THE DISSENT IN THE POSTMATES CASE

This decision is apparently pre-COVID so what will be the future of the "gig economy" is anyone's guess. Only the dissent is being presented here.

MATTER OF VEGA (POSTMATES INC.—COMMISSIONER OF LABOR), 2020 NY Slip Op 2094 - NY: Court of Appeals March 26, 2020:

WILSON, J. (dissenting): NOTE:Judge Wilson dissents and votes to affirm in an opinion in which Judge Garcia concurs.

The majority's opinion suffers from two independent defects. The first is a failure to examine the record to determine whether the findings of the Commissioner of Labor were supported by substantial evidence. Many of those findings were so lacking in support as to appear to have been cut and pasted from the decision in some other matter, or from a form list of all the possible factors that might warrant the conclusion that someone was an employee. Under those circumstances, reversal is required. The second is a failure to recognize that the realities of the contemporary working world have outpaced our jurisprudence. The multitude of factors identified in our caselaw as pertinent to determining whether a claimant is an employee or independent contractor — reflective of a time when employees received a gold watch upon retiring from the sole company at which they spent their entire careers — coupled with our deferential standard of review, has left only two undesirable paths open: either we adhere to the caselaw and standard of review, leaving all agency decisions unreviewable, or we make haphazard reversals without explanation, based on an ad hoc test we do not articulate because it defies explanation. We have chosen the latter approach, which has nothing to recommend it except that it is marginally better than the former.

I.

The majority begins by asserting that "[t]he issue before us is whether the decision of the Unemployment Insurance Appeals Board (the Board) that claimant, a former Postmates, Inc. courier, and others similarly-situated are employees for whom Postmates is required to make contributions to the unemployment insurance fund was supported by substantial evidence" (majority op at 1-2). That is not correct.

At the inception of the hearing before the ALJ, the following colloquy occurred:

ALJ PICHARDO: This two-page document is marked as an exhibit for the record as Hearing Exhibit 1 as of today's date. And so where on this document does it say similarly situated, Ms. Claxton?
MS. CLAXTON [Counsel for the Commissioner]: I don't see it. I don't see it, you're right.
ALJ PICHARDO: So then it's only for —
MS. CLAXTON: For the claimant.
ALJ PICHARDO: — claimant?
MS. CLAXTON: Yes, Judge.
ALJ PICHARDO: All right.

After the close of evidence, the ALJ reiterated that "my decision in this case ... only relates to Mr. Vega and not any other employees." Thus, this appeal concerns only Mr. Vega: whether there is substantial evidence supporting the determination that he was an employee of Postmates. The Board's determination, erroneously purporting to apply its decision to all similarly situated Postmates workers, is itself a freestanding basis to reject the Board's determination (see Pell v Board of Education, 34 NY2d 222, 230-231 [1974] [the action of an administrative tribunal is arbitrary when "taken without regard to the facts"]).^[12]

Anyone can download the Postmates Fleet app to become a courier for Postmates. I could be a Postmates courier, so long as I passed a criminal background check. I could make Postmates deliveries when and where I pleased, and extemporaneously indicate my availability at moments when I need a break from the press of court business. I could make my deliveries by any form of locomotion I choose: walk, bicycle, scooter, car, rollerblade, etc.

Luis Vega, at least briefly, thought more of the Postmates opportunity than did I. He downloaded the app, provided sufficient information to pass the criminal check, and was thereafter authorized to use the Postmates service to make deliveries. Mr. Vega indicated he would be walking to make deliveries. He first logged on to Postmates on June 8, 2015, and last logged on to Postmates on June 15, 2015. At that point, he had worked for Postmates for less than a week and had logged on 12 times for an average of 3 hours and 15 minutes at a time. During those six days, Mr. Vega rejected or ignored about 50% of the assignments offered to him. The record was unequivocal that, even if Mr. Vega requested to make a specific delivery assignment and obtained it, he could thereafter change his mind and reject it, causing it to be dumped back into the pool of assignments available to others. Mr. Vega had no set schedule; he had no supervisor^[13]; and, he chose what deliveries interested him, how to perform deliveries, the route he would take, and the times at which he would log on and off. Of the jobs he accepted over those six days, "a lot of requesters' feedback" indicated that "they weren't receiving the items requested." Postmates therefore blocked Mr. Vega from the app. Mr. Vega then filed for unemployment benefits.

In its determination of the matter in August 2015, the Department of Labor classified Mr. Vega as an employee for the purposes of New York Unemployment Insurance Law. The Department of Labor's determination that Mr. Vega was Postmates' employee lists 24 factors supporting that determination.

