MICHIGAN FAMILY RESOURCES INC v. SERVICE EMPLOYEES INTERNATIONAL UNION LOCAL 517M

Reset A A Font size: Print

United States Court of Appeals,Sixth Circuit.

MICHIGAN FAMILY RESOURCES, INC., Plaintiff-Appellee, v. SERVICE EMPLOYEES INTERNATIONAL UNION LOCAL 517M, Defendant-Appellant.

No. 04-2564.

Decided: January 26, 2007

Before BOGGS, Chief Judge;  MARTIN, BATCHELDER, DAUGHTREY, COLE, CLAY, GILMAN, GIBBONS, ROGERS, SUTTON, COOK, McKEAGUE, and GRIFFIN, Circuit Judges. ARGUED:  Mary Ellen Gurewitz, Sachs Waldman, Detroit, Michigan, for Appellant.  Timothy J. Ryan, Ryan & Lykins, Grand Rapids, Michigan, for Appellee.  James B. Coppess, AFL-CIO Legal Department, Washington, D.C., for Amici Curiae.   ON BRIEF:  Mary Ellen Gurewitz, Sachs Waldman, Detroit, Michigan, Howard F. Gordon, Michigan Public Employees, Service Employees International Union Local 517M, Lansing, Michigan, for Appellant.  Timothy J. Ryan, Ryan & Lykins, Grand Rapids, Michigan, for Appellee.  James B. Coppess, AFL-CIO Legal Department, Washington, D.C., Terry A. Bethel, National Academy of Arbitrators, Indiana University School of Law, Bloomington, Indiana, for Amici Curiae.

OPINION

Local 517M of the Service Employees International Union challenges the decision of the district court vacating an arbitration award in its favor.   Because the arbitrator was “acting within the scope of his authority” in resolving this dispute, because the company has not charged the arbitrator with fraud or dishonesty in making the award, because the arbitrator was “arguably construing ․ the contract” when he awarded union employees a 4% cost-of-living increase for 2003 and because the company has shown no more than that the arbitrator made an error, perhaps even a “serious error,” in interpreting the contract, we reverse and direct the district court to enter an order enforcing the award.   See United Paperworkers Int'l Union, AFL-CIO v. Misco, 484 U.S. 29, 38-39, 108 S.Ct. 364, 98 L.Ed.2d 286 (1987).

I.

Michigan Family Resources (MFR) runs the federal Headstart Program that serves Kent County, which lies in western Michigan.   Local 517M of the Service Employees International Union represents some of MFR's employees.   On behalf of its members, the union negotiated a collective bargaining agreement with MFR that entitled its members to annual wage increases.   The agreement contains four pertinent provisions.

Article 35(1) of the agreement provides:

Bargaining unit members will receive the same cost of living increases paid to other MFR employees pursuant to the directive of MFR's funding source.   The parties understand that the timing and amount of any such increase is entirely dictated by the funding source.

JA 43.   The “funding source” mentioned in this provision, the parties agree, refers to the federal government.

Article 35(2) provides:

During the fall semester of each program year, bargaining unit members will be reviewed and will be considered for a merit increase․ MFR will guarantee at least that for each bargaining unit employee the sum of any COLA [i.e., cost-of-living increase] paid during the year and the merit increase will be as follows:  2002-4%;  2003-2.5%;  2004-3.5%. For example, if the [cost-of-living] increase for 2004 is 2.5%, effective on September 1, 2004 bargaining unit members will receive at least an additional 1.0%.

JA 43-44.

Article 5 provides an “exclusive method of resolving” disputes arising under the agreement and requires the parties to arbitrate any grievances that they cannot resolve on their own.   The arbitrator, it says, “shall have full authority to render a decision which shall be final and binding upon both parties and the employees, except that the arbitrator shall not have authority to change, alter, amend, or deviate from the terms of this collective bargaining agreement in any respect.”   Article 5(c).  The provision also says that “[i]f the Union requests arbitration, the parties shall choose an arbitrator by selecting from the following list through the alternating strike method:  Mario Chiesa[,] Mark Glazer[,] William Daniel[,] George Roumell and Lamont Stallworth.”  Id.

Article 34 provides that the agreement “expresses the understanding of the parties and it will not be changed, modified, or varied, except by written instrument signed by duly authorized agents of the party thereto,” and that “[t]here are no past practices which are binding upon the parties.”   JA 43.

In May 2003, MFR notified the union employees that they would receive a 2.5% cost-of-living increase for 2003-1.5% from the “funding source” (i.e., the federal government), 1% from MFR-while non-union employees would receive a 4% cost-of-living increase for the year.   Although the 2003 pay increase for union employees satisfied the collective bargaining agreement's minimum requirement for that year (2.5%), the union claimed that the agreement required parity between union and non-union employees in the payment of cost-of-living increases.   The union, as a result, filed a grievance against MFR. In accordance with the agreement, the parties engaged one of the designated arbitrators to resolve the dispute.

