Network Externalities and the Empirical Limits of Efficient Breach: A Covenantal Theory of Commercial Obligation {#network-externalities-and-the-empirical-limits-of-efficient-breach-a-covenantal-theory-of-commercial-obligation .Title}

Seth C. Oranburg¹

Abstract

Courts that enforce relational network obligations as bilateral transactional contracts destroy the club-good value of trust on which every network member depends. Efficient breach doctrine assumes every commercial obligation is bilateral and that expectation damages internalize all harm from non-performance. This Article presses a prior empirical question: whether obligations exist in which the obligee is the trading network itself, and the harm from breach is a network externality that bilateral damages cannot internalize.

Drawing on Bernstein’s fieldwork in the diamond, cotton, grain-feed, and automotive-procurement sectors, Greif’s multilateral reputation mechanism, and Ostrom’s governance studies, the Article documents networks where compliance persists even in terminal transactions and expulsion remains non-compensable. Contemporary platform networks exhibit structural parallels the framework can assess, though their covenantal status requires case-by-case verification.

The covenantal framework identifies three structural features—permanence, unconditionality, and third-party witnessing backed by expulsion—whose presence, verified at the pleading stage by a recognized arbitral body with expulsion authority, marks the obligation as structurally covenantal. Breach in such networks degrades the club good of enforcement capital for every member, and no bilateral damages schedule can restore indifference for the collective obligee. In markets meeting this criterion, courts should apply a rebuttable presumption against incorporating relationship-preserving norms under UCC §§ 1-303 and 2-202, review expulsion under procedural-regularity deference, and treat network-internal expulsion under antitrust rule-of-reason analysis with a presumption of legitimacy. Courts applying bilateral logic to these obligations solve for the wrong harm and the wrong obligee, predictably destroying the institutional infrastructure on which every network member depends.

Introduction

A diamond dealer completes a large transaction on West 47th Street in Manhattan. The agreement is oral, sealed by a handshake and the phrase mazel u’bracha. The specific bilateral pair will never deal again; the transaction is bilateral-pair terminal. Yet both parties belong to the New York Diamond Dealers Club, or count as vouched associates of members, and every counterparty on the exchange floor is either a direct network participant or a member’s referred agent. Stranger transactions do not exist in this market.

This transaction is not an isolated curiosity. The diamond market processes billions of dollars in annual trade without written contracts or ordinary legal enforcement, a pattern efficient breach doctrine cannot explain.

Bilateral theory supplies an unambiguous prediction for the terminal pair. If the gain from selling the stones to a higher bidder exceeds the counterparty’s recoverable expectation damages—the difference between contract price and market price—the rational promisor breaches and pays damages. The promisee ends whole; the promisor captures the net surplus; social welfare increases; and the law remains indifferent between performance and breach-plus-damages. The shadow of future bilateral dealings between this specific pair is zero, so no iterated-game mechanism alters the calculus.

Yet the dealer performs. Lisa Bernstein’s fieldwork in the New York Diamond Dealers Club documents systematic compliance rates approaching those of formally enforced contracts, with court filings between members “virtually nonexistent” despite high-value, volatile transactions in which the parties will not deal again bilaterally.² The iterated-game account cannot explain this pattern, because it fails precisely where bilateral future dealings are absent.

The operative mechanism is network-wide expulsion, communicated through the World Federation of Diamond Bourses to affiliated exchanges in Antwerp, Tel Aviv, Ramat Gan, Mumbai, and elsewhere. Expulsion closes global market access permanently and without compensation.³ The dealer performs not to preserve a future bilateral relationship with this specific counterparty; that relationship may end with this transaction regardless. He performs because the network watches. Breach would trigger a status event that degrades the club good of collective enforcement capital shared by every member of the exchange, including the breaching dealer himself. The network’s third-party witnessing apparatus observes the transaction, and the permanent threat of expulsion internalizes the externalities that bilateral expectation damages are structurally blind to. Whether the dealer ever transacts with this particular counterparty again is irrelevant to the compliance mechanism; what matters is that the dealer remains inside the club boundary where the shared enforcement resource operates.

This observation marks where efficient breach doctrine’s descriptive universality claim breaks down. The doctrine’s indifference premise rests on four structural commitments: (1) the relevant parties are bilateral; (2) the harm from breach is a recoverable economic loss measurable in money and payable from promisor to promisee; (3) the promisee is the relevant injured party, and making that party whole through expectation damages exhausts the law’s remedial obligation; and (4) performance and breach-plus-damages are remedially equivalent from the promisee’s perspective.⁴ The welfare claim requires all four to hold simultaneously. If any fails—if the relevant injured party is the network and the relevant harm is a club-good externality that no bilateral transfer can internalize—the calculus cannot establish indifference for the actual obligee.

A concrete illustration clarifies the mismatch. Suppose the breaching dealer can realize a $50,000 gain by selling to a higher bidder while the bilateral counterparty’s expectation damages are only $30,000. Bilateral theory predicts breach-plus-damages: the promisor nets $20,000, the promisee is indifferent, and the law should be indifferent as well. Yet Bernstein documents compliance. The gap is not measurement error in bilateral damages; it is the existence of network externalities imposed on the club good of network enforcement capital—excludable to non-members, non-rivalrous within membership capacity.⁵ The dealer performs not because he fears this particular counterparty, but because breach would degrade the club good for every member of the exchange, including himself, regardless of whether he ever transacts with this specific person again.

Breach generates three such externalities simultaneously. First, an informational externality: every trader must now discount the reliability signal conveyed by “DDC-member,” raising search and counterparty-selection costs across the market. Second, a verification/monitoring externality: all members rationally increase costly screening of stones, creditworthiness, and performance, because one unsanctioned breach demonstrates that membership no longer guarantees compliance. Third, an enforcement-capacity externality: each breach that escapes terminal sanction lowers the collective credibility of expulsion, making future sanctions less deterrent and raising transaction costs for the network as a whole.

These costs are borne diffusely by every member at once; they are not offset by any payment the breaching dealer makes to the single counterparty. Just as a bilateral tort judgment between two riparian owners cannot restore water quality for the dozens of other users downstream,⁶ expectation damages here do not engage, let alone exhaust, the network’s interest in preserving its enforcement infrastructure. Courts awarding expectation damages to the bilateral counterparty in such settings are not “doing their best” for the relevant injured party. They are solving for the wrong harm and the wrong obligee—and in doing so, they predictably destroy the very institutional infrastructure that made the network valuable to every participant, including the bilateral parties before the court.

The destruction is not incidental. When a court incorporates a network norm as a bilateral legal term, it claims the witnessing role that the community had been performing. The community, now observing that the court has taken over enforcement, stops supplying the discretionary gestures and graduated sanctions whose authority derived from community observation rather than state power. The norm’s supply source shifts from collective governance to bilateral legal liability. Screening costs rise across the market. The enforcement-capacity externality is exported from the network—where it was internalized efficiently—to the public courts, where it is not. The bilateral remedy does not merely undercompensate the promisee. It disables the mechanism.

Three structural features organize this empirical pattern. Henry Maine observed that the movement of progressive societies has been from status to contract, but he did not claim the movement was complete or irreversible.⁷ Efficient breach doctrine presupposes its completion for commercial obligation: all such obligations arise from bilateral voluntary agreement, run to the immediate counterparty, and can be discharged by monetary payment of that counterparty’s economic loss. Bernstein’s merchant networks are evidence that the presupposition is incomplete. Membership in the diamond exchange constitutes the trader’s professional identity; expulsion is a status event that dissolves that identity rather than merely transferring resources.

Permanence means membership is not terminable at will without consequence and that obligations persist across the member’s tenure, defining the temporal boundary of the club good. Unconditionality means each trader’s duty runs to the network as collective obligee rather than to the bilateral counterparty as reciprocal promisor; the tribunal does not ask whether the counterparty also behaved badly. Third-party witnessing backed by expulsion means an institutionalized authority observes compliance, records it accessibly, and imposes sanctions that affect the member’s standing with all network participants independently of whether the bilateral counterparty forgives the breach. These features are not characterizations imposed on the data; they are what the data reveal once the bilateral lens is set aside.

This Article demonstrates that efficient breach doctrine, taken as a universal account of commercial obligation, reaches a boundary in networks exhibiting these three covenantal features. Where the diagnostic criterion—a recognized arbitral body with expulsion authority to which the parties have agreed to submit disputes—is met and verifiable at the pleading stage, the doctrine’s indifference and remedial-equivalence premises fail as a descriptive matter, because the doctrine has been solving for the wrong harm and the wrong obligee. The claim is narrow and descriptive in its domain: this Article does not show that most commerce is covenantal, nor does it offer a normative critique of efficient breach where the bilateral assumptions hold. But within its domain, the claim is strong. Where the diagnostic criterion is met, courts should interpret and review in one way; where it is absent, the standard bilateral tools remain appropriate.

The Article proceeds as follows. Part II restates efficient breach in its strongest contemporary form and identifies the internalization premise on which its indifference claim depends. Part III presents the empirical evidence organized around the two direct falsifications of that claim: terminal-transaction compliance and non-compensable expulsion. Part IV develops the covenantal framework, operationalizes the three-feature diagnostic, distinguishes it from relational-contract theory and transaction cost economics, and grounds the status-event concept in Maine. Part V translates the framework into concrete doctrinal rules for four recurring legal contexts, with step-by-step application guidance. Part VI shows how public regulatory regimes inadvertently erode private covenantal structures when they apply bilateral intervention logic to triadic governance settings. Part VII sets out falsifiable predictions and a research agenda. Part VIII addresses objections. The Conclusion reframes the domain in which efficient breach remains descriptively accurate.