A large number of those factual findings are directly contradicted by the record^[14]. Among those factors are: Mr. Vega was told "when, where, and how the work was to be performed"; he was required to report to a supervisor and work an established schedule; he was required to deliver the packages within a set time; his work would be reviewed; he could not take time off without Postmates' approval; he was covered by a workers' compensation policy; he was not free to determine the route of the delivery; and, he could not perform other deliveries while on route with Postmates.
None of these factors has support in the record. Each of those factors was undermined by the testimony before the ALJ, who found that the Postmates couriers chose when they worked, how they worked, and where they worked; couriers could and did choose their own routes; Mr. Vega was not covered by workers' compensation; couriers were free to reject, ignore, or accept assignments as they chose; and, couriers were free to work for other companies at the same time as they worked for Postmates (id.). Although the Commissioner found that Mr. Vega "could not engage substitutes or other couriers without your permission," the record evidence was that Mr. Vega could "hand his phone to a complete stranger" to complete deliveries. Additionally, the Department noted as a factor that: "Individuals performing such services as couriers were previously determined to be your employees." However, the ALJ expressly stated that she was "not going to be considering it" because Postmates and its counsel "weren't aware of this determination." Moreover, counsel for the Commissioner stated, "I didn't really want to make it a part of the record," after which the ALJ reiterated, "that's not going to be before me."

Based on the record, which included both exhibits and testimony, the ALJ held that Mr. Vega was an independent contractor not entitled to unemployment benefits. Although the Board did not make any finding of fact contrary to the ALJ's findings, the Board reversed the ALJ's determination. The Appellate Division reversed the Board's determination for lack of substantial evidence, citing our decision in Yoga Vida (Matter of Vega v Postmates Inc., 162 AD3d 1337 [3d Dept 2018]). The court noted the lack of application or review process, the lack of supervision, the courier's choice to log on and accept delivery requests, the courier's choice of route and mode of transportation, the lack of a required uniform or identification, the payment system (allowing for payment only upon the completion of a delivery), and the lack of reimbursements for delivery-related expenses (id. at 1338-1339). Although some indicia of control remained, the court concluded it only showed incidental control, insufficient to render Mr. Vega an employee (id. at 1339). To recap: the Commissioner, the Board and majority conclude that Mr. Vega's slapdash week of activity made him Postmates' employee; the ALJ and Third Department concluded, as do I, that it did not^[15]. What accounts for that disagreement? The failure of our precedent to keep up with the times.

II.

"Any employer shall become liable for contributions under [Article 18, the Unemployment Insurance law] if it has paid remuneration of three hundred dollars or more in any calendar quarter" (Labor Law § 560 [1]). "Remuneration" in this part of the Labor Law means "every form of compensation for employment paid by an employer to his employee" (Labor Law § 517 [1]) and "employment" is defined as "any service under any contract of employment for hire, express or implied, written or oral" (Labor Law § 511 [1] [a]), subject to many additions not relevant here^[16]. Thus, Postmates' obligation to pay unemployment insurance contributions for Mr. Vega turns on whether its agreement with him was "a contract of employment for hire." That definition of "employment" is circular, so we have interpreted it to apply what is described as the "common law test" of employee status (In re Morton, 284 NY 167, 173 [1940], cf. Community for Creative Non-Violence v Reid, 490 US 730, 740 [1989]), often described as the "control test," when determining liability for unemployment insurance payments.^[17]

Under the control test of employee status, "the critical inquiry in determining whether an employment relationship exists pertains to the degree of control exercised by the purported employer over the results produced or the means used to achieve the results" (Bynog v Cipriani Group, Inc., 1 NY3d 193, 198 [2003]). We further explained that "control over the means is the more important factor to be considered," and distinguished "incidental control over the results produced without further indicia of control over the means employed to achieve the results" from other forms of control that would evince an employment relationship (Matter of Ted Is Back Corp. [Roberts], 64 NY2d at 726; see Matter of Hertz Corp. [Commissioner of Labor], 2 NY3d at 735). However, that means-ends test coexists uneasily with a separate test, the "overall control" test, where "substantial evidence of control over important aspects of the services performed other than results or means is sufficient to establish an employer-employee relationship" (Matter of Empire State Towing and Recovery Assn., Inc., 15 NY3d 433, 437-438 [2010]; Matter of Concourse Ophthalmology Assocs., P.C., 60 NY2d 734 [1983] ["The board's determination is ... supported by substantial evidence of control over important aspects of the services performed other than results or means"]). The "overall control" test is said to apply only "where the details of the work performed are difficult to control because of considerations such as professional and ethical responsibilities" (Empire State Towing, 15 NY3d at 437-438).