On December 10, 2003, the arbitrator issued a ten-page opinion resolving the dispute in favor of the union.   The question, as he saw it, was “whether Article 35 requires MFR to provide parity in [cost-of-living] payments for its [union] employees when non-Union employees receive higher ․ payments.”   Arb. Op. at 2. On this point, the arbitrator reasoned, Article 35 was not entirely clear.   While it required union members to “receive the same payments from the federal funding source as other employees,” it did not directly address “the cost of living increases from other sources, such as from the Employer.”   Id. at 7. The arbitrator then noted that before and after the adoption of the collective bargaining agreement, MFR granted the same cost-of-living increase to all employees, regardless of union affiliation.  Id. at 8. MFR never held merit reviews, he added, and in a 2002 memo (dealing with pay increases for the first year of the collective bargaining agreement) it characterized the entire wage increase to union employees as a cost-of-living increase.  “I am persuaded,” the arbitrator then concluded, “that [Article 35] becomes ambiguous because of the Employer's prior decision to characterize both its individual payment and its payment from the federal funding source as [cost of living].”  Id. In the end he resolved this ambiguity based on the employer's prior practice of granting identical cost-of-living increases to all employees and awarded union members an equivalent 4% cost-of-living increase.

On January 9, 2004, invoking § 301 of the Labor Management Relations Act, 29 U.S.C. § 185, MFR filed a complaint in federal court seeking to vacate the award.   On November 10, 2004, 380 F.Supp.2d 886, the district court granted MFR's motion for summary judgment, holding that “the Arbitrator's award does not draw its essence from the [collective bargaining agreement] because the Arbitrator considered evidence to aid in construing the [collective bargaining agreement] when, in fact, no construction was necessary.”  380 F.Supp.2d at 890. “[T]he Arbitrator,” the court concluded, “went beyond the express terms of the [collective bargaining agreement] by imposing additional requirements upon the parties and considering past practices, which are specifically disclaimed by the [collective bargaining agreement's] waiver provision.”  Id.

A panel of this court affirmed, Mich. Family Res., Inc. v. Serv. Employees Int'l Union Local 517M, 438 F.3d 653 (6th Cir.2006), and applied this court's four-part test for assessing arbitration awards in doing so, see Cement Divs., Nat'l Gypsum Co. v. United Steelworkers of America, AFL-CIO-CLC, Local 135, 793 F.2d 759 (6th Cir.1986).   The union petitioned for and we granted en banc review.   See 2006 U.S.App. LEXIS 11479 (6th Cir. May 5, 2006).

II.

A.

In 1960, the Supreme Court issued three decisions designed to end the federal courts' hostility to labor-arbitration awards.   In one of those decisions, United Steelworkers v. Warrior & Gulf Navigation Co., 363 U.S. 574, 80 S.Ct. 1347, 4 L.Ed.2d 1409 (1960), the Court held that while federal courts, not arbitrators, should determine whether a collective bargaining agreement requires a dispute to be resolved through arbitration, “[d]oubts should be resolved in favor of coverage,” and “[a]n order to arbitrate ․ should not be denied unless it may be said with positive assurance that the arbitration clause is not susceptible of an interpretation that covers the asserted dispute.”  Id. at 582-83, 80 S.Ct. 1347.   In another decision, United Steelworkers v. American Manufacturing Co., 363 U.S. 564, 80 S.Ct. 1363, 4 L.Ed.2d 1432 (1960), the Court held that if the collective bargaining agreement provided that a dispute should be submitted to arbitration, the underlying “question of contract interpretation [is] for the arbitrator,” even though the claim in the court's eyes appears to be a “frivolous” one.  Id. at 568, 80 S.Ct. 1363.  “The courts,” it pointed out, “have no business weighing the merits of the grievance, considering whether there is equity in a particular claim, or determining whether there is particular language in the written instrument which will support the claim.”  Id.

In the final case, the one most pertinent here, United Steelworkers v. Enterprise Wheel & Car Corp., 363 U.S. 593, 80 S.Ct. 1358, 4 L.Ed.2d 1424 (1960), the Court considered the role of the federal courts in enforcing arbitration awards.  “The refusal of courts to review the merits of an arbitration award,” it began, “is the proper approach to arbitration under collective bargaining agreements” because “[t]he federal policy of settling labor disputes by arbitration would be undermined if courts had the final say on the merits of the awards.”  Id. at 596, 80 S.Ct. 1358.   Observing that the court of appeals' refusal to enforce the award in that case was not “based upon any finding that the arbitrator did not premise his award on his construction of the contract,” but was based “merely” on “disagree[ment] with the arbitrator's construction of it,” the Court reversed.  Id. at 598, 80 S.Ct. 1358.   The company's primary argument in favor of declining to enforce the award, the Court explained, was that “by applying correct principles of law to the interpretation of the collective bargaining agreement it can be determined that the agreement did not ․ provide” for this kind of award, “and that therefore the arbitrator's decision was not based upon the contract.”   Id. Yet that approach, the Court countered, “would require courts, even under the standard arbitration clause, to review the merits of every construction of the contract” and “would make meaningless the provisions that the arbitrator's decision is final” because it never would be.  Id. at 598-99, 80 S.Ct. 1358.  “[S]o far as the arbitrator's decision concerns construction of the contract, the courts have no business overruling him because their interpretation of the contract is different from his.”  Id. at 599, 80 S.Ct. 1358.   Only when an arbitrator declines to confine himself “to interpretation and application” of the agreement and instead opts to “dispense his own brand of industrial justice” may the courts refuse to enforce an award.  Id. at 597, 80 S.Ct. 1358.   Thus while the arbitrator may “look for guidance from many sources” in construing the agreement, “his award is legitimate only so long as it draws its essence from the collective bargaining agreement.”  Id.

