Opinion
23-60175
09-18-2024
Hudson Institute of Process Research Incorporated, formerly known as Hudson, a Professional Corporation, formerly known as H.I.P.R. Pacsoft Technologies, Incorporated, Petitioner/Cross-Respondent, v. National Labor Relations Board, Respondent/Cross-Petitioner.
Appeal from the National Labor Relations Board Agency No. 06-CA-306766.
Before CLEMENT, SOUTHWICK, and HO, Circuit Judges.
EDITH BROWN CLEMENT, Circuit Judge
Hudson Institute of Process Research Incorporated petitions for review of two National Labor Relations Board determinations affecting its employees' union election: (1) the NLRB's findings that certain personnel are not "supervisors" statutorily excluded from collective bargaining units; and (2) the NLRB's certification of an employer-wide bargaining unit. The NLRB cross-petitions for enforcement of its order directing Hudson to recognize and bargain with the union. Because the NLRB lacks substantial evidence that certain personnel are not supervisors, we GRANT Hudson's petition for review, REVERSE the bargaining order, and DENY enforcement.
I.
A.
In July 2021, the United Electrical, Radio, and Machine Workers of America petitioned for an election to represent the employees of Hudson, a legal outsourcing and staffing company that provides legal support services to companies and individuals seeking employment-based visas.
The NLRB was then tasked with determining the "bargaining unit," i.e., "the group of workers that will vote on union representation." Kindred Nursing Ctrs. E., LLC v. NLRB, 727 F.3d 552, 554 (5th Cir. 2013). By statute, the "bargaining unit" must exclude "supervisors." See 29 U.S.C. § 152(3) (excluding supervisors from definition of "employee"); 29 U.S.C. § 157 (granting "employees" the right to unionize).
We use the "Board" when referring to the quasi-judicial body consisting of up to five members, who are appointed by the President and confirmed by the Senate and "NLRB" when referring to the National Labor Relations Board as an agency and party to this proceeding. See 29 U.S.C. §153(a); see also 29 U.S.C. § 153(b) (authorizing the Board to delegate its authority).
Hudson and the union disagreed about whether certain personnel- specifically, team leads, team lead assistants, floating team lead assistants, and the entire revision specialist team-should be considered supervisors. Hudson also objected to an employer-wide bargaining unit.
The NLRB held a representation hearing in September 2021. Hudson sought to show that the individuals in the disputed positions were supervisors because they possessed authority to assign work, to discipline or recommend discipline, and to recommend rewards such as bonuses and wage increases.
On December 2, 2021, the NLRB regional director issued a Decision and Direction of Election finding that Hudson "failed to meet[] its burden to establish, by a preponderance of the evidence, that team leads, team lead assistants, floating team lead assistants and members of the revision specialist team are supervisors." The regional director also determined that Hudson had failed to show that an employer-wide bargaining unit was inappropriate.
Hudson appealed the regional director's decision to the Board. While Hudson's appeal was pending, the NLRB conducted an election by mail from December 20, 2021, through January 18, 2022. Due to Hudson's pending appeal, ballots were not immediately counted at the close of voting.
A regional director's decisions are subject to the Board's discretionary review. See 29 U.S.C. § 153(b).
In March 2022, the Board issued a single-sentence order (accompanied by a page-and-a-half long footnote) denying Hudson's request for review. Member John Ring dissented, explaining that he would grant Hudson's request for review to consider the team leads', team lead assistants', and floating team lead assistants' authority to reward employees.
The Board's denial of a request for review constitutes an affirmance of the regional director's decision. 29 C.F.R. § 102.67(g).
The ballots were eventually counted in May 2022, and on September 15, 2022, the NLRB regional director certified the election in favor of the union. Hudson again appealed to the Board, which declined its request for review on October 26, 2022. To preserve its right to judicial review, Hudson refused to bargain with the union and as a result the NLRB found that it had committed an unfair labor practice, in violation of the National Labor Relations Act ("NLRA").
B.
