e4-ad12-3734c461eab6_story.html (quoting Prof. Chris Hoofnagle: “We have to think about how the service provider itself can be a threat”); Letter from Sen. Al Franken to Travis Kalanick (Nov. 19, 2014), available at http://www.franken.senate.gov/ files/letter/141119UberLetter.pdf. 246. What Is Verified ID?, AIRBNB, https://www.airbnb.com/help/ article/450/ (last visited Nov. 22, 2014). 247. In a poignant example, Uber mined user data to discover the percentage of riders using the service for “one-night stands,” dubbing such trips “Ride[s] of Glory” and riders “RoGers.” Uber Crunches User Data to Determine Where the Most ‘One-Night Stands’ Come From, CBS SF BAY AREA (Nov. 18, 2014), http://sanfrancisco.cbslocal.com/2014/11/18/uber-crunches-user-data-to-determine-where- the-most-one-night-stands-come-from/. Uber originally published the results on the company blog, but has since removed the post. Rides of Glory, UBER, http://blog.uber.com/ridesofglory, available at http://web.archive.org/web/ 20140828024924/http://blog.uber.com/ridesofglory. 248. See Jaclyn Trop, The Next Data Privacy Battle May Be Waged Inside Your Car, N.Y. TIMES (Jan. 10, 2014), http://www.nytimes.com/2014/01/11/business/the-next- privacy-battle-may-be-waged-inside-your-car.html (describing privacy concerns implicated by the collection of data resulting from GPS devices, cameras, and other technologies that automakers increasingly integrate into personal vehicle designs).

2015] REGULATING THE SHARING ECONOMY 1115 aggressive riders.249 Disclosing a rider’s full name can also create risks for users, and some commentators have reported incidents of drivers stalking riders through social media.250 Similar issues can easily arise in the context of per-task contractor services and short-term rental services, where contractors and guests gain access to homes or office spaces. Despite these risks, disclosure of personal data may be necessary or beneficial in certain contexts. For TNCs, keeping a dummy number active helps riders to contact drivers in the event the rider forgets a personal item in the vehicle.251 Regulators may not have sufficient experience with sharing platforms to set viable guidelines for privacy policies. Moreover, users and providers consent to some level of disclosure by joining a sharing platform. Regulators might therefore address novel privacy and data security issues on platforms by ensuring the platforms uphold their stated privacy policies,252 and by adding flexible provisions requiring that platforms do not disclose personal information where doing so would unreasonably threaten the safety of users or providers.253

249. Polly Mosendz, Uber Drivers Have Privacy Problems Too, NEWSWEEK (Nov. 19, 2014), http://www.newsweek.com/uber-taxi-e-hailing-riding-app-travis-kalanick-emil- michael-josh-mohrer-uber-285642/. 250. A Daily Beast reporter describes her personal experience of harassment, including a driver who photographed her on the street prior to her trip and later emailed the photograph to her employer, and a driver who messaged the author’s contacts on Facebook. Olivia Nuzzi, Uber’s Biggest Problem Isn’t Surge Pricing. What If It’s Sexual Harassment by Drivers?, DAILY BEAST (Mar. 28, 2014), http://www.thedailybeast.com/articles/2014/03/28/uber-s-biggest-problem-isn-t-surge- pricing-what-if-it-s-sexual-harassment-by-drivers.html. 251. Polly Mosendz, Uber Drivers Have Privacy Problems Too, NEWSWEEK (Nov. 19, 2014), http://www.newsweek.com/uber-taxi-e-hailing-riding-app-travis-kalanick-emil- michael-josh-mohrer-uber-285642. 252. The FTC already brings enforcement actions against other online service providers for failures to uphold a stated privacy policy. Enforcing Privacy Promises, FTC, http://www.ftc.gov/news-events/media-resources/protecting-consumer-privacy/ enforcing-privacy-promises. 253. Some ordinances do expressly consider privacy concerns. For example, the District of Columbia ordinance on TNCs specifies that any administrative rule requiring TNCs to maintain records on drivers must not interfere with the legitimate privacy interests of users. D.C. CODE § 50-329 (2015). Legislatures could also specify some duty of reasonable care for platforms.

1116 BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 30:385

2. Reputation Systems: Regulators Should Encourage Responsible Private Ordering Regulators have focused on consumer welfare in proposing ordinances to govern sharing platforms, especially TNCs.254 Proposed rules for TNCs invariably detail required background checks for drivers and inspection procedures for vehicles.255 Many also specify limitations on surge pricing— a dynamic price scheme that increases fares at peak hours to encourage drivers to log on and fulfill the demand for rides.256 In some cases, ordinances specify that TNCs must make the method for calculating fares transparent, and that the TNC app must place the proposed fare prominently on screen in relation to the ride request confirmation button.257 Yet few regulations have addressed reputation systems on sharing platforms. In the context of sharing platforms, a reputation system usually refers to a record of qualitative reviews or numerical ratings tied to a user or provider’s profile.258 Most TNCs have a numerical rating system; at the end of a ride, riders rate drivers and vice versa259 on a scale of one to five.260 Short-term rental and per-task contractor platforms, by contrast, generally

254. E.g., San Miguel County Land Use Code, Colo., § 5-3001 (2014) (“The purpose of this Section is to promote public health safety and general welfare by establishing standards for rental of a Primary Residence for less than 30 days.”). 255. Cal. P.U.C. Decision, supra note 18, at 72–75. 256. E.g., D.C. CODE § 50-329.02(b)(13) (2015) (“During a state of emergency . . . [a TNC] that engages in surge pricing shall limit the multiplier by which its base fare is multiplied to the next highest multiple below the 3 highest multiples set on different days in the 60 days preceding the declaration of a state of emergency for the same type of service”). 257. E.g., COLO. REV. STAT. § 40-10.1-605(b) (2014) (“A transportation network company shall make available to prospective riders and drivers the method by which the transportation network company calculates fares . . . .”); CHI., ILL., CODE § 9-115- 200(b) (2014) (“[A]ny licensee shall display a button for displaying a fare quote for any requested trip on the licensee's Internet-enabled application or digital platform in the same size and graphics as the licensee's trip request button.”). 258. Tom Slee, Some Obvious Things About Internet Reputation Systems, WHIMSLEY (Sept. 29, 2013), http://tomslee.net/2013/09/some-obvious-things-about-internet- reputation-systems.html.; see also Chrysanthos Dellarocas, Designing Reputation Systems for the Social Web 2 (Bos. Univ. Sch. Mgmt., Research Paper Series No. 2010-18, 2010). 259. Julie Weed, For Uber, Airbnb and Other Companies, Customer Ratings Go Both Ways, N.Y. TIMES (Dec. 1, 2014), http://www.nytimes.com/2014/ 12/02/business/for-uber-airbnb-and-other-companies-customer-ratings-go-both- ways.html. 260. See, e.g., The Star Rating System, LYFT, https://www.lyft.com/help/ article/1453135/ (last visited Feb. 26, 2015).

2015] REGULATING THE SHARING ECONOMY 1117 include qualitative reviews on the user or provider’s profile.261 It is important to note that both user and provider reputations can influence access on the platform—platforms may terminate or block providers with low scores, and providers may decline to accept requests from users with low scores.262 These systems therefore give both users and providers an incentive to behave well on the platform. The exact implementation of a reputation system can also encourage different strategic behaviors on the part of users and providers, and certain system designs will result in more valuable or reliable information than others.263 While ratings and reviews cannot perform the same protective screening function as background checks and safety inspections,264 reputation systems do play a complementary and vital role in online marketplaces. Even if a provider meets the minimum legal standards within a jurisdiction, that assurance does not help consumers to discern the quality of providers in the market. For example, a car purchaser may know that both a Volvo and a Toyota meet the minimum safety standards for personal vehicles, but the purchaser may require additional information about the quality of the car to make an informed decision. In that sense, reputation systems serve a function similar to a brand or trademark for providers. The provider profile serves as an indication of source, and the reputation system allows the provider to build goodwill within the platform’s community of users.265 On the other hand, reputation systems function more closely to credit ratings from the user’s perspective.266 Having a higher reputation score can allow users to gain access to more—

261. See, e.g., How Do I leave a Review or Feedback for the Host?, ROOMORAMA, https://help.roomorama.com/hc/en-us/articles/200497795-How-do-I-leave-a-review-or- feedback-for-the-host-/ (last visited Feb. 26, 2015); How Do I Leave Review?, TASKRABBIT, https://taskrabbit.zendesk.com/entries/59906534-How-do-I-leave-review-/ (last visited Oct. 27, 2015). 262. Dara Kerr, Should Uber and Lyft Keep Passenger Ratings Secret?, CNET (Sept. 25, 2014, 7:00 AM), http://www.cnet.com/news/should-uber-and-lyft-keep-passenger- ratings-secret/. 263. See generally Sergio Marti & Hector Garcia-Molina, Taxonomy of Trust: Categorizing P2P Reputation Systems, 50 COMPUTER NETWORKS 472 (2006) (discussing the importance of design choice is creating reputation systems). 264. Some have argued that reputation systems can replace governmental regulation, and that the sharing economy does not require government intervention. KOOPMAN, supra note 20, at 16–19. 265. See Beth Simone Noveck, Trademark Law and The Social Construction Of Trust: Creating the Legal Framework for Online Identity, 83 WASH. U. L.Q. 1733, 1777–84 (2005). 266. See Danielle Keats Citron & Frank Pasquale, The Scored Society: Due Process for Automated Predictions, 89 WASH. L. REV. 1, 8–18 (2014).

1118 BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 30:385 and potentially higher quality—services.267 Consequently, both users and providers have a strong interest in earning a high reputation score. Precisely because users and providers have a substantial financial and personal interest in their scores, reputation systems may raise consumer protection issues. In particular, the design of reputation systems can inadvertently promote discriminatory behavior. For example, a recent study indicated that non-black hosts on Airbnb charge approximately twelve percent more than black hosts for similar listings.268 Since reputation systems can effectively—or literally269—lock out low-rated users and providers, the system design also presents a concern.270 Many platforms do not embed anti-abuse protections or quality control measures into reputation systems.271 Moreover, most platforms do not provide a

267. Seth Porges, Read These Tips, or Nobody Will Ever Let You Be an Airbnb Guest Again, GIZMODO (June 12, 2012, 11:00 AM), http://gizmodo.com/5918204/read-these- tips-or-nobody-will-ever-let-you-be-an-airbnb-guest-again. 268. Benjamin Edelman & Michael Luca, Digital Discrimination: The Case of Airbnb.com (Harv. Bus. Sch., Working Paper No. 14-054, 2014); see also Ian Ayres et al., Race Effects on eBay (Working Paper, 2011). 269. Some TNCs deactivate drivers who fall below a certain rating threshold. Jeff Bercovici, Uber’s Ratings Terrorize Drivers and Trick Riders. Why Not Fix Them?, FORBES (Aug. 14, 2014, 12:31 PM), http://www.forbes.com/ sites/jeffbercovici/2014/08/14/what-are-we-actually-rating-when-we-rate-other-people/. 270. On platforms where users search for providers in an open market, higher rated providers usually appear first on the list. By contrast, TNCs act as true matchmaking services rather than open market places, so the reputation system serves largely to flag bad actors on the system. But users and providers can still decline matches based on a low score or profile picture. Harry Campbell, Does Race Affect Your Driver Rating?, RIDESHARE GUY (Aug. 6, 2014), http://therideshareguy.com/does-race-affect-your- driver-rating/. 271. Tom Slee has noted that quantitative rating systems on TNC platforms have extremely skewed distribution patterns—users and providers almost universally rank each other at five stars, with a small percentage of one star ratings. Tom Slee, Some Obvious Things About Internet Reputation Systems, WHIMSLEY (Sept. 29, 2013), http://tomslee.net/2013/09/some-obvious-things-about-internet-reputation-systems.html. The score therefore functions more like a “thumbs up” or “thumbs down.” Some platforms intentionally rely on a “thumbs up” system. Thumbs Up., FANCY HANDS, https://www.fancyhands.com/ blog/thumbs-up/ (last visited Feb. 26, 2015). The highly personal interaction involved in many sharing services may discourage users and providers from giving honest feedback on the spot. See, e.g., Dan Harmon & Duncan Trussell, I Make A 5 Stars, HARMONTOWN, http://www.harmontown.com/2014/07/episode-108-i-make-a-5-stars/. In other cases, strategic behavior may motivate users and providers to “trade” positive ratings. Feedback Manipulation Policy, EBAY, http://pages.ebay.com/help/policies/feedback- manipulation.html (last visited Feb. 26, 2015).

2015] REGULATING THE SHARING ECONOMY 1119 clear forum for providers and users to challenge false or unfair reviews.272 Platforms do generally provide a complaint hotline for resolving small damages claims or other specific grievances, but some platforms have notably poor track records in responding to user and provider complaints.273 Of course, platforms have an interest in maintaining the integrity of their reputation systems, because accurate reviews can improve the quality of the platform and thus encourage more users to join.274 Some platforms have taken affirmative steps to improve the design of their reputation systems. Airbnb, for example, recently changed its review system to encourage honest feedback. Before the change, as soon as users and providers submitted a review, Airbnb would publish it. Users and providers could wait up to fourteen days to submit a review, and at the end of the review period the platform would automatically publish any drafted reviews.275 This system encouraged users and providers to avoid negative reviews, for fear of retaliation by the other party.276 Now Airbnb publishes reviews only after both users and providers have submitted a review. The system protects reviewers in that the other side cannot respond in kind to negative reviews.277

272. See, e.g., Alatraqchi v. Uber Techs., Inc., No. C-13-03156 JSC, 2013 WL 4517756 (N.D. Cal. Aug. 22, 2013) (dismissing a deactivated Uber driver’s wrongful termination claim); Can I Remove or Reply to a Review, AIRBNB, https://www.airbnb.com/help/article/32/. 273. Patrick Hoge, Uber and Lyft Both Get an “F” from the Better Business Bureau; So Does Yellow Cab, S.F. BUS. TIMES (Oct. 9, 2014, 10:55 AM), http://www.bizjournals.com/sanfrancisco/blog/2014/10/uber-lyft-f-better-business- bureau.html (noting that both TNCs and a major taxi dispatcher have received failing grades from the San Francisco Bay Area and Northern Coastal California Better Business Bureau for failing to respond to consumer complaints). 274. Jason Tanz, Is It Evil to Give a Bad Airbnb Review if Your Host Was Perfectly Nice?, WIRED (Nov. 21, 2013), http://www.wired.com/2013/11/qq_kia/ (“Bad reviews are ‘frankly foundational to the site,’ says Chip Conley, Airbnb’s head of hospitality.”). 275. Seth Porges, The Strange Game Theory of Airbnb Reviews, FORBES (Oct. 17, 2014), http://www.forbes.com/sites/sethporges/2014/10/17/the-strange-game-theory-of- airbnb-reviews/. 276. Maryam Saeedi has researched retaliatory feedback in a similar context based on eBay’s reputation system. Maryam Saeedi et al., The Value of Feedback: An Analysis of Reputation System (Working Paper, Nov. 6, 2013), available at http://econ.ohio- state.edu/saeedi/noneg_sss.pdf. 277. Seth Porges, The Strange Game Theory of Airbnb Reviews, FORBES (Oct. 17, 2014), http://www.forbes.com/sites/sethporges/2014/10/17/the-strange-game-theory-of- airbnb-reviews/.

1120 BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 30:385

Platforms can also take steps to combat discriminatory behavior through the design of reputation systems.278 While platforms cannot fully control user and provider discretion, altering the content or layout of individual profiles can affect conduct on the platform. Removing profile pictures or making those pictures less prominent can reduce opportunities for biased decision-making.279 Uber has addressed this problem in the context of underserved neighborhoods by preventing drivers from viewing rider destinations until after the driver has accepted a request.280 Platforms could even adopt stronger measures, such as using algorithmic filtering to detect unfair or biased reviewers.281 Although platforms can and should continue to protect users and providers through private ordering, regulators should still take an active role in monitoring reputation systems. Absent regulatory scrutiny, platforms have an incentive to keep reputation systems proprietary.282 Some do not even allow users to view their own reputation score.283 This prevents users and providers from discovering and addressing abuses of the rating system.284 Legislatures should thus consider enacting rules that require greater transparency.285 For example, the Vehicle for Hire Innovation Amendment Act in the District of Columbia requires that

278. Chrysanthos Dellarocas, Designing Reputation Systems for the Social Web (Bos. Univ. Sch. Mgmt., Research Paper Series No. 2010-18, 2010). 279. Benjamin Edelman & Michael Luca, Digital Discrimination: The Case of Airbnb.com (Harv. Bus. Sch., Working Paper No. 14-054, 2014). 280. Arielle, With Uber, Everyone Rides, UBER (Aug. 20, 2014), http://blog. uber.com/everyone-rides/ (last visited Feb. 26, 2015). 281. Greg Harman, The Sharing Economy Is Not as Open as You Might Think, GUARDIAN (Nov. 12, 2014), http://www.theguardian.com/sustainable- business/2014/nov/12/algorithms-race-discrimination-uber-lyft-airbnb-peer. 282. Jason Tanz, The Sharing Economy Needs to Start Sharing Its Data Too, WIRED (May 05, 2014, 6 :30 AM), http://www.wired.com/2014/05/sharing-economy-fico/. 283. Dara Kerr, Should Uber and Lyft Keep Passenger Ratings Secret?, CNET (Sept. 25, 2014, 7:00 AM), http://www.cnet.com/news/should-uber-and-lyft-keep-passenger- ratings-secret/. 284. PAM DIXON & ROBERT GELLMAN, THE SCORING OF AMERICA: HOW SECRET CONSUMER SCORES THREATEN YOUR PRIVACY AND YOUR FUTURE 14 (2014), available at http://www.worldprivacyforum.org/wp- content/uploads/2014/04/WPF_Scoring_of_ America_April2014_fs.pdf. 285. Danielle Keats Citron and Frank Pasquale have argued that automated scoring systems including reputation systems should be subject to procedural safeguards, granting consumers a form of “technological due process.” Citron & Pasquale, supra note 266, at 18–30.

2015] REGULATING THE SHARING ECONOMY 1121 users have access to their own reputation score.286 In addition, regulators should expressly prohibit discriminatory ratings, as in the Chicago TNC ordinance.287 Legislatures might consider adding language guaranteeing a private right of action under a state civil rights statute based on violation of the rating discrimination provision. 3. Promoting Competition: Regulators Should Ensure Provider Mobility and Protect Maverick Platforms Though regulators should take action to reduce new consumer protection concerns, they should also consider the potential risks of overly restricting competition. Thus far, regulators have focused on competition between sharing platforms and traditional service providers. In particular, the FTC has issued statements to both the Chicago legislature and the District of Columbia Taxicab Commission arguing that proposed ordinances for TNCs might impose “unwarranted restrictions.”288 Many have applauded TNCs for disrupting local taxi monopolies.289

286. D.C. CODE § 50-331(b)(8) (2015) (“[A] digital dispatch service may rate a customer so long as the customer’s rating may be viewed by the customer and may not be disclosed to a driver until after the driver accepts a ride request from that customer.”). 287. CHI., ILL., CODE § 9-115-140(b) (2014) (TNCs “shall train their drivers not to discriminate against people with disabilities in their passenger ratings. It shall be a violation of this chapter for a driver to rate a passenger based upon a disability.”). 288. Letter from Andrew I. Gavil, Dir. of Office of Policy Planning, FTC, to Brendan Reilly, Alderman 42nd Ward, City of Chicago (Apr. 15, 2014), available at http://www.ftc.gov/system/files/documents/advocacy_documents/ftc-staff-comment- honorable-brendan-reilly-concerning-chicago-proposed-ordinance-o2014- 1367/140421chicagoridesharing.pdf; Letter from Andrew I. Gavil, Dir. of Office of Policy Planning, FTC, to Jacques P. Lerner, Gen. Counsel, D.C. Taxicab Comm’n (June 7, 2013), available at http://www.ftc.gov/sites/default/files/documents/ advocacy_documents/ftc-staff-comments-district-columbia-taxicab-commission-concerning- proposed-rulemakings-passenger/130612dctaxicab.pdf. 289. E.g., Sharon Pian Chan, End the City’s Taxi Monopoly and Let Uber Roll, SEATTLE TIMES (Aug. 12, 2013), http://seattletimes.com/html/opinion/ 2021599082_sharonpianchancolumntaxisuber13xml.html; Jim Edwards, Uber is Destroying the Value of Taxi Monopolies in a Bunch of American Cities, BUS. INSIDER (Nov. 28, 2014, 7:27 AM), http://www.businessinsider.com/ uberdestroyingvalueoftaximonopoliescartels201411. Most cities have set strict safety regulations on licensing caps on local taxi industries since the early twentieth century. Some cities experimented with deregulation in during the 1970s, but these efforts often ran into serious problems. Today, economists continue to debate whether taxi deregulation is desirable, and if so what kind of deregulation. See generally Hardaway, supra note 66; Paul Stephen Dempsey, Taxi Industry Regulation, Deregulation & Reregulation: the Paradox of Market Failure, 24 TRANSP. L.J. 73 (1996); Harris, supra note 70.

1122 BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 30:385

Yet TNC market share may be even more concentrated than the taxi industry. A study by Future Advisor reported that Uber earned twelve times as much as its closest competitor, Lyft, in early 2014.290 To the extent that Uber attracts new users by providing superior services, this growth simply reflects “competition on the merits.”291 But dominant platforms can also distort markets by creating artificially high barriers to entry for new competitors292 and eliminating maverick firms.293 First, dominant platforms can raise the cost of entry for new firms by locking providers into their network. Sharing platforms benefit from indirect network effects—the more providers operate on the platform, the more valuable the service becomes for users.294 Platforms therefore compete to attract top providers, often by offering sign up bonuses or other incentives.295 Uber and Lyft, for example, have adopted aggressive

290. Chris Nicholson, Study: Uber Pulls Ahead of Lyft in Riders and Revenue With 12x Lead in U.S., FUTUREADVISOR (Sept. 11, 2014), http://blog.futureadvisor.com/study- uber-pulls-ahead-of-lyft-in-riders-and-revenue-with-12x-lead-in-u-s/. Uber CEO Travis Kalanick has claimed that Uber is twenty times larger than Lyft “on a booking basis” and does ten times as many trips. Liz Gannes, Travis Kalanick: Uber Is Raising More Money to Fight Lyft and the “Asshole” Taxi Industry, RE/CODE (May 28, 2014), http://recode.net/2014/05/28/travis-kalanick-uber-is-raising-more-money-to-fight-lyft- and-the-asshole-taxi-industry/. 291. Verizon Commc’ns Inc. v. Law Offices of Curtis v. Trinko LLP, 540 U.S. 398, 407 (2004) (“To safeguard the incentive to innovate, the possession of monopoly power will not be found unlawful unless it is accompanied by an element of anticompetitive conduct.”); see also South Park: Handicar (South Park Digital Studios 2014) (addressing a group of taxi drivers frustrated by competition from a fictional TNC named Handicar: “Why don’t you guys just make your cars cleaner and nicer and try to be better to your customers so that you can compete with Handicar’s popularity in the marketplace?”). 292. Under limited circumstances, a dominant firm can engage in actionable exclusionary conduct under section 2 of the Sherman Act by attempting to block “nascent” technologies from entering the market. U.S. v. Microsoft Corp., 253 F.3d 34, 79 (D.C. Cir. 2001). 293. A maverick is a firm that disproportionately affects the intensity of competition in a market in comparison to its market share. See David T. Scheffman & Mary Coleman, Quantitative Analyses of Potential Competitive Effects From a Merger, 12 GEO. MASON L. REV. 319, 321, 325–26 (2003). Removal of a maverick can constitute a theory of harm under the DOJ’s Merger Guidelines. U.S. DEP’T JUSTICE & FED. TRADE COMM’N, HORIZONTAL MERGER GUIDELINES § 2.1.5 (rev. ed. 2010), available at http://www.ftc.gov/sites/default/files/attachments/merger-review/100819hmg.pdf/. 294. See generally David S. Evans & Michael Noel, Defining Antitrust Markets When Firms Operate Two-Sided Platforms, 2005 COLUM. BUS. L. REV. 667 (2005). 295. Michelle Regner, Want to Be the Next Airbnb or Taskrabbit? Don’t fall for These Marketplace Myths, VENTUREBEAT (Dec. 14, 2014, 8:25 AM), http://venturebeat.com/2014/12/14/want-to-be-the-next-airbnb-or-taskrabbit-dont-fall- for-these-marketplace-myths/ (arguing that even if sharing businesses “won’t be buying traditional inventory per se, they will still be buying another type of inventory—

2015] REGULATING THE SHARING ECONOMY 1123 tactics for recruiting drivers.296 Some of these recruitment campaigns may constitute predatory pricing behavior.297 Such tactics potentially serve as a means of excluding competitor platforms or new entrants from the market.298 Second, dominant platforms can target maverick firms to reduce price competition.299 For example, some TNC and P2P car rental services specialize in trips to the airport.300 These rides are often more lucrative opportunities for drivers than short trips.301 These specialized services may offer lower prices than other TNCs, and therefore create procompetitive efficiencies in the market as a whole.302 In the long term, dominant TNCs

marketplace vendors”). Actively recruiting providers to join a platform is not anticompetitive behavior; in fact, agreements not to recruit may in some cases create antitrust liability. Jonathan Stempel, Apple, Google, Intel, Adobe to Pay $325 Million to Settle Hiring Lawsuit, REUTERS (May 23, 2014), http://www.reuters.com/article/ 2014/05/23/us-apple-google-settlement-idUSBREA4M0MY20140523/. 296. Serena Saitto, Uber Woos Drivers With $1,000 Bonuses in Tussle With Lyft, BLOOMBERG (July 13, 2014, 9:01 PM) http://www.bloomberg.com/ news/2014-07-14/uber-tussles-with-lyft-wooing-drivers-with-1-000-bonuses.html; Ellen Huet, How Uber And Lyft Are Trying To Kill Each Other, FORBES (May 30, 2014), http://www.forbes.com/sites/ellenhuet/2014/ 05/30/how-uber-and-lyft-are-trying-to-kill-each-other/. 297. Predatory pricing occurs where a firm (1) sets prices “below an appropriate measure of its rival’s costs,” and (2) the firm’s predatory pricing creates “a dangerous probability” of eliminating competition and ultimately allowing the firm to recoup losses through supra-competitive pricing. Brooke Grope Ltd. v. Brown & Williamson Tobacco Corp., 509 U.S. 209, 222–24 (1993). Some reports suggest that Uber has offered fares at below cost to drive Lyft out of certain markets. Ellen Huet, Uber’s Newest Tactic: Pay Drivers More Than They Earn, FORBES (July 2, 2014, 6:13 PM), http://www.forbes.com/ sites/ellenhuet/2014/07/02/ubers-newest-tactic-pay-drivers-more-than-they-earn/. 298. Hailo, an app that allows riders to request rides from partnering taxis and black cars, has shut down its operations in North America as a result of price competition from TNCs. Chris Welch, Hailo Is Leaving North America to Escape the War Between Uber and Lyft, VERGE (Oct. 14, 2014, 12:41 PM), http://www.theverge.com/2014/ 10/14/6975049/hailo-taxi-app-leaving-north-america/. 299. See supra text accompanying note 293. 300. WINGZ, https://wingz.me/ (last visited Feb. 26, 2015); FLIGHTCAR, https://flightcar.com/ (last visited Feb. 26, 2015) (P2P car rental service). 301. See, e.g., Fallon Glick, Louisville International Airport steps up enforcement of ridesharing ban, WDRB (Nov. 21, 2014, 3:28 PM), http://www.wdrb.com/story/ 27454057/louisville-international-airport-steps-up-enforcement-of-ridesharing-ban. 302. For example, Wingz charges a flat fee for airport trips. Carolyn Said, Wingz gets PUC green light as official TNC provider, SFGATE (Mar. 18, 2014), http://blog.sfgate.com/techchron/2014/03/18/wingz-gets-puc-green-light-as-official- tnc-provider/.

1124 BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 30:385 will have strong incentives to eliminate these mavericks by acquiring them,303 or by entering into exclusive contracts with airports or airlines.304 The DOJ and FTC have the authority to police such anticompetitive mergers and other predatory behaviors as needed, but regulators can also design rules to preemptively combat both of these threats.305 For example, states can require that airports award licenses for TNCs to access drop off points on a nondiscriminatory basis.306 Regulators can also enact tiered legislation, creating tax breaks or other allowances for small platforms. Most importantly, regulators should ensure that providers can freely move between platforms. In particular, the Chicago ordinance on TNCs may unnecessarily restrain drivers by prohibiting licensed taxis from operating as TNC vehicles.307 These restrictions may be necessary in limited circumstances to prevent consumer confusion with the trade dress of taxi dispatch services.308 However, ordinances can at least specify that providers have the right to contract with more than one TNC platform.309

303. Airbnb has acquired smaller competitor services to expand its service network. E.g., Dara Kerr, Airbnb Buys Crashpadder, Its Largest U.K. Competitor, CNET (Mar. 20, 2012, 7:38 PM), http://www.cnet.com/news/airbnb-buys-crashpadder-its-largest-u-k- competitor/. Fewer mergers have taken place between TNCs. Ellen Huet, Lyft Buys Carpooling Startup Hitch To Grow Lyft Line, FORBES (Sept. 22, 2014), http://www.forbes.com/sites/ellenhuet/2014/09/22/lyft-buys-carpooling-startup-hitch- to-grow-lyft-line/; Robert Hof, Uber CEO Travis Kalanick: We’re Not In Acquisition Mode (Sorry, Lyft), FORBES (Sept. 8, 2014, 1:42 PM), http://www.forbes.com/sites/roberthof/ 2014/09/08/uber-ceo-travis-kalanick-were-not-in-acquisition-mode-sorry-lyft/. 304. Joe Sharkey, United’s Deal With Uber Raises Concerns, N.Y. TIMES (Sept. 29, 2014), http://www.nytimes.com/2014/09/30/business/uniteds-deal-with-uber-raises- concerns.html. 305. In addition, regulators can consider deregulating or amending regulations for traditional for-hire services. Catherine Rampell, The Familiar Cycle of the Taxi Industry Wars, WASH. POST (Dec. 8, 2014), http://www.washingtonpost.com/opinions/catherine-rampell-thoughtful-taxi- regulations-should-consider-the-consumer/2014/12/08/d742cd76-7f19-11e4-8882- 03cf08410beb_story.html. 306. Airports in several cities have approved TNC permits for airport drop offs and pickups. Katherine Driessen, City to Allow Ride-Share Operators at Houston Airports, HOUSTON CHRONICLE (Nov. 12, 2014), http://www.chron.com/news/houston- texas/houston/article/City-to-allow-ride-share-operators-at-Houston-5888089.php. 307. See CHI., ILL., CODE § 9-115-100(c) (2014). 308. Dispatch services have argued that allowing TNCs to suggest an affiliation with licensed taxi drivers, or to allude to the distinctive trade dress of another for-hire transportation service constitutes a misrepresentation of association under the Lanham Act. Yellow Grp. LLC v. Uber Techs. Inc., No. 12 C 7967, 2014 WL 3396055, at *5 (E.D. Ill. July 10, 2014). 309. For example, the District of Columbia regulation provides that a “private or public vehicle-for-hire operator may affiliate with more than one company for the

2015] REGULATING THE SHARING ECONOMY 1125

Ultimately, the benefits of scale in some sharing markets may naturally lead to consolidation. However, taking appropriate steps to preserve a competitive marketplace can improve consumer choice, and encourage local and low-cost competitor platforms to enter the market. In enacting rules to protect consumers, regulators should also consider whether particular proposals would disproportionately affect large or small platforms. Nor should regulators shy away from addressing new threats to consumers, including privacy concerns and reputation systems. As Professor Cortez has argued, regulators should address regulatory disruptions through experimentation and prompt action.310

V. CONCLUSION Sharing platforms benefit consumers by increasing the availability of service providers, lowering costs, and providing altogether new services.311 In addition, these services do not pose an unacceptable danger to users or third parties. Many risks posed by the sharing economy are just as present in the market for traditional service providers.312 Regulators can also address these risks without forcing platforms to conform to the same rules as traditional service providers. In some cases, particularly the for-hire vehicle market, loosening restrictions for traditional service providers may be just as warranted as increasing protections for users and providers in the sharing economy.313 Though sharing platforms pose new consumer protection issues, regulators can confront these concerns without shutting down platforms.314 Yet even if one takes the position that the harms caused by the sharing economy outweigh the opportunities, most commentators would agree

purpose of using digital dispatch unless otherwise provided for by an agreement between the company and the operator.” D.C. CODE § 50-329.02(b)(15) (2015). 310. See generally Cortez, supra note 13. 311. Many regulators and commentators likewise adopt this position. See generally, e.g., BUS. INNOVATION OBSERVATORY, EUR. COMM’N, THE SHARING ECONOMY: ACCESSIBILITY BASED BUSINESS MODELS FOR PEER-TO-PEER MARKETS (2013), available at http://ec.europa.eu/enterprise/policies/innovation/policy/business- innovation-observatory/files/case-studies/12-she-accessibility-based-business-models-for- peer-to-peer-markets_en.pdf; WOSSKOW, supra note 243. 312. Supra Section IV.A. 313. E.g., Eric Roper, City Will Consider Major Reforms to Taxi Regulations, STAR TRIBUNE (June 30, 2014, 12:51 PM), http://www.startribune.com/politics/ statelocal/265132561.html (describing Minneapolis’s plans to revise taxi ordinances to more closely approximate new rules for TNCs). 314. Supra Section IV.B.

1126 BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 30:385 that the sharing economy is here to stay.315 In other words, we can no longer close Pandora’s box. However, regulators should not simply allow the sharing economy to grow in the shadow of the law. Allowing the sharing economy to self-regulate would not adequately safeguard consumers. Thus, responsible regulation of sharing platforms is a necessity, not a choice. Regulatory authorities and legislators have already begun to experiment with balanced solutions, and the outcomes of these efforts will continue to inform future regulations.

315. Even strong critics of the sharing economy tend to believe that the rise of the sharing economy is inevitable, or at least acknowledge the pervasiveness of that opinion. E.g., Catherine Rampell, The Dark Side of ‘Sharing Economy’ Jobs, WASH. POST (Jan. 26, 2014), http://www.washingtonpost.com/opinions/catherine-rampell-the-dark-side-of- sharing-economy-jobs/2015/01/26/4e05daec-a59f-11e4-a7c2-03d37af98440_story.html (describing the sharing economy as part of a larger—seemingly irreversible—trend of deregulation, most notably the decoupling of employment contracts from safety-net programs such as healthcare); see also Tom Slee, Why Canada Should De-Activate Uber, WHIMSLEY (Nov. 22, 2014), http://tomslee.net/2014/11/why-canada-should-de -activate-uber.html (“Contrary to the way some articles are written, we do have a choice here. A lot of the links above talk as if Uber were some kind of inevitable future. . . . Conflating Uber with the broad advance of technology is just wrong, and it’s also exactly what Uber wants us to do.”).

