Sanitary Truck Drivers v. NLRB

United States Court of Appeals FOR THE DISTRICT OF COLUMBIA CIRCUIT Argued February 25, 2022 Decided July 29, 2022 No. 21-1093 SANITARY TRUCK DRIVERS AND HELPERS LOCAL 350, INTERNATIONAL BROTHERHOOD OF TEAMSTERS, PETITIONER v. NATIONAL LABOR RELATIONS BOARD, RESPONDENT BROWNING-FERRIS INDUSTRIES OF CALIFORNIA, INC., D/B/A NEWBY ISLAND RECYCLERY, INTERVENOR On Petition for Review of Orders of the National Labor Relations Board Maneesh Sharma argued the cause for petitioner. With him on the briefs were Susan K. Garea and Harold Craig Becker. Milakshmi V. Rajapakse, Attorney, National Labor Relations Board, argued the cause for respondent. With her on the brief were Jennifer A. Abruzzo, General Counsel, Ruth E. Burdick, Deputy Associate General Counsel, David — 2 — Habenstreit, Assistant General Counsel, and Julie Brock Broido, Supervisory Attorney. Joshua L. Ditelberg and Stuart Newman were on the brief for intervenor for respondent Browning-Ferris Industries of California, Inc. Before: MILLETT, WILKINS, and JACKSON*, Circuit Judges. Opinion for the Court filed by Circuit Judge WILKINS. WILKINS, Circuit Judge: Intervenor Browning-Ferris Industries of California, Inc. (“Browning-Ferris”) operates a recycling plant in Milpitas, California, where it employs about 60 workers. Browning-Ferris contracts with Leadpoint Business Services (“Leadpoint”), which provides BrowningFerris with approximately 240 additional recyclery workers. In July 2013, Petitioner Sanitary Truck Drivers and Helpers Local 350, International Brotherhood of Teamsters (the “Union”) filed a petition with the National Labor Relations Board (“NLRB” or “Board”) to represent Leadpoint’s recyclery workers, asserting that Browning-Ferris and Leadpoint are joint employers of Leadpoint’s workers. The rights and duties of the various parties to this dispute have been extensively litigated before the NLRB. The Union now urges us to vacate two of the NLRB’s recent orders, in which the Board declined to hold BrowningFerris to be a joint employer under the National Labor Relations Act (“NLRA” or “Act”). For the reasons set forth * Circuit Judge, now Justice, Jackson was a member of the panel at the time the case was argued but did not participate in this opinion. — 3 — below, we grant the Union’s petition and vacate the challenged orders. I. We previously set forth facts relevant to the instant petition in a prior opinion. See Browning-Ferris Indus. of Cal., Inc. v. NLRB, 911 F.3d 1195 (D.C. Cir. 2018). Assuming familiarity with the factual and procedural history of the case, we repeat only those details necessary to our resolution of this petition. As mentioned, in July 2013, the Union filed a petition with the NLRB to represent Leadpoint’s recyclery workers, contending that Leadpoint and Browning-Ferris are joint employers. In August 2013, the Board’s Acting Regional Director—applying the NLRB’s then-standard “joint employer” test from TLI, Inc., 271 NLRB 798 , 798–99 (1984), enforced mem., 772 F.2d 894 (3d Cir. 1985), as modified by subsequent cases—found that Leadpoint is the sole employer of its recyclery workers at Browning-Ferris’s facility. The Acting Regional Director reasoned that Browning-Ferris was not a joint employer for purposes of the NLRA because it did not exercise “direct and immediate” control over Leadpoint workers’ essential terms and conditions of employment—such as hiring, firing, or discipline. J.A. 263 (citing TLI, Inc., 271 NLRB at 798–99). The Union sought the Board’s review of the Acting Regional Director’s decision. In August 2015, the Board issued an order in which it overruled TLI and announced a revised joint-employer test. Browning-Ferris Indus. of Cal., Inc., 362 NLRB 1599 , 1613–15, (2015) (“Browning-Ferris I”). In Browning-Ferris I, the Board reasoned that evidence of indirect control can establish joint-employer status. Id. at 1600. The Board also determined that a putative employer’s reserved — 4 — powers of control—even when unexercised—are “clearly relevant to the joint-employment inquiry.” Id. To that end, the Board determined—“based on a full assessment of the facts”— that Browning-Ferris is a joint employer of Leadpoint’s employees because it exercises reserved, indirect, and direct control over them. Id. at 1614–15. Thereafter, the