Hudson provides immigration services to businesses and individual clients seeking employment-based visas. Prior to the pandemic, Hudson maintained offices throughout the country. But as of the administrative proceeding, Hudson only had open offices in Ann Arbor, Michigan (its headquarters at the time) and Pittsburgh, Pennsylvania, with a third office in Chicago that was closed. By the time the company filed its petition for review, Hudson had moved its headquarters to Dallas, Texas.
Hudson has various teams of paralegals who assist in preparing employment-based visas for its clients. The teams generally focus on one stage of the immigration process, generating work product that an attorney reviews and submits to U.S. Citizenship and Immigration Services ("USCIS").
There are six types of teams relevant here: (1) I-140 teams, which handle the initial immigration petitions and draft supporting letters; (2) package (or legal evidence) teams, which gather and process clients' evidence; (3) the forms team, which fills out immigration forms; (4) requests for evidence ("RFE") teams, which handle cases in which USCIS requests additional evidence to support the I-140 petition; (5) I-485 teams, which handle the step that occurs after an I-140 petition has been approved; and (6) the revision specialist team, which handles sensitive editing projects (e.g., if a client flags an issue in a document drafted by an I-140 team).
Except for the revision specialist team, all teams are led by team leads, who are in turn managed by attorneys. Most team leads oversee around four or five writers or paralegals. The I-140 teams, which make up most of Hudson's teams, also have team lead assistants, who share many of the same responsibilities as the team lead. Hudson also employs floating team lead assistants, who rotate among the I-140 teams depending on absences and workflow needs. The revision specialist team is made up of more advanced writers and is overseen by training managers rather than team leads.
While their specific responsibilities vary by team and are described in greater detail below, Hudson's team leads generally assign work within their team (some of whom recently started to use a project management software to do so) and assess and evaluate the writers and paralegals they oversee. Hudson says that the resulting evaluations are considered when promoting and compensating the team members. Most team leads may also place their team members on performance improvement plans. Team leads also approve time-off and overtime requests.
II.
While neither Hudson nor the NLRB challenges our ability to review the petitions, we must confirm that we have jurisdiction as a threshold matter. See Pervasive Software Inc. v. Lexware GmbH &Co., 688 F.3d 214, 23132 (5th Cir. 2012).
Under the NLRA, "[a]ny person aggrieved" by an NLRB final order may file a petition for judicial review in "any United States court of appeals in the circuit wherein the unfair labor practice in question was alleged to have been engaged in or wherein such person resides or transacts business, or in the United States Court of Appeals for the District of Columbia." 29 U.S.C. § 160(f). The NLRB may file petitions (including cross-petitions) for enforcement of its final orders in "any court of appeals of the United States . . . wherein the unfair labor practice in question occurred or wherein such person resides or transacts business." 29 U.S.C. § 160(e).
Several of our sister circuits have held that all courts of appeals have jurisdiction to review and enforce NLRB orders and that the "where" requirements of § 160(e) and (f) "go to venue, not subject-matter jurisdiction." E.g., Brentwood at Hobart v. NLRB, 675 F.3d 999, 1002 (6th Cir. 2012); NLRB v. Wilder Mfg. Co., 454 F.2d 995, 998 n.12 (D.C. Cir. 1972) But we have not expressly addressed the issue.
Although we referred to § 160(f) as granting "jurisdiction" in Bally's Park Place, Inc. v. NLRB, that decision did not directly resolve the question of whether the "where" provisions of § 160(e) and (f) speak to jurisdiction or venue. 546 F.3d 318, 319 (5th Cir. 2008). And "[w]here an opinion fails to address a question squarely, we will not treat it as binding precedent." Thomas v. Tex. Dep't of Crim. Just., 297 F.3d 361, 370 n.11 (5th Cir. 2002); see also Steel Co. v. Citizens for a Better Env't, 523 U.S. 83, 91 (1998) ("[D]rive-by jurisdictional rulings . . . have no precedential effect.").