COM/MP1/avs Date of Issuance 9/23/2013

Decision 13-09-045 September 19, 2013

BEFORE THE PUBLIC UTILITIES COMMISSION OF THE STATE OF CALIFORNIA

Order Instituting Rulemaking on Regulations Relating to Passenger Carriers, Ridesharing, and Rulemaking 12-12-011 New Online-Enabled Transportation Services. (Filed December 20, 2012)

DECISION ADOPTING RULES AND REGULATIONS TO PROTECT PUBLIC SAFETY WHILE ALLOWING NEW ENTRANTS TO THE TRANSPORTATION INDUSTRY

77192335 - 1 - R.12-12-011 COM/MP1/avs

TABLE OF CONTENTS Title Page

DECISION ADOPTING RULES AND REGULATIONS TO PROTECT PUBLIC SAFETY WHILE ALLOWING NEW ENTRANTS TO THE TRANSPORTATION INDUSTRY ...... 2 Summary ...... 2 1. Procedural History ...... 4 2. Jurisdiction ...... 7 2.1. Comments on the Rulemaking ...... 8 2.2. Discussion ...... 11 2.2.1. Neither the Federal Telecommunications Act of 1996 nor Public Utilities Code Section 710 Exempts TNCs from State Jurisdiction ...... 12 2.2.2. TNCs Transport Passengers for Compensation ...... 18 2.2.3. TNCs Operate on a Prearranged Basis ...... 20 2.2.4. The Commission Has the Jurisdiction and the Duty to Establish Regulations Governing the Provision of TNC Services ...... 21 3. Safety ...... 35 3.1. Comments on the Rulemaking ...... 36 3.2. Discussion ...... 39 4. Ridesharing ...... 44 4.1. Comments on the Rulemaking ...... 44 4.2. Discussion ...... 48 5. Transportation Access ...... 52 5.1. Comments on the Rulemaking ...... 52 5.2. Discussion ...... 54 6. Insurance ...... 56 6.1. Comments on the Rulemaking ...... 57 6.2. Discussion ...... 58 7. Workshop Report...... 59 7.1. Discussion ...... 62 8. Comments on Proposed Decision ...... 63 9. Assignment of Proceeding ...... 64 Findings of Fact ...... 64 Conclusions of Law ...... 70 ORDER ...... 72

- i - R.12-12-011 COM/MP1/avs

DECISION ADOPTING RULES AND REGULATIONS TO PROTECT PUBLIC SAFETY WHILE ALLOWING NEW ENTRANTS TO THE TRANSPORTATION INDUSTRY

Summary This decision adopts rules and regulations for New Online Enabled Transportation Services, referred to hereafter as a Transportation Network Company1 (TNC), to ensure that public safety is not compromised by the operation of this new transportation business model. TNCs are not just Lyft, SideCar, InstantCab, and UberX.2 This Commission defines a TNC as an organization whether a corporation, partnership, sole proprietor, or other form, operating in California that provides prearranged transportation services for compensation using an online-enabled application (app) or platform to connect passengers with drivers using their personal vehicles.3 Among other

1 In the Rulemaking, we referred to these companies as New Online-Enabled Transportation Services (NOETS). We are changing the acronym to Transportation Network Company (TNC) for ease of use. 2 The Commission’s Safety and Enforcement Division issued cease and desist letters and $20,000 citations against Uber, Lyft, and SideCar for operating without authority and other violations of state law. However, in 2013, the Safety and Enforcement Division entered into settlement agreements intended to ensure the public safety of both riders and drivers with Uber, Lyft, and SideCar, allowing the companies to operate while the Commission’s TNC rulemaking is underway. http://www.cpuc.ca.gov/PUC/transportation/Passengers/CarrierInvestigations/.

3 There are eleven exemptions to the Passenger Charter-party Carriers’ Act contained in Public Utilities Code § 5353. Our definition of a TNC does not in any way usurp those existing exemptions. For example, one of the exemptions is passenger vehicles carrying passengers on a non-commercial enterprise basis. This exception has been defined by the Commission to mean non-profit organizations. See D.91.-06-025 (“The term ‘noncommercial enterprise basis’ in PU Code Section 5353(f) includes operations conducted on a not-for-profit, tax-exempt basis, as authorized by federal or state law.”). Another exemption is the rideshare exemption itself, which exempts: Transportation of

Footnote continued on next page

R.12-12-011 COM/MP1/avs

requirements established in this decision, we require each TNC (not the individual drivers) to obtain a permit from the California Public Utilities Commission (Commission), require criminal background checks for each driver, establish a driver training program, implement a zero-tolerance policy on drugs and alcohol, and require insurance coverage as detailed below. This decision orders a second phase to this proceeding to review the Commission’s existing regulations over limousines and other charter-party carriers to ensure that the public safety rules are up to date, and that the rules are responsive to the needs of today’s transportation market. In addition, the second phase will consider the potential impact of any legislative changes that could affect our ability to regulate the TNC industry. When the second phase is complete, the Commission will initiate the Commission’s resolution process to update the General Order (GO) 115 and 157 series to include the new regulations relating to the charter-party carrier subclass of TNC. Finally, the Commission is aware that TNCs are a nascent industry. Innovation does not, however, alter the Commission’s obligation to protect public safety, especially where, as here, the core service being provided -- passenger transportation on public roadways -- has safety impacts for third parties and property. The Commission is familiar with and confident in its ability to protect public safety in the face of rapid technological change. Consequently, while the Commission adopts these rules and regulations, it will

persons between home and work locations or of persons having a common work-related trip purpose in a vehicle having a seating capacity of 15 passengers or less, including the driver, which are used for the purpose of ridesharing, as defined in Section 522 of the Vehicle Code, when the ridesharing is incidental to another purpose of the driver.

- 3 - R.12-12-011 COM/MP1/avs

also look for further guidance from the legislature should it decide that there is a need for legislation to provide guidance in regulating this new industry. 1. Procedural History On December 20, 2012, the Commission opened this Rulemaking in order to determine whether and how TNC services arranged through online-enabled apps such as Uber, SideCar, and Lyft might affect public safety.4 In the Order Instituting Rulemaking (Rulemaking), the Commission stated that: We initiate this proceeding to protect public safety and encourage innovators to use technology to improve the lives of Californians.5 The purpose of this Rulemaking is not to stifle innovation and the provision of new services that consumers want, but rather to assess public safety risks, and to ensure that the safety of the public is not compromised in the operation of these business models. The Commission invites all interested parties to participate in this proceeding to ensure that regulation is not a hindrance, but continues to be the safety net that the public can rely on for its protection.6 The Commission sought comment on issues including: how the Commission’s existing jurisdiction should be applied to businesses such as Uber, SideCar, and Lyft; the consumer protection and safety implications of these new

4 The Commission’s Safety and Enforcement Division issued cease and desist letters and $20,000 citations against Uber, Lyft, and SideCar for operating without authority and other violations of state law. However, in 2013, the Safety and Enforcement Division entered into settlement agreements intended to ensure the public safety of both riders and drivers with Uber, Lyft, and SideCar, allowing the companies to operate while the Commission’s TNC rulemaking is underway. http://www.cpuc.ca.gov/PUC/transportation/Passengers/CarrierInvestigations/. 5 R.12-12-011, Rulemaking at 1. 6 R.12-12-011, Rulemaking at 2.

- 4 - R.12-12-011 COM/MP1/avs

methods for arranging transportation services; whether and how the new transportation business models differ from longstanding forms of ridesharing; and the new transportation business models’ potential effect on insurance and transportation access. On January 28, 2013, opening comments were filed by: Willie L. Brown, Jr., Luxor Cab Company, Greater California Livery Association, San Francisco Airport Commission, International Association of Transportation Regulators, Uber Technologies, Personal Insurance Federation of California (PIFC), Center for Accessible Technology (CforAT), Zimride, TransForm, SideCar Technologies, San Francisco Municipal Transportation Agency, Ed Healy, United Taxicab Workers, San Francisco Cab Drivers Association, Taxicab Limousine and Paratransit Association, and Taxicab Paratransit Association of California. On February 11, 2013, reply comments were filed by: Electronic Frontier Foundation, International Association of Transportation Regulators, United Taxicab Workers, Zimride, CforAT, Luxor Cab Company, San Francisco Municipal Transportation Agency, Transform, SideCar Technologies, Taxicab Paratransit Association of California, Ed Healy, Willie J. Brown, Jr., eRideshare, and San Francisco Cab Drivers Association. On February 15, 2013, the Commission held a Prehearing Conference in order to, inter alia, establish the service list, determine the positions of the parties, identify issues for inclusion in the April 2, 2013 Assigned Commissioner and Administrative Law Judge’s Scoping Memo and Ruling (Scoping Memo), and discuss the procedural schedule. Prehearing Conference Statements were filed by: United Taxicab Workers, International Association of Transportation Regulators, Willie J. Brown, Jr., Transform, Taxicab Paratransit Association of

- 5 - R.12-12-011 COM/MP1/avs

California, San Francisco Municipal Transportation Agency, Zimride, Uber Technologies, CforAT, and San Francisco Airport Commission. On March 7, 2013, the Administrative Law Judge (ALJ) issued a notice to the parties via e-mail, setting a workshop schedule and directing parties to file workshop statements answering specific questions about the following issues: TNC operations; jurisdiction; public safety; insurance; background checks; accessibility and equal access; and how Commission regulations may enhance or impede access to public roadways. On April 2, 2013, the assigned Commissioner and ALJ issued the Scoping Memo which established the scope and schedule of the Rulemaking, categorized the Rulemaking as quasi-legislative, and determined that hearings were not necessary. On April 3, 2013, workshop statements were filed by: Willie L. Brown, Jr., The Utility Reform Network, San Francisco Cab Drivers Association, Zimride, SideCar Technologies, TransForm, San Francisco Airport Commission and San Francisco Municipal Transportation Agency, Uber Technologies, Taxicab Paratransit Association of California, United Taxicab Workers, Luxor Cab Company, and CforAT. On April 10 and 11, 2013, the Commission held a workshop to facilitate dialogue among the parties on issues including: jurisdiction, public safety, accessibility, insurance, and proposed modifications for California statutes or Commission regulations. Two parties, TransForm and Taxicab Paratransit Association of California, took notes during the workshop and prepared a draft report summarizing all parties’ positions as articulated during the workshop. Parties reviewed the draft report to ensure that their positions were captured

- 6 - R.12-12-011 COM/MP1/avs

correctly, and on May 17, 2013, TransForm and Taxicab Paratransit Association of California filed the final workshop report with the Commission. On April 25, 2013, CforAT filed a motion requesting an additional round of comments on the issues raised in the Scoping Memo. On May 10, 2013, the ALJ granted the motion, determining that opening comments were due on June 3, 2013 and reply comments were due on June 10, 2013. On July 17, 2013, the California Highway Patrol (CHP) filed its comments.7 The purpose of this Rulemaking is not to stifle innovation and the provision of new services that consumers want, but rather to assess public safety risks, and to ensure that the safety of the public is not compromised in the operation of these business models. The Commission invited all interested parties to participate in this proceeding to ensure that regulation is not a hindrance, but continues to be the safety net that the public can rely on for its protection.8 2. Jurisdiction As noted in the Rulemaking,9 the Commission’s jurisdiction over charter-party carriers is clear. Nevertheless, new technology and innovation require that the Commission continually review its regulations and policies to ensure that the law and the Commission’s safety oversight reflect the current state of the industry and that these regulations are just and fair for all passenger carriers.

7 R.12-12-011, Rulemaking at 1. 8 R.12-12-011, Rulemaking at 2. 9 R.12-12-011, Rulemaking at 2-3.

- 7 - R.12-12-011 COM/MP1/avs

The Commission sought comment on how the Commission’s existing jurisdiction pursuant to the California Constitution and the Public Utilities Code (PU Code) should be applied to businesses like Uber, Sidecar, and Lyft and the drivers employed or utilized by these or similar entities. The Commission also sought comment on whether any existing legislation should be modified or if new legislation should be enacted. 2.1. Comments on the Rulemaking The parties that filed opening comments all addressed jurisdiction in varying degrees. The summaries of the positions of parties below capture all the positions that have been voiced in this Rulemaking on the subject of jurisdiction. The CHP asserts that TNCs fall under existing Commission jurisdiction, because the CHP views TNCs as for-hire passenger carriers.10 The CHP views a donation for transportation service equivalent to direct compensation, because the intent is to conduct a for-hire operation.11 Luxor Cab asserts that these businesses should be regulated the same as all other passenger carriers. Furthermore, it asserts that the presence of new technology for summoning a car does not in any way change the nature of the business that they are engaged in.12 Greater California Livery Association (GCLA) asserts that, based on their experience, these transportation technology companies should be subject to the same Commission regulation and enforcement as charter party carriers.13

10 California Highway Patrol comments filed on 07/17/13 at 1-2. 11 California Highway Patrol comments filed on 07/17/13 at 1. 12 Luxor Cab Opening Comments filed on 01/28/13 at 1. 13 GCLA Opening Comments filed on 01/28/13 at 2.

- 8 - R.12-12-011 COM/MP1/avs

Uber suggests that the Commission does not currently have jurisdiction over Uber because Uber is not a charter-party carrier within the meaning of PU Code § 5351 et seq. Further, Uber advocates against extending the Commission’s jurisdiction to companies like Uber because: 1) no public policy or public interest is advanced by such an extension of the law; 2) the Legislature has recently enacted new legislation exempting Internet Protocol-enabled (IP-enabled) services from regulation by the Commission; and 3) extending Commission regulation to Uber would conflict with Federal and State policies promoting further development of, and innovation in, information services provided over the Internet by prohibiting regulation of information services providers.14 TransForm acknowledges that the Commission has jurisdiction over charter-party carriers not meeting the statutory exemptions for taxicabs and work-related ridesharing, and has exercised this jurisdiction to ensure consumer protection and safety for traditional chartered transportation services.15 TransForm further asserts that the Commission should exercise its jurisdiction carefully so that it is applied in a way that allows growth of technology-enabled ridesharing services rather than eliminating an innovative tool to help address transportation access and climate change. The Commission should recommend to the legislature any necessary modifications to existing statutory exemptions to create a coherent regulatory framework that allows for ridesharing services to grow, while ensuring that consumer protection and safety is addressed. At the same time it is important for high-volume services to consult and coordinate

14 Uber Opening Comments filed on 01/28/13 at 5. 15 TransForm Opening Comments filed on 01/28/13 at 2.

- 9 - R.12-12-011 COM/MP1/avs

with local cities, counties, and public transit agencies to avoid potential impacts.16 The San Francisco Municipal Transportation Authority (SFMTA) says state law defines a charter-party carrier as any “person engaged in the transportation of persons by motor vehicle for compensation, whether in common or contract carriage, over any public highway in this state.”17 Drivers affiliated with businesses like Lyft and Sidecar drive passengers to destinations of their choice in exchange for payment. These businesses collect payments from passengers, share revenue with the drivers, and manage the exchange of information between passengers and drivers to facilitate interactions and commerce between drivers and passengers. SFMTA goes on to say that although certain transportation providers that would otherwise meet the definition of a “charter-party carrier” are exempted by statute from the Commission’s regulatory oversight, services like Lyft and SideCar do not fall within any of these exemptions.18 SideCar asserts that it is neither a charter-party carrier nor a transportation service, but rather it is a technology platform that facilitates exempt ridesharing and, to that extent, should be exempt from Commission jurisdiction under PU Code § 5353(f) and (h).19

16 TransForm Opening Comments filed on 01/28/13 at 4. 17 SFMTA Opening Comments filed on 01/28/13 at 2, citing PU Code § 5360. 18 SFMTA Opening Comments filed on 01/28/13 at 2. 19 SideCar Opening Comments filed on 01/28/13 at 9.

- 10 - R.12-12-011 COM/MP1/avs

Lyft asserts that the Commission should solely focus on regulation necessary to fulfill its responsibility for public safety.20 Lyft cautions the Commission to not force-fit existing regulations onto such an emerging industry. International Association of Transportation Regulators (IATR) recommends that the Commission should conduct further investigation to determine whether TNCs operate without a profit. IATR believes that companies that operate for-profit, and that use on-line apps that directly connect passengers to drivers, clearly fall under the Commission’s definition of a charter-party carrier, and should be subject to all the existing regulations.21 Taxicab Paratransit Association of California asserts that TNCs operate as on demand services and therefore fail to comply with the legal requirements for operation as a Transportation Charter Party (TCP).22 2.2. Discussion California law currently recognizes and regulates three modes of passenger transportation for compensation: taxi services, regulated by cities and/or counties; and charter-party carrier services, and passenger-stage companies, regulated by the Commission. In recent years, the communications revolution in wireless service, smartphones, and on-line apps has further facilitated the development and adoption of passenger transportation for compensation to a point where passengers seeking rides can be readily connected with drivers willing to provide rides in private vehicles. This

20 Zimride (Lyft) Opening Comments filed on 01/28/13 at 4. 21 IATR Opening Comments filed on 01/28/13 at 3. 22 TPAC Opening Comments filed on 02/04/13 at 5. The term TCP is defined and discussed, infra, in this Decision.

- 11 - R.12-12-011 COM/MP1/avs

development in passenger transportation for compensation, referred to in this proceeding as TNCs and associated with companies including UberX, Lyft, and Sidecar, does not fit neatly into the conventional understandings of either taxis or limousines, but that does not mean that this Commission’s responsibility to public safety in the transportation industry should be ignored and/or left for individual companies or the market place to control. 2.2.1. Neither the Federal Telecommunications Act of 1996 nor Public Utilities Code Section 710 Exempts TNCs from State Jurisdiction We reject Uber’s assertion that TNCs are nothing more than an application on smart phones, rather than part of the transportation industry. Uber is the means by which the transportation service is arranged, and performs essentially the same function as a limousine or shuttle company dispatch office. Accordingly, Uber is not exempt from the Commission’s jurisdiction over charter-party carriers. Nonetheless, because of the novelty of these new services, we will address Uber’s jurisdictional arguments here. As Uber notes in its comments, the 1996 Federal Telecommunications Act23 (FTA) distinguishes between “telecommunications” and “information services.” In so doing, Congress codified the Federal Communications Commission’s (FCC) historical determination that “basic” services were to be treated differently from “enhanced” services. Uber seeks to convince the Commission further with a detailed discussion of a Vonage case, in which the FCC concluded that nomadic Voice over Internet Protocol (VoIP) service is a purely interstate service, not subject to state jurisdiction. Uber recounts a California Court of Appeal case

23 P.L. No. 104-104, 110 Stat. 56 (1996).

- 12 - R.12-12-011 COM/MP1/avs

involving actions brought against eBay, where the court held eBay immune from state causes of action. In addition, Uber notes passage of Senate Bill 1161 in 2011 codified §§ 239 and 710 of the PU Code. Section 710 prohibits the Commission from “exercising any regulatory jurisdiction” over VoIP or IP-enabled services, subject to a delegation of federal authority, other express statutory authority, or exceptions contained in § 710. Uber’s citations are beside the point as none of the cited statutes or precedents prevent this Commission from regulating passenger transportation over public roadways. Specifically, we reject the argument that TNCs are simply providers of IP-enabled services and therefore exempt from our jurisdiction. We find this argument to be factually and legally flawed and, therefore, do not accept that the method by which information is communicated, or the transportation service arranged, changes the underlying nature of the transportation service being offered. First, the Commission is not attempting to enact rules that would impose regulations on the smart phone applications used to connect passengers with drivers. Instead, the Commission is promulgating rules that will govern the transportation service itself. Second, we do not believe that this Commission loses its jurisdiction over transportation services simply because a smart phone application is used to facilitate the transportation service. Nothing Uber has cited in California or federal law would mandate that result based on the facts here. Indeed Uber and Sidecar’s position would effectively obviate the Commission’s authority under PU Code § 5371.6(a) to prevent TCPs from operating illegally in order to protect the public and prevent unfair competition:

- 13 - R.12-12-011 COM/MP1/avs

The Legislature finds and declares that advertising and use of telephone service is essential for charter-party carriers of passengers to obtain business and to conduct intrastate passenger transportation services. Unlawful advertisements by unlicensed charter-party carriers of passengers has resulted in properly licensed and regulated charter-party carriers of passengers competing with unlicensed charter-party carriers of passengers using unfair business practices. Unlicensed charter-party carriers of passengers have also exposed citizens of the state to unscrupulous persons who portray themselves as properly licensed, qualified, and insured charter-party carriers of passengers. Many of these unlicensed charter-party carriers of passengers have been found to have operated their vehicles without insurance or in an unsafe manner, placing the citizens of the state at risk. Similarly, the Legislature has created additional safeguards in Government Code § 53075.8(b)(1) that allow for the termination of a taxicab’s telephone service if the taxi is operating without proper authority: The Legislature further finds and declares that the termination of telephone service utilized by taxicabs operating without proper authority is essential to ensure the public safety and welfare. Therefore, local agencies should take enforcement action, as specified in this section, to disconnect telephone service of unauthorized taxicab operators who unlawfully advertise passenger transportation services in yellow page directories and other publications. The enforcement actions provided for by this section are consistent with the decision of the California Supreme Court in Goldin v. Public Utilities Commission (1979) 23 Cal. 3d 638. We deem it is inconsistent with our grant of authority over transportation services to be barred from regulating a transportation service provided by TNCs based on the means of communication used to arrange the service.

- 14 - R.12-12-011 COM/MP1/avs

Moreover, to date neither the FCC, nor a court of higher jurisdiction, has ruled that this Commission, or any other state commission, is precluded by the FTA from regulating TNCs. It is interesting to note that the Federal Trade Commission (FTC) has intervened in state proceedings by filing comments but has not, to date, gone so far as to claim that state-regulatory efforts to assert jurisdiction over TNCs is preempted by the FTA. For instance, on June 7, 2013, the FTC sent a letter to General Counsel of the District of Columbia Taxicab Commission that offered comments in the proposed TNC-related rulemaking. Previously, the FTC filed comments in TNC-related rulemaking proceedings in Alaska24 and Colorado.25 Tellingly, neither the FTC nor the FCC has claimed that the state regulatory bodies are preempted from promulgating regulations to deal with the growing TNC business. In response to the proposed decision, Uber continued its argument by comparing itself to Google PowerMeter. In its August 19, 2013 comments to this decision, Uber stated that in the same way that Google did not become an energy utility by developing the Google PowerMeter software application, Uber does not become a transportation company by developing the Uber Software Application. The major difference between Uber and Google PowerMeter is that Uber controls the financial transaction between the customer and the company. Uber receives the customer fare and then transfers those funds to the driver

24 FTC comments dated April 19, 2013 to the Honorable Debbie Ossiander Concerning AO NO. 2013-36 Regarding the Regulatory Framework for the Licensing and Permitting of Taxicabs, Limousines, and Other Vehicles for Hire in Anchorage, Alaska. 25 FTC comments dated March 6, 2013 to the Colorado Public Utilities Commission In The Matter of the Proposed Rules Regulating Transportation by Motor Vehicle, 4 Code of Colorado Regulations 723-6.

- 15 - R.12-12-011 COM/MP1/avs

minus its share, while Google PowerMeter does not take any money from the customer. Google PowerMeter was a tool that allowed an electricity consumer to view his or her electricity usage. The data displayed by Google PowerMeter was measured by a measurement device installed by the customer with his or her consent. The goal of the Google PowerMeter was to inform the energy customers of their energy use, which could help the consumer identify ways to save energy. The customer was not charged a fare, and Google did not generate other revenues from the tool. If all Uber did was to show customers maps of available cars, without giving them a way to book a ride and without controlling or taking a share of the fare, then the analogy might be more appropriate. The Commission elects to use a more appropriate analogy involving Google. Google Search is an app and a software platform, and uses that software to provide a product: search listings. In 2011, Google agreed to pay a settlement of $500 million for allowing fraudulent pharmaceutical advertisements.26 In the case of pharmaceutical listings, Google Search was connecting people with products that were harmful or fraudulent, and which represented a threat to public safety. The people selling the illegal drugs had to be held accountable, but so did the software platform that connected people with the illegal drugs. The same is true with Uber. The Uber brand is now a known brand for car service. It is expected that a passenger requesting an Uber car will get a black town car or something of similar stature. It is expected that this service may cost more, but it is a higher service with professional drivers. Passengers may call Uber more frequently because of its name recognition . Uber by its name alone is selling a

26 See http://www.wired.com/threatlevel/2013/05/google-pharma-whitaker-sting/all/.

- 16 - R.12-12-011 COM/MP1/avs

type of car service. Because Uber is profiting from this service it should also be held responsible if the driver is negligent or not applying Uber safe practices. The same way Google was held responsible for allowing fraudulent advertisements is the same reason why Uber should be held responsible for its drivers. Uber argues that the taxi cabs and limousines that arrange rides on the Uber platform are already regulated and insured, and that no additional regulation of Uber itself is necessary to protect the public interest. Perversely, however, the fact that regulated forms of transportation arrange rides through the Uber platform injects a considerable degree of uncertainty into the question of whether a taxi cab or limousine’s insurance coverage would cover a claim. For example, if a limousine driver uses Uber’s method of fare calculation and billing rather than the method otherwise required by TCP rules or limousine company policy, in the event of an incident the limousine’s existing insurance policy may deny a claim on the grounds that the limousine had stopped operating, strictly speaking, and for insurance purposes, as a covered vehicle. In this same hypothetical incident, based on Uber’s comments in this proceeding, we anticipate that Uber would deny that it has any obligation to insure the parties injured in the accident, on the grounds that Uber is an app and the limousine driver was already insured. Until this Decision becomes effective, there is a real possibility that parties suffering losses in an incident would find that there is no insurance available to cover their potential claim. Due to the considerable uncertainty that exists concerning the insurance coverage applicable to rides (other than UberX rides) arranged through the Uber app, and the threat to public safety and well-being created by this uncertainty,

- 17 - R.12-12-011 COM/MP1/avs

the Commission is strongly inclined to require Uber to obtain a TCP permit in order to continue operating in California. As discussed elsewhere in this Decision, the Commission intends to open a second phase of this proceeding (Phase II) to consider the rules applicable to TCPs in California. In order to ensure the greatest possible evidentiary record, the Commission would prefer to leave all non-TCN issues, including Uber’s potential TCP status, to Phase II. However, the Commission will not allow the uncertainty regarding Uber’s insurance to persist during the pendency of Phase II. We require Uber to demonstrate to the Commission within 30 days of the issuance of this decision that it maintains commercial liability insurance policies providing not less than $1,000,000 (one million dollars) per-incident coverage for incidents involving vehicles and drivers in transit to or during trips arranged through the Uber app, the Commission reserves the right to require Uber to obtain a TCP permit through Commission resolution. while they are providing Uber services. The insurance coverage shall be available to cover claims regardless of whether an Uber driver maintains insurance adequate to cover any portion of the claim. 2.2.2. TNCs Transport Passengers for Compensation Public Utilities Code § 5360 states in part: Subject to the exclusions of Section 5353, “charter-party carrier of passengers” means every person engaged in the transportation of persons by motor vehicle for compensation, whether in common or contract carriage, over any public highway in this state. We reject the arguments made by Lyft and SideCar that any payment for rides arranged through their apps is voluntary and find that current TNCs are engaged in the transportation of persons for compensation. Although the phrase “for compensation” is not defined by PU Code § 5360, the plain-meaning

- 18 - R.12-12-011 COM/MP1/avs

interpretation of PU Code § 5360 in D. 69231 (June 15, 1965) informs our decision in this proceeding. In D.69231, a skate arena owner was ordered to cease and desist transporting passengers to his skate arena until he obtained his TCP certificate. While the record was unclear as to whether the owner would charge a fee for the proposed service, the Commission determined that even if the transportation was for free, “transportation furnished by business enterprises without charge is also ‘for compensation’ if the organization sponsoring the trip receives a business benefit.”27 The Commission reiterated this interpretation in D.81805 (August 28, 1973) where we reasoned that “it was not necessary for the staff to prove that respondent actually received money consideration for the transportation in question. It is enough that he received an economic benefit.”28 Clearly each TNC is receiving either an economic benefit or a business benefit. At a minimum, they are receiving increased patronage with the growth of their businesses. This possibility was an important factor for the Commission in rendering its decision in D.69231 that the skate arena owner’s status was a TCP: “Applicant would receive a business benefit and compensation from the

27 D.69231 at 409. 28 D.69231 at 493. The Commission has reached a similar conclusion with respect to free service provided by PSCs, finding that the service was for compensation. (See Peter J. Van Loben Sels (Valley Transit Lines) v. B.J. Smith et al., copartners (Cal. Transit Lines), 49 Cal. P.U.C. 290 (1950); and Richard Chala v. Morris Gordon of Gordon’s Outlet Store, et al., Decision No. 57356 in Case No. 6152 (1958), unreported. Our reasoning is also similar the Legislature’s when it added Section 17510.1 to the Business and Professions Code: “As used in this article, ‘sale’ shall include a gift made with the hope or expectation of monetary compensation.” Thus, a donation or a gift can still be considered a form of compensation.

- 19 - R.12-12-011 COM/MP1/avs

increased patronage for his skate arena business resulting from the advertising.”29 2.2.3. TNCs Operate on a Prearranged Basis Unlike taxi cabs, which may pick up passengers via street hails, PU Code § 5360.5 requires that charter party carriers operate on a prearranged basis. We find that TNCs operate on a prearranged basis. PU Code § 5360.5 does not define “prearranged,” and we are reluctant to impose a minimum time requirement as some other jurisdictions have done.30 Instead, we are guided by the plain meaning of “prearranged” as something arranged in advance, which has been our custom and practice in interpreting “prearranged” at the Commission. For example, our information packet for prospective TCP applicants says that all transportation performed by TCPs must be arranged beforehand, and the driver must have a completed waybill in his or her possession at all times during the trip.31 We believe TNCs satisfy the “prearranged” requirement in two ways: first, before a passenger can request a ride, the passenger must download the app and agree to the TNC service agreement. Examples can be found in the TNC written

29 409. 30 For example, the Washington Administrative Code requires that for-hire vehicles must be prearranged for at least 15 minutes. (Washington Rev. Code Section 308-83- 200.) The International Association of Transportation Regulators issued proposed model regulations for smartphone applications in the for-hire industry and suggested that the “prearranged or prearrangement” should require “a minimum of thirty (30) minutes between the request for transportation service and the arrival of the vehicle at the transportation origin location.” 31 Basic Information for passenger carriers and applicants (Rev. /28/11) issued by the Transportation License Section of the Commission.

- 20 - R.12-12-011 COM/MP1/avs

terms of use.32 Uber makes our point clearly in its description of its service that “persons who use the Uber App to request prearranged transportations have sole discretion over whether or not to use the Uber App, if ever.”33 Second, for a particular trip, the passenger must input information such as current location. A TNC driver cannot be hailed like a cab where no information is exchanged until the passenger enters the vehicle. As such, each TNC is offering transportation on a “prearranged” basis. Prearrangement has typically been verified through the use of a waybill. TCPs must possess a waybill for each ride that includes information on the driver’s name, vehicle license plate number, and time and date when the charter was arranged, and similar information.34 Pursuant to more recent legislation, waybills may be kept in an electronic format beginning January 1, 2014.35 In order to comply with the applicable statutes and regulations, all TNC drivers must be able to prove that a ride was matched on the TNC software application as evidence of prearrangement. In other words, information in the software application must be the equivalent of an electronic waybill. 2.2.4. The Commission Has the Jurisdiction and the Duty to Establish Regulations Governing the Provision of TNC Services Based on the record in this proceeding, and as the Rulemaking originally made clear, this Commission regulates charter party passenger carriers pursuant

32 See Exhibits B (Uber), D (SideCar), F (Lyft), and H (Tickengo) to the Workshop brief, filed on April 3 by TPAC. 33 Pre-Workshop Statement, 4, filed on April 3, 2013 by Uber. (Italics added.) 34 General Order 157-D, Part 3.01. 35 See PU Code § 5381.5.

- 21 - R.12-12-011 COM/MP1/avs

to Article XII of the California Constitution and the Charter-party Carriers’ Act, PU Code § 5351 et seq. (the Act). Section 5360 states in part: Subject to the exclusions of Section 5353, “charter-party carrier of passengers” means every person engaged in the transportation of persons by motor vehicle for compensation, whether in common or contract carriage, over any public highway in this state. Section 5381 states in part: …(t)he commission may supervise and regulate every charter-party carrier of passengers in the State and may do all things…necessary and convenient in the exercise of such power and jurisdiction. We are persuaded by the comments made by the CHP, TransForm, and to a certain extent Lyft. Our focus is public safety and secondarily ensuring that regulations reflect changing technology and ways of doing business to ensure that rules are in place to improve the lives of Californians. We agree with the CHP that a “donation” for passenger transportation service is equivalent to direct compensation for the service provided, which falls under the jurisdiction of this Commission. TransForm states in their comments in part: TransForm believes that all people deserve affordable, safe, and easy access to jobs, housing, services, and nature on foot, bicycle, or public transportation. TransForm envisions that in the future transportation will be redefined in terms of access and sustainability, and residents will be able to quickly get where they want to go in ways that fully meet their needs, whether these needs are health, happiness, saving time, or saving money. Our transportation system will provide the public with choices that amount to a system that is exceptional and state-of-the-art. TransForm believes that rideshare services have the potential to advance several California policy goals, including improving transportation access, reducing

- 22 - R.12-12-011 COM/MP1/avs

greenhouse gas emissions, reducing vehicle miles travelled, and reducing congestion. When the legislature passed the landmark transportation law SB 375 in 2008, the legislature found that “[w]ithout improved land use and transportation policy, California will not be able to achieve the goals of AB 32,” the Global Warming Solutions Act. The legislature also found that the transportation sector contributes over 40 percent of the greenhouse gas emissions in the State of California, the largest of any sector, with automobiles and light trucks alone contributing almost 30 percent. The California Air Resources Board, in setting regional greenhouse gas reduction targets, adopted targets requiring each region’s Sustainable Communities Strategy and Regional Transportation Plan to achieve specified reductions in the transportation sector by the years 2020 and 2035.36 We agree with TransForm with respect to the above two points. Additionally, Lyft has been the only TNC that has acknowledged that safety is not only a priority, but there should also be some overarching rules and regulations. We applaud Lyft for its leadership in this area and we certainly agree with Lyft in this area. For the reasons discussed supra, we find that TNCs are charter-party passenger carriers, and therefore we will exercise our existing jurisdiction pursuant to Article XII of the California Constitution and the Passenger Charter-party Carriers’ Act, PU Code §§ 5351, et seq. (the Act). Additionally, the Commission has very broad powers under PU Code § 701 which gives the Commission the ability (via a rulemaking process) to develop new categories of regulation when a new technology is introduced into an existing industry. In

36 TransForm Opening Comments filed on 01/28/13 at 1.

- 23 - R.12-12-011 COM/MP1/avs

this Decision, under the broad grant of authority pursuant to PU Code §§ 5381 and 701, we create the category of Transportation Network Company (TNC) to accompany the existing category of TCP.37 Again, a TNC is defined as an organization, whether a corporation, partnership, sole proprietor, or other form, operating in California that provides transportation services for compensation using an online-enabled app or platform to connect passengers with drivers using their personal vehicles. The primary distinction between a TNC and other TCPs is that a TNC connects riders to drivers who drive their personal vehicle, not a vehicle such as a limousine purchased primarily for a commercial purpose. To that end, a TNC is not permitted to itself own vehicles used in its operation or own fleets of vehicles. With this definition in mind, the Commission finds that Uber (in contrast to UberX) is not a TNC. Uber connects riders with drivers who do not drive their own personal vehicle, but typically operate in town cars or limousines, which the driver may often as well use to transport customers for another limousine/town car company. As such, Uber does not meet the definition of a TNC. As discussed elsewhere in this Decision, the Commission intends to open a second phase of this proceeding (Phase II) to consider the rules applicable to TCPs in

37 The Commission has previously developed new types of transportation services with unique rules relevant to that specific form of transportation. Namely, in D.97-07-063, the Commission “adopt[ed] rules for a new niche form of passenger stage corporation (PSC) that specializes in the common carriage of infants and children . . .” The Commission required such carriers to apply for a PSC permit, but developed a special set of rules applicable to these forms of transportation. D.97-07-063 stated, “This is a restricted class of PSC carrier not previously designated by this Commission, and special requirements need to be imposed on these carriers.” In creating these new rules, the Commission relied on its broad power under § 701, and the Passenger-Stage Corporation provisions of the Public Utilities Code § 5351.