NLRB’s General Counsel issued an unfair labor practice complaint alleging that Browning-Ferris and Leadpoint, as joint employers, violated the NLRA by refusing to bargain with the Union. Browning-Ferris petitioned for judicial review of Browning-Ferris I. This Court mostly upheld the Board’s ruling in Browning-Ferris I, but it remanded to the Board to “rearticulat[e]” the indirect-control element of its new joint-employer test in a way that would conform with the common law, to “meaningfully apply” and explain the second part of its new two-step test, and to consider whether retroactive application of its new test was proper. Browning-Ferris, 911 F.3d at 1221–22. Upon this Court’s limited remand, the Board took a different course. In July 2020, the Board held that it was manifestly unjust to apply its new rule in Browning-Ferris I to these parties, affirmed the Acting Regional Director’s original decision finding that Browning-Ferris was not a joint employer, and dismissed the General Counsel’s unfair labor practice complaint against Browning-Ferris. Browning-Ferris Indus. of Cal., Inc., 369 NLRB No. 139 , at *6 (2020) (“Browning-Ferris II”). Instead, the Board issued an order in which it announced a revised joint-employer test, held that Browning-Ferris is not a joint employer of Leadpoint’s employees, and dismissed the General Counsel’s unfair labor practice complaint against Browning-Ferris. Id. at *6. — 5 — The Union moved for reconsideration of Browning-Ferris II, but the Board denied the motion. Browning-Ferris Indus. of Cal., Inc., 370 NLRB No. 86 , at *1 (2021) (“Browning-Ferris III”). The Union then filed the instant petition seeking judicial review of the Board’s rulings in Browning-Ferris II and Browning-Ferris III. For the reasons discussed below, we grant the Union’s petition and vacate the challenged orders. Prior to the 2015 Browning-Ferris I decision, the Board’s “longstanding joint-employer standard” was one “under which ‘two or more statutory employers are joint employers of the same statutory employees if they share or codetermine those matters governing the essential terms and conditions of employment.’” Browning-Ferris, 911 F.3d at 1205 (quoting Browning-Ferris I, 362 NLRB at 1600) (some internal quotation marks omitted). But in Browning-Ferris I, the Board announced that it would apply that standard in a new way: In determining whether a putative joint employer meets this standard, the initial inquiry is whether there is a common-law employment relationship with the employees in question. If this common-law employment relationship exists, the inquiry then turns to whether the putative joint employer possesses sufficient control over employees’ essential terms and conditions of employment to permit meaningful collective bargaining. 362 NLRB at 1600. The Board clarified that it would not require that “a statutory employer’s control must be exercised directly and immediately.” Id. Rather, “[i]f otherwise sufficient, control exercised indirectly--such as through an intermediary--may” be sufficient to “establish joint-employer status.” Id. — 6 — In our 2018 ruling, we held that “[t]he Board . . . correctly determined that the common-law inquiry is not woodenly confined to indicia of direct and immediate control; an employer’s indirect control over employees can be a relevant consideration.” Browning-Ferris, 911 F.3d at 1209 . But “[i]n applying the indirect-control factor in this case . . . the Board failed to confine it to indirect control over the essential terms and conditions of the workers’ employment.” Id. Accordingly, we remanded “that aspect of the decision to the Board for it to explain and apply its test in a manner that hews to the common law of agency.” Id. Importantly, before the Court issued its decision in 2018, the Board initiated a rulemaking proceeding to establish standards for determining joint-employer status. See The Standard for Determining Joint-Employer Status, 83 Fed. Reg. 46,681 (Sept. 14, 2018). In February 2020, the Board issued a final rule that reinstated a clarified version of the jointemployer standard that was in place prior to Browning-Ferris I. See Browning-Ferris II, 369 NLRB at *1 n.3 (citing Joint Employer Status Under the National Labor Relations Act, 85 Fed. Reg. 11,184 (Feb. 26, 2020) (“2020 