*127Almost four decades ago, in N.L.R.B. v. Yeshiva University , 444 U.S. 672, 100 S.Ct. 856, 63 L.Ed.2d 115 (1980), the Supreme Court sustained the National Labor Relation Board's extension of the protections of the National Labor Relations Act to the faculty of private universities. The Court concluded, however, that the full-time faculty seeking recognition in that case qualified as managerial employees exempt from the NLRA's coverage because, as at other " 'mature' private universit[ies]" where "authority ... is divided between a central administration and one or more collegial bodies," the faculty at Yeshiva exercised "effective[ ] control" over central university policies. Id. at 680, 683, 100 S.Ct. 856. In this case, the Board, applying standards it set forth in its post- Yeshiva decision, Pacific Lutheran University , 361 N.L.R.B. 1404 (2014), ruled that the full- and part-time non-tenure-track faculty of the University of Southern California's (USC's) Roski School of Art and Design exercised no effective control over university policies and, as non-managerial employees, were therefore eligible to join a union. USC petitions for review, arguing, among other things, that the Pacific Lutheran framework conflicts in several different ways with Yeshiva . Because we agree that one aspect of the Board's decision here-namely, its extension of Pacific Lutheran 's"majority status rule" to faculty subgroups-conflicts with Yeshiva , we grant the petition in part and deny the Board's cross-application for enforcement.

I.

Designed by Congress to quell "industrial strife" and its harmful effects on the "channels of commerce," the National Labor Relations Act aimed to stabilize industry by vesting industrial workforces with new labor rights. See Pub. L. No. 74-198, § 1, 49 Stat. 449, 449 (1935) (codified as amended at 29 U.S.C. § 151 et seq. ). Congress sought to "redress the perceived imbalance of economic power between labor and management ... by conferring certain affirmative rights on employees and by placing certain enumerated restrictions on the activities of employers." American Ship Building Co. v. N.L.R.B. , 380 U.S. 300, 316, 85 S.Ct. 955, 13 L.Ed.2d 855 (1965). "The central purpose of [the NLRA] was to protect employee self-organization and the process of collective bargaining from disruptive interferences by employers." Id. at 317, 85 S.Ct. 955.

A.

Years after Congress passed the NLRA, the Supreme Court issued two opinions central to the question before us in this case. First, in *128N.L.R.B. v. Bell Aerospace Co. Division of Textron Inc. , the Court held that although the NLRA, by its terms, covers all employees (except for supervisors and other exemptions immaterial to this case, see 29 U.S.C. § 152(3), (12) ), it nonetheless excludes "managerial" employees from its protections. 416 U.S. 267, 283-84, 94 S.Ct. 1757, 40 L.Ed.2d 134 (1974). As the Court explained, Congress "regarded [managers] as so clearly outside the [NLRA] that no specific exclusionary provision was thought necessary." Id. at 283, 94 S.Ct. 1757.

Second, in N.L.R.B. v. Yeshiva University , the Court clarified for the first time that the NLRA covers university employees and provided guidance about when university faculties constitute managerial employees exempt from the NLRA's coverage. See 444 U.S. at 682-90, 100 S.Ct. 856. In doing so, the Court distinguished between the "type of management-employee relations that prevail in the pyramidal hierarchies of private industry" and those that exist within a "typical 'mature' private university" where "authority ... is divided between a central administration and one or more collegial bodies" composed of academic faculty. Id. at 680, 100 S.Ct. 856. Together, administration and faculty may manage a university through a system of shared governance-" 'the process by which various constituents (traditionally governing boards, senior administration, and faculty ...) contribute to decision making related to college or university policy and procedure.' " American Council on Education's ("ACE") Br. 4 (alteration in original) (quoting Association of Governing Boards of Universities and Colleges, Shared Governance: Changing with the Times 3 (Mar. 2017)). It was respect for the "shared authority" that can be embedded in university governance, paired with the need to safeguard the protections afforded by the NLRA, that guided the Court in Yeshiva . See 444 U.S. at 680, 100 S.Ct. 856.

The facts of Yeshiva are straightforward. A union seeking to represent most of the university's full-time faculty filed a representation petition, requesting that the Board certify it as the faculty's bargaining agent. Id. at 674-75, 100 S.Ct. 856. The university argued that the faculty were managerial, but the Board found otherwise. Id. at 678, 100 S.Ct. 856. Grounding its decision in the NLRA's protection of professional employees-a designation encompassing employees engaged in "predominantly intellectual" work, 29 U.S.C. § 152(12) -the Board found that members of Yeshiva's faculty were professional employees who enjoy the NLRA's protections, rather than exempted managerial ones. Yeshiva , 444 U.S. at 678, 100 S.Ct. 856.

The Supreme Court disagreed. It pointed out that "professionals, like other employees, may be exempted from coverage ... under the judicially implied exclusion for 'managerial employees.' " Id. at 681-82, 100 S.Ct. 856. In non-university settings, moreover, the Board classifies as exempted managerial employees those individuals who "represent[ ] management interests by taking or recommending discretionary actions that effectively control or implement employer policy." Id. at 683, 100 S.Ct. 856. Therefore, the Court explained, the same general analysis must guide the Board in the university setting. Id. at 686, 100 S.Ct. 856. As relevant here, the Court dismissed as inconsistent with these precedents the Board's arguments that faculty members were non-managerial because they engaged in collective decision making (true of many corporate managers working through boards and committees) and because they enjoyed less than "ultimate" authority (held in corporate settings by the *129board of directors). Id. at 684-85, 685 nn. 20-21, 100 S.Ct. 856.

