Reasons are critical to justice and the rule of law. They connect discrete holdings and show that a judge or arbitrator’s decision is not merely a gut reaction to one set of facts on a given day. Reasons embody and transmit the law. They state the principles that separate a society built on whimsy or arbitrary power from one grounded on a core set of written values. Reasons explain decisions, and they should legitimate them.
Traditionally, American arbitrators were not required to give reasons unless the parties requested them. As arbitration has nosed its way into a wider range of disputes, however, reasons have become more important. Today most arbitration rules require a reasoned award, one with true explanations. And that is what most parties expect, too.
Unfortunately, arbitrators sometimes fail to provide sufficient reasons for their decisions. But they alone should not take all the blame. Reviewing courts too often do not recognize and vacate unreasoned awards. These failures are rooted in three crucial opinions: an 11th Circuit opinion in Cat Charter, LLC v. Schurtenberger,1 a Fifth Circuit follow-on opinion in Rain CII Carbon, LLC v. ConocoPhillips,2 and a Second Circuit opinion in Tully Construction Company v. Canam Steel Corporation.3 None of the three awards contained anything recognizable as reasoned. Yet all three courts confirmed the unreasoned awards as if they are reasoned — with the Cat Charter opinion blazing the trail for the other two circuits.
The Cat Charter award. The Cat Charter award is the basis for the most influential appellate opinion in the United States on reviewing awards for reasons.4 A Massachusetts couple, the Ryans, retired to Florida and asked a boatbuilder, Walter Schurtenberger, to build them a boat for $1.2 million. But costs overran the budget. Even after Schurtenberger had collected $2 million from the Ryans, the boat still was not finished. Schurtenberger increased his estimate to $2.6 million. The Ryans thereupon sued.5 Schurtenberger responded that costs had risen because the Ryans kept changing their plans and that they had in fact approved the increases.6
The Ryans brought six claims, including breach of contract and fraud. The parties requested a reasoned award. The panel found for the Ryans on two claims, neither of them fraud. It awarded $1,934,555.00 in actual damages.7
The award did not discuss any details of the Ryans’ six claims, nor the facts, nor the law. It gave no indication that the Ryans felt they were defrauded. It did not discuss the respondents’ counterclaims or their affirmative defenses. The entire substantive discussion stated, in two one-sentence paragraphs, that the Ryans won on each of two claims “by the greater weight of the evidence”; on other claims and defenses, all the award said was that “[a]ll other claims of the Claimants are hereby denied. All counter-claims of the Respondents . . . are denied.”8 That’s it. Some reasons!
The Rain CII Carbon award. A chemical products supplier, Rain CII Carbon, and ConocoPhillips (Conoco) signed a contract for Rain to supply Conoco with an industrial product called green anode coke. Conoco argued it paid an above-market price. The arbitration was a baseball arbitration, so the sole arbitrator chose between the parties’ positions. Rain won: The arbitrator decided to leave the existing contract formula in effect. He awarded Rain $17,702,585.33 (and his ruling will pay Rain many more millions above what Conoco wanted to pay moving forward).
The Rain CII Carbon award was five pages longer than Cat Charter’s. Yet all it offered as a “reason” was a short, single-paragraph summary of each side’s position and a third paragraph saying the formula price stayed in effect. It did not say why Rain won. The hearing was a battle of experts, but the award did not name them or say which testimony the arbitrator found credible.9
The Tully awards. Neither the first Tully award nor an on-remand revised award was reasoned. Tully Construction Company was refurbishing the Whitestone Bridge in New York City. It had various disagreements over steel supplied by subcontractor Canam Steel Corporation. Tully pled nine claims, Canam seven counterclaims. Hearing the 16 claims took a sole arbitrator 17 days. The parties filed over 800 exhibits.
The first unreasoned Tully award. The first award was merely a “list” award. It registered victory or defeat on each claim and, for victories, listed a damage amount but nothing more. It contained no reasons for liability or the damage amount. On the nine claims that failed, the arbitrator simply wrote “0.00.” On the seven claims that succeeded, he listed a dollar amount but no explanation of why the prevailing party should receive it.