We agree with our sister circuits that § 160(e) and (f) govern venue, not jurisdiction. As the Seventh Circuit explained:
As geographic limitations, [§ 160(e) and (f)] ask the "where"-the venue-"question." And the answer they give turns on classic venue concerns-"choosing a convenient forum." By generally permitting the action to proceed in the circuit where "the unfair labor practice in question" occurred or where the company "resides or transacts business," the provisions ensure that the company will not be forced to defend an action in a faraway circuit.Brentwood, 675 F.3d at 1002 (citations omitted). This approach is in line with the Supreme Court's effort to "bring some discipline to the use of the term 'jurisdictional.'" Gonzales v. Thaler, 565 U.S. 134, 141 (2012) (internal quotation marks omitted); see also Arbaugh v. Y &H Corp., 546 U.S. 500, 515 (2006) (holding that a statute is only jurisdictional "[i]f the Legislature clearly states that a threshold limitation on a statute's scope shall count as jurisdictional").
As a court of appeals, we have jurisdiction over the parties' petitions. See Wilder Mfg., 454 F.2d at 998 n.12 ("[A]ll intermediate federal courts have jurisdiction to review and enforce orders of the NLRB."). And the parties agree that venue is proper. This case is therefore properly before our court.
III.
"Whether an employee is a supervisor is a question of fact." Entergy Gulf States, Inc. v. NLRB, 253 F.3d 203, 208 (5th Cir. 2001). The party asserting supervisory status bears the burden of proof. Entergy Miss., Inc. v. NLRB, 810 F.3d 287, 293 (5th Cir. 2015) (citing NLRB v. Ky. River Cmty. Care, Inc., 532 U.S. 706, 711-12 (2001)).
The NLRB's findings of fact are "conclusive" if they are "supported by substantial evidence on the record considered as a whole." 29 U.S.C. § 160(e), (f); STP Nuclear Operating Co. v. NLRB, 975 F.3d 507, 513 (5th Cir. 2020). "Substantial evidence" "means such relevant evidence as a reasonable mind might accept as adequate to support a conclusion." Universal Camera Corp. v. NLRB, 340 U.S. 474, 477 (1951). "It is more than a mere scintilla, and less than a preponderance." STP Nuclear, 975 F.3d. at 513 (quoting El Paso Elec. Co. v. NLRB, 681 F.3d 651, 656 (5th Cir. 2012)). While a reviewing court "may not reweigh the evidence . . . or substitute [its] judgment for that of the [NLRB]," its review is not "pro forma" or "merely a rubber stamp." Creative Vision Res., LLC v. NLRB, 882 F.3d 510, 515 (5th Cir. 2018) (quotation marks and citations omitted). In other words, "[o]ur deference . . . has limits." Carey Salt Co. v. NLRB, 736 F.3d 405, 410 (5th Cir. 2013). For instance, "a decision by the [NLRB] that ignores a portion of the record cannot survive review under the substantial evidence standard." Id. (quotation marks and citation omitted).
We will affirm the NLRB's legal conclusions "if they have a reasonable basis in the law and are not inconsistent with the [NLRA]." Entergy Miss., 810 F.3d at 292 (citation omitted). The Supreme Court has recently clarified that we use traditional tools of statutory interpretation when so assessing; we do not simply defer to an agency's interpretation of "ambiguous" provisions of their enabling acts. See Loper Bright Enterprises v. Raimondo, 144 S.Ct. 2244, 2266 (2024). Loper Bright, however, "d[id] not call into question prior cases that relied on the Chevron framework." Id. at 2273. Rather it explained that "[m]ere reliance on Chevron cannot constitute a 'special justification'" for overruling precedent. Id. Further, where a statute delegates discretionary authority to an agency, "the role of the reviewing court . . . is, as always, to independently interpret the statute and effectuate the will of Congress subject to constitutional limits." Id. at 2263.
Where, as here, the Board adopts the regional director's findings and conclusions, we apply the same standards to the regional director's decision itself. See In-N-Out Burger, Inc. v. NLRB, 894 F.3d 707, 714 (5th Cir. 2018).
IV.