- 24 - R.12-12-011 COM/MP1/avs

California. In order to ensure the greatest possible evidentiary record, the Commission would prefer to leave all non-TNC issues, including Uber’s potential TCP status, to Phase II. UberX, however, does meet the TNC definition and must apply for a TNC license. A company or individual wishing to provide transportation or facilitate transportation of passengers can choose to either get a TCP certificate/permit or a TNC permit.38 Further, TNCs need not apply for a certificate of public convenience and necessity pursuant to PU Code § 5371. TNCs are exempted from this requirement, as are many charter-party carriers regulated by the Commission, pursuant to PU Code § 5384(b), which authorizes the Commission to issue permits to passenger carrier operations who use only vehicles with seating capacities of under 15-passengers. TNC permits will only be granted to companies utilizing smart phone technology applications to facilitate transportation of passengers in the driver’s personal vehicle. Within 45 days after the effective date of this Decision, the Commission’s Safety Enforcement Division (SED) will post a TNC Application Packet on its website, and TNCs currently operating in California are required to file their TNC Applications with SED 60 days thereafter if they wish to continue operating. The TCP requirements are already in place, although as suggested supra the Commission will open a second phase to this Rulemaking to update those rules and regulations to ensure that safety requirements are up to date. Based on the record of this proceeding and the safety and other concerns expressed by parties, the settlement agreements that were entered into with Lyft,

38 There is also a third choice and that is to apply for a taxicab license.

- 25 - R.12-12-011 COM/MP1/avs

SideCar, and Uber, and our existing TCP rules we have created the following rules and regulations for all TNCs. The following rules and regulations shall be applied for all TNCs effective immediately: Safety Requirements a) TNCs shall maintain commercial liability insurance policies providing not less than $1,000,000 (one million dollars) per-incident coverage for incidents involving vehicles and drivers while they are providing TNC services. The insurance coverage shall be available to cover claims regardless of whether a TNC driver maintains insurance adequate to cover any portion of the claim.39 b) TNC drivers shall be required to provide proof of both their personal insurance and the commercial insurance in the case of an accident. c) TNCs shall perform criminal background checks on each TNC driver before the driver begins offering service. In order to protect public safety, any person who has been convicted, within the past seven years, of driving under the influence of drugs or alcohol, fraud, sexual offenses, use of a motor vehicle to commit a felony, a crime involving property damage, and/or theft, acts of violence, or acts of terror shall not be permitted to provide TNC services. d) TNCs shall institute a zero tolerance intoxicating substance policy with respect to drivers as follows: 1. The TNC shall include on its website, mobile application and riders’ receipts, notice/information on the TNC’s zero-tolerance

39 TNCs must make their certificate of insurance public and the Commission will put this certificate on its website.

- 26 - R.12-12-011 COM/MP1/avs

policy and the methods to report a driver whom the rider reasonably suspects was under the influence of drugs or alcohol during the course of the ride. 2. The website and mobile application must include a phone number or in-app call function and email address to contact to report the zero-tolerance complaint. 3. Promptly after a zero-tolerance complaint is filed, the TNC shall suspend the driver for further investigation. 4. The website and mobile application must also include the phone number and email address of the Commission’s Passenger Section: 1-800-894- 9444 and [email protected]. e) TNCs shall obtain each TNC driver’s driving record before the driver begins providing service and quarterly thereafter. Drivers with convictions for reckless driving, driving under the influence, hit and run, or driving with a suspended or revoked license shall not be permitted to be a TNC driver. Drivers may have a maximum of two points on their driving records for lesser offenses, e.g., equipment problems, speeding, or child safety seat violations. f) TNCs shall establish a driver training program to ensure that all drivers are safely operating the vehicle prior to the driver being able to offer service. This program must be filed with the Commission within 45 days of the adoption of this decision. TNCs must report to the Commission on an annual basis the number of drivers that became eligible and completed the course. g) TNC drivers must possess a valid California driver’s license, be at least 21 years of age, and must provide at least one year of driving history before providing TNC services.

- 27 - R.12-12-011 COM/MP1/avs

h) TNCs may only use street-legal coupes, sedans, or light-duty vehicles including vans, minivans, sport utility vehicles (SUVs) and pickup trucks. Hatchbacks and convertibles are acceptable. i) TNC drivers are prohibited from transporting more than 7 passengers on any given ride.40 j) The app used by a TNC to connect drivers and passengers must display for the passenger: 1) a picture of the driver, and 2) a picture of the vehicle the driver is approved to use, including the license plate number to identify the vehicle. k) TNC vehicles shall not be significantly modified from factory specifications, e.g., no “stretch” vehicles. l) Prior to allowing each TNC driver to operate a vehicle, and annually thereafter, a TNC must inspect the driver’s vehicle, or have the vehicle inspected at a facility licensed by the California Bureau of Automotive Repair, and maintain complete documentation of such inspections. A TNC driver’s vehicle must, at a minimum, pass a 19 point inspection prior to allowing the driver to operate the vehicle under the TNC’s platform: 1. Foot brakes; 2. Emergency brakes; 3. Steering mechanism; 4. Windshield;

40 If a TNC elects to carry insurance up to $1.5 million per incident for all of its drivers, then pursuant to PU Code § 5391 and General Order 115-F, the TNC vehicles can include up to 10 people including the driver. However, no TNC driver is permitted to operate a bus, which is defined by California Vehicle Code § 233(b) as “a vehicle designed, used, or maintained for carrying more than 10 persons, including the driver, which is used to transport persons for compensation or profit . . .”

- 28 - R.12-12-011 COM/MP1/avs

5. Rear window and other glass; 6. Windshield wipers; 7. Headlights; 8. Tail lights; 9. Turn indicator lights; 10. Stop lights; 11. Front seat adjustment mechanism; 12. Doors (open, close, lock); 13. Horn; 14. Speedometer; 15. Bumpers; 16. Muffler and exhaust system; 17. Condition of tires, including tread depth; 18. Interior and exterior rear view mirrors; and 19. Safety belts for driver and passenger(s). Regulatory Requirements For all reports identified below required to be provided by TNCs, the reports must be verified. Verification consists of provision of a signature of a corporate officer of the TNC verifying under penalty of perjury under the laws of the State of California that the report is accurate and contains no material omissions. a. TNCs (not the drivers) must be permitted by this Commission before operating as a TNC.41 b. TNCs shall clearly disclose, on their app and website, that TNCs facilitate rides between

41 There are six types of charter party carrier permits/certificates. TNCs shall apply for a class P permit.

- 29 - R.12-12-011 COM/MP1/avs

passengers and private drivers using their own personal vehicles. Additionally, the disclosure should state that each TNC is required to maintain insurance policies providing a minimum of $1,000,000 (one million dollars) per-incident coverage for incidents involving vehicles and drivers while they are providing TNC services. c. TNC drivers may only transport passengers on a prearranged basis. For the purpose of TNC services, a ride is considered prearranged if the ride is solicited and accepted via a TNC digital platform before the ride commences. TNC drivers are strictly prohibited from accepting street hails. d. TNCs shall participate in the California Department of Motor Vehicle’s Employer Pull Notice Program to obtain timely notice when any of the following are added to a TNC driver’s driving record: i. Convictions; ii. Accidents; iii. Failures to appear; iv. Driver’s license suspension or revocation; and v. Any other action taken against the driving privilege. e. TNCs shall obtain proof of insurance from each TNC driver before the driver begins providing service and for as long as the driver remains available to provide service. f. TNCs shall allow passengers to indicate whether they require a wheelchair-accessible vehicle or a vehicle otherwise accessible to individuals with disabilities. g. One year from the effective date of these rules and annually thereafter, each TNC shall submit to the Safety and Enforcement Division a report detailing the number and percentage of their customers who

- 30 - R.12-12-011 COM/MP1/avs

requested accessible vehicles, and how often the TNC was able to comply with requests for accessible vehicles. h. TNC vehicles shall display consistent trade dress (i.e., distinctive signage or display on the vehicle) when providing TNC services that is sufficiently large and color contrasted as to be readable during daylight hours at a distance of at least 50 feet. The trade dress shall be sufficient to allow a passenger, government official, or member of the public to associate a vehicle with a particular TNC (or licensed transportation provider). Acceptable forms of trade dress include, but are not limited to, symbols or signs on vehicle doors, roofs, or grills. Magnetic or removable trade dress is acceptable. TNC shall file a photograph of their trade dress with the Safety and Enforcement Division. i. Although TNCs may provide platforms allowing drivers and passengers to “rate” each other, TNCs shall ensure that such ratings are not based on unlawful discrimination, and that drivers do not discriminate against passengers or potential passengers on the basis of geographic endpoints of the ride, race, color, national origin, religion, sex, disability, age, or sexual orientation/identity. j. One year from the effective date of these rules and annually thereafter, each TNC shall submit to the Safety and Enforcement Division a verified report detailing the number of rides requested and accepted by TNC drivers within each zip code where the TNC operates; and the number of rides that were requested but not accepted by TNC drivers within each zip code where the TNC operates. The verified report provided by TNCs must contain the above ride information in electronic Excel or other spreadsheet format with information, separated by columns, of the date, time, and zip code of each request and the concomitant date, time, and zip code

- 31 - R.12-12-011 COM/MP1/avs

of each ride that was subsequently accepted or not accepted. In addition, for each ride that was requested and accepted, the information must also contain a column that displays the zip code of where the ride began, a column where the ride ended, the miles travelled, and the amount paid/donated. Also, each report must contain information aggregated by zip code and by total California of the number of rides requested and accepted by TNC drivers within each zip code where the TNC operates and the number of rides that were requested but not accepted by TNC drivers. k. One year from the effective date of these rules and annually thereafter, each TNC shall submit to the Safety and Enforcement Division a verified report in electronic Excel or other spreadsheet format detailing the number of drivers that were found to have committed a violation and/or suspended, including a list of zero tolerance complaints and the outcome of the investigation into those complaints. Each TNC shall also provide a verified report, in electronic Excel or other spreadsheet format, of each accident or other incident that involved a TNC driver and was reported to the TNC, the cause of the incident, and the amount paid, if any, for compensation to any party in each incident. The verified report will contain information of the date of the incident, the time of the incident, and the amount that was paid by the driver’s insurance, the TNC’s insurance, or any other source. Also, the report will provide the total number of incidents during the year. l. One year from the effective date of these rules and annually thereafter, each TNC shall submit to the Safety and Enforcement Division a verified report

- 32 - R.12-12-011 COM/MP1/avs

detailing the average and mean number of hours and miles each TNC driver spent driving for the TNC. 42 m. Upon request, drivers shall display to Commission or airport enforcement officers, law enforcement, or city or county officials a physical or electronic record of a ride in progress sufficient to establish that it was prearranged. To the extent that trip records are contained on electronic devices, TNC drivers are not required to relinquish custody of the devices in order to make the required display. n. If a passenger files a complaint against a TNC or TNC driver with the Commission, Commission staff shall have the right to inspect TNC records and vehicles as necessary to investigate and resolve the complaint to the same extent the Commission and Commission staff is permitted to inspect all other charter-party carriers. o. Operations at Airports. TNCs shall not conduct any operations on the property of or into any airport unless such operations are authorized by the airport authority involved. p. Similar to our regulations over limousines one-third of one percent of the total revenues from TNC services in California shall be collected by this Commission on a quarterly basis as part of overall fees. The Commission will convene a workshop one year after the issuance of this decision to hear from all stakeholders on the impacts of this new mode of transportation and the accompanying regulations. Workshops topics will

42 For the requested reporting requirements, TNCs shall file these reports confidentially unless in Phase II of this decision we require public reporting from TCP companies as well.

- 33 - R.12-12-011 COM/MP1/avs

include, but not necessarily be limited to, a consideration of safety, competition, innovation, accessibility, congestion, the California Environmental Quality Act, and other pollution related issues. Specifically, the Commission will be interested to get an update on TNCs’ commercial insurance policies and how these policies have performed. The Commission may choose to open a new proceeding to update its rules based on the information learned in this workshop. TNCs that fail to adhere to these requirements may have their permits revoked or be otherwise subject to sanctions by the Commission. The Commission is authorized to conduct inspections of charter-party carriers including TNCs. For instance, PU Code § 5371.5 states that: “Upon receipt of a complaint containing sufficient information to warrant conducting an investigation, the commission shall investigate any business that advertises limousine-for-hire or passenger charter transportation service for compensation in motor vehicles.” Therefore, each TNC must keep records of all trips made by its TNC drivers. The Commission is also authorized to “cancel, revoke, or suspend any operating permit or certificate” if the carrier violates any of the provisions of the Act, provisions of the operating permit or certificate issued thereunder, or any order, decision, rule, regulation, direction, demand, or requirement established by the Commission.43 The Commission is also authorized to issue fines.44 Sections 5411 to 5420 of the Act contain relevant provisions regarding issuing fines and penalties. In addition, the Commission has established a

43 PU Code § 5378. 44 See e.g., PU Code § 5378(b).

- 34 - R.12-12-011 COM/MP1/avs

citation program in Resolution ALJ-187, which provides a process by which the Commission may issue fines, carriers may appeal fines, and the Commission may hold a hearing pursuant to that appeal. These provisions authorizing the Commission to inspect, investigate, and issue fines and other penalties apply in equal measure to all TNCs as they do to other charter-party carriers. Therefore, the Commission must have access to a TNC’s records whenever it requests them. Parties have raised a number of concerns regarding the Terms & Conditions used by certain TNCs, which include general disclaimers of liability. No Term & Condition in a TNC’s Terms of Service or elsewhere, can be inconsistent with this decision. Nor can any Term & Condition in a TNC’s Terms of Service be used or relied on by the TNC to deny insurance coverage, or otherwise evade the insurance requirements established in this decision. Moreover, the Terms of Service does not absolve the TNC of its responsibilities to comply with the stated regulations in this decision to ensure safety of the public. As stated earlier in this decision, the Commission will open a Phase II to consider updating its regulations over TCP certificate holders. Phase II will also consider the standard and appropriate language for Terms & Conditions for both TCP and TNC certificate holders. 3. Safety The Commission opened this proceeding to protect public safety and secondarily encourage innovators to use technology to improve the lives of Californians. The Commission has a responsibility for determining whether and how public safety might be affected by these TNCs. In opening this Rulemaking,

- 35 - R.12-12-011 COM/MP1/avs

the Commission wanted to assess public safety risks, and to ensure that the safety of the public is not compromised in the operation of TNCs. 3.1. Comments on the Rulemaking As with the issue of jurisdiction a number of parties filed comments about the effect of TNC service on public safety. In this section we will summarize all the positions filed. The CHP asserts that it is too early to determine the effect of this type of service on both the passengers and public safety. It goes on to caution, however, that passenger transportation left unregulated unnecessarily increases the potential for operation of unsafe vehicles, unqualified drivers, and uninsured transportation drivers.45 Luxor Cab’s comments focus more on the need to keep drivers safe. Luxor Cab asserts that taxicab drivers have the highest risk of occupational homicide of all US occupations, and that this is why taxi regulators require safety equipment such as bullet-resistant partitions and digital security cameras, as well as crime- prevention training for drivers.46 The GCLA believes that the transportation technology companies can put the public at risk of potential dangers arising from having unregulated and perhaps even unlicensed drivers and unsafe vehicles providing for-hire transportation services without oversight or enforcement.47 The San Francisco Airport Commission believes that lack of adequate liability insurance, criminal background checks, driver training and regular

45 CHP Comments filed on 7/17/13 at 2. 46 Luxor Cab Comments filed on 01/28/13 at 2. 47 GCLA Comments filed on 01/28/13 at 2.

- 36 - R.12-12-011 COM/MP1/avs

vehicle inspections all decrease public safety, and although some TNCs represent that they do all of the above, the Airport Commission is asking for regulatory verification.48 The SFMTA asserts that TNCs have a negative effect on public safety because of a lack of regulatory oversight. The SFMTA asserts that at the state and local level, California regulators of taxi and limousine service protect the public with the following kinds of requirements: 1. Criminal background checks of drivers; 2. Drug and alcohol testing of drivers; 3. DMV “pull notice” checks to enable suspension of drivers with new safety related moving violations; 4. Driver training for local geography, traffic safety and customer service values; 5. Vehicle age and mileage limitations; 6. Routine, professional vehicle inspections; and 7. Transparent pricing regulations.49 The San Francisco Cab Drivers Association asserts that the proliferation and acceptance of private vehicles and unlicensed public passenger drivers for hire creates a false sense of trust by the general public. Furthermore, it asserts that they are witnessing private vehicles being flagged down and soliciting passengers on the street which will result in an assault or worse, on a passenger or a driver, unprotected by security cameras, dispatch or a shield, and no readily identifiable markings on the vehicle.50

48 San Francisco Airport Commission Opening Comments filed on 01/28/13 at 2. 49 SFMTA Opening Comments filed on 01/28/13 at 8. 50 San Francisco Cab Association’s Opening Comments filed on 01/29/13 at 2.

- 37 - R.12-12-011 COM/MP1/avs

In their comments, Lyft notes that ridesharing is nothing new and has been occurring on a relatively large scale for many decades – from casual carpools and bulletin boards to more recent on-line forums – without any regulation and with few if any institutional safety mechanisms. Lyft goes on to say that rather than creating a new activity requiring scrutiny as a public safety concern, responsible peer-to-peer platforms such as Lyft have introduced innovative and highly effective institutional safety mechanisms that increase public safety over existing alternatives. New tools made available by modern technologies – online criminal background checks, mobile application photo identification, and Global Positioning System (GPS) positioning – can advance public safety beyond existing measures.51 SideCar asserts that TNCs are mission-driven and have strong incentives to protect the trust and safety of their communities and the public. SideCar goes on to claim that its safety program and rules aim to reduce and prevent accidents or other incidents, and it has implemented a 10-point safety program to create a safe experience for drivers and riders alike. Under this safety program, all drivers are required to undergo thorough background checks and safety training.52 United Taxicab Workers assert that TNCs provide service through non- professional drivers of private vehicles, and since they claim that they are not regulated by the state or local authorities, the public can only take the word of the company. United Taxicab Workers go on to note that safety is the paramount concern in the taxi regulation and that taxis are inspected regularly and are

51 Lyft Opening Comments filed on 01/28/13 at 4-5. 52 SideCar Opening Comments filed on 01/28/13 at 17.

- 38 - R.12-12-011 COM/MP1/avs

subject to age and mileage requirements. Furthermore, drivers receive training and must go through background checks prior to becoming a taxi driver.53 In its comments, TPAC asserts that the primary reason for regulation of the passenger transportation industry is the need to ensure safety. It goes on to say that public safety is promoted through the screening of drivers, and by ensuring that those who take on the responsibility of transporting passengers can be held accountable for their actions.54 3.2. Discussion We agree that protecting and enhancing public safety is the paramount purpose behind regulating this industry. We initiated this Rulemaking for the sole purpose of determining how TNCs affect public safety. We further agree with the CHP, the San Francisco Airport Commission, the SFMTA, and other parties who have urged us to adopt safety rules and regulations that will hold TNCs accountable for safety. We also agree with Lyft that ridesharing is nothing new and has been occurring on a relatively large scale for many decades – from casual carpools and bulletin boards to more recent on-line forums. We note, however, that there is a specific exemption for the true form of ridesharing in the PU Code. PU Code § 5353(h) exempts: Transportation of persons between home and work locations or of persons having a common work-related trip in a vehicle having a seating capacity of 15 passengers or less, including the driver, which are used for the purpose of ridesharing, as defined in Section 522 of the Vehicle Code, when the ridesharing is incidental to another purpose of the driver.

53 United Taxicab Workers Opening Comments filed on 01/29/13 at 4-5. 54 TPAC Opening Comments filed on 02/04/13 at 6.

- 39 - R.12-12-011 COM/MP1/avs

The section also states: This exemption does not apply if the primary purpose for the transportation of those persons is to make a profit. “Profit,” as used in this subdivision does not include the recovery of actual costs incurred in owning and operating a vanpool vehicle, as defined in Section 668 of the Vehicle Code. In our view the Commission firmly believes that TNCs do not meet the rideshare exemption and actually are providing transportation services for compensation. Lyft and SideCar have both entered into settlement agreements with the Commission’s Safety and Enforcement Division as stated above and have complied with the safety requirements in those agreements. Therefore, it is not entirely correct to state (as some parties have in their comments) that the public must only rely on the company’s word. These agreements, however, are interim arrangements pending the conclusion of this Rulemaking. Therefore, in this decision we adopt strict safety regulations and guidelines that are similar in nature and in some cases more stringent than current and past practice in the transportation industry as a whole. The regulations for TNCs will require the company to conduct criminal background checks, establish a driver training program, maintain a zero-tolerance policy on drugs and alcohol, register in the Department of Motor Vehicle (DMV) Pull Notice program, conduct a 19-point car inspection, and require a one-year driving history from the driver. These regulations along with other requirements are stated above in the summary section as well as the jurisdiction section. Regarding the criminal background checks, we will require each TNC to conduct a criminal background check for each driver prior to that applicant becoming a TNC driver. The criminal background check must be a national

- 40 - R.12-12-011 COM/MP1/avs

criminal background check including the national sex offender database. The criminal background check should be using the applicant’s social security number and not just the applicant’s name. Any felony criminal conviction within seven years prior to the date of the background check for violent crime, a sexual offense, a crime involving property damage, and/or theft will make the applicant ineligible to be a TNC driver. Regarding the 19-point vehicle inspection, we require the TNC or an authorized third party facility licensed by the California Bureau of Automotive Repair to conduct the car inspections and for the TNC to maintain the record of such inspections in case of an audit. Regarding the DMV Pull Notice Program, we are aware that the California DMV does not currently permit TNCs to enroll non-employee drivers in the Employer Pull Notice Program. We are also aware that it was established to provide employers and regulatory agencies with a means of promoting driver safety through the ongoing review of driver records. An employer enrolled in the program is assigned a requester code. The requester code is added to an employee's driver license (DL) record. When an employee's DL is updated to record an action/activity, a check is made electronically to determine if a pull notice is on file. If the action/activity is one that is specified to be reported under the program, a driver record is generated and mailed to that employer. The DMV Pull Notice program allows a transportation company to monitor DL records of employees. This monitoring accomplishes the following:

 Improves public safety;  Determines if each driver has a valid DL;  Reveals problem drivers or driving behavior; and  Helps to minimize the transportation company’s liability.

- 41 - R.12-12-011 COM/MP1/avs

The Commission began enrolling owner operators into this program in 1990. We are similarly hopeful that the DMV is able to amend the requirements of the program to allow TNCs to participate automatically in the program once they have completed the other requirements for the driver to begin providing service. Specifically, we encourage the DMV to modify the language about employers being the only entity to qualify for this automatic service. We understand that currently TNCs can manually enter into the program, but automatic enrollment improves public safety in that the notification to TNCs will be automatic and timely. We are hoping to work with the DMV to find a solution that improves public safety as we have added new rules and regulations to allow TNCs to provide transportation services. Until the DMV Employer Pull Notice Program is available for use by TNCs, TNCs shall perform, prior to allowing a driver on the platform and quarterly thereafter, driving record checks through the DMV in order to ensure that drivers meet applicable requirements. The DMV check criteria shall provide that a user may have no more than three points within the preceding three years, no “major violations” (reckless driving, hit and run, or driving with a suspended license conviction) within the preceding three years, and no driving under the influence conviction within the past seven years. Regarding the accessibility plan which each TNC is required to file within 45 days of the issuance of this decision, each plan shall include the following: a. A timeline for modifying apps so that they allow passengers to indicate their access needs, including but not limited to the need for a wheelchair accessible vehicle. A passenger should be allowed to state other access needs, either from a drop-down menu with room for comments or through a field requesting information.

- 42 - R.12-12-011 COM/MP1/avs

b. A plan for how the TNC will work to provide appropriate vehicles for passengers who specify access needs, including but not limited to a plan to provide incentive to individuals with accessible vehicles to become TNC drivers. c. A timeline for modifying apps and TNC websites so that they meet accessibility standards. The relevant standard for web access is WCAG 2.0 AA. Guidance on accessibility standards for iPhone apps can be found at http://developer.apple.com/library/ios/documentation/ UserExperience/Conceptual/iPhone Accessibility and http://developer.apple.com/library/ios/documentation/ UserExperience/Conceptual/iPhone Accessibility/Making Application Accessible/Making Application Accessible.html. Guidance on accessibility standards for Android apps can be found at http://developer.android.com/training/accessibility/acce ssible-app.html. d. A timeline for modifying apps so that they allow passengers to indicate that they are accompanied by a service animal, and for adopting a policy that service animals will be accommodated. e. A plan for ensuring that drivers’ review of customers will not be used in a manner that results in discrimination, including any policies that will be adopted and any monitoring that will take place by the TNC to enforce this requirement. Each aspect of the accessibility plan will be addressed in the annual reports required of each TNC regarding compliance, necessary improvements (if any) and additional steps to be taken by the TNC to ensure that there is no divide between service provided to the able and disabled communities. These reports will be served by SED on the service list for this proceeding, and input from interested parties will be invited. Based on SED’s review of the annual reports as well as input from interested parties, the Commission will determine what, if

- 43 - R.12-12-011 COM/MP1/avs

any, changes need to be made in the TNC business model, or new regulations adopted, in order to ensure that TNCs are accessible to, and do not discriminate against, persons with disabilities. 4. Ridesharing The definition of ridesharing does not permit transportation performed for profit.55 Recovery of actual costs incurred only applies to vanpool vehicles, which is defined by the Vehicle Code as seating more than 10 passengers, but less than 15 passengers, including the driver. The Commission sought comment on whether the TNCs’ business models qualify as ridesharing for the purpose of the PU Code § 5353(h) exemption and, with respect to its passenger carrier regulation, whether the Commission should recommend a broader or narrower definition of ridesharing than that contained in the California Vehicle Code. 4.1. Comments on the Rulemaking Various parties filed comments in response to the questions asked in the Rulemaking. This section will summarize all the various positions. We may not cite every party that filed comments, but we will cite every position. Opening comments filed by former San Francisco Mayor Willie L. Brown Jr. proposes a mandatory cap on TNC driver earnings and an updated definition that includes this cap in the PU Code § 5353 (f).56 These comments further state that the issue for sites such as Tickengo and 511.org is that there is no clear definition of vehicles carrying passengers on a noncommercial enterprise basis, and that a clear definition of ridesharing would help eliminate confusion with TCPs, fill empty seats in cars, and reduce pollution and congestion while

55 Rulemaking at 7. 56 Comments from Willie Brown filed on 01/18/13 at 1-2.

- 44 - R.12-12-011 COM/MP1/avs

lowering the cost of door-to-door transportation.57 Tickengo proposes that we limit the maximum share-the-expense carpool amount drivers can collect on a yearly basis to the American Automobile Association’s (AAA) official annual cost of vehicle ownership (currently $8,776 per year).58 Luxor Cab, on the other hand, asserts that the statutory definition of ridesharing is adequate, but what is lacking is compliance with regulations by unlicensed for-hire TNCs.59 Luxor Cab further comments that legitimate ridesharing does not include the transportation of a passenger on a trip the driver was not otherwise planning to take. Luxor asserts that it is the very nature of taxicab service that the ride is offered on demand and in accordance with the passenger’s desired location. Finally, Luxor Cab comments that the amount of compensation should not determine the need for compliance with regulations, but rather it is the nature of the service that ought to be determinative.60 The SFMTA asserts that there is no reason for the Commission to change the definition of ridesharing under the Vehicle Code in order to accommodate for-profit transportation services delivered through smartphone applications. It further asserts that there is nothing about the ‘new business model’ of offering for-hire transportation services through the mechanism of a smartphone application that justifies abandoning the fundamental regulatory infrastructure of the transportation for-hire industry, or that changes the level of regulatory concern when members of the public place themselves in the care and control of

57 Comments of Willie Brown filed on 01/18/13 at 2. 58 Comments of Willie Brown filed on 01/18/13 at 3. 59 Luxor Cab comments filed on 01/28/13 at 3. 60 Luxor Cab comments filed on 01/28/13 at 3.

- 45 - R.12-12-011 COM/MP1/avs

a private individual who they pay to carry them safely to their destination in a motor vehicle over the public right of way.61 Lyft asserts that the Commission is reading the PU Code too narrowly and recommends that the Commission explicitly acknowledge and clarify that: 1) a voluntary donation, regardless of the amount, does not constitute “compensation” as the term is used in § 5360 and that 2) the “primary purpose” of any driver that only receives voluntary donations from riders and no other pay from the company operating the rideshare platform is not to make a “profit,” as defined in § 5353(h). Lyft also suggests that the Commission consider recommending that the Legislature clarify or broaden the definition of ridesharing.62 SideCar urges the Commission to clarify the rideshare exemption in PU Code § 5353(h) and establish a bright line “safe harbor” for ridesharing drivers and authentic peer-to-peer rideshare technology providers. It goes on to say that while the Public Utilities Code currently has no provision for the recovery of the costs incurred in owning and operating a vehicle, except a vanpool vehicle, SideCar believes that a standard should be adopted for ridesharing in regular passenger vehicles.63 The San Francisco Cab Drivers Association asserts that businesses like Sidecar and Lyft clearly do not qualify for exemption from charter carrier laws under the definition of ridesharing as defined in § 522 of the Vehicle Code. This transportation is not between home and work locations or of persons having a

61 SFMTA comments filed on 01/28/13 at 9. 62 Lyft comments filed on 01/28/13 at 7. 63 SideCar comments filed on 01/28/13 at 11.

- 46 - R.12-12-011 COM/MP1/avs

common work-related trip. The sole purpose of these trips is to convey passengers to their requested destination, for profit.64 IATR asserts that while the PU Code exempts from regulation passenger vehicles that carry passengers on a “noncommercial enterprise basis,” this term is not defined. It goes on to say that TNCs fail to meet the definition for ridesharing (as they operate outside of strictly work and home locations, and transport passengers on trips that are NOT incidental to the driver) and fail to qualify for the Commission exemption because they are operating for profit/compensation.65 IATR further suggests that the definition of ridesharing be narrowed whereas Lyft says that the Commission is reading the definition too narrowly. IATR says that the Commission should act to clarify the regulatory exemption and to make clear that to qualify for the exemption, a driver is prohibited from making any profit and/or accepting compensation.66 The CHP asserts that the term “ridesharing” is a term-of-art within the lexicon of transportation – notwithstanding the vehicle used, ridesharing is essentially deemed to be reserved for like-minded individuals with a transportation motivation incidental to another purpose and not seated in profit-making derived from the transportation.67

64 San Francisco Cab Drivers Association comments filed on 01/28/13 at 3. 65 IATR Comments filed on 01/28/13 at 4. 66 IATR Comments filed on 01/28/13 at 5. 67 CHP comments filed on 7/17/13 at 4-5.

- 47 - R.12-12-011 COM/MP1/avs

4.2. Discussion We agree with the vast majority of the parties that filed comments that TNCs do not qualify for the rideshare exemption under PU Code § 5353(h). PU Code § 5353(h) exempts from Commission regulation: Transportation of persons between home and work locations or of persons having a common work-related trip purpose in a vehicle having a seating capacity of 15 passengers or less, including the driver, which are used for the purpose of ridesharing, as defined in Section 522 of the Vehicle Code, when the ridesharing is incidental to another purpose of the driver. This exemption also applies to a vehicle having a seating capacity of more than 15 passengers if the driver files with the commission evidence of liability insurance protection in the same amount and in the same manner as required for a passenger stage corporation, and the vehicle undergoes and passes an annual safety inspection by the Department of the California Highway Patrol. The insurance filing shall be accompanied by a one-time filing fee of seventy-five dollars ($75). This exemption does not apply if the primary purpose for the transportation of those persons is to make a profit. "Profit," as used in this subdivision, does not include the recovery of the actual costs incurred in owning and operating a vanpool vehicle, as defined in Section 668 of the Vehicle Code.68

68 Vehicle Code § 522 defines “ridesharing” as “two or more persons traveling by any mode, including, but not limited to, carpooling, vanpooling, bus pooling, taxi pooling, jitney, and public transit.”

- 48 - R.12-12-011 COM/MP1/avs

Section 5353(h) provides two opportunities to qualify for the rideshare exemption: Transportation of persons between home and work locations or of persons having a common work-related trip purpose in a vehicle having a seating capacity of 15 passengers or less, including the driver, which are used for the purpose of ridesharing, as defined in Section 522 of the Vehicle Code, when the ridesharing is incidental to another purpose of the driver. TNCs fail to satisfy either of these requirements. In our review of the filings and supporting documents, there is no evidence that TNC drivers have a common work-related or incidental purpose with their passengers. Instead, drivers transport passengers entirely at the convenience of the passenger: Lyft is recruiting drivers with the following language: “Be a Lyft Driver” material states that “drivers are making up to $35/hour + choosing their own hours!”69 Uber’s service is defined as “your on-demand private driver.”70 SideCar offers the following pitch to its prospective drivers: “Drive where you want, when you want, and who you want. You are your own boss. Some of our SideCar drivers are earning $30+ per hour.”71 InstantCab tells prospective drivers that it makes “it easy for customers and cab drivers to find each other. We’re looking for drivers to help us launch and provide high quality service to anyone who needs a taxi. We’re not a taxi company, you

69 http://www.lyft.me/drivers. 70 Exhibit A, 34, Workshop Brief, filed by TPAC on April 3, 2013. 71 Exhibit C, 48, Workshop Brief, filed by TPAC on April 3, 2013.

