H.S. Care L.L.C., d/b/a Oakwood Care Center and N&W Agency, Inc.
overrules M.B. Sturgis;
No. 76 (Nov. 19, 2004)
REVIEW AND ORDER
On October 20, 2003, the Regional Director for Region 29 issued a Decision and Direction of Election, in which he found appropriate a petitioned-for unit of nonprofessional employees at Oakwood’s facility in Oakdale, New York. The petitioned-for unit includes both employees who are solely employed by Oakwood and employees who are jointly employed by Oakwood and a personnel staffing agency, N&W.
On February 11, 2004, we granted Oakwood’s request for review of the Regional Director’s decision. Oakwood argued, among other things, that the unit combining the two groups of employees is inappropriate under the Act. Oakwood urged the Board to overrule its decision in M.B. Sturgis, on which the Regional Director relied. In Sturgis, the Board found that bargaining units that combine employees who are solely employed by a user employer and employees who are jointly employed by the user employer and a supplier employer are permissible under the Act.
For the reasons discussed below, we hold, contrary to the Board’s decision in Sturgis, that such units constitute multiemployer units, which, in accordance with the statute, may be appropriate only with the consent of the parties. Therefore, we overrule the Board’s decision in Sturgis and return to the Board’s longstanding prior precedent.
Oakwood has operated the Oakdale long-term residential care facility (the Home) since August 2002. At the time of the hearing in this proceeding, Oakwood had filled 152 beds of the Home’s 280-bed capacity and had hired approximately 55 percent of its expected number of employees, including at least some employees in each relevant classification. There is no dispute that some of the employees are solely employed by Oakwood, and other employees, i.e., those supplied by N&W, are jointly employed by Oakwood and N&W.
The jointly employed employees as well as those solely employed by Oakwood perform duties that are part of the normal functioning of the Home. Oakwood and N&W together determine the pay and benefits of the jointly employed employees. Oakwood supervisors supervise and direct these employees, determine their work schedules, approve schedule changes, assign them overtime work, and approve their requests for time off, both paid and unpaid. In addition, Oakwood supervisors discipline these employees and evaluate their work performance. The jointly employed employees wear identification tags issued by Oakwood and identifying them as employees of the Home.
The parties stipulated that, if the
Board applies its holding in Sturgis
to the facts of this case, the petitioned-for unit would be an appropriate
unit. However, Oakwood urges the
Board to reverse Sturgis,
contending that it was wrongly decided.
Board Precedent: Greenhoot and
For decades prior to Sturgis, the Board’s decisions in Greenhoot, Inc., and Lee Hospital provided the controlling precedent regarding a unit that would combine groups of employees who had an employer in common but did not have identical employers because one or more groups had an additional, joint employer not shared by the others. In such cases, the Board relied on the principle that involuntary combinations of those employees (and their respective employers) constituted inappropriate bargaining units.
In Greenhoot, the Regional Director directed an election in a petitioned-for unit of engineers and maintenance employees working at 14 office buildings. Each building was owned by a separate employer, and all of the buildings were managed by Greenhoot. The employees in each building were jointly employed by their building owner and by Greenhoot. Thus, in buildings owned respectively by A, B, and C, the employees were employed by Greenhoot and A, Greenhoot and B, and Greenhoot and C. The unit sought combined all 14 groups of these employees into a single unit. Greenhoot argued, inter alia, that such a unit was inappropriate.
Agreeing with Greenhoot’s argument, the Board found 14 separate joint employer relationships consisting of Greenhoot and the building owner and treated the petitioned-for unit as a multiemployer unit requiring consent. The Board held that:
. . . there is no legal basis for establishing a multiemployer unit absent a showing that the several employers have expressly conferred on a joint bargaining agent the power to bind them in negotiations or that they have by an established course of conduct unequivocally manifested a desire to be bound in future collective bargaining by group rather than individual action.
Finding no evidence of such consent by the various employers, the Board determined that 14 separate units were appropriate. The Board remanded the proceeding to the Regional Director for the purpose of conducting elections in each of the 14 units.
In Lee Hospital, the
Petitioner sought a unit of all certified registered nurse anesthetists (CRNAs).
The CRNAs were employed in the anesthesia department, which was
operated by Anesthesiology Associates, Inc. (AAI), under a contract with the
hospital. The Regional Director
found that the CRNAs did not have the disparity of interests required to
justify a unit separate from the professional employees employed solely by the
hospital. The Regional Director
found it unnecessary to determine whether the hospital and AAI were joint
employers of the CRNAs. The
Board found that the initial inquiry was not whether there was a disparity of
interests between the CRNAs but rather whether the CRNAs could legally be in
the same unit as the other professionals.
This issue, in turn, depended on whether Lee and AAI were joint
employers of the CRNAs. The Board
As a general rule, the Board does not include employees in the same unit if they do not have the same employer, absent employer consent. Thus, if AAI is a joint employer, the CRNAs could be included in the unit with other professionals employed by Lee Hospital only with the hospital’s consent.
Thus, consistent with Greenhoot, the Board in Lee Hospital found that sharing one employer was insufficient to permit including jointly employed CRNAs in the existing unit of solely employed professional employees without employer consent. Ultimately, however, the Board found that AAI was not a joint employer of the CRNAs and affirmed the Regional Director’s dismissal of the petition on the basis of insufficient disparity of interests.
In Sturgis, the Board overturned these settled principles. Applying a novel definition of “employer” fashioned for the purpose of deciding the case, the Sturgis majority, for the first time in the history of the Act, stated that some units combining jointly employed and solely employed employees were nevertheless single employer units. We have carefully considered the arguments advanced by the Sturgis Board and respectfully disagree. We have concluded that Sturgis was wrongly decided and overrule it today.
In Sturgis, the Board reviewed two decisions by Regional Directors. One decision found appropriate a petitioned-for unit that included Sturgis’ approximately 35 solely employed employees but excluded 10-15 employees supplied to Sturgis by Interim, a temporary help agency. The employer requested review of the Regional Director’s decision, arguing that the unit was not appropriate unless the supplied employees were included. In the other case, the petitioner sought to clarify an existing unit of approximately 600 production and maintenance employees solely employed by Jeffboat Division, American Commercial Marine Service Company (Jeffboat), to accrete 30 steamfitters and welders supplied to Jeffboat by supplier TT&O. The Regional Director dismissed the petition in Jeffboat, finding that the combined unit would be inappropriate without the employers’ consent.
