What it means to “work” is changing. Many have eschewed the traditional nine-to-five job for more flexible, often one-time “gigs”: delivering groceries, fixing a sink, driving a stranger to the airport. Supported by online intermediaries such as Uber and Task Rabbit, this growing gig economy^1×1. See Seth D. Harris & Alan B. Krueger, The Hamilton Project, A Proposal for Modernizing Labor Laws for Twenty-First-Century Work: The “Independent Worker” 12 (2015) (estimating that at least 600,000 workers participate in the online gig economy); McKinsey Global Institute, A Labor Market That Works 33 (2015) (noting that “contingent workers” in “digital marketplaces” are a “rapidly growing” segment of the U.S. working-age population). offers greater freedom for workers and businesses alike. But at the same time, it has raised questions about the existing categories of employment law. Already, disputes have ignited over whether gig workers are employees or merely independent contractors under the law.^2×2. The distinction is important. Employees are entitled to all the benefits and protections associated with this status under various laws, including minimum wages, overtime compensation, and the right to collective bargaining. Contractors are not. As such, many gig workers have initiated misclassification suits against online intermediaries such as Uber and Lyft. See, e.g., O’Connor v. Uber Techs., Inc., No. C-13-3826, 2015 WL 5138097 (N.D. Cal. Sept. 1, 2015); Cotter v. Lyft, Inc., 60 F. Supp. 3d 1067 (N.D. Cal. 2015). Traditionally, one of the most crucial factors indicating the existence of an employment relationship has been the worker’s loss of autonomy.^3×3. See Harris & Krueger, supra note 1, at 8 tbl.1. The challenge in the gig economy — where everyone gets to “be their own boss”^4×4. Uber and the American Worker: Remarks from David Plouffe, Uber: Newsroom (Nov. 3, 2015), https://newsroom.uber.com/1776 [http://perma.cc/XZ6F-BLCV]. Uber has reported that “nearly 90 percent of [its] drivers choose Uber because they want to be their own boss and set their own schedule.” Id. — is to determine who, if anyone, still counts as an employee under this test. Recently, in Sisters’ Camelot,^5×5. 363 N.L.R.B. No. 13 (Sept. 25, 2015). the National Labor Relations Board did just that, finding that door-to-door canvassers with flexible work schedules were employees within the meaning of the National Labor Relations Act.^6×6. 29 U.S.C. §§ 151–169 (2012); id. at 1. The Board concluded that the canvassers — while free to choose when and how much to work — still faced many of the same constraints as traditional employees and were therefore entitled to the same protections.^7×7. See id. at 2–6. The Board’s decision demonstrates how existing legal categories can and do incorporate nontraditional workers. Flexible or not, many relationships in the gig economy are still shaped by the very power asymmetries that define an employment relationship — and should be recognized as such.

Sisters’ Camelot is a nonprofit organization in Minnesota that distributes food to low-income individuals.^8×8. Id. at 1. It funds its operations almost entirely through donations collected by door-to-door canvassers, who operate on a flexible schedule and can choose whether to work on any given day.^9×9. Id. Those who do decide to work must show up at the Camelot facility at a particular time to be transported to their designated canvassing area.^10×10. Id. Canvassers can solicit donations only within their assigned area, and only for Camelot — they are prohibited from soliciting for other organizations at the same time, and can be disciplined for doing so.^11×11. Id. at 2, 4. They are also required to keep detailed records on each house they visit.^12×12. Id. at 1, 2. At the end of their shift, the canvassers deliver their collected donations and are transported back to the facility.^13×13. Id. at 1. Each canvasser receives a nonnegotiable commission based on individual donations collected.^14×14. Id. at 4.

Christopher Allison was one of these canvassers. In 2013, Camelot learned that Allison was involved in efforts to organize a canvassers’ union.^15×15. Id. at 6. He was fired.^16×16. Id. His fellow canvassers were warned that, despite their efforts, they could not “force [Camelot] into the role of bosses.”^17×17. Id. The canvassers filed a complaint with the Board, alleging violations of their right to organize under the National Labor Relations Act.^18×18. The canvassers alleged that Sisters’ Camelot committed unfair labor practices by terminating Allison, by warning the canvassers that their organizing efforts would be futile, and by granting them certain benefits in order to dissuade them from further union activities. Id. at 1; see also 29 U.S.C. §§ 157–158 (2012). An essential premise of their argument was that they were employees within the meaning of the Act and entitled, as such, to protection from unfair labor practices.^19×19. See 29 U.S.C. § 157 (“Employees shall have the right to self-organization . . . .” (emphasis added)).