In the years after what came to be known as the Steelworkers Trilogy, our court refereed numerous disputes over what it means for an award to “draw its essence” from the collective bargaining agreement.   Attempting to summarize those decisions, we eventually announced a four-part inquiry for determining whether to enforce an arbitration award.  “An award fails to draw its essence from the agreement,” we stated, when (1) it “conflicts with express terms of the collective bargaining agreement;  (2)[it] imposes additional requirements that are not expressly provided in the agreement;  (3)[it] is without rational support or cannot be rationally derived from the terms of the agreement;  and (4)[it] is based on general considerations of fairness and equity instead of the precise terms of the agreement.”  Cement Divs., Nat'l Gypsum Co. v. United Steelworkers, Local 135, 793 F.2d 759, 766 (6th Cir.1986) (citations omitted).   If a dissatisfied party to an arbitration satisfied any one of these tests, we held, the court should not enforce the award.   See Wyandot, Inc. v. Local 227, United Food & Commercial Workers Union, 205 F.3d 922, 929 n. 3 (6th Cir.2000).

During the 20 years since Cement Divisions, the Supreme Court has refined the standard of review in this area in two cases, both of which suggest that Cement Divisions gives federal courts more latitude to review the merits of an arbitration award than the Supreme Court permits.   In 1987, just one year after Cement Divisions, the Court considered “when a federal court may refuse to enforce an arbitration award rendered under a collective-bargaining agreement” in United Paperworkers Int'l Union, AFL-CIO v. Misco, Inc., 484 U.S. 29, 31, 108 S.Ct. 364, 98 L.Ed.2d 286 (1987).   Elaborating on the Steelworkers Trilogy, Misco reminded litigants that the courts have a “limited role” when the losing party to an arbitration seeks judicial intervention.  Id. at 36, 108 S.Ct. 364.   The federal courts retain authority to ensure that the arbitration award grows out of a legitimate process:  that the collective bargaining agreement commits the dispute at hand to arbitration, id. at 38, 108 S.Ct. 364 (requiring the arbitrator to “act [ ] within the scope of his authority”);  that the prevailing party did not procure the decision through “fraud,” id.;   that the arbitrator did not suffer from a conflict of interest or exercise “dishonesty” in reaching his decision, id.;   and that the arbitration award did not merely reflect “the arbitrator's own notions of industrial justice,” id.

Otherwise, Misco explained, “[t]he courts are not authorized to reconsider the merits of an award even though the parties may allege that the award rests on errors of fact or on misinterpretation of the contract.”  Id. at 36, 108 S.Ct. 364.  “Because the parties have contracted to have disputes settled by an arbitrator chosen by them rather than by a judge, it is the arbitrator's view of the facts and of the meaning of the contract that they have agreed to accept.”  Id. at 37-38, 108 S.Ct. 364.   In attempting to clarify the line between an award that permissibly “draws its essence from the contract” and one that impermissibly “reflect[s] the arbitrator's own notions of industrial justice,” the Court gave this guidance:  “[A]s long as the arbitrator is even arguably construing or applying the contract and acting within the scope of his authority, that a court is convinced he committed serious error does not suffice to overturn his decision.”  Id. at 38, 108 S.Ct. 364.   Consistent with this frame of reference, the Court upheld the award in that case because the arbitrator arguably was construing and applying the contract and because the decision involved at worst “improvident, even silly,” decisionmaking.   Id. at 39, 108 S.Ct. 364.

Major League Baseball Players Ass'n v. Garvey, 532 U.S. 504, 121 S.Ct. 1724, 149 L.Ed.2d 740 (2001), reinforced the federal courts' modest role in this area.   In summarily reversing a court of appeals' decision, the Court explained that the federal courts' review of labor-arbitration decisions is not just “limited,” Misco, 484 U.S. at 36, 108 S.Ct. 364;  it is “very limited,” Garvey, 532 U.S. at 509, 121 S.Ct. 1724.   Rather than asking whether the arbitration award “draws its essence from the contract,” the Court asked whether the arbitrator was “ ‘even arguably construing or applying the contract.’ ” Id. at 509, 121 S.Ct. 1724 (quoting Misco, 484 U.S. at 38, 108 S.Ct. 364).  “[O]nly when the arbitrator strays from interpretation and application,” the Court explained, does he enter the forbidden world of “effectively ‘dispens[ing] his own brand of industrial justice,’ ” making the arbitrator's decision “unenforceable.”  Id. at 509, 108 S.Ct. 364 (quoting Enterprise Wheel, 363 U.S. at 597, 80 S.Ct. 1358).   But when “an arbitrator resolves disputes regarding the application of a contract, and no dishonesty is alleged, the arbitrator's ‘improvident, even silly, factfinding’ does not provide a basis for a reviewing court to refuse to enforce the award.”  Id. (quoting Misco, 484 U.S. at 39, 108 S.Ct. 364).