I. Efficient Breach and Its Commitments

This Part restates the efficient-breach doctrine in its strongest contemporary form, isolates the four structural commitments that do the load-bearing work of its indifference and welfare claims, and surfaces the internalization premise on which the entire edifice rests. Efficient breach doctrine treats every commercial obligation as a bilateral exchange between identified parties whose relevant harm is a measurable economic loss payable from promisor to promisee. When those commitments hold, performance and breach-plus-damages are remedially equivalent for the promisee, and the law is indifferent between them; the promisor who can put the subject matter to higher-valued use should therefore breach and pay damages. The boundary this framework reaches is the point where the obligee is not bilateral but collective, the harm is a non-rivalrous club-good externality borne simultaneously by every network member, and no bilateral damages schedule can restore indifference for the actual injured party—the trading community itself.

A. The Doctrine in Its Strongest Form

Oliver Wendell Holmes reduced the law of contract to a prediction: “The duty to keep a contract at common law means a prediction that you must pay damages if you do not keep it—and nothing else.”⁸ If performance and breach-plus-damages are legal equivalents—if the law is indifferent between them because the promisee is indifferent—then a promisor who can put resources to higher-valued uses by breaching and paying damages should do so, because breach in such circumstances increases total welfare without reducing the promisee’s position.

Robert Birmingham articulated the prescription formally.⁹ Charles Goetz and Robert Scott supplied the economic mechanism: expectation damages, set at the level that makes the promisee indifferent between performance and breach, optimally deter inefficient breach while permitting efficient breach to occur.¹⁰ Richard Posner popularized the prescription through memorable hypotheticals and embedded it in the dominant law-and-economics curriculum.¹¹

The doctrine rests on four structural commitments that do the load-bearing work of its indifference claim. First, the relevant parties are bilateral: there is a promisor and a promisee, and third-party effects are assumed negligible or externalized from the analysis. Second, the harm from breach is a recoverable economic loss measurable in money and payable from promisor to promisee. Third, the promisee is the relevant injured party, and making the promisee economically whole through expectation damages exhausts the law’s remedial obligation. Fourth, performance and breach-plus-damages are remedially equivalent from the promisee’s perspective. The doctrine’s welfare claim depends on all four simultaneously. If the first commitment fails—if the relevant injured party is not the bilateral counterparty—the calculus fails at its foundation, because it has been solving for the wrong party.

Daniel Markovits and Alan Schwartz, in their comprehensive defense of expectation damages, add a fifth structural commitment that clarifies the doctrine’s scope: the “dual performance hypothesis” holds that promisees implicitly authorize promisors to choose between performance and a money transfer, so that paying expectation damages is itself a mode of performance rather than a breach.¹² This is the strongest possible defense of the expectation remedy because it eliminates the normative gap between performance and breach-plus-damages entirely. But it depends on the promisee being someone who would be content with a money transfer—someone whose interest in the promisor’s conduct is exhausted by its economic yield. A trading community whose enforcement capital is degraded by breach is not such a promisee.

These commitments share an implicit internalization premise: all socially relevant costs of breach must be capable of being internalized to the bilateral promisor-promisee pair through a money transfer that leaves the promisee indifferent. The premise treats any uninternalized cost as either negligible or irrelevant to the doctrine’s descriptive claim. Where breach instead imposes negative externalities on a club good—excludable to non-members yet non-rivalrous within membership capacity—internalization through bilateral damages becomes structurally impossible.¹³ The club good’s degradation (informational, monitoring, and enforcement-capacity externalities) is borne diffusely by every member simultaneously and cannot be restored by any payment the breaching promisor makes to the single counterparty. In such settings, the indifference and remedial-equivalence premises do not hold for the actual obligee, even if they hold for the bilateral counterparty.

This Article need not resolve the debate over the dual-performance hypothesis. Even if the DPH correctly characterizes bilateral commercial exchange—even if promisees in ordinary contracts do implicitly authorize promisors to elect breach-plus-damages as an alternative mode of performance—the hypothesis presupposes a promisee whose interest in the promisor’s conduct is exhausted by its monetary yield. In a covenantal network, the promisee is the collective, and the interest is the non-rivalrous enforcement capital whose degradation no monetary transfer to any individual counterparty can restore. A trading community that suffers diffuse informational, monitoring, and enforcement-capacity externalities when a member breaches is not the kind of promisee the DPH describes. The covenantal framework applies regardless of whether one accepts the DPH for standard bilateral contracts. The DPH simply has no application to an obligee whose interest cannot be exhausted by a money transfer.

B. The Normative Challenges and the Administrability Retreat

Seana Shiffrin argues that breach is a moral wrong regardless of whether expectation damages are paid, because the promisee’s entitlement is to the specific performance promised, not merely to its monetary equivalent.¹⁴ Charles Fried’s earlier account grounded this intuition in a theory of promissory obligation: a promise creates a duty to perform, not a duty to perform-or-pay, and a legal system that treats these as equivalent fails to honor the moral significance of the promissory relationship.¹⁵ Steven Thel and Peter Siegelman have argued that willful breach warrants distinct treatment, on the ground that deliberate repudiation of a promise wrongs the promisee in a way that accidental non-performance does not.¹⁶

The efficient-breach defenders answer with the administrability retreat. Conceding for argument’s sake that breach may be morally wrongful, courts cannot remedy the wrong more fully than expectation damages allow. Specific performance is administratively costly and productive of inefficient continuance of unwanted relationships.¹⁷ Disgorgement requires calculating the promisor’s profits, which is often less reliable than calculating the promisee’s loss.¹⁸ The administrability retreat is persuasive within its own frame: courts doing their best in bilateral proceedings, with bilateral remedial tools, for bilateral parties before them.

The retreat succeeds against Shiffrin and Fried because both arguments operate inside the bilateral frame. They accept that the promisee is the immediate counterparty; they contest whether that counterparty’s entitlement extends beyond the monetary value of the promised performance. Courts can always respond that even if the promisee deserves more, bilateral proceedings with available tools can deliver only expectation damages. Markovits and Schwartz make the same point with greater precision: the dual-performance hypothesis implies that the promisee contractually authorized the money transfer as an alternative mode of performance, so courts awarding expectation damages are not “settling for less” but enforcing the full contract.¹⁹

This Article sidesteps that normative debate. The prior question is descriptive: have we identified the obligee correctly? The question is not whether the individual promisee deserves more than money—that is Shiffrin and Fried’s concern—but whether the entity suffering the harm has been correctly identified at all. Where the obligee is the collective network and the harm is club-good degradation, the normative debate about bilateral remedies becomes irrelevant. We are solving for the wrong party. The analysis that follows is therefore grounded in group-level institutional dynamics and network externalities, not in individual moral philosophy.

C. The Prior Empirical Question

Both the normative challenge and the defense against it share a structural assumption that neither side has examined: the promisee is the bilateral counterparty. The question of whether the promisee deserves more than expectation damages presupposes that we have identified the promisee correctly. If the entity to whom the obligation runs is not the bilateral counterparty but the trading community as a collective—if the harmed party upon breach is the network of all traders whose monitoring costs rise, whose signaling capital is degraded, and whose enforcement infrastructure is eroded—then expectation damages do not undercompensate. They are simply addressed to the wrong party for the wrong harm.

This is the prior empirical question: is the bilateral-exchange model descriptively accurate for all commercial obligations? Are there commercially significant categories of obligation in which the obligee is a collective entity and the harm from breach is a network externality imposed on a club good rather than a bilateral economic loss?

A concrete illustration clarifies why the internalization premise fails in such settings. Suppose a diamond dealer in a bilateral-pair terminal transaction—a deal with a counterparty he will not deal with again, though both remain members of the same exchange—can realize a $50,000 gain by breaching (selling the same stones to a higher bidder) while the bilateral counterparty’s expectation damages, measured by the difference between contract and market price, are only $30,000. Bilateral theory predicts breach-plus-damages: the promisor captures the $20,000 net surplus, the promisee is made whole, and social welfare increases. The law should therefore be indifferent. Yet Bernstein documents systematic compliance even in settings where this specific pairing will not recur.²⁰ The gap is not measurement error in bilateral damages; it is the existence of network externalities imposed on the club good of network enforcement capital—excludable to non-members, non-rivalrous within membership capacity.²¹

Breach generates three such externalities simultaneously. First, an informational externality: every other trader must now discount the value of “DDC-member” as a reliability signal, raising search and counterparty-selection costs across the entire market. Second, a verification/monitoring externality: once a breach demonstrates that membership no longer guarantees compliance, every member must increase screening expenditures to assess counterparty reliability—costs that were previously internalized by the network’s collective enforcement capital. Third, an enforcement-capacity externality: unsanctioned breach degrades the credibility of the terminal expulsion sanction, making the entire compliance equilibrium more costly to sustain for the network as a whole.

These costs are borne diffusely by every member simultaneously; they are not offset by any payment the breaching dealer makes to the single counterparty. Just as a bilateral tort judgment between two riparian owners cannot restore water quality for the dozens of other users downstream, expectation damages here do not engage, let alone exhaust, the network’s interest in preserving its enforcement infrastructure. The administrability retreat therefore does not save the doctrine for this category: courts are not “doing their best” for the relevant injured party; they are solving for the wrong harm and the wrong obligee altogether. The internalization premise implicit in efficient breach’s indifference claim is simply inapplicable where breach degrades a non-rivalrous club good for the collective obligee.

