Picture two companies that signed the same contract on the same afternoon and later fell into the same fight over the same unpaid invoice. One ends up in front of a federal jury; the other ends up in front of a single arbitrator in a downtown conference room. Both walk out with a piece of paper that says the loser owes the winner $4.2 million. Here is the strange part, and the reason this entire article exists: those two pieces of paper are wildly different animals. The jury verdict can be appealed, dissected, reversed for legal error, and trimmed by a remittitur. The arbitration award, by contrast, is almost bulletproof. A judge who is convinced the arbitrator got the law flat wrong will, in the overwhelming majority of cases, still sign an order turning that award into an enforceable judgment.

That asymmetry is the most important "damage statistic" in arbitration, and it is not a number at all. It is a structural fact. When people ask "how much do arbitrators award?" the honest answer is that the dollar figure matters far less than the durability of the figure. A modest award you cannot escape is worth more than a generous verdict your opponent can tie up in appeals for three years.

This article is a tour of how damages work in arbitration from front to back. We will cover what arbitrators are actually allowed to award (compensatory damages, lost profits, interest, sometimes punitive damages, and attorney's fees), the surprisingly tricky rules about punitive and exemplary damages, and then the part most guides skip: what happens to that award when one side runs to court to either confirm it or kill it. We will be candid about the so-called "statistics," because the temptation to invent a tidy national average is strong and the data simply does not support one. By the end you should be able to read an arbitration clause and predict, in broad strokes, what kinds of money are on the table and how hard it will be to make that money stick. If you want the procedural mechanics that surround all of this, pair it with our companion guides on the intricacies of AAA commercial arbitration and arbitration as a form of alternative dispute resolution.

A Note Before We Talk Numbers: Why "Damage Statistics" in Arbitration Are Genuinely Hard

Let us deal with the elephant in the room first, because it shapes everything that follows. There is no reliable, comprehensive, public database of arbitration award amounts the way there is for, say, federal court judgments. This is not an oversight. It is baked into the design of arbitration itself.

Arbitration is private. That is one of its selling points. Most commercial awards are confidential by agreement, are never filed in any public court, and disappear into the parties' filing cabinets. An award only becomes visible to the outside world when one party asks a court to confirm it (turn it into a judgment) or to vacate it (throw it out). At that point a redacted or partial version may land on a public docket. But that means the awards we can actually see are a self-selected, non-representative slice: they skew toward disputes contentious enough that the loser fought confirmation. The quiet awards that everyone simply paid never show up in any dataset. Any "average award" computed from confirmation dockets is therefore measuring the wrong population, the way a study of car crashes drawn only from emergency rooms tells you nothing about the average drive to work.

Three categories of "statistics" do exist, and each has real value if you understand its limits.

First, the major arbitral institutions publish operational data. The American Arbitration Association (AAA), JAMS, and FINRA (which runs the mandatory arbitration forum for the securities industry) periodically release figures on caseloads, the number of cases filed by subject matter, median time to disposition, and, in FINRA's case, the percentage of customer claims that result in some monetary recovery. These are honest and useful, but they are about throughput and process, not about the size of awards in your kind of dispute. FINRA, for instance, has historically reported that a meaningful share of customer claims that go to a full hearing end with the customer being awarded something, but those figures move year to year and FINRA's investor-protection context is very different from a commercial contract fight.

Second, academic and government studies exist. Empirical scholars (the work of Professors Christopher Drahozal, Thomas Stipanowich, and the late Lisa Bingham is foundational here), the U.S. Government Accountability Office, and the Consumer Financial Protection Bureau (whose 2015 arbitration study examined consumer-finance disputes) have all crunched samples of real awards. These studies are gold, but they are snapshots of specific sectors (consumer finance, securities, employment) at specific times, and they routinely warn the reader not to generalize.

Third, commercial analytics vendors now sell datasets scraped from confirmation/vacatur dockets. Useful for litigators, but subject to the same self-selection problem above.

So here is our rule for the rest of this article, and the rule any honest practitioner should adopt: we will describe ranges, mechanisms, and tendencies, and we will label every illustrative dollar figure as a clearly invented hypothetical. When someone confidently tells you "the average arbitration award is $X," ask them which population, which year, and which forum, and watch the confidence evaporate. For a broader framework on valuing any dispute before you ever pick a forum, see our guide to evaluating and assessing a civil case.