Rule”)). The 2020 Rule applied prospectively, so the Board did not apply it to this dispute. Id. In July 2020, on remand from this Court, the Board issued Browning-Ferris II, in which it held that it would be manifestly unjust to apply the rule announced in Browning-Ferris I retroactively to find Browning-Ferris a joint employer. Browning-Ferris II, 369 NLRB at *5 . The Board declared in Browning-Ferris II that retroactive application of BrowningFerris I was improper because “for at least 30 years preceding [Browning-Ferris I,] . . . there was a clear rule of law requiring proof of direct and immediate control under the applicable — 7 — joint-employer test.” Browning-Ferris II, 369 NLRB at *5 . “Indeed, numerous comments filed in our recent jointemployer rulemaking proceeding made abundantly clear that many businesses did rely on that legal standard.” Id. Therefore, “the new standard adopted in the 2015 decision would substantially affect reasonable, settled expectations for relationships established on the basis of the prior standard.” Id. The Board acknowledged that this Court’s remand “sought clarification and redress of two critical shortcomings in the Board’s discussion of its new joint-employer standard,” but the Board concluded that “there is no variation or explanation of that standard that would not incorporate its substantial departure from the prior direct and immediate control legal standard.” Id. As such, the Board was reticent to apply Browning-Ferris I retroactively to the dispute. “In determining whether retroactive application will work a manifest injustice, the Board typically considers the reliance of the parties on preexisting law, the effect of retroactivity on accomplishment of the purposes of the Act, and any particular injustice arising from retroactive application.” Id. at *4 (citing SNE Enters., 344 NLRB 673 , 673 (2005)). “Retroactive application of that new standard would mean that” BrowningFerris “would be suddenly confronted with the new reality that preexisting business relationships with other entities, such as Leadpoint—relationships formed in reliance on a decades-old direct-and-immediate-control standard for determining jointemployer status—thrust upon them unanticipated and unintended duties and liabilities under the Act.” Id. at *5. The Board reasoned that “such a change represents a substitution of ‘new law for old law that was reasonably clear,’” id. (quoting Browning-Ferris, 911 F.3d at 1222 ), and therefore, “it would be manifestly unjust to fail to give [Browning-Ferris] and — 8 — similarly affected businesses reasonable warning before imposing such significant new duties and liabilities.” Id. The Board supported its conclusion by citing to a 1984 decision in which the agency held that a union election must be re-run because two entities had been listed on the ballot as employers, one of which the Board later found not to be a joint employer. See id. (citing H&W Motor Express, 271 NLRB 466 (1984)). The Board found that, because only Leadpoint had been on the ballot at the time of the union election here, H&W Motor Express further undermined the case for retroactively applying Browning-Ferris I. Thus, the Board in Browning-Ferris II declined to apply the test from Browning-Ferris I retroactively, dismissed the NLRB General Counsel’s unfair labor practice complaint, and held that “[t]he joint-employer issue must be resolved under the prior longstanding standard requiring proof of direct and immediate control.” 369 NLRB at *6 . II. The NLRA guarantees employees “the right to selforganization, to form, join, or assist labor organizations, to bargain collectively through representatives of their own choosing, and to engage in other concerted activities for the purpose of collective bargaining or other mutual aid or protection. . . .” 29 U.S.C. § 157 ; Midwest Div.