"Supervisor[s]" are not "employees" under the NLRA and are therefore not entitled to collective bargaining rights. See 29 U.S.C. § 152(3) (excluding "any individual employed as a supervisor" from the NLRA's definition of "employee"); 29 U.S.C. § 157 ("Employees shall have the right to self-organization . . .") (emphasis added). "Supervisor" is defined as:
[A]ny individual having authority, in the interest of the employer, to hire, transfer, suspend, lay off, recall, promote, discharge, assign, reward, or discipline other employees, or responsibly to direct them, or to adjust their grievances, or effectively to recommend such action, if in connection with the foregoing the exercise of such authority is not of a merely routine or clerical nature, but requires the use of independent judgment.29 U.S.C. § 152(11).
The Supreme Court has distilled this definition into a three-part test. Individuals are supervisors if "(1) they hold the authority to engage in any 1 of the 12 listed supervisory functions, (2) their exercise of such authority is not of a merely routine or clerical nature, but requires the use of independent judgment, and (3) their authority is held in the interest of the employer." Ky. River, 532 U.S. at 712-13 (quotation marks and citation omitted). "[T]he possession of authority to engage in any of these functions-even if this authority has not yet been exercised-is what determines whether an individual is a supervisor." STP Nuclear, 975 F.3d at 514; see also Berry Schs. v. NLRB, 627 F.2d 692, 697 (5th Cir. 1980) ("[A] person is a supervisor even if he does not have complete authority to perform any of these functions but can 'effectively recommend' the performance of one or more.").
Hudson contends that the NLRB erred in finding that the personnel in the contested roles did not have authority to (1) assign work and (2) reward and promote employees and/or recommend the same, using their independent judgment. The evidence that Hudson presented to the NLRB and the responsibilities of the relevant positions vary by team, so we address Hudson's teams separately. As explained below, the NLRB's findings regarding the I-140, RFE, and I-485 team leads, team lead assistants ("TLA"), and floating TLAs are not supported by substantial evidence.
The third element of the supervisor test-that the individuals act in the interest of the employers-is not in dispute here.
We decline to address whether the NLRB's determinations that package and form team leads and the revision specialist team are supervisors are supported by substantial evidence. Hudson also challenges the NLRB's findings that the personnel in the contested possessions do not possess supervisory authority to promote, discipline, or discharge employees. Because the authority to engage in any one of the twelve supervisory functions listed in the statute is sufficient to establish supervisor status, see Ky. River, 532 U.S. at 713, we need not reach the validity of these findings.
A.
Hudson's I-140 teams handle the initial employment-based immigration petitions and supporting letters. Each I-140 team has a team lead, a TLA, and four to five legal writing specialists (i.e., writers). The NLRB found that "the [team lead] is usually responsible for assigning work to the writers and the TLA typically reviews and edits the writer's documents. However, [team leads] and TLAs can decide between them who will handle those specific tasks." Because the team lead and TLA generally share the same authority, we assess the roles together, along with "floating" TLAs (I-140 team leads, TLAs, and floating TLAs are referred to collectively as the "I-140 Leaders").
In addition to TLAs who are "anchored" to a given I-140 team, Hudson employs floating TLAs who rotate among I-140 teams depending on the caseload of a particular team. Floating TLAs can fill in for a team lead "in the short term or during an extended leave of absence." When doing so, floating TLAs "have the authority to complete quarterly evaluations, to assign work, to give recommendations regarding discharges, wage increases and [performance improvement plans], and essentially possess all the duties of a [team lead] or TLA."
1.
Despite explaining that an I-140 Leader is "usually responsible for assigning work to the writers," the NLRB found that I-140 Leaders do not "assign" within the meaning of 29 U.S.C. § 152(11), but rather that Hudson's project management software (or PMS) did.
The project management system works as follows: the I-140 Leaders identify the skill of a writer, using a scale from one to three, and likewise identify a project's difficulty on a scale of one to three, and the software then matches projects to correspond with a writer's skill level. The NLRB concluded that, due to the project management software, "the role of the [I- 140 Leaders] in the assignment of work is quite limited. They categorize the type of work product and level of writer, based on the writers' training or skill, and input the document for assignment." While recognizing that I-140 Leaders sometimes assign work independently of the project management software, the NLRB concluded that the I-140 Leaders did not use independent judgment in doing so, explaining that these assignments "merely reflect[ed] that the [I-140 Leaders] underst[ood] the complexity level of various documents and [were] aware of which employees ha[d] been trained to draft what types of documents."