What Damages Can an Arbitrator Actually Award?

Start with a foundational principle that surprises a lot of newcomers: an arbitrator's power to award damages comes almost entirely from the contract, not from a statute or some inherent judicial authority. Arbitration is a creature of consent. When two parties agree to arbitrate "any dispute arising out of this agreement," they are handing a private decision-maker the authority to resolve that dispute, and the scope of the remedies the arbitrator may grant is defined by (1) the arbitration agreement itself, (2) the rules of any institution the parties chose (AAA, JAMS, ICDR, CPR, FINRA), and (3) the substantive law the contract says applies.

Within those boundaries, arbitrators command a remedial toolkit that looks a great deal like a court's.

Compensatory damages are the core of the system, just as they are in litigation. The animating goal of contract damages is to put the non-breaching party in the position it would have occupied had the contract been performed, the "expectation" interest familiar to every first-year law student from Hadley v. Baxendale, (1854) 9 Ex. 341, and its modern descendants. An arbitrator deciding a contract dispute applies the same measures a judge would: the benefit of the bargain, the cost to complete or repair, the difference between contract price and market price, and so on. For the underlying doctrine, our companion piece on contract damage statistics digs into expectation, reliance, and restitution measures in depth, and the analysis transfers cleanly into the arbitral setting because arbitrators are applying the very same body of substantive law.

Consequential and lost-profits damages are available in arbitration on the same terms they are available in court, which means they are governed by foreseeability (the second branch of Hadley) and by whatever the parties' contract says. This is a place where the contract drafting matters enormously: many commercial agreements contain a "consequential damages waiver" barring recovery of lost profits and similar downstream losses. An arbitrator is bound to enforce that waiver. So the single most consequential damages decision in an arbitration is often made years earlier, by the lawyers who drafted the limitation-of-liability clause.

Interest is a quietly large category. Arbitrators routinely award pre-award interest (running from the date of breach to the date of the award), and they have broad discretion over the rate and whether it compounds. Institutional rules grant this power expressly: the AAA Commercial Rules authorize the arbitrator to award "interest at such rate and from such date as the arbitrator(s) may deem appropriate" (AAA Commercial Rule R-47(d)), and the ICDR Rules similarly empower the tribunal to grant pre-award and post-award interest, simple or compound, as it considers appropriate (Article 34(4), ICDR Rules). Over a multi-year dispute on a seven-figure principal, compound pre-award interest can add a substantial fraction to the total. Post-judgment interest, by contrast, generally tracks the statutory federal or state rate once the award is reduced to a judgment.

Specific performance and other equitable relief are also within reach. If the parties' agreement and the governing rules permit it, an arbitrator can order a party to perform, to deliver unique goods, to transfer property, or to refrain from conduct. Arbitrators can also issue declaratory relief and, increasingly, interim measures.

Attorney's fees and costs deserve their own section, and they get one below, because the default rule in arbitration is different from what most people assume.

The key takeaway is that arbitrators are not playing with some impoverished, watered-down version of remedies. In a typical commercial case they can award nearly everything a court could. The real differences live at the edges, punitive damages and the back-end review, which is where the law gets genuinely interesting.

Punitive Damages in Arbitration: The Mastrobuono Problem

Here is one of the most elegant little puzzles in American arbitration law, and it turns on a single, easily overlooked sentence in a contract.

Punitive damages, also called exemplary damages, are not meant to compensate the winner for a loss. They are meant to punish outrageous conduct and deter its repetition. For most of the twentieth century there was real doubt about whether a private arbitrator, exercising contractual rather than sovereign authority, could impose a punishment at all. Some courts, drawing on the idea that punishment is a public function, said no.

The Supreme Court resolved the central federal question in Mastrobuono v. Shearson Lehman Hutton, Inc., 514 U.S. 52 (1995), and the resolution is worth understanding precisely because it is so often misstated. The Mastrobuonos were investors who arbitrated a dispute with their brokerage and won, including a punitive-damages award. Their brokerage account agreement contained two clauses that pulled in opposite directions. One was a choice-of-law clause selecting New York law, and at the time New York's Garrity rule held that arbitrators could not award punitive damages. The other was an arbitration clause incorporating the rules of the National Association of Securities Dealers, which allowed arbitrators to award any relief a court could, punitive damages included. So which clause wins?

