Institutions rarely stay the size of their founding charters. Samuel Huntington measured institutionalization by four qualities: adaptability, complexity, autonomy, and coherence. He argued that organizations survive by taking on functions beyond those they were created to perform. Douglass North added a corollary. Such growth is path dependent, because early choices establish the incentives, constituencies, and mental frames that make some expansions cheap and others nearly unthinkable.
The international trade regime offers one of the clearest twentieth-century demonstrations of both claims. In 1947 it began as a provisional tariff bargain among twenty-three governments. By 1995 it had become a permanent organization whose rules reached into patent law, financial services regulation, food safety standards, and the administrative procedures of domestic agencies. The institution meant to discipline the border came to discipline the state behind it.
This expansion is often told as mission creep, a story of drift. The historical record suggests something more structured. The regime grew along channels cut by its own founding compromises: its provisional status, its reliance on reciprocity, and its preference for legal technique over political deliberation. Each stage of growth solved a problem created by the previous one, and each solution generated new demands. That sequence matters for anyone weighing reform today. The regime's present difficulties over dispute settlement, regulatory sovereignty, and whom the rules serve are less a departure from its history than its logical culmination.
GATT Origins: A Provisional Bargain by Design
The General Agreement on Tariffs and Trade was never meant to stand alone. It was negotiated as an interim instrument, a tariff schedule with a set of commercial policy rules attached. It was meant to be absorbed into the International Trade Organization envisioned by the Havana Charter of 1948, which would have governed employment policy, commodity agreements, restrictive business practices, and investment. The United States Congress never ratified the Charter, and the Truman administration abandoned it in 1950. The stopgap became the system.
That accident of birth shaped everything that followed. GATT took effect through a Protocol of Provisional Application. The Protocol allowed contracting parties to keep existing legislation that conflicted with the agreement, under the so-called grandfather clause. For years GATT had no secretariat of its own and borrowed staff from the Interim Commission for the ITO. It had no members, only contracting parties. Legally, it was a contract among governments rather than an organization with independent standing.
The substantive rules carried equally deliberate limits. John Ruggie later called the settlement embedded liberalism: open multilateral trade was meant to coexist with interventionist domestic policy, not override it. Article XX preserved the right to protect health and conserve natural resources. Article XIX permitted emergency safeguards against import surges. Article XXI deferred almost entirely to national judgments about security. Agriculture was effectively carved out, a status confirmed by the broad waiver granted to the United States in 1955. Textiles were soon managed through a cartel of quotas that culminated in the Multi-Fibre Arrangement.
These constraints were not flaws in the design. They were the design. Governments emerging from depression and war wanted predictable tariffs without giving up the capacity to manage employment, stabilize farm incomes, or plan reconstruction. A thin agreement with generous exits secured participation precisely because it demanded so little of the state's internal workings. A regime that asked less could include more, and its legitimacy rested on its narrowness.
Yet the founding choices also planted the seeds of later growth. Reciprocal bargaining created exporter constituencies with a stake in further negotiating rounds. The most-favored-nation principle extended every concession to all parties, so each round raised the cost of leaving. And because there was no strong political body, the regime would develop under pressure through legal and technical channels rather than open constitutional amendment. Those early incentives would set the path along which the institution later expanded.
TakeawayAn institution's founding limits are often the price of its existence. The narrowness that secures initial consent tends to determine the channels through which it later grows.
Scope Expansion: From the Border to Behind It
The tariff rounds succeeded on their own terms. Industrial tariffs among the major economies fell steeply between the late 1940s and the end of the Tokyo Round in 1979, from levels commonly estimated in the double digits to low single digits. Success produced a classic institutional problem. An organization that has largely finished its original task must either fade in relevance or find new functions. As tariffs fell, the binding constraints on trade became non-tariff barriers: subsidies, technical standards, procurement preferences, and customs valuation.
The Tokyo Round responded with a set of codes covering subsidies, technical barriers, government procurement, and related matters. Each code bound only the parties that chose to sign it. This 'GATT à la carte' approach kept the voluntary character of the founding bargain while extending the regime into domestic regulation. It also fragmented the system and left most developing countries outside the newer rules. Powerful exporting states increasingly found that outcome unsatisfactory.
