“Private” Governance Is Actually a Club Good
                                     Seth C. Oranburg*

Abstract Network governance is a club good. Courts that displace a network’s authority to ostra- cize rule-breakers weaken the excludability that makes governance valuable to members and non-members alike. When the ostracism mechanism loses credibility, members de — ect — rom governance obligations, governance quality declines, and the positive externalities govern- ance creates — or non-members disappear. Courts should there — ore recognize doctrines that de — er to network decisions as implicit Pigouvian subsidies that reduce governance costs and help prevent the undersupply o — socially valuable governance.

* Pro — essor o — Law, University o — New Hampshire Franklin Pierce School o — Law; Director, Program on Organizations, Business and Markets at NYU Law’s Classical Liberal Institute. Introduction Debates about “governance”1 usually treat “public” governance as what sovereign states do through coercion,2 while treating “private” governance as what networks and organiza- tions do through voluntary a —

iliation.3 But public versus private is a — alse dichotomy. Eco- nomic theory recognizes not two types o — goods but — our types: private goods, public goods, common-pool resources, and club goods.4 This Article shows that so-called private govern- ance has — eatures o — a club good: it is excludable through ostracism, nonrivalrous up to con- gestion, voluntarily joined, and — inanced by its members. Club goods can generate positive externalities in the — orm o — bene — its to non-members.5 But, perhaps ironically, public bene — its o — private clubs depend on that club’s ability to ex- clude the public. When a club’s ostracism mechanism is weakened, members de — ect, govern- ance quality declines, the club good decays into a public good, and the positive externalities governance creates — or non-members disappear.6 Re — raming private governance as a club good reveals why its network governance may be undersupplied. Litigation regarding ostracism — rom clubs is not a mere disputes over private ordering. Awarding damages to excluded members might produce public costs that outweigh private bene — its. When court displace a club’s ostracism mechanism, this triggers a com- pound wel — are loss in which both the internal good and its external spillovers are destroyed. Non-members can bear the heaviest costs.

1 Governance generally re — ers to any system o — re — ereeing that allows a group o — people with disparate in- terests to collaborate. 2 See Thomas Hobbes, LEVIATHAN 117–22 (Richard Tuck ed., Cambridge Univerity Press 1996) (1651) (con- tending that stable social order requires individuals in the state o — nature to con — er authority on a sovereign whose coercive power makes compliance with civil laws and covenants rational, since “covenants, without the sword, are but words”). Hobbes’s sovereign is the canonical model o — what I mean by “public” govern- ance through coercion. 3 See F.A. Hayek, Law, LEGISLATION AND LIBERTY: A NEW STATEMENT OF THE LIBERAL PRINCIPLES OF JUSTICE AND POLITICAL ECONOMY 35–54 (University o — Chicago Press 1973) (developing the idea o — “spontaneous order,” in which social rules and institutions emerge — rom decentralized interactions rather than — rom centralized coercive design). Hayek’s account o — norms and institutions arising — rom voluntary arrangements provides the canonical contrast to Hobbesian governance by a coercive sovereign. 4 See James M. Buchanan, An Economic Theory o — Clubs, 32 ECONOMICA 1, 1–14 (1965) (introducing “club goods” as excludable but nonrivalrous up to congestion, alongside private and public goods); Elinor Ostrom, GOVERNING THE COMMONS: THE EVOLUTION OF INSTITUTIONS FOR COLLECTIVE ACTION 30–33 (Cambridge University Press 1990) (distinguishing private goods, public goods, common‑pool resources, and toll/club goods using excludability and rivalry as the two dimensions). 5 E.g., Paul G. Mahoney, The Exchange as Regulator, 83 VA. L. REV. 1453, 1457–58, 1466–67 (1997) (arguing that exchange rules can prevent — ree riding by nonmembers on the exchange’s prices and other assets, thereby linking member governance to broader market bene — its); see also Mark Koyama, Prosecution As- sociations in Industrial Revolution England, 28 J. LEGAL STUD. 95, 95–96 (2012) (describing associations that bundled a private good — or members with the public good o — crime deterrence) 6 See Mancur Olson, THE LOGIC OF COLLECTIVE ACTION: PUBLIC GOODS AND THE THEORY OF GROUPS 2, 44–52 (Harvard Univ. Press 1965) (explaining that collective goods deteriorate when members can — ree ride and no selective incentive or coercive mechanism makes contribution individually rational); see also Barak D. Richman, Firms, Courts, and Reputation Mechanisms: Towards a Positive Theory o — Private Ordering, 104 COLUM. L. REV. 2328, 2332–39 (2004) (explaining that private-ordering systems depend on credible reputa- tional sanctions and exclusion to sustain cooperation) 2026-03-18] 3

 Part I establishes that governance satis --- ies Buchanan’s criteria  --- or club goods and docu- ments positive externalities to non-members across three domains.7 Part II demonstrates the compound wel --- are loss and shows when external losses dominate internal ones. Part III re- characterizes  --- our doctrines that de --- er to network decisions to exclude non-members or os- tracize members as Pigouvian subsidies and stress-tests the  --- ramework against Loper Bright’s overruling o ---  Chevron de --- erence. Part IV con --- ronts the cartel, discrimination, and judicial- externality objections and proposes a standard o ---  calibrated de --- erence.
 Consider the New York Diamond Dealers Club:  --- or over a century, a complete private legal system governing billions o ---  dollars in annual unsecured credit, en --- orced through a sin- gle mechanism — the credible threat o ---  ostracism propagated to every a ---

iliated bourse worldwide.8 For members, this governance reduced transaction costs. For downstream buy- ers, retailers, and consumers who never joined the DDC, it reduced — raud. When the system eroded, those — raud costs spread to banks, retailers, and the consuming public.9 The club good was destroyed, and non-members bore the heaviest costs. Part I develops this para- digm case and two others in detail. Consider next Maine’s lobster gangs. James Acheson documented governance operating at the opposite end o — institutional — ormality — rom the DDC.10 Territory en — orcement emerg- es through graduated social sanctions: verbal warnings escalate to gear inter — erence, cutting o — trap lines, and ultimately exclusion — rom — ishing territories. The scale is substantial — Maine lobster landings have exceeded 130 million pounds annually, worth over $500 million. Yet this system operates with no written bylaws, no — ormal tribunal, no headquarters — governance sustained entirely through the credible threat o — social ostracism and economic exclusion. The governing body can expel a — isher — rom the community, and that expulsion is communicated through networks spanning the entire coast, e —

ectively ending the — isher’s access to productive territory. The mechanism operates identically to the DDC’s — ormal ap- paratus: gradualism, community awareness, and irreversibility o — exclusion render the ostra- cism threat credible and member cooperation rational. The ostracism mechanism maintains governance at every level o — institutional — ormality, — rom Manhattan diamond exchanges to Maine — ishing communities. Mapping this mechanism’s operation across contexts — — ormal and in — ormal, urban and rural, global and local — shows that the analytical move is not do- main-speci — ic but general. Part IV’s sliding-scale procedural standard re — lects this generality,

7 Prior work has recognized governance institutions as clubs without modeling the wel — are consequences o — governance degradation or drawing implications — or judicial de — erence. See Edward Peter Stringham, Private Governance: Creating Order in Economic and Social Li — e 21–36 (2015); Mark Koyama, Prosecution Associations in Industrial Revolution England: Private Providers o — Public Goods?, 41 J. Legal Stud. 95 (2012); L. Lynne Kiesling, The Promise and Perils o — Exclusion, J. Institutional Econ. (2026). Part I’s “Novel Classi — ication” subsection engages these predecessors in detail. 8 Lisa Bernstein, Opting Out o — the Legal System: Extralegal Contractual Relations in the Diamond Industry, 21 J. LEGAL STUD. 115, 119–43 (1992). 9 Barak D. Richman, An Autopsy o — Cooperation: Diamond Dealers and the Limits o — Trust-Based Exchange, 9 J. Legal Analysis 247, 257–76 (2017) (documenting the DDC’s decline and the spread o — costs to banks, retailers, and consumers). 10 James M. Acheson, THE LOBSTER GANGS OF MAINE: TERRITORIAL ENFORCEMENT IN A NEAR-OPEN ACCESS RESOURCE (University Press o — New England, 1988). 4 “Private” Governance Is Actually a Club Good

calibrating the governance presumption to the institutional context in which ostracism oper- ates.

Governance as a Club Good Mapping governance onto Buchanan’s — our criteria reveals not only that the classi — ication holds but that it holds across institutional domains that share no common — eatures except the ostracism mechanism. Diamond bourses, stock exchanges, and commons communities di —

er in size, — ormality, subject matter, geographic scope, and cultural context. What they share is a governance apparatus that satis — ies each o — Buchanan’s de — ining conditions — ex- cludable through the authority to ostracize rule-breakers, partially rivalrous through the con- gestion that determines optimal group size, voluntarily joined, and — inanced by members who bear its costs11 — and that generates positive externalities — or non-members who never participate in the governing body. Prior scholars have recognized that governance institu- tions can — unction as clubs, and that some club arrangements generate spillover bene — its.12 What no prior work has done is model the wel — are consequences o — degrading a governance club good that simultaneously generates positive externalities to non-members, or draw the implication — or how courts should treat governance decisions. This Part establishes the clas- si — ication, documents the externalities, and identi — ies the analytical move that prior scholar- ship has not made.

Buchanan’s Framework In 1965, James Buchanan identi — ied a category o — goods that — its neither the private nor the public ideal type.13 A pure private good is both excludable (nonpayers can be denied ac- cess) and rivalrous (one person’s consumption diminishes another’s). A pure public good is neither excludable nor rivalrous — national de — ense and clean air are consumed by everyone regardless o — contribution. Buchanan observed that most goods — all between these poles. His canonical example was a swimming pool: members can be excluded through the gate, but the pool is nonrivalrous up to a point — adding swimmers does not diminish anyone’s en- joyment until crowding sets in. The optimal club size balances the cost savings — rom sharing against the wel — are losses — rom congestion.14

11 James M. Buchanan, An Economic Theory o — Clubs, 32 ECONOMICA 1, 1–14 (1965). 12 See Edward Peter Stringham, PRIVATE GOVERNANCE: CREATING ORDER IN ECONOMIC AND SOCIAL LIFE 21–36 (2015) (analyzing governance through a club — ramework but without modeling externalities to non-members or wel — are consequences o — degradation); L. Lynne Kiesling, The Promise and Perils o — Exclusion: Using Institutional Design Principles and the Theory o — Clubs to Analyse Regional Transmission Organization Governance, J. INSTITUTIONAL ECON. (2026) (integrating Buchanan’s club theory with Ostrom’s CPR

ramework — or electricity grid governance); Mark Koyama, Prosecution Associations in Industrial Revolution England: Private Providers o — Public Goods?, 41 J. LEGAL STUD. 95 (2012) (documenting prosecution associations that bundled private insurance with the public good o — deterrence). 13 Id. at 1–2. 14 Id. at 6–10. Buchanan’s — ormal model optimizes over two variables simultaneously: the quantity o — the shared good and the number o — members. The optimal membership is the point at which the marginal cost savings — rom adding a member equal the marginal congestion costs that member imposes. 2026-03-18] 5