The list award was only the start, not the end, of the arbitrator’s resistance to explaining himself. Canam Steel, unhappy to find itself ordered to pay a net $6.5 million, wrote the arbitrator challenging the award’s total lack of reasons. The arbitrator responded that his award was reasoned — stating only that it had “sufficiently and specially incorporate[d] all credible evidence adduced during the hearings, detailed the liability for each item of claim and counterclaim, and, as such, [was] a ‘reasoned award.’”10
This dismissive language makes no sense. The award mentioned neither evidence nor liability. If silence “incorporates” evidence, the term incorporate is meaningless. If silence “details” liability, when an award says nothing at all about liability except who wins, then that term, too, is meaningless.11
The still unreasoned second Tully award. The trial court remanded the first award for clarification.12 The arbitrator thereupon expanded his short list award into an 11-page award. But he still did not explain his thinking. All he added was a page and a half of undisputed background; a short statement for each of the 16 claims and counterclaims, each beginning with a one-sentence boilerplate attestation that he reviewed the “relevant, related, or both, information” and that this “information” justified the resolution that followed; then a one-sentence paragraph of great generality listing a few exhibit numbers or record pages cited by the moving party on the claim or counterclaim, followed by a similarly cursory sentence about the other side’s position. He finished each discussion with a boilerplate sentence that the “credible preponderance” of “testimonial” and “documentary evidence” did, or did not, establish the claim or counterclaim. On seven claims it did; on nine it did not. It remains impossible to know what this arbitrator thought about the parties’ arguments and evidence. All he disclosed is who won on each claim.
This hearing took 17 days. Surely the arbitrator had some reasons for ruling as he did!
No one reading any of these three awards can point to words that explain the decision. Yet the three federal appellate courts reviewing them were happy to guess at reasons, supply their own reasons, confuse contentions or vague conclusions about meeting burdens of proof with reasons — in short, to do everything possible to affirm.
If these awards trouble arbitrators, their confirmation by three federal courts of appeals should trouble judges.
The 11th Circuit confirms in Cat Charter. The test created by the 11th Circuit in Cat Charter will confirm almost all awards regardless of missing reasoning. Yet in spite of its gaping flaws, it has become the dominant judicial approach to testing whether awards have true reasons.
The Cat Charter trial judge correctly vacated the uninformative award for failure to provide reasons.13 Astonishingly, the appellate court reversed and remanded for the trial court to reinstate the award.14 It had four unpersuasive arguments for doing so.
First, the court defined reasoned awards through some limited federal case law about a “spectrum of increasingly reasoned awards,”15 with a standard award on the low end of the spectrum (requiring no explanation and not being reasoned), and an award that calls for findings of fact and conclusions of law on the upper end (requiring the most amount of explanation and being the most reasoned).16 According to Cat Charter, anything more than a “simple result” should be treated as reasoned: “A reasoned award is something short of findings and conclusions but more than a simple result.”17 Taken literally, that standard will bless any award saying that “Party A wins by the weight of the evidence,” “Party A’s evidence [witnesses, exhibits, etc.] is more credible,” or “Party B met its burden of proof.”
The problem with spectrum analysis, which puts awards on a single line of increasing reasons, is that it assumes that a reasoned award need not be fully reasoned. In practice, both “reasoned awards” and “findings and conclusions” must be adequately reasoned. The difference between these two forms is in their structure, not whether they are fully, only halfheartedly, or even infinitesimally reasoned. A reasoned award is usually a narrative award, one that often addresses a full set of facts or a broad legal point in a single paragraph. Findings and conclusions, by contrast, typically separate the fact and legal sections into short numbered paragraphs, each discussing just one small piece of evidence or law. What a reasoned award covers in a dozen narrative paragraphs, findings and conclusions may cover (usually but not always in a bit more detail) in a hundred or more very short paragraphs. Findings and conclusions naturally take more time to write, are accordingly more expensive for the parties, often read awkwardly, and not surprisingly are rarely requested.