- 49 - R.12-12-011 COM/MP1/avs

can work for any existing taxi company and use our app to find guaranteed customers.”72 Tickengo tells its prospective drivers that they can “accept any ride if you want to go to the same destination, or if you just want to help.”73 Services provided by TNCs are thus very different from traditional, longstanding forms of ridesharing.74 TNCs are clearly designed to provide a car service for compensation. There is no requirement that there be a common purpose. Instead, TNCs operate as an alternative to other traditional car services. Several parties in comments on the proposed decision expressed concern that the proposed decision would, as former San Francisco Mayor Brown described in his comments, limit the ability of “a regular citizen [to] request a ride from a family member who may wish to give them a ride to the airport for free.”75 Similarly, eRideShare, which has provided an online carpool matching service since 1999, expressed concerns that the proposed decision would override existing statutory exemptions for ridesharing services.76 These concerns are ill founded. We reiterate that our Decision in no way impacts the exemptions in Section 5353 of the Public Utilities Code. To the extent that services such as Rideshare meet

72 https://instantcab.wordpress.com/join/. 73 https://tickengo.com/a/becomedriver/. (Italics added.) 74 The TNCs should be contrasted with http://www.511.org, a ridesharing service which is managed by a partnership of public agencies led by the Metropolitan Transportation Commission, the California Highway Patrol, and the California Department of Transportation. There are no references to Terms and Conditions, donations, and other forms of compensation. 75 Comments on Proposed Decision – from former San Francisco Mayor Willie L. Brown Jr. on 8/12/2013. 76 Final Opening Comments of eRideShare Inc. on 08/19/2013.

- 50 - R.12-12-011 COM/MP1/avs

either the “non-commercial enterprise” or rideshare exemption under Section 5353, or other exemptions as applicable, such services would be exempt from Commission regulation. The Commission has never regulated the ability of a “regular citizen [to] request a ride from a family member who may wish to give them a ride to the airport for free,” and nothing in the Public Utilities Code or our Decision would extend the Commission’s jurisdictional reach to such lengths. Further, the Commission would again note that the basis for regulating TNCs is that they meet the definition of a charter-party carrier under the Public Utilities Code. That is, they are “engaged in the transportation of persons by motor vehicle for compensation.”77 We agree with SFMTA that there is no reason for the Commission to change the definition of ridesharing under the Vehicle Code in order to accommodate for-profit transportation services delivered through smartphone applications. Furthermore, there is nothing about the ‘new business model’ of offering transportation services for compensation through the mechanism of a smartphone application that justifies abandoning the fundamental regulatory infrastructure of the transportation for compensation industry, or that changes the level of regulatory concern. The underlying principal continues to be ensuring public safety. Regulation is the safety net that the public should rely on for its protection. We are not persuaded by the TNCs that would like us to create a regulatory gap because they are using a smartphone to facilitate transportation for compensation. We are, however, encouraged by the TNC’s embrace of

77 PU Code § 5360 (emphasis added).

- 51 - R.12-12-011 COM/MP1/avs

technology and innovation to bring choice and convenience to the public in a safe manner. 5. Transportation Access The Commission’s authority over passenger carriers is grounded in the need to protect the public’s safe and reliable access to California’s roadways. Section 5352 of the Act states: The use of the public highways for the transportation of passengers for compensation is a business affected with a public interest. It is the purpose of this chapter to preserve for the public full benefit and use of public highways consistent with the needs of commerce without unnecessary congestion or wear and tear upon the highways; to secure to the people adequate and dependable transportation by carriers operating upon the highways; to secure full and unrestricted flow of traffic by motor carriers over the highways which will adequately meet reasonable public demands by providing for the regulation of all transportation agencies with respect to accident indemnity so that adequate and dependable service by all necessary transportation agencies shall be maintained and the full use of the highways preserved to the public; and to promote carrier and public safety through its safety enforcement regulations. PU Code § 5352 places public safety as a key goal in ensuring that the public enjoys full access to the roadways. In this Rulemaking the Commission sought comment on the ways that safety regulations may enhance or impede public access to the roadways. 5.1. Comments on the Rulemaking Many parties filed comments in response to this issue and there were some that remained silent. We will summarize those positions that were submitted in this section.

- 52 - R.12-12-011 COM/MP1/avs

Luxor Cab asserts that unlicensed for-hire carriers such as Uber, Lyft, and SideCar do not invest in safety equipment and crime-prevention training for drivers. It goes on to say that TNCs and their drivers try to compensate for the lack of professional safety measures by cherry-picking the customers whom they believe are safest to convey. Luxor Cab then cautions that the result of this type of cherry-picking is de facto red-lining of low-income neighborhoods and discrimination against customers based on drivers’ profiling that may be little more than stereotyping according to ethnicity or disability. Luxor Cab also says that such practices are illegal for licensed operators because they have the effect of reducing public access to the roadways.78 The CHP asserts that the Commission’s oversight responsibilities relative to transportation access are rooted in two essential areas. First, the regulation of accident indemnity to ensure adequate and dependable service by transportation operators and preservation of full use of the highways; and secondly, to promote public and operator safety through enforcement regulations.79 Perhaps the most detailed and focused comments on this issue came from Center for Accessible Technology (CforAT). CforAT rightly reminds us that any demand-response transit service must also comply with state and federal anti- discrimination statutes, including requirements that such services be accessible to people with disabilities.80 San Francisco Cab Drivers Association asserts that they have personally witnessed an abundance of Lyft and other private vehicles transporting people in

78 Luxor Cab opening comments filed on 01/28/13 at 3-4. 79 CHP comments filed on 07/17/13 at 3. 80 CforAT comments filed on 01/28/13 at 1-2.

- 53 - R.12-12-011 COM/MP1/avs

the back seat, blocking up traffic and making illegal maneuvers, while legal taxicabs drive around empty. They go on to say that this adds to traffic congestion. Additionally, the assertion is made that a Lyft driver nearly ran into the individual head-on while making an illegal left turn across Van Ness Avenue in San Francisco onto California Street and a professional driver would not do that.81 5.2. Discussion We agree with CforAT that TNCs must endeavor to provide equal access to all consumers. Because TNCs are in their infancy we cannot determine at this point whether equal access is being hampered. As a threshold matter, TNCs must do the following: a. TNCs shall allow passengers to indicate whether they require a wheelchair-accessible vehicle or a vehicle otherwise accessible to individuals with disabilities. b. One year from the effective date of these rules and annually thereafter, each TNC shall submit to the Safety and Enforcement Division a report detailing the number and percentage of their customers who requested accessible vehicles, and how often the TNC was able to comply with requests for accessible vehicles. Upon receipt this report shall be made public by the Safety and Enforcement Division. This report shall also contain a description of any instances or complaints of unfair treatment or discrimination of persons with disabilities. The above information will be used by the Commission to determine what, if any, changes need to be made to the regulations in order to ensure that TNCs are accessible to, and do not discriminate against, persons with disabilities. The

81 San Francisco Cab Drivers Association comments filed on 01/29/13 at 3-4.

- 54 - R.12-12-011 COM/MP1/avs

Commission also notes it currently has few provisions or protections to ensure equal access for passengers with disabilities under its current TCP regulations.82 Updating any regulations in this area, as found to be needed, may also be something the Commission should consider in Phase 2 of this rulemaking. We also agree with the CHP that the Commission must regulate TNCs to ensure adequate and dependable service by transportation operators and to promote public and operator safety. Consequently, we require TNCs to follow the safety and regulatory requirements stated above in section 3.2 of this decision. And we also agree with Luxor Cab that discrimination against customers based on drivers’ profiling that may be little more than stereotyping by ethnicity, disability, or economic class, will not be tolerated. It is noteworthy that, although not a party to this proceeding, Homobiles was created to serve a community that may not have been adequately served by the existing transportation forms. According to Homobiles’ website, it was formed to serve underserved communities who experience stress or discrimination on various forms of transportation for hire due to their gender or sexual identity.83 The Commission notes that while some parties argue that TNCs such as Lyft, UberX, and SideCar must be regulated either as taxi cabs or limousines in order to ensure nondiscrimination and public safety, Homobiles was formed to meet the needs of consumers whose transportation needs are not being adequately met by

82 For instance, the Commission requires every carrier to maintain on file with the Commission an equipment list of all vehicles in use including whether each vehicle is handicap accessible. (GO 157-D, Section 4.01.) 83 http://www.homobiles.org/terms/.

- 55 - R.12-12-011 COM/MP1/avs

either taxi cabs or limousines. We applaud the founders of Homobiles for establishing a non-profit 501(c)(3) volunteer organization that caters to the underserved communities of San Francisco. We agree with CforAT that the Commission should be informed by the legacy of transit discrimination and should work to ensure that the new services mark a break from this problematic history. Just as it would be unacceptable to allow any form of transit service to operate if it were to engage in racial discrimination, new forms of online-enabled transit services cannot be permitted to exclude people with disabilities. We agree. Therefore, we direct TNCs to submit a plan within 90 days of the effective date of this decision to tell us how they plan to ensure that TNCs will avoid creating a divide between the able and disabled communities. TNCs must explain how they plan to provide incentives to individuals with accessible vehicles to become TNC drivers. Furthermore, TNCs should ensure accessibility accommodations for their apps and websites to enable the disabled public access to the same services as clients who are not disabled.84 6. Insurance California Insurance Code § 11580.1(b) requires that non-commercial vehicles have a minimum liability coverage of $15,000 for injury/death to one person, $30,000 for injury/death to more than one person, and $5,000 for damage to property. The Commission’s GO 115-F requires that any charter party carrier vehicle with a seating capacity of seven passengers or fewer have a minimum

84 Title III of the Americans with Disabilities Act (ADA) requires that businesses and nonprofit services providers make accessibility accommodations to enable the disabled public to access the same services as clients who are not disabled.

- 56 - R.12-12-011 COM/MP1/avs

commercial coverage of $750,000. In the Rulemaking, the Commission sought comments on, inter alia, the insurance aspects of this new transportation model. For instance, if a vehicle is insured as a private vehicle, but involved in an incident while transporting passengers for compensation, the Rulemaking asked what type of coverage would the insurance offer for injuries/damages to the driver, the paying passenger, and any other people or property involved in the incident, and whether the insurance industry had an opinion on the insurance coverage available for private vehicles used to transport passengers for compensation. 6.1. Comments on the Rulemaking This Rulemaking has at least 18 parties who filed comments. No party claimed that TNCs should not have insurance or that liability insurance in the transportation business was not a key component of their business model. In this section we will note the PIFC’s comments.85 We also note that many parties claimed either in their comments or during the workshop that TNCs are uninsured. In its comments, PIFC asserts that it surveyed its member insurance companies, finding that “the industry standard for personal auto insurance policy contracts is to exempt from insurance coverage claims involving vehicles used for transporting passengers for a charge.”86 PIFC goes on to say that in situations where a vehicle is insured as a private vehicle and is used to transport

85 According to comments filed by PIFC on 01/28/13, the PIFC members represent six of the nation’s largest insurance companies (State Farm, Farmers, Liberty Mutual Group, Progressive, Allstate and Mercury) which collectively write a majority of the personal lines of auto insurance in California. 86 PIFC comments filed on 01/28/13 at 1-2.

- 57 - R.12-12-011 COM/MP1/avs

passengers for a fee, no insurance coverage would exist.87 The Commission also inquired about the sufficiency of the minimum liability coverage required under California Insurance Code § 1158.1(b). PIFC asserts that since there would be no coverage for the type of situations at issue, the minimum amount of coverage would be irrelevant.88 Finally, with respect to California Insurance Code § 11580.24, PIFC notes that the legislature encouraged car sharing programs (i.e., renting out one’s personal vehicle to another driver), as long as the owner does not earn more than the annual cost of owning the vehicle from the car sharing program. PIFC goes on to say that in doing so, it shields private passenger car insurers from any liability by shifting the responsibility for coverage to the private vehicle ridesharing program. The PIFC notes that the issue before the Commission is not ridesharing, but instead it is one of using a private passenger vehicle in a livery service. This is clearly not covered under a standard policy; if an incident occurs, coverage would not exist.89 6.2. Discussion We will require TNCs to maintain commercial liability insurance policies providing not less than $1,000,000 (one million dollars) per-incident coverage for incidents involving vehicles and drivers while they are providing TNC services. The insurance coverage shall be available to cover claims regardless of whether a TNC driver maintains insurance adequate to cover any portion of the claim. This level of liability insurance is above what the Commission currently requires of

87 PIFC comments filed on 01/28/13 at 1-2. 88 Id. 89 Id.

- 58 - R.12-12-011 COM/MP1/avs

TCP drivers. It is equal to the insurance that the SFMTA requires of taxicab companies. We reject the claim that Lyft, SideCar, and Uber/UberX do not have insurance. The Commission’s Safety & Enforcement Division, in entering into settlement agreements with these entities, made sure that each of these companies maintained excess liability insurance policies providing a minimum of $1 million per incident. We note PIFC’s comments in this Rulemaking, and note that, even if a TNC driver’s personal insurance does not apply in the event of an accident, the insurance required by the Commission will apply. We require that each TNC file their insurance policies under seal with the Commission as part of applying for a license. Furthermore, the license for the TNC will automatically expire upon expiration of the insurance policy unless and until the TNC provides an updated insurance policy and applies to renew its license. In Phase II of this proceeding we will consider whether these policies for both TCP as well as TNC certificate holders should be made public and included in the Commission’s website. 7. Workshop Report As part of the Scoping Memo, parties were invited to attend a workshop to consider issues including but not limited to jurisdiction, safety, transportation access, and proposed modifications to existing rules and regulations. On April 10 and 11, 2013, the parties attended the Commission’s workshop in San Francisco at the Commission’s offices. The workshop sessions were publicly noticed and open to the public. Two parties that we’d like to thank and extend our appreciation to for drafting the workshop report are TPAC and TransForm. On May 17th these two parties filed the Workshop Report on behalf of those parties who attended

- 59 - R.12-12-011 COM/MP1/avs

the workshop.90 The Workshop Report summarizes party positions as articulated during the workshop. Most of the issues such as jurisdiction, safety, access, and the definition of ridesharing have already been discussed in the above sections of this decision. There are, however, two issues not addressed above that we will address in this section. During the workshop, Commission staff asked whether there was a third way to regulate TNCs that protected public safety, but also allowed innovation and technology to bring choice and convenience to the public. The SFMTA/IATR stated that the idea that there is some third way to regulate these TNCs is offensive to the men and women who work as regulators to protect public safety and access. The SFMTA/IATR pointed out that the taxi industry is a highly managed transportation network that requires regulations to ensure universal access to door to door transportation in an urban environment.91 TPAC stated that it believed that the Commission had inappropriately provided preapproval to a third-way regulatory approach via its settlement agreements with companies such as Uber and Lyft. TPAC stated that the third-way regulatory approach affected by the TNCs’ settlement agreements amounted to the deregulation of the taxicab industry, and as such violated state law.92 Counsel for the SFMTA and the San Francisco Airport Commission stated that

90 TPAC, TransForm, CforAT, GCLA, Luxor Cab, IATR, PIFC, the San Francisco Cab Drivers Association, the San Francisco Limo Union, the San Francisco Medallion Association, SFMTA, The San Francisco Airport Commission, SideCar, Tickengo, Uber, The United Taxicab Workers, TURN, and Lyft. 91 Workshop Report at 14. 92 Id.

- 60 - R.12-12-011 COM/MP1/avs

TNCs have presented no credible argument for a third way. The SFMTA and San Francisco Airport Commission stated that there are two possible regulatory schemes, the local system for taxicabs and the state system for charter-party carriers, but there is no justification for subjecting TNCs to lesser standards than those applicable to all other charter party carriers.93 Luxor Cab stated that the topic of a third way to regulate TNCs is misleading because it assumes that there is something new about the TNCs, when taxi companies have been using similar technological services for several years before the inception of Uber, Lyft, and SideCar.94 SideCar asserted the need for regulatory recognition of the innovative combination of services offered by communications platforms such as SideCar, in combination with noncommercial ridesharing.95 Lyft stated that, to the extent the Commission finds that it should regulate to protect public safety interests, it is supportive of a third way regulatory approach because, if applied to TNCs, the current regulatory scheme would create unreasonable barriers for ridesharing services to enter the market.96 A second issue that was discussed during the workshops and does not neatly fit into any of the discussion above is the notion of fair competition among regulated and unregulated entities. TPAC commented that the goal of the Commission should be to create a fair system. They argue that where both a regulated system and an unregulated system exist, the natural inclination of the industry will be to move towards deregulation in order to avoid all of the costs of

93 Workshop Report at 15. 94 Id. 95 Id. 96 Id.

- 61 - R.12-12-011 COM/MP1/avs

regulatory compliance. Consequently there will be no room left for a regulated industry.97 Several parties including the SFMTA, San Francisco Airport Commission, TPAC, United Taxicab Workers, and the SF Cab Drivers Association contend that regulated taxis cannot compete with TNCs. United Taxicab Workers argue that to allow TNCs to exist in their current unregulated form or subject to minimal regulation essentially creates a race towards the bottom with negative impact on safety and service. These groups contended that professional drivers will be pushed towards the TNC business model because of lower operational costs. The representative from the SFMTA/IATR states that when this unregulated system devastates the regulated environment, no one will be left to provide safe and accessible door to door service to city residents and visitors.98 7.1. Discussion We are not persuaded by the position taken by the SFMTA that updating regulation is offensive to those currently working to regulate public safety and access. Regulatory bodies must always look to update their rules and regulations in order to keep pace with time and technology. The Commission’s goal in this Rulemaking is to strike the proper balance between safety and innovation, so that regulation provides a safety net that the public can rely on for its protection while new businesses innovate and use technology to better the lives of Californians. The regulations that we are adopting for TNCs are similar to what the SFMTA requires of taxicab drivers. Namely, we require a license for each TNC, require a criminal background check to be completed for each driver,

97 Workshop Report at 26. 98 Id.

- 62 - R.12-12-011 COM/MP1/avs

require that each TNC establish a driver training program, and require liability insurance that is equal to what the SFMTA requires of taxicab drivers. We will not, however, meddle into their business model by forcing TNCs to designate each driver an employee or contractor. Again, our role is to protect public safety, not to dictate the business models of these companies. We reject TPAC’s allegation that a third way of regulation is the same as deregulation. The settlement agreements that SED entered into with three of the companies were a first step toward regulation. The regulations that we establish in this decision will ensure that safety is foundational to a TNC’s business. Additionally, we support choice not only for passengers, but also drivers. Going forward, a company may either apply for a TNC license or a TCP license with the Commission. We accept those party’s comments calling for regulation of TNCs. As such, in this decision we exercise our existing jurisdiction pursuant to Article XII of the California Constitution and the Act. In this decision under the broad grant of authority pursuant to PU Code § 5381, we create the category of TNC to accompany the existing category of TCP. A company or individual wishing to provide transportation or facilitate transportation of passengers can choose to either get a TCP license or a TNC license. The TCP requirements are already in place, although as indicated, supra, the Commission will open a second phase to this Rulemaking to update those rules and regulations to ensure that safety requirements are up to date. 8. Comments on Proposed Decision The proposed decision of Commissioner Michael R. Peevey in this matter was mailed to the parties in accordance with Section 311 of the Public Utilities Code and comments were allowed under Rule 14.3 of the Commission’s Rules of

- 63 - R.12-12-011 COM/MP1/avs

Practice and Procedure. Comments were filed on August 19, 2013 by Tickengo, CforAT, SideCar, Lyft, Uber, TPAC, IATR, Los Angeles Department of Transportation, GCLA, TransForm, Luxor Cab, eRideshare, SFMTA, California Airports Council, TLPA, San Francisco Cab Drivers Association (SFCDA), United Taxicab Workers, SFMTA/SFO, PIFC and Consumer Attorneys of California, and reply comments were filed on August 26, 2013 by TPAC, Luxor Cab, United Taxicab Workers, Lyft, IATR, CforAT, TLPA, SFMTA/SFO, SideCar, Uber, PIFC and SFCDA. In response to comments, the proposed decision has been revised to further explain the definition of what constitutes a TNC. It is further noted that the existing exemptions under the Commission’s Charter Party Carrier authority are not usurped by the creation of this new category. All of the existing eleven exemptions still apply. The proposed decision has also been revised to clarify what kind of a criminal background check is expected, the insurance requirements and what specifics should be included in the TNC plans to ensure accessibility. Other revisions in response to comments have been made as appropriate. 9. Assignment of Proceeding Michael R. Peevey is the assigned Commissioner and Robert Mason III is the assigned ALJ in this proceeding. Findings of Fact 1. The Commission opened this Rulemaking on December 20, 2012, to protect public safety and to encourage innovators to use technology to improve the lives of Californians. 2. The Commission has a responsibility for determining whether and how public safety might be affected by these TNCs.

- 64 - R.12-12-011 COM/MP1/avs

3. Parties filed comments in this proceeding on January 28, 2013 and reply comments were filed on February 11, 2013. 4. On February 15, 2013, the Commission held a Prehearing Conference and on April 2, 2013, the assigned Commissioner and ALJ issued a Scoping Memo. 5. Workshops were held on April 11 and 12, 2013, at the Commission’s auditorium. 6. In the Rulemaking we referred to these companies as New Online-Enabled Transportation Services. We are changing the abbreviation to TNC for ease of use. 7. TNCs are not just Lyft, SideCar, InstantCab, and UberX. 8. A TNC is defined as an organization whether a corporation, partnership, sole proprietor, or other form, operating in California that provides prearranged transportation services for compensation using an online-enabled application (app) or platform to connect passengers with drivers using their personal vehicles. 9. California law currently recognizes and regulates three modes of passenger transportation for compensation: taxi services, regulated by cities and/or counties; and charter party carrier services, and passenger stage companies, regulated by the California Public Utilities Commission. 10. It is reasonable to conclude that in recent years, the communications revolution in wireless service, smartphones and apps has further facilitated the development and adoption of passenger transportation for compensation, to a point where passengers seeking rides are readily connected with drivers willing to provide rides in private vehicles. 11. It is reasonable to conclude that current TNCs are providing passenger transportation for compensation.

- 65 - R.12-12-011 COM/MP1/avs

12. TNCs do not fit neatly into the conventional understandings or statutory definitions of either taxis or limousines, but that does not mean that this Commission’s responsibility to public safety in the transportation industry should be ignored and/or left for individual companies to dictate. 13. TNCs operate on a prearranged basis because 1) before a passenger can request a ride, the passenger must download the software application, provide identification information and agree to the TNC service agreement, and 2) for a particular trip, the passenger must input information regarding current location, and finally 3) a TNC driver cannot be hailed on the street similar to a taxicab where no information is shared until the passenger enters the vehicle. 14. In order to comply with the applicable statutes and regulations, all TNC drivers must be able to prove that a ride was matched on the TNC software application as evidence of prearrangement. 15. The California DMV does not currently permit TNCs to enroll non- employee drivers in the Employer Pull Notice Program. Until the DMV Employer Pull Notice Program is available for use by TNCs, TNCs should perform, prior to allowing a driver on the platform and quarterly thereafter, driving record checks through DMV in order to ensure that drivers meet applicable requirements. The DMV check criteria shall provide that a user may have no more than 3 points within the preceding 3 years, no “major violations” (reckless driving, hit and run, or driving with a suspended license conviction) within the preceding 3 years, and no driving under the influence conviction within the past 7 years. 16. It is reasonable to conclude that TNCs are charter party passenger carriers, and therefore we will exercise our existing jurisdiction over these services

- 66 - R.12-12-011 COM/MP1/avs

pursuant to Article XII of the California Constitution and the Passenger Charter-party Carriers’ Act, PU Code § 5351 et seq. 17. It is reasonable to exercise this Commission’s broad grant of authority pursuant to PU Codes §§ 5381 and 701 to create the category of TNC to accompany the existing category of TCP. A company or individual wishing to provide transportation or facilitate transportation of passengers can choose to either get a TCP license or a TNC permit. 18. The definition of ridesharing does not permit transportation performed for profit. 19. Recovery of actual costs incurred only applies to vanpool vehicles, which is defined by the Vehicle Code as seating more than 10 passengers, but less than 15 passengers, including the driver. 20. It is reasonable to conclude that TNCs do not qualify for the rideshare exemption under PU Code § 5353(h), because § 5353(h) provides two opportunities to qualify for the rideshare exemption: either the transportation must have a common work-related purpose; or the transportation must be incidental to another purpose of the driver. TNCs fail to satisfy either of these requirements. 21. Pursuant to PU Code § 5352 the Commission’s authority over passenger carriers is grounded in the need to protect the public’s safe and reliable access to California’s roadways. 22. PU Code § 5352 positions public safety as a key goal in ensuring that the public enjoys full access to the roadways. 23. The primary distinction between a TNC and other TCPs is that a TNC connects riders to drivers who drive their personal vehicle, not a vehicle such as a limousine purchased primarily for a commercial purpose.

- 67 - R.12-12-011 COM/MP1/avs

24. A TNC shall not be permitted to accept street hails. 25. A TNC is not permitted to itself own vehicles used in its operation or own fleets of vehicles. With this definition in mind, the Commission finds that Uber (in contrast to UberX) is not a TNC. 26. Uber connects riders with drivers who do not drive their own personal vehicle, but typically operate in town cars or limousines, which the driver may often as well use to transport customers for another limousine/town car company. 27. In order to ensure the greatest possible evidentiary record, the Commission would prefer to leave all non-TNC issues, including Uber’s potential TCP status, to Phase II. 28. The Commission will not allow the uncertainty regarding Uber’s insurance to persist during the pendency of Phase II.Uber should be required to demonstrate to the Commission within 30 days of the issuance of this decision that it maintains commercial liability insurance policies providing not less than $1,000,000 (one million dollars) per-incident coverage for incidents involving vehicles and drivers while they are providing Uber services. The insurance coverage shall be available to cover claims regardless of whether an Uber driver maintains insurance adequate to cover any portion of the claim. 29. UberX does meet the TNC definition and should apply for a TNC license. 30. In this decision we will require TNCs to maintain commercial liability insurance policies providing not less than $1,000,000 (one million dollars) per- incident coverage for incidents involving vehicles and drivers while they are providing TNC services. The insurance coverage shall be available to cover claims regardless of whether a TNC driver maintains insurance adequate to cover any portion of the claim.

- 68 - R.12-12-011 COM/MP1/avs

31. The criminal background check must be a national criminal background check including the national sex offender database. The criminal background check should be using the applicant’s social security number and not just the applicant’s name. Any felony criminal conviction within seven years prior to the date of the background check for violent crime, a sexual offense, a crime involving property damage, and/or theft will make the applicant ineligible to be a TNC driver. 32. The Commission is authorized to conduct inspections of charter-party carriers which will now include TNCs. For instance, § 5371.5 of the Act states that: “Upon receipt of a complaint containing sufficient information to warrant conducting an investigation, the commission shall investigate any business that advertises limousine-for-hire or passenger charter transportation service for compensation in motor vehicles.” 33. The Commission is also authorized to issue fines pursuant to PU Code § 5378(b). 34. PU Code § 5411 to 5420 of the Act contain relevant provisions regarding issuing fines and penalties. These provisions allow the Commission to issue fines to carriers who have violated one or more provisions of the California Public Utilities Code. In addition, the Commission has established a citation program in Resolution ALJ-187. 35. The Commission’s purpose in this Rulemaking is to ensure that regulation is the safety net that the public relies on for its protection and secondarily encouraging innovation and utilization of technology to better the lives of Californians. 36. No Term and Condition in a TNC’s Terms of Service or elsewhere, can be inconsistent with this decision’s commercial liability insurance requirements for

- 69 - R.12-12-011 COM/MP1/avs

TNCs. Nor can any Term and Condition in a TNC’s Terms of Service be used or relied on by the TNC to deny insurance coverage, or otherwise evade the insurance requirements established in this decision. 37. The Commission will open a Phase II to consider updating its regulations over TCP certificate holders. Phase II will also consider the standard and appropriate language for Terms & Conditions for both TCP and TNC certificate holders. Conclusions of Law 1. The Federal Telecommunications Act of 1996 and recently adopted California legislation (Senate Bill 1161 authored by Senator Alex Padilla) limit California’s ability to regulate IP-enabled services, but they do not prevent California from regulating passenger transportation over public roadways. 2. TNCs are not providers of IP-enabled services and are not exempt from our jurisdiction. 3. To date neither the FCC, nor a court of higher jurisdiction, has ruled that this Commission, or any other state commission, is precluded by the Federal Telecommunication Act of 1996 from regulating TNCs. 4. The Commission regulates charter party passenger carriers pursuant to Article XII of the California Constitution and the Passenger Charter-party Carriers’ Act, PU Code, §§ 5351, et seq. Section 5360 states in part: Subject to the exclusions of Section 5353, “charter-party carrier of passengers” means every person engaged in the transportation of persons by motor vehicle for compensation, whether in common or contract carriage, over any public highway in this state. Section 5381 states in part: …(t)he commission may supervise and regulate every charter-party carrier of passengers in the State and may do all

- 70 - R.12-12-011 COM/MP1/avs

things…necessary and convenient in the exercise of such power and jurisdiction. 5. The Commission has very broad powers under PU Code § 701 which suggests that the Commission has the ability (via a rulemaking process) to develop new categories of regulation when a new technology disrupts an existing industry. 6. We find that TNCs are charter party passenger carriers, and therefore we will exercise our existing jurisdiction pursuant to Article XII of the California Constitution and the Passenger Charter-party Carriers’ Act, PU Code § 5351 et seq. (the Act). In this decision, under the broad grant of authority pursuant to PU Codes § 5381 and 701, we create the category of TNC to accompany the existing category of TCP. 7. Section 5353(h) provides two opportunities to qualify for the rideshare exemption: Transportation of persons between home and work locations or of persons having a common work-related trip purpose in a vehicle having a seating capacity of 15 passengers or less, including the driver, which are used for the purpose of ridesharing, as defined in § 522 of the Vehicle Code, when the ridesharing is incidental to another purpose of the driver. 8. PU Code § 5353(h) exempts transportation of persons between home and work locations or of persons having a common work-related trip in a vehicle having a seating capacity of 15 passengers or less, including the driver, which are used for the purpose of ridesharing, as defined in § 522 of the Vehicle Code, when the ridesharing is incidental to another purpose of the driver. 9. The section also states the exemption does not apply if the primary purpose for the transportation of those persons is to make a profit. “Profit,” as

- 71 - R.12-12-011 COM/MP1/avs

used in this subdivision does not include the recovery of actual costs incurred in owning and operating a vanpool vehicle, as defined in § 668 of the Vehicle Code. 10. Current TNCs do not fulfill the rideshare exemption and actually are providing transportation services for compensation. 11. PU Code § 5352 positions public safety as a key goal in ensuring that the public enjoys full access to the roadways.

ORDER

IT IS ORDERED that: 1. Transportation Network Companies shall follow the safety and regulatory requirements as detailed in Section 2.2.4 of this decision. 2. All reports required by this decision to be submitted by Transportation Network Companies must be verified by the provision of a signature of an officer of the corporation stating under penalty of perjury under the laws of the State of California that the report is accurate and contains no material omissions. 3. Each Transportation Network Company (TNC) (not the driver) must have a license with this Commission. There are six types of charter party carrier permits/certificates. TNCs shall apply for a class P permit. 4. Each Transportation Network Company (TNC) is required to conduct a criminal background check for each driver prior to that applicant becoming a TNC driver. The criminal background check must be a national criminal background check including the national sex offender database. The criminal background check must use the applicant’s social security number and not just the applicant’s name. Any felony criminal conviction within seven years prior to the date of the background check for driving under the influence of drugs or alcohol, fraud, use of a motor vehicle to commit a felony, a violent crime or act of

- 72 - R.12-12-011 COM/MP1/avs

terror, a sexual offense, a crime involving property damage, and/or theft will make the applicant ineligible to be a TNC driver. 5. We require the Transportation Network Company (TNC) or an authorized third party facility licensed by the California Bureau of Automotive Repair to conduct and ensure that each vehicle passes a 19-point vehicle inspection prior to allowing a vehicle to be driven as part of the TNC’s service, and annually thereafter, and for the TNC to maintain the record of such inspections in case of an audit. 6. We require TNCs to maintain commercial liability insurance policies providing not less than $1,000,000 (one million dollars) per-incident coverage for incidents involving vehicles and drivers while they are providing TNC services. The insurance coverage shall be available to cover claims regardless of whether a TNC driver maintains insurance adequate to cover any portion of the claim. This insurance requirement shall be disclosed on each TNC’s app and website. 7. Until the Department of Motor Vehicle (DMV) Employer Pull Notice Program is available for use by Transportation Network Companies (TNC), TNCs shall perform, prior to allowing a driver on the platform and quarterly thereafter, driving record checks through the DMV in order to ensure that drivers meet applicable requirements. The DMV check criteria shall provide that a user may have no more than 3 points within the preceding 3 years, no “major violations” (reckless driving, hit and run, or driving with a suspended license conviction) within the preceding 3 years, and no driving under the influence conviction within the past 7 years. 8. Drivers for Transportation Network Companies are prohibited from accepting street hails from potential passengers.

- 73 - R.12-12-011 COM/MP1/avs

9. This decision orders a second phase to this proceeding to review the Commission’s existing regulations over limousines and other charter party carriers in order to ensure that these rules have kept pace with the needs of today’s transportation market, and that the public safety rules are up to date. In addition, the second phase will consider the potential impact of any legislative changes that could affect our ability to regulate the Transportation Network Company industry. 10. The Commission will convene a workshop one year after the issuance of this decision to hear from all stakeholders on the impacts of this new mode of transportation and accompanying regulations. Workshops topics will include, but not necessarily be limited to, a consideration of safety, competition, innovation, accessibility, congestion, the California Environmental Quality Act, and other pollution related issues. 11. Transportation Network Companies must submit a plan within 90 days of the issuance of this decision to the Safety and Enforcement Division to explain how they plan to ensure that this new form of transportation service does not create a divide between the able and disabled communities. 12. Within 45 days after the effective date of this Decision, the Commission will post a Transportation Network Company Application Packet on its website, and Transportation Network Companies currently operating in California must file their Transportation Network Company Applications with the Safety and

Enforcement Division 60 days thereafter if they wish to continue operating. 13. Uber is required to demonstrate to the Commission within 30 days of the issuance of this decision that it maintains commercial liability insurance policies providing not less than $1,000,000 (one million dollars) per-incident coverage for incidents involving vehicles and drivers while they are providing Uber services.