 Four criteria de --- ine a club good. First, excludability: the club can deny access to the good by denying membership or expelling violators. Second, partial rivalry through congestion: the good is nonrivalrous within capacity but degrades as use increases, creating an optimal size beyond which additional users reduce quality  --- or existing members. Third, voluntary member- ship: individuals choose to join and may exit. Fourth, sel --- - --- inancing: members bear the costs o ---

provision through dues,

ees, or contributions, rather than relying on taxation or external subsidy.15 Goods satis — ying these criteria — — rom gol — courses to satellite television to gated communities — occupy the analytical space between markets and government that Buchan- an’s — ramework was designed to illuminate.16 A potential objection warrants address at the threshold. I — governance consumes — inite en — orcement resources — every arbitration hearing occupies a tribunal, every monitoring action costs sta —

time — then governance may be rivalrous — rom the start, placing it in the common-pool resource quadrant o — Ostrom’s matrix rather than the club good quadrant. The distinction turns on the primary source o — quality degradation. For a common-pool re- source like a — ishery, each unit o — extraction directly subtracts — rom the stock available to others; rivalry is intrinsic to consumption. For governance, quality degrades not because each member’s use o — the governance mechanism subtracts — rom the mechanism itsel — , but be- cause the mechanism’s credibility depends on the — raction o — members who cooperate. A DDC member who submits a dispute to arbitration does not diminish the arbitration sys- tem’s capacity in the way that a — isher’s catch diminishes the — ish stock. The system can hear the dispute without reducing its ability to hear the next one — up to the point o — congestion. What degrades governance is not overuse but de — ection: members who cheat, — ree-ride on others’ compliance, or re — use to accept arbitral outcomes. This is Buchanan’s congestion pat- tern (quality declines with overuse beyond a threshold), not Ostrom’s subtractability pattern (each unit o — consumption subtracts — rom the total).17

Three Domains o

Governance Externalities Governance mechanisms satis — y Buchanan’s criteria across a range o — institutional set- tings. More importantly, they generate positive externalities — bene — its that — low to non- members who never participate in the governing body and cannot be charged — or the bene-

its they receive. Three domains illustrate both the classi — ication and the externality.

15 Richard Cornes & Todd Sandler, THE THEORY OF EXTERNALITIES, PUBLIC GOODS, AND CLUB GOODS 347–97 (2d ed. 1996) (extending Buchanan’s model to incorporate heterogeneous members, multiple goods, and institutional variation); Todd Sandler & John T. Tschirhart, The Economic Theory o — Clubs: An Evaluative Survey, 18 J. ECON. LITERATURE 1481, 1481–1521 (1980) (surveying the — irst — i — teen years o — club theory and identi — ying extensions to local public goods, alliances, and — acility-sharing). 16 Buchanan’s insight has been applied to NATO burden-sharing, Tiebout jurisdictional competition, telecommunications networks, and environmental programs. See Sandler & Tschirhart, supra note 9, at 1504–14. What it has not been applied to — until now — is the governance mechanism itsel — . 17 The distinction matters — or the wel — are analysis in Part II. I — governance were a common-pool resource, degradation would — ollow — rom overuse — each member consuming more governance than the system can sustain. Because governance is a club good, degradation — ollows — rom undermined excludability — members de — ecting because the ostracism threat is no longer credible. The policy implication di —

ers: CPR degradation calls — or use restrictions; club good degradation calls — or restoring excludability. Judicial de — erence restores excludability. It does not restrict use. 6 “Private” Governance Is Actually a Club Good

Diamond Bourses The New York Diamond Dealers Club operates a complete private legal system. Mem- bers transact on the basis o — oral agreements sealed with the phrase mazal u’bracha, with dis- putes resolved through internal arbitration rather than litigation.18 Lisa Bernstein’s — ounda- tional study documented a governance mechanism encompassing rules o — conduct, binding arbitration, graduated sanctions, and in — ormation sharing propagated worldwide through the World Federation o — Diamond Bourses.19 The mechanism satis — ies each o — Buchanan’s crite- ria. Excludability operates through ostracism: a member — ound to have cheated — aces expul- sion — rom the DDC, and that expulsion is communicated to every a —

iliated bourse world- wide, e —

ectively ending the trader’s career.20 Congestion constrains optimal size: as Richard Cooter and Janet Landa demonstrated, en — orcement costs rise as trading networks expand beyond the group’s capacity to monitor behavior through personal knowledge and reputa- tion, degrading governance quality in a pattern structurally identical to Buchanan’s conges- tion — unction.21 Membership is voluntary — traders choose to join the DDC and accept its rules — and sel — - — inancing through dues and arbitration — ees. What sustains this system is not the act o — expulsion itsel — — actual expulsions are rare — but the credible threat o — expulsion and the cost it imposes. A trader expelled — rom the DDC loses access to the only marketplace where billions o — dollars in diamonds change hands on oral agreements. No other institution o —

ers equivalent access at equivalent cost. Because every member knows this, every member has reason to comply with DDC rules, accept arbitration outcomes, and deal honestly — not out o — altruism but because the alter- native is economic exile. The threat o — ostracism converts what would otherwise be a collec- tive action problem (each trader has an individual incentive to cheat) into a cooperative equi- librium (no trader cheats because the cost o — getting caught exceeds the gain — rom cheating). It is this equilibrium — sustained by the credible threat o — exclusion, not by any external en-

orcement authority — that produces the governance quality — rom which non-members bene — it. The governance mechanism generates positive externalities — or non-members. DDC ar- bitration and en — orcement reduce — raud rates throughout the diamond supply chain, bene — it- ing downstream buyers — retailers, jewelers, insurers, consumers — who never join the

18 Lisa Bernstein, Opting Out o — the Legal System: Extralegal Contractual Relations in the Diamond Industry, 21 J. LEGAL STUD. 115, 119–25 (1992). 19 Id. at 130–38 (describing the DDC’s arbitration system, its relationship to the WFDB, and the propagation o — sanctions across a —

iliated bourses). 20 Id. at 138–43 (documenting how expulsion — rom one bourse triggers worldwide exclusion through the WFDB network). Janet Landa described these sanctions as the — unctional equivalent o — contract law: “(a) withdrawal o — credit so that the trader has to deal on a cash basis; (b) exclusion — rom — uture dealings; and (c) ‘expulsion’ — rom the group via bankruptcy proceedings.” Janet T. Landa, A Theory o — the Ethnically Homogeneous Middleman Group: An Institutional Alternative to Contract Law, 10 J. LEGAL STUD. 349, 356 (1981). 21 Richard Cooter & Janet Landa, Personal Versus Impersonal Trade: The Size o — Trading Groups and Contract Law, 4 INT’L REV. L. & ECON. 15, 19–24 (1984) (extending Buchanan’s optimal-size analysis to trading networks and showing that en — orcement costs rise sharply at ethnic boundaries where monitoring capacity declines). 2026-03-18] 7

bourse and pay nothing

or its governance.22 Barak Richman’s study o — the DDC’s decline con — irmed the externality by documenting the counter — actual. When the trust-based trading system eroded under pressure — rom globalization, new market entrants, and technological change, — raud costs spread to banks, retailers, and consumers who had previously bene — ited

rom the DDC’s governance without bearing its costs.23 Vertical integration replaced trust- based exchange as the dominant transactional — orm — a costlier institutional arrangement that eliminated the externality-generating mechanism entirely.24

Stock Exchange Sel

-Regulation Stock exchanges have regulated their members since long be — ore the Securities and Ex- change Commission existed.25 The governance mechanism includes listing standards, trading surveillance, member discipline, and en — orcement actions — a regulatory apparatus operated by the exchange — or its members and en — orced through the threat o — expulsion — rom trading privileges.26 Excludability operates through membership revocation and the denial o — trading access. Congestion is inherent: monitoring quality degrades as market complexity increases and the number o — regulated entities grows relative to en — orcement capacity. Membership is voluntary in the relevant sense — FINRA membership is mandatory — or registered broker- dealers, but the decision to become a broker-dealer is itsel — voluntary, and the mandatory character re — lects Congress’s judgment that sel — -regulatory governance produces bene — its worth preserving.27 The externality to non-members is market integrity. Exchange sel — -regulation reduces in- sider trading, deters — raud, and improves price discovery — bene — its that accrue to every in-

22 The externality operates through the — raud-reduction channel. When the DDC credibly sanctions cheaters, the expected cost o —

raud in the diamond supply chain — alls — or all participants, not just DDC members. A retailer purchasing diamonds — rom a DDC-a —

iliated dealer bene — its — rom the lower — raud probability even though the retailer is not a DDC member and pays no DDC dues. 23 Barak D. Richman, An Autopsy o — Cooperation: Diamond Dealers and the Limits o — Trust-Based Exchange, 9 J. Legal Analysis 247, 269 (2017) (“[T]he rash o — bankruptcies among diamond processing companies today isn’t simply bad news — or these companies and their creditors. It’s bad news — or all o — us. The banks don’t trust the industry any more.”). 24 Id. at 275 (documenting how vertical integration strategies arose to mitigate transactional hazards when the trust-based system no longer provided reliable governance). 25 Paul G. Mahoney, The Exchange as Regulator, 83 VA. L. REV. 1453, 1457–62 (1997) (documenting NYSE sel — -regulatory rules dating to the nineteenth century and arguing that Congress built the Securities Exchange Act o — 1934 on existing sel — -regulatory in — rastructure rather than replacing it). 26 FINRA reported approximately 700 disciplinary actions and roughly 350–400 individuals barred — rom the securities industry in 2023. See FINRA, 2023 Annual Report; c — . SEC Division o — En — orcement, Annual Report FY 2023 (reporting 784 total en — orcement actions). The volume o — SRO en — orcement is comparable to or exceeds the SEC’s own output, supporting the claim that sel — -regulatory governance produces substantial en — orcement at scale. 27 The Securities Exchange Act o — 1934 § 15A, 15 U.S.C. § 78o-3, requires broker-dealers to register with a national securities association (e —

ectively FINRA). Congress imposed this requirement precisely because it recognized that sel — -regulatory governance generates bene — its — market integrity, investor con — idence, e —

icient price discovery — that the SEC alone could not produce at equivalent scale or cost. See Mahoney, supra note 19, at 1462–70. 8 “Private” Governance Is Actually a Club Good

vestor in the market, not just the member

irms subject to SRO rules.28 Mahoney argued that exchange rules against — raud and manipulation — unction to “prevent — ree riding on the ex- change’s prices and other assets by nonmembers,” directly connecting exchange governance to the club goods — ramework.29 When the Supreme Court imposed procedural requirements on NYSE governance decisions in Silver v. New York Stock Exchange, it acknowledged both the value o — sel — -regulation and the risk o — its abuse — holding that “the entire public policy o — sel — -regulation, beginning with the idea that the Exchange may set up barriers to member- ship, contemplates that the Exchange will engage in restraints o — trade which might well be unreasonable absent sanction by the Securities Exchange Act.”30

Commons Governance Community-managed — orests, irrigation systems, and — isheries provide a third domain where governance satis — ies Buchanan’s criteria and generates externalities beyond the gov- erning community.31 Elinor Ostrom’s research documented how local communities develop monitoring, sanctioning, and con — lict-resolution mechanisms to manage common-pool re- sources — governance apparatus that is excludable (non-contributors — ace graduated sanc- tions up to exclusion), congestible (governance quality degrades as group size exceeds moni- toring capacity), voluntary within the community context, and sel — - — inancing through com- munity labor and contributions.32 The externality to non-members is environmental. Ashwini Chhatre and Arun Agrawal studied eighty — orest commons across ten countries and — ound that — orests managed by larg- er groups o — local users stored signi — icantly more carbon than those under government man- agement.33 Carbon sequestration is a textbook global positive externality — the governance o — local — orest commons produces climate bene — its — or the entire planet, well beyond the governing community. Nepal’s community — orestry program provides a second data point at national scale: approximately 22,000 Community Forest User Groups managing roughly 2.2 million hectares o —

orest have contributed to an increase in Nepal’s — orest cover — rom ap- proximately 29% in the 1990s to roughly 44% by 2016, with documented downstream bene-