A.

Putting aside the clear lack of record support for a multitude of the factors relied on by the Department of Labor, the majority's conceptual error in resolving this case (and, I suspect, a good deal of the confusion below on the application of the control test) stems from a lack of decisional clarity about what factors matter when and why. Every contractual relationship for the provision of services will involve some control — whether overall or over means-ends — being ceded to the service-provider and some control being held by the requestor. No sensible enterprise gives even an indisputably independent contractor complete control over the "results produced or the means used to achieve the results." Imagine instructing a contractor to build a house, with no specification as to the size, layout, style or features to be included (ends) or a requirement that the contractor comply with local building laws (means). Absent a more defined legal standard it is unclear how much control the employer may have over an independent contractor before that contractor becomes an "employee," or, for that matter, what makes control "incidental" as compared to non-incidental.^[18]

Matters are especially unclear in the semi-professional world of the "overall control" test, which tells us only that sometimes control over "important aspects" other than ends or means matters. But even the means-ends test is no panacea. Means and ends are not perfectly polarized. Here, for example, Postmates allows its couriers to choose whatever method of delivery they wish, but had Mr. Vega opted to deliver by pogo stick, turning sushi into a poke bowl or burritos into taco salads, surely he would have been bounced from the app expeditiously. Postmates undoubtedly cares that its customers receive their dinners intact, but Postmates' concern for that end, or a ban on the means of pogo stick deliveries, does not address the question of control for employment purposes. Both the means (no pogo) and ends (no mush) would be required whether the delivery person was an independent contractor or an employee.

Because "control" standing alone is relatively unhelpful, we have responded by creating a multifactor analysis where no one factor is determinative and where, as the majority correctly observes, "no enumerated list of factors can apply to every situation faced by a worker" (majority op at 5). The Supreme Court of the United States, reviewing a similar proliferation of factors, noted dryly that "the traditional agency law criteria offer no paradigm of determinacy" (Nationwide Mut. Ins. Co. v Darden, 503 US 318, 327 [1992]). This approach has given us a hash of factors that may be held more or less probative to the "control" determination depending on who is performing the analysis; indeed, in the briefing on this case, the parties point to more than twenty factors, each supported by one or more of our cases, none of which overruled the other, all of which are claimed to bear on the control analysis. Yet without providing guidance as to which factors ought to be weighed, and how weighed, and when weighed — without providing some precision and clarity as to what the factors are meant to support— the existing paradigm ends up giving the agency free rein to make whatever legal determinations it pleases until, usually with little explanation, we reverse the Board's findings for want of "substantial evidence" in an opinion that usually wants for substantial explanation (cf. Matter of Charles A. Field Delivery Services [Roberts], 55 NY2d 516, 517 [1985]).^[19]

B.

Our inconsistent, summary precedent makes it nearly impossible to arrive at a decision in this case that seems in perfect harmony with what has come before. Reading today's decision, one might think that, in the future, the Board's employment-status determinations will unfailingly be affirmed on the grounds of substantial evidence. Instead, today's decision is just one more bounce of the ball in the opposite direction whence it came.

In 12 Cornelia St. (56 NY2d 895 [1982]), a memorandum opinion, we held that the Unemployment Insurance Appeal Board's determination that real estate salespersons were employees was not supported by substantial evidence, because that "determination must rest on evidence that the company exercises control over the results produced by its salespersons or the means used to achieve the results" — and "such control [wa]s lacking" where the salespersons: (1) were paid commissions; (2) worked whatever hours they chose; (3) were free to engage in outside employment; (4) bore their own expenses; (5) were not required to attend meetings or trainings; (6) paid their own premiums for health insurance; and, (7) found their own leads. Those salespeople were independent contractors even though the real estate corporation supplied them with business cards, held regular sales meetings, and provided them with workers' compensation (12 Cornelia St., Inc. v Ross, 83 AD2d 681, 682 [3d Dept 1981]). Mr. Vega — like the salespeople — was paid by commission (he received a percentage of the fee charged to the customer by Postmates for each delivery, which fee varied by distance), worked whatever hours he chose, was free to engage in outside employment, bore his own expenses, was provided no health insurance, and was not required to attend meetings or trainings (other than one initial meeting on how to use the app). Yet, despite the "evidence in the record that would have supported a contrary conclusion" (majority op at 4, citing Concourse Ophthalmology, 60 NY2d 734 [1983]), in 12 Cornelia St., we reversed the Board's determination as not supported by substantial evidence.