An administrative law judge dismissed their complaint,^20×20. Sisters’ Camelot, No. 18-CA-100514, slip op. at 7 (NLRB Div. of Judges Aug. 7, 2013). ruling that the canvassers were independent contractors, not employees.^21×21. Id. Recognizing, however, that “a reviewing body may conclude otherwise,” id., the judge also addressed the merits of the complaint, finding that — if the canvassers were employees — Camelot did indeed violate their rights under the Act, id. at 9. Crucial to this conclusion were the judge’s findings that the canvassers were in control of their own schedules and lacked direct supervision.^22×22. See id. at 7. The judge emphasized that the canvassers were free to make important choices: whether to show up for their shifts and, during their shifts, whether to “work . . . or goof-off.”^23×23. Id. at 4. Under his analysis, these facts “strongly indicate[d] independent contractor status.”^24×24. Id. The judge applied the Board’s longstanding multifactor test for determining whether an individual is an employee or an independent contractor. See id. At the time of his decision, the test consisted of ten factors: (1) the extent of control by the employer; (2) whether the individual is engaged in a distinct occupation or business; (3) the extent of supervision by the employer; (4) the skill required in the occupation; (5) whether the employer supplies the instrumentalities, tools, and place of work; (6) the length of time the individual is employed; (7) the method of payment; (8) whether the work is part of the regular business of the employer; (9) the understanding between the parties; and (10) whether the employer is in the business. See Roadway Package Sys., Inc., 326 N.L.R.B. 842, 849 n.32 (1998). The judge found that all but two factors (the skill required and whether the employer supplied the instrumentalities of work) favored independent contractor status. Sisters’ Camelot, slip op. at 3–7.

On review, a three-member panel of the Board^25×25. The panel consisted of Chairman Mark Gaston Pearce, Kent Hirozawa, and Lauren McFerran. reversed the judge’s dismissal, finding that the canvassers were employees protected under the Act.^26×26. See Sisters’ Camelot, 363 N.L.R.B. No. 13, at 1, 5–6. The Board also agreed with the judge that Camelot had violated the Act by terminating Allison and by its statement of futility. Id. at 6. Consistent with its longstanding approach, the Board undertook a holistic consideration of “all of the incidents of the relationship”^27×27. Id. at 2 (quoting FedEx Home Delivery, 361 N.L.R.B. No. 55, at 1 (Sept. 30, 2014)); see also NLRB v. United Ins. Co. of Am., 390 U.S. 254, 258 (1968). between Camelot and its canvassers. It applied a multifactor test and concluded that nine out of the eleven relevant factors indicated the existence of an employment relationship.^28×28. The Board applied the multifactor test as articulated and refined in FedEx Home Delivery, 361 N.L.R.B. No. 55, a decision issued after the judge’s ruling in Sisters’ Camelot. See Sisters’ Camelot, 363 N.L.R.B. No. 13, at 2. Analyzing the same ten factors as the judge below, the Board found that all but two (length of employment and the parties’ understanding of the relationship) weighed in favor of employee status. Id. at 5. The Board also considered an eleventh factor set forth in FedEx — whether the individual in question is rendering services as an independent business — and found that this, too, indicated an employment relationship. See id. In stark contrast to the judge’s analysis below, the Board determined that — notwithstanding the canvassers’ “freedom to work or not work as they choose”^29×29. Sisters’ Camelot, 363 N.L.R.B. No. 13, at 2. — Camelot still imposed significant constraints on the canvassers’ choices and opportunities. The Board emphasized that the canvassers, once on the clock, worked at set times and locations.^30×30. See id. Their compensation was nonnegotiable.^31×31. See id. at 4. Detailed records of their canvassing allowed Camelot to supervise their activities with “close scrutiny.”^32×32. Id. at 5; see also id. at 3. And canvassers had no control over “important business decisions” that could affect their profit or loss: where, how, and from whom to solicit.^33×33. Id. at 5. Considered together, these facts all pointed toward an employment relationship.