Steven Shavell has argued that breach of contract is generally not immoral because markets require flexibility and legal enforcement of all promises at face value would impede efficient resource allocation.²² His argument, like the administrability retreat, is calibrated to bilateral exchange. The network-externality analysis does not contest Shavell’s general claim. It argues that certain commercial obligations are not bilateral exchanges at all, and that Shavell’s framework simply does not apply to them.

II. Merchant Networks as Empirical Anomaly

The bilateral model underlying efficient breach doctrine predicts that, in any commercial obligation, a promisor will breach whenever the gain from non-performance exceeds the expectation damages payable to the identified bilateral counterparty. Where the transaction is terminal between the specific parties—where no bilateral future exists and the shadow of future bilateral dealings is zero—the prediction is especially clear: breach-plus-damages should occur whenever it is privately profitable, because no iterated-game mechanism alters the calculus and expectation damages render the promisee indifferent. The doctrine further assumes that any sanction for breach can be commuted into a monetary transfer that restores indifference for the relevant obligee.

Where the data systematically diverge from these predictions, the internalization premise fails. The harm is not a bilateral economic loss but a degradation of the club good of network enforcement capital—excludable to non-members, non-rivalrous within membership capacity—borne diffusely through informational, verification/monitoring, and enforcement-capacity externalities that no bilateral damages schedule can reach.

A. A Contemporary Illustration: Amazon’s Seller Network

The pattern is not a relic of pre-digital commerce. Amazon’s third-party seller network exhibits the compliance dynamic the covenantal thesis predicts: sellers systematically forgo individually profitable deviations from quality standards because the network’s monitoring and expulsion capacity internalizes externalities that bilateral damages between buyer and seller cannot reach. Amazon maintains a club good—consumer trust in product quality, delivery reliability, and dispute resolution—through continuous algorithmic monitoring of seller performance metrics (order defect rate, late shipment rate, valid tracking rate) and permanent suspension of sellers who fall below threshold standards.²³

The structural parallels to Bernstein’s merchant networks are significant. Permanence: a seller’s account history, customer reviews, and Buy Box eligibility accumulate over years and constitute commercial identity within the platform. Unconditionality: Amazon enforces performance standards without regard to bilateral excuses—a seller cannot justify late shipments by citing a particular buyer’s unreasonable expectations, because the obligation runs to the marketplace’s collective reliability signal. Third-party witnessing backed by expulsion: algorithmic monitoring serves the witnessing function, suspension records constitute a global compliance history, and suspension operates as a status event affecting the seller’s standing with every buyer and competitor on the platform simultaneously.

Whether Amazon’s seller network is fully covenantal under the diagnostic criterion requires case-by-case verification rather than presumption. The witnessing function is algorithmic rather than institutionally peer-governed, which raises questions about whether the monitoring capacity is independent in the sense the covenantal framework requires or merely an exercise of platform proprietorship. Permanence is qualified by the possibility of creating new accounts, though Amazon’s increasingly sophisticated identity-verification systems make this progressively more difficult. And suspension may be partially compensable through reinstatement litigation or account transfers, which complicates the non-compensable-expulsion criterion.

Amazon’s seller network is therefore a candidate covenantal network—one that satisfies the structural features in significant respects but whose full diagnostic status requires the case-by-case analysis the framework prescribes. This contemporary illustration is a digital instantiation of the same structural dynamics that four independent research programs have documented in traditional merchant networks. This Part presents that evidence, organized around two direct falsifications of efficient breach’s descriptive claims: terminal-transaction compliance (where bilateral theory predicts defection) and non-compensable expulsion (where bilateral theory predicts that any sanction can be priced).

B. Two Direct Falsifications

The terminal transaction problem is the first direct falsification. Bilateral theory predicts defection in transactions where no bilateral future between the counterparty pair exists. With no future dealings between the specific parties, the promisor faces only the counterparty’s expectation damages; if the gain from breach exceeds that amount, breach-plus-damages increases joint welfare and the law should be indifferent.²⁴ The data contradict this prediction.

Lisa Bernstein’s fieldwork in the New York Diamond Dealers Club documents oral agreements sealed by handshake and the phrase mazel u’bracha, high-value volatile transactions, and compliance rates approaching those of formally enforced contracts, with court filings between members “virtually nonexistent.”²⁵ Many transactions are terminal as between the two parties: a dealer completing a final exchange with a counterparty within the same network has no bilateral future with that specific individual to protect. Yet compliance persists.

Avner Greif’s historical analysis of medieval Maghribi traders documents the same pattern across a different institutional setting: merchants performed on contracts in distant ports where no bilateral future existed and legal enforcement was unavailable, because the community’s multilateral reputation mechanism—observing conduct, sharing information, and collectively sanctioning defection—supplied compliance incentives that bilateral models cannot generate.²⁶ Greif’s game-theoretic formalization proves that the multilateral reputation equilibrium sustains compliance in precisely the terminal-transaction settings where bilateral reputation fails.²⁷

The non-compensable expulsion problem is the second direct falsification. Bilateral theory assumes that any sanction for breach can be commuted into a monetary transfer that restores indifference for the relevant obligee. Where the sanction is non-compensable—where no payment restores membership and the excluded party simply loses market access permanently—bilateral theory’s remedial-equivalence premise fails.

Expulsion from the New York Diamond Dealers Club and the World Federation of Diamond Bourses is non-compensable: no fine or payment restores membership, and the expelled dealer loses access to affiliated exchanges globally.²⁸ NGFA arbitration imposes non-compensable expulsion for persistent non-compliance; the defaulting member is excluded from the national network without monetary compensation for that exclusion, though the underlying trade dispute is resolved by an arbitration award.²⁹ Automotive original equipment manufacturers exclude non-compliant suppliers from the sector-wide sourcing pool without compensating them for the lost business opportunity.³⁰

These are not close cases. They are direct falsifications: bilateral theory predicts compensable sanctions because it assumes the relevant harm is a bilateral economic loss payable from promisor to promisee. The data show non-compensable status sanctions that affect the expelled party’s standing with every network participant, not just the bilateral counterparty whose transaction triggered the expulsion.

C. Grain, Feed, and Cotton Markets: Misidentifying the Obligee

Bernstein’s study of the National Grain and Feed Association documents networks in which the obligee is the trading community rather than the bilateral counterparty.³¹ The NGFA operates an arbitration system whose awards are backed by the threat of expulsion, communicated to all NGFA members and affiliated regional associations. The critical institutional feature is unconditionality: NGFA arbitration does not ask whether both parties behaved badly. The tribunal’s inquiry is limited to whether the respondent breached the trade rules, not whether the complainant also violated norms or whether the breach was provoked. The member’s duty runs to the network as collective obligee, not to the bilateral counterparty as reciprocal promisor.

Nanakuli Paving & Rock Co. v. Shell Oil Co.³² illustrates the mismatch between bilateral legal treatment and covenantal network structure. Shell’s consistent price-protection practice reflected cooperative norms of the Hawaiian asphaltic-paving trade. The Ninth Circuit incorporated that practice as trade usage under UCC section 2-202, over Shell’s written contract language to the contrary. But as Bernstein’s subsequent study of NGFA contracts documents, traders in analogous industries deliberately keep relationship-preserving accommodations outside written agreements and outside courts specifically to prevent judicial incorporation. One grain trader explained the logic with precision: bringing accommodation evidence into a legal proceeding would signal that the norm is a bilateral legal entitlement, collapsing the institutional separation on which its authority depends.³³ Bilateral theory treats such accommodations as potential course-of-performance evidence; the data show traders deliberately keeping them outside bilateral proceedings to preserve the club good.

Bernstein’s cotton industry study confirms the pattern in a fourth independent empirical domain.³⁴ The American Cotton Shippers Association and the Liverpool Cotton Association exhibit the same three covenantal features: career-length membership, collective obligation enforced through association arbitration, and non-compensable expulsion from the trading community. The cotton study is significant for a specific reason: it demonstrates that the covenantal pattern is not an artifact of ethnic or religious community homogeneity. The mechanism arises wherever the institutional conditions obtain, independent of the social characteristics of the membership. Compliance is sustained by the architecture, not by the community’s demographics.

The key insight from the grain and cotton evidence is that the obligee is plural: both the bilateral counterparty (who receives the arbitration award for the underlying transactional breach) and the collective network (whose enforcement capital is protected by the expulsion mechanism that operates independently of bilateral forgiveness). Courts that award expectation damages to the bilateral counterparty without deferring to the network’s expulsion decision are solving for the wrong harm—or more precisely, for only one of two harms, while rendering the second invisible.

D. Automotive Procurement: Graduated Status Enforcement

Bernstein’s subsequent procurement study, and Richman’s extension of the diamond analysis to global supply networks, document the same institutional pattern in automotive original equipment manufacturer supply chains.³⁵ OEMs maintain networks of suppliers competing for sourcing contracts. The enforcement system is graduated: warnings, reduced contract volume, exclusion from new sourcing opportunities, and ultimately expulsion from the sector-wide sourcing pool.

The expulsion is not bilateral. An OEM that excludes a supplier from its own sourcing does not merely end a bilateral relationship; it communicates the exclusion through procurement-manager networks and industry associations to other OEMs. The excluded supplier loses access to the entire automotive-procurement sector, not just the specific OEM that imposed the sanction.³⁶ This is third-party witnessing backed by expulsion: an institutionalized authority—the network of procurement managers sharing performance histories—observes compliance, records it accessibly, and imposes sanctions affecting the supplier’s standing with all network participants simultaneously.