Instead, the Court looked to the principles underlying the managerial exception-namely, that managers fall outside the NLRA's protections because "an employer is entitled to the undivided loyalty of its representatives." Id. at 682, 100 S.Ct. 856. Unions divide that loyalty, and the fear of compromised loyalty was "particularly acute" at a university like Yeshiva, where the faculty exercised "pervasive[ ]"-even "absolute"-control over academic matters. Id. at 679, 686, 689, 100 S.Ct. 856. "The controlling consideration in this case," the Court explained, "is that the faculty of Yeshiva University exercise authority which in any other context unquestionably would be managerial." Id. at 686, 100 S.Ct. 856. Elaborating, the Court stated:

[The faculty's] authority in academic matters is absolute. They decide what courses will be offered, when they will be scheduled, and to whom they will be taught. They debate and determine teaching methods, grading policies, and matriculation standards. They effectively decide which students will be admitted, retained, and graduated. On occasion their views have determined the size of the student body, the tuition to be charged, and the location of a school. When one considers the function of a university, it is difficult to imagine decisions more managerial than these. To the extent the industrial analogy applies, the faculty determines within each school the product to be produced, the terms upon which it will be offered, and the customers who will be served.

Id.

With the Yeshiva faculty's "absolute" academic authority as a backdrop, the Court explained to the Board-in language important to the issue before us today-how to identify managerial faculty in future university cases. Faculty qualify as managerial when they exercise "effective recommendation or control" over central employer policies. Id. at 683 n.17, 100 S.Ct. 856. Faculty exercise such control when a university "depends on the [faculty's] professional judgment ... to formulate and apply crucial policies constrained only by necessarily general institutional goals," particularly when the university "requires faculty participation in governance." Id. at 689, 100 S.Ct. 856.

That said, the Court took care to avoid "sweep[ing] all professionals outside the [NLRA] in derogation of Congress' expressed intent to protect them." Id. at 690, 100 S.Ct. 856. Faculty are protected by the NLRA if their "decisionmaking is limited to the routine discharge of professional duties in projects to which they have been assigned." Id . Further cabining its holding, the Court recognized that faculties are heterogeneous, and that non-managerial subsets may exist within a faculty entrusted with managerial authority. For instance, the Board might draw a "rational line" between "tenured and untenured faculty members." Id. at 690 n.31, 100 S.Ct. 856. Such a distinction turns on "how a faculty is structured and operates." Id. In saying this, the Court made clear that "this is a starting point only, and that other factors not present here may enter into the analysis in other contexts." Id.

B.

For several decades following Yeshiva , the Board, when determining a given faculty's managerial status, applied a totality-of-the-circumstances approach informed by the contours of that case. See, e.g. , American International College , 282 N.L.R.B. 189, 190-202 (1986) (discussing at length whether the faculty at issue fell "within the scope of the Supreme Court's decision *130in [ Yeshiva ]"). But in LeMoyne-Owen College v. N.L.R.B. , 357 F.3d 55 (D.C. Cir. 2004), we found the Board's approach inadequate. There the Board had classified the full-time faculty of a small private institution in Memphis as non-managerial. Id. at 55-56, 60. Although the Board had relied on the fact that circumstances at LeMoyne-Owen were similar to those at other colleges where it had classified faculty as non-managerial, we were concerned that the Board failed to "discuss or even mention a single one of the precedents on which the College relied." Id. at 60. Ending our analysis before reaching the managerial status of the college's faculty, we observed that, although the Board's totality-of-the-circumstances approach followed naturally from Yeshiva 's"long list of relevant factors" paired with "the exquisite variety of academic institutions across the country," the Board must provide an "adequate explanation for the result it reached in this case." Id. at 57, 61. We instructed the Board to delineate "which factors are significant and which less so, and why" in managerial-faculty determinations. Id. at 61. Doing so, we explained, would afford faculty, universities, and reviewing courts "predictability and intelligibility" concerning the Board's "[application of] the test to varied fact situations." Id. ; see also Point Park University v. N.L.R.B. , 457 F.3d 42, 50 (D.C. Cir. 2006) (reiterating this requirement in another faculty case).

The Board endeavored to satisfy LeMoyne-Owen in Pacific Lutheran University . Unlike Yeshiva , which involved full-time faculty, Pacific Lutheran concerned a faculty subgroup-specifically, full-time "contingent" (non-tenure eligible) faculty. 361 N.L.R.B. at 1417. Drawing on its "30-plus years applying Yeshiva " and echoing the Supreme Court's language, the Board used Pacific Lutheran as a platform to launch a new test designed "to answer the question whether faculty in a university setting actually or effectively exercise control over decision making pertaining to central policies of the university such that they are aligned with management." Id .

Under the Pacific Lutheran framework for determining whether faculty qualify as managerial, the Board "organize[s] [its] review of faculty decision-making into five general areas": academic programs, enrollment management policies, finances, academic policies, and personnel policies and decisions. Id . The Board classifies the first three as "primary" and the last two as "secondary." Id . at 1420. Although faculty participation in these areas takes various forms, under the Pacific Lutheran framework the Board trains its focus on participation through service on university committees, which specialize in a given field and then propose policies to the full faculty or to the university's administration. See id. at 1424-28 (consistently analyzing faculty participation by looking to faculty committees).

Again echoing Yeshiva , the Board then considers whether the faculty exercise "actual control or effective recommendation" authority over each of these five areas. Id. at 1421. The test for "effective" control is demanding: "recommendations must almost always be followed by the administration. Further, faculty recommendations are 'effective' if they routinely become operative without independent review by the administration." Id . (internal citation omitted). Intent on ensuring that faculty exercise "actual-rather than mere paper-authority," Pacific Lutheran requires "specific evidence" of the process by which the administration adopts faculty recommendations. Id. Under Pacific Lutheran , the Board looks at "both the structure of university decisionmaking and where the faculty at issue fit within that structure, including the nature of the employment relationship." Id. at 1421-22.

