Multiparty and Multicontract Arbitration: Consolidation, Joinder, and the Non-Signatory Problem

Document type: Article Practice area: Arbitration — International Arbitration Jurisdiction: United States Last reviewed: 5 September 2026

The mismatch

Arbitration rests on consent. Section 2 of the Federal Arbitration Act makes written arbitration agreements "valid, irrevocable, and enforceable, save upon such grounds as exist at law or in equity for the revocation of any contract," and the whole edifice is built on the proposition that parties are bound because they agreed to be.

Commerce does not organize itself that way. A construction project involves an owner, a contractor, a dozen subcontractors, designers, suppliers, and sureties, connected by a web of contracts that were negotiated separately and may contain different dispute resolution clauses. A corporate acquisition involves a purchase agreement, a transition services agreement, an escrow agreement, employment agreements, and a shareholders' agreement, signed by overlapping but not identical parties. A financing involves a credit agreement, guarantees, security documents, and intercreditor arrangements.

When something goes wrong, the dispute crosses those boundaries. The owner's claim against the contractor depends on what a subcontractor did. The buyer's indemnity claim under the purchase agreement overlaps with its claim under the escrow agreement. And the parties discover that they have contracted into a system that resolves bilateral disputes efficiently and multilateral ones badly.

The mechanisms that address this — consolidation, joinder, and the non-signatory doctrines — are the subject of this article, and the single most important practical point about all of them is that they must be provided for in advance. A tribunal cannot conjure jurisdiction over a party that did not agree, and institutional rules cannot override the clauses the parties actually signed.

Consolidation

Consolidation combines two or more pending arbitrations into a single proceeding.

The source of the power is the parties' agreement, either directly or through incorporation of institutional rules. Modern institutional rules contain consolidation provisions, and while their formulations differ, they converge on similar requirements:

All parties agree; or
All claims are made under the same arbitration agreement; or
The claims are made under different arbitration agreements, but the arbitrations are between the same parties, the disputes arise in connection with the same legal relationship or a related series of transactions, and the arbitration agreements are compatible.

Compatibility is the requirement that does the work, and it is where drafting matters. Two clauses are compatible when they can be operated together: same institution, same seat, same governing law of the arbitration agreement, same number of arbitrators, same appointment mechanism, same language. A clause providing for three arbitrators seated in New York cannot readily be consolidated with one providing for a sole arbitrator seated in Singapore.

Who decides. Under most institutional rules, the institution itself decides consolidation, at least where a tribunal has not yet been constituted, and its decision is administrative rather than jurisdictional — it does not preclude a tribunal from later deciding it lacks jurisdiction.

The timing problem. Consolidation is far easier before tribunals are constituted. Once two tribunals are sitting, consolidation requires reconstituting at least one, which the parties who nominated those arbitrators will resist.

Court-ordered consolidation. United States federal courts have generally held that they may not order consolidation absent an agreement permitting it. The reasoning follows from consent: a party that agreed to arbitrate with A did not agree to arbitrate with A and B together, and imposing consolidation alters the bargain. Some state arbitration statutes take a different view, and the seat's law matters.

Joinder

Joinder adds a party to an existing arbitration.

Institutional rules generally permit joinder where the additional party is bound by the same arbitration agreement, or where all parties, including the additional one, consent. Some rules permit joinder by the tribunal after constitution; others restrict it to before, precisely because of the appointment problem discussed below.

The consent constraint is real. A subcontractor that never signed the main contract's arbitration clause cannot be joined to the owner-contractor arbitration merely because its work is at issue. The owner's claim against it, if any, must be pursued separately unless a non-signatory theory applies or the contracts were drafted to permit it.

Multi-tier structures. Where a project's contracts are drafted as a suite — main contract, subcontracts, and supply agreements all containing compatible clauses referencing the same institution, seat, and rules, and each expressly permitting joinder and consolidation of related disputes — the machinery works. Where each contract was negotiated separately with different counsel using different templates, it does not.

The appointment problem

Multiparty arbitration creates a difficulty that took the arbitration world a famous case to notice.