28 See Mahoney, supra note 19, at 1468–73 (arguing that exchange rules against — raud and manipulation produce market-wide bene — its). The CFA Institute has documented how SRO rules generate “heightened public trust” bene — iting all market participants, not only members. 29 Id. at 1472. 30 Silver v. New York Stock Exch., 373 U.S. 341, 349 (1963). Silver is analyzed in detail in Part IV as the

oundational case — or calibrated de — erence to private governance. 31 Elinor Ostrom, Governing the Commons: The Evolution o — Institutions — or Collective Action (1990) (establishing the empirical and theoretical — oundations — or commons governance); Elinor Ostrom, Beyond Markets and States: Polycentric Governance o — Complex Economic Systems, 100 Am. Econ. Rev. 641, 641–72 (2010) (Nobel Prize lecture synthesizing the case — or polycentric governance). 32 Ostrom, Beyond Markets and States, supra note 25, at 653–55 (describing Design Principles — or success — ul commons governance, including clear boundaries, graduated sanctions, monitoring, and con — lict-resolution mechanisms). Ostrom’s Design Principle #8 explicitly addresses the relationship between local governance and larger-scale institutions, acknowledging that commons governance generates e —

ects requiring coordination beyond the community — an implicit recognition o — the externality this Article makes explicit. 33 Ashwini Chhatre & Arun Agrawal, Trade-o —

s and Synergies Between Carbon Storage and Livelihood Bene — its — rom Forest Commons, 106 PNAS 17667, 17667–70 (2009). 2026-03-18] 9

its including reduced soil erosion and improved water quality — or non-member communi- ties.34 Community irrigation governance in systems like Valencia’s Tribunal de las Aguas pre- serves downstream water quality — or non-members who take no part in the upstream gov- ernance decisions.35 These externalities are not incidental byproducts. Their magnitude tracks governance quality: better-monitored — orests store more carbon and deliver cleaner water downstream, because the governance mechanism (graduated sanctions, monitoring, con — lict resolution) determines how e —

ectively the resource is managed.

Macroeconomic Evidence The three domains examined above demonstrate the externality mechanism at institu- tional levels where governance mechanisms can be observed and documented. Broader em- pirical support emerges — rom macroeconomic evidence. La Porta, Lopez-de-Silanes, Shlei — er, and Vishny’s Law and Finance (106 J. Pol. Econ. 1113 (1998)) documented across ninety-nine countries that the strength o — investor protection law and the quality o — law en — orcement ex- plain substantial variation in stock market development, debt markets, and economic growth — a cross-country — inding con — irming what the case studies show at institutional scale. Op- tionally, Acemoglu, Johnson, and Robinson’s Colonial Origins o — Comparative Development (91 Am. Econ. Rev. 1369 (2001)) demonstrates that institutions determining the quality o — gov- ernance (extractive vs. inclusive institutions) predict divergent economic outcomes across regions with identical geographies, natural resources, and disease environments. What the three case studies demonstrate at the level o — speci — ic governance mechanisms — that insti- tutional quality tracks externality magnitude — cross-country evidence con — irms at the mac- roeconomic level: societies that maintain higher-quality governance institutions experience higher growth, broader — inancial development, and greater stability. The governance-as-club- good — ramework explains this pattern: governance generates positive externalities, and insti- tutional arrangements that protect governance quality (through property-rule protection, en-

orceable rules, and credible ostracism) there — ore sustain the economic growth that those externalities enable.

The Novel Classi

ication Prior scholars have applied club theory to governance settings, but each stopped at the same analytical boundary. Janet Landa characterized the ethnically homogeneous middleman group as “a club-like structural arrangement, an alternative to contract law and to the verti- cally integrated — irm” — with contract en — orcement as the output the club-like group pro- vides, not governance itsel — as the good within Buchanan’s typology.36 Landa’s — ramework

34 Nepal Department o — Forests, Community Forestry Statistics (2016). The causal attribution requires hedging: — orest cover increases re — lect multiple — actors including government policy and demographic change, not community governance alone. The point is that community governance is a signi — icant contributing — actor and that the environmental bene — its — carbon sequestration, watershed protection, biodiversity — — low to non-members o — the Community Forest User Groups. 35 Ostrom, Governing the Commons, supra note 25, at 69–82 (documenting the Valencian irrigation system’s 500-year governance record). 36 Landa, supra note 14, at 361. 10 “Private” Governance Is Actually a Club Good

internalizes externalities within the group: “once the code o

ethics emerges in a Chinese middleman economy, all externalities are internalized.”37 Bene — its — low to insiders. Outsiders are simply excluded. Edward Stringham titled a chapter o — his 2015 book “Governance as a Club Good,” the closest prior use o — this Article’s central phrase.38 But Stringham’s project is libertarian politi- cal philosophy, not wel — are economics. He argues that “governance can be analyzed as a club good in which private provision o — rules is already prevalent” and that “private governance could be more prevalent and could substitute — or government.”39 His analysis treats govern- ance as a product that private parties already supply — a descriptive observation about insti- tutional prevalence — without placing governance in Buchanan’s — ormal — ramework, without modeling externalities to non-members, and without analyzing the wel — are consequences o —

governance degradation. Stringham does not discuss judicial de

erence. L. Lynne Kiesling’s recent work integrates Buchanan’s club theory with Ostrom’s com- mon-pool resource — ramework, arguing that Regional Transmission Organizations “trans-

orm a reliability commons into a rule-de — ined club.”40 Her analysis is domain-speci — ic (elec- tricity grid governance) and identi — ies a problem — the “non-replicable club” where physical network constraints prevent exit and thus disable the competitive discipline that normally constrains clubs — that is directly relevant to this Article’s analysis o — when de — erence should be reduced. 41 Kiesling models the governance institution as a club. She does not model governance as a club good that generates externalities to non-members. Asher Prakash and Matthew Potoski modeled voluntary environmental programs as clubs that provide reputational club goods — the credible signal o — environmental steward- ship that members share.42 The club good in their — ramework is the brand, not the govern- ance mechanism (monitoring, en — orcement, certi — ication) that maintains the brand’s credibil- ity. Mark Koyama documented prosecution associations in Industrial Revolution England that bundled a private good (mutual insurance — or members) with a public good (deterrence o — crime bene — iting the entire community), coming closest to this Article’s externality claim in a speci — ic historical domain.43 Peter Leeson argued that governance should be organized through clubs because clubs have residual claimants and — ace competitive pressure — a

37 Id. at 357. 38 Stringham, supra note 6, at 21–36. 39 Id. at 36. 40 Kiesling, supra note 6. 41 Id. (arguing that when clubs cannot be replicated, the exclusion rules that maintain governance quality can harden into barriers to innovation — a “pacing problem” that Ostrom’s adaptive governance principles can address). This Article takes up Kiesling’s non-replicability insight in Part IV, where it in — orms the calibration o — judicial de — erence. 42 Asher Prakash & Matthew Potoski, The Voluntary Environmentalists: Green Clubs, ISO 14001, and Voluntary Environmental Regulations (2006); Asher Prakash & Matthew Potoski, Collective Action Through Voluntary Environmental Programs: A Club Theory Perspective, 37 POL’Y STUD. J. 773 (2009). 43 Koyama, supra note 6, at 95–96 (“Consistent with the reasoning o — Demsetz (1970), I — ind that prosecution associations were economic clubs that bundled the private good o — insurance with the public good o — deterrence.”). Koyama’s associations illustrate this Article’s thesis in microcosm, but his contribution is historical and descriptive rather than theoretical — he does not generalize the observation into a — ramework with normative implications — or courts. 2026-03-18] 11

normative argument about institutional design, not a classi

icatory claim about the goods typology.44 The persistent pattern across six decades o — scholarship is this: every author treats gov- ernance as the mechanism that produces club goods, or treats governance institutions as clubs that provide goods to members. Nobody classi — ies the governance mechanism itsel — — the monitoring, arbitration, and en — orcement apparatus — within the goods typology. Col- lapsing this distinction is not merely a relabeling. Once governance is classi — ied as a club good with positive externalities, three analytical consequences — ollow that none o — the prede- cessor — rameworks produced. First, governance becomes subject to Buchanan’s optimization conditions, which predict that it will be provided at less than the socially optimal level when externalities are not internalized. Second, governance becomes subject to Olson’s — ree-rider prediction: i — excludability is lost, rational members will de — ect — rom governance obligations, and quality will degrade.45 Third, governance becomes subject to Pigouvian subsidy logic: because it generates positive externalities, the market will undersupply it, and legal doctrines that reduce its production costs — unction as corrective subsidies.46 These three consequences — undersupply, — ragility, and the subsidy rationale — are the subjects o — Parts II and III.

The Compound Wel

are Loss When courts displace the ostracism mechanism that makes governance excludable, two wel — are losses occur simultaneously. The internal governance good degrades as — ree-riding replaces cooperation — Mancur Olson’s standard prediction — or any collective good that loses excludability.47 At the same time, the positive externalities that governance generated

or non-members disappear, because those externalities were proportional to governance quality. This compound e —

ect — compound because two distinct losses — low — rom the same event — has no precedent in the — ormal literature. Scholars have modeled the undersupply o — public goods, the congestion o — club goods, and the positive externalities o — various insti- tutional arrangements. Nobody has modeled the wel — are consequences o — converting a club good with positive externalities into an undersupplied public good by stripping away its ex- cludability — the speci — ic mechanism through which judicial displacement o — governance destroys value. This Part develops the compound wel — are loss and shows that the external component dominates the internal one.

Internal Degradation The internal loss — ollows directly — rom Olson. Governance quality depends on member cooperation: compliance with rules, participation in monitoring, acceptance o — arbitration

44 Peter T. Leeson, Government, Clubs, and Constitutions, 80 J. ECON. BEHAV. & ORG. 301 (2011). 45 Mancur Olson, The Logic o — Collective Action: Public Goods and the Theory o — Groups 53–65 (1965). 46 A.C. Pigou, THE ECONOMICS OF WELFARE (1920); William J. Baumol & Wallace E. Oates, THE THEORY OF ENVIRONMENTAL POLICY 21–35 (1975). Part III develops this point at length, recharacterizing de — erence doctrines as implicit Pigouvian subsidies. 47 Olson, supra note 39, at 53–65. 12 “Private” Governance Is Actually a Club Good

outcomes, and willingness to bear the costs o

en — orcement.48 When members can de — ect

rom these obligations without — acing credible exclusion, rational sel — -interest predicts that they will. A diamond trader who knows that courts will reverse any expulsion decision has diminished incentive to comply with DDC arbitration awards. A broker-dealer that knows FINRA’s disciplinary sanctions will be overturned on judicial review has diminished incen- tive to maintain compliance systems beyond the minimum required to survive litigation. The cooperation rate — alls, governance quality declines, and the club good degrades toward the undersupplied equilibrium that characterizes public goods.49 This result is not controversial. It is Olson applied to a speci — ic institutional setting. The

ree-riding that concerns Olson here is not the bene — icial consumption o — governance exter- nalities by non-members — which is the positive spillover Part I documented — but the destructive de — ection o — members who stop cooperating because they can no longer be ex- cluded — or doing so. What has not been recognized is the second, simultaneous loss.