The Uruguay Round, launched at Punta del Este in 1986, changed the logic of expansion. Coalitions of firms in services, pharmaceuticals, software, and entertainment recast weak foreign patent and copyright protection as a trade distortion. In the United States they organized through bodies such as the Intellectual Property Committee. Washington backed the argument with unilateral pressure under Section 301 of the Trade Act of 1974. Developing countries accepted the General Agreement on Trade in Services (GATS) and the Agreement on Trade-Related Aspects of Intellectual Property Rights (TRIPS). In return, they secured the phase-out of textile quotas and a first framework for agricultural reform.
The decisive innovation was procedural: the single undertaking. Members of the new World Trade Organization had to accept nearly all agreements as a package, with only a few plurilateral exceptions such as the government procurement code. The agreements on sanitary and phytosanitary measures and on technical barriers to trade tied domestic standards to scientific justification and international benchmarks. The question was no longer whether a state discriminated at the border. It was whether its internal regulation was necessary, proportionate, and grounded in evidence.
Historical institutionalists call this pattern layering: new rules are added on top of old ones rather than replacing them. The original GATT text survived intact as GATT 1994, while new mandates accumulated around it. Each layer was justified in the original vocabulary of market access, non-discrimination, and trade distortion. The cumulative effect, however, was a regime claiming jurisdiction over subjects its founders had deliberately left to national politics. Continuity of language masked a profound change in the institution's constitutional character.
TakeawayWhen an institution exhausts its original task, its founding vocabulary becomes an elastic license. Once the concept of distortion is allowed to travel, almost any domestic policy can be redescribed as a barrier.
Dispute Resolution Power: The Judicialization of Trade
GATT's early approach to disputes reflected its diplomatic character. Complaints first went to the chair or to working parties that included the disputing governments themselves. From the 1950s, panels of independent experts became the norm, and a body of reasoned decisions slowly accumulated. But establishing a panel and adopting its report both required consensus, so the losing party could block the outcome. Robert Hudec showed that compliance was nonetheless surprisingly strong, sustained by reputation and reciprocity rather than compulsion.
That weakness was tolerable while the stakes were limited to tariffs. It became intolerable once obligations reached into services and intellectual property. Rights-holders in wealthy states would not accept substantive commitments without reliable enforcement. Developing countries, having conceded TRIPS, had their own reason to want limits on unilateral retaliation under tools such as Section 301. Both camps arrived at legalization, for opposite reasons.
The Dispute Settlement Understanding of 1994 inverted the decision rule. Under negative consensus, panels were established and reports adopted automatically unless every member, including the winning party, agreed otherwise. A standing Appellate Body of seven members was created to review legal interpretations. Retaliation now required authorization, and it carried the weight of law rather than informal pressure. Joseph Weiler described the shift as a move from the ethos of diplomats to the rule of lawyers. The result was, in effect, the busiest compulsory jurisdiction in international law.
Courts with automatic jurisdiction and open-ended mandates rarely stay modest. In US–Gasoline and US–Shrimp, the Appellate Body read Article XX's exceptions in light of general international law and evolving environmental norms. In EC–Hormones, it scrutinized the scientific basis of European food regulation. Its reasoning often showed genuine sensitivity to regulatory autonomy, yet in practice its rulings worked as binding precedent. In Huntington's terms, the institution gained autonomy faster than its political bodies gained the coherence to correct it, because rule-making still required consensus.
That imbalance proved decisive. Members unhappy with judicial interpretations could not legislate around them. The principal objector, the United States across several administrations, used the one lever available and blocked new appointments. In December 2019 the Appellate Body lost its quorum. Several members built the Multi-Party Interim Appeal Arbitration Arrangement as a workaround. The episode showed that delegated authority lasts only as long as the states that delegated it continue to accept the original bargain.
TakeawayJudicial authority grows fastest where legislative correction is hardest. An institution that can interpret but cannot be overruled invites its members to attack its existence rather than its rulings.
The trajectory of trade institutions illustrates a pattern that recurs across institutional history. Founding compromises that win consent through narrowness create the conditions for later expansion. Expansion achieved through technical and legal channels then outruns the political mechanisms that might legitimate it.
GATT's provisional status pushed development toward legal technique. Tariff success pushed it toward regulation behind the border. Regulatory reach pushed it toward compulsory adjudication. At each stage the regime grew more complex and more autonomous, meeting two of Huntington's criteria while straining the coherence and adaptability that depend on renewed political consent.
The lesson for reformers is neither that expansion was illegitimate nor that retrenchment is inevitable. It is that institutional authority, however well reasoned, rests on a political settlement that must be periodically renegotiated. The deeper question for any governance regime is whether it can build the capacity to revise its own bargain before its members decide to walk away from it.