Recent US court decisions concerning so-called "evergreen" standby letters of credit have exposed a drafting issue that could have consequences far beyond the two disputes involved.

Clauses intended to keep a standby alive through automatic annual extensions may not achieve that result if the language provides only for an extension of one year.

As highlighted by GTR Ventures, drawing on an excellent article by Jacob Atkins in Global Trade Review, two US courts have recently considered wording that the market had apparently regarded as creating continuing automatic renewals. The courts were less persuaded by how the clauses had been understood in practice and concentrated instead on what the instruments actually provided.

In Starr Indemnity & Liability Co. v. Midwest Mortgage Associates Corp., the standby provided that it would be automatically extended "for one (1) year" from its expiration date unless notice of non-renewal was given. The beneficiary maintained that this mechanism repeated annually, keeping the standby in force until the issuer eventually gave notice.

The court reached a different conclusion. In its interpretation, the clause provided for one automatic extension of one year, not an indefinite succession of annual extensions.

A separate Illinois case involving a standby originally issued in 1985 reached a broadly similar result. Once again, wording apparently intended to create an evergreen undertaking was found insufficient to keep the standby alive through successive extensions.

The scale of the potential exposure gives these decisions significance well beyond an interesting point of drafting.

Global Trade Review reports that US banks had more than US$340 billion of standby letters of credit outstanding as of March 2026, while variations of the language considered by the courts have apparently been used widely. It would be wrong to assume that every standby containing similar wording will necessarily receive the same judicial interpretation, but the decisions provide ample reason for issuers and beneficiaries to look again at their instruments.

None of this points to a weakness in the standby letter of credit itself. The problem lies in allowing a generally accepted commercial understanding to take the place of precise drafting. If the intention is that an undertaking should continue through repeated annual extensions until notice of non-renewal is given, the instrument needs to say precisely that.

Compare an undertaking that is automatically extended "for one year" with one that is automatically extended "for successive one-year periods". To those involved in the original transaction, the intended result may have seemed identical. Years later, when the instrument is being examined in the context of a dispute, that shared understanding may be of little assistance if it cannot be found in the wording of the standby.

BAFT has responded by encouraging banks to review their standard templates and outstanding instruments to establish whether their auto-extension provisions actually produce the intended result.

That seems a sensible exercise not only for issuers but also for beneficiaries relying upon long-standing standbys. An undertaking assumed to remain available is of limited comfort if its wording allows an argument that it expired years earlier.

There is another aspect of the discussion that deserves attention.

The GTR Ventures commentary connects the problem with the continuing digitisation of trade finance and suggests that greater use of standardised, machine-readable terms could help reduce ambiguity. Greater standardisation can certainly help, particularly where commonly used provisions can be expressed consistently rather than being recreated for individual transactions. Structured data also offers obvious advantages as trade finance moves towards greater automation.

It does not, however, remove the drafting problem. Technology can reproduce an unclear provision perfectly. Converting ambiguous language into structured data does not make the underlying obligation any clearer, and automating its processing may simply allow the ambiguity to travel further through the transaction without being questioned.

In fact, greater automation raises the standard we should expect from drafting. A human practitioner reading an awkwardly expressed clause may recognise the probable commercial intention and question the wording. A system can only work with the information and logic it has been given.

If an auto-extension provision is intended to operate repeatedly, its language needs to express that outcome without depending upon assumptions about what an "evergreen" clause would normally be expected to do.

For banks, applicants and beneficiaries, the US decisions provide a reason to re-visit wording that may have remained untouched for many years. The fact that a clause has routinely been described as "evergreen", or that everyone involved expected it to operate indefinitely, cannot substitute for what the undertaking actually says.

In the cases now attracting attention, the difference was remarkably small. "One year" and "successive one-year periods" may look like minor variations in drafting, but they can describe very different obligations.

That is reason enough to check which one the standby actually creates.

Source acknowledgement: This article draws upon the GTR Ventures commentary and Jacob Atkins' reporting for Global Trade Review, "US court rulings cast doubt on 'evergreen' standby letters of credit", published 12 August 2026. https://www.gtreview.com/news/americas/us-court-rulings-cast-doubt-on-evergreen-standby-letters-of-credit/

This article represents the views of the author and not necessarily those of ICC.