The automotive network exhibits all three covenantal features. Permanence: suppliers make relationship-specific investments in tooling, training, and co-location; obligations persist across multiple sourcing cycles and the membership constitutes commercial identity within the sector. Unconditionality: OEMs enforce sector-wide compliance standards without inquiring whether the supplier’s breach was provoked by the OEM’s own conduct; the focus is sector-level compliance, not bilateral reciprocity. Third-party witnessing: performance histories are shared across the procurement-manager network, and exclusion from one OEM’s sourcing pool affects the supplier’s standing with all OEMs simultaneously.

This graduated status enforcement exhibits Ostrom’s design principles for successful common-pool governance: monitoring capacity through shared compliance histories and graduated sanctions sustain the club good of network trust.³⁷ Bilateral theory predicts reliance on court-enforceable terms or bilateral reputation; the data show compliance sustained by the network’s institutional memory, whose degradation would impose diffuse externalities on all members’ transaction costs.

Richman’s observation that some DDC functions declined as parallel institutions developed is itself evidence for the covenantal thesis rather than against it: when the club good erodes—whether through institutional competition, regulatory displacement, or internal decay—compliance erodes with it, precisely as the framework predicts. The covenantal account does not require any particular network to persist indefinitely; it requires only that wherever the three features obtain, the club-good externalities arise and bilateral internalization fails.

E. The Scope of the Descriptive Claim

The empirical claim is limited and conditional. This Article does not assert that most commerce is covenantal, nor that bilateral contract theory is generally false. The claim is that there exist networks—identifiable through the three-feature diagnostic—where the obligee is collective rather than bilateral, the harm is club-good degradation rather than bilateral economic loss, and efficient breach doctrine’s indifference premise fails as a descriptive matter.

The networks Bernstein studied—diamond, cotton, grain-feed, and automotive procurement—are not marginal. They process billions of dollars in annual trade and have operated for decades without systematic resort to legal enforcement. The pattern is not industry-specific; it arises wherever the institutional conditions obtain.

The boundary is sharp. Networks lacking permanent membership, conditional obligations, or expulsion authority remain within efficient breach’s domain. A spot market with anonymous traders, a bilateral supply contract with no network affiliation, or a relational contract enforced through bilateral reputation rather than collective expulsion—all are accurately described by bilateral theory. The covenantal framework applies only where all three features are present and verifiable at the pleading stage through documentary evidence: membership agreements, arbitral body rules and bylaws, and evidence of expulsion authority.

Contemporary platform networks exhibit structural parallels that the three-feature diagnostic can assess, though their covenantal status requires case-by-case verification rather than categorical presumption. The framework is not a historical curiosity. It is a recurring structural response to the club-good externality problem, and Part IV now develops the theoretical account that explains why.

III. The Covenantal Framework

Bilateral theory, including its efficient-breach variant, predicts defection in terminal transactions and assumes any sanction can be commuted into a bilateral damages payment that restores indifference for the identified promisee. Part III showed that the merchant-network data contradict both predictions: compliance persists in terminal transactions, and expulsion is non-compensable. This Part develops the conceptual framework that organizes those divergences.

The framework is descriptive and institutional, not normative or moral. It draws on Henry Maine’s historical observation that the movement from status to contract is incomplete, operationalizes three observable features drawn from Brinig and Nock’s analysis of covenant relationships, and maps those features directly onto the economic character of the club-good externality problem. It applies only where the diagnostic criterion is met: the existence of a recognized arbitral body with expulsion authority to which the parties have agreed to submit disputes—an ascertainable fact at the pleading stage.

A. The Status-Event Concept

Expulsion from a trading network is a status event, not a contractual breach, and the distinction matters for the internalization premise of efficient breach. In a pure contractual setting, breach affects only the bilateral parties’ resource positions, which damages can rebalance. In a status-based network, breach degrades the collective recognition that membership confers—the shared signal of reliability that reduces transaction costs for all members simultaneously. That degradation is a club-good externality, non-rivalrous within the membership boundary yet degraded for every member by each unsanctioned breach. No bilateral damages payment restores the status or the collective enforcement capital it sustains.³⁸

Henry Maine observed in Ancient Law that “the movement of the progressive societies has hitherto been a movement from Status to Contract.”³⁹ Maine described a historical tendency, not a completed or irreversible process. Efficient breach doctrine presupposes that the tendency has run to completion for commercial obligation: all such obligations arise from bilateral voluntary exchange, run to an identified counterparty, and can be discharged by a monetary payment that leaves that counterparty indifferent. The merchant-network evidence shows the presupposition is incomplete. Membership in the diamond exchange or NGFA constitutes a trader’s professional identity over an extended career horizon; expulsion dissolves that identity rather than merely transferring resources.

Maine’s observation is used here solely for its descriptive historical insight that status-based obligation can persist within modern commercial economies. No normative claim about the relative desirability of status versus contract follows from the analysis.

B. The Three-Feature Diagnostic

The covenantal framework operationalizes three structural features that distinguish these obligations from bilateral contracts. The features are drawn from the sociological literature on covenant relationships, of which Margaret Brinig and Steven Nock’s analysis of covenant marriage statutes is among the most careful.⁴⁰ Brinig and Nock’s work is an empirical study of divorce rates under covenant-marriage statutes; the three-feature diagnostic as applied to commercial networks is this Article’s analytical extension, not a direct borrowing of their theoretical architecture for family law.⁴¹ The features are observable institutional facts, verifiable at the pleading stage by the presence of a recognized arbitral body with expulsion authority.

Permanence means that network membership is not terminable at will without consequence and that obligations persist across the member’s tenure rather than resetting with each discrete transaction. In the diamond bourse or NGFA, a trader’s compliance history accumulates over a career and defines professional identity within the club. This feature sets the temporal boundary of the club good: degradation of enforcement capital is persistent rather than one-off, so that each breach imposes long-term informational and monitoring externalities on all current and future members.

Unconditionality means that duties run to the network as collective obligee rather than as reciprocal consideration for bilateral performance. The exchange tribunal does not inquire whether the counterparty also behaved badly or whether bilateral reciprocity would have excused non-performance. This feature identifies the harm as collective: breach degrades the shared enforcement resource for every member simultaneously, generating enforcement-capacity externalities that cannot be offset by any bilateral transfer.

Third-party witnessing backed by expulsion means that an institutionalized authority observes compliance, records it accessibly to multiple members, and imposes sanctions affecting the member’s standing with the entire network independently of whether the bilateral counterparty forgives the breach. An expelled diamond dealer loses market access even if his counterparty would have accepted performance or damages. This feature supplies the non-replicable monitoring and graduated-sanction technology that manages the club-good externalities internally. Expulsion is the terminal sanction whose credibility sustains the entire enforcement capital stock.

These features are not characterizations imposed on the data; they are what Bernstein, Greif, and Ostrom document when the bilateral lens is removed. The diagnostic triggers on the institutional fact of a recognized arbitral body with expulsion authority—an ascertainable pleading-stage criterion that correlates with the three features without presupposing the covenantal conclusion.⁴²

C. A Typology of Commercial Obligation

The three-feature framework generates a diagnostic typology courts can apply without circularity.

Ordinary bilateral contracts involve identified parties exchanging reciprocal promises for consideration, obligations bounded by the discrete transaction, breach excusing the non-breaching party’s further performance, and bilateral expectation damages as the operative remedy. Bilateral theory and efficient breach describe this category accurately; the internalization premise holds because all relevant costs are bilateral.

Relational contracts in Macneil’s sense involve ongoing relationships with implicit norms, relationship-preserving accommodations, and compliance sustained partly by expectation of future bilateral dealings. These obligations exceed discrete bilateral terms but retain a bilateral remedial horizon. The relevant bilateral counterparties are identifiable, and norms emerge from the relationship itself rather than from an institutional governance structure independent of the bilateral pair.

Covenantal obligations arise where all three structural features are demonstrably present: career-length membership, unconditional duty running to the collective, and third-party witnessing backed by expulsion. The obligee is the trading community. The harm from breach is degradation of the club good of enforcement capital. No bilateral damages schedule reaches the diffuse informational, monitoring, and enforcement-capacity externalities the breach imposes on non-party members.

The diagnostic difference between relational and covenantal obligation is structural, not quantitative. Where compliance would persist in a bilateral-pair terminal transaction—between network members who will never deal with each other again—by virtue of the network’s expulsion threat rather than the shadow of future bilateral dealings, and where expulsion is non-compensable, the obligation is covenantal. This criterion is falsifiable: if terminal-transaction compliance rates within a claimed covenantal network do not significantly exceed those in comparable bilateral markets, the covenantal characterization is not supported.

The original article’s comparative table mapping each category onto the DDC, NGFA, and ordinary bilateral settings is reproduced in the Technical Appendix; the prose typology above is sufficient for the doctrinal analysis that follows. This account applies only where the three features are demonstrably present; it leaves relational and bilateral categories untouched.

D. Distinguishing Relational Contract Theory

Ian Macneil’s relational-contract theory is the leading account of how commercial obligation exceeds the discrete bilateral model.⁴³ Macneil correctly observed that long-term relationships generate solidarity norms, role integrity, and preservation-of-relation norms that operate independently of written terms. Stewart Macaulay’s earlier empirical work similarly documented non-contractual relations in business.⁴⁴ Robert Scott has shown that courts often err when incorporating relational norms, leading sophisticated parties to draft more rigidly.⁴⁵

The covenantal framework builds on these insights while marking a narrower domain and reaching a contrary prescriptive conclusion on judicial incorporation. The two accounts diverge in three specific and consequential respects.