—MMC, LLC v. NLRB, 867 F.3d 1288, 1293 (D.C. Cir. 2017). As we previously held, the NLRA’s “test for joint-employer status is determined by the common law of agency.” Browning-Ferris, 911 F.3d at 1206 . We review the Board’s retroactivity decision by assessing its factfinding for substantial evidence. “Substantial evidence — 9 — ‘means such relevant evidence as a reasonable mind might accept as adequate to support a conclusion.’” NLRB v. Ingredion Inc., 930 F.3d 509, 514 (D.C. Cir. 2019) (quoting Universal Camera Corp. v. NLRB, 340 U.S. 474, 477 (1951)). In short, we “affirm the Board’s findings unless ‘no reasonable factfinder’ could find as it did.” Id. (quoting Alden Leeds, Inc. v. NLRB, 812 F.3d 159, 165 (D.C. Cir. 2016)). With respect to whether retroactive application would advance the policies of the NLRA, we review the Board’s determination with a degree of deference. NLRB v. Food Store Emps. Union, Loc. 347, 417 U.S. 1 , 10 n.10 (1974) (“a court reviewing an agency decision following an intervening change of policy by the agency should remand to permit the agency to decide in the first instance whether giving the change retrospective effect will best effectuate the policies underlying the agency’s governing act”); Cadillac of Naperville, Inc. v. NLRB, 14 F.4th 703, 714 (D.C. Cir. 2021) (per curiam). Finally, we review de novo the Board’s determination as to whether it replaced clear, settled law with new law. See Qwest Servs. Corp. v. FCC, 509 F.3d 531, 539 (D.C. Cir. 2007). The Union contends that in the challenged orders the Board defied our orders in Browning-Ferris and acted arbitrarily in making its retroactivity decision. We need not decide whether the Board properly heeded our remand instructions because, as we explain below, the Board made multiple overlapping errors in its retroactivity analysis in its orders on remand that require vacatur of Browning-Ferris II and Browning-Ferris III. III. In Browning-Ferris II, the Board asserted that BrowningFerris I generated a sea change in the joint-employer test, for which there had been “a clear rule of law requiring proof of — 10 — direct and immediate control” “for at least 30 years.” 369 NLRB at *5. As such, the Board held that it would be manifestly unjust to apply any variation of Browning-Ferris I retroactively. Id. But the Board’s own precedent belies these assertions. A. As we held in our 2018 decision, the Board’s precedent on what established joint-employer status had varied in the decade preceding Browning-Ferris I. See Browning-Ferris, 911 F.3d at 1200–01. In other words, contrary to the Board’s declaration in Browning-Ferris II, see 369 NLRB at *5, the Board’s precedent on the joint-employer standard was anything but static. Indeed, the Board first announced the “direct and immediate control” standard in 2002. See In re Airborne Freight Co., 338 NLRB 597 , 597 n.1 (2002) (“The essential element in this analysis is whether a putative joint employer’s control over employment matters is direct and immediate.”); see NLRB v. CNN Am., Inc., 865 F.3d 740, 749 (D.C. Cir. 2017) (noting that the Board has acknowledged that Airborne Freight “held that the essential element in the joint-employer analysis is whether a putative joint employer’s control over employment matters is direct and immediate”) (cleaned up) (citation omitted). Moreover, in describing the old, assertedly settled rule, the Board failed to consider its own recent 2020 rulemaking, which explained that it in fact had never actually ceased considering indirect and reserved control, even though it did not consider those factors dispositive standing alone. See 2020 Rule, 85 Fed. Reg. at 11,227. Nor did the agency grapple with the question of how “fact-specific” and “case-by-case” its former — 11 — “style of adjudication” had been—a factor that we told the Board could well be important in assessing whether application of a new law would be manifestly unjust. Browning-Ferris, 911 F.3d at 1222 (quoting AT&T v. FCC, 454 F.3d 329 , 333– 34 (D.C. Cir. 2016)) (internal quotation marks omitted); cf. 2020 Rule, 85 Fed. Reg. at 11,224 (“Joint-employer determinations have always been fact-intensive, and they will continue to be so.”). Thus, in Browning-Ferris II—a decision issued just five months after the Board announced the 2020 Rule—the Board inexplicably overlooked the longstanding role of indirect control in the Board’s joint-employer inquiry. We therefore conclude that the Board’s retroactivity analysis in Browning-Ferris II was erroneous. The Board failed to establish that Browning-Ferris I represented the kind of clear departure from longstanding and settled law that the agency said justified its retroactivity conclusion. B. Next, the Board erred in holding that there was “no variation or explanation” of