Because spectrum analysis incorrectly assumes at the outset that these two award forms are arranged on a line of increasing reasoning, rather than just a line of increasing detail, it has to posit a middle category of awards that are not as well reasoned as findings and conclusions. It assumes that there are awards with no reasons, reasoned awards with a sort of half reasoning, and findings and conclusions that are fully reasoned. But arbitration practice and rules do not share this understanding at all. They use “reasoned award” as a term for a fully reasoned award. Courts are supposed to read words in their plain meaning, and there is nothing ambiguous about “reasoned award.” They have no excuse for not understanding this.
Judges rely on this concept of spectrum analysis to call conclusory statements — for instance, that a party wins by the weight of the evidence or wins because it has more credibility — “reasoned.” But this standard fails to effectively distinguish unreasoned from reasoned awards.18Every party that wins on a claim or defense satisfies this test. A test for reasons that all winners pass is meaningless.
The assumption that awards lie on a spectrum in essence lets courts replace “reasoned award” with “findings and conclusions” as the sole guardian of truly reasoned awards. This is an unwarranted judicial colonization of arbitration practices. The arbitration rules that applied in Cat Charter were a prior version of today’s Rule R-48 on “Form of Award” in the American Arbitration Association’s (AAA’s) Commercial Arbitration Rules. The only award form that rule actually describes is a “reasoned award,” the form that applies if the parties or the arbitrator choose it as the parties did in Cat Charter. Most AAA arbitrators in practice give reasons, sometimes brief reasons, even if the parties make no request. But, as everybody in the arbitration community knows, if a reasoned award is not selected, the arbitrators are free to write a standard award with no reasons. Nothing in these rules or practices authorizes courts to demand that the parties request findings and conclusions before the courts will require the arbitrators to issue a truly reasoned award.
The AAA’s international division, the International Centre for Dispute Resolution, provides in article 33.1 of its International Dispute Resolution Procedures, which applied in Rain CII Carbon, that “The Tribunal shall state the reasons upon which an award is based,” unless the parties ask for a different form. There was no excuse for the parties not getting a reasoned award. Nor did anything in the rules or facts authorize the Fifth Circuit to confirm the unreasoned award because the parties did not ask for findings and conclusions.
Finally, Rule R-44 of the AAA’s 2010 Construction Rules, which applied in Tully, let the parties choose between forms, including a “reasoned award” and “findings and conclusions.” The Tully parties chose the former. Nothing in these rules or the record gave any indication that the Tully parties actually expected to get either a list award with no reasons (the arbitrator’s first attempt) or a longer, cryptic description that still failed to describe his reasons (his second).
But the Cat Charter court did not stop with spectrum analysis. As a second argument, the court unpersuasively crafted a definition of “reasoned” by combining dictionary definitions into the following:
A “reasoned award” [is] an award that is provided with or marked by the detailed listing or mention of expressions or statements offered as a justification of an act — the “act” here being, of course, the decision of the Panel.19
Unfortunately, this test is so all-encompassing that the slightest statements will satisfy it. A “mention” of an “expression” offered as “justification”? That shelters “You win on the weight of the evidence.” It probably even shelters “After hearing the evidence, I just feel that claimant is right in this case” or “Well, after that, I lean toward respondent.” This definition requires nothing specific about the evidence introduced, the legal principles involved, or the parties’ arguments.
Third, the court leaned on credibility: “Put simply,” the court boldly and without any foundation pronounced, “the controversy here turned primarily upon credibility determinations made by the Panel.”20 And, it later embellished, “[i] n essence, this dispute was a swearing match, and its resolution necessarily depended on credibility determinations made by the arbitrators.”21 But how can the court know that? The Cat Charter award says nothing about credibility. Neither “credible” nor “credibility” nor any synonyms appear in the award. The award does not discuss witnesses or testimony. Yet somehow the court of appeals could divine that the arbitrators just made some global credibility judgment? This is arbitrary judicial decision-making indeed.