Misco and Garvey refine the Steelworkers Trilogy in two ways.   One, they define the line between a permissible award (that “draws its essence from the contract”) and an impermissible award (that “simply reflect[s] the arbitrator's own notion[ ] of industrial justice”) based on whether “the arbitrator is even arguably construing or applying the contract and acting within the scope of his authority.”  Misco, 484 U.S. at 38, 108 S.Ct. 364;  see Garvey, 532 U.S. at 509, 121 S.Ct. 1724.   Two, Misco and Garvey show that the Court means what it is saying.   In both cases, the Court held that once it was established that the arbitrator was construing or applying the contract (and acting within the scope of his authority), it made no difference whether the arbitrator had committed “serious,” “improvident” or even “silly” errors in resolving the merits of the dispute.  Misco, 484 U.S. at 38-39, 108 S.Ct. 364;  Garvey, 532 U.S. at 509-10, 121 S.Ct. 1724.

 Our decision in Cement Divisions of course does not reflect these refinements of the Steelworkers Trilogy because the court had no reason to know they existed.   Accordingly, instead of continuing to apply Cement Divisions' four-part inquiry, a test we now overrule, we will consider the questions of “procedural aberration” that Misco and Garvey identify.  Misco, 484 U.S. at 40 n. 10, 108 S.Ct. 364.   Did the arbitrator act “outside his authority” by resolving a dispute not committed to arbitration?   Did the arbitrator commit fraud, have a conflict of interest or otherwise act dishonestly in issuing the award?   And in resolving any legal or factual disputes in the case, was the arbitrator “arguably construing or applying the contract”?   So long as the arbitrator does not offend any of these requirements, the request for judicial intervention should be resisted even though the arbitrator made “serious,” “improvident” or “silly” errors in resolving the merits of the dispute.

These relatively straightforward inquiries, we acknowledge, mask a harder question:  What role, if any, still remains for a court to consider the arbitrator's resolution of the merits in determining whether he was “arguably construing” the contract?   The union, supported by the national AFL-CIO, argues that the court's view of the merits has nothing to do with this process-driven question.   So long as the arbitrator purported to construe the contract, that is the end of the matter.   MFR argues by contrast that federal courts always must consider the merits of a dispute to determine whether the arbitrator was arguably construing the contract.

 We respectfully disagree with both of them.   The Court's repeated insistence that the federal courts must tolerate “serious” arbitral errors suggests that judicial consideration of the merits of a dispute is the rare exception, not the rule.   At the same time, we cannot ignore the specter that an arbitration decision could be so “ignor[ant]” of the contract's “plain language,” Misco, 484 U.S. at 38, 108 S.Ct. 364, as to make implausible any contention that the arbitrator was construing the contract.   An interpretation of a contract thus could be “so untethered to” the terms of the agreement, to borrow a phrase from our Circuit Justice, Garvey, 532 U.S. at 512, 121 S.Ct. 1724 (Stevens, J., dissenting), that it would cast doubt on whether the arbitrator indeed was engaged in interpretation.   Such an exception of course is reserved for the rare case.   For in most cases, it will suffice to enforce the award that the arbitrator appeared to be engaged in interpretation, and if there is doubt we will presume that the arbitrator was doing just that.   Cf. Warrior & Gulf, 363 U.S. at 582-83, 80 S.Ct. 1347 (“An order to arbitrate the particular grievance should not be denied unless it may be said with positive assurance that the arbitration clause is not susceptible of an interpretation that covers the asserted dispute.   Doubts should be resolved in favor of coverage.”).

This view of the “arguably construing” inquiry no doubt will permit only the most egregious awards to be vacated.   But it is a view that respects the parties' decision to hire their own judge to resolve their disputes, a view that respects the finality clause in most arbitration agreements, see JA 27 (stating that “the arbitrator shall have full authority to render a decision which shall be final and binding upon both parties”), and a view whose imperfections can be remedied by selecting better arbitrators.

B.

Gauged by these modest requirements, the arbitrator's award in this case must be enforced.   Nothing in the record, nothing the parties have told us, suggests that fraud, a conflict of interest or dishonesty infected the arbitrator's decision or the arbitral process.   And no one disputes that the collective bargaining agreement committed this grievance to arbitration or disputes that this arbitrator was one of the five arbitrators selected by the parties to be eligible to resolve this dispute.   The arbitrator, in short, was acting within the scope of his authority.

 That leaves the question whether the arbitrator was engaged in interpretation:  Was he “arguably construing” the collective bargaining agreement?   The arbitrator's ten-page opinion has all the hallmarks of interpretation.   He refers to, quotes from and analyzes the pertinent provisions of the agreement, and at no point does he say anything indicating that he was doing anything other than trying to reach a good-faith interpretation of the contract.

Neither can it be said that the arbitrator's decision on the merits was so untethered from the agreement that it casts doubt on whether he was engaged in interpretation, as opposed to the implementation of his “own brand of industrial justice.”   When it comes to parity between union and non-union cost-of-living increases-the issue before the arbitrator-the agreement is clear in one respect but silent in another.   On the one side of the ledger, Article 35(1) of the agreement says that government-funded cost-of-living increases must be distributed equally, providing that “[b]argaining unit members will receive the same cost of living increases paid to other MFR employees pursuant to the directive of MFR's funding source.”   JA 43.   Whatever cost-of-living increase “MFR's funding source” (namely, the federal government) provides in a given year, then, that funding must be allocated equally in terms of percentage salary increases to union and non-union members alike.   On the other side of the ledger, Article 35(1) says nothing about parity when it comes to employer-funded cost-of-living increases.   Viewed from the vantage point of Article 35(1), the agreement is silent about whether parity must exist between union and non-union cost-of-living increases funded by the employer.