The Supreme Court, in an opinion by Justice Stevens, held that the punitive award stood. The Court read the two clauses together and harmonized them: the New York choice-of-law clause was best understood to incorporate New York's substantive rights and obligations, not New York's special allocation-of-powers rule about what arbitrators specifically may do. Because the contract was at best ambiguous on the punitive-damages point, the Court applied two thumbs on the scale: the federal policy favoring arbitration embodied in the Federal Arbitration Act, and the venerable rule that ambiguities in a contract are construed against the drafter (here, the brokerage that wrote the form). The result is a clean doctrine. Arbitrators may award punitive damages where the parties' agreement, fairly read, permits it, and a generic choice-of-law clause selecting a state whose courts limit arbitral punitive awards will not, by itself, strip the arbitrator of that power.

Two practical lessons fall out of Mastrobuono, and both are about drafting.

First, if a party genuinely wants to keep punitive damages off the table, it must say so explicitly and unambiguously in the arbitration agreement, not hope that a buried choice-of-law clause will do the work. A clear sentence, "the arbitrator shall have no authority to award punitive, exemplary, or multiple damages," is generally enforceable. Courts respect the parties' freedom to limit the remedies their chosen decision-maker can grant; that is the flip side of arbitration being a creature of consent.

Second, and this is the subtle one, the enforceability of a punitive-damages waiver is not absolute. If the underlying claim is one Congress created with punitive or statutory damages as part of a deliberate enforcement scheme, a contractual waiver that operates to gut the statute can run into the "effective vindication" doctrine that the Supreme Court discussed in American Express Co. v. Italian Colors Restaurant, 570 U.S. 228 (2013). The doctrine is narrow after Italian Colors, but it survives for the case where a remedy waiver would effectively forbid the assertion of a federal statutory right altogether. This is a genuinely unsettled, fact-specific corner of the law, and it varies with the statute and the circuit, so flag it and get specific advice rather than assuming a boilerplate waiver is invincible.

A clean (invented) illustration. Suppose Northwind Software, Inc. and Acme Retail Group arbitrate a fraud-in-the-inducement claim under an AAA clause that incorporates the Commercial Rules and is silent on punitive damages, with a generic "governed by the laws of the State of New York" clause. Under Mastrobuono, the arbitrator may award punitive damages if the governing substantive law and the facts support them, because the choice-of-law clause does not, standing alone, take that power away. Now suppose the same contract had instead said, in its own numbered paragraph, "the arbitrator may not award punitive or exemplary damages of any kind." That sentence almost certainly would be enforced, and Acme's fraud claim, however sympathetic, would be capped at its actual losses. One paragraph; a multimillion-dollar swing.

Attorney's Fees, Costs, and Fee-Shifting

In American litigation the default is the "American Rule": each side pays its own lawyers unless a statute or contract shifts fees. That default carries into arbitration, but with a twist that catches people off guard.

An arbitrator may award attorney's fees and costs when one of three things is true: (1) the parties' contract provides for fee-shifting (a "prevailing party" fee clause), (2) the governing statute authorizes fee-shifting (many consumer-protection, civil-rights, and antitrust statutes do), or (3) the applicable arbitration rules or law authorize it. Crucially, several institutional rule sets allow an arbitrator to award fees when both parties have requested them, on the theory that by each asking for its own fees, both sides have consented to the arbitrator deciding the fee question. The AAA Commercial Rules, for example, permit the arbitrator to award attorney's fees if all parties have requested them or if authorized by law or the agreement. So two combatants who each reflexively demand fees in their pleadings may, without quite intending to, hand the arbitrator the power to make one of them pay the other's lawyers.

This matters for the "damage statistics" picture because fee awards can dwarf the underlying claim. In a hard-fought commercial arbitration, fees and costs (including the arbitrators' own compensation and the institution's administrative fees, which are real money in arbitration in a way court filing fees are not) can run to six or even seven figures. When you read about a "$3 million arbitration award," it is worth asking how much of that is principal, how much is interest, and how much is fees, because those components behave very differently and are recovered under very different rules. Our guide to writing a demand letter discusses how to frame a fee-shifting demand at the outset, and the same framing discipline pays off in an arbitration prayer for relief.