The Vanishing Spillovers The positive externalities that governance generates — or non-members are proportional to governance quality. When the DDC’s governance system — unctioned e —

ectively, down- stream buyers bene — ited — rom reduced — raud rates throughout the diamond supply chain. When exchange sel — -regulation maintains market integrity, all investors bene — it — rom im- proved price discovery and reduced manipulation. When commons governance sustains en- vironmental quality, non-members o — the governing community bene — it — rom carbon seques- tration, clean water, and ecosystem stability. In each case, the externality is a — unction o —

governance quality — not a

ixed endowment that persists regardless o — institutional per-

ormance. When governance quality degrades through the internal mechanism Olson predicts, the externalities degrade with it. Fraud rates rise in the diamond supply chain, harming down- stream buyers. Market integrity declines, harming all investors. Environmental quality deteri- orates, harming non-members o — the commons community. The external loss compounds the internal one: the club good’s members su —

er — rom degraded governance, and non- members who previously received spillover bene — its lose those bene — its as well. The external loss is larger than the internal loss when non-members outnumber mem- bers — which they typically do by orders o — magnitude. The DDC has approximately 1,800 members; the downstream diamond supply chain serves millions o — retailers and hundreds o —

48 See David Hirshlei — er & Eric Rasmusen, Cooperation in a Repeated Prisoners’ Dilemma with Ostracism, 12 J. ECON. BEHAV. & ORG. 87, 87–106 (1989) (modeling how the threat o — ostracism sustains cooperation in repeated games and showing that cooperation collapses when the credibility o — ostracism is undermined). 49 The degradation path is not necessarily — rom club good to pure public good. Because governance involves both diminished excludability and congestion ( — inite en — orcement resources shared among more

ree-riders), the degraded state may more precisely be a common-pool resource — a good that is rivalrous but nonexcludable. See Ostrom, Governing the Commons, supra note 25, at 30–33 (de — ining common-pool resources). The wel — are implications are similar: undersupply, overuse o — the en — orcement mechanism by those who do participate, and eventual collapse. The distinction between degradation to public good and degradation to common-pool resource is analytically precise but does not change the direction o — the wel — are prediction. 2026-03-18] 13

millions o

consumers. 50 FINRA oversees approximately 3,400 broker-dealer — irms and 620,000 registered representatives; the U.S. equity market serves tens o — millions o — individu- al investors and trillions o — dollars in institutional capital.51 Any given commons community numbers in the hundreds or thousands; the bene — iciaries o — carbon sequestration and water- shed protection encompass the global population. Because each unit o — governance quality generates spillover bene — its to every non-member, a marginal reduction in governance quali- ty produces a small per-member internal loss and a large aggregate external loss — the prod- uct o — a small per-non-member e —

ect multiplied by a much larger non-member population.52 The compound wel — are loss di —

ers — rom the standard Pigouvian undersupply problem in a critical respect. The standard problem is that a positive-externality good is undersupplied at market equilibrium — producers cannot capture the — ull social bene — it, so they produce less than the socially optimal quantity. The compound wel — are loss is worse. The good is not merely undersupplied but destroyed, because the en — orcement mechanism that maintained it is disabled. Undersupply leaves a diminished version o — the good in place; destruction elimi- nates it. A swimming pool with too-high membership — ees is undersupplied — some people who would bene — it — rom access are excluded. A swimming pool whose gate has been re- moved is destroyed as a club good — nonpayers — lood in, the pool becomes congested, maintenance — unding disappears, and the asset degrades to the point o — unusability. Govern- ance de — erence is the gate.

The Counter

actual The compound wel — are loss is a theoretical prediction. Three episodes provide empirical traction. Richman’s study o — the DDC’s decline is the most detailed account o — governance deg- radation producing spillover harm.53 The trust-based exchange system that Bernstein docu- mented eroded under pressure — rom globalization, the entry o — Indian diamond cutters who operated outside the DDC’s ethnic en — orcement in — rastructure, the growth o — online retail that disrupted traditional distribution channels, and generational changes that weakened communal bonds. 54 As governance quality declined, the costs spread beyond the DDC’s

50 See Richman, supra note 3, at 250–55 (describing the DDC’s membership and the diamond industry’s downstream structure). 51 FINRA, 2023 Annual Report (reporting membership and registered representative totals). Total U.S. equity market capitalization exceeded $50 trillion in 2024. 52 A — ormal model extending the Hirshlei — er-Rasmusen ostracism — ramework to incorporate court intervention probability con — irms this mechanism. See Appendix. The model shows that as the probability o — judicial override increases above a threshold, cooperation collapses and governance quality degrades. Because non-members outnumber members and cannot substitute alternative governance, the aggregate external wel — are loss exceeds the aggregate internal loss. The speci — ic ratio is parameter-dependent — it varies with the relative size o — the member and non-member populations, the marginal externality per unit o — governance quality, and the shape o — the cooperation-collapse — unction — but the directional — inding is robust: external losses dominate internal losses across a wide range o — plausible parameterizations. 53 Richman, supra note 3, at 264–76. 54 Id. at 257–64 (documenting — ive — actors in the erosion o — diamond-industry cooperation: globalization o —

cutting centers, entry o

new ethnic groups into the trade, development o — online retail, declining margins, and generational shi — ts in communal identity). 14 “Private” Governance Is Actually a Club Good

membership. Banks reduced credit to diamond dealers, increasing

inancing costs through- out the industry. Retailers investing in branded jewelry became vulnerable to misrepresenta- tion o — synthetic diamonds — a — raud risk that the DDC’s governance had previously sup- pressed. Vertical integration replaced trust-based exchange as the dominant transactional

orm, eliminating the low-cost governance mechanism and substituting a costlier institutional structure that served members’ narrow interests but generated no comparable externalities

or non-members.55 The collapse o — FTX in November 2022 illustrates the magnitude o — the externality that governance prevents, even though its causal mechanism di —

ers — rom the thesis. FTX operat- ed as a centralized cryptocurrency exchange with a governance system that set trading rules and managed custody o — customer assets.56 FTX’s governance did not degrade through judi- cial intervention; it was — raudulent — rom inception. But the scale o — spillover harm demon- strates why the externality matters: customer losses exceeded $8 billion; contagion bankrupt- ed BlockFi, Genesis Global, and multiple hedge — unds; institutional investors including the Ontario Teachers’ Pension Plan and Sequoia Capital su —

ered hundreds o — millions in losses; and the Solana ecosystem lost roughly 60% o — its token value, harming projects with no di- rect FTX relationship.57 When governance is absent or — raudulent, the external costs dwar —

the internal ones — exactly the asymmetry the

ramework predicts would result — rom judicial degradation o —

unctioning governance. The overruling o — Chevron de — erence in Loper Light Enterprises v. Raimondo provides a third case, though it is too recent — or empirical measurement o — governance degradation.58 The

ramework predicts that removing the Pigouvian subsidy Chevron provided will shi — t agency resources — rom substantive governance to de — ensive litigation, narrowing interpretations and leaving regulatory gaps. Coglianese and Walters have identi — ied a paradox consistent with this prediction: agencies have begun invoking Loper Light itsel — to justi — y regulatory changes without notice-and-comment procedures, using the removal o — de — erence as a reason to by- pass the procedural constraints that previously accompanied agency governance.59 Whether the compound wel — are loss materializes is an empirical question. That the — ramework gener- ates a testable prediction — or the most signi — icant administrative law development in — our decades is itsel — evidence o — the classi — ication’s analytical utility.

55 Id. at 275. The replacement o — trust-based governance with vertical integration is itsel — evidence o — the compound wel — are loss. Vertical integration serves the integrated — irm’s members but does not generate the — raud-reduction externality that trust-based governance produced — or the entire supply chain. The shi — t

rom club good to private good eliminates the spillover. 56 In re FTX Trading Ltd., Case No. 22-11068 (Bankr. D. Del.) ( — irst-day declaration o — John J. Ray III: “Never in my career have I seen such a complete — ailure o — corporate controls.”). 57 These — igures are drawn — rom bankruptcy — ilings, public company disclosures, and contemporaneous reporting. See Hilary J. Allen, DeFi: Shadow Banking 2.0?, 64 Wm. & Mary L. Rev. 919 (2023) (analyzing crypto governance — ailures and their systemic e —

ects). The FTX illustration is not a controlled experiment — the collapse involved — raud and mismanagement, not judicial displacement o — governance. But it demonstrates the mechanism: when private governance — ails, the external costs spread — ar beyond the governed community. 58 Loper Light Enters. v. Raimondo, 144 S. Ct. 2244 (2024). 59 Cary Coglianese & Daniel E. Walters, The Great Unsettling: Administrative Governance A — ter Loper Light, 77 ADMIN. L. REV. 101 (2025). 2026-03-18] 15

De

erence as Pigouvian Subsidy I — the compound wel — are loss is real, one would expect the legal system to have devel- oped some response to it — even without a theoretical — ramework explaining why. Courts are not economists, and common-law judges do not read Pigou be — ore deciding whether to review an expulsion — rom a trade association. But judges do observe consequences. A judge who reviews a DDC arbitration award and orders reinstatement o — an expelled member ob- serves, over repeated cases, that the bourse’s en — orcement authority weakens, that compli- ance declines, and that the — raud costs the bourse previously suppressed begin to spread through the supply chain. A judge who re — uses to second-guess a FINRA disciplinary action observes that the exchange’s en — orcement remains credible and that the market — unctions with — ewer regulatory interventions — rom the SEC. Courts that repeatedly con — ronted these patterns developed, over decades, a set o — doctrines that share a common structure: pre- sumptive de — erence to governance decisions, rebuttable under speci — ic conditions. The busi- ness judgment rule, the FAA’s presumption o — arbitrability, the antitrust rule o — reason, and common-law de — erence to voluntary associations each emerged — rom di —

erent doctrinal soil, justi — ied by di —

erent rationales, in di —

erent areas o — law. None was designed as an economic intervention. Yet each per — orms the same economic — unction: it reduces the cost o — produc- ing governance, partially correcting the undersupply that positive-externality theory predicts. Standard Pigouvian theory identi — ies the mechanism. Goods generating positive external- ities will be undersupplied at market equilibrium, because producers cannot capture the — ull social bene — it o — their output.60 The corrective is a subsidy equal to the marginal external bene — it, bringing production to the socially optimal level.61 The intuition is straight — orward: i —

a beekeeper’s hives pollinate neighboring

arms, the beekeeper will keep — ewer hives than the community needs because she cannot charge the — armers — or the pollination. A Pigouvian subsidy compensates the beekeeper — or the external bene — it, increasing production to the level that accounts — or the — armers’ gain. I — governance is a club good that generates positive externalities — as Part I established — then governance will be undersupplied in the absence o — a corrective mechanism. The — ormal model in the Appendix con — irms this intuition and adds a structural prediction: total wel — are is maximized neither at blanket de — erence (which permits governance abuse without check) nor at plenary review (which destroys the ostra- cism mechanism’s credibility), but at an interior point where courts intervene in a small — rac- tion o — governance decisions. This is precisely the structure these doctrines embody — each creates a legal presumption rather than an absolute immunity, allowing judicial override un- der de — ined conditions while protecting governance production as the de — ault. The doctrines are not Pigouvian subsidies by design. They are Pigouvian subsidies by — unction. This Part makes that — unction explicit.