The Board’s decision offers important lessons for the gig economy. By recognizing the canvassers as employees, the Board signaled its continued commitment^34×34. Sisters’ Camelot is consistent with a series of NLRB decisions conferring employee status on part-time or otherwise flexible workers. See, e.g., FedEx Home Delivery, 361 N.L.R.B. No. 55 (Sept. 30, 2014) (FedEx delivery drivers); Lancaster Symphony Orchestra, 357 N.L.R.B. 1761 (2011) (orchestra musicians); Roadway Package Sys., Inc., 326 N.L.R.B. 842 (1998) (delivery drivers). to accommodating — and protecting — nontraditional forms of work. The flexible relationship between Camelot and its canvassers invites comparisons to the broader gig economy.^35×35. Some commentators have already picked up on the decision’s potential implications for Uber drivers. See, e.g., Brian Mahoney, NLRB Speaks on Indy Contractors, Politico (Sept. 28, 2015, 10:00 AM), http://www.politico.com/tipsheets/morning-shift/2015/09/nlrb-v-gig-economy-uaw-fiat-deal-in-jeopardy-joint-employer-update-210419 [http://perma.cc/8USC-SBUM]; Benjamin Sachs, New NLRB Decision on Independent Contractors, On Labor (Sept. 28, 2015), http://onlabor.org/2015/09/28/new-nlrb-decision-on-independent-contractors [http://perma.cc/EH4R-3HLU]. And while the Board’s specific findings are highly fact bound, its broader reasoning provides useful guidance on how to categorize workers in the gig economy. In Sisters’ Camelot, the Board made clear that, despite shifting boundaries, the existing categories of “employee” and “independent contractor” are still viable — capacious enough to capture new, unconventional work patterns, but narrow enough to confer protection only on those relationships that warrant it.

The Board’s premise — a fundamental one in employment law — is that all workers fall on one side of a distinct dichotomy: employee, or independent contractor. But when it comes to gig workers, this premise has been challenged. Commentators have been quick to point out that many aspects of the gig economy — the flexible schedules, the lack of in-person supervision, the ability to work for many businesses at the same time — defy simple classification.^36×36. See, e.g., Harris & Krueger, supra note 1, at 6–10; Greg Ip, As the Gig Economy Changes Work, So Should Rules, Wall St. J. (Dec. 9, 2015, 12:49 PM), http://www.wsj.com/articles/as-the-gig-economy-changes-work-so-should-rules-1449683384. Some decisionmakers agree, contending that the current legal framework is too rigid to accommodate the now-countless variations of work. In a lawsuit against Lyft, a federal judge lamented that the categorization of gig workers is an untenable endeavor — like being “handed a square peg and asked to choose between two round holes.”^37×37. Cotter v. Lyft, Inc., 60 F. Supp. 3d 1067, 1081 (N.D. Cal. 2015). Professors Seth Harris and Alan Krueger have proposed creating a new legal category, the “independent worker,” for those who occupy the “middle ground between traditional employees and independent contractors.”^38×38. Harris & Krueger, supra note 1, at 5. Their paradigmatic example is the Uber driver. Id. Their argument is that the existing legal framework cannot adequately identify who deserves which benefits and protections under the law, creating uncertainty, inefficiency, and opportunities for regulatory arbitrage.^39×39. See id. at 6–7. In their view, working relationships in the gig economy are “not so dependent, deep, extensive, or long lasting” that businesses like Uber should be compelled to “assume responsibility for all aspects of . . . workers’ economic security”^40×40. Id. at 8. Harris and Krueger argue that independent workers should be entitled to some, but not all, of the benefits associated with employee status. See id. at 15–21. — that is, to take on “the role of bosses.”^41×41. Sisters’ Camelot, 363 N.L.R.B. No. 13, at 6.

What complicates this view is that, in reality, many workers in the gig economy are deeply dependent on the businesses that hire them. Harris and Krueger characterize the many laws that regulate the employment relationship as reflecting a “social compact between employees and employers,”^42×42. Harris & Krueger, supra note 1, at 6. with employees sacrificing their autonomy in exchange for some degree of economic security.^43×43. See id. at 7. Some gig workers find themselves making exactly this kind of sacrifice, but without receiving protection in return.^44×44. Professor Benjamin Sachs has applied this argument to Uber drivers. See, e.g., Benjamin Sachs, Do We Need an “Independent Worker” Category?, On Labor (Dec. 8, 2015), https://onlabor.org/2015/12/08/do-we-need-an-independent-worker-category/ [http://perma.cc/VN3U-3RTE].