First, relational-contract theory retains a bilateral remedial horizon and favors judicial incorporation of relational norms as a way to honor commercial practice. The covenantal account reaches the opposite prescriptive conclusion: incorporation claims the witnessing role for the court, collapsing the institutional separation on which the norms’ binding authority depends and exporting the three club-good externalities back onto the membership.

Second, relational theory does not predict terminal-transaction compliance. Its mechanisms still rely on some bilateral or ongoing relational horizon; the covenantal frame explains compliance in terminal settings because the obligation runs to the network whose permanence and third-party witnessing internalize the externalities independently of any bilateral future.

Third, relational theory does not theorize the institutionalized observation and expulsion capacity as the source of norm authority. Macneil saw community solidarity; the covenantal frame identifies the non-replicable monitoring and graduated-sanction technology that manages the club good internally. This mechanism explains why Nanakuli-style incorporation predictably reduces the supply of discretionary accommodations—an effect neither bilateral reputational models nor relational-contract theory predicts.

The two accounts are complementary rather than competing. Relational theory identifies the richness of commercial relationships; the covenantal frame isolates the subset in which the club-good externality problem requires a triadic institutional solution. This distinction is falsifiable: covenantal networks should display terminal-transaction compliance and reduced accommodations after judicial incorporation—effects relational theory does not predict.

E. Distinguishing Transaction Cost Economics

Oliver Williamson’s transaction cost economics offers a distinct and powerful alternative explanation for the institutional structures documented in Part III: sophisticated parties facing the risk of bilateral hold-up—where relationship-specific investments expose one party to opportunistic renegotiation by the other—choose governance structures with credible commitment devices to protect those investments.⁴⁶ Under TCE, the DDC’s expulsion mechanism is such a device: it deters bilateral opportunism by raising the cost of defection for any member who has sunk reputational capital into network membership. Membership itself is the relationship-specific asset; the DDC’s enforcement authority is the governance structure that protects it against bilateral hold-up.

The covenantal account does not dispute this analysis as far as it goes. TCE correctly identifies asset specificity as a governance-choice driver and correctly predicts that parties facing hold-up risk will select credible commitment technologies. Where TCE differs from the covenantal account is in the direction of its remedial logic, which remains fundamentally bilateral: the safeguarding arrangement is selected by the bilateral parties to protect each other’s relationship-specific investments against each other’s opportunism. The analysis remains promisor-promisee dyadic.

The covenantal account’s divergence is sharpest in terminal transactions, where—by definition—neither party has made relationship-specific investments with respect to the other, so TCE’s bilateral safeguarding rationale is absent.⁴⁷ Yet compliance persists. The covenantal account explains compliance through the network-externality mechanism rather than bilateral hold-up protection: the obligation runs to the collective network whose enforcement capital is degraded by breach, independently of the bilateral parties’ investment positions.

The two frameworks are complementary at the level of institutional design. TCE explains why merchants build expulsion-backed governance structures; the covenantal account explains why those structures generate obligations that run to the collective rather than to the bilateral counterparty, with the remedial implications that follow.

F. Network Externalities as Economic Foundation

The externalities literature establishes that bilateral legal rules systematically fail when harms are diffuse and fall on non-parties that bilateral remedies cannot reach.⁴⁸ Network breach in covenantal obligations generates precisely such harms to the club good of enforcement capital—excludable to non-members (non-members cannot trade on the membership signal) yet non-rivalrous within membership capacity (one trader’s use of the signal does not diminish others’ use, at least until capacity is strained by widespread defection).⁴⁹

Breach imposes three specific externalities. First, informational: each unsanctioned breach devalues the membership reliability signal, raising search and verification costs for every member seeking to rely on that signal in future transactions. Second, verification and monitoring: once a breach demonstrates that membership no longer guarantees compliance, every member must increase screening expenditures to assess counterparty reliability—costs that were previously internalized by the network’s collective enforcement capital. Third, enforcement-capacity: unsanctioned breach degrades the credibility of the terminal expulsion sanction, making the entire compliance equilibrium more costly to sustain for the network as a whole.

The three covenantal features manage these externalities internally. Permanence ensures that long-term degradation affects the entire club boundary rather than any single bilateral relationship. Unconditionality prevents bilateral reciprocity from offsetting collective harm—a promisor cannot escape the network’s sanction by pointing to the counterparty’s prior bad conduct. Third-party witnessing supplies the monitoring and graduated-sanction technology whose design principles Ostrom identified in her study of common-pool governance.⁵⁰

The Ostrom-Buchanan analogy requires a precise statement. Ostrom’s design principles were developed for common-pool resources—non-excludable and rivalrous goods. The enforcement capital at stake in covenantal networks is excludable and non-rivalrous: a Buchanan club good, not an Ostrom commons. The analogy runs to the institutional design mechanisms (monitoring, graduated sanctions, defined boundaries), not to the underlying resource type. The formal result is that bilateral judicial intervention—by claiming the witnessing role for the court—degrades the detection probability, weakens the compliance condition, and reduces the enforcement capital stock through a cascading institutional failure that bilateral remedy logic cannot perceive because it focuses on the individual dispute rather than the network’s governance infrastructure.

IV. Doctrinal Applications

The covenantal framework translates directly into four recurring legal contexts. Each prescription follows mechanically from the three-feature diagnostic and the club-good externalities it produces. Courts apply the framework in two steps: (1) determine at the pleading stage whether a recognized arbitral body with expulsion authority exists (the diagnostic trigger); and (2) if the trigger is satisfied, shift to the covenantal rules set out below. Where the trigger is absent, ordinary bilateral doctrine governs unchanged.

A. UCC §§ 1-303 and 2-202 Incorporation: Rebuttable Presumption Against Judicialization

When a party seeks to incorporate a network norm (course of performance, course of dealing, or trade usage) to override or supplement written terms, courts normally apply UCC §§ 1-303 and 2-202. The covenantal framework adds a rebuttable presumption against incorporation once the diagnostic trigger is met. The presumption rests on the displacement-of-witnessing mechanism: judicial incorporation converts a collectively governed norm into a bilateral legal entitlement, exporting verification/monitoring externalities to the entire membership and discouraging future accommodations.

Step-by-step application

1. Confirm the diagnostic trigger: a recognized arbitral body with expulsion authority to which the parties agreed.

2. Apply the presumption: the court treats the network’s deliberate maintenance of institutional separation (e.g., keeping accommodations off-paper) as the functional equivalent of “careful negation” under UCC § 1-303(e).

3. Rebuttal opportunity: the party seeking incorporation must produce clear evidence that the parties specifically intended judicial enforcement of the norm as a bilateral term.

4. If unrebutted, the court excludes the norm from the contract; the network’s internal tribunal retains interpretive primacy.

Nanakuli Paving & Rock Co. v. Shell Oil Co.⁵¹ illustrates the error the presumption prevents. The Ninth Circuit incorporated Shell’s price-protection practice as trade usage despite contrary written terms. Under the covenantal rule, the Hawaiian asphaltic-paving network’s norm would have triggered the presumption; absent explicit contractual consent to judicial enforcement, the accommodation would remain outside the contract, preserving the club good for all participants.

B. Arbitration and Expulsion Review: Procedural-Regularity Deference

Federal courts review trade-association expulsion decisions under the Federal Arbitration Act (FAA). The covenantal framework supplies the precise standard of review once the diagnostic trigger is met: procedural-regularity deference. Courts ask only whether (1) the tribunal followed its own announced procedures, (2) the member received adequate notice and a genuine opportunity to respond, and (3) the stated basis for expulsion bears a rational connection to a legitimate network-governance interest (preservation of enforcement capital). Courts can operationalize this deference via three steps:

1. Confirm the diagnostic trigger.

2. Limit review to the three procedural prongs above; substantive merits are off-limits.

3. If any prong fails, vacate under 9 U.S.C. § 10; otherwise, confirm the award.

This standard flows directly from Silver v. New York Stock Exchange⁵² and aligns with Stolt-Nielsen, Oxford Health, and Hoffman v. Cargill.⁵³⁵⁴⁵⁵ It protects the expelled member against arbitrary process while leaving the network’s witnessing function intact.

C. Antitrust Treatment of Covenantal Expulsion: Two Distinct Analyses

Covenantal expulsion raises Sherman Act § 1 concerns. The framework separates two analytically distinct questions:

Category 1 – Challenge to the governance mechanism itself. Courts apply rule-of-reason analysis with a presumption of legitimacy. Expulsion is the terminal sanction that sustains the enforcement-capacity externality; disabling it exports three club-good externalities to public courts. The presumption is rebutted only by clear evidence that the expulsion was not a genuine governance sanction.

Category 2 – Challenge to market power. Where the network also possesses monopoly or monopsony power, courts conduct full market-structure scrutiny under the rule of reason. Covenantal deference protects the expulsion tool; it does not immunize anticompetitive use of that tool. The two analyses run in parallel: procedural-regularity review for the governance decision, antitrust review for market effects.⁵⁶

Northwest Wholesale Stationers⁵⁷ and Silver⁵⁸ supply the doctrinal anchors. The framework clarifies when each line of cases applies.