One fears that perhaps the judges knew and respected the arbitrators and, believing them to be sophisticated, skilled arbitrators, gave them a pass even though they did not really explain their award. But this would be a rule of men and women, not of law. A court should never affirm or confirm a decision below just because it trusts the judge or arbitrator without any regard for the merits of the challenge. The only competent evidence about this decision is the award, and it doesn’t contain any reasons.
Finally, even though the parties had requested reasons, the court blamed them for not asking for findings and conclusions. “[H]ad the parties wished for a greater explanation, they could have requested that the Panel provide findings of fact and conclusions of law.”22 Yet, as described above, the parties had agreed on a reasoned award, a term that is part of the AAA rules that governed in Cat Charter. There is nothing ambiguous about the word “reasoned.” The Eleventh Circuit had no right to alter the word’s meaning just because the parties did not request a more detailed form that courts sometimes use. In essence, they blamed two parties in arbitration, a process that is designed to get away from excess formalities and rigidities in court practice, for not using a secret judicial code. There is no logic here: “Reasoned” cannot mean that you will receive reasons only if you ask for findings and conclusions.
The Fifth Circuit confirms in Rain CII Carbon. The Fifth Circuit applied the Cat Charter test to confirm in Rain. Unlike in Cat Charter, the Rain trial court confirmed the award. It decided that, given the two paragraphs summarizing the parties’ arguments followed by the one-sentence finding for Rain, “one could certainly distill some level of reasoning” from the award.23 But if distillation is needed, that is judicial guesswork, not arbitrator articulation.
The Fifth Circuit, affirming, also fixated on the fact that the award listed contentions before announcing the winner, complaining that Conoco did not acknowledge the contentions:
Conoco ignores that the preceding paragraph thoroughly delineates Rain’s contention that Conoco had failed to show that the initial formula failed to yield market price, a contention that the arbitrator obviously accepted. Conoco would have this court vacate the arbitration award merely because the arbitrator did not reiterate this reason in the following paragraph.24
Yet the contentions summarized in the award describe two ships passing in the night in such general terms that they cannot explain why the arbitrator boarded one ship and not the other.25
The Fifth Circuit said the arbitrator must have obviously agreed that Conoco lost because it “failed to show that the initial formula failed to yield market price,” but that is just a recital of the ultimate conclusion, not the reasons for it. A reasoned award would have explained why the arbitrator rejected Conoco’s “show[ing]” and accepted Rain’s. The parties’ briefs list multiple arguments on why the contract formula did, or did not, generate a market price.26 Those underlying disputes required three days of evidence. What did the arbitrator find persuasive in them, and why?
Rain argued that Conoco’s novice expert didn’t know what he was doing.27 Did the arbitrator agree? Did he find the expert generally competent but the substance of his work in this case unconvincing? Or could the arbitrator not decide between experts, but nonetheless found Rain’s position overall more credible for other reasons? Or did the arbitrator simply believe that Conoco failed to sufficiently respond to Rain’s detailed criticisms of its expert?
The court wrote as if the short contention paragraphs contained the arbitrator’s thinking and reasoning, characterizing Conoco’s complaint as being that the arbitrator “did not reiterate this reason [implicitly, the arbitrator’s reason] in the following paragraph.”28 But the court was wrong — the arbitrator did not actually write the contentions himself. He simply lifted them from one side’s proposed award.29 Moreover, he plucked the language from the draft of the losing party, Conoco! If he agreed with Conoco’s phrasing, indeed, on almost all of the award, why did he declare Rain the winner?
In asking for a reasoned award, the parties told the arbitrator to decide which side prevailed and explain why. They did not ask him to repeat party positions they surely already knew better than he did (and always will) and then just declare a winner. Yet this was all the “explanation” they got:
Based on the testimony, evidence, exhibits, arguments, and submissions presented to me in this matter, I find that the price formula . . . shall remain in effect for the balance of the term as stated in the contract.