Hudson argues that the NLRB erred in two ways: first, by ignoring evidence that the I-140 Leaders retain authority to depart from the software's writer-to-project pairings; and second, by failing to recognize that I-140 Leaders use independent judgment to match writers to projects based on skill level and complexity.
As to point one, Hudson points out that its managing partner testified at the hearing that the I-140 Leaders may "change almost everything about a project" including the project's start date, the writer, and the deadline. The company contends that the NLRB reached its conclusion by wrongly focusing on the I-140 Leaders' exercise of authority rather than their mere possession of it. Indeed, the NLRB's decision largely hinged on the fact that "[t]here were no examples of a [I-140 Leader] rejecting or overriding the computer assignment." And on appeal, the NLRB contends that "Hudson does not identify any . . . instances of [I-140 Leaders] exercising independent judgement" and that I-140 Leaders possess mere "paper authority" to reassign work independent of the project management software. Accordingly, Hudson argues, by failing to recognize that the I-140 Leaders maintain authority to override, modify, or reject the software's generated project matches, or "change almost everything about a project," the NLRB committed a legal error.
As to point two, Hudson argues that the NLRB ignored record evidence that it was I-140 Leaders-not the PMS-exercising ultimate assignment authority. As the company describes it, the software is "better understood as a tool that confirms decisions made by the supervisors, not the supervisors' replacement."
We agree with Hudson that the NLRB's conclusion that I-140 Leaders do not use independent judgment to assign tasks is both legally erroneous and unsupported by substantial evidence.
First, the NLRB's decision conflated "having authority," as required by the NLRA, with "exercising authority," which the statute does not require. See 29 U.S.C. § 152(11). As we explained in STP Nuclear, "the possession of authority to engage in [supervisory] functions-even if this authority has not yet been exercised-is what determines whether an individual is a supervisor." 975 F.3d at 514; see also Lakeland Health Care Assocs. v. NLRB, 696 F.3d 1332, 1343 n.11 (11th Cir. 2012) ("Whether an individual qualifies as a 'supervisor' under the [NLRA] does not necessarily rest on his or her employer's ability to provide actual examples of [exercising a listed] authority. Rather, . . . the [NLRA] directs the [NLRB] to evaluate the putative supervisor's authority [to act]." (emphases in original)). Put another way, "the actual exercise of the enumerated power is irrelevant so long as the authority to do so is present." Multimedia KSDK, Inc. v. NLRB, 303 F.3d 896, 899 (8th Cir. 2002).
To be sure, the NLRB has long held that "conclusory evidence is not sufficient to establish supervisory authority." I.H.S. Acquisition No. 114, Inc., 350 N.L.R.B. 489, 490 (2007). And evidence that individuals exercise authority certainly illustrates that such authority exists. But that does not mean that such evidence is required for a finding of supervisory status.
This error is particularly problematic here because Hudson only started using the project management software less than one month before the NLRB's hearing and, perhaps unsurprisingly, the record reflects that many I-140 Leaders were still becoming familiar with the new system. The record contains evidence of the I-140 Leaders assigning work before Hudson started using the project management software; the company simply did not provide specific "examples" of the I-140 Leaders "rejecting or overriding" the computer's matches in the one month since that system went into use. The record also shows that the I-140 Leaders still possessed the authority to assign work even after Hudson adopted the software. As the regional director explained, the I-140 Leaders still have "the ability to reassign the assignment or override the assignment generated by the system."