In Concourse Ophthalmology (60 NY2d at 736 [1983]), another memorandum decision, we upheld the Board's determination that doctors were employees, laying stress on the alleged employer's control of: (1) employee schedules; (2) the place of employment; (3) the appointment-making process; (4) the fee schedule; (5) ownership of key equipment; (6) administration of bills; and, (7) allocation of record-keeping responsibilities, even though the purported employer did not control the "results or means" of the doctors' work. However, those doctors maintained their own malpractice insurance, operated substantial outside practices, and "functioned completely autonomously" (Matter of Concourse Ophthalmology Assocs., P.C., 89 AD2d 1047, 1048 [3d Dept 1982] [Levine, J., dissenting]). We did not dispute the putative employer's contention that "the record is devoid of evidence of control over results or means." Instead, we brushed the means-ends test aside because "professional services do not lend themselves to such control." Thus, we affirmed the Board's decision on the grounds that substantial evidence existed in the record. Following that precedent, then, where the factors cut in different directions, the Board may have had "substantial evidence" for its determination here, because Postmates controlled the delivery-assignment process, the fee schedule, administration of bills and allocation of record-keeping responsibilities, even though the first two factors we cited in Concourse Ophthalmology (regular work schedule and employer determination of place of work) cut against finding Mr. Vega an employee.

In yet another memorandum, we reversed the Board's decision that salespeople for an aluminum siding installation company were employees, holding that decision was not supported by substantial evidence (Ted Is Back, 64 NY2d at 726; see also Matter of Ted Is Back Corp., 103 AD2d 932, 932 [3d Dept 1984]). We held that "incidental control over the results produced without further indicia of control over the means employed to achieve the results will not constitute substantial evidence of an employer-employee relationship" (Ted Is Back, 64 NY2d at 726). That "incidental control" included the corporation's supply of form contracts and its retained right to approve contracts (id.). That the salespeople were "agents" of the company was not "decisive, for this is equally true where salespeople are determined to be independent contractors" (id.). The Court cited factors including that the salespeople: (1) worked at their own convenience; (2) were free to hold outside employment; (3) were not limited to a particular territory; and, (4) were paid on commissions. Applying that analysis to this case, we would affirm the order of the Appellate Division that the Board lacked substantial evidence for its determination because each of the above factors from Ted Is Back applies to Mr. Vega; that "incidental control" is insufficient, as a matter of law, to deem Mr. Vega an employee.
We yet again reversed the decision of the Board that a salesperson was an employee, on the ground that substantial evidence did not support its determination in Hertz (2 NY3d at 733). There, the claimant was a salesperson for Hertz: she visited travel agencies to promote Hertz's products by making presentations and distributing sales materials (id. at 734-735). She was compensated per-visit, was not required to attend meetings, and had minimal supervision (id. at 735). On the other hand, Hertz told her what to wear during her visits, instructed her on what products to promote, barred her from promoting competing products, and instructed her on how to present Hertz products (id.). Despite those factors, which could have supported the Board's determination under the substantial evidence standard, we said that she was not an employee as a matter of law, because Hertz exerted merely "incidental control" (id.). Mr. Vega was compensated per delivery; the Hertz salesperson per visit. Mr. Vega was told where and what to pick up and deliver, just as the Hertz representative was told what products to promote. Neither was supervised; neither had to attend meetings. But Mr. Vega was not prohibited from delivering for competing services even while engaged in a Postmates delivery and could wear whatever he liked. Hertz would lead one to conclude that the order of the Appellate Division must be affirmed here because the factors determining Mr. Vega's status point more strongly to independent contractor than the factors on which we reversed the Board's determination in Hertz.