A self-attestation by an arbitrator, or a judge, that he did his job thoroughly in picking a winner is not a reason.30 “Trust me” cannot be a reason in a principled system of law.
The Second Circuit confirms in Tully. To its credit, the Tully trial court vacated the first award by noting its total absence of reasons.31
When that skimpy, barely page-and-a-half award returned after remand as an 11-page award, however, the same judge confirmed it.32 Yet all the arbitrator had done on remand was insert an initial sentence on each claim attesting that he had done his job, two one-sentence paragraphs each vaguely summarizing one party’s position in terms too broad to make a fair decision on that claim, and then end with a boilerplate conclusion that the claimant either did, or did not, meet its burden. He gave no hint why.
The Second Circuit, however, held that the new award contained “key factual findings” and explained “why Tully was entitled to damages on some claims and not others.”33 But where?
***
Stunningly, the Cat Charter test — quickly taken up in Rain CII Carbon and in two Second Circuit opinions, including the Tully opinion — is the leading test for reasons in American law34 There was not a large body of law on reasoned awards before Cat Charter. Since then, many federal and state courts have adopted its standard.35
Despite Cat Charter’s failure to require reasoned awards, it nonetheless seems to have encouraged losing parties to challenge awards as lacking reasons. Perhaps lawyers previously assumed courts would be so deferential to the arbitrators that appeal was futile. Even though the opinion did not devise a good test for reasoned awards, Cat Charter at least did publicize that a failure to provide reasons, when they are requested, exceeds arbitral powers under FAA section 10(a)(4) and thus provides another basis for losing parties to seek vacatur. In that way, the opinion has helped advance the law by publicizing the right to reasons, even if its ineffective test does not help actually realize that right.
Creating a Test for Whether Awards Are Reasoned
Unreasoned arbitration awards fall into patterns.36
Announcement awards simply announce the winner, perhaps with a little introductory or background language, but nothing on liability or damages.
Attestation awards contain a bit more, attesting that the arbitrator heard the evidence and mastered the record (often mentioning specific pleadings, exhibits, and pre- and post-hearing briefs when available).
Burden of proof and weight of evidence awards, like the Cat Charter award, along with credibility awards, are just boilerplate to say that the prevailing party won. But this provides no real explanation, because every winning party — by definition — has met the burden of proof, prevailed on the weight of the evidence, and presented a more credible case. The question for a reasoned award is why this is so.
Contention and issue-spotting awards state the parties’ positions, usually in summary form, but do not explain why the arbitrator chose one side’s position over the other’s. The Rain award is a classic example.
Evidentiary list awards list exhibit numbers and testimony pages without discussion, as in the second Tully award. They may even detail a bit of the evidence. But they do not give the arbitrators’ analysis or say what compels the result.
Finally, volumetric awards include enough mass — often long contentions or listings of evidence — that a court decides there must be a “there there.” That the award in Tully 2 was notably longer likely helped both reviewing courts conclude that it contained sufficient reasons. Yet even a minimal substantive analysis by either court should have shown that the arbitrator still had not explained his decisions.
These common patterns lend themselves to a test that effectively identifies inadequately reasoned awards. A definition good enough to police awards for reasons — without involving courts in judging the merits — needs to clearly state that arbitrators must explain their decision on each potentially dispositive claim, counterclaim, defense, and remedy.37
Here is a detailed example of such a test:
A reasoned award should explain who won and why by clearly explaining:
The better practice is for a reasoned award to use a level of detail that addresses, even if briefly, the major party arguments on both sides of the issues above, and explains at least briefly the resolution of cumulative alternative claims and defenses as well.39
A number of the clauses in this definition — including the ones about jurisdiction, rule of law, choice of law, and burden of proof — are based upon cases vacating awards on these issues for being insufficiently explained. A simplified single-sentence test, losing something in comprehensiveness but perhaps gaining in clarity, would be:
A reasoned award will clearly explain why the arbitrator ruled for one side and rejected the other’s position on each claim, counterclaim, defense, and remedy that, if granted, would have altered all or part of the outcome.