In each case, the Board found that the supplied employees were jointly employed by the user (i.e., Sturgis or Jeffboat) and its supplier. Moreover, the Board held that units including these jointly employed employees and the solely employed employees of the user were permissible under Section 9(b) without the consent of the employers. The Board reasoned:
That a unit of all of the user’s employees, both those solely employed by the user and those jointly employed by the user and the supplier, is an “employer unit” within the meaning of Section 9(b) is logical and consistent with precedent. The scope of a bargaining unit is delineated by the work being performed for a particular employer. In a unit combining the user employer’s solely employed employees with those jointly employed by it and a supplier employer, all of the work is being performed for the user employer. Further, all of the employees in the unit are employed, either solely or jointly, by the user employer. Thus, it follows that a unit of employees performing work for one user employer is an “employer unit” for purposes of Section 9(b).
331 NLRB at 1304–1305. The Board therefore overruled Lee Hospital and held that employees jointly employed by employers A and B could be in a unit with employees solely employed by one of them. The issue of whether they are in fact to be in one unit would turn on a “community of interest” analysis.
The Sturgis Board also reaffirmed, however, the continuing validity of the Board’s holding in Greenhoot that a unit comprised of employees jointly employed by a single supplier employer and multiple user employers was a multiemployer unit impermissible absent consent of all parties. The Board stated:
[C]ases like Greenhoot involve multiple user employers whose only relationship to each other is that they obtain employees from a common supplier employer. In such cases, the union seeks to represent a unit that includes employees of all of the users. Thus, it is clear that the unit is a multiemployer unit and therefore consent of the separate user employers would be required before the Board could direct an election.
In dicta, the Sturgis
Board “clarified” Greenhoot
by stating that if the union in Greenhoot
had sought to bargain with and named only Greenhoot (the supplier employer) in
its election petition, the unit, if certified, would not be a multiemployer
Thus, the Sturgis Board found that the proposed combined units were permissible under Section 9(b). It remanded both proceedings to the respective Regional Directors for a determination whether the petitioned-for unit at Sturgis and the accretion at Jeffboat were appropriate under applicable community of interest standards.
The Sturgis Board’s
reinterpretation of the concept of an “employer unit” severed that term
from its statutory moorings. This
loss of direction gave rise to anomalous decisions such as Gourmet
Award Foods, 336
NLRB 872 (2001). There, the
employer expanded its use of temporary employees in response to increased
business, and notified the union representing its drivers and warehousemen
that some temporary employees would work at the facility for 4 to 5 months.
The union demanded that the employer apply the collective-bargaining agreement
to these temporary employees, and the Board agreed.
Finding them to be jointly employed by the employer and the temporary
agencies that supplied them, the Gourmet Award Foods Board concluded
that the temporary employees were automatically added to a unit of solely
employed employees, without any vote by either group, because, under Sturgis,
they were “new employees” “hired” into a unit classification.
Thus, Sturgis gave rise not only to a new definition of
“employer,” but, as extended in Gourmet Award Foods, also a new
conception of “new employees” and “hire” as well.
The Gourmet Award Foods Board acknowledged that the employer had no reason to expect that the temporary employees supplied by other employers would be included in the existing bargaining unit. Faulting the employer for failing to foresee the unforeseeable, however, the Board proceeded to find that those temporary employees were included in the unit because the unit definition did not expressly exclude them.
Further extending the strained logic of Sturgis, the Gourmet Award Foods Board held that the employer was required to apply its collective-bargaining agreement to the temporary employees only with respect to those terms that it controlled. By requiring the employer to apply some but not all of the terms of its collective-bargaining agreement to these “employees,” the Board effectively modified the parties’ agreement as well.
Reconsideration of Sturgis
Section 9(b) of the Act establishes the Board’s authority to determine appropriate units, providing in relevant part:
The Board shall decide in each case whether, in order to assure to employees the fullest freedom in exercising the rights guaranteed by this Act, the unit appropriate for the purposes of collective bargaining shall be the employer unit, craft unit, plant unit, or subdivision thereof . . .
Of these permissible categories of units, the broadest is the “employer unit,” with each of the other delineated types of appropriate units representing subgroups of the work force of an employer. Thus, the text of the Act reflects that Congress has not authorized the Board to direct elections in units encompassing the employees of more than one employer. The legislative history supports this interpretation of the plain language of the Act. Specifically, Congress included the phrase “or subdivision thereof” to authorize other units “not as broad as ‘employer unit,’ yet not necessarily coincident with the phrases ‘craft unit’ or ‘plant unit.’”
Where the parties voluntarily agree
to multiemployer bargaining, however, the Board has long recognized the
legitimacy of such units.
In cases where all parties voluntarily agree to bargain on a
multiemployer basis, the Supreme Court has recognized that “Congress
intended that the Board should continue its established administrative
practice of certifying multiemployer units.”
By ignoring the bright line between employer and multiemployer units, Sturgis departed from the statutory directive of Section 9(b) as well as decades of Board precedent. We find that the new approach adopted in Sturgis, however well intentioned, was misguided both as a matter of statutory interpretation and sound national labor policy.
Sturgis left in place the fundamental principle that Section 9(b) permits the Board to find multiemployer units appropriate only with the consent of the parties but restricted the situations in which that principle would be applied through a strained interpretation of the phrase “employer unit.” The Sturgis majority held that a unit composed of the employees of a user employer and employees jointly employed by the user and a supplier employer consists of employees of one employer. We disagree with this redefinition of terms, as it is inconsistent with the plain meaning of “employer unit” in the Act. Returning to prior precedent under both Greenhoot and Lee Hospital, we conclude that solely employed employees and jointly employed employees are employees of different employers and that their inclusion in the same bargaining unit creates a multiemployer unit.
A joint employer, under the Board’s traditional definition, is comprised of two or more employers (e.g., A and B) that “share or codetermine those matters governing essential terms and conditions of employment” for bargaining unit employees. All of the unit employees work for a single employer, i.e., the joint employer entity A/B. Therefore, a joint employer unit of A/B is not a multiemployer unit. In a Sturgis unit, in contrast, some of the employees are employed by A, and others are employed by A/B. It may be that, as to the latter group, A and B jointly set all terms and conditions of employment. Or, it may be that, as to that group, A sets some terms and B sets others. The critical point is that the one group has its terms set by A/B. The other group has its terms set only by A. Thus, the entity that the two groups of employees look to as their employer is not the same. No amount of legal legerdemain can alter this fact.