D. Preliminary Injunctions Against Expulsion: Heightened Showing Requirement

Courts asked to enjoin expulsion pending litigation apply the Winter factors.⁵⁹ In covenantal settings the framework imposes a heightened showing requirement tied to the club-good externalities:

1. Likelihood of success: plaintiff must demonstrate procedural irregularity under the three-prong standard in Part V.B. Substantive disagreement is insufficient.

2. Irreparable harm: plaintiff’s harm is concrete; the network’s harm (degraded enforcement credibility) is collective and must be weighed.

3. Balance of equities: court considers the diffuse harm to non-party members whose transaction costs rise if expulsion loses credibility.

4. Public interest: preserving private enforcement mechanisms that internalize externalities courts cannot efficiently manage.

A preliminary injunction issues only if the plaintiff satisfies all four factors after the heightened showing on likelihood of success. This rule prevents premature judicial override of the witnessing function while still protecting members against arbitrary process.

V. Regulatory Consequences: The Club-Good Cost of Bilateral Intervention

Just as common-law doctrines can erode network trust by importing bilateral remedies into triadic governance settings, public regulatory regimes often intervene in commercial relationships under the assumption that they are strictly bilateral, power-imbalanced contracts. When regulators mandate expanded due process for termination, portability of reputational capital, or good-cause requirements before expulsion, they attempt to protect an identified individual. When those tools are applied to covenantal networks, however, they predictably trigger the degradation of the underlying club good by undermining the expulsion finality that sustains the enforcement-capacity externality. The network cannot internalize the costs of free-riding membership; it must substitute expensive ex ante screening for efficient ex post expulsion, contracting the club good’s value for all compliant members.

This Part examines two regulatory domains where this pattern is already visible: franchise good-cause termination statutes and platform worker deactivation regulation. Both illustrate the consequence of applying bilateral intervention logic to triadic governance structures, and both generate natural experiments for testing Hypothesis H3 in Part VII.

A. Franchise Good-Cause Termination Statutes

Franchising is a modern, highly formalized covenantal network. The franchisor and franchisees are bound by long-term contracts with high exit costs (permanence), owe duties to maintain brand standards that protect the reputation of all franchisees simultaneously (unconditionality to the collective), and are subject to systematic monitoring backed by the threat of termination (third-party witnessing and expulsion).⁶⁰ The brand itself is the club good: excludable (only franchisees may use it), non-rivalrous across geographic territories (one franchisee’s use does not diminish another’s), and degradable by free-riding members who cut corners on cleanliness, product quality, or service standards.

A franchisee who cuts corners on any of these dimensions degrades the brand’s reputation—a negative externality borne by every other franchisee in the system who invested in the same brand signal. The degradation is non-rivalrous (it affects all franchisees simultaneously) and, from the network’s perspective, non-compensable through bilateral remedies: the compliant franchisee in the next territory over cannot sue the corner-cutting franchisee for diminution of brand value. The network’s only effective tool is swift expulsion before the degradation spreads.

Beginning in the 1970s, many states enacted franchise relationship laws prohibiting franchisors from terminating a franchise agreement without “good cause,” typically defined as a material breach with a right to cure over thirty to ninety days.⁶¹ The regulatory logic was bilateral: legislators assumed franchisors possessed superior bargaining power and might opportunistically terminate profitable franchisees to capture locations—the “cream-skimming” hypothesis. Whatever validity this hypothesis has in some franchise contexts, the intervention is structurally mismatched to the triadic reality of franchise networks.

By imposing a good-cause requirement and mandating a right to cure, state statutes make it vastly more expensive and time-consuming for franchisors to expel free-riding members, effectively forcing the network to tolerate degradation of the club good that compliant franchisees cannot individually remedy. The formal model in the Technical Appendix captures this dynamic through the regulatory constraint parameter phi: as phi increases (representing stronger good-cause requirements), the effective sanction probability falls, the compliance condition weakens, formal litigation increases, and the enforcement capital stock degrades. Proposition 5 of the Appendix proves that each of these effects is strictly monotonic in phi.

The empirical evidence is consistent with the predicted effect. Klick, Kobayashi, and Ribstein’s panel-data study of fast-food establishments found that state laws restricting franchisor termination rights correlate with statistically significant reductions in franchising activity and franchise employment.⁶² Because franchisors could not efficiently manage the club good of brand integrity through swift expulsion, they curtailed network expansion or shifted to less efficient, company-owned models—effectively capturing internally what the franchise network would have governed collectively. The covenantal framework explains this result: bilateral regulatory intervention eroded the enforcement-capacity externality by impairing expulsion finality, forcing the substitution of costly ex ante screening and internal ownership for efficient ex post sanction.⁶³

B. Platform Worker Deactivation and Castellanos v. State of California

A structurally identical conflict has emerged in the regulation of gig-economy platforms. Digital platforms like Uber and Lyft are candidate covenantal networks whose diagnostic status requires explicit application of the three-feature test rather than presumption. The platform maintains a club good—consumer trust in safety and service reliability—through continuous algorithmic monitoring of driver ratings and permanent deactivation of drivers who fall below threshold performance metrics.

Applying the three-feature diagnostic reveals both structural parallels and significant complications. Permanence is contested: unlike the DDC or NGFA, drivers can potentially re-enter the network under a new account, which qualifies the career-defining permanence the full diagnostic requires. Unconditionality is substantially present: platforms enforce performance standards without regard to bilateral excuses, and the obligation runs to the platform’s collective consumer-reliability signal rather than to any bilateral counterparty. Third-party witnessing is algorithmic rather than institutionally peer-governed, which raises questions about whether monitoring is independent or merely an exercise of platform proprietorship. Whether deactivation is non-compensable is subject to litigation risk: if drivers can challenge deactivation in court and obtain reinstatement, the terminal sanction loses the finality that sustains the enforcement-capacity externality.

Digital platforms are therefore best treated as candidate covenantal networks requiring case-by-case diagnostic application, not presumptive members of the category. In some configurations—a dominant platform in a thin market with credible, non-reversible deactivation—the three features may be substantially satisfied. In others—competitive multi-platform markets with reinstatement litigation—the features are insufficiently present for the covenantal framework to apply.

California’s Assembly Bill 5 (AB5), enacted in 2019, reclassified app-based drivers as presumptive employees under the ABC test, subjecting driver deactivations to standard labor-law just-cause requirements.⁶⁴ The regulatory logic was bilateral: the perceived power imbalance between platform corporations and individual drivers justified protective employment status. But where a platform network satisfies the three covenantal features, bilateral employment-law deactivation requirements produce exactly the degradation of the enforcement-capacity externality the covenantal account predicts: if platforms must litigate good cause in public proceedings for each deactivation, the cost of maintaining the witnessing function becomes prohibitive and the club good of consumer safety is exported to courts rather than internally managed.

The platform industry recognized that deactivation finality was constitutive of the club-good governance structure and campaigned for Proposition 22, a ballot initiative maintaining app-based driver status as independent contractors while providing a discrete benefits floor. Castellanos v. State of California upheld Proposition 22 unanimously.⁶⁵ The California Supreme Court’s opinion reasoned primarily in labor-classification terms, but the structural logic underlying its conclusion aligns with the covenantal account: preserving the platforms’ deactivation authority preserved the institutional mechanism on which consumer trust depends, while bilateral employment-law just-cause requirements would have imposed a governance structure designed for dyadic employment relationships onto a network that cannot function without expulsion finality.

A related regulatory proposal—portability mandates requiring platforms to make driver ratings transferable across competing networks—would cause an analogous degradation through a different mechanism.⁶⁶ Portability removes the non-compensable character of expulsion: a driver permanently deactivated from one platform for unsafe behavior could transfer a curated rating history to a competitor network, eliminating the finality on which the enforcement-capacity externality depends. The covenantal framework predicts that portability mandates will increase monitoring costs on all platforms (each must now re-screen incoming drivers whose rating history may be curated rather than genuine) and reduce the credibility of deactivation as a deterrent.

The lesson from both franchise and platform settings is the same: when regulators identify a bilateral power imbalance and impose bilateral protective remedies, they may inadvertently destroy the triadic governance mechanism that makes the network valuable to all participants—including the very members those remedies are designed to protect. A platform or franchise network that cannot efficiently expel free-riders must either expand ex ante screening costs, contract the club-good membership to a level where internal monitoring is manageable, or abandon the network governance structure for corporate hierarchy. None of these outcomes benefits the franchise or platform participants whom good-cause statutes and employment-reclassification mandates were designed to help. This conclusion is conditional: it applies only where the three covenantal features are present in the specific platform or franchise context, and the degree to which they are present requires the case-by-case diagnostic application developed above.

VI. Empirical Predictions and a Research Agenda

The covenantal framework does not rest the descriptive boundary claim solely on the empirical patterns surveyed in Part III. Those patterns supply the anomaly; the framework supplies the institutional mechanism—permanence defining the club boundary over extended horizons, unconditionality making harm collective, and third-party witnessing supplying non-replicable monitoring and graduated sanctions—that explains why bilateral remedies cannot internalize the three club-good externalities when the diagnostic criterion is met. From this mechanism follow four falsifiable predictions that distinguish the covenantal account from standard bilateral reputational models and from relational-contract theory.

The predictions are hypotheses that follow if the structural account is correct; they are not claims that existing data already confirm them. They are testable with arbitration records, trader surveys, and natural experiments created by regulatory interventions. The agenda acknowledges the age of Bernstein’s primary fieldwork and treats the covenantal frame as a lens for new empirical inquiry rather than a conclusive proof based on a handful of classic studies.