Other provisions of the agreement do not eliminate this ambiguity. Article 35(2) addresses merit increases and says that “bargaining unit members will be reviewed and will be considered for a merit increase” each year.   But this sentence says nothing about cost-of-living increases, and MFR premised its 2003 salary increases on changes to the cost of living.

Article 35(2) also “guarantee[s]” each union employee a minimum salary increase for each of the three years of the agreement, so that “the sum of any COLA [i.e., cost-of-living increase] paid during the year and the merit increase will be as follows:  2002-4%;  2003-2.5%;  2004-3.5%. For example, if the [cost-of-living] increase for 2004 is 2.5%, effective on September 1, 2004 bargaining unit members will receive at least an additional 1.0%.” JA 43-44.   But by setting a floor for salary increases to union employees during these three years, the agreement did not impose a ceiling on those increases.   The agreement thus is silent about increases over the floor for union employees and about whether increases over the floor should be given to union and non-union employees on an equivalent basis.

In the face of this contractual silence, the arbitrator did what all adjudicators do under these circumstances:  looked for other indicators of meaning.   In attempting to resolve this ambiguity, he had at least three paths open to him.   In the first place, he could have noted that the agreement is utterly silent about whether there must be parity in employer-funded cost-of-living increases, then drawn an inference from the employer's application of the agreement during its first year.   Because MFR had awarded equal employer-funded cost-of-living increases in 2002, the arbitrator could have drawn the inference that the parties' application of the agreement reflected their understanding of it.   This approach would not have violated Article 34's instruction that “[t]here are no past practices which are binding upon the parties,” JA 43 (emphasis added), and indeed Article 34's prohibition only on using past practices to “bind[ ]” the parties also might well have permitted the arbitrator to draw an inference from the fact that the employer's practice before this collective bargaining agreement had been the same-to give equal cost-of-living adjustments.

Instead of relying on this inference, the arbitrator took a different tack.   He determined that the “above language”-Article 35(1)'s mandate that “[b]argaining unit members will receive the same cost of living increases paid to other MFR employees pursuant to the directive of MFR's funding source,” id.-“becomes ambiguous because of the Employer's prior decision to characterize both its individual payment and its payment from the federal funding source as [cost of living].”   Arb. Op. at 8 (emphasis added).   In other words, because the employer traditionally gave federally funded and its own cost-of-living increases, the arbitrator took the “same cost of living increases” language in Article 35(1) to require parity for all such increases.   This train of reasoning, however, contains two serious flaws.   Later language in this sentence (“pursuant to the directive of MFR's funding source”) shows that the provision refers only to parity for federally funded cost-of-living increases.   And the clear meaning of this sentence does not “become [ ] ambiguous” based on a party's interpretation of it.

 Whether the “becomes ambiguous” phrase was a slip of the pen or a slip in thought, it bears emphasizing that it was still “the above language”-the contract language-that he was trying to figure out.   And the fact remains that the agreement was utterly silent, and thus already ambiguous, on the point at hand:  Did the contract tell MFR whether there should be parity in employer-funded cost-of-living increases?   That ambiguity permitted, indeed required, the arbitrator to engage in construction of the agreement, which is all we are asked to determine.   That he chose the wrong path in justifying the award, however, does not give us a warrant to vacate it.   If it is true that appellate courts generally “review[ ] judgments, not opinions” while performing de novo review, Chevron USA Inc. v. Nat'l Res. Def. Council, 467 U.S. 837, 842, 104 S.Ct. 2778, 81 L.Ed.2d 694 (1984), it is assuredly true that we review outcomes, not opinions, while performing the exceedingly deferential task of considering whether to vacate an arbitration award.   Add to this the fact that arbitrators in this area may be, and frequently are, non-lawyers (and therefore hardly inculcated in all of the nuances of contract interpretation), that arbitrators have no obligation to issue a written decision to justify their awards, see Enterprise Wheel, 363 U.S. at 598, 80 S.Ct. 1358, and that vacating awards based on flawed chains of reasoning may well encourage arbitrators to “play it safe by writing no supporting opinions,” id., and it becomes clear that this award should be enforced.

Through all of this, it needs to be emphasized, we do not endorse either of the above paths of reasoning-either the one the arbitrator took or the one he could have taken.   For our part, the better course for the arbitrator would have been to say that the agreement's explicit parity requirements for government-funded, cost-of-living increases implied that the parties did not mean to create other parity requirements.   And indeed had a district court in a diversity case faced this question, we might well have demanded that answer in conducting plenary review of the issue.   But that of course is not the point.   That the deciphering of this contract required implications and inferences suffices by itself to show that the arbitrator was permissibly engaged in interpretation.   The arbitrator, it is true, made a legal error, perhaps even a serious legal error, but an error of interpretation nonetheless, which does not authorize us to vacate the award.   It was the “arbitrator's construction,” not three layers of federal judicial review, that the parties “bargained for,” and that delegation of decision-making authority must be respected even when time and further review show that the parties in the end have bargained for nothing more than error.