In finding otherwise, the Sturgis Board relied on earlier cases involving department stores, where the Board found units combining employees solely employed by the store with employees jointly employed by the store and its licensees to be appropriate. However, these cases are not precedent for the holding in Sturgis, because no party raised, and the Board and the reviewing courts did not consider, the statutory restrictions imposed by Section 9(b) on nonconsensual units that are multiemployer in scope. Therefore, these cases do not answer the statutory question presented in Sturgis and in this proceeding and provide no precedential support for the sea-change in Board law wrought by Sturgis.
In addition, the department store
cases are distinguishable on their facts.
A series of cases involving K-Mart illustrates the typical pattern.
In each, licensees operated sales departments within K-Mart stores
under license agreements that vested K-Mart with wide-ranging control over the
licensees’ business operations and labor relations.
K-Mart controlled advertising, merchandising, credit decisions, and the
physical arrangement of the department.
Under the licenses, “the public is given the impression of a single,
integrated enterprise since . . . each licensee must ‘conduct sales on the
premises solely under the name of K-Mart.’”
K-Mart also controlled the licensees’ labor relations to such an extent that
“it would be impossible for a licensee to enter into a binding collective
bargaining agreement without [K-Mart’s] consent.”
In short, K-Mart and its licensees were not only joint employers but a
unique relationship not present in Sturgis or this case.
The policy implications of Sturgis are as problematic as its interpretation of Section 9(b). As Member Brame pointed out in his Sturgis dissent, the bargaining structure contemplated in that decision gives rise to significant conflicts among the various employers and groups of employees participating in the process. These are precisely the types of conflicts that Section 9(b) and the Board’s community-of-interest test are designed to avoid.
Although Sturgis anticipates
that each employer will bargain with respect to employees whom it employs and
as to the terms and conditions of employment that it controls, the reality of
collective bargaining defies such neat classifications.
Two examples illustrate this point.
First, the wages paid to the jointly employed employees, which are
frequently controlled by the supplier, could certainly have an effect on the
negotiation of the wages of the solely employed employees, a matter controlled
by the user. Second, the user
employer would likely determine the holiday schedule for its facility, but the
supplier might control whether the jointly employed employees are paid for
those holidays. These are merely
examples of how the bifurcation of bargaining regarding employees in the same
unit hampers the give-and-take process of negotiation between a union and an
employer, and places the employers in the position of negotiating with one
another as well as with the union. The
user employer’s status as the customer of the supplier may effectively
restrict the supplier’s options
in bargaining as to those subjects that appear to be within its purview.
This fragmentation thus undermines effective bargaining.
The bargaining regime posited in Sturgis also fails to adequately protect employee rights. It combines jointly employed and solely employed employees in a single unit, with a single union negotiating with two different employers, each of which controls only a portion of the terms and conditions of employment for the unit. Such a structure subjects employees to fragmented bargaining and inherently conflicting interests, a result that is inconsistent with the Act’s animating principles.
For example, in many situations, the wages of the supplied employees are set by the supplier (A), and the wages of the solely employed employees are set by the user (B). The result is that the wages of the employees of A/B may be traded away, in bargaining, for the sake of employees of B, or vice-versa. In order for employees to enjoy the full prospect of effective representation, the Act contemplates that employees be grouped together by common interests and by a common employer. The nonconsensual mixing of employees of different employers vitiates that basic principle.
The Sturgis dicta
authorizing, for the first time, petitions in which the union names only one
of the joint employers creates additional bargaining difficulties for
employees. This concept
profoundly diminishes employee Section 7 rights.
Because the supplier employer generally controls the economic terms of
its employees, a petition that names only the user employer potentially
saddles the jointly employed employees with a representative that will be
unable to bargain with the employer that controls their wages. Presumably,
such a petition also potentially precludes the employees from ever obtaining
union representation for that purpose (unless the Sturgis Board
intended also to sanction multiple “exclusive” representatives for the
For these reasons, we conclude that permitting a combined unit of solely and jointly employed employees, as the Board did in Sturgis, contravenes Section 9(b) by requiring different employers to bargain together regarding employees in the same unit. We hold that combined units of solely and jointly employed employees are multiemployer units and are statutorily permissible only with the parties’ consent.
In the present proceeding, the Petitioner seeks to represent a unit encompassing both the solely employed employees of Oakwood and the jointly employed employees supplied to Oakwood by N&W. We find that the petitioned-for unit is a multiemployer unit. Because neither Oakwood nor N&W has consented to bargaining with the other in a multiemployer unit, the petition must be dismissed.
The Regional Director’s Decision and Direction of Election is reversed, and the petition is dismissed.
Dated, Washington, D.C. November 19, 2004
National Labor Relations Board
Members Liebman and Walsh, dissenting.
The Board now effectively bars yet another group of employees—the sizeable number of workers in alternative work arrangements—from organizing labor unions, by making them get their employers’ permission first.1 That result is surely not what Congress envisioned when it instructed the Board, in deciding whether a particular bargaining unit is appropriate, “to assure to employees the fullest freedom in exercising the rights guaranteed by th[e] Act.” 29 U.S.C. §159(b).
Section 9(b) of the Act authorizes the Board to choose among “the employer unit, craft unit, plant unit, or subdivision thereof.” Id. Overruling M.B. Sturgis, Inc., 331 NLRB 1298 (2000), the majority insists that the statutory phrase “employer unit” plainly prohibits a bargaining unit that covers, in the same workplace, both a supplier employer’s employees and the user employer’s employees—despite the fact that, under the Board’s law, the two employers are joint-employers of the supplied workers, and the supplied workers and their coworkers share a community of interest.
The alternative work arrangements involved here depart from the so-called “traditional relationship” characterized by a social contract between employers and employees. These new arrangements have been caused “in large measure by pressures on employers to restructure themselves, pressures that are massive and appear unlikely to go away.” Peter Capelli, et al., Change at Work (1997).2 Capelli concludes that “[s]ome companies appear to be making increased use of contingent labor as part of their strategic decision to pursue a low-wage, low-skill, high-turnover path to profit making.” Id. at 77. As one recent examination of the working poor observed:
[G]lobalization has thrown the least-skilled into head-on competition with people willing to work for pennies on the dollar. And a torrent of immigration, mainly poor rural Mexicans, has further swelled the low-end labor pool. Together, these trends have shoved many hourly wage occupations into a world-wide, discount labor store stocked with cheap temps, hungry part-timers, and dollar-a-day labor in India, Mexico, and China, all willing to sell their services to the lowest bidder. Against such headwinds, full employment offers only partial protection.
