top of page
Search

The Panama Ports Dispute and the Problem of Spillover

  • Writer: Aişe Gül Akkoyun
    Aişe Gül Akkoyun
  • Apr 10
  • 4 min read

Updated: Aug 3

When an arbitration is only the part of the dispute we can measure.

Background

On 29 January 2026, Panama's Supreme Court handed down a ruling declaring unconstitutional the concession under which Panama Ports Company, a subsidiary of the Hong Kong-based CK Hutchison group, had operated the Balboa and Cristobal terminals at either end of the Panama Canal. The Court held that the arrangement, in place for nearly three decades, conferred exclusive privileges and tax exemptions incompatible with the constitution. The ruling was finalised the following month.

The commercial and legal consequences followed quickly. On 3 February 2026, Panama Ports Company commenced arbitration against the Republic of Panama under the arbitration clause of the concession contract, applying the ICC Rules. By late March, the claimant had supplemented its claims following what it characterised as an unlawful takeover of the terminals and company property, bringing the amount at stake above USD 2 billion. Panama, for its part, assumed control of the facilities by decree, installed A.P. Moller-Maersk and Mediterranean Shipping Company as interim operators, and signalled that a new concession would be awarded within roughly eighteen months.

What kind of dispute is this?

It is worth being precise about the legal architecture, because it is frequently blurred in reporting. This is contract-based commercial arbitration under a concession agreement, administered by the ICC, not an investment treaty claim before ICSID. The distinction matters for what can be argued, what remedies are available, how confidentiality operates, and what enforcement looks like at the end.

It also matters for research design. Treaty-based ISDS is comparatively well documented: registers, caseload statistics, and increasingly public procedural orders. Contract-based arbitration against States, particularly in infrastructure and concessions, is far less visible and correspondingly under-represented in quantitative work on investor-State disputes. A case of this size becomes visible mainly because a listed company must disclose it.

The part that does not fit in the case file

The feature of this dispute that seems most analytically interesting is that its effects did not stay inside the arbitration.

Within weeks of the annulment, terminal operations changed hands, two of the world's largest shipping lines took interim positions in a strategically located asset, and the matter attracted sustained diplomatic commentary from States that were not parties to the concession. None of that is a legal consequence of the ruling in any formal sense. All of it is a consequence of the dispute.

This raises a question that empirical work on arbitration tends to leave in the background: do investment and concession disputes produce measurable external effects, such as reallocation of commercial relationships, regulatory responses in third States, diplomatic signalling, or changes in the risk premium attached to comparable assets, that ought to be treated as part of the dispute rather than as context surrounding it?

The default unit of analysis in most quantitative arbitration research is the case: its parties, its treaty or contract, its duration, its outcome, its quantum. That unit is tractable precisely because it is bounded. But if the substantive consequences of a dispute are distributed across supply chains and diplomatic relationships rather than concentrated in an award, then the bounded unit systematically undercounts what happened.

Why chokepoints make this visible

Disputes over strategic chokepoints tend to expose the phenomenon more clearly than others, simply because the asset is not fungible. A canal terminal cannot be relocated, and its operator is difficult to replace quietly. When the legal status of such an asset becomes contested, the response is not confined to the parties: shipping lines reprice, governments comment, and competing operators position themselves before any tribunal has ruled on anything.

That visibility is methodologically useful. It suggests a way of testing spillover empirically rather than asserting it. Candidate indicators include:

  • Operator turnover: how often does a contested concession change hands, in whole or in part, before the arbitration concludes?

  • Timing: how does the interval between the triggering State measure and the reallocation of the asset compare with the expected duration of the proceedings?

  • Third-party positioning: do competitors acquire interests in the disputed asset while proceedings are pending, and on what terms?

  • Diplomatic response: is there official commentary from non-party States, and does it cluster around procedural milestones?

None of these requires access to confidential filings. All are observable from public sources, and all can be coded across cases.

The awkward implication

If spillovers are real and substantial, then two common ways of evaluating arbitration become harder to sustain.

The first is the assumption that the award is where the outcome lives. In a dispute where the asset has already been transferred and the commercial relationships already reconfigured, an award rendered years later determines compensation but not much else. The substantive allocation happened earlier, outside any tribunal.

The second is the assumption that States and investors bargain in the shadow of the expected award. Where a State can effect a transfer of control quickly and a claimant's realistic remedy is damages some years out, the shadow cast by the award may be considerably shorter than the theory assumes.

The open question

Should arbitration research continue to treat the case as the unit of analysis, clean, comparable and bounded, or should it attempt to measure what happens around the case, at the cost of far messier data and much weaker claims to comparability?

There is a real trade-off here rather than an obvious answer. The bounded approach produces reliable numbers about a narrow question. The broader approach addresses the question people actually care about, with data that may not support the weight placed on it. The Panama proceedings are a useful test case for whether the second approach can be made rigorous enough to be worth the loss of precision.

Related on DLS

Sources

The USD 2 billion figure reflects the claimant's supplemented claims as reported in late March 2026. As with all pending proceedings, the amount in issue is a moving figure rather than a settled one.

Comments

bottom of page