Because the I-140 Leaders possess authority to assign work, the next question is whether the I-140 Leaders use independent judgment in doing so. In concluding that the answer was "no," the NLRB found that reassigning work originally assigned by the PMS "merely reflects that the [I-140 Leaders] understand[] the complexity level of various documents and [are] aware of which employees have been trained to draft what types of documents," which does not amount to independent judgment. Hudson argues that this conflicts with an earlier agency adjudication, which explained that "the mere existence of company policies does not eliminate independent judgment from decision-making if the policies allow for discretionary choices," rather, a supervisor exercises independent judgment if that determination involves a "personal judgment based on personal experience, training, and ability." Oakwood Healthcare, Inc., 348 N.L.R.B. 686, 693 (2006); accord Entergy Miss., 810 F.3d at 296-97.
We agree with Hudson. Hudson's policies clearly allow for I-140 Leaders to make discretionary choices. The record shows that I-140 Leaders have authority to "change almost everything about a project," including the start date, the end date, the nature of the project, and the proposed writer. It is not simply that the I-140 Leaders understand the complexity of the project types and are aware of which employees have been trained on certain types of documents; the record shows that they use their own judgment, based on their personal experience, training, and ability, in assigning "who shall do the job." Oakwood, 348 N.L.R.B. at 693 (alteration adopted); see also STP Nuclear, 975 F.3d at 521 ("Assessing employees' skills . . . commonly supports a finding that assignments are made using independent judgment."). Indeed, the NLRB's own decision recognized that the I-140 Leaders grade each project a writer completes, evaluate the writers on a quarterly basis, and factor this information into how they assign projects.
We considered similar facts in STP Nuclear. There, maintenance supervisors assigned tasks to employees based on work set out in an "Authorized Work Schedule," which "comprehensively describe[d], in detail and hour-by-hour, the work to be performed by every crew on every shift." 975 F.3d at 519. The NLRB had concluded that "maintenance supervisors fulfill[ed] a purely ministerial role in implementing already-established assignments" and "their occasional delegation of tasks to specific employees based on experience or certifications [was] not an exercise of 'independent judgment.'" Id. at 520. But our court reversed, finding that the NLRB failed to consider certain tasks that maintenance supervisors performed that illustrated their authority to assign work, including their delegation of tasks to crew members based on the supervisor's assessment of crew members' skills and certifications. Id. at 521. We explained, "no evidence in the record support[ed] the [NLRB]'s claim that maintenance supervisors merely follow[ed] pre-planned procedures when delegating tasks." Id.
While the NLRB is correct that "[t]he PMS performs the further step of actually matching appropriate assignment/employee pairs," differentiating it from the system at issue in STP Nuclear, this argument does not account for the I-140 Leaders' authority to depart from the software's recommendations. And our prior precedent shows that this discretionary authority is enough to establish that the I-140 Leaders assign work using independent judgment. Accordingly, the NLRB's finding to the contrary is unsupported by substantial evidence.
2.
Regarding the I-140 Leaders' purported authority to reward and promote, Hudson's principal argument on appeal is that the NLRB ignored evidence that the I-140 Leaders' feedback to writers directly impacted those employees' compensation via Hudson's "Key Performance Index" (or KPI).
Hudson's KPI, which it uses to monitor employee performance, works as follows. For each project a writer completes, an I-140 Leader fills out a project feedback form that includes comments for the writer and a grade (exceptional, excellent, satisfactory, or unsatisfactory) that is converted into points and factored into the KPI. The form includes guidelines for determining the score of the project, which is based on the number of errors found and whether the error is major or minor. It also contains language indicating that I-140 Leaders still have discretion when completing the form. For instance, it states that "the guidelines . . . outline the most common mistakes and major concerns. However, be advised that the final distinction between 'major' or 'minor' error is still up to the [I-140 Leader's] discretion." A company witness explained at the hearing that these feedback forms could impact a writer's KPI by up to seventeen percent, and that KPI directly affects the quarterly bonus a writer receives, is considered when setting salaries, and may impact whether a writer gets promoted. The NLRB also concluded that higher Hudson management does not review the underlying work product to verify the accuracy of the grade.