A. Hypotheses Distinguishing Covenantal from Bilateral Reputational and Relational-Contract Models

H1 (Terminal-Transaction Compliance). Bilateral reputational models predict measurably lower compliance rates in terminal transactions, where the shadow of future bilateral dealings is zero and expectation damages can render the counterparty indifferent.⁶⁷ Relational-contract theory does not generate a contrary prediction because it still relies on some bilateral or ongoing relational horizon. The covenantal account predicts that compliance rates will be statistically indistinguishable between terminal and non-terminal transactions where the network’s expulsion authority is credible. Breach degrades the club good for all members simultaneously through the three externalities; the obligation runs to the network, whose witnessing function internalizes those costs independently of any bilateral future. This prediction is also a direct test between the covenantal account and TCE’s bilateral-safeguarding mechanism: TCE predicts weakening compliance in terminal transactions, while the covenantal account predicts no such weakening. H1 is testable by comparing arbitration records that distinguish terminal from ongoing-relationship disputes with matched bilateral-market litigation rates.

H2 (Judicialization and Accommodation Supply). In covenantal networks, judicial incorporation of relationship-preserving norms under UCC sections 1-303 and 2-202 should produce a measurable decrease in the frequency of voluntary accommodations compared to matched markets that preserve institutional separation. Incorporation claims the witnessing role for the court, converting discretionary community gestures into bilateral legal liabilities and exporting the verification/monitoring externality to all members. Bilateral reputational models treat incorporation as neutral or efficiency-enhancing; relational-contract theory predicts neutral or beneficial clarification. The covenantal account predicts the supply reduction through the displacement of the third-party witnessing feature. It is testable through trader surveys or analysis of contract amendments in NGFA-style networks versus purely bilateral commodity markets.

H3 (Regulatory Portability and Litigation Effects). Regulatory interventions that weaken expulsion finality—reputational portability mandates or statutory good-cause termination requirements—should correlate with increased formal litigation rates and decreased internal arbitration utilization in affected covenantal networks. By undermining the enforcement-capacity externality, such rules export the club-good degradation to courts or competing networks and reduce the credibility of network sanctions. Neither bilateral nor relational models predict this litigation increase; the covenantal account does because portability and good-cause requirements eliminate the finality that sustains the non-rivalrous enforcement capital within the club boundary. Testable through difference-in-differences analysis of California’s AB5 experience or the patchwork of state franchise good-cause statutes enacted since the 1970s.⁶⁸

H4 (Network Density and Efficient-Breach Rates). Markets with higher network density—measured by membership concentration, arbitral-body utilization, and expulsion incidence—should display lower rates of privately profitable breach in situations where bilateral expectation damages would fully compensate the counterparty. Higher density strengthens all three covenantal features, enhancing the network’s capacity to internalize the three externalities internally; the collective cost of degrading the shared enforcement capital exceeds the bilateral private surplus. Bilateral models predict efficient breach whenever the promisor’s gain exceeds bilateral damages regardless of density. This prediction is testable through survey research of commercial actors or comparison of court-filing rates across industries with and without recognized arbitral bodies.

B. What Would Falsify the Covenantal Account

The covenantal framework generates not only positive predictions but specific falsification conditions. H1 fails—and the covenantal account requires significant qualification—if terminal-transaction breach rates within established covenantal networks are significantly and systematically higher than non-terminal-transaction rates, demonstrating that network externalities do not in fact sustain compliance independently of bilateral future dealings. H2 fails if post-incorporation accommodation supply in matched covenantal networks does not decrease relative to preserved-separation controls, suggesting that the displacement-of-witnessing mechanism overstates the court’s substitution effect. H3 fails if regulatory interventions weakening expulsion finality do not correlate with increased formal litigation, suggesting that the enforcement-capacity externality is less central to network governance than the framework claims. H4 fails if network density does not predict lower efficient-breach rates, suggesting that the club-good degradation mechanism is not in fact load-bearing for compliance.

These falsification conditions are design specifications for the research agenda below, not concessions about the framework’s current standing. The empirical patterns documented in Part III are consistent with the covenantal account; the hypotheses and their falsification conditions are the structured predictions that determine whether that consistency reflects correlation or confirmation. Only accounts that could be wrong are capable of being right.

C. Research Design

Testing these hypotheses requires data not compiled in standard public databases but in principle observable. The NGFA maintains arbitration records accessible to researchers under appropriate non-disclosure agreements; diamond-industry records through the Diamond Dealers Club and World Federation archives offer another source; Bernstein’s cotton-industry data provide a third. Quasi-experimental designs are available through regulatory shocks: AB5 and related gig-platform rules, or the variation in state franchise termination statutes. Difference-in-differences comparisons of litigation rates before and after such interventions in covenantal versus non-covenantal industries would isolate H3. Direct elicitation from commodity traders, automotive procurement managers, and platform operators about decision-making in hypothetical terminal-transaction scenarios—where the parties will not deal again bilaterally, with varying expulsion credibility—could separate network externalities from bilateral reputation effects.

Digital platforms provide contemporary settings in which the extent of genuine non-compensable expulsion can be measured and the three features assessed, though the degree of genuine non-compensability requires case-by-case verification before these settings are treated as covenantal rather than merely reputational.⁶⁹ Decentralized autonomous organizations with on-chain governance and wallet-level expulsion through smart-contract execution present a candidate research setting. Permanence in DAOs can be measured by wallet-age and governance-participation history—proxies for the career-length membership that defines the club boundary in traditional merchant networks. But pseudonymity and the possibility of wallet abandonment qualify this feature: a member can exit one wallet and create another, potentially circumventing the permanence condition in ways that a diamond dealer leaving the DDC cannot. DAOs therefore present the covenantal diagnostic’s strongest contemporary test case for the witnessing and expulsion features, with unconditionality as the open empirical question.

This Article treats the Bernstein and Richman studies as the initial anomaly that reveals the limits of the bilateral model and supplies a structural framework for contemporary and future empirical inquiry.⁷⁰ Whether covenantal networks remain common, are declining under regulatory pressure, or are re-emerging in digital contexts are open questions the framework is designed to help answer.

VII. Objections

The covenantal framework identifies a narrow descriptive boundary rather than a universal normative critique. Four anticipated objections merit direct engagement. A fifth—the discrimination objection—is addressed separately given its distinct character.

A. “This Is Just Relational Contracting with Different Vocabulary”

The strongest potential objection is that the covenantal account merely relabels relational-contract theory. Both approaches acknowledge that commercial relationships often involve norms beyond written terms, that compliance can exceed what bilateral enforcement explains, and that trading communities play a role in sustaining performance.

Part IV.D addresses this objection in full. The frameworks diverge in three specific, consequential respects that the club-good externalities analysis makes visible. First, the covenantal account reaches the opposite prescriptive conclusion on judicial incorporation: incorporation claims the witnessing role for the court, collapsing the institutional separation on which the norms’ binding authority depends and exporting the three club-good externalities back onto the membership, rather than honoring commercial practice. Second, the covenantal account predicts terminal-transaction compliance that relational theory cannot predict. Third, the covenantal account theorizes the institutionalized observation and expulsion capacity as the source of norm authority—a mechanism Macneil did not supply and that explains why Nanakuli-style incorporation predictably reduces the supply of discretionary accommodations, an effect relational theory does not predict.

The two accounts are complementary rather than competing, but the distinction is falsifiable: covenantal networks should display terminal-transaction compliance and reduced accommodations after judicial incorporation—effects relational theory does not predict.

B. “The Empirical Foundation Is Dated and Industry-Specific”

Bernstein’s primary diamond fieldwork dates to the early 1990s; Richman’s extensions to the early 2000s. The strongest version of this objection is not merely that the data are old, but that the institutional conditions Bernstein documented may no longer obtain. Digital commerce has disaggregated supply chains, regulatory intervention has expanded judicial oversight of private governance, and globalization has created competitive alternatives to traditional merchant networks. If these developments have eroded the very covenantal structures the Article describes, the framework may be historically accurate but contemporarily irrelevant—a theory of institutions that have already disappeared.

The objection fails to reckon with the breadth and trajectory of the evidence. Bernstein’s own subsequent work extends the pattern well beyond the DDC: her 2001 cotton study documents the same three covenantal features in a commodity market with no ethnic or religious community homogeneity, refuting the claim that the phenomenon depends on culturally specific conditions. Her 2015 procurement study demonstrates the identical institutional dynamics in technologically sophisticated automotive supply chains, where the monitoring is algorithmic and the membership is corporate rather than individual. The pattern is not industry-specific; it arises wherever the institutional conditions obtain, across four independent empirical domains spanning traditional commodity markets and modern manufacturing networks.

Contemporary digital platforms provide further evidence against the staleness objection. Amazon’s seller network, documented in Part III.A, exhibits the compliance dynamic the covenantal thesis predicts: algorithmic monitoring, permanent suspension, and systematic compliance with quality standards that bilateral buyer-seller damages cannot explain. The FTC’s 2023 complaint against Amazon details the seller governance mechanisms—performance metrics, account suspension, and related-account deactivation—that constitute the functional equivalent of DDC expulsion in a digital marketplace.⁷¹ The institutional form has changed; the structural dynamics have not.