*131Relevant distinctions include "tenured vs. tenure eligible vs. nontenure eligible" and "regular vs. contingent" faculty members. Id. at 1422. Rather than requiring that managerial faculty control some set number of decision-making areas, Pacific Lutheran calls for a holistic review of the control exerted over the five areas. See id. at 1423 (noting that to determine managerial status the Board will assess faculty's control "over those areas").

Beyond these context-specific inquiries, and central to this case, Pacific Lutheran sets a bright-line "majority status rule" under which a committee's actual control or effective recommendation authority over a particular decision-making area may be ascribed to faculty only if they constitute a majority of that committee: "[i]f faculty members do not exert majority control, we will not attribute the committee's conduct to the faculty." Id. at 1421 n.36.

Applying this framework to the situation at Pacific Lutheran, the Board determined that the contingent faculty were non-managerial. Although the university had recently allowed contingent faculty to sit on committees, no such faculty members had yet joined a committee and, in any event, it appeared that they would have had no "right to vote within their respective divisions, schools, or constituent departments." Id. at 1424, 1427-28. Contingent faculty did participate in and have a vote at the faculty assembly, but the Board characterized the assembly as "little more than a conduit to transmit previously agreed-upon recommendations to the administration" and observed that contingent faculty comprised only about twenty percent of the faculty assembly's average attendees. Id. at 1428. According to the Board, the nature of the contingent faculty members' employment relationship with the university also counseled against a finding of managerial status. In contrast to Yeshiva , where the university "require[d] faculty participation in governance," 444 U.S. at 689, 100 S.Ct. 856, the Pacific Lutheran administration rarely discussed with contingent faculty its expectations for their university service, such as committee participation, see Pacific Lutheran , 361 N.L.R.B. at 1423. Contingent faculty had to renew their contracts annually and received minimal institutional support (such as funding for professional development or research). Id.

In considering whether faculty held a majority of committee seats, the Board left uncertain whether the majority status rule applied to the contingent faculty only or to the faculty as a whole. Even if contingent faculty members sat on committees, the Board observed, "they would be a minority on the ... committee[s] as their membership is currently structured." Id. at 1428. Because the faculty as a whole constituted a minority on all of these committees, id. at 1424, however, the significance of this observation was unclear.

Two Board members-Phillip Miscimarra and Harry Johnson-separately dissented in part. They "generally agree[d]" with the Board's "admirable effort" to craft distinct, workable decision-making areas. Id. at 1430 (Miscimarra, Member, concurring in part and dissenting in part), 1441 (Johnson, Member, dissenting). Uncertain about how the Board would weigh the "primary" and "secondary" areas in future cases, however, the dissenters believed that "substantial" authority in secondary areas-unaccompanied by effective control over any of the three primary areas-may sometimes be sufficient to trigger managerial status. Id. at 1429-30 (Miscimarra, Member, concurring in part and dissenting in part); see also id. at 1441-42 (Johnson, Member, dissenting).

*132The dissenters focused primarily on Pacific Lutheran 's"treatment of authority, control, and effective recommendation" authority, which they considered "too onerous and inflexible." Id. at 1430 (Miscimarra, Member, concurring in part and dissenting in part). According to the dissenters, the Pacific Lutheran framework demands too much because it requires that the administration "almost always" follow faculty recommendations. Id. at 1430 (Miscimarra, Member, concurring in part and dissenting in part), 1443 (Johnson, Member, dissenting). "[B]y failing to consider the actual, diverse processes of university business operations and governance," Johnson argued, Pacific Lutheran "raised the bar for establishing managerial status of faculty to an unattainable height." Id. at 1442-43. He also believed that the Board had fashioned "a false dichotomy" when it held that committees exercise effective control only if administrators routinely accept the committees' recommendations and do so without independent review. Id. at 1443-44. Such a rule, he contended, flouts the ideal of universities as "places rich in dialogue" between the administration and faculty. Id. at 1444.

C.

The University of Southern California, a large private university in Los Angeles, has twenty-two schools, some 6,600 faculty, and between 30,000 and 40,000 students. One of these schools, the Gayle Garner Roski School of Art and Design ("Roski")-the venue for the dispute before us-offers undergraduate and graduate degrees in fine arts and critical studies, among others. See University of Southern California ("USC "), 365 N.L.R.B. No. 11, at *7 (Dec. 30, 2016). In 2015, the Service Employees International Union, Local 721 ("Union") petitioned to represent Roski's full- and part-time non-tenure-track faculty. Id. at *5.

Reprising the battle lines drawn in Yeshiva , USC contended that the Roski non-tenure-track faculty were managerial, and the Union disagreed. In a decision adopted by the Board without opinion, the Regional Director, applying the Pacific Lutheran framework, found the Roski non-tenure-track faculty non-managerial. See id. at *1, *18.

The Regional Director began by addressing several issues left unresolved in Pacific Lutheran . She first confirmed what the Board had hinted in that case: to exercise effective control over a committee, the faculty subgroup seeking recognition-not just the faculty as a whole-must hold a majority of committee seats. "[E]ven if the faculty on [either committee] could be said to actually or effectively control decisionmaking ..., I would not attribute that control to the nontenure track faculty at issue here, as they do not constitute a majority of either committee." Id. at *16. Second, although Pacific Lutheran left ambiguous whether, for a committee to exercise effective control, the administration must "almost always" accept a committee's recommendations and those recommendations must "routinely become operative without independent review," Pacific Lutheran , 361 N.L.R.B. at 1421, or whether just one or the other would do, the Regional Director indisputably required both: "The Board also clarified that for faculty recommendations to be 'effective,' the administration must 'almost always' adopt the recommendations, and do so 'routinely' without independent review." USC , 365 N.L.R.B. at *15. Third, the Regional Director determined that exercising effective control over one secondary area "alone" is insufficient to trigger faculty managerial status. Id. at *17. By affirming the Regional Director's decision, the Board *133has, in effect, adopted these interpretations as its own. See id. at *1 n.1 ("The Regional Director's decision properly applied [ Pacific Lutheran ].").