- 74 - R.12-12-011 COM/MP1/avs

The insurance coverage shall be available to cover claims regardless of whether an Uber driver maintains insurance adequate to cover any portion of the claim. 14. UberX meets the Transportation Network Company (TNC) definition and must apply for a TNC license. 15. No Term and Condition in a TNC’s Terms of Service or elsewhere, can be inconsistent with this decision. Nor can any Term and Condition in a TNC’s Terms of Service be used or relied on by the TNC to deny insurance coverage, or otherwise evade the insurance requirements established in this decision. 16. Taxicab Paratransit Association of California’s motion to compel discovery is denied without prejudice. 17. Rulemaking 12-12-011 remains open. This order is effective today. Dated September 19, 2013, at San Francisco, California.

MICHAEL R. PEEVEY President MICHEL PETER FLORIO CATHERINE J.K. SANDOVAL MARK J. FERRON CARLA J. PETERMAN Commissioners

- 75 - START WITH

A GUIDE FOR BUSINESS

LESSONS LEARNED FROM FTC CASES

FEDERAL TRADE COMMISSION

START WITH

1. Start with security.

2. Control access to data sensibly.

3. Require secure passwords and authentication.

4. Store sensitive personal information securely and protect it during transmission.

5. Segment your network and monitor who’s trying to get in and out.

6. Secure remote access to your network.

7. Apply sound security practices when developing new products.

8. Make sure your service providers implement reasonable security measures.

9. Put procedures in place to keep your security current and address vulnerabilities that may arise.

10. Secure paper, physical media, and devices.

When managing your network, developing an app, or even organizing paper files, sound security is no accident. Companies that consider security from the start assess their options and make reasonable choices based on the nature of their business and the sensitivity of the information involved. Threats to data may transform over time, but the fundamentals of sound security remain constant. As the Federal Trade Commission outlined in Protecting Personal Information: A Guide for Business, you should know what personal information you have in your files and on your computers, and keep only what you need for your business. You should protect the information that you keep, and properly dispose of what you no longer need. And, of course, you should create a plan to respond to security incidents.

In addition to Protecting Personal Information, the FTC has resources to help you think through how those principles apply to your business. There’s an online tutorial to help train your employees; publications to address particular data security challenges; and news releases, blog posts, and guidance to help you identify – and possibly prevent – pitfalls.

There’s another source of information about keeping sensitive data secure: the lessons learned from the more than 50 law enforcement actions the FTC has announced so far. These are settlements – no findings have been made by a court – and the specifics of the orders apply just to those companies, of course. But learning about alleged lapses that led to law enforcement can help your company improve its practices. And most of these alleged practices involve basic, fundamental security missteps. Distilling the facts of those cases down to their essence, here are ten lessons to learn that touch on vulnerabilities that could affect your company, along with practical guidance on how to reduce the risks they pose.

1 1 Start with security.

From personal data on employment applications to network files with customers’ credit card numbers, sensitive information pervades every part of many companies. Business executives often ask how to manage confidential information. Experts agree on the key first step: Start with security. Factor it into the decisionmaking in every department of your business – personnel, sales, accounting, information technology, etc. Collecting and maintaining information “just because” is no longer a sound business strategy. Savvy companies think through the implication of their data decisions. By making conscious choices about the kind of information you collect, how long you keep it, and who can access it, you can reduce the risk of a data compromise down the road. Of course, all of those decisions will depend on the nature of your business. Lessons from FTC cases illustrate the benefits of building security in from the start by going lean and mean in your data collection, retention, and use policies.

Don’t collect personal information you don’t need. Here’s a foundational principle to inform your initial decision-making: No one can steal what you don’t have. When does your company ask people for sensitive information? Perhaps when they’re registering online or setting up a new account. When was the last time you looked at that process to make sure you really need everything you ask for? That’s the lesson to learn from a number of FTC cases. For example, the FTC’s complaint against RockYou charged that the company collected lots of information during the site registration process, including the user’s email address and email password. By collecting email passwords – not something the business needed – and then storing them in clear text, the FTC said the company created an unnecessary risk to people’s email accounts. The business could have avoided that risk simply by not collecting sensitive information in the first place.

Hold on to information only as long as you have a legitimate business need. Sometimes it’s necessary to collect personal data as part of a transaction. But once the deal is done, it may be unwise to keep it. In the FTC’s BJ’s Wholesale Club case, the company collected customers’ credit and debit card information to process transactions in its retail stores. But according to the complaint, it continued to store that data for up to 30 days – long after the sale was complete. Not only did that violate bank rules, but by holding on to the information without a legitimate business need, the FTC said BJ’s Wholesale Club created an unreasonable risk. By exploiting other weaknesses in the company’s security practices, hackers stole the account data and used it to make counterfeit credit and debit cards. The business could have limited its risk by securely disposing of the financial information once it no longer had a legitimate need for it.

2 Don’t use personal information when it’s not necessary. You wouldn’t juggle with a Ming vase. Nor should businesses use personal information in contexts that create unnecessary risks. In the Accretive case, the FTC alleged that the company used real people’s personal information in employee training sessions, and then failed to remove the information from employees’ computers after the sessions were over. Similarly, in foru International, the FTC charged that the company gave access to sensitive consumer data to service providers who were developing applications for the company. In both cases, the risk could have been avoided by using fictitious information for training or development purposes.

2 Control access to data sensibly.

Once you’ve decided you have a legitimate business need to hold on to sensitive data, take reasonable steps to keep it secure. You’ll want to keep it from the prying eyes of outsiders, of course, but what about your own employees? Not everyone on your staff needs unrestricted access to your network and the information stored on it. Put controls in place to make sure employees have access only on a “need to know” basis. For your network, consider steps such as separate user accounts to limit access to the places where personal data is stored or to control who can use particular databases. For paper files, external drives, disks, etc., an access control could be as simple as a locked file cabinet. When thinking about how to control access to sensitive information in your possession, consider these lessons from FTC cases.

Restrict access to sensitive data. If employees don’t have to use personal information as part of their job, there’s no need for them to have access to it. For example, in Goal Financial, the FTC alleged that the company failed to restrict employee access to personal information stored in paper files and on its network. As a result, a group of employees transferred more than 7,000 consumer files containing sensitive information to third parties without authorization. The company could have prevented that misstep by implementing proper controls and ensuring that only authorized employees with a business need had access to people’s personal information.

3 Limit administrative access. Administrative access, which allows a user to make system-wide changes to your system, should be limited to the employees tasked to do that job. In its action against Twitter, for example, the FTC alleged that the company granted almost all of its employees administrative control over Twitter’s system, including the ability to reset user account passwords, view users’ nonpublic tweets, and send tweets on users’ behalf. According to the complaint, by providing administrative access to just about everybody in-house, Twitter increased the risk that a compromise of any of its employees’ credentials could result in a serious breach. How could the company have reduced that risk? By ensuring that employees’ access to the system’s administrative controls was tailored to their job needs.

3 Require secure passwords and authentication.

If you have personal information stored on your network, strong authentication procedures – including sensible password “hygiene” – can help ensure that only authorized individuals can access the data. When developing your company’s policies, here are tips to take from FTC cases.

Insist on complex and unique passwords. “Passwords” like 121212 or qwerty aren’t much better than no passwords at all. That’s why it’s wise to give some thought to the password standards you implement. In the Twitter case, for example, the company let employees use common dictionary words as administrative passwords, as well as passwords they were already using for other accounts. According to the FTC, those lax practices left Twitter’s system vulnerable to hackers who used password-guessing tools, or tried passwords stolen from other services in the hope that Twitter employees used the same password to access the company’s system. Twitter could have limited those risks by implementing a more secure password system – for example, by requiring employees to choose complex passwords and training them not to use the same or similar passwords for both business and personal accounts.

4 Store passwords securely. Don’t make it easy for interlopers to access passwords. In Guidance Software, the FTC alleged that the company stored network user credentials in clear, readable text that helped a hacker access customer credit card information on the network. Similarly, in Reed Elsevier, the FTC charged that the business allowed customers to store user credentials in a vulnerable format in cookies on their computers. In Twitter, too, the FTC said the company failed to establish policies that prohibited employees from storing administrative passwords in plain text in personal email accounts. In each of those cases, the risks could have been reduced if the companies had policies and procedures in place to store credentials securely. Businesses also may want to consider other protections – two-factor authentication, for example – that can help protect against password compromises.

Guard against brute force attacks. Remember that adage about an infinite number of monkeys at an infinite number of typewriters? Hackers use automated programs that perform a similar function. These brute force attacks work by typing endless combinations of characters until hackers luck into someone’s password. In the Lookout Services, Twitter, and Reed Elsevier cases, the FTC alleged that the businesses didn’t suspend or disable user credentials after a certain number of unsuccessful login attempts. By not adequately restricting the number of tries, the companies placed their networks at risk. Implementing a policy to suspend or disable accounts after repeated login attempts would have helped to eliminate that risk.

Protect against authentication bypass. Locking the front door doesn’t offer much protection if the back door is left open. In Lookout Services, the FTC charged that the company failed to adequately test its web application for widely-known security flaws, including one called “predictable resource location.” As a result, a hacker could easily predict patterns and manipulate URLs to bypass the web app’s authentication screen and gain unauthorized access to the company’s databases. The company could have improved the security of its authentication mechanism by testing for common vulnerabilities.

5 4 Store sensitive personal information securely and protect it during transmission.

For many companies, storing sensitive data is a business necessity. And even if you take appropriate steps to secure your network, sometimes you have to send that data elsewhere. Use strong cryptography to secure confidential material during storage and transmission. The method will depend on the types of information your business collects, how you collect it, and how you process it. Given the nature of your business, some possibilities may include Transport Layer Security/Secure Sockets Layer (TLS/SSL) encryption, data-at-rest encryption, or an iterative cryptographic hash. But regardless of the method, it’s only as good as the personnel who implement it. Make sure the people you designate to do that job understand how your company uses sensitive data and have the know-how to determine what’s appropriate for each situation. With that in mind, here are a few lessons from FTC cases to consider when securing sensitive information during storage and transmission.

Keep sensitive information secure throughout its lifecycle. Data doesn’t stay in one place. That’s why it’s important to consider security at all stages, if transmitting information is a necessity for your business. In Superior Mortgage Corporation, for example, the FTC alleged that the company used SSL encryption to secure the transmission of sensitive personal information between the customer’s web browser and the business’s website server. But once the information reached the server, the company’s service provider decrypted it and emailed it in clear, readable text to the company’s headquarters and branch offices. That risk could have been prevented by ensuring the data was secure throughout its lifecycle, and not just during the initial transmission.

Use industry-tested and accepted methods. When considering what technical standards to follow, keep in mind that experts already may have developed effective standards that can apply to your business. Savvy companies don’t start from scratch when it isn’t necessary. Instead, they take advantage of that collected wisdom. The ValueClick case illustrates that principle. According to the FTC, the company stored sensitive customer information collected through its e-commerce sites in a database that used a non-standard, proprietary form of encryption. Unlike widely-accepted encryption algorithms that are extensively tested, the complaint charged that ValueClick’s method used a simple alphabetic substitution system subject to significant vulnerabilities. The company could have avoided those weaknesses by using tried-and-true industry-tested and accepted methods for securing data.

6 Ensure proper configuration. Encryption – even strong methods – won’t protect your users if you don’t configure it properly. That’s one message businesses can take from the FTC’s actions against Fandango and Credit Karma. In those cases, the FTC alleged that the companies used SSL encryption in their mobile apps, but turned off a critical process known as SSL certificate validation without implementing other compensating security measures. That made the apps vulnerable to man-in-the-middle attacks, which could allow hackers to decrypt sensitive information the apps transmitted. Those risks could have been prevented if the companies’ implementations of SSL had been properly configured.

5 Segment your network and monitor who’s trying to get in and out.

When designing your network, consider using tools like firewalls to segment your network, thereby limiting access between computers on your network and between your computers and the internet. Another useful safeguard: intrusion detection and prevention tools to monitor your network for malicious activity. Here are some lessons from FTC cases to consider when designing your network.

Segment your network. Not every computer in your system needs to be able to communicate with every other one. You can help protect particularly sensitive data by housing it in a separate secure place on your network. That’s a lesson from the DSW case. The FTC alleged that the company didn’t sufficiently limit computers from one in-store network from connecting to computers on other in-store and corporate networks. As a result, hackers could use one in-store network to connect to, and access personal information on, other in-store and corporate networks. The company could have reduced that risk by sufficiently segmenting its network.

7 Monitor activity on your network. “Who’s that knocking on my door?” That’s what an effective intrusion detection tool asks when it detects unauthorized activity on your network. In the Dave & Buster’s case, the FTC alleged that the company didn’t use an intrusion detection system and didn’t monitor system logs for suspicious activity. The FTC says something similar happened in Cardsystem Solutions. The business didn’t use sufficient measures to detect unauthorized access to its network. Hackers exploited weaknesses, installing programs on the company’s network that collected stored sensitive data and sent it outside the network every four days. In each of these cases, the businesses could have reduced the risk of a data compromise or its breadth by using tools to monitor activity on their networks.

6 Secure remote access to your network.

Business doesn’t just happen in the office. While a mobile workforce can increase productivity, it also can pose new security challenges. If you give employees, clients, or service providers remote access to your network, have you taken steps to secure those access points? FTC cases suggest some factors to consider when developing your remote access policies.

Ensure endpoint security. Just as a chain is only as strong as its weakest link, your network security is only as strong as the weakest security on a computer with remote access to it. That’s the message of FTC cases in which companies failed to ensure that computers with remote access to their networks had appropriate endpoint security. For example, in Premier Capital Lending, the company allegedly activated a remote login account for a business client to obtain consumer reports, without first assessing the business’s security. When hackers accessed the client’s system, they stole its remote login credentials and used them to grab consumers’ personal information. According to the complaint in Settlement One, the business allowed clients that didn’t have basic security measures, like firewalls and updated antivirus software, to access consumer reports through its online portal. And in Lifelock, the FTC charged that the company failed to install antivirus programs on the computers that employees used to remotely access its network. These businesses could have reduced those risks by securing computers that had remote access to their networks.

8 Put sensible access limits in place. Not everyone who might occasionally need to get on your network should have an all- access, backstage pass. That’s why it’s wise to limit access to what’s needed to get the job done. In the Dave & Buster’s case, for example, the FTC charged that the company failed to adequately restrict third-party access to its network. By exploiting security weaknesses in the third-party company’s system, an intruder allegedly connected to the network numerous times and intercepted personal information. What could the company have done to reduce that risk? It could have placed limits on third-party access to its network – for example, by restricting connections to specified IP addresses or granting temporary, limited access.

7 Apply sound security practices when developing new products.

So you have a great new app or innovative software on the drawing board. Early in the development process, think through how customers will likely use the product. If they’ll be storing or sending sensitive information, is your product up to the task of handling that data securely? Before going to market, consider the lessons from FTC cases involving product development, design, testing, and roll-out.

Train your engineers in secure coding. Have you explained to your developers the need to keep security at the forefront? In cases like MTS, HTC America, and TRENDnet, the FTC alleged that the companies failed to train their employees in secure coding practices. The upshot: questionable design decisions, including the introduction of vulnerabilities into the software. For example, according to the complaint in HTC America, the company failed to implement readily available secure communications mechanisms in the logging applications it pre-installed on its mobile devices. As a result, malicious third-party apps could communicate with the logging applications, placing consumers’ text messages, location data, and other sensitive information at risk. The company could have reduced the risk of vulnerabilities like that by adequately training its engineers in secure coding practices.

9 Follow platform guidelines for security. When it comes to security, there may not be a need to reinvent the wheel. Sometimes the wisest course is to listen to the experts. In actions against HTC America, Fandango, and Credit Karma, the FTC alleged that the companies failed to follow explicit platform guidelines about secure development practices. For example, Fandango and Credit Karma turned off a critical process known as SSL certificate validation in their mobile apps, leaving the sensitive information consumers transmitted through those apps open to interception through man-in-the-middle attacks. The companies could have prevented this vulnerability by following the iOS and Android guidelines for developers, which explicitly warn against turning off SSL certificate validation.

Verify that privacy and security features work. If your software offers a privacy or security feature, verify that the feature works as advertised. In TRENDnet, for example, the FTC charged that the company failed to test that an option to make a consumer’s camera feed private would, in fact, restrict access to that feed. As a result, hundreds of “private” camera feeds were publicly available. Similarly, in Snapchat, the company advertised that messages would “disappear forever,” but the FTC says it failed to ensure the accuracy of that claim. Among other things, the app saved video files to a location outside of the app’s sandbox, making it easy to recover the video files with common file browsing tools. The lesson for other companies: When offering privacy and security features, ensure that your product lives up to your advertising claims.

Test for common vulnerabilities. There is no way to anticipate every threat, but some vulnerabilities are commonly known and reasonably foreseeable. In more than a dozen FTC cases, businesses failed to adequately assess their applications for well-known vulnerabilities. For example, in the Guess? case, the FTC alleged that the business failed to assess whether its web application was vulnerable to Structured Query Language (SQL) injection attacks. As a result, hackers were able to use SQL attacks to gain access to databases with consumers’ credit card information. That’s a risk that could have been avoided by testing for commonly-known vulnerabilities, like those identified by the Open Web Application Security Project (OWASP).

10 8 Make sure your service providers implement reasonable security measures.

When it comes to security, keep a watchful eye on your service providers – for example, companies you hire to process personal information collected from customers or to develop apps. Before hiring someone, be candid about your security expectations. Take reasonable steps to select providers able to implement appropriate security measures and monitor that they’re meeting your requirements. FTC cases offer advice on what to consider when hiring and overseeing service providers.

Put it in writing. Insist that appropriate security standards are part of your contracts. In GMR Transcription, for example, the FTC alleged that the company hired service providers to transcribe sensitive audio files, but failed to require the service provider to take reasonable security measures. As a result, the files – many containing highly confidential health-related information – were widely exposed on the internet. For starters, the business could have included contract provisions that required service providers to adopt reasonable security precautions – for example, encryption.

Verify compliance. Security can’t be a “take our word for it” thing. Including security expectations in contracts with service providers is an important first step, but it’s also important to build oversight into the process. The Upromise case illustrates that point. There, the company hired a service provider to develop a browser toolbar. Upromise claimed that the toolbar, which collected consumers’ browsing information to provide personalized offers, would use a filter to “remove any personally identifiable information” before transmission. But, according to the FTC, Upromise failed to verify that the service provider had implemented the information collection program in a manner consistent with Upromise’s privacy and security policies and the terms in the contract designed to protect consumer information. As a result, the toolbar collected sensitive personal information – including financial account numbers and security codes from secure web pages – and transmitted it in clear text. How could the company have reduced that risk? By asking questions and following up with the service provider during the development process.

11 9 Put procedures in place to keep your security current and address vulnerabilities that may arise.

Securing your software and networks isn’t a one-and-done deal. It’s an ongoing process that requires you to keep your guard up. If you use third-party software on your networks, or you include third-party software libraries in your applications, apply updates as they’re issued. If you develop your own software, how will people let you know if they spot a vulnerability, and how will you make things right? FTC cases offer points to consider in thinking through vulnerability management.

Update and patch third-party software. Outdated software undermines security. The solution is to update it regularly and implement third-party patches. In the TJX Companies case, for example, the FTC alleged that the company didn’t update its anti-virus software, increasing the risk that hackers could exploit known vulnerabilities or overcome the business’s defenses. Depending on the complexity of your network or software, you may need to prioritize patches by severity; nonetheless, having a reasonable process in place to update and patch third- party software is an important step to reducing the risk of a compromise.

Heed credible security warnings and move quickly to fix them. When vulnerabilities come to your attention, listen carefully and then get a move on. In the HTC America case, the FTC charged that the company didn’t have a process for receiving and addressing reports about security vulnerabilities. HTC’s alleged delay in responding to warnings meant that the vulnerabilities found their way onto even more devices across multiple operating system versions. Sometimes, companies receive security alerts, but they get lost in the shuffle. InFandango , for example, the company relied on its general customer service system to respond to warnings about security risks. According to the complaint, when a researcher contacted the business about a vulnerability, the system incorrectly categorized the report as a password reset request, sent an automated response, and marked the message as “resolved” without flagging it for further review. As a result, Fandango didn’t learn about the vulnerability until FTC staff contacted the company. The lesson for other businesses? Have an effective process in place to receive and address security vulnerability reports. Consider a clearly publicized and effective channel (for example, a dedicated email address like [email protected]) for receiving reports and flagging them for your security staff.

12 10 Secure paper, physical media, and devices.

Network security is a critical consideration, but many of the same lessons apply to paperwork and physical media like hard drives, laptops, flash drives, and disks. FTC cases offer some things to consider when evaluating physical security at your business.

Securely store sensitive files. If it’s necessary to retain important paperwork, take steps to keep it secure. In the Gregory Navone case, the FTC alleged that the defendant maintained sensitive consumer information, collected by his former businesses, in boxes in his garage. In Lifelock, the complaint charged that the company left faxed documents that included consumers’ personal information in an open and easily accessible area. In each case, the business could have reduced the risk to their customers by implementing policies to store documents securely.

Protect devices that process personal information. Securing information stored on your network won’t protect your customers if the data has already been stolen through the device that collects it. In the 2007 Dollar Tree investigation, FTC staff said that the business’s PIN entry devices were vulnerable to tampering and theft. As a result, unauthorized persons could capture consumer’s payment card data, including the magnetic stripe data and PIN, through an attack known as “PED skimming.” Given the novelty of this type of attack at the time, and a number of other factors, staff closed the investigation. However, attacks targeting point-of-sale devices are now common and well-known, and businesses should take reasonable steps to protect such devices from compromise.

Keep safety standards in place when data is en route. Savvy businesses understand the importance of securing sensitive information when it’s outside the office. InAccretive , for example, the FTC alleged that an employee left a laptop containing more than 600 files, with 20 million pieces of information related to 23,000 patients, in the locked passenger compartment of a car, which was then stolen. The CBR Systems case concerned alleged unencrypted backup tapes, a laptop, and an external hard drive – all of which contained sensitive information – that were lifted from an employee’s car. In each case, the business could have reduced the risk to consumers’ personal information by implementing reasonable security policies when data is en route. For example, when sending files, drives, disks, etc., use a mailing method that lets you track where the package is. Limit the instances when employees need to be out and about with sensitive data in their possession. But when there’s a legitimate business need to travel with confidential information, employees should keep it out of sight and under lock and key whenever possible.

13 Dispose of sensitive data securely. Paperwork or equipment you no longer need may look like trash, but it’s treasure to identity thieves if it includes personal information about consumers or employees. For example, according to the FTC complaints in Rite Aid and CVS Caremark, the companies tossed sensitive personal information – like prescriptions – in dumpsters. In Goal Financial, the FTC alleged that an employee sold surplus hard drives that contained the sensitive personal information of approximately 34,000 customers in clear text. The companies could have prevented the risk to consumers’ personal information by shredding, burning, or pulverizing documents to make them unreadable and by using available technology to wipe devices that aren’t in use.

Looking for more information? The FTC’s Business Center (business.ftc.gov) has a Data Security section with an up-to-date listing of relevant cases and other free resources.

About the FTC The FTC works for the consumer to prevent fraudulent, deceptive, and unfair practices in the marketplace. The Business Center gives you and your business tools to understand and comply with the law. Regardless of the size of your organization or the industry you’re in, knowing – and fulfilling – your compliance responsibilities is smart, sound business. Visit the Business Center at business.ftc.gov.

Your Opportunity to Comment The National Small Business Ombudsman and 10 Regional Fairness Boards collect comments from small businesses about federal compliance and enforcement activities. Each year, the Ombudsman evaluates the conduct of these activities and rates each agency’s responsiveness to small businesses. Small businesses can comment to the Ombudsman without fear of reprisal. To comment, call toll-free 1-888-REGFAIR (1-888-734-3247) or go to sba.gov/ombudsman.

14

Federal Trade Commission business.ftc.gov June 2015 Cotter v. Lyft, Inc., 60 F.Supp.3d 1067 (2015) 80 Cal. Comp. Cases 329, 24 Wage & Hour Cas.2d (BNA) 964

To determine whether a worker is an employee or an independent contractor under 60 F.Supp.3d 1067 California law, courts should apply the United States District Court, test with an eye towards the purposes the N.D. California. California wage and hour statutes were meant Patrick Cotter, et al., Plaintiffs, to serve, and the type of person they were v. meant to protect. Cal. Lab. Code §§ 226.7, 1194, 2802, 3700; Cal. Unemp. Ins. Code §§ Lyft, Inc., Defendant. 976, 13020. Case No. 13–cv–04065–VC Cases that cite this headnote | Signed 03/11/2015 [3] Labor and Employment Synopsis Independent Contractors and Their Background: Former drivers brought putative class action Employees against carrier, alleging that carrier improperly classified To determine whether a worker is an drivers as independent contractors, and thus denied them employee or an independent contractor under benefits of being classified as employees under California's California law, the principal question is wage and hour laws. Parties cross-moved for summary whether the person or company to whom judgment. service is rendered has the right to control the manner and means of accomplishing the result desired. [Holding:] The District Court, Vince Chhabria, J., held that genuine issue of material fact existed as to whether 1 Cases that cite this headnote carrier retained right to control drivers. [4] Labor and Employment Nature, Creation, and Existence of Motions denied. Employment Relation Under California law, the company need not exercise its full right of control for a worker to West Headnotes (16) be deemed an employee.

Cases that cite this headnote [1] Labor and Employment Presumptions and burden of proof [5] Labor and Employment Under California law, if someone performs Independent Contractors and Their a service for a company, or for another Employees individual, the person performing the service is generally presumed to be an employee; if The determination of whether a worker is the company contends the person is not an an employee or an independent contractor employee, but rather is performing services as under California law is not a question an independent contractor, the burden is on of interference, or non-interference, not the company to prove this. a question of whether there have been suggestions, or even orders, as to the conduct Cases that cite this headnote of the work; but a question of the right to act, as distinguished from the act itself or the [2] Labor and Employment failure to act. Independent Contractors Cases that cite this headnote

© 2016 Thomson Reuters. No claim to original U.S. Government Works. 1 Cotter v. Lyft, Inc., 60 F.Supp.3d 1067 (2015) 80 Cal. Comp. Cases 329, 24 Wage & Hour Cas.2d (BNA) 964

[6] Labor and Employment Cases that cite this headnote Independent Contractors and Their Employees [10] Labor and Employment Under California law, although the focus of Independent Contractors and Their the test for determining whether a worker is Employees an employee or an independent contractor is Beyond the primary question whether the on the company's right to control, a finding principal retains the right to control the of employee status for a particular worker or manner and details of the work, California group of workers does not require that the courts look to a number of secondary company retain the right to control every last indicia of the nature of a service relationship detail; employee status may still exist where a to determine whether an employment certain amount of freedom is inherent in the relationship exists, including: (1) whether work. the one performing services is engaged in a Cases that cite this headnote distinct occupation or business; (2) the kind of occupation, with reference to whether, in the locality, the work is usually done [7] Labor and Employment under the direction of the principal or by Independent Contractors and Their a specialist without supervision; (3) the skill Employees required in the particular occupation; (4) Under California law, what matters in the whether the principal or the worker supplies determination of whether a worker is an the instrumentalities, tools, and the place of employee or an independent contractor is work for the person doing the work; (5) the whether the entity retains all necessary control length of time for which the services are to over the relevant portion of its operations. be performed; (6) the method of payment, whether by the time or by the job; (7) whether Cases that cite this headnote or not the work is a part of the regular business of the principal; and (8) whether or [8] Labor and Employment not the parties believe they are creating the Nature, Creation, and Existence of relationship of employer-employee. Employment Relation Cases that cite this headnote Under California law, the right to terminate at will, without cause, is strong evidence in support of an employment relationship. [11] Labor and Employment Independent Contractors and Their Cases that cite this headnote Employees Under California law, the secondary factors [9] Labor and Employment for determining whether an employment Independent Contractors and Their relationship exists generally cannot be applied Employees mechanically as separate tests; they are Under California law, whether a right of intertwined and their weight depends often on control exists, such that a worker would particular combinations. be deemed an employee rather than an Cases that cite this headnote independent contractor, may be measured by asking whether or not, if instructions were given, they would have to be obeyed on pain [12] Labor and Employment of at-will discharge for disobedience. Independent Contractors and Their Employees

© 2016 Thomson Reuters. No claim to original U.S. Government Works. 2 Cotter v. Lyft, Inc., 60 F.Supp.3d 1067 (2015) 80 Cal. Comp. Cases 329, 24 Wage & Hour Cas.2d (BNA) 964

To determine whether a worker is an employee or independent contractor under California [16] Federal Civil Procedure law, a court should evaluate each service Fair Labor Standards Act cases; wages arrangement on its facts, recognizing that the and hours regulations dispositive circumstances may vary from case Genuine issue of material fact existed to case. as to whether carrier retained right to control drivers, and thus whether drivers 1 Cases that cite this headnote were employees or independent contractors, precluding summary judgment on drivers' [13] Labor and Employment claims under California's wage and hour laws. Questions of law and fact as to Cal. Lab. Code §§ 226.7, 1194, 2802, 3700; Cal. employment status Unemp. Ins. Code §§ 976, 13020.

Under California law, if reasonable people 2 Cases that cite this headnote could differ on whether a worker is an employee or an independent contractor based on the evidence in the case, the question is not for a court to decide, and it must go to the jury; Attorneys and Law Firms this is true even if no significant dispute exists about the underlying facts, because the act of *1069 Matthew David Carlson, Carlson Legal Services, weighing and applying numerous intertwined San Francisco, CA, for Plaintiffs. factors, based on particular facts, is itself generally the job of the jury. Alex Santana, Christopher M. Ahearn, Thomas Michael McInerney, Ogletree Deakins Nash Smoak & Stewart, Cases that cite this headnote P.C., San Francisco, CA, for Defendant.