Richman’s observation that some DDC functions declined as parallel institutions developed is itself evidence for the covenantal thesis rather than against it: when the club good erodes—whether through institutional competition, regulatory displacement, or internal decay—compliance erodes with it, precisely as the framework predicts. The covenantal account does not require any particular network to persist indefinitely; it requires only that wherever the three features obtain, the club-good externalities arise and bilateral internalization fails. The falsifiable predictions in Part VII are designed to resolve the staleness question empirically. H1 through H4 specify testable hypotheses that distinguish the covenantal account from bilateral alternatives using contemporary data—arbitration records, trader surveys, and natural experiments created by regulatory interventions like AB5 and state franchise good-cause statutes. The framework is not a post-hoc rationalization of classic studies; it is a lens for new empirical inquiry whose predictions are testable with current institutional data.

C. “The Diagnostic Is Circular”

A third objection holds that the diagnostic criterion—a recognized arbitral body with expulsion authority—is too closely related to the theoretical conclusion—that obligations run to the collective and expulsion is non-compensable—to function as an independent test. If the criterion just picks out networks with the right properties by definition, the diagnostic adds nothing.

The objection confuses correlation with identity. The pleading-stage criterion is an observable institutional fact: the existence of a recognized arbitral body with expulsion authority is binary and documentable—the NGFA Trade Rules, DDC bylaws, or equivalent constitute recognized arbitral bodies if they are established, memorialized in the parties’ agreement, and provide for expulsion as a sanction. This fact correlates with the three features without presupposing them and is ascertainable at the pleading stage from contract documents and course of dealings. If evidence later shows that one or more features are absent notwithstanding the institutional arrangement—if expulsion is in practice compensable, or if duties run only bilaterally—the presumption is rebutted and the standard bilateral toolkit applies.⁷² The diagnostic functions as a verifiable threshold rather than a tautology.

D. “Expulsion Destroys Livelihoods and Therefore Demands Judicial Intervention”

The most powerful equity objection is that expulsion from a trade association or platform network can destroy a merchant’s professional livelihood, and that judicial restraint in the face of such a severe sanction appears callous. A diamond dealer expelled from the global exchange network or an automotive supplier progressively excluded from OEM procurement faces not a financial setback but professional extinction. Courts asked to stand aside while this occurs will naturally resist.

The covenantal account does not deny the severity of the sanction or assert that expulsion is never unjust or that expelled members never suffer disproportionately. It argues something more specific and structural: judicial restoration of membership—beyond procedural review—erodes the third-party witnessing function whose integrity is the precondition for every other member’s low-cost exchange. A court that reinstates an expelled member over the network’s objection provides real equity to the individual plaintiff but simultaneously degrades the club good of enforcement capital for the entire membership: informational signals weaken, monitoring costs rise across the market, and the enforcement-capacity externality is exported to courts. The procedural-regularity standard protects against arbitrary expulsion process without claiming the witnessing role. It provides the expelled member with review of notice, opportunity to respond, and rational connection to a legitimate governance interest—all the process the covenantal structure can absorb without disabling the mechanism that makes the network valuable to the compliant majority.⁷³

E. “Covenantal Networks Can Discriminate Against Protected Classes”

A fifth objection is the most pointed from a civil rights perspective: covenantal deference, by insulating expulsion from judicial review of substantive merits, may also insulate discriminatory exclusions. A network that expels a member on pretextual governance grounds while the actual motivation is race, religion, or national origin would receive procedural deference that a purely bilateral employer would not receive under Title VII or section 1981.

The covenantal account addresses this objection in three structural steps. First, civil rights statutes apply independently of and in parallel with covenantal deference. Title VII, section 1981, and their state equivalents impose obligations that are not displaced by the covenantal framework any more than they are displaced by FAA arbitration deference. The covenantal analysis does not create a civil-rights-free zone; it identifies the appropriate standard for judicial review of governance decisions within that zone.⁷⁴

Second, the procedural-regularity standard itself screens for a category of discriminatory expulsion. The third prong of procedural regularity—that the stated basis for expulsion must be rationally connected to a legitimate network-governance interest—excludes expulsions whose stated rationale has no plausible connection to preserving the club good. A pretextual governance rationale offered to conceal discriminatory motivation fails this prong, because discriminatory animus is not a legitimate network-governance interest and the rational-connection requirement can expose gaps between stated reason and actual conduct.

Third, the unconditionality feature itself provides structural evidence against systematic discrimination. A network that enforces membership standards unconditionally—without inquiring into bilateral excuses, bilateral provocation, or bilateral equities—is applying a norm whose authority derives from universal application within the club boundary. Systematic discrimination based on protected characteristics is structurally inconsistent with unconditionality: it introduces precisely the bilateral, identity-based inquiries that the unconditionality feature is designed to exclude. A network that departs from unconditional enforcement by applying membership standards selectively to members of certain backgrounds is exhibiting conditional rather than covenantal obligation—which means the three-feature diagnostic may not be satisfied, and the covenantal framework would not apply in the first instance.

Conclusion

The diamond dealer on West 47th Street performs not because he fears his counterparty, but because the network is watching. He performs in a bilateral-pair terminal transaction—where the two parties will never deal with each other again, where no iterated-game mechanism sustains compliance, and where expectation damages payable to the bilateral counterparty would leave his net gain positive. He performs because breach would degrade the club good of collective enforcement capital shared by every member of the exchange. The network’s third-party witnessing apparatus observes the transaction; expulsion is non-compensable and global. The obligation runs to the trading community, not to the individual counterparty. Efficient breach doctrine cannot explain this pattern because it cannot see the obligee.

This Article has identified the point at which efficient breach’s descriptive universality claim reaches a boundary. The doctrine’s indifference premise rests on four structural commitments: the relevant parties are bilateral; the harm is a recoverable bilateral economic loss; the promisee is the relevant injured party; and performance and breach-plus-damages are remedially equivalent. Where all four hold, the bilateral toolkit is accurate and sufficient. Where the relevant injured party is the trading community and the relevant harm is degradation of a club good through diffuse informational, verification/monitoring, and enforcement-capacity externalities, none of the four holds and the calculus cannot establish indifference for the actual obligee.

The covenantal framework organizes the divergence. It is descriptive and institutional, not normative or moral. It requires no judgment about whether breach is wrongful, no endorsement of status-based obligation as a normative ideal, and no displacement of bilateral contract doctrine in the domains where bilateral assumptions hold. It identifies, through three observable structural features—permanence, unconditionality, and third-party witnessing backed by expulsion—the subset of commercial obligations in which the obligee is collective, the harm is a club-good externality, and bilateral remedies predictably destroy the institutional infrastructure they purport to protect.

The non-universalist scope of the framework is itself a substantive claim about efficient breach doctrine: the doctrine is not wrong in the domains where the four structural commitments hold. It is descriptively incomplete, in the sense that there exists a commercially significant category of obligation the doctrine cannot perceive. Courts and commentators who have defended efficient breach—Holmes, Birmingham, Goetz and Scott, Posner, Markovits and Schwartz—are correct in their own domain. The covenantal account does not contest their analysis; it identifies the boundary at which that analysis becomes inapplicable. The claim is the legal analogue of a domain restriction in economic theory: the indifference claim holds for bilateral promisor-promisee pairs whose harm is a measurable bilateral economic loss; it fails for collective obligees whose harm is a club-good externality. Identifying the boundary is the contribution; contesting the doctrine within its own domain is not.

The doctrinal and regulatory implications are conditional and narrow. In markets satisfying the diagnostic criterion, courts should apply a rebuttable presumption against incorporating relationship-preserving norms under UCC sections 1-303 and 2-202, review expulsion decisions under the procedural-regularity standard anchored in Silver v. New York Stock Exchange and the Stolt-Nielsen/Oxford Health line, and treat network-internal expulsion under antitrust rule-of-reason analysis with a presumption of legitimacy when it enforces membership norms. Where the expelling network also possesses monopoly or monopsony power, covenantal deference protects the governance mechanism while antitrust scrutiny addresses the market structure; the two analyses run in parallel. Where regulatory intervention applies bilateral good-cause or portability requirements to covenantal networks, the predicted consequence is degradation of the enforcement-capacity externality, substitution of costly ex ante screening for efficient ex post expulsion, and contraction of the club good’s value for all compliant members.

The framework generates four falsifiable predictions that distinguish it from bilateral reputational models, relational-contract theory, and transaction cost economics: terminal-transaction compliance rates should be statistically indistinguishable from non-terminal rates within covenantal networks; judicial incorporation of covenantal norms should reduce voluntary accommodation supply; regulatory interventions weakening expulsion finality should correlate with increased formal litigation; and network density should predict lower efficient-breach rates. These are design specifications for a research agenda, not post-hoc rationalizations of the classic studies on which the framework initially relies.

Efficient breach doctrine is not wrong. It is right about most commercial obligation. What it cannot perceive—because its bilateral lens excludes the obligee who is actually watching—is the category of obligations where breach is not a bilateral resource reallocation but a status event that degrades a collective good. When courts and regulators apply bilateral tools to those obligations, they do not merely get the remedy wrong. They destroy the governance infrastructure that made the network valuable in the first place. The formal model in the Technical Appendix proves, under a range of specifications, that each increment of judicialization produces this cascading institutional failure. The math is not merely consistent with the qualitative argument; it provides the mechanism.

The diamond dealer on West 47th Street already knows this. He performs because the network is watching and because the network’s watching is what makes the exchange possible at all. The task for doctrine and regulation is to recognize what he already understands: that the relevant obligee is not standing across the table from him, but is constituted by every member of the exchange, present and future, whose transaction costs depend on the integrity of the signal that his performance preserves.