Applying this interpretation of Pacific Lutheran , the Regional Director classified the Roski non-tenure-track faculty as non-managerial. For purposes of determining effective control, the Regional Director centered her discussion around a dozen or so university-wide committees and other USC faculty bodies, including the Roski school-level faculty council, that USC contends provide faculty with opportunities to weigh in on university governance. Id. at *15-18. The committees conduct studies, write reports, and make recommendations to either the Academic Senate or the provost. Id. at *8. The University Committee on Curriculum, for instance, approves, rejects, or modifies every proposed USC course, program, major, minor, and degree offering. Id. at *9, *15. Although the Committee on Finance and Enrollment has made only a few recommendations, those recommendations have involved significant issues, including the creation of a university-wide plan to address graduate student enrollment, whether to maintain a holistic review process for undergraduate student admissions, and the payout rate of USC's endowment. Id. at *10, *16-17. Other committees make recommendations on academic programs, teaching guidelines, software upgrades, grading policies, appointments, promotions, and tenure decisions. Id. at *8-13.

Most USC university-wide and school-specific committees are open to all faculty members, including nontenure track, though non-tenure-track faculty are excluded from decisions concerning tenure. Id. at *11. Committees are almost 100 percent faculty, although some faculty also have administrative appointments. Administrators occasionally sit "ex officio " on committees where they have no vote. See id. at *8-13.

Looking at all committees relevant to each of the five decision-making areas, the Regional Director determined that USC had failed to show that the faculty serving on these committees exercised effective control over any decisionmaking area. Id. at *18 ; see also Pacific Lutheran , 361 N.L.R.B. at 1427 n.65 (noting that the party asserting managerial status assumes the burden of proof). In support, she found that USC had failed to provide "specific evidence," as required by Pacific Lutheran , 361 N.L.R.B. at 1421, of the type of review the administration conducted of the committees' recommendations before accepting them. See USC , 365 N.L.R.B. at *15-18. She also found that the testimony concerning the work of some committees was vague, see id. at *15, 17-18, as was the testimony about the decision-making authority of other committees, see id. For still other committees, she found that they had offered too few recommendations to characterize their recommendations as "routinely" followed. See id. at *16.

In further support of her conclusion that the Roski nontenure-track faculty exercised no effective control, the Regional Director found that the nature of their "tenuous employment" relationship "limit[s]" their role at the university. Id. at *18. Though non-tenure-track faculty are eligible to join any committee and make up about three-fourths of the USC faculty, they represent a minority on almost every committee. Id. at *7-9. They receive little to no institutional support in the form of funding, evaluations, or other career guidance; "[i]n fact, [non-tenure-track faculty witnesses] testified that administrators in their departments or schools have never met with them to discuss expectations about their teaching, their scholarship or *134artistic work, or their service to the University." Id. at *8.

Following her discussion of most every committee (including all committees in the three primary areas), the Regional Director also found that, even if the relevant committees exercised effective control, "nontenure track faculty do not constitute a majority of [those] committee[s]." Id. at *17. Therefore, "they cannot be found to possess any managerial control." Id.

Upon finding both that no committee exercised effective control and that the Roski non-tenure-track faculty held a minority of seats on the key committees, the Regional Director concluded that USC had "failed to establish that the full-time and/or part-time nontenure track faculty at ... the Roski School actually or effectively exercise control over decision making pertaining to central policies of the university." Id. at *18. As noted, on a request for review, the Board adopted the Regional Director's opinion in full. Id. at *1 & n.1.

Member Miscimarra, one of the Pacific Lutheran dissenters, again disagreed. Asserting that the Board holds managers in other industries to a lower standard, Miscimarra faulted the Board for ascribing managerial status to only those faculty with "unreviewable authority." Id. at *2 (Miscimarra, Member, dissenting). He also believed that the Court in Yeshiva had rejected the idea that possessing a majority of committee seats was a prerequisite for exercising effective control over a committee. He put it this way:

[ Yeshiva ] held that a faculty member may possess managerial authority even though he or she cannot individually establish policy separate from the committees on which he or she serves. Similarly, faculty members in an individual department or program may be managerial, even if as a group they are a minority of the total faculty and are outnumbered and outvoted on every issue.

Id. at *3-4. In Miscimarra's view, extension of the majority status rule to determine the managerial status of a faculty subgroup conflicts with "the principle of collegial managerial authority that the Supreme Court recognized in Yeshiva ." Id. at *3.

Following the Regional Director's decision and the Board's adoption of it, the Roski non-tenure-track faculty voted to form a bargaining unit represented by the Union. When USC refused to bargain-an unfair labor practice under the NLRA, see 29 U.S.C. § 158(a)(5) -the Union filed a complaint with the Board. USC acknowledged its refusal to bargain, and challenged the certification on the grounds that Pacific Lutheran conflicts with Yeshiva and that the Roski nontenure-track faculty in fact exercise managerial authority. The Board granted summary judgment to the Union. See University of Southern California , 365 N.L.R.B. No. 89, at *1 (June 7, 2017).

In its petition for review, USC makes three basic arguments. First, it argues that several elements of the Pacific Lutheran framework, as applied in this case, conflict with Yeshiva : its requirement that a faculty subgroup must hold a majority of committee seats in order to exercise effective control through a committee, its standard for "effective" control, and its classification of the five decision-making areas as primary or secondary. Whether the Pacific Lutheran framework comports with Yeshiva is an issue of first impression given that the university in Pacific Lutheran never petitioned for review and that no other circuit has yet considered the question. Second, USC argues that, contrary to LeMoyne-Owen , the Pacific Lutheran framework fails to provide a workable standard to determine the managerial status of faculty.