 MFR persists that if the arbitrator erred, then he must have exceeded his authority because Article 5(c) of the collective bargaining agreement says “that the arbitrator shall not have authority to change, alter, amend, or deviate from the terms of this collective bargaining agreement in any respect.”   Not so.   An arbitrator does not exceed his authority every time he makes an interpretive error;  he exceeds that authority only when the collective bargaining agreement does not commit the dispute to arbitration.   Otherwise, every error would be grounds for judicial intervention, which is inconsistent with the Supreme Court's insistence that we must tolerate “serious,” “improvident” and “silly” legal and factual arbitral errors and which is inconsistent with another provision of this collective bargaining agreement (and most collective bargaining agreements):  that the arbitrator “shall have full authority to render a decision which shall be final and binding upon both parties.”   No decision would be final, to say nothing of “speedy,” Misco, 484 U.S. at 38, 108 S.Ct. 364, if the losing party needed only to allege an interpretive error to invoke the intervention of the federal courts.   True, no arbitration agreement (and certainly not this one) explicitly gives the arbitrator jurisdiction to misapprehend the agreement, and neither for that matter does Title 28 of the United States Code explicitly give appellate courts jurisdiction to affirm legal errors. But that is the import of having a finality clause in an arbitration agreement, the import of the Supreme Court's decisions in this area over the last 47 years and, so far as our jurisdiction is concerned, the import of asking fallible human beings to make final sense of imperfectly worded documents.   In the last analysis, we have an arbitrator who plainly was “arguably construing” the contract and who perhaps just as plainly made a “serious error” in construing the contract, a confluence of circumstances that does not invest us with authority to “overturn [the] decision.”  Garvey, 532 U.S. at 509, 121 S.Ct. 1724.   The award must be enforced.

III.

For these reasons we reverse and remand the case so that the district court may enter an order enforcing the arbitration award.

I.

My dissent is a narrow one:  I agree with the majority in principle, but not in application.   As to principle, I agree with the majority's conclusion that this Court's Cement Divisions test be dropped.   See Maj. Op. at 753.   The four-part Cement Divisions inquiry has allowed this Court too much latitude to review the merits of arbitrator interpretations of collective bargaining agreements, in contravention of the dictates of the Supreme Court's Steelworkers Trilogy as clarified in Misco and, to a lesser extent, Garvey.  Cement Divisions should consequently be replaced by a test more reflective of Misco and Garvey, i.e., something to the tune of “as long as the arbitrator is arguably construing the contract or applying the contract and acting within the scope of his authority, this court will not overturn his decision.”

Of course, simply replacing this Court's test with that of the Supreme Court does not advance matters much, especially since the Supreme Court has provided us with only two post-Trilogy cases for guidance.   See Garvey, 532 U.S. at 512, 121 S.Ct. 1724 (noting that the Supreme Court's cases “do not provide significant guidance as to what standards a federal court should use in assessing whether an arbitrator's behavior is so untethered to either the agreement of the parties or the factual record so as to constitute an attempt to ‘dispense his own brand of industrial justice’ ”) (Stevens, J., dissenting).   To be fair, the majority appears to replace Cement Divisions with an inquiry that looks only for “procedural aberration” committed by the arbitrator, not for legal error.   Maj. Op. at 753.   The majority then posits a series of hypotheticals to indicate when such procedural aberrations might have occurred.   This is an admirable start, especially since the majority writes from a clean slate.   But the contours of “procedural aberration”-or, if one prefers, when an arbitrator is not “arguably construing or applying the collective bargaining agreement, and acting within the scope of his authority”-can only be defined as we apply this new test to new, discrete cases.   After all, “[t]he glory of the Anglo-American system of adjudication is that general principles are tested in the crucible of concrete controversies.   A court cannot be assumed to address and resolve in the case in which it first lays down a rule every controversy within the semantic reach of the rule.”   Malhotra v. Cotter & Co., 885 F.2d 1305, 1312 (7th Cir.1989).  Misco was a relatively easy case, because the appeals court there applied vague notions of public policy to overturn the arbitrator's decision.  Garvey was also easier than the case we review now, because in Garvey the appeals court attempted to substitute its own credibility determination (a factual finding) for that of the arbitrator.   Here, however, we deal with an arbitrator's decision that a term of the collective bargaining agreement “becomes ambiguous.”   This is completely non-sensical because it implies that the term was unambiguous to begin with.

Thus I turn to the application of the majority's new standard.   Arbitrator Glazer began his analysis by quoting the relevant pay raise parity language in Article 35 of the collective bargaining agreement:  “Bargaining unit members will receive the same cost of living [“COLA”] increases paid to other MFR employees pursuant to the directive of MFR's funding source.”   Arb. Op. at 7. As the majority correctly points out, the plain meaning of this parity provision does not jump off the page.   Because MFR's “funding source” is the federal government, Article 35 clearly requires union/nonunion parity as to government-funded cost-of-living increases.   But Article 35 says little about union/non-union parity as to employer-funded cost-of-living increases, the key issue here.   Had Glazer found that the Article 35 parity clause was inherently ambiguous, then there is some merit to the majority's argument that this finding, even if erroneous, is unreviewable inasmuch as it represents an arbitrator's attempt to “arguably construe” the collective bargaining agreement.   But Glazer ruled, instead, that Article becomes ambiguous “because of [MFR's] prior decision to characterize both its [employer] payment and its payment from the federal funding source as ‘COLA.’ ” Arb. Op. at 8. This is where Glazer's opinion is woefully flawed.