Our most recent decision in this area, also a memorandum, is Yoga Vida. There, as here, the Department of Labor initially determined Yoga Vida was liable for unemployment insurance for its non-staff yoga instructors; as here, an ALJ overruled that determination; next, the Unemployment Insurance Appeal Board reversed the ALJ's decision, sustaining the Department's initial determination (Yoga Vida, 28 NY3d at 1015). The Appellate Division sustained the Board's determination that the non-staff yoga instructors were employees. We reversed, holding the Board's determination was not supported by substantial evidence because the non-staff instructors made their own schedules, were paid only if students attended their classes, could work for competitors, and were not required to attend meetings or trainings (id.)^[20]. Each of those factors — which are the factors that justified our reversal as a matter of law — is true as to Mr. Vega: Mr. Vega made his own schedule, was paid only for deliveries he made, could work and perform deliveries for other companies even while making deliveries for Postmates, and was not required to attend meetings or trainings. Moreover, Mr. Vega did not have to work on a prearranged schedule; the non-staff instructors did. Mr. Vega did not have to work at his employer's place of business; the non-staff instructors did. In Yoga Vida, we emphasized that Yoga Vida's determination and collection of the fees that the non-staff instructors received did "not supply sufficient indicia of control" (id. at 1016). So too, Postmates' control over the fees Mr. Vega received should be insufficient to support an employer-employee relationship^[21]. That the outcome today is the diametric opposite of the outcome arrived at just three years ago in Yoga Vida, an outcome that the majority contends was reached by the same standard as applied here (majority op at 8), shows just how inconstant our "test" has become.

The facts surrounding Mr. Vega's six-day adventure as a courier neatly fit into the exertion of mere "incidental control," which does not provide substantial evidence for a Board's determination of employee status (see e.g. Ted Is Back, 64 NY2d at 726). Mr. Vega retained more than just "some independence to choose [his] work schedule and delivery route" (majority op at 6) — he had complete control over his schedule, the hours he logged on, the jobs he accepted or rejected (or rejected even after accepting them), and the routes he took when, having accepted a job, he actually made the promised delivery. Nor does Mr. Vega need to have actual control over "the service Postmates provides its customers" (majority op at 6 [emphasis added]), for that is not the test. The control test applies only to the company's control over the worker's labor; the worker need not have control over, or anything to do with, the service provided by the company. Our precedent may not be consistent, and it certainly makes it difficult for litigants and lower courts to apply the control test, but at least it provides this answer: Mr. Vega is far more an "independent contractor" than the real estate salespeople given business cards by their company in 12 Cornelia St., the product promoter who was told what to wear and how to present in Hertz, or the yoga instructors told when, where and what to teach in Yoga Vida.

III.

The overarching problem in our control test jurisprudence is exemplified by Mr. Vega's case. The parties here attempt to rely on two dozen different factors, all of which derive from previous Court of Appeals' decisions and all of which we have suggested may have some bearing on the control analysis. However, those factors fail to aid us in this decision. Factors that seem in some of our cases to provide support for the determination of an employer-employee relationship do not always do so (compare Concourse Ophthalmology, 60 NY2d at 736 [where the company collected and determined fees, supporting the decision that ophthalmologists were employees] with Yoga Vida, 28 NY3d at 1017 [Fahey, J., dissenting] [where Yoga Vida collected and fixed the fees of the non-staff instructors, but substantial evidence did not support the Board's decision that the non-staff instructors were employees]),^[22] whereas factors that we previously identified as showing merely incidental control insufficient to establish an employment relationship, such as setting one's own schedule, the ability to work for other companies, and being paid on commission (see e.g. Ted Is Back, 64 NY2d at 726) are today eschewed by the majority.

It is no surprise, then, that reading the procedural history of this case or our prior decisions feels like watching a ping-pong match: no one, not even this Court, knows which combinations of factors are determinative and which are not. We have held that no one factor is determinative (Concourse Ophthalmology, 60 NY2d at 736), and the majority emphasizes the formlessness of our jurisprudence by observing that "no enumerated list of factors can apply to every situation faced by a worker" (majority op at 5). At the same time, the majority adds a new factor to the list: "the nature of the work" (id. ). It wields that new factor based on unsupported judgments, contrasting couriers as "low-paid workers performing unskilled labor who possess limited discretion over how to do their jobs" (id. at 6), with yoga instructors who perform work that "is not comparable to that of a delivery person," because a yoga instructor provides a "service ... unique to that of the instructor and his or her personal characteristics" (id. at 7). We have never before mentioned the "nature of the work" in the employment context, let alone held that factor determinative in such cases.^[23]