*135And third, even assuming the validity of the Pacific Lutheran framework, USC argues that the Regional Director lacked substantial evidence to classify the Roski non-tenure-track faculty as non-managerial.

In considering USC's challenges, we defer to the Board's judgment, born of its expertise, so long as its decision is "consistent with controlling precedent," based on "substantial evidence," and neither "arbitrary" nor "capricious." International Union of Operating Engineers, Local 147 v. N.L.R.B. , 294 F.3d 186, 188 (D.C. Cir. 2002) ; see also 5 U.S.C. § 706(2)(A) ; N.L.R.B. v. Curtin Matheson Scientific, Inc. , 494 U.S. 775, 786, 110 S.Ct. 1542, 108 L.Ed.2d 801 (1990) (explaining that because "the NLRB has the primary responsibility for developing and applying national labor policy," courts will uphold any rule that is "rational and consistent with the [NLRA], even if we would have formulated a different rule had we sat on the Board" (internal citation omitted)).

II.

We begin with USC's argument that the Pacific Lutheran framework-and consequently, the decision at issue here-is "[in]consistent with controlling precedent," i.e., that it conflicts with the Supreme Court's decision in Yeshiva . But as even its dissenters acknowledge, Pacific Lutheran represents an "admirable effort" by the Board to tame a thicket of case law that touches on numerous interrelated features of the faculty experience at universities. Pacific Lutheran , 361 N.L.R.B. at 1441 (Johnson, Member, dissenting). And for the most part, that effort succeeds. There is, however, a major problem.

Under Pacific Lutheran , as interpreted by the Board in this case, a faculty subgroup seeking recognition exercises effective control over a decision-making area through its participation on a committee only when that subgroup constitutes a majority of the committee. See Pacific Lutheran , 361 N.L.R.B. at 1421 n.36 ; USC , 365 N.L.R.B. at *16. This rule, which we shall refer to as the "subgroup majority status rule," is critical because, as discussed above, see supra at 130, the Board generally looks to a faculty's authority at the committee level when determining whether the faculty exercise effective control over Pacific Lutheran 's five decision-making areas.

USC first claims that the subgroup majority status rule resurrects the theory, foreclosed by Yeshiva , that faculty can attain managerial status based only on authority wielded individually, as opposed to authority shared collectively with others. See Yeshiva , 444 U.S. at 685 & n.21, 100 S.Ct. 856 (criticizing the theory as "flatly inconsistent" with Board precedents that credit managerial authority to "supervisors who work through committees"). But in neither Pacific Lutheran nor here did the Board run afoul of that precise holding. Indeed, the Board agrees that faculty can exercise managerial authority collectively. See Respondent's Br. 37 ("Faculty who constitute the majority of a committee can be managerial based on their service on that committee, even though the authority they exercise there is collective."). The extension of Pacific Lutheran 's majority status rule to faculty subgroups instead reflects the Board's belief that a subgroup, acting collectively, is unable to exercise managerial authority through a committee when that subgroup holds a minority of committee seats.

Assuming that understanding of the Board's subgroup majority status rule, USC next argues that the rule "simply disregard[s]" the contributions that minority subgroups can make to university governance. Petitioner's Br. 34. Under the *136majority status theory, USC continues, even when a university draws "no distinction" between faculty subgroups "in their roles in faculty governance," including situations where the subgroups "all participate in the effective management of the University," the Board will "arbitrarily say that one category must be in the majority[,] ... necessarily ... render[ing] the other category non-managerial-when in fact, all are managerial." Id. at 38 (emphases omitted). Because a subgroup's "ability to invoke the rights or protections of the [NLRA] should not hinge on whether some other group of faculty control or effectively recommend university policy," the Board's response goes, it would be inappropriate to deny a subgroup the NLRA's protections when an entirely different group of faculty, concerned about advancing its own priorities rather than the subgroup's, runs the show. Respondent's Br. 35 (emphasis in original).

In our view, the Board's subgroup majority status rule rests on a fundamental misunderstanding of Yeshiva . To explain why, we return to the Court's opinion and specifically to three key themes.

First, as noted above, the Court drew a sharp distinction between the "pyramidal" hierarchies characteristic of private industry and the faculty "bodies" at universities. Yeshiva , 444 U.S. at 680, 100 S.Ct. 856. "[T]he authority structure of a university does not," the Court explained, "fit neatly within the statutory scheme we are asked to interpret," and for that reason, "the Board has recognized that principles developed for use in the industrial setting cannot be 'imposed blindly on the academic world.' " Id. at 680-81, 100 S.Ct. 856 (quoting Syracuse University , 204 N.L.R.B. 641, 643 (1973) ). In the "pyramidal" industrial arena, the Board looks to whether individuals, or groups of individuals, exercise managerial authority. Id. at 680, 100 S.Ct. 856. By contrast, the Court in Yeshiva returned time and again to the question of whether the "faculty" as a body-as opposed to individual professors or subgroups-exercises managerial authority. To the Court, the "controlling consideration" was that the "faculty at Yeshiva University exercise authority which in any other context unquestionably would be managerial." Id. at 686, 100 S.Ct. 856 (emphasis added). "The 'business' of a university is education," the Court explained, "and its vitality ultimately must depend on academic policies that largely are formulated and generally are implemented by faculty governance decisions." Id. at 688, 100 S.Ct. 856 (emphasis added). Continuing its focus on the "faculty" as a body, the Court observed that "[t]he university requires faculty participation in governance"; that "Yeshiva and like universities must rely on their faculties to participate in the making and implementation of their policies"; and that "[t]he faculty participate in University-wide governance through their representatives on an elected student-faculty advisory council." Id. at 675-76, 689, 100 S.Ct. 856 (emphasis added).