Michelle Conlin & Aaron Bernstein, “Working ... and Poor,” Business Week, May 31, 2004 (Issue 3885), at 62.
It is a mistake to approach this case—involving the rights of temps, part-timers, and other contingent workers to improve their working conditions through union representation—without this frame of reference. And however real the competitive pressures on American firms, their need to respond to economic uncertainty should not be permitted to erode their employees’ rights to union representation. Today’s decision does just that, rather than ensure that the Board’s law keeps pace with changing conditions, as Congress envisioned. Insisting on a supposed strict construction, the majority wrongly concludes that the statute itself does not permit the approach taken by the Board in Sturgis. As for the policy reasons the majority offers, no empirical support is cited. Remarkably, the Board did not even solicit the public’s views on the practical effects of Sturgis. The majority never explains how granting employers veto power over union representation advances the “fullest freedom” of employees to pursue collective bargaining. Protecting employers who choose to create alternative work arrangements—and who now are essentially encouraged to do so, to frustrate union organizing—is not the Act’s overriding concern.
Below, we summarize the Sturgis decision, describe the majority’s decision here, and then explain why Sturgis was correctly decided and why today’s decision is misguided. As will become clear, the majority’s claim that its decision represents a “return to the Board’s long-standing prior precedent” is simply wrong.
the sturgis holding
For reasons it carefully explained, the Sturgis Board overruled Lee Hospital, 300 NLRB 947 (1990)—a 10-year-old decision, missing any rationale, which itself broke with precedent. Lee Hospital barred units that, without the employers’ consent, combine jointly and solely employed employees, even where all the employees at issue perform similar work for the user employer at a common worksite.3 331 NLRB at 1304–1308.
The key finding in Sturgis was that joint employers are fundamentally distinct from groups of independent employers who bargain in the traditional multiemployer setting. In the Sturgis Board’s words:
We decline to accept the faulty logic of Lee Hospital . . . that a user employer and a supplier employer—both of which employ employees who perform work on behalf of the user employer—are equivalent to the completely independent employers in multiemployer bargaining units. No pre-Lee Hospital Board conceived of such units as multiemployer units, and neither do we.
331 NLRB at 1305.
Significantly, Sturgis did not mandate any form of bargaining unit in a joint-employer situation. It merely authorized bargaining units combining jointly with solely employed employees, where such units would best enable all of the employees concerned to bargain with their employers, based on their community of interest. As Sturgis discussed, it had become critical that this form of unit be available, because more and more employees were working in alternative work arrangements. Id. at 1298. It remained for employees to decide whether to seek such a unit—and for the Board to find that the unit was appropriate, applying traditional tests.4
Finally, to provide an additional way for employees to achieve union representation, Sturgis limited Greenhoot, supra, so as to permit supplied employees to seek to bargain only with the supplier employer, to the extent that that employer controlled their terms and conditions of employment. 331 NLRB at 1308.
the proper resolution of this case
There is no dispute that the Employers here constitute a joint employer. There is also no dispute that the employees whom N&W Agency supplies to Oakwood are jointly employed by N&W and Oakwood; that they perform the same functions as Oakwood’s own employees; that both groups are intermixed virtually without distinction; or that both groups have common supervision and wear the same identification tags. In fact, the parties have quite properly stipulated that a joint bargaining unit would be appropriate here unless Sturgis is overruled. Accordingly, for all the reasons explained above, the Regional Director was correct in approving a joint unit per Sturgis.
the majority’s position
The new majority’s position is simple and stark: Sturgis “departed from the statutory directive of Section 9(b)” by “[a]pplying a novel definition of ‘employer’ fashioned for the purpose of deciding the case” and “ignoring the bright line between employer and multiemployer units.”5 According to the majority, permitting joint employees to be combined in a unit with sole employees, without the consent of their employer, is “inconsistent with the plain meaning of ‘employer unit’ in the Act.” Because two or more employers are involved, the Act prohibits any bargaining unit that involves both of them, unless they each consent.
The majority, then, insists that there is no difference between (1) two employers who enter into an arrangement with each other and who share control over certain employees at the same worksite; and (2) competing employers who have no arrangement with each other and who employ their own workers separately, at different worksites. In each case, in the majority’s view, the Act demands the consent of all the employers before bargaining with them can occur.
In addition to concluding that the Act precludes a Sturgis bargaining unit, the majority finds Sturgis deficient as a matter of “sound national labor policy.” Citing former Member Brame’s dissent, the majority asserts that the common bargaining unit authorized in Sturgis “gives rise to significant conflicts among the various employers and groups of employees participating in the process.” The majority cites no evidence at all—from either before or after Sturgis—for this assertion.
Finally, the new majority seems to reject the alternative form of bargaining unit authorized by Sturgis: a unit of all of a supplier employer’s employees to bargain only with the supplier employer—the classic case, it appears to us, of an “employer-wide unit.”
The majority is mistaken in every critical respect. Section 9(b) of the Act does not, by its terms, rule out a Sturgis unit. Such a unit is permitted by the statute, and, in fact, is necessary to enable the growing number of employees in alternative work arrangements to benefit from collective bargaining, if they so choose. None of the practical or policy objections raised by the majority stand up to scrutiny. Its approach, moreover, will hasten obsolescence of this statute.
A. Sturgis Is Consistent with Section 9(b)
Section 9(b) of the Act authorizes the Board to designate an appropriate bargaining unit consisting of “the employer unit, craft unit, plant unit, or subdivision thereof.” The majority asserts that “the text of the Act reflects that Congress has not authorized the Board to direct elections in units encompassing the employees of more than one employer.” But this view misreads the Act. The provision defines the scope of the bargaining unit, not the source of its members.