First, to say that a clause becomes ambiguous makes little sense in the realm of contract interpretation, because it implies that the clause was un ambiguous to begin with.   And if that were the case, then Glazer would have no business bringing in extrinsic evidence to construe it.   The majority views Glazer's use of the phrase “becomes ambiguous” as either a slip of the pen or a slip of thought.   Maj. Op. at 755.   This may or may not be true, but I am less willing than the majority to give the arbitrator the benefit of the doubt.   The majority opines that either way, Glazer was faced with “contractual silence,” and thus it was proper for him to look for “other indicators of meaning.”   Maj. Op. at 755.   Another plausible reading of Glazer's opinion, however, is that he viewed-rightly or wrongly-the Article 35 plain language as unambiguous, and then went beyond the scope of his authority in subsequently “rendering” it ambiguous.

Furthermore, even if the majority is correct that Glazer was faced with “contractual silence,” that does not automatically mean his subsequent interpretation of this silence must evade our review.   In this case, cost-of-living increases from sources other than the “funding source”-here, the federal government-are simply not provided for by the collective bargaining agreement, and so it is not at all clear why union employees should have a right to such a benefit in the absence of express language creating that benefit in the agreement itself.   In other words, Article 35 does not require further construction as to cost-of-living increases from non-federal sources.   To say that Glazer was “arguably construing” something that requires no further construction is to say that Glazer was acting beyond the scope of his authority as an arbitrator.   Contractual silence does not always equal contractual ambiguity.

Second, the alleged “past practice” that Glazer uses to render the Article 35 parity provision ambiguous, Arb. Op. at 8, is actually a “past memorandum” discussing raises for contract year 2002 (as opposed to contract year 2003, the dispute over which spawned the arbitration at issue in this case).   And yet the memorandum clearly states that the 2002 raises are being given on a “one-time, non-precedent setting basis.”   In light of this language, for Glazer to have used the 2002 memorandum as a precedent to render Article 35 ambiguous simply contravenes common sense.   In my opinion, then, Arbitrator Glazer clearly acted beyond the scope of his authority in finding new meaning in Article 35 of the collective bargaining agreement.   Or if the majority prefers, Glazer's actions were procedurally aberrant for those of a professional labor arbitrator.

II.

Let me be clear about what I am not saying.   I am not suggesting that we decide an arbitration case such as this one on the merits.   I am simply suggesting that we affirm the judgment of the district court, which was to vacate the award and remand for further proceedings-i.e., for re-arbitration, or at the very least for a clarification by Arbitrator Glazer.   And I am not suggesting that upon re-arbitration, or upon further clarification, the decision of the arbitrator has to be different.1  And contrary to what the majority suggests, a decision simply to vacate and remand an arbitration award, as opposed to a decision that reverses the award on the merits, is not inconsistent with the high level of deference that we must show to arbitrators' decisions regarding collective bargaining agreements.   The majority reasons that since appellate courts generally “review judgments, not opinions” while performing de novo review, therefore we must assuredly “review outcomes, not opinions” while performing hyper-deferential arbitrator award review.   Maj. Op. at 755-56.   I disagree.   This Court routinely reviews a lower court's reasoning (i.e., process) on de novo review, even if such review has no bearing on our affirmance of the outcome of the case.   Appellate courts constantly write de novo review opinions including such language as:  “While we disagree with the district court's analysis of the _ issue, we nevertheless affirm.”   Why do we do this?   Because process and reasoning are important, as they provide the justification and backbone for outcomes.   And in a hyper-deferential review situation, as opposed to de novo review, process is often the only thing we can review.

Some may think that my position seems contrary to that outlined by the Supreme Court in its Steelworkers Trilogy:

A mere ambiguity in the opinion accompanying an award, which permits the inference that the arbitrator may have exceeded his authority, is not a reason for refusing to enforce the award.   Arbitrators have no obligation to the court to give their reasons for an award.   To require opinions free of ambiguity may lead arbitrators to play it safe by writing no supporting opinions.   This would be undesirable for a well-reasoned opinion tends to engender confidence in the integrity of the process and aids in clarifying the underlying agreement.

United Steelworkers v. Enterprise Wheel & Car Corp., 363 U.S. 593, 597, 80 S.Ct. 1358, 4 L.Ed.2d 1424 (1960).   Here, however, we deal not with a “mere ambiguity” in Arbitrator Glazer's opinion, but rather a completely nonsensical construction.   I do not believe that even the Supreme Court would wish to see such errors go uncorrected, or at the very least, unchallenged.   And the Court is not entirely correct in its prediction that searching review of arbitrators' reasoning will lead them to “play it safe” by issuing summary decisions unsupported by any writing at all, see also Maj. Op. at 756, for any ability of arbitrators to avoid scrutiny in this way is ultimately circumscribed by the desires of the parties who select and agree to be bound by them.   Losing parties will no longer agree to arbitration schemes if they are never informed in writing, at least on a basic level, why it is that they have lost.   Furthermore, no matter how deferential an “arguably construing” test this Court wishes to apply to arbitral review cases, the test is rendered impotent if the arbitrator issues only an up-or-down decision.   Our deference is not so great that we allow the arbitrator to do all the “construing” in his head only, without committing anything to writing.