In any event, adding that factor to our litany makes no sense. The purpose of the control test, and even the variegated factor analysis, is to prevent the over-reliance on the kind of work in the determination of employment status. Forcing that factor into the control test provides the opportunity for every courier — regardless of any analysis of the factors or the degree of control exerted over them — to be an employee because of the nature of delivery work: if Mr. Vega is an employee, so is everyone who delivers a tangible good. Had the "nature of work" been determinative in Yoga Vida, non-staff yoga instructors and staff instructors alike would need to have been deemed employees. Instead, however, this Court held that substantial evidence did not support the Board's determination that non-staff yoga instructors were employees by relying on dissimilarities — having nothing to do with the nature of the work — between the non-staff instructors and the staff instructors (Yoga Vida, 28 NY3d at 1015). If the nature of the work holds as much importance as the majority declares it does today, the dissimilarities between the non-staff and staff instructors would have been irrelevant.
For that reason, the majority's reliance on Matter of Rivera (69 NY2d 679 [1986]), is misplaced. Simply because the employees at issue in Rivera were delivery persons does not make that case "indistinguishable" (majority op at 7). Although similar in many respects, there are several notable differences: Mr. Vega could accept a delivery assignment and then later change his mind at any time, even after a customer had been told that Mr. Vega would deliver it, whereas there is no indication that the Rivera couriers could turn back a delivery assignment after accepting it; Mr. Vega was not given a time limit for completing deliveries, whereas the Rivera couriers were; Mr. Vega was free to choose any means of transportation he wished (which he could vary at will with no need to inform Postmates), whereas Rivera required its couriers to use motor vehicles and to purchase both ordinary insurance as well as special cargo insurance for their vehicles (Claim of Rivera, 120 AD2d 852, 854 [3d Dept 1986]).

The addition of a new factor to the control test illustrates the most concerning aspect of our ever-changing employment determination decisions. Because the test depends on innumerable factors, which vary from case to case and opinion to opinion, and we review the Board's determinations for substantial evidence, the Board is given unbounded discretion. We will never be able adequately to review their determinations because they will always rely on factors that we — at one point or another — have sanctified. "Substantial evidence" review of a smorgasbord of flavorless factors authorizes unrestrained agency decision-making.

Without regular examination, modification and explication of our common law, common-law tests risk falling into incoherence or vagueness. In the broader sense, this case implicates many factors upon which our Court and the Appellate Division have relied to reverse an agency determination as "unsupported by substantial evidence" because the courts have concluded that the facts relied on by the agency do not really speak to the ultimate issue of control (see e.g. Empire State Towing, 15 NY3d at 437; 12 Cornelia St., 56 NY2d at 897-898; Ted Is Back, 64 NY2d at 726; Hertz, 2 NY3d at 733; Yoga Vida, 28 NY3d at 1015-1016). Our several prior reversals of the Board's decisions, though offering that "the determination of the appeal board, if supported by substantial evidence on the record as a whole, is beyond further judicial review even though there is evidence in the record that would have supported a contrary decision'" (Empire State Towing, 15 NY3d at 437 [2010]), are best understood as reflecting our recognition that we cannot allow the deferential standard of review to shield the Board from all review, even if the Board has relied exclusively on factors we have deemed relevant in determining who is an employee. Of course, the most important concern for us is to ensure that the law is clear and consistent as to what employers, workers and the lower courts must consider in forming and evaluating work relationships. Our terse memorandum decisions do not serve that purpose.

We need a clear understanding, comprehending the modern realities of our rapidly evolving economy, of who should be an employee and who an independent contractor, including whether work relationships should continue to be measured on that dichotomy. It is past time for the law to reexamine the definition of work and its application to different forms of entitlement and obligation. Our current framework, as inconsistently applied, fails to provide clarity to anyone involved. The agencies tasked with applying our cases and the courts that attempt to review those decisions for substantial evidence oscillate in result and rationale. The common-law test for status as an employee developed in a vastly different time, when employment was monotonic. Now, it is cacophonic. The number of workers performing multiple or alternative jobs has grown dramatically. New technology and the rise of the sharing economy have driven further changes, including the crowdsourcing of flexible and low-barrier-to-entry jobs upon which many workers are less reliant than our traditional notion of career employees. The challenge is how to apply our inconsistent common-law test in a world where work looks much different than it did when that test was developed and where we cannot, if we hold true to the deferential standard of review, reverse an agency determination so long as it has relied on some of the factors we have identified as relevant.