The language of Section 9(b) does not foreclose a unit comprising more than one “employer,” much less multiple employers who constitute a joint employer, a concept that has no specific basis in the Act’s language, but is well-established. E.g., Boire v. Greyhound Corp., 376 U.S. 473 (1964) (rejecting direct review of Board’s direction of election in joint-employer unit). It is, of course, undisputed that all the employees of one “employer” can constitute an appropriate bargaining unit. And the majority itself acknowledges that a “joint employer, under the Board’s traditional definition, is comprised of two or more employers, [and] all of the unit employees work for a single employer, i.e., the joint employer entity.” The majority concedes that a joint-employer unit consisting solely of the joint employees is appropriate, even absent the employers’ consent. Why the Act’s language permits such a unit, but forecloses a Sturgis unit, is unexplained.
The majority observes that a “craft unit, plant unit, or subdivision” of an employer unit are each, by definition, a sub-unit of an “employer unit.” But this fact tells us nothing about the outer scope of an “employer” or an “employer unit.”6 Similarly, a “plant unit” presumably can include all employees at a particular worksite, at least if they have an employer in common.
Contrary to the majority’s suggestion that Congress intended to restrict the Board’s authority, the Supreme Court has made clear that the “Board is accorded broad discretion” in making unit determinations. Allied Chemical and Alkali Workers v. Pittsburgh Plate Glass, 404 U.S. 157, 171 (1971).7 The aim, as the language of Section 9(b) does make clear, is “to assure employees the fullest freedom in exercising the rights guaranteed by th[e] Act.”
The legislative history cited by the majority does not support its position. Rather, it establishes only that Congress, in Section 9(b), intended to authorize the Board to find multiemployer units and employer sub-units appropriate. The only concern expressed by either the Wagner Act Congress or the Taft-Hartley Congress with respect to bargaining units that included more than one employer was focused on industrywide or anti-competitive bargaining units and on multiple-worksite situations.8 There was no expression of disapproval with respect to joint-employer units.
The majority’s restrictive application of Section 9(b) is based entirely on the premise that a joint-employer situation is no different than a multiemployer, multi-worksite situation. That premise is false. The critical difference, noted in Sturgis, is that where one or more supplier employers provides employees to a single user employer at a common worksite, all of the employees at the site work for the user employer. 331 NLRB at 1304–1305. Hence the unit scope is employerwide. Surely employees who are working side by side, for employers who have voluntarily created that arrangement, should be able to join together in the same bargaining unit, if they choose to. They are part of a common enterprise and, absent a common union representative, they are potential competitors with each other with respect to the terms and conditions of their work. Accordingly, where the Board’s other criteria for determining community of interest are met, it is appropriate for the joint employees to be combined with the user employer’s sole employees in a joint bargaining unit.9
By contrast, in the traditional multiemployer bargaining situation, the employers are entirely independent businesses, often compete with each other, operate at separate locations on different work projects, and hire their own employees. They have nothing in common except that they operate in the same industry. As Sturgis recognized, the Board developed the consent requirement in such cases precisely because the employers at issue were physically and economically separate from each other, their operations were not intermingled, and their employees were not jointly controlled.10
B. Sturgis Is
Consistent with Economic Realities and
In a series of recent decisions, the Board has demonstrated a disturbing reluctance to recognize changes in the economy and the workplace and to ensure that our law reflects economic realities and continues to further the goals that Congress has set.11 The reversal of Sturgis is part of this unfortunate trend. It represents a failure of the Board to fulfill its duty.
1. The Board must interpret the act in light of the rise of alternative work arrangements
As an administrative agency
responsible for enforcing Congressional policy, the Board has a fundamental
As the Board itself long has emphasized, specifically with respect to unit determinations:
Because the scope of the unit is basic to and permeates the whole of the collective bargaining relationship, each unit determination, in order to further effective expression of the statutory purposes, must have a direct relevancy to the circumstances within which collective bargaining is to take place. For, if the unit determination fails to relate to the factual situation with which the parties must deal, efficient and stable collective bargaining is undermined rather than fostered.
Kalamazoo Paper Box,
136 NLRB 134, 137 (1962). In
short, the Board has a statutory obligation under Section 9(b) to update its
approach in determining appropriate bargaining units in accordance with
current workplace realities.
In Sturgis, the Board met this obligation. The decision noted at the outset that according to the General Accounting Office’s (GAO) interpretation of Bureau of Labor Statistics (BLS) numbers, employment in the temporary help supply industry increased by 577 percent from 1982 to 1998, while total employment increased by only 41 percent. 331 NLRB at 1298. Given this significant economic development, failure of the Board to adjust its law of bargaining units within the terms of Section 9(b) would have been arbitrary and contrary to the Board’s duty.
Information subsequently available
only confirms that the Sturgis Board was right.14
In a report prepared for the U.S. Department of Labor, two RAND
Corporation researchers observe that “less-permanent employment arrangements
have taken root and are likely to become more prevalent in the face of rapid
technological change and competitive market pressures.”15
In many cases, however, it is practically impossible for supplied employees who work in a joint-employer setting to organize or to bargain effectively with their joint employer, unless they can join with the user employer’s sole employees who work alongside them. By the same token, the solely-employed employees in that setting can be handicapped in improving their conditions of employment if they are barred from organizing and bargaining with the supplied workers.
There is no dispute that many employers use contingent workers simply to reduce their labor costs.16 Employers get the biggest savings through the limited fringe benefits provided to contingent workers.17 Another motivation, of course, is to prevent core and contingent employees alike from organizing and bargaining effectively.18 It is surely no coincidence that, to date, we can cite no instance where contingent employees sought and obtained the consent of both user and supplier employers to be included in a unit with the user employer’s sole employees.
Given this reality, it makes no sense for the majority to say that traditional multiemployer bargaining units help to “bring about industrial stability for the benefit of all,” but that Sturgis units would not. The Act does not envision that “industrial stability” will be achieved by frustrating the ability of workers to organize. But this is precisely what the majority’s decision will do, at worst accelerating the expansion of a permanent underclass of workers.19 Ironically, the American Staffing Association itself has cited Sturgis to rebut allegations that temporary employees are not protected by federal labor law.20
bargaining units facilitate rather
In the majority’s view, of course, the ability of workers to organize is trumped by other concerns. Our colleagues argue that under Sturgis the resulting “bargaining structure . . . gives rise to significant conflicts among the various employers and groups of employees participating in the process.” They cite the prospect that the employers involved would be forced to “negotiat[e] with one another as well as with the union,” or that the interests of the solely or the jointly employed group of employees “may be traded away” for the other’s, with resulting “fragmentation” of bargaining.21 The majority also posits several examples of such “conflicts.”