These repeated references to the "faculty" as a body are not linguistic accidents; they are central to the Court's reasoning. Take the Court's discussion of the bedrock principle underlying the managerial exception: that employers deserve the loyalty of employees who exercise discretionary authority over central employer policies. Id. at 687-88, 100 S.Ct. 856. Highlighting that the faculty functions as a single body, the Court observed that if, consistent with principles of shared governance, the university "depends on the professional judgment of its faculty to formulate and apply crucial policies," then the university deserves those employees' "undivided loyalty," which in turn triggers *137managerial status. Id. at 682, 689, 100 S.Ct. 856 (emphasis added). The Court's analysis turned not on an aggregation of the power delegated to a series of individuals or a mosaic of subgroups-the focus of the Board's subgroup majority status rule-but rather on the role played by the faculty as a body.

Reinforcing this idea-and this is Yeshiva 's second theme-the Court repeatedly stressed the importance of collegiality. "[A]uthority in the typical 'mature' private university," the Court explained, is split between the administration and "one or more collegial bodies." Id. at 680, 100 S.Ct. 856. The Court observed that "[t]he Board itself has noted that the concept of collegiality 'does not square with the traditional authority structures with which th[e] [NLRA] was designed to cope,' " and that "traditions of collegiality continue to play a significant role at many universities." Id. (second alteration in original) (quoting Adelphi University , 195 N.L.R.B. 639, 648 (1972) ).

The Board's subgroup majority status rule is unfaithful to this critical aspect of Yeshiva . It ignores the possibility that faculty subgroups, despite holding different status within the university, may share common interests and therefore effectively participate together as a body on some or all of the issues relevant to managerial status. It is entirely plausible that, for example, non-tenure-track faculty, especially full-time nontenure-track faculty, would agree with tenure-track faculty on questions of course offerings, academic integrity, and grading policies. Yet the Board's subgroup majority status rule presupposes that non-tenure-track faculty have no authority over such matters unless they constitute a majority of the relevant committee. Of course, the interests of some minority faculty subgroups may differ from the majority on certain issues, such as with respect to promotion and salary. That, however, is precisely where the concept of collegiality comes into play. Disregarding this fundamental characteristic of university governance, the Board's subgroup majority status rule assumes that minority subgroups can never work out their differences with the majority.

Taken together, these two themes of Yeshiva -a focus on the faculty as a body and an emphasis on collegiality-demonstrate that the question the Board must ask is not whether a particular subgroup can force policies through based on crude headcounts, but rather whether that subgroup is structurally included within a collegial faculty body to which the university has delegated managerial authority. This reading finds support in Yeshiva 's third theme: the Court's guidance on how the Board should approach situations where the faculty members seeking recognition constitute a subgroup, such as in this case (non-tenure-track) and in Pacific Lutheran (contingent). Employees, the Court explained, may not be considered management and thus excluded from the NLRA's coverage if their "decisionmaking is limited to the routine discharge of professional duties in projects to which they have been assigned." Yeshiva , 444 U.S. at 690, 100 S.Ct. 856. Indeed, the Court noted that there may have been "faculty members at Yeshiva" itself "who properly could be included in a bargaining unit." Id. at 690 n.31, 100 S.Ct. 856. For example, "a rational line could be drawn between tenured and untenured faculty members, depending upon how a faculty is structured and operates." Id . Put another way, the question the Board must ask is not a numerical one-does the subgroup seeking recognition comprise a majority of a committee-but rather a broader, structural one: has the university included the subgroup in a faculty body vested with managerial responsibilities? To answer this question, the *138Board must, as required by Yeshiva , examine how the faculty is "structured" and "operates," as well as the duties employees have been "assigned." For example, as in Yeshiva , does the university "require" faculty members to participate in committees? If this analysis demonstrates that the subgroup seeking recognition is structurally part of a faculty body to which the university has delegated managerial authority, then, apart from several exceptions we shall discuss below, that ends the matter: the subgroup is managerial regardless of whether its members constitute a majority of the relevant committees or whether they even participate at all.

Two additional considerations support this understanding of Yeshiva .

First, the Court relied on a line of Board decisions from the private sector holding that it is possible for minority employee shareholders "who owned enough stock to give them, as a group, a substantial voice in the employer's affairs" to exercise effective control, even absent majority control. Yeshiva , 444 U.S. at 685 n.21, 100 S.Ct. 856. Indeed, the Board itself has recognized that it "has excluded groups of shareholder-employees from bargaining units even though they collectively owned less than half of the employers' stock." Upper Great Lakes Pilots, Inc. , 311 N.L.R.B. 131, 132 n.9 (1993). Although the Board cites these cases in its brief while discussing this very issue, it fails to acknowledge their inconsistency with its subgroup majority status rule.

Second, and quite apart from Yeshiva , we just cannot see how the subgroup majority status rule will work. As USC notes, "the make-up of a particular committee will necessarily change from year to year. If only the majority category is managerial, then from one year to the next, different categories of perfectly managerial faculty could be deemed non-managerial for no other reason than a one- or two-person swing on a committee roster." Petitioner's Br. 38. In his dissent, Member Miscimarra pointed out that, under the subgroup majority status rule, slicing and dicing the faculty in different ways-by tenure status, school, seniority, etc.-may allow "even faculty who indisputably exercise managerial authority" to qualify as non-managerial. USC , 365 N.L.R.B. at *4 & n.7 (Miscimarra, Member, dissenting). In its amicus brief, the American Council on Education likewise warns that "any group of faculty can be sub-divided until it no longer commands a majority." ACE Br. 27. The subgroup majority status rule contributes to this strategic division of faculties; Yeshiva 's structural approach avoids it.