[14] Labor and Employment Questions of law and fact as to ORDER DENYING CROSS–MOTIONS employment status FOR SUMMARY JUDGMENT Under California law, only when a court concludes that from all the facts only a Re: Dkt. Nos. 69, 74 single inference and one conclusion may be drawn on the issue of whether a worker is an VINCE CHHABRIA, District Judge employee or an independent contractor may the court rule on the question as a matter of I. law, without the need for a jury trial. The question in this case is whether Lyft drivers Cases that cite this headnote are “employees” or “independent contractors” under California law. The answer is of great consequence for the [15] Labor and Employment drivers, because the California Legislature has conferred Questions of law and fact as to many protections on employees, while independent employment status contractors receive virtually none. The answer is also of Under California law, all factors need not great import to Lyft, because its business model assumes point in one direction for a court to rule the drivers are independent contractors. as a matter of law about a worker's proper classification. At first glance, Lyft drivers don't seem much like employees. We generally understand an employee to be Cases that cite this headnote someone who works under the direction of a supervisor, for an extended or indefinite period of time, with fairly

© 2016 Thomson Reuters. No claim to original U.S. Government Works. 3 Cotter v. Lyft, Inc., 60 F.Supp.3d 1067 (2015) 80 Cal. Comp. Cases 329, 24 Wage & Hour Cas.2d (BNA) 964 regular hours, receiving most or all his income from that for Lyft in California since the company's inception in one employer (or perhaps two employers). Lyft drivers can 2012. The plaintiffs and Lyft have filed cross-motions work as little or as much as they want, and can schedule for summary judgment, with the plaintiffs urging the their driving around their other activities. A person might Court to declare them “employees” as a matter of law, treat driving for Lyft as a side activity, to be fit into his and Lyft urging the Court to declare them “independent schedule when time permits and when he needs a little contractors” as a matter of law. But under California extra income. law, the question of how to classify a worker is typically for a jury. A court may only decide the question as But Lyft drivers don't seem much like independent a matter of law if application of the multitude of contractors either. We generally understand an relevant factors would require any reasonable juror to independent contractor to be someone with a special skill reach the same conclusion. Here, because the numerous (and with the bargaining power to negotiate a rate for the factors for deciding whether a worker is an employee or use of that skill), who serves multiple clients, performing an independent contractor point in decidedly different discrete tasks for limited periods, while exercising great directions, a reasonable jury could go either way. discretion over the way the work is actually done. Accordingly, there must be a trial. Traditionally, an independent contractor is someone a principal might have found in the Yellow Pages to perform a task that the principal or the principal's own employees II. were unable to perform—often something tangential to the day-to-day operations of the principal's business. See Lyft operates a smartphone application, or “app,” Antelope Valley Press v. Poizner, 162 Cal.App.4th 839, through which passengers are matched with nearby 75 Cal.Rptr.3d 887, 900 (2008) (describing the traditional drivers who are available to transport people in their “notion [of] an independent contractor [as] someone hired personal automobiles. Lyft markets itself as “Your friend to achieve a specific result that is attainable within a finite with a car.” Carlson Decl., Ex F at LYFT 000088. Cars period of time, such as plumbing work, tax service, or the that transport passengers for Lyft are easily recognizable, creation of a work of art for a building's lobby”). Lyft because Lyft gives each driver a “Carstache” (a big fuzzy drivers use no special skill when they give rides. Their work pink mustache) to attach to the front of his car when is central, not tangential, to Lyft's business. Lyft might not using it to give “Lyfts.” 1 To be a Lyft driver, a person control when the drivers work, but it has a great deal of must download the app, submit his car for inspection, power over how they actually do their work, including the undergo some form of background check, and submit power to fire them if they don't meet Lyft's specifications to an in-person interview with a Lyft representative. See about how to give rides. And some Lyft drivers no doubt Cotter Depo. 40:21–46:13. To be a Lyft rider, a person treat their work as a full-time job—their livelihood may must simply download the app onto her smartphone and depend solely or primarily on weekly payments from Lyft, register her credit card information. See Kirtikar Decl., even while they lack any power to negotiate their rate of p.1; Carlson Decl., Ex F at LYFT 000108. pay. Indeed, this type of Lyft driver—the driver who gives “Lyfts” 50 hours a week and relies on the income to feed 1 his family—looks very much like the kind of worker the See Christine Lagorio–Chafkin, The Origin— California Legislature has always intended to protect as and Evolution—of Lyft's Pink Mustache, Inc.com (Aug. 1, 2014), http://www.inc.com/christine-lagorio/ an “employee.” evolution-of-lyft-mustache.html; see also Carlson Decl., Ex. UU. *1070 The plaintiffs in this case were once Lyft drivers. They contend Lyft owes them money because The rider uses the app to hail a ride. Kirtikar Decl., p.1. it should have paid them as employees rather than Lyft's system forwards the request to the nearest driver independent contractors. For example, they argue that, who is logged in to the app. That driver may then accept, under California law, Lyft should have reimbursed them decline, or ignore the ride request. Id. If the driver declines for expenses, and that, at least sometimes, Lyft failed the request or ignores it for a specified period, Lyft's to pay them minimum wage. The plaintiffs propose to system sends the request to the next closest driver who is represent a class consisting of all people who have driven logged on. If that driver accepts the ride, he is “matched” with the rider and generally proceeds to pick her up and

© 2016 Thomson Reuters. No claim to original U.S. Government Works. 4 Cotter v. Lyft, Inc., 60 F.Supp.3d 1067 (2015) 80 Cal. Comp. Cases 329, 24 Wage & Hour Cas.2d (BNA) 964 drive her to her destination. However, the driver may Lyft tracks each driver's acceptance rate, and tells cancel his acceptance before picking up the rider, or upon drivers that an acceptance rate above 90 percent is meeting the rider but before commencing the ride. Id. “excellent,” while an acceptance rate below 75 percent “needs improvement.” It tells drivers that an acceptance At the time the plaintiffs drove for Lyft, the company rate “well below the community standard” will result operated on a “donation” system. After the driver in an email warning, and after three such warnings the dropped off the rider at her destination, the app driver's account will be deactivated. Carlson Decl., Ex. recommended a donation for the ride. The rider could T at LYFT001733–34. Lyft also tracks each driver's decide whether to accept this amount, pay a different cancellations, and may terminate a driver for high amount, or pay nothing at all. If the rider took no cancellation rates. Kirtikar Depo 159:15–160:14. Lyft action within 24 hours after the end of the ride, Lyft instructs drivers to “call support” before canceling a ride. automatically charged the recommended amount to the Carlson Decl., Ex. Z at LYFT000038. rider's credit card. But if the rider chose to pay a different amount, Lyft instead charged that amount to the rider's At the end of a ride, the app prompts riders to rate the ride credit card (or charged nothing, if the rider so chose). on a scale of one to five stars, with one star being “awful” Lyft retained a 20 percent “administrative fee” from each and five being “awesome.” Id. at LYFT 000041. Drivers *1071 charge and paid the remainder to the driver, with whose average star rating falls below a certain threshold the driver receiving payment from Lyft on a weekly basis are subject to termination. 3 for all rides given during that week. Goldin Depo. at 67:7– 12. Lyft has since changed its system in some markets, 3 Over time, Lyft's policy shifted from an individualized including California, so that rather than recommending a review of drivers' ratings to a system where drivers donation, Lyft charges riders a minimum fare for every are automatically deactivated if their average rating 2 ride. either fell below a certain threshold—most recently, 4.6 stars—or fell within the bottom few percent of drivers. Kirtikar Depo. at 83:5–84:15. 2 See Salvador Rodriguez, Lyft also will instate fares in California, ditching donation Lyft's relationship with its drivers is governed in part by its system, L.A. Times (Nov. 15, 2013), Terms of Service, which drivers must accept if they wish to http:// articles.latimes.com/2013/nov/15/business/la– give “Lyfts.” The Terms of Service apply to both drivers fi–tn–lyft–minimum–fares–california–20131115. and riders. Under a section titled “Driver Representations To accept rides, a driver must log onto the Lyft app in and Warranties,” each driver agrees that: “driver mode.” Lyft uses projected demand to determine how many drivers it will allow to log onto the app in • he is at least 23 years old driver mode at any one time. In the first several months • he has a valid driver's license after the Lyft app became operational in May 2012, Lyft provided two methods by which drivers could arrange to • he owns or has the legal right to operate the vehicle and drive for Lyft on a given day: (i) drivers could submit is named on the insurance policy covering the vehicle requests for particular hours one week in advance and Lyft would approve some or all of these hours; or (ii) if it was • he will only use the vehicle that has been registered less than one week in advance, drivers could log onto a with Lyft website and reserve any available hours. In early 2013, • his vehicle is in good operating condition Lyft added a third method—allowing drivers to simply log onto driver mode at any time, without having a particular • he will not “offer or provide transportation services time slot approved or reserved. But this third method for profit, as a public carrier or taxi service, charge is only available when Lyft determines that demand is for rides or otherwise seek non-voluntary *1072 not already being met by the drivers who reserved time compensation from Riders, or engage in any other through the reservation system. Kirtikar Depo. at 61:04– activity in a manner that is inconsistent with such 62:04. Driver's obligations under this Agreement”

• he will not offer rides exceeding 60 miles

© 2016 Thomson Reuters. No claim to original U.S. Government Works. 5 Cotter v. Lyft, Inc., 60 F.Supp.3d 1067 (2015) 80 Cal. Comp. Cases 329, 24 Wage & Hour Cas.2d (BNA) 964

Santana Decl, Ex. D at LYFT 000147–48. Elsewhere in • “Go above and beyond with good service such the Terms of Service, Lyft disclaims its “control over as helping passengers with luggage or holding an the quality or safety of the transportation that occurs as umbrella for passengers when it's raining” a result of the Service.” Id. at LYFT000153. Similarly, in the “Limitation of Liability” section, the Agreement • “If you ever need to cancel a Lyft, call support first.” states that “LYFT HAS NO RESPONSIBILITY Carlson Decl., Ex. Z at LYFT000038. Another section of WHATSOVER FOR THE ACTIONS OR CONDUCT the guide instructed drivers to “Make sure your Carstache OF DRIVERS OR RIDERS.... LYFT HAS NO is on whenever you're driving for Lyft,” “[o]ffer your CONTROL OVER THE IDENTIY OR ACTIONS OF passenger a phone charge, and ask what kind of music they THE RIDERS AND DRIVERS....” Id. at LYFT000154. would like to listen to,” and “[f]irst ask your passenger if there is a preferred route they would like to take. If not, But the Terms of Service tell drivers that Lyft use [a GPS navigation system], even if you think you know “reserve[s] the right ... to investigate and terminate Your where you're going.” Id. at LYFT000039. participation in the Lyft Platform if You ... behaved in a way which could be regarded as inappropriate[.]” Id. at In July 2013, Lyft replaced the guide with a frequently LYFT 000148. In addition, the Terms of Service state that asked questions (“FAQs”) section on its website. Like the “Either You or We may terminate Your participation in guide, the FAQs instruct drivers about such things as the the Lyft Platform ... at any time, for any or no reason, cleanliness of their vehicles and the use of GPS navigation without explanation, effective upon sending written or while driving, but the FAQs include more detail. For email notice to the other party.... We maintain sole example, under the question “What are Lyft's standards discretion to bar Your use of the Services in the future, for for keeping my car clean?” drivers are instructed that any or no reason.” Id. at LYFT 000153. “[h]aving a clean car is a key part of the Lyft *1073 experience. It's relaxing for passengers (and you!) to ride Between December 2012 and July 2013, Lyft also gave in a pleasantly spotless and nice smelling car ... Wash and drivers a guide. This included a section called “Lyft Rules vacuum your car at least once a week to keep it in tip- of the Road,” which gave drivers a list of “rules to live by,” top condition.” Carlson Decl., Ex. RR at LYFT001705. including: The FAQs also cover a number of issues not addressed in • “Phone should always be mounted and plugged into the guide. For example, under the question “Can I ask ... charger” my passenger for their phone number or other contact information?” drivers are told that they may never ask • “No talking on the phone (unless it's the passenger)” for the contact information of a passenger. Carlson Decl., Ex. EE at LYFT001495. And under the question “Can I • “Only pick up Lyft passengers, don't pick up smoke in my car? Can passengers smoke in my car?” Lyft passengers who hail from the street or who use other warns drivers it will disable their accounts if a passenger mobile apps.” reports that the driver's car smells like smoke. Carlson Decl., Ex. GG at LYFT001498. • “You should be the only non-passenger in the car. (no friends, children or pets can ride along with you)”

• “Greet every passenger with a big smile and fist bump” III.

• “Keep your car clean on the inside and outside” Plaintiff Patrick Cotter drove for Lyft from early September 2012 to January 30, 2013, when the drivers' • “Keep your seats and trunk clear for use by your guide was in effect. In the roughly four months he drove passengers” for Lyft, Cotter gave 173 rides, which covered a total of • “Do not request tips. If asked by the passenger, let 410 miles. Carlson Supp. Decl., Ex. DDDD. During this them know that the app will suggest a price” time, Cotter also worked at Facebook, and arranged his driving schedule so it would not conflict with that work. • “Do not accept any cash” Cotter Depo. at 147:9–16. Cotter was terminated after

© 2016 Thomson Reuters. No claim to original U.S. Government Works. 6 Cotter v. Lyft, Inc., 60 F.Supp.3d 1067 (2015) 80 Cal. Comp. Cases 329, 24 Wage & Hour Cas.2d (BNA) 964 using a substitute vehicle to give rides, rather than the car California law gives many benefits and protections to that Lyft had approved. Carlson Supp. Decl., Ex. UUU. employees; independent contractors get virtually none. Employees are generally entitled to, among other things, Plaintiff Alejandra Maciel drove for Lyft from August minimum wage and overtime pay, Cal. Lab.Code § 2, 2013 to September 13, 2013. By this time, Lyft had 1194, meal and rest breaks, Cal. Lab.Code § 226.7, replaced the drivers' guide with the FAQs. In the roughly reimbursement for work-related expenses, Cal. Lab.Code six-week period she worked for Lyft, Maciel drove § 2802, workers' compensation, Cal. Lab.Code § 3700, anywhere from 5 to 20 hours per week, averaging roughly and employer contributions to unemployment insurance. 30 completed rides per week. This does not include time Cal. Unemp. Ins.Code § 976. Employers are also required spent logged onto the app waiting for ride requests. Lyft under the California Unemployment Insurance Code to terminated Maciel after she received an average passenger withhold and remit to the state their employees' state rating of 4.54 stars, a rating that placed her in the bottom income tax payments. Cal. Unemp. Ins.Code § 13020. 5 percent of drivers. Carlson Decl., Ex. J. These statutes are designed to protect workers. For The lawsuit is a proposed class action. Cotter and example, the minimum wage statute seeks to guarantee Maciel originally asserted claims under the California that the “weakest and most helpless class” of workers Labor Code on behalf of all Lyft drivers in the country, receive “a wage that insures for them the necessary shelter, regardless of the state in which they drove for Lyft. But wholesome food and sufficient clothing.” Martinez v. the Court ruled that they could not pursue a nationwide Combs, 49 Cal.4th 35, 109 Cal.Rptr.3d 514, 231 P.3d class action under California's wage and hour laws. 259, 271 (2010). The rule that employees be reimbursed (Docket No. 51.) The plaintiffs subsequently amended for costs ensures that employers don't undercut wages by their complaint, and now seek to represent only people passing the cost of doing business on to their employees. who have driven for Lyft in California since the company's See Janken v. GM Hughes Elecs., 46 Cal.App.4th 55, 53 inception in 2012. The plaintiffs allege that because Lyft Cal.Rptr.2d 741, 753 (1996); see also Grissom v. Vons misclassifies its drivers as independent contractors, the Companies, Inc., 1 Cal.App.4th 52, 1 Cal.Rptr.2d 808, drivers have been deprived of California's minimum wage, 812 (1991) ( “[T]he obvious purpose of [section 2802] is reimbursement for work-related expenses, and other to protect employees from suffering expenses in direct protections that California law confers upon employees. consequence of doing their jobs.”). And “[t]he purpose The question whether the case may proceed as a class of the unemployment insurance program is to provide action has not yet been presented. The parties first benefits for ‘persons unemployed through no fault of their filed these cross-motions for summary judgment on own, and to reduce involuntary unemployment and the whether Cotter and Maciel themselves were employees or suffering caused thereby to a minimum.’ ” W. Hollywood independent contractors. Cmty. Health & Fitness Ctr. v. Cal. Unemployment Ins. Appeals Bd., 232 Cal.App.4th 12, 181 Cal.Rptr.3d 196, 198 (2014) (quoting Cal. Unemp. Ins.Code § 100). IV. The California Legislature has decided that employees [1] Under California law, if someone performs a service need these protections as a check against the bargaining for a company (or for another individual), the person advantage employers have over employees—particularly performing the service is generally presumed to be an unskilled, lower-wage employees—and the corresponding employee. If the company contends the person is not ability employers would otherwise have to dictate the an employee, but rather is performing services as an terms and conditions of the work. See C.S. Smith Metro. independent contractor, the burden is on the company to Mkt. Co. v. Lyons, 16 Cal.2d 389, 106 P.2d 414, 420– prove this. See Narayan v. EGL, Inc., 616 F.3d 895, 900 21 (1940) (“The inequality of bargaining power between (9th Cir.2010); see also Bemis v. People, 109 Cal.App.2d employer and employee has long been fully recognized by 253, 240 P.2d 638, 644 (1952). legislation curtailing the employer's freedom to bargain with his employees as he chooses.”); see also Narayan, 616 Whether a worker is classified as an employee or an F.3d at 897 (“[S]tatutes enacted to confer special benefits independent contractor has *1074 great consequences.

© 2016 Thomson Reuters. No claim to original U.S. Government Works. 7 Cotter v. Lyft, Inc., 60 F.Supp.3d 1067 (2015) 80 Cal. Comp. Cases 329, 24 Wage & Hour Cas.2d (BNA) 964 on workers are designed to defeat rather than implement Superior Court, 27 Cal.3d 690, 166 Cal.Rptr. 331, 613 contractual arrangements.”). P.2d 579, 585 (1980)); see also Borello, 256 Cal.Rptr. 543, 769 P.2d at 406 (“[T]he ‘control-of-work-details' test Independent contractors do not receive these protections for determining whether a person rendering service to because they generally are in a far more advantageous another is an ‘employee’ or an excluded ‘independent position: contractor’ must be applied with deference to the purposes of the protective legislation.”); id., 256 Cal.Rptr. 543, [C]ontractors who are truly 769 P.2d at 400 (explaining that the court's decision “has independent readily can sever the implications for the employer-employee relationship upon business relationship and take their which other state social legislation depends”); Truesdale services and equipment elsewhere v. Workers' Comp. Appeals Bd., 190 Cal.App.3d 608, 235 when faced with unfair or arbitrary Cal.Rptr. 754, 759 (1987) (“Persons such as Truesdale ... treatment, or unfavorable working are certainly within the class of persons the legislative conditions. They usually have scheme was meant to protect; and there are no applicable contracts with more than one statutory exclusions. Truesdale had little or no bargaining company, or contract with one position; he was simply required to sign the so-called company on a full-time basis for independent contractor agreement prepared by the alleged short durations, and consequently employer's attorney. There were no negotiations, no are not dependent on a single options.”). employer in the same all-or-nothing fashion as traditional employees [3] Turning to the actual test, the “principal” question who tend to work on a full- “is whether the person [or company] to whom service is time basis for an indefinite term. rendered has the right to control the manner and means of Because of these characteristics of accomplishing the result desired.” Borello, 256 Cal.Rptr. independence, a true contractor 543, 769 P.2d at 404 (quoting Tieberg v. Unemployment does not suffer the effects of unequal Ins. Appeals Bd., 2 Cal.3d 943, 88 Cal.Rptr. 175, 471 P.2d bargaining power to any degree 975, 977 (1970) (alteration and internal quotation marks comparable to that suffered by omitted)). See also Ayala v. Antelope Valley Newspapers, employees. 4 Inc., 59 Cal.4th 522, 173 Cal.Rptr.3d 332, 327 P.3d 165, 170–71 (2014); Alexander v. FedEx Ground Package Sys., Inc., 765 F.3d 981, 988 (9th Cir.2014); Ruiz v. Affinity 4 Ruth Burdick, Principles of Agency Permit the NLRB Logistics Corp., 754 F.3d 1093, 1101 (9th Cir.2014), cert. to Consider Additional Factors of Entrepreneurial Independence and the Relative Dependence of denied, ––– U.S. ––––, 135 S.Ct. 877, 190 L.Ed.2d 704 Employees When Determining Independent Contractor (2014); Millsap v. Fed. Express Corp., 227 Cal.App.3d Status Under Section 2(3), 15 Hofstra Lab. & Emp. 425, 277 Cal.Rptr. 807, 811 (1991) (“If control may be L.J. 75, 130 (1997). exercised only as to the result of the work and not the means by which it is accomplished, an independent [2] *1075 Accordingly, although it's easy to get lost contractor relationship is established.”). in the weeds when applying California's test for deciding whether a worker is an employee or an independent [4] [5] The company need not exercise its full right of contractor, courts should apply the test with an eye control for a worker to be deemed an employee. “ ‘It is towards the purposes those statutes were meant to serve, not a question of interference, or non-interference, not a and the type of person they were meant to protect. question of whether there have been suggestions, or even “[P]ast decisions ... teach that in light of the remedial orders, as to the conduct of the work; but a question of nature of the legislative enactments authorizing the the right to act, as distinguished from the act itself or the regulation of wages, hours and working conditions for failure to act.’ ” Ayala, 173 Cal.Rptr.3d 332, 327 P.3d at the protection and benefit of employees, the statutory 172 (quoting Hillen v. Indus. Acc. Comm'n, 199 Cal. 577, provisions are to be liberally construed with an eye to 250 P. 570, 571 (1926)). promoting such protection.” Martinez, 109 Cal.Rptr.3d 514, 231 P.3d at 275 (quoting Indus. Welfare Com. v.

© 2016 Thomson Reuters. No claim to original U.S. Government Works. 8 Cotter v. Lyft, Inc., 60 F.Supp.3d 1067 (2015) 80 Cal. Comp. Cases 329, 24 Wage & Hour Cas.2d (BNA) 964

[6] [7] And although the focus of the test is on the principal; and (h) whether or not the company's right to control, a finding of employee status parties believe they are creating the for a particular worker or group of workers does not relationship of employer-employee. require that the company retain the right to *1076 control every last detail. Employee status may still exist Id. Borello also recognized, as “logically pertinent to the where “[a] certain amount of ... freedom is inherent in inherently difficult determination” of employee status, the work.” Air Couriers Int'l v. Emp't Dev. Dep't, 150 the six-factor test employed by other jurisdictions, whose Cal.App.4th 923, 59 Cal.Rptr.3d 37, 44 (2007). What factors include: “(1) the alleged employee's opportunity matters is whether the entity retains “all necessary control for profit or loss depending on his managerial skill; [and] over the [relevant] portion of its operations.” Borello, 256 (2) the alleged employee's investment in equipment or Cal.Rptr. 543, 769 P.2d at 408. materials required for his task, or his employment of helpers[.]” Id., 256 Cal.Rptr. 543, 769 P.2d at 404. [8] [9] The right to terminate at will, without cause, is “[s]trong evidence in support of an employment [11] [12] These factors “[g]enerally ... cannot be applied relationship.” Borello, 256 Cal.Rptr. 543, 769 P.2d at mechanically as separate tests; they are intertwined and 404 (quoting Tieberg, 88 Cal.Rptr. 175, 471 P.2d at 979 their weight depends often on particular combinations.” (internal quotation marks omitted)). “Whether a right of Id., 256 Cal.Rptr. 543, 769 P.2d at 404 (quoting control exists may be measured by asking whether or not, Germann v. Worker's Comp. Appeals Bd., 123 Cal.App.3d if instructions were given, they would have to be obeyed on 776, 176 Cal.Rptr. 868, 871 (1981) (internal quotation pain of at-will discharge[ ] for disobedience.” Ayala, 173 marks omitted)); see also Ruiz, 754 F.3d at 1100. Cal.Rptr.3d 332, 327 P.3d at 172 (quoting Toyota Motor “[T]o determine whether a worker is an employee or Sales U.S.A., Inc. v. Superior Court, 220 Cal.App.3d 864, independent contractor, a court should evaluate ‘[e]ach 269 Cal.Rptr. 647, 653 (1990) (internal quotation marks service arrangement ... on its facts, [recognizing that] the omitted)). dispositive circumstances may vary from case to case.’ ” Ruiz, 754 F.3d at 1100 (quoting Borello, 256 Cal.Rptr. [10] Beyond the primary question whether the principal 543, 769 P.2d at 407). retains the right to control the manner and details of the work, California courts look to a number of “ ‘secondary’ indicia of the nature of a service relationship.” Borello, 256 V. Cal.Rptr. 543, 769 P.2d at 404. These include: [13] [14] Under California law, if reasonable people (a) whether the one performing could differ on whether a worker is an employee or services is engaged in a distinct an independent contractor based on the evidence in the occupation or business; (b) the case, the question is not for a court to decide; it must kind of occupation, with reference go to the jury. See Angelotti v. Walt Disney Co., 192 to whether, in the locality, the Cal.App.4th 1394, 121 Cal.Rptr.3d 863, 870 (2011). This work is usually done under the is true even if no significant dispute exists about the direction of the principal or by underlying facts, because *1077 the act of weighing a specialist without supervision; and applying numerous intertwined factors, based on (c) the skill required in the particular facts, is itself generally the job of the jury. particular occupation; (d) whether See, e.g., Narayan, 616 F.3d at 901 (“The drawing of the principal or the worker supplies inferences from subordinate to ‘ultimate’ facts is a task the instrumentalities, tools, and for the trier of fact—if, under the governing legal rule, the place of work for the person the inferences are subject to legitimate dispute.” (quoting doing the work; (e) the length of Sec'y of Labor v. Lauritzen, 835 F.2d 1529, 1543 (7th time for which the services are to Cir.1987) (Easterbrook, J., concurring))). Only when a be performed; (f) the method of court concludes that “from all the facts only a single payment, whether by the time or by inference and one conclusion may be drawn” may the the job; (g) whether or not the work court rule on the question as a matter of law, without the is a part of the regular business of the

© 2016 Thomson Reuters. No claim to original U.S. Government Works. 9 Cotter v. Lyft, Inc., 60 F.Supp.3d 1067 (2015) 80 Cal. Comp. Cases 329, 24 Wage & Hour Cas.2d (BNA) 964 need for a jury trial. Borello, 256 Cal.Rptr. 543,769 P.2d (2011) (“Even if one or two of the individual factors at 416 (internal quotation marks omitted). 5 might suggest an employment relationship, summary judgment is nevertheless proper when ... all the factors weighed and considered as a whole establish that [an 5 California cases sometimes label this a “question individual] was an independent contractor and not an of fact,” and they sometimes label it as a “mixed employee.”). Accordingly, several recent California Court question of law and fact.” Compare Hillen, 250 P. at 571 (mixed question of law and fact), with Borello, 256 of Appeal cases have held that plaintiffs were independent Cal.Rptr. 543, 769 P.2d at 416 (question of fact). In contractors as a matter of law, even while acknowledging the context of the parties' cross-motions for summary that certain factors cut in favor of employee status. judgment in this case, there is no difference. See See, e.g., Beaumont–Jacques v. Farmers Grp., Inc., 217 Narayan, 616 F.3d at 901 (“That there is a legal Cal.App.4th 1138, 159 Cal.Rptr.3d 102, 108 (2013); overlay to the factual question does not affect the role Arnold, 135 Cal.Rptr.3d at 220. And on the flip side, of the trier of fact.” (quoting Lauritzen, 835 F.2d at two recent Ninth Circuit cases have *1078 held that 1543)). Regardless of whether it is characterized as a plaintiffs were employees as a matter of law, even when question of fact or a mixed question, the question is some factors cut in favor of independent contractor status. one for the jury unless the evidence is so one-sided See Alexander, 765 F.3d at 994–97; Ruiz, 754 F.3d at that from it there is “but one inference [that] can 1103–05. But those rulings were based on a conclusion reasonably be drawn.” See Burlingham v. Gray, 22 that the arrow pointed so strongly in the direction of one Cal.2d 87, 137 P.2d 9, 16 (1943); see also Hana Fin., status or the other that no reasonable juror could have Inc. v. Hana Bank, ––– U.S. ––––, 135 S.Ct. 907, 911, ––– L.Ed.2d –––– (2015) (mixed questions of law and pointed the arrow in the opposite direction after applying fact are typically questions for the jury). California's multi-factor test. The Ninth Circuit has observed that, given California's “multi-faceted test” for employee status, and given the presumption in favor of employee status under California VI. law, establishing independent contractor status as a matter of law presents a “particularly difficult” hurdle. Under the standards discussed above, a reasonable Narayan, 616 F.3d at 900. Even though in Narayan jury could conclude that the plaintiff Lyft drivers were many of the underlying facts were undisputed, the Ninth employees. But because a reasonable jury could also Circuit concluded it could not “readily say ... that conclude that they were independent contractors, there the ‘ultimate conclusion as to whether the workers are must be a trial. employees or independent contractors' is one of law.” Id. at 901 (quoting Lauritzen, 835 F.2d at 1543). Likewise, in Arzate v. Bridge Terminal Transport Inc., 192 Cal.App.4th A. 419, 121 Cal.Rptr.3d 400 (2011), the California Court of Appeal reversed a trial court's grant of summary As a threshold matter, Lyft tepidly asserts there is no judgment in favor of a transportation company where need to decide how to classify the drivers, because they there was little dispute about the underlying facts. The don't perform services for Lyft in the first place. Under court noted that even though there was little indication the this theory, Lyft drivers perform services only for their company controlled the “manner and means” by which riders, while Lyft is an uninterested bystander of sorts, its drivers hauled their loads, several of the secondary merely furnishing a platform that allows drivers and factors weighed in favor of employee status. Id. at 405. riders to connect, analogous perhaps to a company like The court concluded that weighing this “competing, if not eBay. But that is obviously wrong. Lyft concerns itself necessarily conflicting, evidence” fell within the province with far more than simply connecting random users of of a trier of fact. Id. its platform. It markets itself to customers as an on- demand ride service, and it actively seeks out those [15] To be sure, all factors need not point in one direction customers. See Hartman Depo. at 64:17–65:24 (describing for a court to rule as a matter of law about a worker's Lyft's marketing efforts aimed at attracting riders). It proper classification. See Arnold v. Mut. of Omaha Ins. gives drivers detailed instructions about how to conduct Co., 202 Cal.App.4th 580, 135 Cal.Rptr.3d 213, 221 themselves. Notably, Lyft's own drivers' guide and FAQs

© 2016 Thomson Reuters. No claim to original U.S. Government Works. 10 Cotter v. Lyft, Inc., 60 F.Supp.3d 1067 (2015) 80 Cal. Comp. Cases 329, 24 Wage & Hour Cas.2d (BNA) 964 state that drivers are “driving for Lyft.” Carlson Decl., as commands or prohibitions, not suggestions. The title Ex. U, Ex. Z at LYFT000039. Therefore, the argument “Rules of the Road” does not sound like a list of that Lyft is merely a platform, and that drivers perform suggestions. Nor do Lyft's answers to the questions posed no service for Lyft, is not a serious one. See Yellow Cab in the FAQs, which include statements such as “do not v. Workers' Comp. Appeals Bd., 277 Cal.Rptr. 434, 437, take personal calls,” “always use navigation,” “you must 226 Cal.App.3d 1288 (1991) (noting that, contrary to keep your phone mounted at all times,” “[m]ake sure to Yellow Cab's insistence that it merely served as a lessor you always have music playing,” and “always wear your of taxicabs, Yellow “cultivated the passenger market by ‘stache on the grill!” See Carlson Decl., Exs. DD, JJ, KK, soliciting riders, process[ed] requests for service through OO, SS. a dispatching system,” and instructed the drivers in “service” and “courtesy”). Cf. Decision Adopting Rules Even if these instructions were ambiguous about whether and Regulations to Protect Public Safety While Allowing they are suggestive or mandatory, the record tends to New Entrants to the Transportation Industry, California resolve that ambiguity in favor of the latter, by indicating Public Utilities Commission, Decision 13–09–045, pp. 63– that Lyft reserves the right to penalize (or even terminate) 68 (Sept. 19, 2013) (concluding that companies such as drivers who don't follow them. Indeed, Lyft reprimanded Lyft are engaged in the business of providing passenger both Cotter and Maciel after passengers reported to Lyft transportation for compensation). 6 each plaintiff's respective violations of Lyft policies— Cotter's driving with a car other than the one Lyft had approved and Maciel's driving with her husband in the car. 6 Pltfs' RJN, Ex. A at 63–68. The Court grants the See Cotter Depo. at 111:10–116:18; see also Carlson Decl., plaintiffs' request for judicial notice of this document. Exs. PPP, SSS. In the Terms of Service, Lyft reserves See Fed. R. Evid. 201. the right to “investigate” and “terminate” drivers who [16] With respect to Lyft's more reasonable argument— have “behaved in a way which could be regarded as that the plaintiffs are independent contractors as a matter inappropriate.” The FAQs tell drivers Lyft will disable of law—the primary issue is the degree to which Lyft their accounts if their cars smell like smoke. Lyft also retains the right to control drivers. As discussed in Section tells drivers it may terminate them for declining too many IV, under California law, whether Lyft actually exercises ride requests, or for accepting and then canceling too this control is less important than whether it retains the many ride requests. And, as happened to Maciel, Lyft right to do so. terminates drivers whose passenger ratings fall below a certain threshold. See supra Section II. Although Lyft drivers enjoy great flexibility in when and how often to work, once they do accept ride requests, Lyft What's more, the Terms of Service not only give Lyft retains a good deal of control over how they proceed. a broad right to terminate drivers for cause, but also Lyft instructed the plaintiffs (in the “Rules of the Road” the right to bar drivers from the platform “for any or section of the driver guide and later in the FAQs on no reason.” As the California Supreme Court recently its website) not to do a number of things—not to talk explained in Ayala, a hirer's ability to discharge a worker on the phone with a passenger present, not to pick up without cause is “[p]erhaps the strongest evidence of the non-Lyft passengers, not to have anyone else in the car, right to control.” 173 Cal.Rptr.3d 332, 327 P.3d at 171; not to request tips, not to smoke or to allow the car to see also id. (“ ‘The power of the principal to terminate the *1079 smell like smoke, and not to ask for a passenger's services of the agent gives him the means of controlling contact information. Lyft also affirmatively instructed the the agent's activities.’ ” (quoting Malloy v. Fong, 37 Cal.2d plaintiffs to do a number of things—to wash and vacuum 356, 232 P.2d 241, 249 (1951)). the car once a week, to greet passengers with a smile and a fist-bump, to ask passengers what type of music they'd It would be difficult to rule as a matter of law that like to hear, to offer passengers a cell phone charge, and the plaintiffs were independent contractors when the to use the route given by a GPS navigation system if the most important factor for discerning the relationship passenger does not have a preference. See supra Section II. under California law, namely, the right of control, tends to cut the other way. But even beyond that, the Lyft insists these are suggestions, with no real consequence “secondary” factors cut in both directions. Several tend for a driver who ignores them. But most are written

© 2016 Thomson Reuters. No claim to original U.S. Government Works. 11 Cotter v. Lyft, Inc., 60 F.Supp.3d 1067 (2015) 80 Cal. Comp. Cases 329, 24 Wage & Hour Cas.2d (BNA) 964 to support Lyft's position. For example, the parties say they were available to accept packages for delivery; evidently believed they were entering into an independent (2) were free to decline to perform a particular delivery contractor relationship, as evidenced by the statement to when contacted by the dispatcher, even if they had made that effect in the Terms of Service (although the parties' themselves available; (3) chose their own driving routes; own perception of their relationship is not dispositive). (4) took time off whenever they wanted and did not See Alexander, 765 F.3d at 998, 997. Other factors tend need permission to do so; (5) used their own vehicles to support the plaintiffs' position. For example, the work (which did not bear any JKH marking or logo) to make performed by the drivers is “wholly integrated” into Lyft's the deliveries and paid for their own gas, maintenance, business—after all, Lyft could not exist without its drivers and insurance; (6) did not wear uniforms or badges —and “[t]he [riders] are [Lyft's] customers, not the drivers' that evidenced their affiliation or relationship with JKH; customers,” *1080 Estrada v. FedEx Ground Package (7) performed delivery services for other companies; (8) System, Inc., 154 Cal.App.4th 1, 64 Cal.Rptr.3d 327, 334, were not supervised by JKH; (9) earned their money 336 (2007). And driving for Lyft requires no special skill by splitting the fee that JKH charged its customers —something we often expect independent contractors to for each delivery; and (10) filled out application forms have. See JKH Enters. v. Dep't of Indus. Relations 142 acknowledging their status as independent contractor. 48 Cal.App.4th 1046, 48 Cal.Rptr.3d 563, 579 (2006). Cal.Rptr.3d at 568–70. And in Air Couriers, the court upheld a trial court's determination that drivers for Sonic Many of the remaining secondary factors are equivocal. Couriers were employees where the drivers: (1) decided The primary instrumentality needed to drive for Lyft, when and how long to work; (2) worked other jobs while a 2000 model year or newer car, is provided by the driving for Sonic; (3) were not required to accept every drivers, but in contrast to a truck or other commercial job, but instead rejected jobs for a variety of reasons and vehicle, providing a car often does not require a significant were not required to give reasons for doing so; (4) did not investment. See Gonzalez v. Workers' Comp. Appeals Bd., suffer repercussions for rejecting jobs; (5) were paid by the 46 Cal.App.4th 1584, 54 Cal.Rptr.2d 308, 312–13 (1996) job, and were able to negotiate higher rates on some jobs (noting that a car used by a worker “frequently serves for a variety of reasons; (6) supplied their own vehicles, also as a family all-purpose vehicle”). Drivers are paid supplies, and equipment when delivering for Sonic; (7) based on individual rides, but they have no apparent were not required to wear uniforms; and (8) received no ability to negotiate the rates at which Lyft's suggested formal training. 59 Cal.Rptr.3d at 38–39. donations are calculated, or the percentage Lyft retains as an “administrative fee.” Ultimately, none of these In light of these California cases (particularly the two secondary factors is dispositive. See Narayan, 616 F.3d at discussed immediately above), and given the evidence 901 (“ ‘We must assess and weigh all of the incidents of the here, Lyft's motion for summary judgment must be relationship with the understanding that no one factor is denied. While the evidence is far from *1081 conclusive, decisive[.]’ ” (quoting NLRB v. Friendly Cab Co., 512 F.3d “there exist[s] at the very least sufficient indicia of an 1090, 1097 (9th Cir.2008))). employment relationship between the plaintiff [d]rivers and [Lyft] such that a reasonable jury could find the Finally and most importantly, two relatively recent cases, existence of such a relationship.” Narayan, 616 F.3d at JKH Enterprises v. Department of Industrial Relations and 904. Air Couriers International v. Employment Development Department, undermine Lyft's argument that its drivers are independent contractors as a matter of law. Both B. cases involved facts quite similar to this one, and in both cases the California Court of Appeal upheld rulings that For similar reasons, the record precludes a summary delivery drivers were employees. In JKH, the court upheld judgment ruling in favor of the plaintiffs. As already the Department of Industrial Relations' conclusion that discussed, Lyft drivers enjoy great flexibility in when and “special” package delivery drivers were employees of how often to work—far more flexibility than the typical JKH, even though the special drivers: (1) were not employee. Indeed, Cotter and Maciel both testified that required to work either at all or on any particular they felt free to set their work hours as they saw fit, and schedule, but instead called in to a JKH dispatcher to that they could accept or reject individual ride requests.