60 A.C. Pigou, THE ECONOMICS OF WELFARE 172–203 (1920). 61 William J. Baumol & Wallace E. Oates, THE THEORY OF ENVIRONMENTAL POLICY 21–35 (1975). The Pigouvian subsidy is the mirror image o — the Pigouvian tax: where negative externalities warrant a tax to reduce production to the social optimum, positive externalities warrant a subsidy to increase production to the social optimum. 16 “Private” Governance Is Actually a Club Good

The Economic Logic Courts do not write checks to private governance bodies. The subsidy operates indirect- ly, by reducing governance production costs. Undersupply, in this context, means that gov- erning bodies invest less in governance than the social optimum — weaker monitoring, less vigorous en — orcement, — ewer sanctions, thinner institutional in — rastructure — because they cannot charge non-members — or the — raud reduction, market integrity, or environmental quality those non-members receive. Producing governance is expensive: governing bodies must invest in monitoring systems, arbitration in — rastructure, en — orcement capacity, and rep- utational capital. These investments are at risk every time a governance decision is challenged in court. Litigation costs — attorney — ees, discovery, the management distraction o — de — end- ing a governance decision, the uncertainty o — outcome — are direct costs o — governance production that are incurred even when the governance body prevails. The prospect o — litiga- tion also generates indirect costs: risk-averse governance bodies moderate their en — orcement, reduce the — requency or severity o — sanctions, and invest in de — ensive documentation rather than substantive monitoring.62 De — erence doctrines reduce these costs by limiting the circumstances under which gov- ernance decisions can be challenged and the intensity o — judicial review when challenges oc- cur. A governance body operating under a de — erence regime invests in governance with greater con — idence that its decisions will stand, — aces lower expected litigation costs, and can allocate resources to substantive monitoring rather than de — ensive procedures. The subsidy is the di —

erence between governance production costs with de — erence and governance produc- tion costs without it. An economist might object that cost avoidance is not the same as a direct trans — er — that shielding a governance body — rom litigation is di —

erent — rom writing it a check. For the governance producer’s investment decision, the distinction is immaterial. A

irm deciding whether to invest in monitoring, arbitration, and en — orcement capacity re- sponds identically to a dollar o — reduced litigation cost and a dollar o — direct subsidy: both lower the marginal cost o — governance production, increasing the quantity produced toward the social optimum. What makes the mechanism Pigouvian is not the — orm o — the trans — er

62 C — . Aurelio Gurrea-Martínez, Re-Examining the Law and Economics o — the Business Judgment Rule — rom a Comparative Perspective, 18 J. CORP. L. STUD. 417 (2018) (observing that the BJR “may also create a positive externality — or society: the promotion o — innovation and development” — the closest prior observation that a de — erence doctrine generates positive externalities, though not developed into a systematic — ramework). The cost-reduction mechanism this Article identi — ies is more general: every de — erence doctrine reduces the expected cost o — producing governance, and this cost reduction — unctions as a subsidy — or a positive-externality good. 2026-03-18] 17

but the

unction — correcting the undersupply o — a positive-externality good by reducing the private cost o — producing it.63

De

erence Doctrines Recharacterized Four doctrines that currently lack a common theoretical — oundation share this structure. Each was developed — or domain-speci — ic reasons — corporate law, arbitration policy, anti- trust, associational — reedom — and each is justi — ied by its own doctrinal rationale. The gov- ernance-as-club-good — ramework reveals them as instances o — a single economic mechanism: subsidizing the production o — a positive-externality good.

The Business Judgment Rule The business judgment rule insulates directors — rom personal liability — or honest business decisions made on an in — ormed basis, in good — aith, and without con — licts o — interest.64 Its conventional justi — ications are domain-speci — ic: reducing managerial risk aversion, preventing judicial incompetence in business decisions, avoiding hindsight bias, and providing legal cer- tainty. 65 Each o — these rationales explains why de — erence bene — its the corporation and its shareholders. None explains why de — erence bene — its anyone outside the — irm. The club-goods — ramework supplies the missing rationale. Board governance is a club good: excludable (directors can be removed, o —

icers terminated, shareholders who challenge board authority — ace procedural barriers), congestible (board deliberation quality degrades with excessive outside inter — erence), and generating positive externalities — or non- shareholders — innovation, employment, economic growth, and the commercial stability that — unctional corporate governance contributes to the broader economy. The BJR reduces governance production costs by insulating directors — rom the litigation risk that would oth-

63 The Pigouvian characterization is — unctional rather than — ormal. Courts do not calculate marginal external bene — its or calibrate subsidy magnitudes. But the economic literature on implicit subsidies supports the equivalence: Stanley Surrey’s — oundational work on tax expenditures demonstrated that tax deductions, exemptions, and credits — unction as government spending by other means, with identical e —

ects on investment incentives. See Stanley S. Surrey, Pathways to Tax Re — orm: The Concept o — Tax Expenditures (1973). Legal doctrines that shield producers — rom liability — charitable immunity, sovereign immunity, statutory caps on damages — similarly reduce the cost o — activity by eliminating litigation risk, — unctioning as implicit subsidies — or the shielded conduct. De — erence doctrines operate through the same mechanism. A — ormal wel — are model speci — ying the magnitude o — the implicit subsidy each de — erence doctrine provides, comparing it to the marginal external bene — it o — governance, and determining whether current de — erence levels approximate the social optimum is a subject — or — uture work. 64 See Aronson v. Lewis, 473 A.2d 805, 812 (Del. 1984) (establishing the BJR as a “presumption that in making a business decision the directors o — a corporation acted on an in — ormed basis, in good — aith and in the honest belie — that the action taken was in the best interests o — the corporation”). 65 See Stephen M. Bainbridge, The Business Judgment Rule as Abstention Doctrine, 57 VAND. L. REV. 83, 108–30 (2004) (authority-preserving rationale); Frank H. Easterbrook, Managers’ Discretion and Investors’ Wel — are: Theories and Evidence, 9 DEL. J. CORP. L. 540 (1984) (risk-aversion rationale). 18 “Private” Governance Is Actually a Club Good

erwise chill bold decision-making, independent judgment, and long-term investment in mon- itoring and compliance. It is an implicit Pigouvian subsidy — or a positive-externality good.66

The Federal Arbitration Act The FAA’s “liberal — ederal policy — avoring arbitration agreements” reduces the cost o —

maintaining private dispute resolution systems by en

orcing arbitration clauses and limiting judicial review o — arbitral awards.67 Without the FAA, parties could agree to arbitrate and then litigate when the arbitral outcome proved un — avorable — — ree-riding on the arbitration system’s investment in in — rastructure, expertise, and procedural — airness while re — using to be bound by its results. Arbitration systems are governance club goods: excludable (participation requires agree- ment), congestible (tribunal capacity is — inite), and generating positive externalities — or the judicial system and the broader commercial community — dispute resolution that reduces court congestion, develops specialized commercial norms, and provides — aster, cheaper reso- lution that bene — its all market participants who rely on the stability o — commercial relation- ships.68 The FAA’s presumption o — arbitrability is a Pigouvian subsidy that reduces arbitra- tion production costs by eliminating the — ree-rider problem that would otherwise undermine investment in private dispute resolution.

The Antitrust Rule o

Reason The rule o — reason reduces antitrust exposure — or cooperative governance by applying a balancing test rather than per se condemnation to coordinated conduct that may enhance e —

iciency.69 Without it, any governance mechanism that involves coordinated behavior — joint standard-setting, collective en — orcement, membership criteria, shared monitoring — would be per se illegal under Section 1 o — the Sherman Act, because all such mechanisms require agreement among competitors. Cooperative governance generates positive externalities — or the markets it regulates: qual- ity standards reduce search costs — or all buyers, not just cooperative members; collective en-

orcement deters — raud beyond the cooperative’s membership; shared monitoring produces in — ormation that bene — its regulators, investors, and consumers. The rule o — reason is a

66 This re — raming does not change the BJR’s operational content — directors still receive de — erence — or in — ormed, good- — aith, disinterested decisions. What it changes is the justi — ication — or that de — erence. The conventional rationale is e —

iciency internal to the — irm: shareholders are better o —

when directors are protected — rom hindsight bias. The club-goods rationale is wel — are external to the — irm: society is better o —

because the BJR enables governance that produces spillover bene

its no one else would provide at equivalent quality. 67 Moses H. Cone Mem’l Hosp. v. Mercury Constr. Corp., 460 U.S. 1, 24 (1983); CompuCredit Corp. v. Greenwood, 565 U.S. 95, 98 (2012) (FAA “requires courts to en — orce agreements to arbitrate according to their terms” unless “overridden by a contrary congressional command”). 68 Drahozal, supra note 56, at 585–86 (examining the “positive externalities hypothesis” and — inding that the concern about lost precedent is overstated because arbitration may displace settlements rather than trials, and arbitrators develop their own body o — specialized precedent). 69 See Broadcast Music, Inc. v. Columbia Broad. Sys., Inc., 441 U.S. 1, 19–20 (1979) (blanket licensing integrating “sales, monitoring, and en — orcement” is not per se price- — ixing when it creates e —

iciencies individual actors could not achieve); Northwest Wholesale Stationers, 472 U.S. at 295. 2026-03-18] 19

Pigouvian subsidy that enables cooperative governance to exist by reducing the antitrust cost o — producing it. Per se treatment is the withdrawal o — that subsidy — and AP and FOGA show that withdrawal is appropriate when the governance mechanism’s true purpose is competition-suppressing rather than externality-producing.70

De

erence to Voluntary Associations Common-law de — erence to voluntary associations — the principle that courts re — rain

rom inter — ering in associations’ internal a —

airs absent bad — aith, procedural un — airness, or violation o — the association’s own rules — is the oldest and least theorized o — the — our doc- trines.71 It rests on — reedom o — association and contract principles, but these justi — ications explain why members should be — ree to — orm associations, not why courts should de — er to associations’ governance decisions once — ormed. Association governance is the paradigm governance club good. Pro — essional associations maintain ethical standards that bene — it all clients o — the pro — ession, not just the association’s members. Trade associations develop industry norms that reduce transaction costs — or all market participants. Religious communities sustain social institutions — charity, mutual aid, moral education — whose bene — its radiate beyond the congregation. De — erence reduces the litigation cost o — maintaining these governance systems. When Pinsker and Falcone increased judicial oversight o — association membership decisions, they increased governance produc- tion costs — or every association operating within those jurisdictions — costs that the exter- nality — ramework predicts will reduce governance supply at the margin.72 The — our doctrines share a common structure despite their doctrinal diversity. Each re- duces the expected cost o — producing governance, enabling governance bodies to invest in the monitoring, en — orcement, and institutional in — rastructure that generates positive external- ities. Each is justi — ied on domain-speci — ic grounds that do not re — erence externalities. And each produces the wel — are consequence that Pigouvian theory predicts: by subsidizing a posi- tive-externality good, it brings production closer to the social optimum than unsubsidized production would achieve. The governance-as-club-good — ramework does not argue that these doctrines were consciously designed as Pigouvian subsidies. It argues that they — unc-

70 The cartel cases thus — it within the — ramework rather than undermining it. Per se treatment is not the absence o — a Pigouvian subsidy; it is the recognition that the activity in question generates negative externalities (reduced competition, higher prices, restricted output) rather than positive ones. The

ramework’s proposed standard — rebuttable presumption o — validity, overcome by showing anticompetitive purpose — operationalizes the distinction between positive-externality governance that warrants the subsidy and negative-externality cartelization that does not. 71 See Auto. Elec. Serv. Corp. v. Ass’n o — Auto. A — termarket Distribs., 747 F. Supp. 1483, 1489 (E.D.N.Y. 1990) (“It is the proper policy o — the Court to re — rain — rom unnecessarily inter — ering in the internal a —

airs o — a private, not- — or-pro — it trade association . . . . Absent bad — aith or bias, this Court would not intervene.”); Pinsker v. Pac. Coast Soc’y o — Orthodontists, 526 P.2d 253, 256 (Cal. 1974) (“Courts should not attempt to

ix a rigid procedure that must invariably be observed. Instead, the associations themselves should retain the initial and primary responsibility — or devising a method.”). 72 See Falcone v. Middlesex Cnty. Med. Soc’y, 170 A.2d 791 (N.J. 1961). The — ramework does not argue that Pinsker and Falcone were wrongly decided — both involved monopolistic associations where the exit mechanism that disciplines club governance was absent, and both imposed procedural — loors rather than substantive review. The point is that the increased governance production costs those decisions impose are real, and the externality — ramework explains why courts should be parsimonious in expanding them. 20 “Private” Governance Is Actually a Club Good

tion as such, and that recognizing this

unction provides a uni — ied rationale — or calibrating de — erence. Whether this pattern extends beyond private governance to public agency gov- ernance is the question the next subsection takes up.