But these conflicts are entirely hypothetical. Neither the majority nor the employer parties have cited evidence that any Sturgis unit has created such conflicts or disrupted a bargaining process. Regardless, the solution the majority endorses—making it virtually impossible for certain workers to organize—is worse than the supposed problem. Perhaps the simplest refutation of the majority’s “policy” rationale is provided by no less reliable a source than the National Association of Professional Employer Organizations (NAPEO). NAPEO, on its website, publicly affirms that its supplier-employer members operate as “co-employers” with their clients, and states in the light of Sturgis:
The National Labor Relations Board recognizes that in co-employment relationships, worksite employees are appropriately included in the client employer’s collective bargaining unit. Where a collective bargaining agreement exists, [professional employer organizations] fully abide by the agreement’s terms.
http://www.napeo.org/peoindustry/faq.fm#18 (emphasis added). However, even assuming that the “conflicts” posited by the majority did arise, they would not differ from the conflicts that have always existed in any setting where joint or multiple employers bargain jointly, or even (with respect to inter-employee conflicts) in the setting of a single employer. Employees are always required to rank their interests and to resolve competing interests between employee sub-groups. That this might occur in Sturgis units, too, is no reason to prevent employees from choosing them.22
In any case, the Board has long recognized that “the fact that two or more groups of employees may have some different interests does not render a combined unit inappropriate if there is a sufficient community of interest among all of the employees.” Holiday Inn City Center, supra, 332 NLRB at 1250; Berea Publishing Co., 140 NLRB 516, 518 (1963).
Similarly, there are competing employer interests wherever joint or multiple employers are involved, and some employers in the group will exert greater leverage than others. This, again, is the normal byproduct of give-and-take group bargaining. In a contingent-employee, joint-employer setting, the user employer may have the economic power to dictate the management position on joint employees’ terms of employment. Alternatively, the employers involved must negotiate their conflicting interests among themselves, regardless of the nature of the bargaining unit(s) with which they are dealing.
Contrary to the majority’s assertions, in most settings where joint employees seek a Sturgis unit, it is because the user employer, not the supplier employer, holds the real power to set their terms and conditions of employment. The true significance of Sturgis units is to make it possible for all the user employer’s employees—both those solely employed by the user employer and those jointly employed by the user and supplier employers—to organize and bargain effectively to improve their terms of employment.
The Board’s statutory mandate, as we have now said repeatedly, is to “to assure to employees the fullest freedom in exercising the rights guaranteed by th[e] Act.” 29 U.S.C. §159(b). To the extent that joint employers can prevent their employees from seeking the most effective bargaining unit simply by withholding their consent, they can “shape the unit at will.”23 That result—which exalts business flexibility at the complete expense of employee rights—is the opposite of what Congress intended.
Finally, the majority also appears to be barring a petition seeking to represent only the supplier employer’s employees, and naming only the supplier employer. How that holding can be justified under Section 9(b) is mysterious. By its holding, the majority essentially requires employees to bargain in multiple units in every joint employer case, regardless of the circumstances. After today, a contingent employee cannot be a member of a bargaining unit that includes the user employer’s sole employees, even if all of them work at the same worksite and their work functions are indistinguishable. Nor, however, can the contingent employee attempt to organize all, or even some, of his or her fellow-contingent employees to bargain only with the supplier employer to the extent that that employer controls their terms of employment. The only option the contingent employee has is to organize the other contingent employees to bargain with both the supplier and user employers, but in complete isolation from the user employer’s own employees. Similarly, the user employer’s own employees can only organize in a unit of their own, regardless of the contingent employees who are performing identical work right next to them. The majority, then, seems to have gone out of its way to make it impossible for joint employees to exercise their Section 7 rights effectively. Ironically, they also seem to be creating a collective bargaining nightmare for employers.
While Sturgis was decided only 4 years ago, it built on well-established principles, took economic realities into account, and reversed a relatively recent decision that itself was out of step with Board precedent. The majority today not only disagrees with Sturgis on policy grounds, it also argues that Sturgis exceeded the Board’s statutory authority. If the majority were correct, then the National Labor Relations Act itself could not guarantee an important, and growing, segment of American workers the right to collective bargaining. The problem here, however, is not the statute, but the agency that administers it. We dissent.
Dated, Washington, D.C. November 19, 2004
National Labor Relations Board
 The Employer also requested review of the Regional Director’s determination that a substantial and representative complement of unit employees existed, so that an election could be conducted. We find it unnecessary to pass on that issue, however, based on our decision regarding the appropriateness of the unit.
 331 NLRB 1298 (2000).
 The parties stipulated to the facts in this proceeding, and no witnesses testified at the hearing.
 205 NLRB 250 (1973).
 300 NLRB 947 (1990).
Greenhoot, supra at 251.
Lee Hospital, supra at 948,
citing Greenhoot, supra at 251
 The petition concerning the Sturgis unit named only Sturgis as the employer; the Jeffboat petition named both Jeffboat and TT&O.
 Sturgis, supra at 1305.
 Id. at 1308.
Dissenting, Member Brame found that Sec. 9(b) does not authorize the Board
to certify bargaining units that encompass both the solely employed
employees of an employer and employees jointly employed by that and another
 H.R. Statement on Conf. Rep. S. 1958, 79 Cong. Rec. 10297, 10299 (1935), reprinted in 1 Leg. Hist. 3260, 3263 (NLRA 1935).
 See, e.g., Shipowners Assn. of the Pacific Coast, 7 NLRB 1002, 1024–1025 (1938), review denied sub nom. AFL v. NLRB, 103 F.2d 933, affd. 308 U.S. 401 (1940). See also H. Conf. Rep. No. 510 on H.R. 3020, reprinted in 2 Leg. Hist. 535–536 (LMRA 1947).
 NLRB v. Truck Drivers Local 449 (Buffalo Linen Supply Co.), 353 U.S. 87, 96 (1957).
 NLRB v. Greyhound Corp., 368 F.2d 778, 780 (5th Cir. 1966), on remand from Boire v. Greyhound Corp., 376 U.S. 473 (1964).
 See, e.g., S.S. Kresge, 169 NLRB 442, enfd. in pert. part 416 F.2d 1225 (6th Cir. 1969) and S.S. Kresge, 162 NLRB 498 (1966), enfd. in pert. part sub nom. Gallenkamp Stores v. NLRB, 402 F.2d 525 (9th Cir. 1968).