The Reasoned Award: Why the "Why" Matters

In court, you almost always get a written explanation: findings of fact, conclusions of law, a reasoned opinion you can appeal. In arbitration, that is an option, not a default. The parties choose, usually in the arbitration agreement or at the preliminary hearing, what form of award they want.

There are three common forms. A standard (bare) award simply states who wins and how much, with no reasoning, the arbitral equivalent of a jury's general verdict. A reasoned award explains the tribunal's reasoning, the closest thing arbitration offers to a judicial opinion. And findings of fact and conclusions of law is a more formal, detailed variant sometimes used in high-stakes matters.

Why does the form of award belong in an article about damages? Because the form you choose has a direct, often underappreciated effect on the size and reviewability of the money. A reasoned award forces the arbitrator to show the work on damages, which both disciplines the calculation and gives a reviewing court something concrete to look at if the loser challenges the award as, say, exceeding the arbitrator's powers. But a reasoned award is also a double-edged sword: by writing down the reasoning, the arbitrator creates a record that the loser can mine for error. A bare award gives a frustrated loser almost nothing to attack, which is precisely why a winner sometimes prefers it. The strategic choice between a bare and a reasoned award is, in part, a damages-protection decision, and savvy parties make it deliberately rather than by default.

The Back End: Confirmation, Vacatur, and the FAA

Now we get to the heart of why arbitration awards are so durable. Winning the award is only half the job. To make it enforceable, the prevailing party converts it into a court judgment through confirmation. The losing party's only real counter-move is to seek vacatur, asking a court to throw the award out. And here the Federal Arbitration Act (FAA), 9 U.S.C. §§ 1 et seq., sets the rules, the rules are deliberately stingy toward the loser.

Section 9 of the FAA tells a court to confirm an award "unless the award is vacated, modified, or corrected" under the Act's narrow grounds. Confirmation is close to a rubber stamp; the court is not supposed to re-examine the merits.

The vacatur grounds live in 9 U.S.C. § 10(a), and they are short enough to quote in substance. A federal court may vacate an award only:

  1. where the award was procured by corruption, fraud, or undue means;
  2. where there was evident partiality or corruption in the arbitrators;
  3. where the arbitrators were guilty of misconduct in refusing to postpone the hearing upon sufficient cause shown, or in refusing to hear pertinent and material evidence, or of any other misbehavior by which the rights of a party have been prejudiced; or
  4. where the arbitrators exceeded their powers, or so imperfectly executed them that a mutual, final, and definite award upon the subject matter submitted was not made.

Read that list again and notice what is not on it. There is no ground for "the arbitrator misapplied the law." There is no ground for "the arbitrator misunderstood the contract." There is no ground for "the damages award is too big" or "the evidence did not support this number." A court reviewing an arbitration award is policing the integrity of the process, fraud, bias, gross procedural unfairness, and a decision-maker who wandered outside the questions the parties submitted, not the correctness of the result. The companion grounds for modification in 9 U.S.C. § 11 are similarly narrow, reaching evident mathematical miscalculations, awards on matters not submitted, and matters of form. An arbitrator who adds wrong (writes "$1,200,000" where the math clearly yields "$1,020,000") can be corrected under § 11; an arbitrator who simply values the damages differently than the judge would cannot.

This is the legal engine behind the durability we opened with. An arbitrator can get the damages analysis "wrong" in a way that would be reversible error on appeal from a trial court, and the award still stands. That is not a bug; it is the bargain the parties struck when they chose a final, private decision-maker over the appellate system.

Hall Street Associates v. Mattel: You Cannot Buy More Review

For years, sophisticated parties tried an end-run. If FAA review is too narrow, they reasoned, why not write our own broader standard into the contract? A clause might say "the award shall be subject to vacatur if the arbitrator's findings of fact are not supported by substantial evidence or the conclusions of law are erroneous," essentially manufacturing appellate-style review by agreement.