Stress Test: Chevron De

erence and Agency Governance Administrative agencies are not Buchanan clubs. Citizens do not voluntarily join the EPA’s jurisdiction, and agency budgets come — rom taxation rather than member dues. Ap- plying the club-goods — ramework to agency governance requires a structural adjustment that the previous — our doctrines did not need. The adjustment is this: the “club” in the adminis- trative context is not the agency but the regulated industry. Firms voluntarily enter regulated sectors — energy, — inance, pharmaceuticals, telecommunications — and in doing so subject themselves to the agency’s governance, just as a broker-dealer voluntarily enters the securi- ties industry and thereby subjects itsel — to FINRA.73 The agency is the governance mecha- nism — or the industry club, not the club itsel — . This reconceptualization — rames Chevron de — - erence and its overruling as a stress test o — whether the club-good — ramework’s predictions hold in the public governance domain where voluntary membership is contested and institu- tional variety is greatest. Chevron de — erence reduced the cost o — agency governance by allowing agencies to inter- pret ambiguous statutes without the constant risk that courts would substitute their own judgment. Agencies could invest in developing expertise, promulgating detailed regulations, and building consistent en — orcement programs with con — idence that their reasonable inter- pretations would survive judicial review. The subsidy operated through two channels: re- duced litigation costs (agencies did not need to de — end every interpretive choice de novo) and reduced uncertainty costs (regulated parties could rely on agency interpretations, reduc- ing compliance costs and enabling long-term planning). The positive externalities o — agency governance extend well beyond the regulated industry. Specialized regulation produces envi- ronmental protection, — inancial stability, consumer sa — ety, and public health bene — its that ac- crue to the general public — non-members o — the industry club who bear none o — the com- pliance costs. These externalities depended on the governance quality that Chevron de — er- ence enabled. When the Supreme Court overruled Chevron in Loper Light Enterprises v. Raimondo, it re- moved the subsidy.74 The compound wel — are loss — ramework predicts that removing Chev- ron de — erence will degrade agency governance quality and reduce the positive externalities it generates. Agencies — acing de novo judicial review o — every statutory interpretation will shi — t resources — rom substantive governance to de — ensive litigation. Risk-averse agencies will nar- row their interpretations, leaving regulatory gaps that no other institution will — ill. Coglianese

73 The voluntariness o — industry entry distinguishes this application — rom the objection that all citizens are involuntarily subject to regulation. The relevant population is not all citizens but the — irms that chose to operate in a regulated domain. Their entry into the domain is voluntary in the same sense as a diamond trader’s entry into the DDC or a broker-dealer’s entry into the securities industry. The positive externalities o — agency governance — environmental protection, — inancial stability, consumer sa — ety — — low to the non- member public, just as the DDC’s — raud reduction — lows to downstream consumers. 74 Loper Light Enters. v. Raimondo, 144 S. Ct. 2244, 2261–73 (2024) (holding that courts must exercise independent judgment under APA § 706 rather than de — erring to agency statutory interpretations). 2026-03-18] 21

and Walters have identi

ied a pattern consistent with this prediction: agencies invoking Loper Light itsel — to justi — y regulatory changes without notice-and-comment procedures, using the removal o — de — erence as a reason to bypass the procedural constraints that previously ac- companied agency governance.75 I — this pattern holds, the overruling o — Chevron will have degraded agency governance quality while simultaneously weakening the procedural sa — e- guards that legitimized it — a compound loss o — exactly the kind the — ramework predicts. Whether this prediction holds is an empirical question that — uture research can test, making Loper Light’s governance e —

ects the next critical test o — the — ramework’s predictive power.

Calibrated De

erence I — governance is a club good that generates positive externalities, then courts that dis- place the governance mechanism destroy value not only — or members but — or non-members who bene — it — rom its spillovers. Parts II and III develop this claim and its doctrinal implica- tions. Be — ore reaching those arguments, however, the thesis must survive — our serious objec- tions. Private governance mechanisms have — unctioned as cartels, en — orced discriminatory exclusion, and operated without the transparency and error correction that judicial review provides. Courts are not wrong to be skeptical. The question is whether skepticism should be the de — ault or whether a more calibrated approach — presumptive de — erence overcome by speci — ic showings — better accounts — or the wel — are consequences o — displacing govern- ance. This Part takes each objection at — ull strength be — ore proposing a standard.

The Cartel Objection The most doctrinally developed challenge to governance de — erence is that private gov- ernance mechanisms mask anticompetitive conduct. Three Supreme Court decisions estab- lish the point with unmistakable — orce. The Associated Press maintained bylaws that restricted membership and prohibited members — rom selling news to nonmembers — governance rules that, in — orm, resembled the quality-control and exclusivity provisions o — any trade association.76 The Court — ound a per se Sherman Act violation. “Arrangements or combinations designed to sti — le competition cannot be immunized by adopting a membership device accomplishing that purpose.”77 The Fashion Originators’ Guild o — America operated an even more elaborate governance appa- ratus: a design registration system, inspection and auditing procedures, tribunals, and gradu- ated — ines — institutional in — rastructure that, judged by its — ormal attributes alone, was indis- tinguishable — rom the DDC’s private legal system.78 The Court struck it down without in- quiring into the reasonableness o — the Guild’s methods. “The reasonableness o — the methods pursued by the combination to accomplish its unlaw — ul object is no more material than

75 Coglianese & Walters, supra note 53, at 130–40. 76 Associated Press v. United States, 326 U.S. 1 (1945). 77 Id. at 4. 78 Fashion Originators’ Guild o — Am., Inc. v. FTC, 312 U.S. 457, 461–63 (1941) (describing the Guild’s registration, inspection, and en — orcement apparatus). 22 “Private” Governance Is Actually a Club Good

would be the reasonableness o

the prices — ixed by unlaw — ul combination.”79 And in Silver v. New York Stock Exchange, the Court held that the NYSE’s collective termination o — a non- member broker-dealer’s wire connections violated the Sherman Act because the exchange provided no notice, no hearing, and no stated basis — or its action — even though the Securi- ties Exchange Act authorized exchange sel — -regulation.80 These cases pose a genuine problem — or the governance-as-club-good thesis. I — the ana- lytical — ramework cannot distinguish the AP’s membership bylaws — rom the DDC’s, or FO- GA’s tribunals — rom FINRA’s disciplinary process, it provides no operational guidance to courts. The objection is not that cartels exist — everyone knows that — but that the — ormal attributes o — governance mechanisms (rules, exclusion, en — orcement, ostracism) are identical regardless o — whether the mechanism serves governance purposes or anticompetitive ones. The — ramework can make this distinction, but the distinction operates on purpose and e —

ect, not institutional — orm. Governance-serving exclusion targets members who accepted governance obligations and then violated them: the diamond trader who cheated on a trans- action, the broker-dealer who engaged in — raudulent trading, the commons user who over- harvested. The excluded party was a participant in the governance system whose de — ection threatened the club good. Competition-suppressing exclusion targets parties who never sought to participate in governance but merely competed in the same market: the newspaper that wanted to publish AP-sourced stories, the retailer that sold garments the Guild had not registered, the broker-dealer that competed with NYSE members — or order — low. The ex- cluded party was not a de — ector — rom governance but a competitor whose presence was commercially inconvenient. This distinction tracks the rule o — reason as the Supreme Court already applies it. In Northwest Wholesale Stationers, the Court held that expulsion — rom a purchasing cooperative was not per se illegal absent a showing o — market power or exclusive access to essential — acili- ties, precisely because cooperatives are “designed to increase economic e —

iciency and render markets more, rather than less, competitive.” 81 The governance-as-club-good — ramework provides the economic rationale — or that holding: cooperative governance is a positive- externality good, and expulsion decisions that maintain governance quality are presumptively wel — are-enhancing. The presumption is overcome when the challenger demonstrates that exclusion targeted competitive conduct rather than governance violations — which is exactly what the AP, FOGA, and Silver plainti —

s showed. The — ramework thus changes the burden, not the test. Under the current rule o — reason, courts evaluate the competitive e —

ects o — governance decisions without any presumptive direction. Under the proposed — ramework, governance exclusion decisions carry a rebuttable presumption o — validity because they maintain a positive-externality good. The challenger must demonstrate anticompetitive purpose — a showing that shi — ts the analysis — rom de — er-

79 Id. at 468. 80 Silver v. New York Stock Exch., 373 U.S. 341, 347–49 (1963). 81 Northwest Wholesale Stationers, Inc. v. Pac. Stationery & Printing Co., 472 U.S. 284, 295 (1985). 2026-03-18] 23

ence to scrutiny. This is not a novel doctrinal innovation; it is an economic justi

ication — or the structure that Northwest Wholesale already implies.82

The Discrimination Objection The most morally serious challenge to governance de — erence is that the most e —

ective private governance systems in the empirical record — diamond bourses, Maghribi trader coalitions, ethnic Chinese middleman groups — rely on ethnic or religious homogeneity as their core en — orcement mechanism.83 De — erence to exclusion decisions necessarily risks de — - erence to exclusion along protected characteristics. Two Supreme Court decisions — rame the doctrinal collision. In Roberts v. United States Jay- cees, the Court held that Minnesota’s anti-discrimination statute could compel the Jaycees to accept women as regular members without violating — reedom o — association, because the state’s compelling interest in eradicating gender discrimination outweighed the burden on associational — reedom — particularly where the Jaycees were large, nonselective, and had produced no evidence that women’s admission would impede the organization’s protected activities.84 In Boy Scouts o — America v. Dale, the Court reached the opposite result, holding that New Jersey’s public accommodations law could not compel the Boy Scouts to readmit a gay assistant scoutmaster because — orced membership would “signi — icantly burden” the organiza- tion’s expressive message.85 The Court de — erred to the Scouts’ own characterization o — its values: “We are not, as we must not be, guided by our views o — whether the Boy Scouts’ teachings with respect to homosexual conduct are right or wrong.”86 The governance-as-club-good — ramework does not de — end ethnic exclusion. It draws a distinction that Roberts and Dale together support: exclusion based on governance-relevant conduct warrants de — erence; exclusion based on ascriptive characteristics does not. When the DDC expels a member — or cheating on a diamond transaction, the exclusion criterion is gov- ernance-relevant — the member violated rules that maintain the club good. When a medical society excludes a quali — ied physician because o — his ethnicity, the exclusion criterion is as- criptive — it bears no relationship to governance — unction.87 The — act that ethnic homogene- ity happens to — acilitate governance en — orcement, by reducing monitoring costs and increasing