the joint-employer test in Browning-Ferris I that would not result in manifest injustice. Browning-Ferris II, 369 NLRB at *5. As the Board explained in Browning-Ferris I, and as we echoed in our 2018 opinion, the agency’s holding that Browning-Ferris was a joint employer rested on finding that the company had “reserved, direct, and indirect control over Leadpoint employees.” Browning-Ferris I, 362 NLRB at 1615; see Browning-Ferris, 911 F.3d at 1218 (“[T]he Board’s decision turned on its finding that Browning-Ferris exercised control both directly and indirectly.”) (internal quotation marks omitted). The Board failed to explain how it would be a manifest injustice for the Board to consider all of those factors here in light of the agency’s assertion in its 2020 rulemaking that it had — 12 — previously considered reserved and indirect control in assessing joint-employer status. See 2020 Rule, 85 Fed. Reg. at 11,227; Resp. Br. 47 (“Board precedent predating BrowningFerris I” is consistent with “finding joint-employer status based on evidence of direct, indirect, and reserved control”). Those two positions are in direct conflict, and the Board’s failure to address the contradiction was arbitrary and capricious. Gilbert v. NLRB, 56 F.3d 1438 , 1445 (D.C. Cir. 1995) (“It is, of course, elementary that an agency must conform to its prior decisions or explain the reason for its departure from such precedent.”). C. Furthermore, in its decisions on remand, the Board also neglected to consider the second part of the Browning-Ferris I test. At the second step, the Board assesses whether the “putative joint employer possesses sufficient control over employees’ essential terms and conditions of employment to permit meaningful collective bargaining.” Browning-Ferris I, 362 NLRB at 1600; see also Browning-Ferris, 911 F.3d at 1221 (noting that this is the “second half to the Board’s new test” and calling on the agency to “meaningfully apply” it if it concludes that Browning-Ferris is a joint employer under the common law). In its rulemaking, the agency observed that the second step may significantly limit the number of parties deemed joint employers. See 2020 Rule, 85 Fed. Reg. at 11,211. The Board thus made clear that the second step could render BrowningFerris I no more than a clarification of the agency’s legal rule, rather than “a substitution of new law for old law that was reasonably clear,” yet it is the latter that typically justifies refraining from applying a rule retroactively. Verizon Tel. Cos. v. FCC, 269 F.3d 1098, 1109 (D.C. Cir. 2001) (internal — 13 — quotation marks and citation omitted). The challenged orders have nothing to say about the second part of the test, leaving the Board’s conclusion that it would be manifestly unjust to apply any variant of Browning-Ferris I retroactively unreasoned, illogical, and inconsistent with the Board’s position in the 2020 Rule. D. Next, the Board made no specific finding that BrowningFerris had itself actually relied to its detriment on the preBrowning-Ferris I joint-employer standard. Instead, the agency found it would have been reasonable for the firm to have done so and pointed to comments filed by other parties in the 2020 rulemaking asserting that they had relied on the old test. Browning-Ferris II, 369 NLRB at *5. Yet the Board did not explain how it is manifestly unjust to apply a new rule retroactively to a party that it did not find had in fact relied on the old rule. And the Board’s retroactivity test, as well as ours, consistently focuses on the reliance of the parties before the tribunal. See SNE Enters., 344 NLRB at 673 (“In determining whether the retroactive application of a Board rule will cause manifest injustice, the Board will consider the reliance of the parties on preexisting law, the effect of retroactivity on accomplishment of the purposes of the [NLRA], and any particular injustice arising from retroactive application.”); AT&T, 454 F.3d at 332 (“AT & T does not and indeed cannot point us to a settled rule on which it reasonably relied”); Pub. Serv. Co. of Colorado v. FERC, 91 F.3d 1478, 1490 (D.C. Cir. 1996) (“[T]he apparent lack of detrimental reliance on the part of the [parties] is the crucial point” supporting retroactivity); CHARLES H. KOCH & RICHARD MURPHY, 2 ADMIN. L. & PRAC. § 5:67 (3d ed. 2022) (“In reviewing retroactive application of a rule created in adjudication, the court will look to the reliance of the parties.”). Indeed, the Board pointed to no evidence, let — 14 — alone substantial evidence, that Browning-Ferris had relied on the joint-employer test that predated Browning-Ferris I to its detriment. In sum, no reasonable factfinder could find as the Board did. See Ingredion, 930 F.3d at 514 . E. Next, the Board’s invocation of its ruling in H&W Motor Express, 271 NLRB No. 80 (1984), cannot redeem its retroactivity decision. See Browning-Ferris II, 369 NLRB at *5. In H&W, the Board determined that a company was not a joint employer and ordered a re-run of a union election in which that company had been on the ballot because “employees . . . cast their ballots based on the Regional Director’s finding that a joint employer relationship existed.” 271 NLRB at 468. In Browning-Ferris II, the Board reasoned that “the reverse situation exists” here, and “the principles stated in H&W Motor Express militate against retroactivity all the same with respect to imposition of a bargaining obligation on [Browning-Ferris] when the employees here cast their ballots on the assumption that a joint-employer relationship did not exist.” 369 NLRB at *5. Even assuming that H&W provides that an employer must be on a union ballot before it can be ordered to the bargaining table—and the Union cites contrary caselaw on this question— H&W does not justify the Board’s decision here. See Pet’r Reply Br. 17 (“The Board regularly adds employers through unfair labor practice proceedings, and without an election, to established bargaining relationships if it determines the added employer is a joint employer.”) (citing Branch Int’l Servs., Inc., 427 NLRB 209 (1998)). If anything, H&W would have supported requiring another union election, not holding that — 15 — Browning-Ferris is not a joint employer, which is what the Board did here. F. Finally, the Board’s adoption of the NLRB Acting Regional Director’s decision was erroneous in multiple respects. To begin with, the agency was obligated to explain how the test the Acting Regional Director applied was consistent with the common law, or if it was not, then why its application was nevertheless proper. Our court’s 2018 decision made clear that “the right-to-control element of the Board’s joint-employer standard [discussed in Browning-Ferris I] has deep roots in the common law,” see Browning-Ferris, 911 F.3d at 1199 , and that the “common-law rule” is that “unexercised control bears on employer status.” Id. at 1210 ; see id. at 1211 (“[T]hat ‘common-law element of control is the principal guidepost’ in determining whether an entity is an employer of another.”) (citation omitted and emphasis added). Further, we held that “there is no sound reason that the . . . joint-employer inquiry would give [indirect control] a cold shoulder.” Id. at 1218 (“[The] argument that the common law of agency closes its mind to evidence of indirect control is unsupported by law or logic.”); see id. at 1216 (a “rigid distinction between direct and indirect control has no anchor in the common law”). In other words, in 2018, we held that the common law would not countenance ignoring reserved and indirect control in assessing joint-employer status. As such, we specifically upheld the Board’s articulation of its legal test in BrowningFerris I as including consideration of reserved and indirect control as “fully consistent with the common law.” Id. at 1222 . We remanded for a clearer articulation and application of the indirect control factor to the facts of “this case,” id., in a manner “that hews to the common law of agency,” id. at 1209 . In sum, — 16 — given that we upheld the core components of the jointemployer test in Browning-Ferris I; recognized that the Board’s test at the time of the Acting Regional Director’s decision essentially ignored both indirect and reserved control, see id. at 1201 ; and made clear that the Board’s joint-employer test must “color within the common-law lines identified by the judiciary[,]” the Board at least had to explain in BrowningFerris II how it could lawfully apply a standard that we have said was out of step with the common law. See id. at 1200 , 1208–09. In Browning-Ferris III, the Board attempted to sidestep this problem by insisting that “[t]he D.C. Circuit no more than agreed that indirect and reserved control can be relevant considerations in the common law, not that they must be given weight independent of direct-and-immediate control.” 