The problem. A three-member tribunal is conventionally constituted with each side nominating one arbitrator and the two nominees or the institution selecting the chair. With three parties whose interests diverge, that mechanism fails: if two respondents must jointly nominate one arbitrator while the claimant nominates its own, the respondents' right to participate in constituting the tribunal is unequal to the claimant's.

The response. The influential French decision in the Dutco litigation held that the principle of equality in the constitution of the tribunal is a matter of public policy that cannot be waived in advance. The practical consequence was a redesign of institutional rules.

The modern solution, adopted in substance across the major institutions: where multiple parties on one side cannot agree on a joint nomination, the institution appoints all members of the tribunal, including the chair, disregarding any nomination already made. This preserves equality by removing the nomination right from everyone rather than by giving it unequally.

What this means for a party. In a multiparty case, the right to nominate an arbitrator may evaporate. That is a genuine cost of the multiparty structure and it should be understood before a party insists on joinder or consolidation that will trigger it.

Drafting responses. Sophisticated multi-contract suites sometimes provide expressly for the appointment mechanism in a consolidated or multiparty proceeding: institutional appointment of the entire tribunal, or a specified allocation of nomination rights among defined groups of parties.

Non-signatories: when someone who did not sign is bound

The most litigated question in this area is whether a person who did not sign an arbitration agreement can be compelled to arbitrate, or can compel arbitration against a signatory.

The governing framework comes from Arthur Andersen LLP v. Carlisle, 556 U.S. 624 (2009), which held that a litigant who was not a party to an arbitration agreement may invoke § 3 of the FAA if the relevant state contract law allows enforcement by or against a non-party:

"Because 'traditional principles' of state law allow a contract to be enforced by or against nonparties to the contract through 'assumption, piercing the corporate veil, alter ego, incorporation by reference, third-party beneficiary theories, waiver and estoppel,' the Sixth Circuit's holding that nonparties to a contract are categorically barred from § 3 relief was error."

GE Energy Power Conversion France SAS v. Outokumpu Stainless USA, LLC, 590 U.S. 432 (2020), extended the point to international cases governed by the New York Convention, holding that the Convention does not conflict with domestic equitable estoppel doctrines permitting a non-signatory to compel arbitration.

The theories, and how they actually work:

Incorporation by reference. Contract A incorporates the arbitration clause of Contract B. Common in construction, where subcontracts incorporate the main contract's terms. The question is always whether the incorporating language is broad enough to capture the arbitration clause specifically — general incorporation of "the terms and conditions" of another contract may or may not reach its dispute resolution provision, and courts differ.

Assumption. A non-signatory's conduct manifests assent to be bound — participating in the arbitration without objection, or performing under the contract and claiming its benefits.

Agency. A signatory signed as agent for the non-signatory principal, who is bound.

Alter ego and veil piercing. The corporate separateness of a signatory and a non-signatory affiliate is disregarded. A demanding standard, and the same standard as in any other veil-piercing context.

Third-party beneficiary. A non-signatory intended to benefit from the contract may be bound by its arbitration clause when suing on it.

Equitable estoppel. The most used and most elastic. Two branches:

Direct benefits estoppel: a non-signatory that knowingly exploits and receives direct benefits from a contract cannot avoid its arbitration clause when suing on obligations arising from it.
Intertwined claims estoppel: a signatory's claims against a non-signatory are so intertwined with the contract containing the arbitration clause, and the non-signatory's relationship with the signatory so close, that the signatory is estopped from avoiding arbitration.

The second branch is applied inconsistently and some courts have narrowed or rejected it. Because Arthur Andersen directs the analysis to state contract law, the answer varies by jurisdiction, and a choice-of-law analysis precedes the estoppel analysis.

The asymmetry worth noting. Courts have been more willing to allow a non-signatory to compel arbitration than to allow a signatory to compel a non-signatory into arbitration. The first respects the signatory's own agreement to arbitrate; the second imposes arbitration on someone who never agreed.