82 A bright-line indicator rein — orces the distinction: exclusion o — parties who previously accepted and then violated governance rules is presumptively governance-serving; exclusion o — parties who never sought membership but merely competed with members is presumptively anticompetitive. Courts already employ this distinction implicitly. Making it explicit, and grounding it in club-goods economics, gives the analysis a theoretical — oundation it currently lacks. 83 Landa, supra note 30, at 349–56 (describing how ethnic identity markers — kinship, clan names, religious observance — — unction as low-cost screening devices that enable trust-based exchange); Avner Grei — , Contract En — orceability and Economic Institutions in Early Trade: The Maghribi Traders’ Coalition, 83 AM. ECON. REV. 525, 528–32 (1993) (documenting how the Maghribi traders’ ethnic network sustained cooperation through multilateral punishment strategies). 84 Roberts v. United States Jaycees, 468 U.S. 609, 623–29 (1984). 85 Boy Scouts o — Am. v. Dale, 530 U.S. 640, 655–56 (2000). 86 Id. at 661. 87 See Falcone v. Middlesex Cnty. Med. Soc’y, 170 A.2d 791, 795–96 (N.J. 1961) (holding that a medical society with virtual monopoly over hospital access could not arbitrarily exclude quali — ied physicians; “a monopoly raises duties which may be en — orced against the possessors o — the monopoly”). 24 “Private” Governance Is Actually a Club Good

social sanctions, does not mean it is necessary

or governance. Bernstein documented the DDC’s complete private legal system operating across multiple ethnic communities within the diamond trade; the system’s en — orcement power derived — rom commercial reputation and WFDB-wide ostracism, not — rom ethnic solidarity alone.88 Roberts is correctly decided under this — ramework. The Jaycees’ gender exclusion was not tied to governance — unction — the organization produced no evidence that women’s mem- bership would degrade the quality o — its civic programming, leadership training, or communi- ty service — and gender is an ascriptive characteristic unrelated to the rules the organization en — orces. De — erence was appropriately denied. Dale is also supportable under this — rame- work, though — or a narrower reason than the Court articulated: the Boy Scouts’ exclusion was tied to its stated expressive mission, and the Court de — erred to the organization’s charac- terization o — how — orced membership would a —

ect that mission — a — orm o — calibrated de — - erence to governance substance, not a blanket endorsement o — any exclusion criterion the organization might adopt.89 Non-discrimination principles represent a competing positive-externality good. Dignity, equal citizenship, and equal market access generate social bene — its that extend — ar beyond the individuals directly protected — they sustain the legitimacy o — markets and institutions — or all participants.90 When the governance club good and the anti-discrimination good collide, the anti-discrimination good prevails, because its externalities are broader and its moral — ounda- tion is stronger. The — ramework handles this collision not by abandoning de — erence but by speci — ying that the rebuttable presumption o — validity does not extend to exclusion criteria based on protected characteristics.

Judicial Externalities and the Ostrom Objection Two additional objections challenge the — ramework — rom di —

erent directions. The — irst argues that judicial review itsel — generates positive externalities — precedent, transparency, error correction — that blanket de — erence would destroy.91 The second, drawing on Elinor Ostrom’s research, argues that governance communities sustain collective goods without

88 Bernstein, supra note 12, at 130–38. Lisa Bernstein and Brad Peterson have shown more recently that contract governance regimes can sca —

old inter- — irm trust without pre-existing community ties. See Lisa Bernstein & Brad Peterson, Managerial Contracting: A Preliminary Study, 14 J. LEGAL ANALYSIS 176 (2022). 89 The tension between Roberts and Dale does not undermine the — ramework; it illustrates why calibration is necessary. The relevant question is whether the exclusion criterion tracks governance — unction, not whether the organization labels itsel — “expressive.” A pro — essional association that excludes members — or violating ethical rules warrants de — erence on the substance o — the violation — inding; the same association excluding applicants by race does not, regardless o — how it characterizes its mission. 90 See Roberts, 468 U.S. at 625 (“Acts o — invidious discrimination in the distribution o — publicly available goods, services, and other advantages cause unique evils that government has a compelling interest to prevent — wholly apart — rom the point o — view such conduct may transmit.”). 91 See Owen M. Fiss, Against Settlement, 93 YALE L.J. 1073, 1085–89 (1984) (arguing that adjudication produces public goods — authoritative interpretations o — law — that settlement and private ordering cannot replicate). 2026-03-18] 25

absolute excludability, undermining the claim that excludability is essential to governance quality.92 The judicial-externality objection is real but asymmetric. Governance externalities are de- stroyed when courts displace the ostracism mechanism, because governance quality depends on the credible threat o — exclusion — a credibility that is binary in its e —

ect on member be- havior even i — excludability itsel — is spectral. I — members know that courts will reverse expul- sion decisions, the expected cost o — de — ection — alls, and cooperation unravels.93 Judicial ex- ternalities, by contrast, are merely — oregone under de — erence — they are not destroyed but can be generated through other cases. The judicial system is robust; precedent can be produced by adjudicating cases that do not involve governance exclusion decisions. The governance system is — ragile; once the ostracism mechanism is disabled, there is no substitute institution that can produce the same externalities at equivalent cost.94 Drahozal reached a parallel con- clusion in the arbitration context, — inding that the concern about lost precedent — rom private dispute resolution is overstated because arbitration may displace settlements rather than tri- als, and because arbitrators develop their own body o — precedent within specialized indus- tries.95 The Ostrom objection requires absorption rather than resistance. Ostrom demonstrated that communities sustain governance through graduated sanctions, reputation networks, and social pressure — mechanisms that maintain cooperation without the bright-line exclusion that diamond bourses or stock exchanges employ.96 These are excludability mechanisms, but so — ter ones, operating on a spectrum rather than as an on-o —

switch. The relevant question is not whether governance requires absolute excludability but whether judicial intervention pushes excludability below the threshold needed to sustain cooperation in a given institutional context. Small, tight-knit commons communities operate above this threshold through social mechanisms because members interact repeatedly, in — ormation travels quickly, and reputa- tion costs are high. Larger, more impersonal governance systems require — ormal mechanisms because the social in — rastructure is thinner.97

92 See Ostrom, Beyond Markets and States, supra note 25, at 653–55 (documenting communities that sustain commons governance through graduated sanctions, reputation, and social pressure rather than absolute exclusion). 93 This is the mechanism the game-theoretic model in the Appendix — ormalizes. See Appendix (extending the Hirshlei — er-Rasmusen ostracism — ramework to incorporate court intervention probability). 94 The — ragility asymmetry is not merely theoretical. When Richman documented the DDC’s decline, the trust-based exchange system was not replaced by a superior governance mechanism — it was replaced by vertical integration, a costlier institutional — orm that eliminated the externality-generating mechanism entirely. Richman, supra note 3, at 275. 95 Christopher R. Drahozal, Privatizing Civil Justice: Commercial Arbitration and the Civil Justice System, 9 KAN. J.L. & PUB. POL’Y 578, 585–86 (2000). 96 Ostrom, Governing the Commons, supra note 25, at 90–102. 97 Kiesling’s “non-replicable club” concept identi — ies where the Ostrom objection has the most — orce. When the governed resource cannot be replicated — a transmission grid, a natural monopoly, a single regional hospital — members cannot exit and — orm competing governance systems. In these settings, the exit mechanism that disciplines Buchanan clubs is absent, and the case — or judicial oversight is strongest. Kiesling, supra note 6. The calibrated de — erence — ramework accounts — or this: non-replicable governance systems warrant lower de — erence precisely because their members cannot discipline governance — ailures through exit. 26 “Private” Governance Is Actually a Club Good

One  --- urther objection warrants brie ---  address. Louis Kaplow and Steven Shavell demon- strated that liability rules (judicial intervention with compensation) can be more e ---

icient than property rules (absolute de — erence) when courts can reasonably assess harm.98 Govern- ance is a context where their own — ramework — avors property-rule protection. The value o —

governance to the club is di

icult — or courts to assess because it is intangible — composed o — reputation, trust, monitoring capacity, and en — orcement credibility, none o — which has a market price. The governance mechanism is destroyed rather than merely diminished by in- tervention — ostracism credibility operates as a threshold — unction: below a critical probabil- ity o — judicial override, the threat disciplines cooperation; above that threshold, it does not, and cooperation collapses nonlinearly. And the market — or governance is thin — there is no easy substitute when a particular governance system is disabled. These are precisely the con- ditions under which Calabresi and Melamed concluded that property rules dominate.99

The Proposed Standard The — ramework supports a standard o — calibrated de — erence that mirrors structures courts already employ but provides an economic rationale they currently lack. Courts review- ing governance exclusion decisions should apply a rebuttable presumption o — validity: the decision stands unless the challenger demonstrates one o — three rebuttal conditions. First, anticompetitive purpose. The exclusion served to suppress competition rather than maintain governance standards. The burden — alls on the challenger to show that the excluded party was targeted — or competitive conduct, not governance violations. This tracks the rule o — reason as applied in Northwest Wholesale and distinguishes governance-serving exclusion (Silver’s process — ailure notwithstanding, the NYSE’s sel — -regulatory authority was not itsel —

condemned)

rom competition-suppressing exclusion (AP, FOGA). Second, discriminatory criteria. The membership or exclusion criteria track ascriptive char- acteristics rather than governance-relevant conduct. The burden — alls on the challenger to show that the criteria are pretextual — that they serve to exclude on the basis o — identity ra- ther than to maintain governance quality. This tracks anti-discrimination law as applied in Roberts and Falcone and preserves the governance de — erence recognized in Dale — or exclusion tied to legitimate organizational — unction. Third, absence o — process proportionate to the governance mechanism. The governing body excluded a member without any process consistent with the mechanism’s own institutional character. For — ormal governance bodies — stock exchanges, pro — essional associations, trade organiza- tions with written bylaws — this means Silver’s requirements: notice, an opportunity to re- spond, and a stated basis — or the decision. Congress “cannot be thought to have sanctioned and protected sel — -regulative activity when carried out in a — undamentally un — air manner.”100

98 Louis Kaplow & Steven Shavell, Property Rules Versus Liability Rules: An Economic Analysis, 109 HARV. L. REV. 713, 725–40 (1996). 99 Guido Calabresi & A. Douglas Melamed, Property Rules, Liability Rules, and Inalienability: One View o —

the Cathedral, 85 HARV. L. REV. 1089, 1106–10 (1972). See also Henry E. Smith, Property as the Law o

Things, 125 HARV. L. REV. 1691, 1704–10 (2012) (arguing that property rules protect complex institutional arrangements against destabilizing piecemeal intervention). 100 Silver, 373 U.S. at 361. 2026-03-18] 27