In Hall Street Associates, L.L.C. v. Mattel, Inc., 552 U.S. 576 (2008), the Supreme Court shut that door, at least under the FAA. The parties there had agreed to exactly such an expanded-review clause. The Court, per Justice Souter, held that the grounds for vacatur and modification in §§ 10 and 11 are exclusive, parties cannot contract for broader judicial review of an award under the FAA. The statutory list is the whole list. The Court reasoned that the FAA's text and its design, a streamlined, summary process, would be undone if litigants could bolt on full appellate review by private agreement. Section 9 says the court "must" confirm unless one of the enumerated grounds applies, and the Court took "must" seriously.

Hall Street is a load-bearing case for any discussion of arbitration damages because it tells you, definitively, that the way to control the risk of an aberrant award is not through after-the-fact judicial review. You cannot fix a runaway damages number by writing a fancy review clause; the Supreme Court has foreclosed it. The real levers are on the front end: choosing your arbitrator carefully, capping or excluding categories of damages by contract, requiring a reasoned award, and, if you truly want appellate review, building a private appellate process. Several institutions now offer exactly that, the AAA/ICDR Optional Appellate Arbitration Rules and JAMS's optional appeal procedures let parties agree to a second-tier arbitral panel that reviews the first award for legal error. Hall Street left that route open because it keeps the dispute inside arbitration; what the FAA forbids is conscripting the courts into expanded review. If certainty and reviewability matter more to you than speed and finality, that trade-off is a reason some disputes belong in court in the first place, a calculus our overview of administrative and government litigation damages touches on from a different angle.

"Manifest Disregard of the Law": A Doctrine in Limbo

No discussion of vacatur is complete without the ghost that haunts it: manifest disregard of the law. For decades, courts recognized a judge-made, non-statutory ground for vacating an award where the arbitrator knew the governing law and deliberately ignored it, not mere legal error, but a conscious refusal to apply law the arbitrator understood to control. The phrase traces to dictum in Wilko v. Swan, 346 U.S. 427 (1953).

Then came Hall Street, which said the statutory grounds are exclusive. That immediately raised the question: did Hall Street abolish manifest disregard along with contractual expansion of review? The Supreme Court has pointedly declined to resolve it. In Stolt-Nielsen S.A. v. AnimalFeeds International Corp., 559 U.S. 662 (2010), the Court noted in a footnote that it did not need to decide whether manifest disregard survives as an independent ground or persists, if at all, only as a shorthand for the § 10(a)(4) "exceeded their powers" ground.

The result is a genuine circuit split, and this is one of those places where you must check your circuit before you advise anyone. Broadly: some circuits (the Fifth and Eighth, for instance) have read Hall Street to eliminate manifest disregard as a standalone basis; others (the Second, Ninth, and others) have preserved it, sometimes by recharacterizing it as a gloss on the statutory "exceeded their powers" language rather than a free-floating common-law ground. Either way, the practical reality is the same and worth stating bluntly: manifest disregard almost never succeeds. It is not "the arbitrator was wrong about the law." It requires showing the arbitrator was aware of clearly governing law and consciously chose to flout it, an almost impossible thing to prove from a typical award, and flatly impossible to prove from a bare award that gives no reasons. As a tool for attacking a damages number you dislike, manifest disregard is a very long shot, and lawyers who promise clients otherwise are setting them up for disappointment.

What This Means for the Size and Stability of Awards

Put the pieces together and a clear picture emerges of how damages behave on judicial review.

A frustrated loser who thinks the damages are too high has, in practice, three honest options. It can argue the arbitrator exceeded their powers under § 10(a)(4), for example by awarding a category of damages the contract expressly forbade (this is where a clear punitive-damages waiver does its work). It can argue evident partiality or fraud under § 10(a)(1)-(2). Or it can argue the award violates a well-defined, dominant public policy, a narrow ground the Supreme Court recognized in the labor-arbitration context in W.R. Grace and Eastern Associated Coal Corp. v. United Mine Workers, 531 U.S. 57 (2000), and that courts apply gingerly. What the loser cannot do is relitigate the damages math. That is the deal.