Moreover, Harris and Krueger premise their argument on the idea that gig workers “do not fit easily into the existing legal definitions of ‘employee’ and ‘independent contractor’ status.”^45×45. Harris & Krueger, supra note 1, at 2. Granted, the particular details of work might look different in the gig economy, but, as the Board has consistently recognized, careful consideration of the relevant factors^46×46. The test for determining employee status varies across statutes and jurisdictions, but most courts and agencies emphasize similar factors, such as control. See, e.g., Nationwide Mut. Ins. Co. v. Darden, 503 U.S. 318, 323–24 (1992) (articulating a multifactor test for determining employee status under ERISA, including the “right to control the manner and means” of work, id. at 323 (quoting Cmty. for Creative Non-Violence v. Reid, 490 U.S. 730, 751 (1989))); Sec’y of Labor v. Lauritzen, 835 F.2d 1529, 1534–35 (7th Cir. 1987) (articulating a six-factor “economic reality” test, id. at 1534, for determining employee status under the Fair Labor Standards Act (FLSA), including “the nature and degree of the alleged employer’s control,” id. at 1535); see also S.G. Borello & Sons, Inc. v. Dep’t of Indus. Relations, 769 P.2d 399, 404 (Cal. 1989) (emphasizing “the right to control work details,” among other factors indicating employee status); Rev. Rul. 87-41, 1987-1 C.B. 296 (outlining the Internal Revenue Service’s twenty-factor test for determining “whether sufficient control is present to establish an employer-employee relationship”). can still uncover employment relationships where they exist. In Sisters’ Camelot, the Board’s analysis of three factors in particular — control, supervision, and entrepreneurial opportunity — shows how.^47×47. For a discussion on how the Board’s analysis of these factors sheds light on the classification of Uber drivers, see Sachs, supra note 35.

One of the most important factors indicating an employment relationship is control. Traditionally, employers have exercised control over their employees by dictating wages, hours, and working conditions. In the gig economy, none of these usual mechanisms are a given. Many gig workers today have broad discretion over when, where, and how much to work. The Board’s analysis in Sisters’ Camelot, however, suggests that such discretion does not necessarily deprive a worker of employee status. Like many gig workers, Camelot’s canvassers were “free[] to work or not work as they [chose],”^48×48. Sisters’ Camelot, 363 N.L.R.B. No. 13, at 2. but the Board concluded that “such discretion [was] outweighed by the control”^49×49. Id. at 2 n.5. that Camelot otherwise exercised over their work lives — for example, by setting their daily start and end times, requiring detailed recordkeeping, restricting their canvassing area, and disciplining them.^50×50. Id. at 2. The key inquiry for the Board was not whether Camelot controlled the canvassers’ schedules, but rather how much control it exercised when the canvassers did decide to work. The Board opted for a broad understanding of control, recognizing that, especially in nontraditional work arrangements, it can often take atypical forms.

Another factor often used to distinguish employees from independent contractors is supervision. Assessing this factor can be difficult in the gig economy, where online intermediaries dominate and in-person supervision is rare. Lack of direct supervision does not, however, preclude a worker from employee status. In Sisters’ Camelot, the Board acknowledged that the canvassers were “not generally subject to in-person supervision” — in part because it would have been “highly impractical” — but still found a “significant level of oversight” in the relationship, achieved through detailed recordkeeping requirements.^51×51. Id. at 3. Notably, the Board measured supervision in relative terms, based on what was practicable in that context.^52×52. See id.; see also Mitchell Bros. Truck Lines, 249 N.L.R.B. 476, 481 (1980) (evaluating supervision in the context of “the nature of the occupation”), overruled in part on other grounds by Container Transit, Inc., 281 N.L.R.B. 1039 (1986). This approach offers a means of distinguishing between the many online intermediaries in the gig economy. While many of them maintain a strict hands-off model, some of them — such as Uber and Instacart — have detailed feedback mechanisms to enable remote monitoring,^53×53. See, e.g., O’Connor v. Uber Techs., Inc., No. C-13-3826, 2015 WL 5138097, at *19 (N.D. Cal. Sept. 1, 2015) (“Uber maintains a uniform ability to monitor certain aspects of its drivers’ performance, principally through the star rating system and other rider feedback . . . .”). suggesting the existence of an employment relationship.

One final factor bearing on a worker’s status is his entrepreneurial opportunity for profit or gain. For many gig workers, a significant draw is the potential to “make good money”^54×54. Sign Up to Drive with Uber, Uber, https://get.uber.com/drive [http://perma.cc/U32L-SA7H]. Uber entices drivers with the promise of substantial profits: “Got a car? Turn it into a money machine.” Id. — conditioned, of course, on how much they choose to work. But their opportunities for profit are limited in other ways, too. In Sisters’ Camelot, the Board demonstrated a strong awareness of these constraints, stressing that the relevant test is “actual, not merely theoretical, entrepreneurial opportunity.”^55×55. Sisters’ Camelot, 363 N.L.R.B. No. 13, at 5 (emphasis added) (citing FedEx Home Delivery, 361 N.L.R.B. No. 55, at 1 (Sept. 30, 2014)). Although Camelot’s canvassers were free to work for other organizations, these opportunities for extra profit were outweighed, in the Board’s view, by the substantial limits imposed on entrepreneurial initiative: commission rates were nonnegotiable, and canvassers were prohibited from soliciting outside their assigned areas or for other organizations during their shifts.^56×56. Id. at 4–5. There was little that the canvassers could do — besides working harder, in theory — to increase their profits. The Board’s actual-opportunity test offers another means for distinguishing between the different kinds of relationships in the gig economy. On this factor, setting fixed rates, like Uber does,^57×57. E.g., O’Connor, 2015 WL 5138097, at *17 (“[T]he evidence is clear that Uber sets its drivers’ pay without any input or negotiation from the drivers.”). might indicate an employment relationship, whereas using a bidding model or allowing workers to set their own prices — like Agent Anything^58×58. See How It Works, Agent Anything, http://www.agentanything.com/how-it-works [http://perma.cc/WL7H-DMDX]. and Sidecar^59×59. See Take Control with the New Sidecar, Sidecar, http://marketplace.side.cr/driver [http://perma.cc/9SC8-TYSQ]. do — might not.