Class and collective arbitration

A distinct multiparty problem, and one the Supreme Court has addressed repeatedly with a consistent theme: class arbitration requires an affirmative contractual basis, and it will not be inferred.

Green Tree Financial Corp. v. Bazzle, 539 U.S. 444 (2003), was a fractured decision in which a plurality suggested that whether a contract permits class arbitration is a question for the arbitrator. Later decisions have substantially undermined that premise.

Stolt-Nielsen S.A. v. AnimalFeeds International Corp., 559 U.S. 662 (2010), held that a party may not be compelled to submit to class arbitration unless there is a contractual basis for concluding that it agreed to do so. Where the parties had stipulated that their agreement was silent on class arbitration, imposing it exceeded the panel's powers:

"An implicit agreement to authorize class-action arbitration is not a term that the arbitrator may infer solely from the fact of the parties' agreement to arbitrate."

The Court emphasized the differences between bilateral and class arbitration: the arbitrator no longer resolves a single dispute but adjudicates the rights of absent parties; the stakes are magnified; the commercial privacy of arbitration is lost; and the limited judicial review appropriate to a consensual bilateral process becomes troubling when applied to aggregate claims.

Lamps Plus, Inc. v. Varela, 587 U.S. 176 (2019), went further, holding that an ambiguous agreement cannot provide the necessary contractual basis, and that the state law contra proferentem rule — construing ambiguity against the drafter — could not be applied to find consent to class arbitration:

"Courts may not infer from an ambiguous agreement that parties have consented to arbitrate on a classwide basis."

The practical position. Class arbitration exists only where the agreement affirmatively provides for it, which very few commercial agreements do. Class action waivers in arbitration agreements have been broadly enforced.

What has grown instead is mass arbitration: thousands of individual claimants filing individual arbitrations simultaneously, each triggering the respondent's obligation to pay institutional filing fees. The aggregate fee exposure can be enormous before any merits are reached, and the technique has produced substantial settlements and a wave of drafting responses — batching provisions, bellwether procedures, fee-shifting arrangements, staged filing requirements, and mediation preconditions. Whether particular responses are enforceable is being litigated, and drafters should expect the answers to keep moving.

Who decides: arbitrability and delegation

Underlying every consolidation, joinder, and non-signatory dispute is the question of who decides it.

First Options of Chicago, Inc. v. Kaplan, 514 U.S. 938 (1995), set the framework: courts decide whether the parties agreed to arbitrate a matter unless the parties clearly and unmistakably provided that the arbitrator should decide arbitrability. The standard is deliberately demanding — silence or ambiguity means a court decides.

Howsam v. Dean Witter Reynolds, Inc., 537 U.S. 79 (2002), distinguished substantive arbitrability — whether the parties agreed to arbitrate this dispute, presumptively for the court — from procedural questions such as time limits, notice, laches, and conditions precedent, presumptively for the arbitrator.

Henry Schein, Inc. v. Archer & White Sales, Inc., 586 U.S. 63 (2019), held that where the parties have delegated arbitrability to the arbitrator, a court may not decline to enforce the delegation on the ground that the argument for arbitration is "wholly groundless." The delegation, if clear, is enforced.

Coinbase, Inc. v. Suski, 602 U.S. 143 (2024), addressed a further layer: where parties have entered two contracts, one delegating arbitrability and a later one that arguably sends the dispute to court, a court must decide which contract governs. The Court reasoned that the dispute is about the formation of the parties' arbitration agreement, and that question is always for the court:

"A court, not an arbitrator, must decide which contract governs."

That holding matters directly for multicontract disputes: where a transaction comprises several agreements with different dispute resolution provisions, the threshold question of which clause applies is a judicial one.

How delegation is established. Most institutional rules contain a competence-competence provision empowering the tribunal to rule on its own jurisdiction, and many courts have held that incorporation of such rules constitutes clear and unmistakable delegation. That conclusion is not universal, particularly where one party is an unsophisticated consumer, and the analysis differs by circuit.