This is exactly why the most effective "damages control" in arbitration happens before any dispute exists, in the contract: limitation-of-liability clauses, consequential-damages waivers, liquidated-damages provisions, caps, and clear statements about punitive damages and fees. By the time you are reading the award, your damages exposure was largely fixed years ago by your drafting.

Cross-Border Awards: The New York Convention

If the FAA is the engine for domestic awards, the New York Convention is the engine for international ones, and it is, quietly, one of the most successful treaties in the world.

The Convention on the Recognition and Enforcement of Foreign Arbitral Awards, concluded in 1958 and now joined by over 170 states, is the reason a company can win an arbitration in Singapore and collect against the loser's assets in Texas, or vice versa. The United States acceded in 1970 and implemented the Convention through Chapter 2 of the FAA, 9 U.S.C. §§ 201-208. The Convention's core bargain is simple and powerful: each signatory agrees to recognize and enforce arbitral awards made in other signatory states, subject only to a short, exclusive list of defenses set out in Article V of the Convention.

Those Article V defenses will feel familiar after our tour of FAA § 10, because the FAA's domestic grounds were drawn from the same tradition. A court may refuse to enforce a Convention award only on grounds such as: the arbitration agreement was invalid; the resisting party was not given proper notice or was otherwise unable to present its case; the award exceeds the scope of the submission; the tribunal or procedure did not conform to the parties' agreement; the award is not yet binding or has been set aside at the seat; the subject matter is not arbitrable in the enforcing state; or enforcement would be contrary to that state's public policy. As with the FAA, notice what is missing: there is no "the tribunal got the damages wrong" defense. A foreign award's damages number enjoys roughly the same near-immunity from merits review that a domestic award enjoys, just under a different instrument.

For a party thinking about damages in an international deal, the Convention changes the calculus in a specific way: an arbitral award against a foreign counterparty is, in most of the world, easier to enforce than a court judgment, because no comparable global treaty makes national court judgments routinely enforceable across borders. That enforceability premium is itself a kind of damages enhancement, a $5 million award you can actually collect on three continents is worth more than a $7 million judgment trapped in one. This is one of the strongest practical arguments for choosing arbitration in cross-border contracts, and it belongs in any honest comparison of forums.

Cost and Speed: The Other Half of the "Damages" Question

When clients ask about arbitration "damages," they are often really asking a broader question: what will I net, and how soon? Two awards of identical face value are not economically identical if one took fourteen months and the other took four years, and if one cost $80,000 in forum and arbitrator fees while the other cost $400,000. So a candid treatment of arbitration damages has to address cost and speed, with the same honesty we applied to award sizes.

The conventional wisdom is that arbitration is faster and cheaper than litigation. The reality is more nuanced, and the empirical literature (again, Drahozal and Stipanowich are the careful voices here) supports a more measured claim: arbitration can be faster and cheaper, especially when the parties use expedited procedures, limit discovery, and stay disciplined, but a sprawling, document-heavy commercial arbitration with three arbitrators and full-bore discovery can rival or exceed the cost of litigation. What arbitration almost always saves is the back end: there is no multi-year appellate process, which is precisely the durability point we keep returning to. The savings are less in the trial and more in the absence of an appeal.

There is also a cost asymmetry worth naming. In litigation, the public funds the judge and the courtroom; the filing fee is trivial. In arbitration, the parties pay the arbitrator's hourly rate and the institution's administrative fees. For a small claim that is a real deterrent; for a large one it is a rounding error. This is why consumer and employment arbitration, where one side is an individual, has attracted special scrutiny and special rules (the AAA Consumer and Employment fee schedules cap the individual's share precisely to address this). For a fuller treatment of the procedural cost structure, our AAA commercial arbitration guide walks through the fee tiers.

Worked Hypotheticals: Following the Money

Abstractions only go so far. Let us run two clearly-labeled, invented scenarios to see how all of this fits together.

Hypothetical One: The Software Licensing Dispute. Northwind Software, Inc. licenses an enterprise platform to Acme Retail Group for a fixed annual fee. The contract has an AAA arbitration clause incorporating the Commercial Rules, a New York choice-of-law clause, a "prevailing party shall recover reasonable attorney's fees" clause, and a consequential-damages waiver. Acme stops paying, claiming the software failed; Northwind demands arbitration for the unpaid fees, and Acme counterclaims for the profits it allegedly lost when the platform crashed during its holiday sale.