The NLRB nonetheless concluded that: (1) Hudson had not established that "project feedback forms directly impact writers' salary, raises or promotions" because it remained unclear "whether there is something other than project feedback forms that are part of the KPI or if project feedback forms are the only documents contained in this index and what weight is given to the project feedback forms when making decisions regarding salary, raises and promotions;" and (2) the I-140 Leaders did not use independent judgment when filling out the project feedback forms because, essentially, they used a rubric. Hudson argues that these conclusions rest on "an unreasonable understanding of the record."
As to the NLRB's first conclusion, the agency argues on appeal that Hudson failed to establish a "direct relationship" or "direct correlation" between the evaluations and a writer's compensation. But, as Hudson explains, this argument ignores evidence that the difference between the highest evaluation and the lowest evaluation could make a difference of up to seventeen percent of a writer's KPI and that the KPI is directly linked to quarterly bonuses. Indeed, the NLRB recognized that "project feedback forms do impact the amount of the quarterly bonus a writer receives," but faulted Hudson for not providing more information about exactly how KPI is calculated. Even so, there was not substantial evidence supporting the NLRB's conclusion that I-140 Leaders do not effectively recommend rewards for writers given that Hudson showed a direct link between the project feedback forms, KPI, and quarterly bonuses and that Hudson management did not independently investigate the accuracy of the forms. See First Healthcare Corp., 323 N.L.R.B. 1171, 1171-72 (1997).
As to the finding that I-140 Leaders do not use independent judgment when completing the forms, the NLRB failed to analyze the parts of the rubric that call for the use of discretion. Not only is this inconsistent with the substantial evidence standard, but it is also inconsistent with the NLRB's Oakwood adjudication, which, again, explained that "the mere existence of company policies does not eliminate independent judgment from decisionmaking if the policies allow for discretionary choices." 348 N.L.R.B. at 693; see also Brusco Tug &Barge Co. v. NLRB, 247 F.3d 273, 278 (D.C. Cir. 2001) ("[A]n agency adjudication must either be consistent with prior adjudications or offer a reasoned basis for its departure from precedent."). The regional director failed to account for the I-140 Leaders' discretion to identify errors and determine whether they are major or minor. Nor did the regional director address that I-140 Leaders may count minor errors "more heavily" if they are repeated after feedback, without providing any standards for applying that proviso. As such, the NLRB's finding that I-140 Leaders do not reward or effectively recommend reward is not supported by substantial evidence.
B.
Hudson's RFE teams respond to USCIS requests for additional information and prepare supporting evidentiary documents such as letters of recommendation and proof of employment.
Unlike the I-140 Leaders, RFE team leads did not use a project management software when assigning work. Instead, the NLRB heard testimony from current and former RFE team leads that they assigned work based on case complexity, the turnaround time required, and a writer's ability, and that it was within the sole discretion of the team lead to assign work within his or her team.
Yet the NLRB found that RFE team leads were not supervisors because "assigning or reassigning work based on [an] employee's known skills and availability is not reflective of the exercise of independent authority." But as we explained above, this conclusion cannot be squared with our holding in STP Nuclear, which rejected the NLRB's conclusion that the "occasional delegation of tasks to specific employees based on experience or certifications is not an exercise of 'independent judgment.'" 975 F.3d at 520. Because the NLRB ignored evidence that the RFE team leads assessed the writers' skills and abilities when making assignments, its finding that RFE team leads are not supervisors under the NLRA is unsupported by substantial evidence.
C.
Hudson's I-485 teams fill out forms and collect documentation after an I-140 petition is approved. I-485 team leads "assign[] work to paralegals, check[] paralegals' work, answer[] [paralegals'] questions, and ensur[e] deadlines are met."
The record indicates that I-485 team leads engage in "discretionary assigning" of work using a spreadsheet containing the paralegals' names and the number of clients retained for the day. Hudson management explained at the hearing that new paralegals are assigned simpler tasks like scanning and mailing and then given more complex tasks, such as renewing employment authorization cards, as they gain experience. I-485 team leads have "discretion to stagger the workload" of the paralegals.
