Which Alliances Are Gaining Weight

The question usually asked of an alliance is whether it makes sense. It is the wrong question, or at least a question whose answer predicts very little.

Arrangements that made excellent sense at inception dissolve. Arrangements assembled for reasons nobody would now defend persist for decades. The variable that separates them is not rationality but self-reinforcement: when every party expects every other party to maintain a commitment, the commitment holds — because acting on the expectation is individually sensible for each of them, whatever the merits of the original design.

That converts into something usable. Instead of asking whether an arrangement is optimal, ask whether it is gaining or losing expectational weight. Weight is not sentiment and it is not communiqué language. It is observable in published behaviour, and it moves before anything is announced.

The indicators

Five are visible from the open record.

Seniority of attendance over time. Not who attends once, but the trend. A body whose meetings were once attended by heads of government and are now attended by officials has lost weight, regardless of what the meetings produce. The reverse trend is a stronger signal still, because seniority is scarce and its allocation is a revealed preference.

Whether commitments are being institutionalised or re-announced. A commitment converted into a standing secretariat, a funded budget line or a binding instrument has been absorbed. A commitment restated annually in similar language has not — re-announcement is a substitute for implementation and a reliable indicator of a body that cannot compel its own members.

Whether new members join, and whether they are recruited or apply. Accession is the clearest single measure of expectational weight, because a state joining is making a costly bet that the arrangement will still matter when the costs fall due. The direction of the approach matters too: applicants indicate weight, recruits indicate its absence.

Whether members build dependent infrastructure. The strongest indicator and the slowest. When a member restructures its own institutions around an arrangement — creating domestic bodies to interface with it, training people into it, writing it into national procedure — it has made a commitment that is expensive to unwind. Infrastructure is expectation with money spent on it.

Whether obligations are invoked in unrelated disputes. When parties cite an arrangement's obligations in matters it was not designed for, they are treating it as a general source of authority. That is weight accumulating beyond the original remit.

A worked case, in one institutional field

The clearest illustration sits inside a single system, which removes the difficulty of comparing unlike things.

The WTO Appellate Body was established under Article 17 of the Dispute Settlement Understanding as a standing body of seven. In August 2017 the United States said it would not agree to launch selection processes to fill vacancies until its systemic concerns about dispute settlement were addressed. The body ran on three members through much of 2019. On 10 December 2019 two terms expired, leaving one member — below the three required to serve on an appeal under Article 17(1). By December 2023 members had failed to launch the selection processes on the seventy-second occasion, with one delegation speaking on behalf of 130 members.

Read that against the indicators. This is not an arrangement that lost weight because its members stopped valuing it: 130 of them said so, repeatedly, on the record. It lost function because a single sustained refusal was not answered, and the mechanism had no route around it.

Now the other half. Under Article 25 of the DSU, parties may agree to arbitration as an alternative or complement to panels and the Appellate Body. In April 2020 eighteen members used that provision to create a multi-party interim appeal arrangement. It now has sixty-one participants, with members continuing to join, and appeals have been resolved through it in practice.

That is weight migrating rather than disappearing — accession rising, an existing instrument repurposed as the vehicle, and participants building procedure around it. Neither observation predicts what either body will be in five years, and neither is an endorsement. They are descriptions of trajectory, which is all the test claims to produce.

Why rationality is the wrong lens

It is worth being precise about why the obvious question underperforms, because the answer is not that rationality is irrelevant.

An arrangement that no longer serves a member is under strain, and strain matters. But leaving is a separate decision from disliking, and it carries its own arithmetic: the cost of exit is paid immediately and visibly, while the benefit of exit is diffuse and arrives later. A member that would not join today will frequently not leave either, and will continue to perform the commitment while briefing that it is unsatisfactory. Read the briefing as a leading indicator of departure and you will be wrong for a decade.

The converse also holds. Members can be entirely committed and the arrangement can still fail, because commitment distributed across many parties is not the same thing as a mechanism that functions. That is the case the next section turns on, and it is the more instructive one, since it fails in a way that support cannot fix.

What the test is for

The payoff is timing. A positioning decision taken after a realignment is forced is taken at the worst available price: everyone else is making the same decision simultaneously, the options have narrowed, and the move is legible to every party as a response rather than a choice. The same decision taken while weight is still shifting is cheap, reversible and unremarkable.

That is the whole commercial and sovereign argument for watching expectational weight rather than communiqués. The indicators move first, they are visible in published material, and they are boring enough that almost nobody tracks them.

The limit is worth stating twice, because it is the framework's real vulnerability. Weight is a belief about what others believe, and beliefs of that kind are stable until a single defection goes unanswered. The Appellate Body case is instructive precisely there: the mechanism did not degrade gradually in proportion to support. It functioned, and then it did not, and the gap between those two states was a small number of unfilled positions.