Walk the damages through the framework. Northwind's claim for unpaid fees is straightforward expectation damages, and the arbitrator can award them plus pre-award interest at a rate the arbitrator deems appropriate under Rule R-47(d). Acme's lost-profits counterclaim runs straight into the consequential-damages waiver; the arbitrator is bound to enforce it, so the holiday-sale profits, however large, are off the table. Because both sides demanded fees and the contract has a prevailing-party clause, the arbitrator can shift fees to the loser. If the arbitrator awards Northwind, say, an (invented) $900,000 in unpaid fees plus $110,000 in interest plus $250,000 in fees, the headline "$1.26 million award" is really three different buckets recovered under three different rules. If Acme then runs to court claiming the arbitrator misread the failure-to-perform defense, it will almost certainly lose: that is a merits argument, and merits arguments are not on the § 10 list. The award stands.

Hypothetical Two: The Cross-Border Joint Venture. Northwind and a German manufacturer, Rhein Industrie GmbH, arbitrate a joint-venture dispute seated in London under institutional rules, governed by English law, with a clause silent on punitive damages. The tribunal awards Northwind compensatory damages and compound interest. Rhein's assets are largely in Germany and partly in the United States. Northwind does not need to relitigate anything; it takes the London award to courts in Germany and the U.S. and seeks recognition under the New York Convention. Rhein's only defenses are the narrow Article V grounds, and "the tribunal valued our damages too generously" is not among them. Northwind collects. Note the punitive-damages wrinkle: because the seat is London and the governing law is English, and because Mastrobuono is a domestic-FAA doctrine, the availability of any punitive or exemplary element is governed by the chosen law and the agreement, not by the American case, a reminder that the Mastrobuono analysis is jurisdiction-specific and does not travel automatically across borders.

These two stories illustrate the through-line of this entire article: in arbitration, the damages that matter are the ones the contract and the chosen rules allow, and once awarded, they are extraordinarily hard to dislodge.

How Arbitrators Actually Calculate Damages

It is worth dispelling a myth: arbitrators do not pull damages numbers out of the air, and the good ones are, if anything, more rigorous than a jury because they are professionals (often retired judges or experienced practitioners) who will have to live with their reputations.

In a typical reasoned commercial award, the arbitrator works through the same proof requirements a court imposes. The claimant must prove damages with reasonable certainty, not mathematical precision, but more than speculation. Lost-profits claims, especially for new or unestablished ventures, face the same skepticism in arbitration that they face in court under the "new business rule," and a claimant who shows up without a credible damages expert and a defensible model is asking for trouble. The defending party, in turn, will press the duty to mitigate, attacking the claimant's failure to cover, resell, or otherwise reduce its losses. These are not arbitration-specific doctrines; they are the ordinary substantive law of damages, which the arbitrator is bound to apply. This is exactly why our sibling guides on contract damages and intellectual-property damages are useful companions, the measures they describe are the measures the arbitrator uses.

Where arbitration differs is in procedure and presentation. Discovery is typically narrower, so the damages record is built faster and on fewer documents. The rules of evidence are relaxed; an arbitrator can consider evidence a court might exclude, which cuts both ways for a damages case. And the absence of a jury means the damages story is told to a sophisticated professional rather than to laypeople, which tends to reward disciplined expert analysis over emotional appeal. For very large or technical damages questions, parties sometimes use specialized arbitrators or even a separate damages phase (bifurcation), trying liability first and damages second.

Key Takeaways

If you remember only a handful of things from this guide, make them these.

The most important "statistic" in arbitration is not a dollar figure; it is the near-certainty that whatever the arbitrator awards will survive judicial review. The Federal Arbitration Act's vacatur grounds in 9 U.S.C. § 10 are narrow and exclusive, they police process integrity, not the correctness of the damages, and Hall Street v. Mattel confirmed that parties cannot contract for broader court review.

Arbitrators can award nearly the full range of remedies a court can: compensatory and consequential damages, lost profits, interest (often compound, often substantial), equitable relief, and, under Mastrobuono v. Shearson, punitive damages where the agreement permits them. The flip side is that the contract controls: a clear remedy waiver, consequential-damages exclusion, or punitive-damages bar will be enforced, so your damages exposure is largely fixed at the drafting table.