That said, whether an individual subgroup possesses a majority of committee seats is not always irrelevant. Quite to the contrary, determining whether a subgroup holds a decisive bloc of committee seats may be necessary where a subgroup's interests fundamentally diverge from those of the majority. Although the Court in Yeshiva emphasized the value of faculty collegiality, there may well be issues on which the interests of the subgroup and the faculty as a whole differ so significantly that they cannot be reconciled even through collegial compromise. Under such circumstances, which the Board can identify only through a careful analysis of how the faculty seeking recognition actually functions, the Board might appropriately conclude that the subgroup cannot exercise effective control unless it constitutes a majority of the relevant committees. Likewise, if a subgroup that the university expects to participate in a committee nonetheless fails to do so, this may signal the presence of structural barriers to that group's participation. In this case, for example, the Board may well be correct that there is something about the status of non-tenure-track *139faculty, especially part-time non-tenure-track, that effectively silences any managerial "voice," Pacific Lutheran , 361 N.L.R.B. at 1423, that such faculty otherwise might possess. See Yeshiva , 444 U.S. at 690 n.31, 100 S.Ct. 856 (acknowledging that "other factors not present here," such as tenure status, "may enter into the analysis in other contexts").

To be sure, the Pacific Lutheran framework accounts for these factors (diverging interests and the nature of non-tenure-track employment). See 361 N.L.R.B. at 1421-22 (looking at the "nature of the employment relationship," including "tenured vs. tenure eligible vs. nontenure eligible"). But Pacific Lutheran , as interpreted by the Board in this case, runs afoul of Yeshiva by using these factors as part of a determination focused on whether the petitioning subgroup alone exercises effective control. The Board should instead, as required by Yeshiva , think of this analysis as having two distinct inquiries: whether a faculty body exercises effective control and, if so, whether, based on the faculty's structure and operations, the petitioning subgroup is included in that managerial faculty body. Only as part of the latter analysis should the Board dig into whether a subgroup's actual interests diverge so substantially from those championed by the rest of the faculty that holding a minority of seats on the relevant committees is akin to having no managerial role at all, or whether a subgroup's low participation rates stem from a tenuous employment relationship that vitiates any managerial role the university expects the subgroup to perform. See Yeshiva , 444 U.S. at 674-77, 686, 100 S.Ct. 856 (examining closely the specific operations at Yeshiva).

Of course, the Board has flexibility over how to organize its analysis in any given case. In some situations, it might make sense to address these two inquiries sequentially: is there a managerial faculty body and, if so, is the petitioning subgroup a part of it? But in other situations, it may be unnecessary for the Board to consider whether a managerial faculty body exists because, even assuming one did, the petitioning subgroup is so clearly not included in it-because, for example, university rules prohibit its participation in committees. The critical point is this: however the Board proceeds, it must treat these two inquiries separately and may not conflate them by asking whether the petitioning subgroup alone exercises effective control.

A final observation: in Pacific Lutheran , the Board emphasized that since the Court decided Yeshiva some four decades ago, universities "are increasingly run by administrators" and rely more and more on non-tenure-track faculty "who, unlike traditional faculty, have been appointed with no prospect of tenure and often no guarantee of employment." Pacific Lutheran , 361 N.L.R.B. at 1422. According to the Board, these trends "ha[ve] the effect of concentrating and centering authority away from the faculty." Id. Building on this point, amicus American Association of University Professors points out that "[r]ather than relying on faculty expertise and recommendations, the growing ranks of administrators increasingly make unilateral decisions on university policies and programs, often influenced by considerations of external market forces and revenue generation." American Association of University Professors' Br. 10. By contrast, the American Council on Education, though acknowledging these trends, emphasizes "the continued primacy of shared governance." ACE Br. 13. This is an interesting debate, and it may even be relevant. Regardless of national trends, however, the Board must not lose sight of the fact that the question before it in any case in which a faculty subgroup seeks recognition *140is whether that university has delegated managerial authority to a faculty body and, if so, whether the petitioning faculty subgroup is a part of that body. As we explained in Point Park , this requires "an exacting analysis of the particular institution and faculty at issue." 457 F.3d at 48 (emphasis added).

Having thus concluded that the Board's subgroup majority status rule runs afoul of Yeshiva , we turn to USC's challenge to Pacific Lutheran 's standard for "effective" control.

III.

Recall that Pacific Lutheran announced the following standard for determining when faculty exercise effective control over decision making: "[T]o be 'effective,' recommendations must almost always be followed by the administration. Further, faculty recommendations are 'effective' if they routinely become operative without independent review by the administration." 361 N.L.R.B. at 1421 (internal citations and footnote omitted). Filling a gap left by the Board in Pacific Lutheran , the Regional Director required that both elements-"almost always" followed and "routinely" adopted "without independent review"-must be satisfied to support an effective control finding. USC , 365 N.L.R.B. at *15. USC does not challenge the Regional Director's conjunctive interpretation. Instead, drawing from the Pacific Lutheran dissents, the university argues that "there is no ... logical conclusion" other than that the Pacific Lutheran standard, as interpreted in USC , resurrects the "ultimate authority" requirement for managerial status that Yeshiva forbids. Petitioner's Br. 33 (emphasis omitted); see also Yeshiva , 444 U.S. at 685 & nn.20-21, 100 S.Ct. 856.

In Yeshiva , the Court held that, in order to qualify as managerial, faculty must exercise "effective recommendation or control" over central university policies. Yeshiva , 444 U.S. at 683 n.17, 100 S.Ct. 856. The Yeshiva faculty met that standard, the Court found, because their control was "substantial[ ] and pervasive[ ]": among other things, the faculty "decide[d] [what] students will be admitted, retained, and graduated," "decide[d] what courses will be offered," and "determine[d] teaching methods, grading policies, and matriculation standards." Id. at 679, 686, 100 S.Ct. 856. The Court also explained that faculty need not have "final authority" in order to enjoy managerial status; a "rarely exercised veto power" held by the administration "does not diminish the faculty's effective power." Id at 683 n.17, 685, 100 S.Ct. 856.