The stakes are high. As the California Supreme Court, confronting a similar question, phrased the matter (Dynamex Operations W. v Superior Ct., 4 Cal 5th 903, 912-13 [2018] [Cantil-Sakaute, C.J.]):
"On the one hand, if a worker should properly be classified as an employee, the hiring business bears the responsibility of paying federal Social Security and payroll taxes, unemployment insurance taxes and state employment taxes, providing worker's compensation insurance, and, most relevant for the present case, complying with numerous state and federal statutes and regulations governing the wages, hours, and working conditions of employees. The worker then obtains the protection of the applicable labor laws and regulations. On the other hand, if a worker should properly be classified as an independent contractor, the business does not bear any of those costs or responsibilities, the worker obtains none of the numerous labor law benefits, and the public may be required under applicable laws to assume additional financial burdens with respect to such workers and their families."
"Although in some circumstances classification as an independent contractor may be advantageous to workers as well as to businesses, the risk that workers who should be treated as employees may be improperly misclassified as independent contractors is significant in light of the potentially substantial economic incentives that a business may have in mischaracterizing some workers as independent contractors. Such incentives include the unfair competitive advantage the business may obtain over competitors that properly classify similar workers as employees and that thereby assume the fiscal and other responsibilities and burdens that an employer owes to its employees. In recent years, the relevant regulatory agencies of both the federal and state governments have declared that the misclassification of workers as independent contractors rather than employees is a very serious problem, depriving federal and state governments of billions of dollars in tax revenue and millions of workers of the labor law protections to which they are entitled."^[24]

Although it is well within the purview of the courts to alter a common-law test, that is best done incrementally; the complete overhaul of our common-law employment test to adapt it to the present and future economy is not a task to which courts are well suited^[25]. Whether, to what degree, and on what basis we wish to provide unemployment benefits to Postmates couriers generally, or to other workers in the gig economy, is a policy question best left to the legislature. Whether the test for that entitlement should be the same as the test for Postmates' liability if a courier, speeding on an electrified bicycle to make a timely delivery of a hot dinner, injures a pedestrian, is also a question best suited to legislative determination. The role of the courts is to interpret the law and to clarify it when we can. The accumulation of indecisive, unweighted factors articulated in our past cases, scrutinized under our highly deferential standard of review, typically produces either a de facto lack of review or an uninformative summary reversal. The resulting body of law is difficult to reconcile and does little to advise agencies and lower courts (to say nothing of business enterprises or workers) how any particular work relationship will or should be adjudicated. Today's decision places further stress on that test through its contradiction of our recent decision in Yoga Vida and its incompatibility with several others in which we found the Board lacked substantial evidence in circumstances less compelling than this. Nevertheless, until the legislature steps in, we have an obligation to do our best to reach consistent results so that both businesses and workers can structure their affairs with a sound understanding of when the benefits and obligations of "employment" are imposed or conferred upon them. Whether other Postmates couriers are employees is not before us. Mr. Vega's case is, and he is not.

[12] Not only was Postmates given no opportunity to present evidence as to other couriers, but the parties were expressly informed by the ALJ, at the start of the hearing, that the ALJ's determination would be limited to just Mr. Vega. That neither party addressed the ALJ's limited determination does not change that determination's scope (majority op at 3 n 1). Thus, the Board's subsequent decision, if read to impose a contribution requirement on Postmates for Mr. Vega and all other similarly situated couriers, would implicate due process concerns because Postmates was expressly instructed that the proceeding would relate solely to Mr. Vega's claim and not others similarly situated (see e.g. Martin v Ronan, 47 NY2d 488, 490 [1979]["a requisite of due process (is) the opportunity to be heard before one's rights or interests are adversely affected"]).
[13] Postmates' witness before the ALJ, the regional manager, believed Mr. Vega had been a bicycle courier, which just goes to show how much "supervision" of Mr. Vega actually took place.
[14] Incidentally, Mr. Vega did not appear and presented no evidence at the hearing; the sole evidence presented came from Postmates.
[15] At oral argument, counsel for the Board had difficulty pinpointing when Mr. Vega became an employee of Postmates, eventually settling for the moment he made his first Postmates delivery.
[16] I note that "employment" under Article 18 expressly includes "any service by a person for an employer as an agent-driver or commission-driver engaged in distributing meat, vegetable, fruit, or bakery products; beverages other than milk; or laundry or dry-cleaning services" (Labor Law § 511 [1] [b] [1]), a definition that might arguably sweep in Mr. Vega if the record showed what he had delivered (or promised to deliver), but the parties do not direct any arguments to the possible application of this definition and accordingly I do not consider it.
[17] Of course, the Legislature may, for unemployment compensation purposes, adopt a different definition of "employee." It has done so in other parts of the Labor Law, providing at Labor Law § 2(7) that one is employed simply when one is "permitted or suffered to work" for an employer (cf. Nationwide Mut. Ins. Co. v Darden, 503 US 318, 326 [1992] [discussing similar language in the federal Fair Labor Standards Act]). However, that broader definition does not apply to the unemployment insurance article, which uses the definition discussed above (Labor Law § 510).
[18] We have never defined "incidental control." Doing so might better allow us, even if we cannot reach consensus on what meets the substantial evidence standard, to determine when the facts are not supported by substantial evidence.
[19] Of our opinions in the unemployment insurance cases since 1981, Villa Maria Inst. of Music v Ross (54 NY2d 691 [1981]), 12 Cornelia St., Inc. v Ross (56 NY2d 895 [1982]), Matter of Concourse Ophthalmology Assoc., P.C. (60 NY2d 734 [1983]), Matter of Ted Is Back Corp. (64 NY2d 725 [1984]), Matter of Rivera (69 NY2d 679 [1986]), Matter of Salamanca Nursing Home, Inc. (68 NY2d 901, 903 [1986]), Matter of Hertz Corp. (Commissioner of Labor) (2 NY3d 733 [2004]), and Matter of Yoga Vida NYC, Inc. (28 NY3d 1013 [2016]) were all summary, memoranda judgments. We have treated this question in a full opinion only in Matter of Charles A. Field Delivery Serv., Inc. (66 NY2d 516 [1985]) and Matter of Empire State Towing and Recovery Assn., Inc. (15 NY3d 433 [2010]). The profusion of summary memoranda in this area underscores the incoherence of the conventional test for independent contractor/employee status when applied to an evolving economy that encompasses new styles of working.
[20] The majority's claim that our decision in Yoga Vida "reaffirmed" the substantial evidence standard (majority op at 8) illustrates the amorphousness of that standard when applied in the unemployment insurance context. For example, even though "Yoga Vida generally determines what fee is charged and collects the fee directly from the students [that] does not supply sufficient indicia of control over the
instructors" to constitute substantial evidence of employment (Yoga Vida, 28 NY3d at 1016). If the majority is reaffirming the standard we applied in Yoga Vida, that standard has become so elastic as to be whimsical.