See also Stack & Co., 97 NLRB 1492 (1952) (finding that
department store was an employer of leased department employees without
addressing whether resulting unit was multiemployer in scope).
 Gallenkamp, supra at 529.
 NLRB v. S.S. Kresge, supra, 416 F.2d at 1227.
Id. at 1231.
 The difficulties that arise from the Sturgis holding are necessarily magnified in situations involving multiple supplier employers.
 As Member Brame correctly noted in Sturgis, additional problems could arise in designating whether union activity is primary or secondary under Sec. 8(b)(4)(ii)(B).
 Concededly, in a multiemployer situation, employees of different employers are grouped together in one unit and covered by one contract. But this is because multiemployer bargaining can bring about industrial stability for the benefit of all. See Retail Associates, 120 NLRB 388 (1958). Multiemployer bargaining requires the consent of all, an element missing here.
As noted above, the required consent, consistent with well-established
precedent concerning multiemployer bargaining, must be clear and
unequivocal, as manifested by express agreement or by actually entering into
bargaining on a multiemployer basis. Employer
consent will not be inferred from the joint employer relationship.
Greenhoot, supra, 205 NLRB at 251.
1 These workers thus join university graduate student employees and disabled workers in rehabilitative employment in what amounts to a legal underclass. See Brown University, 342 No. 42 (2004) (graduate teaching assistants); Brevard Achievement Center, 342 NLRB No. 101 (2004) (disabled workers).
2 As Capelli has explained elsewhere:
[T]he common thread in these developments is that they represent a weakening of the traditional bureaucratic employment relationship, with its long-term commitments and internal labor market practices, that buffered employment from outside markets. In its place, more immediate market forces increasingly govern the internal operations of firms and, in turn, the employment relationship. The phrase ‘market-mediated employment relationships’ might describe the new practices that are based around more short-term, contractual relationships shaped by pressures from the outside labor market and individualized incentives.
“Market-Mediated Employment: The Historical Context” in The New
Relationship: Human Capital in the American Corporation, Blair and
Kochan, eds., at 66-69 (2000). See
also Lynn Karoly & Constantijn Panis, The 21st Century at Work
116–119, 192-194 (2004) (discussing “nonstandard work arrangements”).
3 Lee Hospital had improperly extended the holding of Greenhoot, Inc., 205 NLRB 250 (1973), in which the Board found that where a supplier provided maintenance employees to multiple building owners at different locations, resulting in separate joint-employer relationships between the supplier and each owner, a single bargaining unit combining all the owners’ employees with all the supplied employees was impermissible without the owners’ consent. 205 NLRB at 251. Sturgis reaffirmed Greenhoot in this respect. 331 NLRB at 1303-1305.
The majority mischaracterizes Gourmet Award Foods, 336 NLRB
872 (2001), a decision involving the statutory duty to bargain in good
faith, as demonstrating that Sturgis results in “anomalous
decisions.” In fact, Gourmet Award Foods involved the traditional
principle “that when an established bargaining unit expressly encompasses
employees in a specific classification, new employees hired into that
classification are included in the unit.”
Id. at 873. The
newly-hired, jointly-employed employees there fell within the job
classifications covered by an existing collective-bargaining agreement.
Including them in the unit was necessary to preserve the union’s
bargain with the employer. Sturgis,
in turn, established that combining jointly-employed and solely-employed
employees in the same unit was not inappropriate. Id. at 874.
5 Sec. 9(b) of the Act provides, in pertinent part:
Board shall decide in each case whether, in order to assure to employees the
fullest freedom in exercising the rights guaranteed by this Act, the unit
appropriate for the purposes of collective bargaining shall be the employer
unit, craft unit, plant unit, or subdivision thereof.
29 U.S.C. 159(b) (emphasis added).
Correspondingly, the first sentence of Sec. 9(a)—the sentence in
the Act that establishes the right of employees to have an exclusive
bargaining representative–-does not even refer to an “employer.”
The sentence authorizes “a unit appropriate for [collective
bargaining] purposes.” Neither
Sec. 9(a) nor any other part of the Act even suggests that an appropriate
unit cannot include the joint and sole employees of a joint employer.
7 “[The Taft-Hartley] Congress . . . intended to leave to the Board’s specialized judgment the inevitable questions concerning multiemployer bargaining bound to arise in the future.” NLRB v. Truck Drivers Local 449 (Buffalo Linen), 353 U.S. 87, 96 (1957) (emphasis added). “While the exact limits of the Board’s powers or the precise meaning of the terms have not been fully defined, judicially, we know that they lie within the area covered by the words ‘employer,’ ‘plant,’ and ‘craft.’ [footnote omitted] . . . As a standard, the Board must comply, also, with the requirement that the unit selected must be one to effectuate the policy of the act, the policy of efficient collective bargaining. Where the policy of an act is so definitely and elaborately stated, this requirement acts as a permitted measure of delegated authority.” Pittsburgh Plate Glass v. NLRB, 313 U.S. 146, 165-166 (1941).
8 National Labor Relations Board, Legislative History of the National Labor Relations Act, 1935 at 1300, 3219-3221, 3253-3256, 3264-3267 (1985 Reprint); National Labor Relations Board, Legislative History of the Labor Management Relations Act, 1947 at 58-61, 117, 121-122, 299-300, 535–536, 550-551, 584, 612, 636, 643–644, 663, 672-674 (1985 Reprint).
9 The majority also attempts to distinguish cases involving contingent employees from earlier cases, on which Sturgis in part relied, where the Board consistently found combined units of joint and sole employees appropriate in department-store settings. By way of example, the majority characterizes the department store and single-store licensee employers in Gallenkamp Stores v. NLRB, 402 F.2d 525 (9th Cir. 1968) as not only a joint employer but a “joint enterprise—a unique relationship not present in Sturgis or this case.”
The majority has it backwards.
In Gallenkamp, the licensed single-store operations were
integrated into K-Mart’s storewide operation as whole, self-contained
entities. In Sturgis and
in this case, by contrast, the supplier employers retained no separate
identity whatsoever within the user employer’s operation or on the user
employer’s worksite: they were merely an external source of employees who
were absorbed and intermixed among the user employer’s sole employees.
