A trade deal that ships half the cargo
On June 9, 2026, co-convenors Australia, Japan and Singapore told a WTO Joint Statement Initiative meeting that 73 members now support the concluded text of the Agreement on Electronic Commerce (ECA), with 67 members — covering roughly 70% of global trade — already moving to implement it through interim arrangements pending formal entry into force (WTO, June 9, 2026). The pact still needs 45 formal ratifications before it takes legal effect, a threshold the convenors are targeting for mid-2027.
That is real progress after nine years of negotiation — the initiative launched in December 2017 and expanded from 72 to 77 participants by January 2019 before narrowing to the 67 who adopted the interim pathway at MC14 in Yaoundé this March (WTO). But the June 9 announcement was notably candid about what the ECA does not do: it does not ban data localization requirements, and it does not guarantee free cross-border data flows. Those were the two provisions original US negotiators pushed hardest for — and the two the agreement ultimately punted.
What's actually in the text — and what isn't
The ECA runs 38 provisions across roughly two dozen pages and covers customs duties on electronic transmissions, e-signatures and e-contracts, consumer protection, unsolicited commercial messaging, and dispute settlement — genuinely useful plumbing for cross-border digital trade. But provisions on data transfers, localization of data storage, and source-code disclosure were left for "further negotiations," because attempts to reach agreement on basic data-governance principles foundered during talks. Nine of the 91 original negotiating participants declined to sign at all: Brazil, Colombia, El Salvador, Guatemala, Indonesia, Paraguay, Taiwan, Türkiye and the United States, each for different reasons — the US specifically objected that the security exception clause was too weak, while several Latin American holdouts opposed a permanent moratorium on digital-transmission tariffs (WITA).
The case for leaving data flows out — steelmanned
Governments that insisted on carving data governance out of the ECA have a defensible position, not just a protectionist reflex. Countries building domestic cloud and AI infrastructure — India, Indonesia, Nigeria among them — argue that binding free-flow commitments negotiated now would lock in today's data-center geography, permanently advantaging the US and China at the expense of everyone building capacity later. Financial and health regulators in multiple jurisdictions also have legitimate supervisory reasons to require certain records be stored or at least mirrored domestically, independent of any protectionist motive. And a WTO-wide free-flow rule, once ratified by 45+ members, would be far harder to adjust than a national data-protection statute as AI training practices and breach patterns evolve. Treating the data question as too consequential to rush is not an unreasonable instinct.
Why the omission still costs the world money — and clarity
The trouble is that the ECA's own economic case rests on removing exactly the friction its text leaves standing. A WTO staff working paper by Eddy Bekkers, Javier López-González and Roger So projects that full implementation across the WTO membership could add up to $8.7 trillion in cumulative global trade by 2040, with the largest proportional gains flowing to low- and lower-middle-income economies in Sub-Saharan Africa and South Asia — the same economies most likely to justify data localization on infant-industry grounds (Digital Trade4.EU summary of WTO ERSD-2025-06). Under the narrower 67-member interim scenario, the gain drops to roughly $2.4 trillion — most of it still contingent on frictionless movement of data across borders, the one thing the agreement does not secure.
For small exporters and cloud-dependent startups, the practical effect is that the ECA's customs and e-signature harmonization removes minor paperwork costs while the bigger cost — a patchwork of national data-localization mandates that forces duplicate infrastructure and fragments cloud pricing — remains untouched and, if anything, is spreading. A firm serving customers in a dozen ECA-implementing markets still needs to model a dozen separate data-residency regimes.
The proportionate path forward
Regulators are right that data governance deserves its own deliberate process rather than being bundled into a broader trade text under negotiating-deadline pressure. But "further negotiations" cannot become a permanent parking lot. The 45-ratification threshold gives WTO members roughly a year to build momentum before mid-2027; that same window should be used to open a dedicated, narrower-scope track on data transfers — modeled on interoperable frameworks like the Global CBPR Forum rather than a single binding free-flow rule — so the next JSI round has something concrete to ratify instead of another statement of intent. An e-commerce agreement that facilitates commerce but not the data commerce increasingly runs on is a foundation, not a finish line.