The practical drafting point. If delegation is intended, say so expressly in the clause rather than relying on incorporated rules: "The arbitrator shall have exclusive authority to resolve any dispute relating to the interpretation, applicability, enforceability, or formation of this agreement to arbitrate, including any claim that all or part of it is void or voidable."

Drafting the clause suite

Everything above converges on drafting, and the drafting must be done across the whole transaction rather than contract by contract.

The governing principle: in a transaction comprising several agreements among overlapping parties, the dispute resolution provisions should be designed as a single system, not copied from a template into each document by whoever drafted it.

What compatibility requires:

The same institution and the same rules, referenced identically.
The same seat.
The same governing law of the arbitration agreement, stated expressly — this is a distinct choice from the governing law of the contract and it is routinely omitted.
The same number of arbitrators, or a mechanism that resolves a difference.
The same appointment procedure.
The same language.
The same confidentiality regime.

What must be added:

Express consent to consolidation of arbitrations arising under any agreement within the defined transaction suite, with the suite identified by schedule.
Express consent to joinder of any party to any agreement within the suite.
An appointment mechanism for multiparty proceedings — most simply, institutional appointment of the entire tribunal where the parties on one side cannot agree.
A mechanism for related but non-suite contracts, such as subcontracts, requiring the prime party to procure compatible clauses.
A delegation provision, if delegation is intended.
A class waiver, if that is intended, drafted with the mass-arbitration response in mind.

What to avoid:

Different institutions in different agreements of the same deal.
A carve-out in one agreement sending some disputes to court while a related agreement sends the same subject matter to arbitration — this is the Coinbase problem, and it guarantees a threshold fight about which clause governs.
Silence on consolidation and joinder in a transaction that plainly needs both.
A clause that provides for a sole arbitrator in one agreement and three in another.
Reliance on institutional rules to solve a problem the clauses create. Rules operate within the parties' consent; they do not supply it.

A practical discipline. On any transaction with more than two agreements, one lawyer should own the dispute resolution provisions across all of them, review them together at signing, and produce a one-page map: which agreements, which parties, which clause, same institution, same seat, consolidation permitted, joinder permitted. That map takes an hour and prevents the most expensive category of arbitration dispute.

Worked example one: the construction chain

Ines Aguirre is general counsel of an owner developing a processing facility. The prime contract with the contractor provides for arbitration under a named institution, seated in Houston, three arbitrators. The contractor's subcontracts — negotiated separately by the contractor — variously provide for litigation in state court, arbitration under a different institution seated elsewhere, and, in two cases, nothing at all.

What goes wrong. A commissioning failure produces claims flowing in every direction: the owner against the contractor; the contractor against three subcontractors; two subcontractors against a supplier; and everyone against the design engineer, who contracted directly with the owner under a separate agreement with its own clause.

Where the parties end up: four proceedings in three forums, with overlapping evidence, inconsistent findings a real possibility, and no mechanism to bring them together. The design engineer's separate contract means the owner's claim against it proceeds independently of the arbitration in which the technical facts will actually be determined.

What Ines does for the next project. She takes ownership of dispute resolution across the whole contract suite:

One institution, one seat, one set of rules, one language, three arbitrators, stated identically in every agreement.
Governing law of the arbitration agreement stated expressly and identically.
Every agreement in the suite expressly consents to consolidation of arbitrations arising under any suite agreement, and to joinder of any suite party.
The prime contract requires the contractor to include the same clause, verbatim, in every subcontract above a stated value, and to procure the subcontractor's written consent to joinder in any arbitration between the owner and the contractor.
A schedule to each agreement lists the suite documents.
Multiparty appointment: where parties on one side cannot agree a joint nomination within twenty-one days, the institution appoints all three arbitrators.
The design agreement is brought into the suite.

Cost of the change: perhaps a week of drafting and a negotiation with the contractor about the flow-down obligation. Value: on the next failure, one proceeding rather than four.