The NLRB found that the I-485 team leads do not use independent judgment when they assign work because the assignments are simply "based on employee availability and known skills." But, like the RFE team leads and I-140 Leaders, I-485 team leads are also responsible for evaluating and assessing their team members' skills. The NLRB's finding ignores this evidence, and overlooks the I-485 team leads' discretion to assign work to "ensur[e] deadlines are met." Accordingly, the finding that I-485 team leads do not assign work is not supported by substantial evidence.
V.
29 U.S.C. § 159(b) directs the NLRB to "decide . . . the unit appropriate for the purposes of collective bargaining." The selection of the appropriate bargaining unit "lies largely within the discretion of the [NLRB], whose decision, 'if not final, is rarely to be disturbed.'" S. Prairie Constr. Co. v. Loc. No. 627, Int'l Union of Operating Eng'rs, 425 U.S. 800, 805 (1976) (per curiam) (quoting Packard Motor Co. v. NLRB, 330 U.S. 485, 491 (1947)). Our review of the NLRB's determination of an appropriate bargaining unit is therefore "limited to determining whether the decision is arbitrary, capricious, an abuse of discretion, or lacking in evidentiary support." Electronic Data Sys. Corp. v. NLRB, 938 F.2d 570, 573 (5th Cir. 1991) (citation omitted). Such review is "exceedingly narrow" and the employer challenging the NLRB's determination must show that the "designated unit is clearly not appropriate." Macy's, Inc. v. NLRB, 824 F.3d 557, 563 (5th Cir. 2016) (citations omitted).
While the NLRA does not dictate how to determine the appropriate bargaining unit, the NLRB has developed its own guidelines. See id. at 56364. Where, as here, the union petitions for an employer-wide bargaining unit, the NLRB applies a presumption that the petitioned-for unit is appropriate. E.g., Greenhorne &O'Mara, Inc., 326 N.L.R.B. 514, 516 (1998). The party objecting to the unit can rebut this presumption by showing that "the interests of a given classification are so disparate from those of other employees that they cannot be represented in the same unit." Airco, Inc., 273 N.L.R.B. 348, 349 (1984). The NLRB then applies its "community of interest" test, which considers the "similarity in employees' skills, duties, and working conditions; centralized control of management and supervision; functional integration of business operations, including employee interchange; geographic proximity; bargaining history; and extent of union organization and employee choice." Exemplar, Inc., 363 N.L.R.B. 1500, 1501 (2016); Macy's, Inc., 824 F.3d at 570.
Hudson's argument that an employer-wide unit was not appropriate here focuses largely on the fact that the NLRB's decision only considered Hudson's offices in Ann Arbor, Pittsburgh, and Chicago, not the additional offices in Vancouver, Washington, Pasadena, California, and Durham, North Carolina that Hudson operated prior to the pandemic. But as the NLRB points out, Hudson's operations had remained the same for the eighteen months between the onset of the pandemic and the hearing and "Hudson makes no claim that it plans to re-open its many closed offices, end remote work, or otherwise reorganize its operations."
Hudson also contends that the NLRB failed to consider variation as to which teams are present at which office, office integration, and differences between employee benefits across states. But the decision shows that the NLRB did consider each of these factors and simply disagreed that they warranted separate bargaining units.
Hudson argues that the NLRB "improperly discounted" the fact that state law requires some variation in employee benefits, noting that employees have different health insurance providers and that there are differences in paid sick and medical leave. But, under this view, there could virtually never be a bargaining unit encompassing employees from different states. And regardless, there is no evidence that the NLRB reversibly erred in its conclusion that "benefits among . . . employees of different locations are quite similar."
VI.
While the NLRB's certification of an employer-wide unit was lawful, the bargaining unit improperly included supervisors who are not empowered with collective bargaining rights. Thus, Hudson did not violate the NLRA by refusing to bargain, and the unfair labor practice complaint fails. See Pittsburgh Plate Glass Co. v. NLRB, 313 U.S. 146, 152 (1941); STP Nuclear, 975 F.3d at 523. We therefore GRANT Hudson's petition for review, REVERSE the bargaining order, and DENY enforcement.