But demanding

ormal written notice and adversarial hearings — rom an Ostrom-style com- mons community or a tight-knit trading network would be incoherent — it would impose procedural costs that destroy the very governance mechanism the — ramework aims to pro- tect. In in — ormal governance systems where monitoring and sanctions operate through repu- tation, graduated social pressure, and community knowledge, the relevant question is wheth- er the excluded member was blindsided without any prior signal or whether the exclusion

ollowed a pattern o — warnings, reduced cooperation, and social cues that the member’s con- duct was unacceptable. A — isher whose neighbors stop trading with him a — ter repeated over- harvesting has received process proportionate to the mechanism; a trader expelled overnight by a cartel with no prior indication o — noncompliance has not.101 The procedural — loor scales with institutional — ormality: — ormal governance bodies must provide — ormal process; in — ormal governance bodies must provide process recognizable within their own institutional logic. This three-part standard mirrors the business judgment rule’s structure — de — erence as the de — ault, with speci — ic triggers — or heightened scrutiny — and gives courts a de — ined role without destroying the club good. Courts do not review the substance o — governance deci- sions (whether the member actually committed the violation, whether the sanction was pro- portionate to the o —

ense, whether the governance body weighed the evidence correctly). Courts review three threshold questions: was the purpose governance-serving or anticompet- itive, were the criteria conduct-based or ascriptive, and was proportionate process provided? I — all three conditions are met, the governance decision stands. I — any is absent, the presump- tion is overcome and the court reviews the decision under the standard applicable to the speci — ic de — iciency — rule-o — -reason analysis — or anticompetitive purpose, anti-discrimination scrutiny — or ascriptive criteria, or procedural review — or process — ailures. The third prong’s sliding scale ensures that the standard protects Ostrom’s commons communities and Bern- stein’s trading networks as e —

ectively as it protects FINRA and the NYSE — requiring pro- cess that — its the institution rather than imposing a single procedural template that would privilege — ormal organizations over the in — ormal governance systems that generate some o —

the largest positive externalities. The degree o — de — erence within this — ramework is not uni — orm. It should track the quality o — the ostracism mechanism along three dimensions: accuracy in identi — ying rule-breakers (does the governance body investigate be — ore sanctioning?), proportionality in sanctioning them (are graduated sanctions available, or is permanent expulsion the only option?), and reliability in distinguishing governance-serving exclusion — rom anticompetitive or discrimina- tory exclusion (does the governance body have internal checks against abuse?). Higher- quality mechanisms warrant stronger de — erence. Lower-quality mechanisms — and non- replicable governance systems where exit is impossible — warrant weaker de — erence, because

101 See Auto. Elec. Serv. Corp. v. Ass’n o — Auto. A — termarket Distribs., 747 F. Supp. 1483, 1489 (E.D.N.Y. 1990) (“It is the proper policy o — the Court to re — rain — rom unnecessarily inter — ering in the internal a —

airs o — a private, not- — or-pro — it trade association, which is governed by by-laws. . . . Absent bad — aith or bias, this Court would not intervene in the dispute or review the determination o — AAAD to expel the plainti —

.”). The bad- — aith exception, not mere procedural irregularity, is the appropriate trigger. C — . Ostrom, Governing the Commons, supra note 25, at 90–102 (documenting how commons communities en — orce rules through graduated sanctions — verbal warnings, social disapproval, reduced cooperation, and ultimately exclusion — that constitute meaning — ul process within the community’s institutional — ramework even without — ormal written procedures). 28 “Private” Governance Is Actually a Club Good

the costs o

governance — ailure are concentrated on members who cannot escape and non- members who cannot substitute.102

Conclusion This Article has argued that governance is a club good — excludable through ostracism, partially rivalrous through congestion, voluntarily joined, and sel — - — inancing — and that this classi — ication produces three consequences the existing literature has not recognized. Gov- ernance generates positive externalities — or non-members who never participate in the gov- erning body. Courts that displace the ostracism mechanism trigger a compound wel — are loss in which both the internal club good and its external spillovers are destroyed. And de — erence doctrines — unction as implicit Pigouvian subsidies that reduce governance production costs, partially correcting the undersupply that positive-externality theory predicts. The — ramework has immediate implications — or three areas o — active doctrinal contesta- tion. First, post-Loper Light administrative law. The overruling o — Chevron de — erence removed the Pigouvian subsidy — or agency governance at a moment when the positive externalities o —

specialized regulation — environmental protection,

inancial stability, public health — are under political and judicial pressure — rom multiple directions. The compound wel — are loss

ramework predicts that governance quality will degrade and externalities will diminish as agencies shi — t resources — rom substantive regulation to de — ensive litigation. Whether this prediction holds is an empirical question, but the — ramework provides a theoretical basis — or evaluating the consequences that existing administrative law scholarship — — ocused on insti- tutional competence and democratic legitimacy — does not supply. Second, the ongoing debate over the Federal Arbitration Act’s scope. The — ramework provides a principled basis — or distinguishing commercial arbitration — where sophisticated parties voluntarily submit to governance systems that generate positive externalities — or the broader commercial community — — rom consumer arbitration, where the “voluntariness” o —

membership is contested and the externalities may

low in the opposite direction i — arbitra- tion suppresses claims that would otherwise produce precedent and deterrence. The Pigou- vian subsidy — ramework predicts that the FAA should be calibrated, not categorical: stronger de — erence where governance externalities are larger and membership is genuinely voluntary, weaker de — erence where externalities are smaller or negative and membership is — unctionally compelled. Third, antitrust treatment o — pro — essional association governance. Trade associations, standard-setting organizations, and pro — essional licensing bodies operate governance mecha- nisms that generate positive externalities — quality standards, ethical norms, consumer pro-

102 This variable-de — erence approach is already implicit in how courts treat di —

erent governance bodies. Delaware courts give stronger de — erence to board decisions by independent directors with no con — licts o —

interest than to decisions by interested directors. Federal courts give stronger de

erence to FINRA disciplinary proceedings, which include notice, hearing, and appellate review, than to in — ormal expulsion decisions by unincorporated associations. The — ramework makes the implicit calibration explicit and grounds it in club-goods economics. 2026-03-18] 29

tection — while simultaneously creating opportunities

or anticompetitive conduct. The proposed standard o —

ers courts a structure — or navigating this tension: a rebuttable pre- sumption o — validity — or governance decisions, overcome by showing anticompetitive pur- pose, discriminatory criteria, or absence o — procedural sa — eguards. This is not a novel doctri- nal invention; it is an economic justi — ication — or the structure that Northwest Wholesale and Sil- ver already imply. The Article’s central empirical limitation is that the compound wel — are loss is a theoreti- cal prediction, not a measured quantity. The external-dominates-internal — inding — that non- members lose more — rom governance degradation than members do — rests on a — ormal model and on the structural observation that non-members vastly outnumber members in every domain examined, not on direct wel — are measurement. Comparative institutional anal- ysis o —

ers a path — orward: comparing governance outcomes in jurisdictions with stronger versus weaker de — erence regimes, measuring — raud rates and transaction costs be — ore and a — - ter governance breakdowns, and tracking the spillover e —

ects o — Loper Light on regulated in- dustries. The club-goods — ramework generates testable predictions. Testing them is the next step. This Articled demonstrated that governance is a club good. Courts that displace it de- stroy value — or the people least able to protect themselves — the non-members who bene — it

rom governance spillovers without any voice in the decisions that determine whether those spillovers continue. Calibrated de — erence is not judicial abdication. It is the recognition that governance, like any positive-externality good, requires protection — rom the — orces that would cause it to be undersupplied. Appendix: Formal Model o — the Compound Wel — are Loss This Appendix presents a game-theoretic model extending Hirshlei — er and Rasmusen’s ostracism — ramework to incorporate court intervention and positive externalities to non- members. The model — ormalizes the compound wel — are loss described in Part II and con-

irms that external losses dominate internal losses under a wide range o — parameterizations.

Setup N members o — a governing body interact in a repeated game over an in — inite horizon. Each period, each active member chooses to Cooperate (comply with governance rules at cost c) or De — ect (violate rules — or private bene — it b, where b > c). De — ectors — ace ostracism: exclusion — rom the governing body — or T periods. Members discount — uture payo —

s at rate δ ∈ (0,1). Courts override ostracism decisions with probability p ∈ [0,1]. When a court reinstates a de — ector, the e —

ective expected punishment — alls — rom T periods to approximately T(1 − p) periods.

Governance Quality and Externalities Governance quality G is a concave — unction o — the cooperation rate x among active members: G(x) = x α · Q, where α ∈ (0,1) re — lects diminishing returns and Q is a scaling pa- rameter. Concavity captures the observation that the — irst de — ectors are disproportionately damaging — a single act o —

raud in the DDC undermines the reputation o — the entire sys- tem. The governing body produces two value streams: internal wel — are ( — or members) and ex- ternal wel — are ( — or non-members). Internal wel — are — or members is calculated as: Wint = x · [−c + G(x) · β] where β converts governance quality into per-member bene — it. External wel — are — or M non-members is calculated as: Wext = G(x) · γ · M, where γ is the marginal external bene — it per unit o — governance quality per non-member.

Incentive Compatibility A member cooperates i — the expected payo —

rom cooperation exceeds the expected payo —

rom de — ection — ollowed by ostracism. Cooperation requires: b − c < Σs=1T(1−p) δs · [c + G(x) · β] The le — t side is the one-period gain — rom de — ection. The right side is the discounted value o — bene — its — oregone during ostracism. As p increases (courts more likely to override), the 2026-03-18] 31

right side shrinks: the expected punishment

alls, de — ection becomes more attractive, and the cooperation rate x declines.

The Compound Wel

are Loss When p increases and x — alls, both wel — are components decline. Internal wel — are — alls be- cause — ewer members cooperate and governance quality G(x) declines. External wel — are — alls because W ext is proportional to G(x), which has declined. The ratio o — external to internal wel — are loss is: ΔWext / ΔWint ≈ (γ · M) / (β · x) Because M » N in every domain examined (downstream diamond buyers vastly out- number DDC members; all investors outnumber FINRA member — irms; the global popula- tion outnumbers any commons community), the external loss dominates the internal loss whenever γ/β is not extremely small — that is, whenever the marginal externality per non- member is not negligible relative to the marginal internal bene — it per member. This condition holds in all three domains: — raud reduction, market integrity, and environmental quality all generate per-non-member bene — its that are positive and nontrivial.

Calibration and Robustness The speci — ic ratio ΔWext / ΔWint is parameter-dependent. Under illustrative parameteriza- tions (N = 10 members, M = 50–500 non-members, δ = 0.80, b = 4.0, c = 1.5, T = 5 peri- ods, α = 0.7, γ/β = 0.05), the ratio ranges — rom approximately 50:1 to over 200:1. The direc- tional — inding — external losses dominate — is robust across all parameterizations where M

N and γ > 0. The speci

ic magnitudes should not be cited as empirical predictions; they depend on assumed values — or the discount — actor, externality elasticity, and cooperation- collapse — unction. What the model establishes is the structural result: the compound wel — are loss produces external harm that exceeds internal harm by a — actor that grows with the ratio o — non-members to members.

Optimal Override Rate The model also permits analysis o — the optimal court intervention probability p. Total wel — are Wtotal = Wint + Wext is maximized at an interior p that is positive but small. At p = 0 (blanket de — erence), governance operates without any external check on errors or abuse. At high p (plenary review), the ostracism threat lacks credibility and cooperation collapses. The optimum lies between these extremes: courts override a small — raction o — governance deci- sions, providing a check against pathological equilibria while preserving the credibility o — the ostracism mechanism. The model does not generate a speci — ic numerical prescription — or p* because the optimal rate depends on institutional parameters (discount — actor, externality elasticity, membership size) that vary across governance domains. What the model establish- es is the structural result: blanket de — erence and plenary review are both suboptimal, and cal- ibrated de — erence — a rebuttable presumption that most governance decisions stand, with judicial intervention triggered by speci — ic conditions — is the wel — are-maximizing policy.