Worked example two: the corporate group

Tomás Bregović acts for a buyer in a completed acquisition. The purchase agreement, signed by the buyer and the seller's holding company, contains an arbitration clause. Post-closing, the buyer discovers that inventory was overstated, and the misconduct appears to have been carried out by two operating subsidiaries and by the seller's former CFO, none of whom signed the purchase agreement.

The question: can the buyer bring them into the arbitration?

The analysis, party by party:

The holding company — a signatory. Straightforward.

The operating subsidiaries — non-signatories. Tomás considers: agency (did the holding company contract as their agent? unlikely on these facts); alter ego (a demanding standard, and there is no evidence of disregard of corporate formalities); third-party beneficiary (the subsidiaries received no direct benefit under the purchase agreement); and direct benefits estoppel (the subsidiaries were sold, not benefited by the agreement).

The former CFO — a non-signatory individual. Tomás considers whether the claims against him are so intertwined with the purchase agreement, and his relationship with the seller so close, that intertwined-claims estoppel applies. The answer depends on the governing state's law, and it is genuinely uncertain.

What Tomás actually does. He files the arbitration against the holding company only, on the contractual claims, and files a separate court action against the subsidiaries and the former CFO on fraud claims that do not depend on the purchase agreement. He then confronts a parallel proceedings problem of his own making — but a manageable one, because the arbitration will determine the contractual measure and the court action the tort claims.

The lesson for the next deal. The purchase agreement should have named the parties who might be defendants: material subsidiaries as parties to the arbitration clause, or as express third-party beneficiaries bound by it; and key individuals bound through their employment or restrictive covenant agreements, with compatible clauses. Adding parties to a clause at signing costs nothing. Adding them afterward is impossible.

Worked example three: the failed consolidation

Nkechi Obiora represents a supplier in two arbitrations with the same counterparty: one under a 2021 master agreement, one under a 2023 amendment that, on the counterparty's reading, is a separate contract.

Both clauses name the same institution. But: the 2021 clause provides for three arbitrators seated in London; the 2023 clause provides for a sole arbitrator seated in New York. The 2021 clause is silent on the governing law of the arbitration agreement; the 2023 clause specifies New York law.

Nkechi applies to consolidate. The institution refuses. The clauses are not compatible: different seats, different tribunal sizes, different appointment mechanisms, and an unresolved question about the law governing each arbitration agreement.

The consequences. Two proceedings, two tribunals, two sets of costs, overlapping witnesses giving evidence twice, and — the outcome the parties would least have wanted — a genuine risk of inconsistent findings on the same factual question about what the parties agreed in 2023.

The cause. The 2023 amendment was drafted by transactional counsel who took a clause from a recent unrelated deal because it was convenient. Nobody compared it with the clause in the agreement it amended.

The rule that would have prevented it: an amendment to an agreement containing an arbitration clause should either be silent on dispute resolution — leaving the original clause to govern — or should replicate it exactly. A different clause in an amendment creates two arbitration agreements where the parties thought they had one, and it is among the most common and most avoidable errors in the field.

Enforcement consequences

Multiparty complexity does not end with the award. It reappears at enforcement, and the risks are worth naming because they can undo the efficiency the mechanisms were meant to deliver.

The New York Convention grounds. Under the Convention, given effect by 9 U.S.C. § 201 and following, recognition may be refused where the party against whom the award is invoked was not given proper notice or was otherwise unable to present its case; where the award deals with a difference not contemplated by or falling within the submission to arbitration; or where the composition of the arbitral authority or the arbitral procedure was not in accordance with the parties' agreement.

Each of those maps onto a multiparty risk:

Improper joinder or consolidation produces an award against a party that did not agree to arbitrate with the others — a jurisdictional defect and a scope objection.
The appointment problem produces a challenge under the composition ground. A party deprived of its nomination right, where the agreement gave it one, has a real argument. This is precisely why institutional rules now strip nomination rights from everyone rather than allocating them unequally.
Consolidation without consent is a procedure not in accordance with the parties' agreement.

Vacatur in a US-seated arbitration. Section 10 of the FAA permits vacatur where the arbitrators exceeded their powers. Stolt-Nielsen is the paradigm: imposing class arbitration without a contractual basis exceeded the panel's powers. A tribunal that consolidates or joins without authority is exposed to the same argument.