My position therefore does not stray from what the Supreme Court noted in Misco:

[I]n the very rare instances when an arbitrator's procedural aberrations rise to the level of affirmative misconduct, as a rule the court must not foreclose further proceedings by settling the merits according to its own judgment of the appropriate result, since this step would improperly substitute a judicial determination for the arbitrator's decision that the parties bargained for in the collective-bargaining agreement.   Instead, the court should simply vacate the award, thus leaving open the possibility of further proceedings if they are permitted under the terms of the agreement.   The court also has the authority to remand for further proceedings when this step seems appropriate.

484 U.S. at 40 n. 10, 108 S.Ct. 364.   This case presents us with such a rare instance, and I would thus vacate and remand for further proceedings. See also Patten v. Signator Ins. Agency, Inc., 441 F.3d 230, 236 (4th Cir.2006) (vacating and remanding an award because the arbitrator “simply amended or altered the agreement, and thus he acted without authority”);  Cytyc Corp. v. DEKA Products Ltd. P'Ship, 439 F.3d 27, 34 (1st Cir.2006) (noting that an award “unfounded in reason and fact” could properly be vacated even under the highly deferential standard of arbitrator review);  Brentwood Medical Associates v. United Mine Workers of America, 396 F.3d 237, 242 (3d Cir.2005) (while circuit court review of arbitration awards is limited to asking whether the parties “got what they bargained for,” this includes inquiry into whether the arbitrator produced a “rational award”);  Int'l Union of Operating Eng'rs, Local 139, AFL-CIO v. J.H. Findorff & Son, Inc., 393 F.3d 742, 745 (7th Cir.2004) (courts must only enforce “intellectually honest arbitral decisions”).

An arbitrator who writes a ten-page opinion containing “all the hallmarks of interpretation,” Maj. Op. at 754, may indeed be “arguably construing” the collective bargaining contract, but he could still be guilty of a procedural aberration by virtue of his irrational reasoning, reasoning that seriously calls into question whether or not he was acting within the scope of his authority.   In other words, no matter how deferential is our review on the merits, we may still demand rational thought from the arbitrator in reaching those merits.   Otherwise it strikes me as no review at all.   See Cytyc, 439 F.3d at 32 (holding that while the “authority of a federal court to disturb an arbitration award is tightly circumscribed,” such awards are not “utterly impregnable”);  Brentwood Medical, 396 F.3d at 244 (noting that courts are “neither entitled nor encouraged to simply ‘rubber stamp’ the interpretations and decisions of arbitrators”) (Ambro, J., dissenting).

I therefore respectfully dissent.

I concur in parts I and II.A of the majority opinion.   The separation of instances of “procedural aberration” from substantive review of the merits and the recognition that in rare and egregious cases the courts must find that arbitral merits decisions depart so far from the contract language as not to amount to contract construction provide useful guidance for consideration of awards.   I thus do not understand the majority opinion to limit review to “procedural aberration,” as the dissent seems to indicate.   I also agree with many of the majority's observations about the arbitrator's decision in Part II.B. In particular, I believe that the majority correctly characterizes the issue with respect to this award as one of substance, not procedure.   I disagree, however, that application of the analytical framework outlined by the majority for reviewing the substantive merits of arbitration awards results in enforcement of this award.   While certainly the text of the decision reads like the arbitrator is attempting to construe or interpret the contract, what the arbitrator in fact does is ignore the plain language of the contract limiting the parity requirement to increases funded by the federal funding source and bind the parties to past practice despite the contract's clear language to the contrary.   In my view, the window dressing of nice opinion language should not obscure the absence of underpinnings for the decision.   And, unlike the majority, I find it difficult to characterize silence as ambiguity.   I would place this decision in the category of those so “ignor[ant]” of the contract's “plain language,” Misco, 484 U.S. at 38, 108 S.Ct. 364, “as to make implausible any contention that the arbitrator was construing the contract.”  (Maj. Op. at 753.)   For this reason, I would affirm the district court.

FOOTNOTES

1.   This is not an absurd result.   In sentencing cases, for example, we are often fully aware that a trial judge may issue exactly the same sentence, or perhaps even a higher one, on remand for re-sentencing.   What we are typically asking for in such cases is that the trial judge simply clarify his reasons for issuing the sentence (e.g., that he was treating the Sentencing Guidelines as advisory only, or that he had assessed all of the § 3553 factors), not that the trial judge substitute our reasons for his own.

MARTIN, J. (pp. 757-60), delivered a separate opinion concurring in part and dissenting in part, in which CLAY and GILMAN, JJ., joined with GIBBONS, J. (pp. 760-61), also delivering a separate opinion concurring in part and dissenting in part, in which BATCHELDER, J., joined.