The Sturgis employers and the Employers here therefore
comprised a “joint enterprise” at their respective common workplaces to
a greater extent than did the Gallenkamp employers.
See, e.g., Chapman Dehydrator, 51 NLRB 664, 666–667 (1943)
(multi-employer unit not appropriate absent consent, where there was no
evidence of “any managerial interrelationship between members of the
Association,” and employers “operate . . . as separate and distinct
business organizations with no interchange of employees”); Sagamore
Mfg., 39 NLRB 909, 915–916 (1942) (same result where employers were
“independent and competing” with each other); F. E. Booth & Co.,
10 NLRB 1491, 1496 (1939) (same result where interchange of employees
between employers was “not effectuated by any plan of [the union] or
through any common agency of the companies” and “each of the companies
hires its own employees as the conditions of its business require”); Alaska
Packers Assn., 7 NLRB 141, 143–144 (1938) (same result even though
“the three companies constitute an economic [regional industry]
aggregate,” because they are “separate and distinct business
organizations”). See also Bull-Insular Line, 56 NLRB 189, 193–194
(1944) (Puerto Rico-wide employer unit not appropriate absent consent, but
unit of two employers appropriate where they were “interlocking corporate
organizations” and their employees “together are engaged in various
tasks incidental to the loading or unloading of cargo vessels at the [two]
companies’ piers in the harbor of San Juan”).
Cf. Rayonier, Inc., 52 NLRB 1269, 1274 (1943) (multi-employer
unit appropriate in view of implied consent through collective-bargaining
history, in contrast to cases where employers are merely “competing
companies not otherwise related except through membership in an employer
11 See MV Transportation, 337 NLRB 770 (2002); Airborne Express, 338 NLRB No. 72 (2002); Brown University, 342 NLRB No. 42 (2004); Brevard Achievement Center, 342 NLRB No. 101 (2004).
American Trucking Assns. v. Atchison, Topeka & Santa Fe
Railway, 387 U.S. 397, 416 (1967).
13 Ford Motor Co. v. NLRB, 441 U.S. 448, 496 (1979) (emphasis added), quoting NLRB v. Insurance Agents, 361 U.S. 477, 499 (1960); NLRB v. Erie Resistor Corp., 373 U.S. 221, 236 (1963); and NLRB v. Steelworkers, 357 U.S. 357, 362–363 (1958).
Bureau of Labor Statistics, U.S. Department of Labor, “Contingent
and Alternative Employment Arrangements, February 2001,” (May 24, 2001).
Other commentators have estimated that contingent employees comprise
between 16 percent and 29 percent of the national work
force. See Stephen
Befort, “Revisiting the Black Hole of Workplace Regulation: A Historical
and Comparative Perspective of Contingent Work,” 24 Berkeley Journal of
Employment & Labor Law 153, 158-159 fn. 46-49 (2003) (and sources
cited); Tiffany Fonseca, “Collective Bargaining Under the Model of M.B.
Sturgis, Inc.,” 5 U. of Pa. Journal of Labor and Employment Law
167, 172 fn. 39 (2002) (and sources cited).
15 Karoly & Panis, supra, The 21st Century at Work, at 192.
See Michael Hely, “The Impact of Sturgis on Bargaining Power
for Contingent Workers in the U.S. Labor Market,” 11 Washington U. Journal
of Law & Policy 295, 307, fn. 51 (2003)(and sources cited); Patricia
Ball, “The New Traditional Employment Relationship: An Examination of
Proposed Legal and Structural Reforms for Contingent Workers,” 43 Santa
Clara Law Review 901, 915-916 fn. 119-128 (2003) (and sources cited);
Stephen Befort, “Labor and Employment Law at the Millenium: A Historical
Review and Critical Assessment,” 43 Boston College Law Review 351, 369 fn.
121-123 (2002) (and sources cited).
Capelli, Change at Work, supra, at 77. See also Pamela Mendels,
“Allowing Temps to Organize,” BusinessWeek online, http://www.businessweek.com/careers/content/sep2000/ca20000914_957.htm
(“Employers use temps because they are cheaper than permanent staff: A 1999 Bureau of Labor Statistics study found, for example that only about 8.5% of temps get health insurance from their employer, compared to 58% of traditional workers”).
Lobel, supra at 116 fn. 27-28; Befort, Revisiting the Black Hole of
Workplace Regulation,” supra at 163 fn. 80-82; Hely, supra at 308 fn.
55-58; Befort, “Labor and Employment Law at the Millennium,” supra at
370-371 fn. 129–132; Fonseca, supra at 186 fn. 166.
19 According to a 1999 BLS study of alternative employment arrangements, temporary help agency workers (those most like the Sturgis situation) were more likely than other workers to be young, female, African-American or Hispanic. Relatively few of the young workers in this arrangement were going to school: just 16 percent were attending high school or college, compared with 43 percent of young people working in a traditional job. The great majority of temps worked a full-time week of at least 35 hours. Of those who worked part time, 41 per cent would have preferred a full-time job, compared with only 18 per cent of workers in traditional jobs. These workers predominantly worked in clerical and machine operator occupations. Just one in three preferred their arrangement to a traditional job. A majority said they were working in the arrangement because that was the only type of work they could find. Temporary help agency workers had low earnings. At $329 per week for full-time workers, their median earnings were about two-thirds of the earnings of traditional workers. Temporary agencies did not commonly provide health insurance and pension benefits. Just 46 percent of temps had health insurance from any source, compared with 83 percent of traditional workers. Pension coverage was even lower. Sharon R. Cohany, “Workers in Alternative Employment Arrangements: A Second Look,” Monthly Labor Review, Nov. 1998, Vol. 121, No. 11, at 10-15.
20 American Staffing Association, “The Staffing Services Industry: Myth and Reality” (February 15, 2001), available at http://www.staffingtoday.net/staffstats/index.html.
The prospect that the employers involved would be forced to negotiate
with one another as well as with the union is, of course, true in any joint
employer situation, not just the Sturgis scenario.
22 As Sturgis noted, in S.S. Kresge v. NLRB, 416 F.2d 1225 (6th Cir. 1969), which the majority vainly attempts to distinguish, the Sixth Circuit made precisely this point. 331 NLRB at 1307 fn. 18, quoting 416 F.2d at 1232.
23 Waterfront Employers Assn., 71 NLRB 80, 111 (1946) (finding a multiemployer unit appropriate).