370 NLRB at *1 n.2 (emphasis omitted). The Board also reasoned that “the Board’s prior standard fell within the boundaries of the common law as applied in the particular context of the Act,” as demonstrated by “the Board’s recent final rule.” Id. (citing 2020 Rule, 85 Fed. Reg. at 11,184). The Board’s position is dubious. In promulgating the 2020 Rule, the Board explained that it had previously considered reserved and indirect control in the joint-employer inquiry, and cited two orders, both from 2000, to justify that claim. See 2020 Rule, 85 Fed. Reg. at 11,227 (citing Le Rendezvous Restaurant, 332 NLRB 336 (2000) and M.B. Sturgis, Inc., 331 NLRB 1298 (2000)). But the fact that the Board’s standard comported with the common law back in 2000 says nothing about the propriety of the Acting Regional Director’s straitened test in 2013. See Browning-Ferris, 911 F.3d at 1201 . In our 2018 decision, we took great pains to inform the Board that the failure to consider reserved or indirect control is inconsistent with the common law of agency. See id. at 1210–11, 1213, — 17 — 1216–17, 1219. The Board needed to provide a reasoned explanation for adopting the Acting Regional Director’s decision wholesale. It failed to deliver. To make matters worse, in adopting the Acting Regional Director’s recommendation, the Board disregarded the Browning-Ferris I Board’s rejection of many of the Acting Regional Director’s factual findings. In particular, the Board “disagree[d] with the Regional Director’s factual findings that [Browning-Ferris] does not mandate how many employees work on the line, the speed at which they work, where they stand, or how they pick material.” Browning-Ferris I, 362 NLRB at 1604 n.17. Instead, it concluded that there was a “clear and direct connection between [Browning-Ferris’s] decisions and employee work performance” because Browning-Ferris has “unilateral control over the speed of the [material] streams and specific productivity standards for sorting.” Id. at 1616. Unlike the Acting Regional Director, the Board found “multiple examples of reserved, direct, and indirect control over Leadpoint employees.” Id. at 1615. We also noted in our prior decision that “whether indirect control can be dispositive is not at issue in this case because the Board’s decision [in Browning-Ferris I] turned on its finding that Browning-Ferris exercised control both directly and indirectly.” Browning-Ferris, 911 F.3d at 1218 (emphasis added) (internal quotation marks and citation omitted). Yet in the challenged orders the Board adopted the Acting Regional Director’s decision without acknowledging its earlier rejection of important aspects of his factual findings. All the Board said was that the majority in Browning-Ferris I did not overturn the Acting Regional Director’s holding under the old standard. Browning-Ferris II, 369 NLRB at *6. But the Board in Browning-Ferris I had no reason to apply the old standard because it announced and applied a new one. The Board’s — 18 — orders on remand from this court tell us nothing about how it would have ruled in light of its factual disagreements with the Regional Director. And especially given the agency’s current view of its old joint-employer test—under which a finding of direct and immediate control is necessary to deem a party a joint employer but reserved and indirect control can supplement the analysis—the Board’s prior finding that Browning-Ferris had exercised direct control seems critical. Nothing in today’s decision nullifies the essential principle that the National Labor Relations Board is free to change its mind. But the Board must acknowledge when it is doing so and explain its reasoning, see FCC v. Fox Television Stations, Inc., 556 U.S. 502, 515 (2009), and when defining jointemployer status, it “must color within the common-law lines identified by the judiciary[,]” Browning-Ferris, 911 F.3d at 1208 . In summary, we grant the Union’s petition, vacate the challenged orders, and remand for the Board to conduct further proceedings consistent with this opinion. So ordered.