© 2016 Thomson Reuters. No claim to original U.S. Government Works. 12 Cotter v. Lyft, Inc., 60 F.Supp.3d 1067 (2015) 80 Cal. Comp. Cases 329, 24 Wage & Hour Cas.2d (BNA) 964

Both plaintiffs testified that they chose which parts of the undisputed facts provided “overwhelming evidence” San Francisco in which they accepted ride requests, and of control over of the drivers' day-to-day work. See that they were never required to adhere to any appearance Ruiz, 754 F.3d at 1093. For example, Alexander turned standards. Both plaintiffs testified to having minimal on FedEx's “control over every exquisite detail of the contact with Lyft management during their tenure as drivers' performance,” 765 F.3d at 991, including: (i) their appearance; (ii) the appearance and specifications of their drivers. See Cotter Depo. at 150:3–25; Maciel Depo. at 105:25–106:24. They did not drive full time, and it vehicles; (iii) their work schedules; and (iv) their service does not appear that driving for Lyft was their primary areas. Id. at 989–90. The experience of the Lyft driver is much different from the experience of the FedEx driver, source of income. See Cotter Depo. at 147:9–16; Santana Decl., Ex. F (Cotter's resume, describing driving for underscoring why the plaintiffs have not established here Lyft as a “hobby”); Maciel Depo. at 31:22–24, 81:02– that summary judgment should be granted in their favor. 20 (explaining that she worked another job and drove around 10 hours per week, scheduling her driving around other commitments). Indeed, particularly given Cotter VII. and Maciel's relatively sparse work schedules, one could easily imagine a reasonable jury concluding that they As should now be clear, the jury in this case will be were independent contractors, even if other Lyft drivers handed a square peg and asked to choose between two with heavier or more regular schedules might properly be round holes. The test the California courts have developed over the 20th Century for classifying workers isn't very deemed employees. 7 helpful in addressing this 21st Century problem. Some factors point in one direction, some point in the other, 7 On the other hand, a sparse work schedule does not and some are ambiguous. Perhaps Lyft drivers who work necessarily preclude a finding of employee status. more than a certain *1082 number of hours should See Burlingham, 137 P.2d at 16. (“The fact that be employees while the others should be independent the employee chooses his own time to go out and contractors. Or perhaps Lyft drivers should be considered return and is not directed where to go or to whom to sell is not conclusive of the relationship and is a new category of worker altogether, requiring a different not inconsistent with the relation of employer and set of protections. But absent legislative intervention, employee.”). For example, if a server picked up a California's outmoded test for classifying workers will couple shifts a month at a restaurant when convenient apply in cases like this. And because the test provides for both the restaurant and the server, that schedule nothing remotely close to a clear answer, it will often be wouldn't prevent the server from being classified as an for juries to decide. That is certainly true here. employee when working those shifts. The primary cases the plaintiffs cite in support of their IT IS SO ORDERED. motion for summary judgment, Alexander v. FedEx Ground Package Systems and Ruiz v. Affinity Logistics, All Citations are very different from the case at hand. In both cases, the Ninth Circuit held that delivery drivers in California 60 F.Supp.3d 1067, 80 Cal. Comp. Cases 329, 24 Wage & were employees as a matter of law. But in each case, Hour Cas.2d (BNA) 964

End of Document © 2016 Thomson Reuters. No claim to original U.S. Government Works.

© 2016 Thomson Reuters. No claim to original U.S. Government Works. 13

1 2 3 4 UNITED STATES DISTRICT COURT 5 NORTHERN DISTRICT OF CALIFORNIA 6 JANE DOE 1, et al., 7 Case No. 15-cv-04670-SI Plaintiffs, 8 v. ORDER GRANTING IN PART AND 9 DENYING IN PART DEFENDANT'S UBER TECHNOLOGIES, INC., MOTION TO DISMISS 10 Defendant. Re: Dkt. Nos. 49, 52 11

12

13 Now before the Court is defendant Uber Technologies, Inc.’s motion to dismiss the 14 Amended Complaint. Docket No. 49. This matter came on for hearing on April 1, 2016. For the 15 reasons set forth below, the Court hereby GRANTS in part and DENIES in part defendant’s

16 motion.

17 United States District Court District States United

Northern District of California of District Northern 18 BACKGROUND 19 Plaintiffs Jane Doe 1 and Jane Doe 2 bring this tort suit against Uber Technologies, Inc. 20 (“Uber”) for sexual assaults that plaintiffs allege they suffered at the hands of Uber drivers. The 21 following allegations are drawn from plaintiffs’ Amended Complaint. 22 Since 2010, Uber has operated as a “transportation network company.” Docket No. 46,

23 Amended Complaint (“AC”) ¶ 23. Individuals download Uber’s smartphone application and then 24 use the “App” to make a transportation request. Id. ¶¶ 1, 23. “They are then matched with an 25 Uber driver who picks them up and drives them to a destination. App users must pay for the ride 26 through the App with a credit card. Uber pays the driver a share of the fare collected, and retains 27 the remainder.” Id. ¶ 23. 28 Uber solicits and retains non-professional drivers to provide the car rides that customers

1 order through the Uber App. Id. ¶ 25. One who wishes to drive for Uber applies online and 2 uploads photos of a driver’s license, vehicle registration, and proof of insurance. Id. ¶53. Uber 3 then performs a background check through a third party company. Id. ¶¶ 58, 66. This check runs 4 the driver’s social security number through a database, capturing information dating back seven 5 years. Id. ¶¶ 59, 66. Once Uber approves a driver, that driver is available to the public to provide 6 transportation services through the App. Id. ¶ 25. Neither drivers nor riders pay a fee to download 7 the Uber App. Id. ¶ 29. “Uber’s sole source of revenue is from charges to riders for trips taken.” 8 Id.

9 In February 2015, in Boston, Massachusetts, Doe 1 and her friends used the Uber App to 10 arrange a car ride after they had gone to dinner and then to a party. Id. ¶¶ 82-84. Uber driver 11 Abderrahim Dakiri confirmed that he was on his way, and picked up Doe 1 and her friends. Id.

12 ¶¶ 11, 85. After Dakiri dropped off Doe 1’s friends first, Doe 1 gave Dakiri the address of her

13 destination. Id. ¶ 86. Dakiri then began to sexually assault Doe 1. Id. ¶¶ 88-92. Dakiri did not 14 take a direct route to Doe 1’s destination but drove more than 15 minutes off route “in order to 15 increase his opportunity to sexually assault her.” Id. ¶ 93. Dakiri parked the car in a remote area

16 and continued to sexually assault Doe 1 until she was able to unlock the car door and run away.

17 Id. ¶¶ 96-97. United States District Court District States United

Northern District of California of District Northern 18 In August 2015, in Charleston, South Carolina, Doe 2 and a group of friends got a ride 19 from Uber driver Patrick Aiello, after Doe 2’s friend arranged the ride using Uber’s App. Id. 20 ¶¶ 12, 111-114. Aiello drove the group to a bar. Id. ¶ 119. He commented that he would like to 21 give the group a ride home, and someone in the group asked Aiello if he would agree to pick them 22 up later. Id. ¶¶ 122. The group later saw Aiello enter that same bar and observed him sitting at

23 the bar during the night. Id. ¶¶ 123, 125. 24 At the end of the evening, Aiello drove Doe 2 and a friend from the group back to her 25 friend’s apartment. Id. ¶ 128. During the ride, Doe 2 mentioned that she could not find her phone 26 and wanted to look for it at the apartment. Id. ¶ 129. Doe 2 “intended to collect her phone from 27 her friend’s apartment and walk the two blocks home to her apartment.” Id. ¶ 130. After looking 28 for her phone for five to ten minutes, Doe 2 left for her own apartment. Id. ¶ 131. 2

1 “When Ms. Doe 2 went outside, Aiello said he would drive her home.” Id. ¶ 132. “[S]till 2 believing that Aiello was acting in his capacity as an Uber driver,” Doe 2 got into the car and gave 3 Aiello her home address. Id. ¶ 133. Shortly thereafter, Doe 2 realized that Aiello was driving the 4 wrong way. Id. ¶ 134. When she pointed this out, Aiello asked, “How are you going to pay me?” 5 and told Doe 2 that she owed him a blow job. Id. ¶¶ 135-136. Doe 2 tried to get out of the car, but 6 Aiello had locked the doors. Id. ¶ 137. Aiello drove the car to a remote parking lot off a highway 7 area where he “proceeded to viciously rape her and threaten her with harm multiple times.” Id. 8 ¶¶ 138-139. Afterwards, Doe 2 “was able to get onto the highway, crossed to the median, and

9 then started running alongside the highway away from the parking lot.” Id. ¶ 140. A car hit Doe 10 2’s arm while she was waving for help. Id. ¶ 141. The car then stopped and called 911. Id. 11 Police took Doe 2 to the hospital, where she became suicidal and was transferred to a psychiatric

12 unit for three days. Id. ¶¶ 141, 143.

13 The Amended Complaint alleges that Uber’s background check system dates back seven 14 years. Id. ¶¶ 58-59, 66. After her assault, Doe 1 learned that Dakiri had resided in the United 15 States for less than three years. Id. ¶ 100. Aiello had a previous domestic violence arrest,

16 resulting in an assault conviction in April 2003. Id. ¶¶ 115, 117. Aiello applied to become a

17 driver for Uber in 2015. Id. ¶¶ 117-118. United States District Court District States United

Northern District of California of District Northern 18 On October 8, 2015, Doe 1 and Doe 2 filed this lawsuit against Uber. Docket No. 1. The 19 Court has jurisdiction based on diversity jurisdiction under 28 U.S.C. § 1332. Id. ¶ 15. Uber 20 moved to dismiss the complaint on December 3, 2015. Docket No. 34. On January 20, 2016, 21 plaintiffs filed their Amended Complaint, thereby mooting Uber’s motion. Docket Nos. 46, 48. 22 In their Amended Complaint, plaintiffs bring six claims for relief: (1) negligence and negligent

23 hiring, supervision, and retention; (2) fraud; (3) battery; (4) assault; (5) false imprisonment; and 24 (6) intentional infliction of emotional distress. 1 AC ¶¶ 248-300. Plaintiffs bring claims 3 through 25 6 against Uber under a theory of respondeat superior. Id. ¶¶ 269, 276, 284, 291. Plaintiffs seek 26

27 1 Plaintiffs include language under the IIED claim that more properly supports a claim for 28 negligence. See AC ¶¶ 296-298. At the hearing, plaintiffs confirmed that they bring this claim under a theory of intentional infliction of emotional distress. 3

1 declaratory and injunctive relief, damages (including punitive damages), and attorneys’ fees and 2 costs. Id. at 54-55. 3 Uber now moves to dismiss the Amended Complaint for failure to state a claim under 4 Federal Rule of Civil Procedure 12(b)(6). Docket No. 49, Motion to Dismiss (“Mot.”). Uber 5 argues that plaintiffs have failed to state a claim for relief as to all six of their claims. Uber asks 6 that the Court dismiss plaintiffs’ prayer for punitive damages. Uber also seeks to seal an exhibit 7 filed in support of its motion. Docket No. 52. 8 Following the hearing on April 1, 2016, Uber sought leave to file a supplemental brief.

9 Docket No. 61. The Court granted both parties leave to file supplemental briefs, which they filed 10 on April 8, 2016. Docket Nos. 62, 66, 67. 11

12 LEGAL STANDARD

13 Under Federal Rule of Civil Procedure 12(b)(6), a district court must dismiss a complaint 14 if it fails to state a claim upon which relief can be granted. To survive a Rule 12(b)(6) motion to 15 dismiss, the plaintiff must allege “enough facts to state a claim to relief that is plausible on its

16 face.” Bell Atl. Corp. v. Twombly, 550 U.S. 544, 570 (2007). This “facial plausibility” standard

17 requires the plaintiff to allege facts that add up to “more than a sheer possibility that a defendant United States District Court District States United

Northern District of California of District Northern 18 has acted unlawfully.” Ashcroft v. Iqbal, 556 U.S. 662, 678 (2009). While courts do not require 19 “heightened fact pleading of specifics,” a plaintiff must allege facts sufficient to “raise a right to 20 relief above the speculative level.” Twombly, 550 U.S. at 555, 570. 21 In deciding whether a plaintiff has stated a claim upon which relief can be granted, the 22 court must assume that the plaintiff’s allegations are true and must draw all reasonable inferences

23 in the plaintiff’s favor. See Usher v. City of Los Angeles, 828 F.2d 556, 561 (9th Cir. 1987). 24 However, the court is not required to “accept as true allegations that are merely conclusory, 25 unwarranted deductions of fact, or unreasonable inferences.” In re Gilead Scis. Sec. Litig., 536 26 F.3d 1049, 1055 (9th Cir. 2008). 27 If the court dismisses the complaint, it must then decide whether to grant leave to amend. 28 The Ninth Circuit has “repeatedly held that a district court should grant leave to amend even if no 4

1 request to amend the pleading was made, unless it determines that the pleading could not possibly 2 be cured by the allegation of other facts.” Lopez v. Smith, 203 F.3d 1122, 1130 (9th Cir. 2000) 3 (citations and internal quotation marks omitted). 4 5 DISCUSSION 6 I. Claims Relying on Respondeat Superior Theory 7 Uber urges the Court to dismiss plaintiffs’ claims 3 through 6, which rely on a theory of 8 respondeat superior. Uber argues that plaintiffs have not alleged sufficient facts to establish that

9 there is an employment relationship between Uber and drivers Dakiri and Aiello. Mot. at 4. Uber 10 alternatively argues that it cannot be vicariously liable because, it claims, sexual assault falls 11 outside the scope of an employee’s duties. Id. at 9. Uber also disputes plaintiffs’ assertion that

12 Uber is a “common carrier.” Id. at 13.

13 14 A. Employer-Employee Relationship 15 Under California law, “an employer may be held vicariously liable for torts committed by

16 an employee within the scope of employment.” Mary M. v. City of Los Angeles, 54 Cal. 3d 202,

17 208 (1991) (citation omitted). The parties dispute whether Uber drivers are employees of Uber; United States District Court District States United

Northern District of California of District Northern 18 plaintiffs allege that they are and defendants argue that they are not employees but are independent 19 contractors. See, e.g., AC ¶¶ 36-51; Mot. at 6; Docket No. 55, Reply at 4. Whether an individual 20 is classified as an employee or as an independent contractor depends on “whether the person to 21 whom service is rendered has the right to control the manner and means of accomplishing the 22 result desired.” S.G. Borello & Sons, Inc. v. Dep’t of Indus. Relations, 48 Cal. 3d 341, 350 (1989)

23 (citations omitted); see also Bradley v. Cal. Dep’t of Corrs. and Rehab., 158 Cal. App. 4th 1612, 24 1626 (2008) (“The prevailing view is to consider the totality of the circumstances, reflecting upon 25 the nature of the work relationship between the parties, and placing emphasis on the control 26 exercised by the employer over the employee’s performance of employment duties.”) (citing 27 Vernon v. State, 116 Cal. App. 4th 114, 124-25 (2004)). 28 5

1 While control is the key factor, California courts have recognized other indicia as relevant 2 to defining employment status. Derived from the Restatement (Second) of Agency § 220,2 these 3 considerations include:

4 (a) whether the one performing services is engaged in a distinct occupation or business; (b) the kind of occupation, with reference to 5 whether, in the locality, the work is usually done under the direction of the principal or by a specialist without supervision; (c) the skill 6 required in the particular occupation; (d) whether the principal or the worker supplies the instrumentalities, tools, and the place of work 7 for the person doing the work; (e) the length of time for which the services are to be performed; (f) the method of payment, whether by 8 the time or by the job; (g) whether or not the work is a part of the regular business of the principal; and (h) whether or not the parties 9 believe they are creating the relationship of employer-employee. 10 S.G. Borello, 48 Cal. 3d at 351 (citations omitted). Additionally, “[s]trong evidence in support of 11 an employment relationship is the right to discharge at will, without cause.” Id. at 350-51

12 (citations omitted). “The parties’ label is not dispositive and will be ignored if their actual conduct

13 establishes a different relationship.” Estrada v. FedEx Ground Package Sys., Inc., 154 Cal. App. 14 4th 1, 10-11 (2007). 15 As this multiple-factor approach suggests, a person's status as an employee or independent

16 contractor is a question of fact, but may be determined as a matter of law “if from all the facts only

17 a single inference and one conclusion may be drawn . . . .” Borello, 48 Cal. 3d at 367 (citations United States District Court District States United

Northern District of California of District Northern 18 omitted); see also Serv. Employees Int’l Union, 225 Cal. App. 3d 761, 771 (1990) (“[T]he 19 question whether one is an independent contractor, agent or employee is largely one of fact 20 depending on all the circumstances of the relations of the parties.”) (quoting Housewright v. 21 Pacific Far East Line, Inc., 229 Cal. App. 2d 259, 265 (1964)). 22 Here, plaintiffs have alleged sufficient facts to claim plausibly that an employment

23 relationship exists. In support of this assertion, plaintiffs have alleged that Uber sets fare prices 24 without driver input and that drivers may not negotiate fares. AC ¶ 39. If a driver takes a 25 circuitous route, Uber may modify the charges to the customer. Id. ¶ 40. Uber retains control

26 2 27 Although the Restatement (Second) of Agency has now been superseded by the Restatement (Third) of Agency, the factors continue to be used by the courts and remain largely 28 identical, save replacing the terms “master” and “servant” with “employer” and “employee.” See Schmidt v. Burlington N. & Santa Fe Ry. Co., 605 F.3d 686, 690 n.3 (9th Cir. 2010). 6

1 over customer contact information. Id. ¶ 42. Uber’s business model depends upon having a large 2 pool of non-professional drivers. Id. ¶ 25. There are no apparent specialized skills needed to 3 drive for Uber. Id. ¶¶ 36-38, 53-66. Uber retains the right to terminate drivers at will. Id. ¶ 43. 4 Uber also controls various aspects of the manner and means by which drivers may offer 5 rides through the Uber App. Among these, plaintiffs have alleged that Uber requires drivers to 6 accept all ride requests when logged into the App or face potential discipline. Id. ¶ 44. The 7 Amended Complaint also asserts that Uber requires drivers to:

8 dress professionally; send the customer who has ordered a ride a text message when the driver is 1-2 minutes away from the pickup 9 location; keep their radios either off or on “soft jazz or NPR;” open the door for the customer; and pick up the customer on the correct 10 side of the street where the customer is standing. 11 Id. ¶ 45.

12 Certain factors, as alleged, support Uber’s assertion that drivers are independent

13 contractors, though not enough to convert the question into a matter of law. See Borello, 48 Cal. 14 3d at 367. These include that the drivers generally do not receive a salary but are paid by the ride 15 and that the drivers supply their own cars and car insurance. AC ¶¶ 41, 50-51. Even these factors,

16 however, are not necessarily dispositive. See Estrada, 154 Cal. App. at 1, 5 (finding drivers for 17 FedEx to be employees even where drivers supplied their own trucks and maintained their own car

United States District Court District States United 3

Northern District of California of District Northern 18 insurance); see also AC ¶¶ 47-48 (alleging that in certain cities Uber drivers may receive a 19 guaranteed minimum rate, “tantamount to a salary,” and that in January 2016 Uber announced that 20 drivers will have guaranteed earnings, thereby—in plaintiffs’ view—giving “Uber drivers 21 everywhere . . . essentially guaranteed salaries . . . .”). It matters not whether Uber’s licensing 22 agreements label drivers as independent contractors, if their conduct suggests otherwise. See

23 Estrada, 154 Cal. App. at 10-11 (citations omitted); Mot. at 6-7. 24 25 26

27 3 Uber cites a similar case for the opposite proposition, but, as discussed further infra, Uber draws from the dissenting and not the majority opinion. See Mot. at 9 (citing Elijahjuan v. Super. 28 Ct., 210 Cal. App. 4th 15, 31-32 (2012)). 7

1 Other judges in this district have ruled likewise in similar cases. In O’Connor v. Uber 2 Technologies, Inc., Judge Chen found that plaintiff Uber drivers had sufficiently alleged the 3 existence of an employment relationship.4 The complaint in that case alleged that drivers:

4 are required to follow a litany of detailed requirements imposed on them by Uber and they are graded, and are subject to termination, 5 based on their failure to adhere to these requirements (such as rules regarding their conduct with customers, the cleanliness of their 6 vehicles, their timeliness in picking up customers and taking them to their destination, what they are allowed to say to customers, etc.)[.] 7 8 O’Connor v. Uber Techs., Inc., No. 13-cv-3826, 2013 WL 6354534, at *6 (N.D. Cal. Dec. 5,

9 2013). In Cotter v. Lyft, Inc., Judge Chhabria denied a summary judgment motion brought by 10 software app operator Lyft. Judge Chhabria refused to find as a matter of law that drivers who 11 used the Lyft app were independent contractors, where the evidence showed that “Lyft retains a

12 good deal of control over how [the drivers] proceed.” Cotter v. Lyft, Inc., 60 F. Supp. 3d 1067,

13 1078 (N.D. Cal. 2015). This included rules such as not talking on the phone with passengers 14 present, not requesting tips, and not asking for a passenger’s contact information, with Lyft 15 reserving the right to penalize or terminate drivers who did not comply with the rules. Id. at 1078-

16 79.

17 It may be that facts will ultimately be revealed that disprove plaintiffs’ allegations or that United States District Court District States United

Northern District of California of District Northern 18 tilt the scales toward a finding that Uber drivers are independent contractors. However, taking the 19 allegations in the Amended Complaint as true, plaintiffs have alleged sufficient facts that an 20 employment relationship may plausibly exist. 21 22 B. Acts within the Scope of Employment

23 Uber argues in the alternative that it cannot be vicariously liable for Aiello and Dakiri’s 24 acts because “sudden sexual assaults by employees are outside the scope of an employee’s duties 25 and cannot support employer liability.” Mot. at 9. This is not necessarily so under California law. 26

27 4 The Court notes that the plaintiffs have filed a motion for preliminary approval of a class 28 action settlement in that case. O’Connor, No. 13-cv-3826, Docket No. 518. The hearing on the motion is set for June 2, 2016. Id. 8

1 “For the doctrine of respondeat superior to apply, the plaintiff must prove that the 2 employee’s tortious conduct was committed within the scope of employment.” Mary M., 54 Cal. 3 3d at 209. This analysis asks whether “in the context of the particular enterprise an employee’s 4 conduct is not so unusual or startling that it would seem unfair to include the loss resulting from it 5 among other costs of the employer’s business.” Id. (quoting Perez v. Van Groningen & Sons, Inc., 6 41 Cal. 3d 962, 968 (1986)). This “foreseeability” analysis looks not to “statistical frequency, but 7 [to the] relationship between the nature of the work involved and the type of tort committed.” 8 Lisa M. v. Henry Mayo Newhall Mem’l Hosp., 12 Cal. 4th 291, 302 (1995). As guidance, courts

9 in California consult three policy goals underlying the respondeat superior doctrine: “preventing 10 future injuries, assuring compensation to victims, and spreading the losses caused by an enterprise 11 equitably . . . .” Id. at 304. Whether an employee was acting within the scope of employment is a

12 question of fact, unless “the facts are undisputed and no conflicting inferences are possible.”

13 Mary M., 54 Cal. 3d at 213 (quoting Perez, 41 Cal. 3d at 968). 14 With respect to sexual misconduct by an employee, the California Supreme Court has not 15 declared, as Uber would have it, that such acts always bar vicarious liability on the part of the

16 employer. See, e.g., Lisa M., 12 Cal. 4th at 297 n.3, 300 (noting that causal nexus to the

17 employee’s work is required, that such acts are not per se unforeseeable, and that California has United States District Court District States United

Northern District of California of District Northern 18 abandoned the “motive-benefit” test that would preclude vicarious liability for most sexual 19 assaults). In Mary M., the California Supreme Court held the city of Los Angeles liable when a 20 police officer, after detaining a woman during a traffic stop, followed the woman to her home and 21 raped her. Mary M., 54 Cal. 3d at 207. Several years later, however, in Lisa M., the court found 22 that an ultrasound technician was not acting within the scope of his employment when he molested

23 a patient during an examination. Distinguishing its prior holding in Mary M., the court explained 24 that the ultrasound technician “had no legal or coercive authority over plaintiff.” Lisa M., 12 Cal. 25 4th at 303. Rather, “[h]is subsequent battery of the patient was independent of the narrow purpose 26 for which plaintiff was asked to trust him.” Id. at 304. In Xue Lu v. Powell, 621 F.3d 944 (9th 27 Cir. 2010), the Ninth Circuit navigated the tension between the decisions in Mary M. and Lisa M., 28 explaining that “[t]he liability of a private employer in California does not turn on the 9

1 vulnerability of the victim but on the extent to which the tort of the employee is incident to his 2 employment.” Xue Lu, 621 F.3d at 949. After assessing questions of foreseeability and the policy 3 goals underlying respondeat superior, the appeals court found that an asylum officer was acting 4 within the scope of his employment when he molested applicants for asylum. See id. at 947-48. 5 In this case, which is only at the pleading stage, foreseeability and policy rationales weigh 6 in favor of allowing the complaint to move forward on the scope of employment question. It may 7 be that sexual assault by a taxi driver (or a taxi-like driver, as the case may be) “is not so unusual 8 or startling that it would seem unfair to include the loss resulting from it among other costs of the

9 employer’s business.” See Lisa M., 12 Cal. 4th at 299 (citations omitted). Assaults of this nature 10 are exactly why customers would expect taxi companies to perform background checks of their 11 drivers. Holding Uber liable could also forward the underlying policy goals of respondeat

12 superior, including prevention of future injuries and assurance of compensation to victims.

13 Unlike in cases of sexual harassment, for instance, plaintiffs in this case do not have separate 14 remedies under Title VII or California’s Fair Employment and Housing Act. See Farmers Ins. 15 Group v. County of Santa Clara, 11 Cal. 4th 992, 1019-1020 (1995) (finding on summary

16 judgment that sexual harassment was outside scope of employment, such that harasser could not

17 seek indemnity from the county as his employer). Arguably, though perhaps more tenuously, it is United States District Court District States United

Northern District of California of District Northern 18 possible that allowing liability would more equitably spread the losses caused by the enterprise of 19 shuttling customers in private cars. 20 Uber singles out the allegations against driver Aiello in particular, arguing that plaintiffs 21 have failed to allege that Aiello was using the Uber App at the time that he assaulted Doe 2 and 22 therefore that he can’t have been acting within the scope of employment. Mot. at 11-13. The

23 Amended Complaint states that Doe 2’s friend used the Uber App to arrange the initial pick-up 24 from Aiello to a bar but is silent as to whether the App was used to summon Aiello at the end of 25 the night. See AC ¶¶ 113-114. The Amended Complaint alleges that someone in the group asked 26 Aiello at the end of the initial car ride whether he would agree to pick them up later. Id. ¶ 122. 27 Plaintiffs also allege that Doe 2 got back into Aiello’s car at the end of the night, after he dropped 28 off her friend, “still believing that Aiello was acting in his capacity as an Uber driver . . . .” Id. ¶¶ 10

1 128-133. 2 It is no longer a principle under California law that an employer may be vicariously liable 3 for an employee’s assault only when the assault was committed to further the interests of the 4 employer. Lisa M., 12 Cal. 4th at 297 n. 3 (“[T]he ‘motive-benefit’ test, which would preclude 5 respondeat superior liability for most sexual assaults, has been ‘abandoned’ in California.”) (citing 6 LeGrand & Leonard, Civil Suits for Sexual Assault: Compensating Rape Victims, 8 Golden Gate 7 L. Rev. 479, 507 (1979)). The same principles regarding prevention of future injuries, assurance 8 of compensation to victims, and equitably spreading the losses caused by Uber’s business model

9 apply with regard to Aiello. A finder of fact could determine that Aiello was acting within the 10 scope of his employment where he was summoned using the Uber App, even if rides were offered 11 and received later that night without the use of the App. Taking the allegations in the light most

12 favorable to plaintiffs, as is required at this stage of the case, it is plausible that the rides Doe 2

13 took at the end of the night came about only because of Aiello’s affiliation with Uber. 14 In sum, the Court cannot determine -- as Uber effectively argues -- that as a matter of law 15 sexual assault by Uber drivers is always outside the scope of employment, if the drivers are in fact

16 ultimately found to be employees. The California Supreme Court has left this question open. See

17 Mary M., 54 Cal. 3d at 218 n.11. Like a police officer who rapes a detained woman, an employee United States District Court District States United

Northern District of California of District Northern 18 who throws a hammer at a fellow worker in a fit of irritation, or an asylum officer who abuses his 19 role to corner female immigrants and molest them, sexual assault by an Uber driver may be 20 incidental to the operation of its business. See Xue Lu, 621 F.3d at 948-49 (citing, inter alia, Carr 21 v. Wm. C. Crowell Co., 28 Cal. 2d 652 (1946)). At the very least, the pleadings present a close 22 enough call that the Court finds no reason to deviate from the ordinary rule that “the determination

23 whether an employee has acted within the scope of employment presents a question of fact . . . .” 24 See Mary M., 54 Cal. 3d at 213. For the purpose of surviving a motion to dismiss, plaintiffs have 25 plausibly alleged that drivers Dakiri and Aiello were acting within the scope of employment when 26 they assaulted plaintiffs. 27 28 11

1 C. Common Carrier 2 Plaintiffs seek to hold Uber liable as a common carrier with respect to claims 3 through 6. 3 AC ¶¶ 269, 276, 284, 291. They state that “[a]s a common carrier, Defendant is vicariously liable 4 for its employees’ and agents’ intentional and negligent torts, whether or not such acts were 5 committed within the scope of employment.” Id. 6 For the purposes of tort claims, California law defines “common carrier” as follows: 7 “Every one who offers to the public to carry persons, property, or messages, excepting only 8 telegraphic messages, is a common carrier of whatever he thus offers to carry.” Cal. Civ. Code

9 § 2168; see also Squaw Valley Ski Corp. v. Super. Ct., 2 Cal. App. 4th 1499, 1507 (1992). 10 “Hence, a common carrier within the meaning of Civil Code section 2168 is any entity which 11 holds itself out to the public generally and indifferently to transport goods or persons from place to

12 place for profit.” Id. at 1508.

13 In their briefs and at the hearing, Uber has argued that it is not a common carrier but that it 14 is a “broker” of transportation services. The Court is not persuaded by this argument. Uber cites 15 to two cases in support of this proposition. One involved a defendant which was found, on

16 summary judgment, not to be a common carrier but to be a property broker. Reply at 7-8 (citing

17 Chubb Group v. H.A. Transp. Sys., Inc., 243 F. Supp. 2d 1064, 1070 (C.D. Cal. 2002)). There, the United States District Court District States United

Northern District of California of District Northern 18 broker had contracted with a transportation company, which had then subcontracted with a 19 trucking company, for the transportation of cigarettes; the cigarettes were eventually stolen from a 20 truck while the truck was parked at a restaurant. Chubb Group, 243 F. Supp. 2d at 1066-67. 21 Factually, this little resembles the situation at hand. Uber also argues that it is a transportation 22 broker by presenting the dissent of a California Court of Appeal decision as the majority holding.

23 See Mot. at 9 (citing Elijahjuan v. Superior Court, 210 Cal. App. 4th 15, 31-32 (2012)). The 24 majority opinion stated, “The sole substantive issue on appeal is whether the parties agreed to 25 arbitrate their dispute.” Uber’s reliance on this case is therefore misplaced. See also O’Connor, 26 82 F. Supp. 3d at 1137-38, 1141-45 (granting summary judgment against Uber over Uber’s 27 arguments that it was “a ‘technology company,’ not a ‘transportation company’”). 28 12

1 Uber also argues that a common carrier finding “would merely . . . impose a heightened 2 duty of care; it would not displace the rule that an employer cannot be liable for employee 3 misconduct outside the scope of employment.” Mot. at 13. Many of the cases that Uber cites rely 4 on a negligence theory. In Lopez, for instance, the California Supreme Court held that the 5 common carrier bus company owed a higher degree of care to its passengers and was not immune 6 from liability when passengers injured other passengers on the bus. See Lopez v. S. California 7 Rapid Transit Dist., 40 Cal. 3d 780, 783 (1985). In San Francisco v. Superior Court, the Court of 8 Appeal noted that a common carrier “must use the utmost care and diligence for [the] safe carriage

9 [of its passengers].” City & Cty. of San Francisco v. Super. Ct., 31 Cal. App. 4th 45, 47-49 10 (1994). The reason the bus company was not held liable for a passenger’s attack in that case was 11 because no amount of care or diligence by the bus driver could have prevented the sudden

12 stabbing of a passenger, when the assailant gave “no warning or cause for alarm” until he pulled

13 out a knife seconds beforehand. Id. 14 A different analysis is required when the assailant is the common carrier’s own employee. 15 In Berger v. Southern Pacific Co., 144 Cal. App. 2d 1 (1956), the California Court of Appeal

16 found that the jury was properly instructed that the Pullman Company could be liable for the rape

17 that its porter committed upon a passenger. Berger, 144 Cal. App. 2d at 9. Quoting from a United States District Court District States United

Northern District of California of District Northern 18 treatise, the court explained, “The liability of a common carrier for an assault by one of its 19 employees on a passenger is not dependent on the question as to whether the employee was acting 20 within the scope of his authority or in the line of his duty, but is based upon its broad duty as a 21 common carrier to protect its passengers from assault.” Id. at 7. 22 When given an opportunity to brief the matter further, Uber argued the Berger decision "is

23 at odds with Supreme Court authority, ignores comparable cases decided before it, and violates 24 modern tort principles.” Docket No. 66, Def.’s Suppl. Br. at 1. However, Uber relied largely on 25 cases addressing negligence or passenger-on-passenger assault, and on a case decided by the New 26 York Court of Appeals. These cases simply are not analogous to the present facts, where plaintiffs 27 have alleged that an employee intentionally sexually assaulted a passenger. 28 13

1 Plaintiffs have alleged sufficient facts to plausibly claim that Uber is a common carrier. 2 Looking beyond any conclusory assertions, plaintiffs have alleged critical underlying facts: that 3 Uber’s services are available to the general public and that Uber charges customers standardized 4 fees for car rides. AC ¶ 24. Though plaintiffs allege that drivers may have their driving privileges 5 revoked, plaintiffs do not allege, nor does defendant counter, that an Uber customer may lose 6 riding privileges. See AC ¶ 43. Plaintiffs’ allegations support the claim that Uber “offers to the 7 public to carry persons,” thereby bringing it within California’s definition of common carrier for 8 tort purposes. See Cal. Civ. Code § 2168; see also Squaw Valley, 2 Cal. App. 4th at 1508 (finding

9 defendant to be a common carrier because it “indiscriminately offers its . . . chair lift to the public 10 to carry skiers at a fixed rate”). 11 Plaintiffs may, or may not, ultimately prevail on their vicarious liability claims. As Judge

12 Chhabria noted when analyzing the claims of Lyft drivers, “The test the California courts have

13 developed over the 20th Century for classifying workers isn’t very helpful in addressing this 21st 14 Century problem.” Cotter, 60 F. Supp. 3d at 1081. To the extent that these are close questions, 15 the Court finds that they are more appropriately resolved at a later stage of the litigation. For now,

16 plaintiffs have sufficiently alleged the claims, and accordingly, the Court DENIES Uber’s motion

17 to dismiss claims 3 through 6. United States District Court District States United

Northern District of California of District Northern 18 19 II. Claim for Negligent Hiring, Supervision, and Retention 20 In addition to suing Uber vicariously for the torts of Dakiri and Aiello, plaintiffs also seek 21 to hold Uber directly liable for negligent hiring, supervision, and retention. See AC at 46. Setting 22 aside the arguments regarding employment status that the Court has already addressed, Uber’s key

23 argument is that no claim for relief may lie where plaintiffs have not plausibly alleged that Uber 24 knew or should have known that Dakiri and Aiello posed any danger. Mot. at 14. 25 Under California law, an employer may be held directly liable for the behavior of an unfit 26 employee where the employer was negligent in the hiring, training, supervising, or retaining of 27 that employee. Delfino v. Agilent Techs., Inc., 145 Cal. App. 4th 790, 815 (2006). Negligence 28 liability will be imposed upon the employer if it knew or should have known that hiring the 14

1 employee created a particular risk or hazard and that particular harm then materializes. As such, 2 California follows the rule set forth in the Restatement (Second) of Agency Section 213, which 3 provides in pertinent part: “A person conducting an activity through servants or other agents is 4 subject to liability for harm resulting from his conduct if he is negligent or reckless . . . in the 5 employment of improper persons or instrumentalities in work involving risk of harm to others.” 6 Liability may be imposed “either on the basis of . . . action -- for example, the negligent hiring of 7 an agent -- or . . . inaction -- for example, the failure to provide adequate supervision of the agent's 8 work.” Far West Financial Corp. v. D & S Co., 46 Cal. 3d 796, 812 (1988).