The practical implication. A party seeking consolidation or joinder should build the consent record carefully — the clauses, the institutional decision, any party agreement — and a tribunal proceeding over objection should address its jurisdiction expressly in a reasoned decision. A party opposing should preserve the objection at every stage: on the application, in the answer, in the terms of reference or equivalent, and in the post-hearing submissions. An objection not preserved is generally waived, and a party that participates without objecting will find it difficult to raise the point at enforcement.

Partial awards and severability. Where a tribunal has jurisdiction over some parties or claims and not others, an award that separates them cleanly is far easier to enforce. Tribunals in complex cases increasingly issue partial awards for exactly this reason.

The efficiency question

It is worth asking whether the multiparty machinery is actually worth using, because the answer is not always yes.

What consolidation and joinder deliver: one set of evidence, one tribunal, one set of findings, no inconsistent outcomes, and — usually — lower aggregate cost.

What they cost:

The nomination right, which in a multiparty proceeding is likely to be lost to institutional appointment.
Speed. A consolidated proceeding with six parties is slower than a bilateral one, sometimes dramatically. Scheduling alone becomes a project.
Confidentiality. A party's documents and evidence become visible to counterparties it never contracted with.
Control. A claimant that wanted a focused case against one respondent finds itself in a proceeding shaped by other parties' claims and defences.
Settlement dynamics. Bilateral settlement becomes harder when six parties must agree, though it also becomes possible to resolve everything at once.
Enforcement risk, as above.

When to seek it: where the same factual determination will decide multiple claims, where inconsistent findings would be seriously damaging, where the parties are genuinely interdependent, and where the cost of duplicated proceedings is large relative to the amounts at stake.

When to resist it: where a party's own claim is strong and simple and would be slowed by others' complexity; where confidentiality matters more than efficiency; where the nomination right is genuinely valuable given the subject matter; and where a party has a strong jurisdictional objection that consolidation would obscure.

There is no default answer. The mistake is to treat consolidation as automatically desirable, or joinder as automatically resisted, without asking what the party actually wants from the proceeding.

Running a multiparty proceeding

Once a proceeding involves four or more parties, its management becomes a discipline of its own, and tribunals and counsel who have not done it before consistently underestimate what it takes.

The first procedural conference matters more than usual. Items that must be settled early:

Party groupings. Which parties are aligned, for the purposes of submission deadlines, hearing time, and document requests. Groupings that reflect actual alignment work; groupings imposed by the tribunal on parties with divergent interests generate objections.
Submission sequencing. In a bilateral case, memorials alternate. With claims running in several directions — claimant against respondent, respondent against a joined party, cross-claims among respondents — the sequence must be designed. A common structure is simultaneous first-round submissions by all claiming parties, then simultaneous responses, then a single reply round.
Document production. Requests multiply combinatorially. Limits on the number of requests per party pair, and a consolidated redfern-style schedule, are essential.
Hearing time allocation. Not equal shares — shares proportionate to the claims each party is advancing and defending, agreed in advance and enforced with a chess clock.
Cross-examination. Who may cross-examine whom, and in what order. Without a protocol, a witness may face five successive cross-examinations covering the same ground.
Confidentiality. What each party may see. Where parties are commercial competitors, a confidentiality ring may be needed within the arbitration itself.
Costs. How costs will be allocated among multiple parties with different degrees of success is genuinely difficult and should be flagged early.

Settlement of part. A party that settles mid-proceeding must exit cleanly: a consent award or a withdrawal, an allocation of costs to date, provisions dealing with its continuing document and witness obligations, and — importantly — the effect of the settlement on the remaining parties' claims for contribution.

The tribunal's burden. A six-party arbitration is several times the work of a bilateral one, and the tribunal's availability becomes the schedule's binding constraint. Parties selecting arbitrators for a complex multiparty case should ask directly about capacity and should expect longer gaps between hearings.