[21] The majority cites a few other distinctions between the non-staff yoga instructors and Mr. Vega, but those sound more different than they really are. The yoga instructors' interest in maintaining a "customer following to be successful" is akin to the customer ratings Mr. Vega needed to remain successful on the Postmates app (majority op at 7). Likewise, the ability of the yoga instructors to affect their compensation to some degree by choosing hourly compensation or a percentage of class fees corresponds to Mr. Vega's ability to affect his compensation by taking, e.g., high dollar volume jobs or short-haul jobs that he could consolidate and deliver simultaneously. Indeed, Mr. Vega had more control of his compensation than did the yoga instructors, inasmuch as he could accept a job, immediately see the compensation he would receive from it, and reject it if it was not to his liking. There is nothing to suggest the yoga instructors could decline to teach a class or switch from percentage to hourly compensation if only two customers appeared for class.
[22] Judge Rivera's concurring observation that Yoga Vida was wrongly decided further emphasizes the incoherence of our decisional law in this area.
[23] Although the majority points to no prior precedent of ours listing "nature of the work" among the factors to be considered, we are advised that because "indicia of control will vary depending on the nature of the work," it is "obvious" that nature of the work should be included in the list of factors (majority op at 5 n 3, 8). I do not think it "obvious" — one way or the other — that differences in the work of ophthalmologists and aluminum siding salespeople tend to make the former more controllable than the latter, or that our determination that the former were employees and the latter independent contractors had anything to do with the intrinsic nature of their occupations.
[24] See also Joint Task Force on Employee Misclassification, Annual Report 2015, https://www.labor.ny.gov/agencyinfo/PDFs/Misclassification-Task-Force-Report-2-1-2015.pdf (discussing at length the effect of employee misclassification on New York state in particular); Sarah Jeong, Strike All You Want. Uber Won't Pay a Living Wage, NY Times (May 10, 2019), https://www.nytimes.com/2019/05/10/opinion/uber-ipo.html (discussing studies showing that in New York City, about half of ride-hailing drivers are supporting families with children, but earn so little that 40 percent of those drivers qualify for Medicaid and another 18 percent qualify for food stamps); Noam Scheiber, Uber and Other Gig Companies Maneuver to Shape Labor Rules, NY Times (Mar. 26, 2019), https://www.nytimes.com/2019/03/26/business/economy/gig-economy-lobbying.html (discussing the "highly disruptive" effect of proposals to classify large numbers of gig workers as employees).
[25] Our call for a new test need not be preserved (majority op at 5 n 3). In fact, because the employment test is one created by the courts, it is our job to change it if necessary — regardless of if the parties call for it. However, it is the role of the legislature to make policy; here, where the majority's holding overturns a decision of this Court from just three years prior, it is clear we are in sore need of legislative decision-making, not solely judicial re-interpretation.