9 As to driver Aiello, plaintiffs allege as follows. When Aiello signed up as an Uber driver, 10 Uber used a background check company called Accurate Background, Inc. AC ¶ 58. Accurate 11 would run a potential driver’s social security number “through records databases similar to those

12 held by credit agencies, which only go back for a period of seven years and do not capture all

13 arrests and/or convictions.” Id. ¶ 59. Aiello had previously been arrested for domestic violence 14 and was ultimately convicted of assault in 2003. Id. ¶¶ 115, 117. The background check did not 15 capture this conviction. Id. ¶ 118. Plaintiffs allege that Uber knew or should have known that

16 Aiello “would be a danger to passengers and lead to a risk of the very type of danger and harm that

17 occurred on . . . August 9, 2015.” Id. ¶ 250. United States District Court District States United

Northern District of California of District Northern 18 Uber does not directly dispute these allegations but characterizes Aiello’s prior conviction 19 as a “12-year-old disorderly-persons offense that could have been expunged.” Mot. at 15. Uber 20 supports its claim with an exhibit, asking the Court to incorporate the exhibit by reference or to 21 take judicial notice of it.5 See Docket Nos. 50, 51 ¶ 7, Ex. 5. Plaintiffs oppose this request. 22 Docket No. 54, Opposition at 17 n.9. It is not apparent at this time that the document Uber

23 presents is the same offense to which the Amended Complaint refers. The Court therefore 24

25 5 Uber has separately filed a motion to seal this exhibit, the state court record of Aiello’s 26 2003 conviction. Docket No. 52. Uber concedes that this document is a matter of public record. Reply at 11 n.3. At the hearing, Uber stated that it did not think the Court required the document 27 in order to dispose of the motion and indicated that it would withdraw the document. Given that Uber has filed no notice of withdrawal to date, and finding the document inappropriate for sealing, 28 the Court hereby DENIES the motion to seal.

15

1 DENIES Uber’s request for incorporation by reference and judicial notice.6 Taking the facts in the 2 Amended Complaint as true, the Court finds that plaintiffs have sufficiently alleged that Uber 3 should have known about Aiello’s criminal history such that Uber may be liable for negligent 4 hiring, supervision and retention. The motion to dismiss plaintiffs’ first claim for relief is 5 therefore DENIED as to Aiello. 6 Plaintiffs fail, however, to make such allegations as to driver Dakiri. The Amended 7 Complaint alleges that Dakiri had been in the country for less than three years when he assaulted 8 Doe 1 and that Uber used Accurate Background, Inc. to run a background check. AC ¶¶ 58-59,

9 100. The relevant allegations end there. Plaintiffs do not allege that anything existed in Dakiri’s 10 background that Uber knew or should have known and that should have prevented Uber’s 11 approval of Dakiri as a driver. The motion to dismiss plaintiffs’ claim of negligent hiring,

12 supervision, and retention is therefore GRANTED as to Dakiri. The Court grants the motion

13 without prejudice to a later request to amend this claim to add Dakiri back in should further 14 information come to light. 15

16 III. Claim for Fraud

17 For allegations of fraud or mistake, a complaint must meet the heightened pleading United States District Court District States United

Northern District of California of District Northern 18 standard of Rule 9(b), which requires a plaintiff to “state with particularity the circumstances 19 constituting fraud or mistake.” Fed. R. Civ. P. 9(b). A claim for fraud must have the following 20 elements: “(1) a misrepresentation (false representation, concealment, or nondisclosure); (2) 21 knowledge of falsity (or ‘scienter’); (3) intent to defraud, i.e., to induce reliance; (4) justifiable 22 reliance; and (5) resulting damage.” Anderson v. Deloitte & Touche LLP, 56 Cal. App. 4th 1468,

23 1474 (1997) (citation omitted); see also In re Glenfed, Inc. Sec. Litig., 42 F.3d 1541, 1548 (9th 24 Cir. 1994) (“[A] plaintiff must set forth what is false or misleading about a statement, and why it 25 is false.”). “Averments of fraud must be accompanied by the ‘who, what, where, and how’ of the

26 6 27 Plaintiffs do not challenge Uber’s request with respect to Exhibits 1 through 4. Finding no opposition and no reason to doubt that the documents are what they purport to be, the Court 28 GRANTS Uber’s request for judicial notice of Exhibits 1 through 4 of the Cohen declaration. See Docket No. 51. 16

1 misconduct charged.” Vess v. Ciba-Geigy Corp. USA, 317 F.3d 1097, 1106 (9th Cir. 2003) 2 (quoting Cooper v. Pickett,137 F.3d 616, 627 (9th Cir. 1997)). Rule 9(b) is satisfied if the 3 allegations “identif[y] the circumstances constituting fraud (or mistake) so that the defendant can 4 prepare an adequate answer from the allegations.” Moore v. Kayport Package Express, Inc., 885 5 F.2d 531, 540 (9th Cir. 1989). “Rule 9(b)’s particularity requirement applies to state-law causes 6 of action.” Vess, 317 F. 3d at 1103. 7 Plaintiffs base their second claim for relief on fraud. They assert that Uber made false 8 statements that riders would be safe taking rides through Uber, while knowing that Uber “had not

9 adequately screened its drivers.” AC ¶¶ 259-260. They also assert that Uber fraudulently 10 misrepresented its ability to track its drivers “and ensure that the drivers were taking the most 11 direct routes to Plaintiffs’ destinations, instead of going far off-route.” Id. ¶ 261. Uber argues that

12 plaintiffs have failed to plead their fraud claim with particularity, in that: (1) they fail to identify

13 what is false about Uber’s statements regarding driver background checks; (2) they fail to identify 14 what is false about Uber’s ability to track drivers’ routes, and about when, where, or how Uber 15 made such statements; and (3) they fail to establish reliance. Mot. at 16-17.

16 The Amended Complaint is replete with allegations regarding the falsity of Uber’s claims.

17 Plaintiffs describe in detail the process by which individuals become drivers for Uber and how United States District Court District States United

Northern District of California of District Northern 18 Uber screens those applicants. AC ¶¶ 36-38, 53-66. They include extensive quotes that plaintiffs 19 challenge as false, including ones that plaintiffs pulled from Uber’s website and from a report 20 commissioned by Uber and Mothers Against Drunk Driving. Id. ¶¶ 2-3, 70-71, 74-76, 236. Uber 21 clearly appears to understand what statements plaintiffs are challenging, as Uber summarizes and 22 refutes those allegations in its motion. See Mot. at 16-17. There is more than enough information

23 for Uber to “prepare an adequate answer from the allegations.” Moore, 885 F.2d at 540. 24 Moreover, plaintiffs cite to statements that Uber should readily have in its possession and be able 25 to identify. See Susilo v. Wells Fargo Bank, N.A., 796 F. Supp. 2d 1177, 1191 (C.D. Cal. 2011) 26 (“While the already heightened pleading standard is further heightened when a party pleads fraud 27 against a corporation, . . . the requirement is relaxed where ‘the defendant must necessarily 28 17

1 possess full information concerning the facts of the controversy,’ . . . or ‘when the facts lie more in 2 the knowledge of the opposite party[.]’”) (citations omitted). 3 The Amended Complaint also alleges that plaintiffs “actually and reasonably relied on the 4 false facts and misrepresentations provided by Defendant when they agreed to utilize Uber’s 5 services.” AC ¶ 264. At this stage of the case, the Court must draw all reasonable inferences in 6 favor of the non-moving party. See Usher, 828 F.2d at 561. It is reasonable for the Court to infer 7 from the allegations in the complaint that Doe 1 and Doe 2 assert that they accepted rides from 8 Uber drivers in reliance on the statements Uber made regarding rider safety. Uber further argues

9 that Doe 2 in particular could not have relied on Uber’s statements if she was not using the App 10 when she rode with Aiello. Mot. at 17; Reply at 13. If Uber is correct that Doe 2 rode with Aiello 11 without using the Uber App, it is reasonable for the Court to infer that she did so because she

12 relied on Uber’s stamp of approval that Aiello was a safe person from whom to obtain a ride.

13 Accordingly, Uber’s motion to dismiss plaintiffs’ fraud claim is DENIED. 14 15 IV. Punitive Damages

16 Lastly, Uber seeks dismissal of the request for punitive damages. Under California law, a

17 plaintiff may recover punitive damages in connection with a non-contractual claim if she United States District Court District States United

Northern District of California of District Northern 18 establishes by clear and convincing evidence that the defendant is guilty of (1) fraud, (2) 19 oppression or (3) malice. Cal. Civil Code § 3294(a). For an employer to be liable for those 20 damages, it must have “had advance knowledge of the unfitness of the employee and employed 21 him or her with a conscious disregard of the rights or safety of others or authorized or ratified the 22 wrongful conduct . . . or [been] personally guilty of oppression, fraud, or malice.” Id. § 3294(b)

23 (emphasis added). 24 Uber’s argument focuses on the first portion of Section 3294(b), whether Uber had 25 advance knowledge of Dakiri or Aiello’s unfitness. Mot. at 18-19. Uber ignores that punitive 26 damages may also be appropriate under that same subsection if Uber itself is found guilty of fraud. 27 See Cal. Civil Code § 3294(b). At the pleadings stage, a plaintiff need not plead all of the 28 elements of a claim for punitive damages but need only plead such facts as could plausibly support 18

1 the underlying claim of fraud that will support a punitive damages award. See Susilo, 796 F. 2 Supp. 2d at 1196-97.7 Because plaintiffs have successfully pleaded a fraud claim, § III, supra, the 3 Court DENIES Uber’s motion to dismiss the claim for punitive damages. 4 5 CONCLUSION 6 For the foregoing reasons and for good cause shown, the Court hereby: GRANTS Uber’s 7 motion to dismiss the negligent hiring, supervision, and retention claim (claim 1) as it relates to 8 Dakiri, without prejudice, but DENIES the motion as it relates to Aiello; DENIES the motion to

9 dismiss plaintiffs’ fraud claim (claim 2); DENIES Uber’s motion to dismiss claims brought under 10 a respondeat superior theory (claims 3 through 6); and DENIES Uber’s motion to dismiss requests 11 for punitive damages. The Court also DENIES Uber’s motion to seal Exhibit 5 to the Cohen

12 declaration. See Docket No. 52.

13 14 IT IS SO ORDERED. 15 Dated: May 4, 2016

16 ______SUSAN ILLSTON

17 United States District Judge United States District Court District States United

Northern District of California of District Northern 18 19 20 21 22

23 24

25 7 26 Susilo was later called into question by another district court in Crisanto v. Cty. of Tulare, No. 15-cv-1527, 2015 WL 7188165 (C.D. Cal. Nov. 16, 2015). Crisanto questioned whether the 27 court in Susilo correctly applied the legal standard to strike a claim for punitive damages under Rule 12(f). Crisanto, 2015 WL 7188165, at *6 n.5. Because Uber seeks the dismissal of 28 plaintiffs’ punitive damages claim here under Rule 12(b)(6), the Court finds that the Susilo analysis applies. 19 FOR PUBLICATION

UNITED STATES COURT OF APPEALS FOR THE NINTH CIRCUIT

JANE DOE NO. 14, No. 12-56638 Plaintiff-Appellant, D.C. No. v. 2:12-cv-03626-JFW-PJW

INTERNET BRANDS, INC., DBA Modelmayhem.com, ORDER AND OPINION Defendant-Appellee.

Appeal from the United States District Court for the Central District of California John F. Walter, District Judge, Presiding

Argued and Submitted February 7, 2014 Opinion withdrawn February 24, 2015 Re-argued and Submitted April 8, 2015 Pasadena, California

Filed May 31, 2016

Before: Mary M. Schroeder and Richard R. Clifton, Circuit Judges, and Brian M. Cogan, District Judge.*

Opinion by Judge Clifton

* The Honorable Brian M. Cogan, District Judge for the U.S. District Court for the Eastern District of New York, sitting by designation. 2 DOE V. INTERNET BRANDS, INC.

SUMMARY**

Communications Decency Act

The panel withdrew the opinion filed on September 17, 2014, and in a superseding opinion reversed the district court’s Fed. R. Civ. P. 12(b)(6) dismissal, as barred by the Communications Decency Act, of an action against Internet Brands, Inc. alleging liability for negligence under California law based on a failure to warn; and remanded for further proceedings.

Section 230(c) of the Communications Decency Act provides that “[n]o cause of action may be brought and no liability may be imposed under any State or local law that is inconsistent with this section.”

Plaintiff Jane Doe sought to hold Internet Brands liable for failing to warn her about information it obtained from an outside source about how third parties targeted and lured victims through Internet Brand’s website modelmayhem.com, a networking website for people in the modeling industry.

The panel held that the Communications Decency Act did not bar Jane Doe’s failure to warn claim under California law. The panel concluded that Jane Doe’s negligent failure to warn claim did not seek to hold Internet Brands liable as the “publisher or speaker of any information provided by another information content provider,” 47 U.S.C. § 230(c)(1), and therefore the Communications Decency Act did not bar the

** This summary constitutes no part of the opinion of the court. It has been prepared by court staff for the convenience of the reader. DOE V. INTERNET BRANDS, INC. 3 claim. The panel expressed no opinion on the viability of the failure to warn allegations on the merits.

COUNSEL

Jeffrey Herman (argued) and Stuart S. Mermelstein, Herman Law, Boca Raton, Florida, for Plaintiff-Appellant.

Daniel P. Collins (argued), Munger, Tolles & Olson LLP, Los Angeles, California; Jonathan H. Blavin, Munger, Tolles & Olson LLP, San Francisco, California; Wendy E. Giberti, iGeneral Counsel, P.C., Beverly Hills, California; Patrick Fraioli, Ervin Cohen & Jessup LLP, Beverly Hills, California, for Defendant-Appellee.

Patrick J. Carome (argued), Wilmer Cutler Pickering Hale and Dorr LLP, Washington, D.C.; Felicia H. Ellsworth and Brook Hopkins, Wilmer Cutler Pickering Hale and Dorr, Boston, Massachusetts, for Amici Curiae The Computer and Communications Industry Association; The Internet Association; Care.com, Inc.; Craigslist, Inc.; Facebook, Inc.; IAC/Interactivecorp; and Tumblr, Inc.

ORDER

By order entered February 24, 2015, Defendant-Appellee Internet Brands Inc.’s Petition for Rehearing, filed October 31, 2014, was granted, the Petition for Rehearing En Banc was denied as moot, the opinion filed on September 17, 2014 was withdrawn, and the case scheduled for a new oral argument. 4 DOE V. INTERNET BRANDS, INC.

An opinion is filed together with this order. Subsequent petitions for rehearing or rehearing en banc may be filed.

OPINION

CLIFTON, Circuit Judge:

Model Mayhem is a networking website, found at modelmayhem.com, for people in the modeling industry. Plaintiff Jane Doe, an aspiring model who posted information about herself on the website, alleges that two rapists used the website to lure her to a fake audition, where they drugged her, raped her, and recorded her for a pornographic video. She also alleges that Defendant Internet Brands, the company that owns the website, knew about the rapists but did not warn her or the website’s other users. She filed an action against Internet Brands alleging liability for negligence under California law based on that failure to warn.

The district court dismissed the action on the ground that her claim was barred by the Communications Decency Act (“CDA”), 47 U.S.C. § 230(c) (2012). We conclude that the CDA does not bar the claim. We reverse and remand for further proceedings.

I. Background

At the motion to dismiss stage, we assume factual allegations stated in the Complaint filed by Plaintiff to be DOE V. INTERNET BRANDS, INC. 5 true.1 Sprewell v. Golden State Warriors, 266 F.3d 979, 988 (9th Cir. 2001). Plaintiff alleges that Internet Brands owns and operates the website modelmayhem.com, which it purchased in 2008. Model Mayhem is a networking site for professional and aspiring models to market their services. It has over 600,000 members. Plaintiff Jane Doe, a fictitious name, was an aspiring model who became a member of Model Mayhem.

Unbeknownst to Jane Doe, two persons, Lavont Flanders and Emerson Callum, were using Model Mayhem to identify targets for a rape scheme, allegedly as early as 2006. Flanders and Callum are not alleged to have posted their own profiles on the website. Instead, they browsed profiles on Model Mayhem posted by models, contacted potential victims with fake identities posing as talent scouts, and lured the victims to south Florida for modeling auditions. Once a victim arrived, Flanders and Callum used a date rape drug to put her in a semi-catatonic state, raped her, and recorded the activity on videotape for sale and distribution as pornography.

In 2008, Internet Brands purchased Model Mayhem from Donald and Taylor Waitts, the original developers of the site. Shortly after the purchase, Internet Brands learned of how Flanders and Callum were using the website. It is not alleged precisely how Internet Brands obtained that information, but it is alleged that the company “as early as August, 2010, knew that two individuals, Lavont Flanders and Emerson Callum, had been criminally charged in this scheme, and further knew from the criminal charges, the particular details

1 Given the serious nature of the allegations, we note that Internet Brands has specifically denied substantially all of the allegations, including that the assailants contacted Plaintiff through the website. 6 DOE V. INTERNET BRANDS, INC. of the scheme, including how MODELMAYHEM.COM had been used in the scheme and its members victimized.” Specifically, it is alleged that Internet Brands knew that:

a. Lavont Flanders and Emerson C a l l u m w o u l d c o n t a c t f e m a l e MODELMAYHEM.COM members, using fake identities, disguised as talent scouts.

b. Lavont Flanders and Emerson Callum would lure female MODELMAYHEM.COM members to South Florida to participate in fake auditions for a fraudulent modeling contract opportunity.

c. Lavont Flanders and Emerson Callum would drug the female MODELMAYHEM.COM members with a date-rape drug during the fake audition.

d. Emerson Callum would then rape the unknowingly drugged women.

e. Lavont Flanders and Emerson Callum would record the rape on video camera.

f. Lavont Flanders and Emerson Callum would produce the rape videos and distribute the video on the internet, guised as consensual hardcore pornography.

It is also alleged that Internet Brands sued the Waitts in August 2010 for failing to disclose the potential for civil suits arising from the activities of Flanders and Callum. DOE V. INTERNET BRANDS, INC. 7

The reference to criminal charges suggests that the information was obtained by Internet Brands from an outside source, not from monitoring postings on the Model Mayhem website. As noted above, Flanders and Callum did not post on the website.

In February 2011, several months after Internet Brands had learned about the criminal activity, Flanders, pretending to be a talent scout and using a false identity, contacted Jane Doe, in the words of the Complaint, “through” the Model Mayhem website.2 Jane Doe went to south Florida for a purported audition, where Flanders and Callum drugged, raped, and recorded her.

Jane Doe filed this diversity action against Internet Brands in the Central District of California, where Internet Brands is based, asserting one count of negligent failure to warn under California law. She alleges that Internet Brands knew about the activities of Flanders and Callum but failed to warn Model Mayhem users that they were at risk of being victimized. She further alleges that this failure to warn caused her to be a victim of the rape scheme.

Internet Brands filed a motion to dismiss the action under Federal Rule of Civil Procedure 12(b)(6), on the ground that her claim was barred by the CDA. The district court granted the motion to dismiss and dismissed the action with prejudice. It denied leave to amend the complaint on the

2 Internet Brands has contended that Jane Doe was contacted directly by her assailants, not through the website. At oral argument, counsel for Jane Doe may have agreed that the contact was outside the website. This distinction does not affect our conclusion. 8 DOE V. INTERNET BRANDS, INC. ground that any amendment would be futile. Jane Doe appeals.

II. Discussion

We review de novo a district court’s decision to grant a motion to dismiss. Edwards v. Marin Park, Inc., 356 F.3d 1058, 1061 (9th Cir. 2004). We also review de novo questions of statutory interpretation. United States v. Harvey, 659 F.3d 1272, 1274 (9th Cir. 2011).

California law imposes a duty to warn a potential victim of third-party harm when a person has a “special relationship to either the person whose conduct needs to be controlled or . . . to the foreseeable victim of that conduct.” Tarasoff v. Regents of Univ. of California, 17 Cal.3d 425, 435 (1976), superseded by statute, Cal. Civ. Code § 43.92. Jane Doe alleges that Internet Brands had a cognizable “special relationship” with her and that its failure to warn her of Flanders and Callum’s rape scheme caused her to fall victim to it. Internet Brands argues that the CDA precludes the claim. Although we assume that Internet Brands may contest the scope of the duty to warn under California law and, in particular, the existence of the required special relationship, that issue is not before us. The dismissal of the action by the district court was based entirely on the CDA.

The question before us, therefore, is whether the CDA bars Jane Doe’s negligent failure to warn claim under California law. We begin with the language of the statute. Campbell v. Allied Van Lines Inc., 410 F.3d 618, 620 (9th Cir. 2005). DOE V. INTERNET BRANDS, INC. 9

Section 230(c) of the CDA, is titled “Protection for ‘Good Samaritan’ blocking and screening of offensive material.” It provides two types of protection from civil liability, but only the first type is relevant to this case:

(1) Treatment of publisher or speaker

No provider or user of an interactive computer service shall be treated as the publisher or speaker of any information provided by another information content provider.

The preemptive effect of this subsection is express: “No cause of action may be brought and no liability may be imposed under any State or local law that is inconsistent with this section.” Section 230(e)(3).

Separated into its elements, subsection (c)(1) precludes liability for “(1) a provider or user of an interactive computer service (2) whom a plaintiff seeks to treat, under a state law cause of action, as a publisher or speaker (3) of information provided by another information content provider.” Barnes v. Yahoo!, Inc., 570 F.3d 1096, 1100–01 (9th Cir. 2009) (footnote omitted). Thus, section 230(c)(1) precludes liability that treats a website as the publisher or speaker of information users provide on the website. In general, this section protects websites from liability for material posted on the website by someone else.

The first element is satisfied in this case because Internet Brands is a provider of an interactive computer service as that 10 DOE V. INTERNET BRANDS, INC. term is defined in section 230(f)(2).3 The essential question, then, is whether Plaintiff’s failure to warn cause of action “inherently requires the court to treat” Internet Brands “as a publisher or speaker” “of information provided by another information content provider.” Barnes, 570 F.3d at 1100–02. Put differently, the case turns on whether it would be inconsistent with section 230(c)(1) for the State of California to require an interactive computer service provider to warn its users about the threat of a known sexual predator.

A clear illustration of a cause of action that treats a website proprietor as a publisher is a defamation action founded on the hosting of defamatory third-party content. See, e.g., Carafano v. Metrosplash.com, Inc., 339 F.3d 1119 (9th Cir. 2003). In such circumstances, the protections of section 230(c)(1) apply, and they continue to apply even if the website proprietor has not acted to remove offensive content posted by others. For example, this court has held that the CDA barred a negligent undertaking claim against a website that failed to remove an offensive profile posted on the website by the victim’s ex-boyfriend. Barnes, 570 F.3d at 1101–03. Such liability, the court explained, would “treat” the website as the “publisher” of user content because “removing content is something publishers do” and to permit liability for such conduct “necessarily involves treating the liable party as a publisher of the content it failed to remove.” Id. at 1103.

3 “The term ‘interactive computer service’ means any information service, system, or access software provider that provides or enables computer access by multiple users to a computer server, including specifically a service or system that provides access to the Internet and such systems operated or services offered by libraries or educational institutions.” Section 230(f)(2). DOE V. INTERNET BRANDS, INC. 11

Jane Doe’s claim is different, however. She does not seek to hold Internet Brands liable as a “publisher or speaker” of content someone posted on the Model Mayhem website, or for Internet Brands’ failure to remove content posted on the website. Jane Doe herself posted her profile, but she does not seek to hold Internet Brands liable for its content. Nor does she allege that Flanders and Callum posted anything to the website. The Complaint alleges only that “JANE DOE was contacted by Lavont Flanders through MODELMAYHEM.COM using a fake identity.” Jane Doe does not claim to have been lured by any posting that Internet Brands failed to remove. Internet Brands is also not alleged to have learned of the predators’ activity from any monitoring of postings on the website, nor is its failure to monitor postings at issue.

Instead, Jane Doe attempts to hold Internet Brands liable for failing to warn her about information it obtained from an outside source about how third parties targeted and lured victims through Model Mayhem. The duty to warn allegedly imposed by California law would not require Internet Brands to remove any user content or otherwise affect how it publishes or monitors such content.

Any alleged obligation to warn could have been satisfied without changes to the content posted by the website’s users and without conducting a detailed investigation. Internet Brands could have given a warning to Model Mayhem users, perhaps by posting a notice on the website or by informing users by email what it knew about the activities of Flanders and Callum. Posting or emailing such a warning could be deemed an act of publishing information, but section 230(c)(1) bars only liability that treats a website as a publisher or speaker of content provided by somebody else: 12 DOE V. INTERNET BRANDS, INC. in the words of the statute, “information provided by another information content provider.” 47 U.S.C. § 230(c)(1). A post or email warning that Internet Brands generated would involve only content that Internet Brands itself produced. Therefore, an alleged tort based on a duty that would require such a self-produced warning falls outside of section 230(c)(1).

In sum, Jane Doe’s negligent failure to warn claim does not seek to hold Internet Brands liable as the “publisher or speaker of any information provided by another information content provider.” Id. As a result, we conclude that the CDA does not bar this claim.

The core policy of section 230(c)(1) supports this conclusion. As the heading to section 230(c) indicates, the purpose of that section is to provide “[p]rotection for ‘Good Samaritan’ blocking and screening of offensive material.” That means a website should be able to act as a “Good Samaritan” to self-regulate offensive third party content without fear of liability. In particular, section 230 was in part a reaction to Stratton Oakmont, Inc. v. Prodigy Servs. Co., 1995 WL 323710 (N.Y. Sup. Ct. May 24, 1995) (unpublished), a New York state court decision holding that an internet service provider became a “publisher” of offensive content on its message boards because it deleted some offensive posts but not others. Id. at *4. Under Stratton Oakmont’s reasoning, a website had to choose between voluntarily removing some offensive third party content, which would expose the site to liability for the content it did not remove, or filtering nothing, which would prevent liability for all third party content. See id. “In passing section 230, Congress sought to spare interactive computer services this grim choice by allowing them to perform some editing on DOE V. INTERNET BRANDS, INC. 13 user-generated content without thereby becoming liable for all defamatory or otherwise unlawful messages that they didn’t edit or delete.” Fair Housing Council v. Roommates.Com, LLC, 521 F.3d 1157, 1163 (9th Cir. 2008) (en banc) (hereafter Roommates.Com). Simply put, the immunity provision was “enacted to protect websites against the evil of liability for failure to remove offensive content.” Id. at 1174.

Jane Doe’s failure to warn claim has nothing to do with Internet Brands’ efforts, or lack thereof, to edit, monitor, or remove user generated content. Plaintiff’s theory is that Internet Brands should be held liable, based on its knowledge of the rape scheme and its “special relationship” with users like Jane Doe, for failing to generate its own warning. Thus, liability would not discourage the core policy of section 230(c), “Good Samaritan” filtering of third party content.

Another policy of section 230 is to “avoid the chilling effect upon Internet free speech that would be occasioned by the imposition of tort liability upon companies that do not create potentially harmful messages but are simply intermediaries for their delivery.” Delfino v. Agilent Techs., Inc., 52 Cal. Rptr. 3d 376, 387 (Ct. App. 2006). As section 230(b) itself explains, “[i]t is the policy of the United States . . . to promote the continued development of the Internet . . . [and] to preserve the vibrant and competitive free market that presently exists for the Internet and other interactive computer services, unfettered by Federal or State regulation.” Jane Doe’s cause of action does not seek to impose “intermediary” liability. Although Internet Brands may have been an “intermediary” between Jane Doe and the rapists in a broad sense, there is no allegation that Model Mayhem transmitted any potentially harmful messages between Jane 14 DOE V. INTERNET BRANDS, INC.

Doe and Flanders or Callum. There is also no allegation that Flanders or Callum posted their own profiles on the website. That Internet Brands was in some sense an “intermediary” between Jane Doe and the rapists simply does not mean that the failure to warn claim treats Internet Brands as the publisher or speaker of user content.

It may be true that imposing any tort liability on Internet Brands for its role as an interactive computer service could be said to have a “chilling effect” on the internet, if only because such liability would make operating an internet business marginally more expensive. But such a broad policy argument does not persuade us that the CDA should bar the failure to warn claim. We have already held that the CDA does not declare “a general immunity from liability deriving from third-party content.” Barnes, 570 F.3d at 1100. “[T]he Communications Decency Act was not meant to create a lawless no-man’s-land on the Internet.” Roommates.Com, 521 F.3d at 1164. Congress has not provided an all purpose get- out-of-jail-free card for businesses that publish user content on the internet, though any claims might have a marginal chilling effect on internet publishing businesses. Moreover, the argument that our holding will have a chilling effect presupposes that Jane Doe has alleged a viable failure to warn claim under California law. That question is not before us and remains to be answered.

Barring Jane Doe’s failure to warn claim would stretch the CDA beyond its narrow language and its purpose. To be sure, Internet Brands acted as the “publisher or speaker” of user content by hosting Jane Doe’s user profile on the Model Mayhem website, and that action could be described as a “but-for” cause of her injuries. Without it, Flanders and Callum would not have identified her and been able to lure DOE V. INTERNET BRANDS, INC. 15 her to their trap. But that does not mean the failure to warn claim seeks to hold Internet Brands liable as the “publisher or speaker” of user content.

Publishing activity is a but-for cause of just about everything Model Mayhem is involved in. It is an internet publishing business. Without publishing user content, it would not exist. As noted above, however, we held in Barnes that the CDA does not provide a general immunity against all claims derived from third-party content. In that case we affirmed the dismissal of a claim for negligent undertaking as barred under the CDA, as discussed above at 10, but we reversed the dismissal of a claim for promissory estoppel under Oregon law. The publication of the offensive profile posted by the plaintiff’s former boyfriend was a “but-for” cause there, as well, because without that posting the plaintiff would not have suffered any injury. But that did not mean that the CDA immunized the proprietor of the website from all potential liability. As we observed in Roommates.Com, “we must be careful not to exceed the scope of the immunity provided by Congress.” 521 F.3d at 1164 n.15. Congress could have written the statute more broadly, but it did not.

The parties discuss other court decisions regarding the CDA in their briefs. The case law provides no close analogies, though, because the cases are all distinguishable in critical respects. For example, the purported tort duty does not arise from allegations about mishandling the removal of third party content. Barnes, 570 F.3d at 1105–06 (holding that the CDA bars negligent undertaking claim arising from Yahoo’s failure to take reasonable care in removing offensive profiles). Nor is there a contractual duty arising from a promise distinct from tort duty arising from publishing 16 DOE V. INTERNET BRANDS, INC. conduct. Id. at 1108–09 (holding that the CDA does not bar a promissory estoppel claim).

The tort duty asserted here does not arise from an alleged failure to adequately regulate access to user content or to monitor internal communications that might send up red flags about sexual predators. Doe II v. MySpace, Inc., 175 Cal.App.4th 561, 573 (Ct. App. 2009) (holding that the CDA bars tort claims based on a duty to restrict access to minors’ MySpace profiles); Doe v. MySpace, Inc., 528 F.3d 413 (5th Cir. 2008) (holding that CDA bars claims for negligence and gross negligence in not preventing a 13 year old girl from lying about her age to create a personal profile that led to contact by a sexual predator). Jane Doe alleges actual knowledge by Internet Brands from an outside source of information about criminal activity.

This case does not concern an employer-employee relationship giving rise to a negligent supervision claim. Lansing v. Southwest Airlines Co., 980 N.E.2d 630, 639–41 (Ill. Ct. App. 2012) (holding that the CDA does not bar a negligent supervision claim against an airline whose employee used the company email and text messaging systems to harass the plaintiff).

In short, this case presents the novel issue of whether the CDA bars Jane Doe’s failure to warn claim under California law. We conclude that it does not.

III. Conclusion

The CDA does not bar Jane Doe’s failure to warn claim. We express no opinion on the viability of the failure to warn allegations on the merits. We hold only that the CDA is not DOE V. INTERNET BRANDS, INC. 17 a valid basis to dismiss Jane Doe’s complaint. Accordingly, we reverse and remand for proceedings consistent with this opinion.

REVERSED AND REMANDED.