Where this is heading

Three developments are worth watching.

Institutional rule convergence. The major institutions have converged substantially on consolidation and joinder provisions, and successive revisions have generally expanded them — permitting joinder after tribunal constitution in defined circumstances, permitting consolidation of arbitrations under related but non-identical agreements, and providing more detailed multiparty appointment mechanisms. A clause referencing a specific edition of the rules should be considered against whether the parties want the current edition to apply.

Mass arbitration and the drafting response. The volume of individual filings has driven a wave of clause innovation, and courts are now testing which responses are enforceable. Provisions requiring batching, bellwethers, staged filing, or mediation preconditions face challenges on unconscionability and on whether they frustrate the right to arbitrate. Drafters should expect this area to keep moving and should avoid provisions that could invalidate the whole clause if struck.

Non-signatory doctrine. GE Energy confirmed that domestic equitable doctrines survive the New York Convention, but the content of those doctrines is state law and the intertwined-claims branch of estoppel is applied inconsistently. Practitioners should expect the answer to depend on the governing law of the arbitration agreement — which is one more reason to state that law expressly rather than leaving it to be litigated.

Court assistance and the interaction with litigation

Multiparty arbitration frequently coexists with court proceedings, and the interaction is worth understanding.

Compelling arbitration. Section 4 of the FAA permits a party aggrieved by another's refusal to arbitrate to petition for an order compelling arbitration. In a multiparty setting, a party may need to compel some counterparties while defending court claims from others.

Partial referral. Where some claims or parties are arbitrable and others are not, a court may compel arbitration of the arbitrable claims and stay or proceed with the rest. Whether to stay the litigation pending the arbitration is discretionary, and the considerations are familiar: overlap of issues, risk of inconsistent findings, prejudice, and efficiency.

The non-arbitrable residue. Almost every complex multiparty dispute leaves something in court — a claim against a party with no arbitration agreement, a statutory claim, an in rem or lien proceeding, a claim against an insurer. Counsel should map this at the outset rather than discovering it after the arbitration is under way.

Confirmation and vacatur. Section 9 governs confirmation of a domestic award; § 10 vacatur. In a multiparty case, awards may be confirmed against some parties and challenged by others, in different courts, on different grounds. Planning for this is part of designing the proceeding.

Provisional relief. Court-ordered provisional relief in aid of arbitration remains available in most jurisdictions and is frequently the only practical route where a party needs relief against a non-signatory or before a tribunal exists.

A coordination habit. Where an arbitration and a related court action are proceeding, one lawyer should own the interaction: what is said in one forum is discoverable and quotable in the other, findings in one may have preclusive effect in the other, and settlement in one may need to account for the other. Separate teams running separate proceedings on the same facts is how parties end up taking inconsistent positions in front of the two decision-makers most likely to notice.

The ten-minute review that prevents most of this

At the closing of any transaction with more than two agreements, one person should spend ten minutes producing this table and circulating it.

Agreement	Parties	Institution	Seat	Arbitrators	Law of the arb. agreement	Language	Consolidation?	Joinder?	Delegation?
Purchase agreement
Escrow agreement
Transition services
Shareholders' agreement
Employment agreements

Any row that differs from the others is a problem — either an intentional one someone should be able to explain, or an unintentional one that can still be fixed before signing.

Any blank in the "law of the arbitration agreement" column is a future dispute about which law governs the clause's validity, scope, and reach to non-signatories.

Any "no" in the consolidation or joinder columns, in a transaction where the agreements are commercially interdependent, should be a deliberate choice rather than an oversight.

And one question to ask out loud: if this deal goes wrong, who will be sued, and are they all bound by a compatible clause? The answer is very often that the most likely defendants — subsidiaries, individuals, advisers, sureties — are not parties to any arbitration agreement at all. That is sometimes the right outcome. It should never be an accident.

The whole of this article reduces to that table and that question. Arbitration's multiparty machinery works when the consent was engineered in advance and fails when it was not, and the engineering costs an hour at signing against years of parallel proceedings afterward.