Attorney's fees can be shifted in arbitration by contract, by statute, or because both parties asked for them, and fees plus interest can make up a large fraction of a headline award. Always disaggregate the components.

"Manifest disregard of the law" is a doctrine in limbo, recognized in some circuits and not others, and a long shot everywhere; do not bank a strategy on it.

Internationally, the New York Convention makes arbitral awards enforceable across more than 170 countries on narrow Article V defenses, an enforceability premium that can make an arbitral award worth more than a larger court judgment.

And finally: be honest about the numbers. No one can responsibly quote a precise "average arbitration award," because the visible awards are a self-selected slice and the institutional data measures process, not award size. Describe ranges, cite the real studies (the CFPB's 2015 study, FINRA's annual statistics, the empirical scholarship), and label your hypotheticals as hypotheticals.

Frequently Asked Questions

Can an arbitrator award punitive damages? Yes, where the parties' arbitration agreement, fairly read, permits it. The Supreme Court so held in Mastrobuono v. Shearson Lehman Hutton, Inc., 514 U.S. 52 (1995), and added that a generic state choice-of-law clause does not, by itself, strip the arbitrator of that power. Parties can, however, expressly waive punitive damages in the agreement, and such waivers are generally enforced, subject to a narrow "effective vindication" limit for some federal statutory claims.

How hard is it to overturn an arbitration award I think is too high? Very hard. A court can vacate an award only on the narrow grounds in 9 U.S.C. § 10 (fraud, evident partiality, certain procedural misconduct, or the arbitrator exceeding their powers), or modify it for a clear mathematical error under § 11. "The arbitrator got the law or the damages wrong" is not a ground. Hall Street Associates v. Mattel, 552 U.S. 576 (2008), held that parties cannot contract for broader judicial review under the FAA.

What is "manifest disregard of the law"? A judge-made doctrine allowing vacatur where an arbitrator knew the governing law and consciously ignored it, far more than ordinary legal error. After Hall Street, the circuits are split on whether it survives as an independent ground or only as a label for the statutory "exceeded their powers" basis. It rarely succeeds and is essentially impossible to prove from a no-reasons award.

Are arbitration awards public, so I can look up typical amounts? Generally no. Most commercial awards are confidential and never filed publicly. Awards become visible mainly when someone seeks confirmation or vacatur in court, which makes any "average award" computed from court dockets unrepresentative. Institutional data (AAA, JAMS, FINRA) and academic and government studies (such as the CFPB's 2015 arbitration study) tell you about caseloads, timing, and sector-specific outcomes, not a reliable national average award size.

Can I recover attorney's fees in arbitration? Sometimes. The American Rule (each side pays its own lawyers) is the default, but fees can be shifted by a contractual prevailing-party clause, by a fee-shifting statute, or, under many institutional rules, when both parties have requested fees, thereby consenting to the arbitrator deciding the issue. Fees and costs can be a large share of a total award.

Will interest meaningfully increase my award? Often, yes. Arbitrators have broad discretion to award pre-award interest from the date of breach, and they may make it compound (AAA Commercial Rule R-47(d); ICDR Rules Article 34(4)). Over a multi-year, seven-figure dispute, compound interest can add a significant fraction to the principal.

How do I enforce an arbitration award against a foreign company? Through the New York Convention (1958), implemented in the U.S. by 9 U.S.C. §§ 201-208. Over 170 countries have agreed to recognize and enforce each other's arbitral awards, subject only to the narrow defenses in Article V. This makes arbitral awards generally easier to enforce across borders than court judgments.

Is arbitration really cheaper and faster? It can be, especially with expedited procedures and limited discovery, and it almost always saves the years and expense of an appeal. But a large, three-arbitrator commercial arbitration with full discovery can cost as much as litigation, and the parties, not the public, pay the arbitrators. The biggest reliable saving is finality, not the hearing itself.

Related Articles

This article is general information, not legal advice. Arbitration law and damages doctrine vary by jurisdiction, by chosen rules, and over time; consult qualified counsel about your specific situation before acting.