An effective addition would go further by explicitly rejecting the most common forms of unreasoned awards. An agreement, statute, or opinion could say this:
Awards that merely announce winners, attest that the arbitrators reviewed all the facts and arguments and the like, proclaim who prevailed by the weight of the evidence or burden of proof or whose case was more credible, or list the parties’ contentions and then announce a winner, are not reasoned. Awards are not reasoned merely because they are very long and describe a lot of evidence, or because they list exhibit numbers and transcript pages and portions of pleadings, if they do so without explaining how the cited material justifies the outcome. A reasoned award must explain in the arbitrators’ words the factual and legal disagreements presented by the parties and how the arbitrators resolved those disagreements.40
This definition, just like the first one, can be shortened. Any version would be a vast improvement over the Cat Charter test.
Some judges might think: “If parties want reasons, let them go to court.”
But this institutional diminishment of arbitration ignores its important role, especially in light of the Supreme Court’s emphatic promotion of the practice in recent decades.
The Supreme Court may once truly have believed that parties sacrifice their right to a quality decision when they choose arbitration.41 But that changed in the 1980s, when the Court claimed to espy a national policy favoring arbitration in the Federal Arbitration Act. (It is not clear where this policy had been hiding from the Court from 1953, when it was so highly and publicly critical of arbitration, to the early 1980s.)42 Given that reality, courts ought not disparage arbitration.
Some have argued that the Court only embraced arbitration in its desperation to save federal courts from an unmanageable flood of litigation.43 That is a highly jaundiced view of judicial motivation. It ascribes the press of cases as the Court’s sole motivation in a wide range of arbitration cases without giving weight to other considerations, including most importantly the Court’s legal analysis. But arbitration offers much more than just a pressure valve for our court system.
Arbitration is not a single process. Its flexibility offers many things courts cannot. A large part of arbitration’s attractiveness is that parties have a real voice in choosing their “judges.” Parties can seek out arbitrators known for skill in decision-making and managing a hearing, and also look in the pool of arbitrators for those with specialized knowledge and industry background. Because parties pay arbitrators to have enough time to handle the dispute, the arbitrators owe the parties the time to move an arbitration as fast as the parties want and to give it as much attention as the parties desire. Arbitrators have no justification for being too busy to hear or decide motions for months on end, as courts too often are, or to take months or even years to render an opinion on the merits after hearing the evidence. Arbitrators also are in a much better position to guarantee truly reasoned awards. They should never say that the parties do not need much explanation in an award because they already know the facts.44
When parties indicate they want a reasoned award, directly or by choosing rules that require them, they must get them. Reasons should reveal whether the process was fair and the parties heard. Reasons are particularly important to losing parties, who naturally tend to take their loss hard and to be more suspicious that something went wrong in the arbitration. In addition, being forced to write out opinions improves the decision process — it clarifies arguments, and sometimes even requires that arbitrators change their minds.45
When courts enforce the requirement of reasons in arbitration, they uphold the rule of law. Reasons let parties see that their rights were determined under the law’s dictates.
The Cat Charter standard (and its progeny) is a failure. It does not effectively ensure anything close to a truly reasoned award. It confirms all kinds of sloppy awards that merely say the winner won. Courts should foster our national policy in favor of arbitration by holding arbitrators to their responsibility to give reasons when reasons are due.
JOHN BURRITT MCARTHUR was one of the first trial lawyers at Houston’s Susman Godfrey LLP, where he practiced from 1983 to 1992, and later a name partner in San Francisco’s Hosie McArthur LLP. He has focused his time on serving as an arbitrator since 2009, arbitrating almost 150 arbitrations to award. His most recent publication, The Reasoned Arbitration Award in the United States: Its Promise, Problems, Preparation, and Preservation, was published in October 2022 by Juris Publications.