The Board’s analysis of these three factors makes clear that many of the distinctive features of the gig economy — the flexible schedules, the lack of in-person supervision, and the promise of entrepreneurial opportunity — do not disqualify its workers from employee status. On the contrary. When applied with nuance, as in Sisters’ Camelot, the Board’s multifactor test reveals that even nontraditional “gigs” can impose significant constraints on a worker’s autonomy. These constraints are precisely what define an employment relationship.

What will be difficult for decisionmakers, going forward, is to recognize those constraints in new and diverse work arrangements. Gig workers can fit into the traditional dichotomy of employees and independent contractors, but, as the detailed analysis in Sisters’ Camelot shows, not without close scrutiny of the facts.^60×60. Not all of the facts found in Sisters’ Camelot are applicable to all gig workers. For example, the canvassers were unable to choose their “daily start and end times,” Sisters’ Camelot, 363 N.L.R.B. No. 13, at 2, and did not provide their own tools, “only their own pens,” id. at 3. For workers who have more flexible schedules, or who provide their own tools — using, for example, their own cars — questions about their legal status might still remain. Harris and Krueger have argued that the classification of gig workers will require “long, costly, and uncertain legal battles,”^61×61. Harris & Krueger, supra note 1, at 6. and that persistent uncertainty about their status exacerbates the risk of misclassification, with businesses “reorganiz[ing] their work to . . . avoid providing required benefits and protections.”^62×62. Id. at 7. Reliance on complex, multifactor tests only adds to the uncertainty.^63×63. This concern is not new. Judge Easterbrook raised it in Secretary of Labor v. Lauritzen, 835 F.2d 1529 (7th Cir. 1987), where the Seventh Circuit determined that migrant pickle pickers were employees under the FLSA. See id. at 1534–38. In his concurrence, Judge Easterbrook lamented that the court’s multifactor test “offer[ed] little guidance for future cases,” making it difficult for businesses to structure their behavior. Id. at 1539 (Easterbrook, J., concurring). “Why keep cucumber farmers in the dark about the legal consequences of their deeds?” he asked. Id. The existing categories may still work, but for them to work well, legislatures, courts, and agencies may want to consider stronger decisionmaking rules — such as explicit statutory protections for gig workers^64×64. Recently, the Seattle City Council unanimously passed a bill allowing drivers for Uber, Lyft, and other ride-hailing apps to unionize. See Nick Wingfield & Mike Isaac, Seattle Will Allow Uber and Lyft Drivers to Form Unions, N.Y. Times (Dec. 14, 2015), http://www.nytimes.com/2015/12/15/technology/seattle-clears-the-way-for-uber-drivers-to-form-a-union.html. or, more broadly, a default presumption in favor of employee status.^65×65. Some countries, including Mexico, the Netherlands, and Portugal, have adopted statutes establishing a rebuttable presumption in favor of employee status. See Harris & Krueger, supra note 1, at 6. Recently, the U.S. Department of Labor issued a new interpretation of the FLSA that represented a small step in this direction, emphasizing that the Act was “specifically designed to ensure as broad of a scope of statutory coverage as possible” and should be interpreted to include “most workers” as protected employees. See U.S. Dep’t of Labor, Wage & Hour Div., Administrator’s Interpretation No. 2015-1 (July 15, 2015).

Without a doubt, the gig economy will continue to test our understanding of what “work” means. And the classification of its workers will present distinct challenges. But the gig worker is not a “square peg.”^66×66. Contra Cotter v. Lyft, Inc., 60 F. Supp. 3d 1067, 1081 (N.D. Cal. 2015). As the Board demonstrated in Sisters’ Camelot, even nontraditional working relationships can fit comfortably within the existing framework of employment.