We see nothing in the Board's two-part standard for "effective" control that runs afoul of Yeshiva . Although USC argues otherwise, the Board's standard does not require "ultimate authority." Quite to the contrary, its use of the phrase "almost always" allows for occasional, or as the Court put it, "rarely exercised," vetoes. The word "routine" likewise leaves room for some administrative review. True, the standard is demanding, but it comports with Yeshiva , and we agree with the Board that setting a high bar for effective control is necessary to avoid interpreting the managerial exception so broadly that it chips away at the NLRA's protections. See Pacific Lutheran , 361 N.L.R.B. at 1419 n.32 ("[W]e are mindful of the fundamental principle that 'exemptions from [the NLRA's] coverage are not so expansively interpreted as to deny protection to workers the [NLRA] was designed to reach.' " (quoting Holly Farms Corp. v. N.L.R.B. , 517 U.S. 392, 399, 116 S.Ct. 1396, 134 L.Ed.2d 593 (1996) )). That said, the Board must be careful to apply its effective control standard with sensitivity to the notion *141of "collegial managerial authority that the Supreme Court recognized in Yeshiva ." USC , 365 N.L.R.B. at *3 (Miscimarra, Member, dissenting). "[T]raditions of collegiality continue to play a significant role at many universities," Yeshiva , 444 U.S. at 680, 100 S.Ct. 856, and the Board's test for effective control must accommodate these traditions.

USC complains about a stray sentence in the Regional Director's opinion: "I am not convinced by the conclusory evidence in the record that the Board of Trustees, for example, would sign off without second thought on a tuition amount or endowment payout based solely on the recommendation of a newly-formed faculty committee that had never before considered such issues." USC , 365 N.L.R.B. at *17 (emphasis added). According to USC, requiring that the administration accept committee recommendations "without second thought" sets an even higher bar for effective control than the one we understand the Board to have established. Petitioner's Br. 22. Were this characterization of the Regional Director's decision accurate, we would share USC's concerns. The Regional Director, however, never suggested that her "without second thought" statement amounted to a new requirement for establishing effective control. Instead, she made the statement in response to USC's contention that the administration had in fact accepted the newly formed Committee on Finance and Enrollment's recommendations on important financial matters "without second thought."

IV.

We can quickly dispose of USC's remaining arguments.

First, regarding Pacific Lutheran 's classification of the five decision-making areas as primary and secondary, USC acknowledges that Yeshiva affords the Board leeway to design these categories, but nonetheless argues that the Pacific Lutheran framework inappropriately relegates academic and personnel policies to secondary status while elevating finances to primary status. Justifying the classification, the Board explained that the primary areas, more so than the secondary ones, touch the university as a whole, align faculty with management, and affect the "product" the university offers-all considerations taken from the Court's discussion in Yeshiva . Pacific Lutheran , 361 N.L.R.B. at 1420 ; see also Yeshiva , 444 U.S. at 675-77, 686, 690, 100 S.Ct. 856. Although we can imagine different groupings of these factors, nothing in Yeshiva dictates the outcome one way or the other, and the Board's categorization falls well within its discretion under the NLRA. Indeed, this is precisely the type of line-drawing that the Board is "entitled" to make "without our constant second-guessing." N.L.R.B. v. Kentucky River Community Care, Inc. , 532 U.S. 706, 720, 121 S.Ct. 1861, 149 L.Ed.2d 939 (2001).

USC next argues that, by leaving open exactly how many of the five decision-making areas over which faculty must exercise control in order to qualify as managerial, Pacific Lutheran fails to satisfy this court's instruction in LeMoyne-Owen that it identify "which factors are significant and which less so, and why." 357 F.3d at 61. We disagree. At the time of LeMoyne-Owen , the Board was engaging in a totality-of-the-circumstances analysis that we feared might be nothing more than "a cloak for agency whim." Id . In response, the Board developed the Pacific Lutheran framework, which provides detailed standards for effective control, five precisely articulated decision-making areas, a list of the relevant factors within each area, and a prioritization of the areas. True, the Board could have been more precise, but nothing in LeMoyne-Owen requires it to have done so, especially because managerial *142status determinations "do[ ] not lend [themselves] to ex ante line drawing or a mathematical exercise in box checking." Respondent's Br. 39.

Finally, USC argues that, even if the Pacific Lutheran framework complies with Yeshiva , the Board's classification of the Roski non-tenure-track faculty as non-managerial was unsupported by substantial evidence. Relatedly, USC argues that the Regional Director erred in denying its motion to reopen the record concerning allegedly contradictory testimony from a Union witness. But given our conclusion that the Board's subgroup majority status rule runs afoul of Yeshiva , and because we are uncertain whether the Board would have reached the same conclusion absent that rule, we think it best to leave these arguments unaddressed and to instead give the Board an opportunity to reconsider the case afresh under the proper legal standard. See Braniff Airways, Inc. v. Civil Aeronautics Board , 379 F.2d 453, 466 (D.C. Cir. 1967) (remanding to the agency because the "court [wa]s in substantial doubt whether the administrative agency would have made the same ultimate finding with the erroneous findings or inferences removed from the picture" (internal quotation marks omitted)).

V.

For the foregoing reasons, the petition for review is granted in part, the Board's cross-application for enforcement is denied, and the case is remanded to the Board for further proceedings consistent with this opinion. We, of course, express no opinion as to whether the application of the differing standard on remand would lead to the same or a different result.

So ordered.