Digital: EU-Singapore Digital Trade Agreement - making exceptions clear and a slightly different model for DTAs
The EU and Singapore concluded negotiations for their Digital Trade Agreement (DTA) back in July and the pre-scrubbed text of the agreement was released last month. This gives me the chance to note a couple points of interest. In this post I take a look at how EU practice on the data flow rules is increasingly detailed and specific in terms of what the obligations require and what the exceptions allow; and then I set out how this latest DTA takes a slightly different legal form from previous agreements.
For those interested in DTAs, I’d also recommend reading Simon Lester’s observations on the EU-Singapore DTA’s approach to data flows and exceptions.
More is more when it comes to clarifying data flow obligations and exceptions
First, the EU-Singapore DTA highlights again how detailed the EU’s approach to data flow rules are. The EU agreements tend to set out in detail exactly what measures the relevant data flow rules prohibit and then also explicitly clarify the range of exceptions that apply, with examples of precisely what measures are covered by those exceptions.
It also seems that EU agreements with more explicit obligations to allow data to flow, also include more detailed clarification of the policy space that is still preserved. You can see this, for example, in the EU-Japan agreement, which has explicit language requiring the parties to not prohibit “(e)…the transfer of information into [its territory]” or require approval “(f)…prior to the transfer of information to the territory of the other Party” (Article 8.81). This is then accompanied by additional clarifications specifying in detail that these obligations are subject to personal data and privacy exceptions:
The EU-Singapore DTA doesn’t go this far but it does clarify the meaning of “legitimate public policy objective” with an expanded list of examples of ‘legitimate’ objectives (now including reference to artificial intelligence and disinformation):1
FN1: For the purpose of this Article, “legitimate public policy objective” shall be interpreted in an objective manner and shall enable the pursuit of objectives such as to protect public security, public morals, or human, animal or plant life or health, to maintain public order, to protect other fundamental interests of society such as social cohesion, online safety, cybersecurity, safe and trustworthy artificial intelligence, or protecting against the dissemination of disinformation, or other comparable objectives of public interest, taking into account the evolving nature of digital technologies and related challenges.
It also explicitly states that the legitimate public policy exception “does not affect the interpretation of other exceptions…and their application to this Article and the right of the Party to invoke any of them”, and then adds a broad privacy exception allowing for any personal data protection measures that a Party “deems appropriate”.
This goes quite a bit further than, for example, the recently signed Australia-UAE CEPA Digital Trade Chapter which instead adopts a high-level, broad rule on data flows, and a similarly high-level exception for “legitimate public policy objectives” (in addition to the agreement’s general and security exceptions).
As Simon notes, the various exceptions within the data rules, layered on top of the general exceptions will likely be complex to interpret. This probably bolsters the case for greater detail to avoid overly narrow interpretations, particularly given the EU’s higher sensitivity around data and privacy.
The increased level of detail, often couched as ‘clarifications’, in the EU agreements also make clear that these data rules are not rigid impositions upon governments’ policy space requiring a data flow ‘free for all’. It also makes me wonder what the equivalent clarifications might look like for other jurisdictions where the concerns may relate more to putative security or ‘public order’ concerns rather than privacy and data protection. Even if not ultimately reflected in any text, it wouldn’t surprise me if these clarifications provide inspiration for others on how to they should justify their restrictive measures.
A slightly different model for digital trade agreements
Second, the EU-Singapore DTA provides a slightly different model for how to do digital trade agreements. To recap, the approaches for concluding DTAs that have been used to-date are:
Digital trade rules included within a free trade agreement (FTA), such as the digital trade and electronic commerce chapters in agreements such as the CPTPP or EU-New Zealand FTA, or the African Continental Free Area’s Digital Trade Protocol;
This also includes agreements that amend existing FTAs to include or improve digital trade rules,2 such as the Australia-Singapore Digital Economy Agreement, the United Kingdom-Ukraine Digital Trade Agreement, or Korea-Singapore Digital Partnership Agreement; and
Standalone agreements in the form of a entirely separate treaty, this includes the Digital Economy Partnership Agreement and the US-Japan Digital Trade Agreement.
The key difference between these is that it is much harder to argue that the exceptions for FTAs under the WTO apply to standalone agreements. As such, where the intention is to apply the digital trade rules preferentially, the most straightforward route is to include them in a broader FTA.
The EU-Singapore DTA has taken something of a middle ground between a full standalone agreement and amending the existing EU-Singapore FTA to insert these new rules:3
ARTICLE 1 - Objectives
The objective of this Agreement is to facilitate digital trade in goods and services between the Parties in accordance with the provisions of this Agreement. This Agreement shall be applied within the framework of the Partnership and Cooperation Agreement and shall, together with the Free Trade Agreement, form the free trade area consistent with Article XXIV of the GATT 1994 and Article V of the GATS.
ARTICLE 41 - Relations with Other Agreements
1. This Agreement shall be an integral part of the overall relations between the Union and its Member States, of the one part, and Singapore, of the other part, as governed by the Partnership and Cooperation Agreement and the Free Trade Agreement and shall form part of a common institutional framework. It constitutes a specific agreement giving effect to the trade provisions of the Partnership and Cooperation Agreement and, together with the Free Trade Agreement, forms the free trade area consistent with Article XXIV of the GATT 1994 and Article V of the GATS.
These provisions make clear that the DTA itself is a separate agreement to the FTA but still forms part of the overall relationship between the EU and Singapore on trade matters, including under the broader Partnership and Cooperation Agreement.4 This allows the DTA to rely on the services liberalisation provided under the EU-Singapore FTA to ensure that any preferential treatment provided is covered by Article V of GATS.
This seems largely like a change in form without a large impact on substance. Although, there are some minor implications from taking this approach:
A separate agreement can make it much easier to clearly establish the scope and nature of the obligations in the DTA. Trying to integrate new digital trade rules into an FTA that might be over a decade old can require careful consideration of definitions, scope provisions and implications for other chapters. By setting up an entirely new agreement that work can be minimised.
Depending on domestic treaty-making requirements, it may be easier to enter into a new treaty rather than go back and amend an existing one (particularly in the case of the EU given its particularly competencies vis-a-vis its member States).
As a separate agreement, the DTA itself can be terminated independently of the FTA (see Article 39). While it is possible to allow for termination/suspension of specific FTA Chapters or provisions, this is less common and potentially contentious.
It is always interesting to see how practice on digital trade rules is evolving, particularly the variations seen across jurisdictions, where differing policy interests and sensitivities are certainly making themselves evident in the texts.5
Although it is not clear why the list of legitimate policy objectives contained in the separate Article on the right to regulate does not seem to have been similarly expanded (these right to regulate provisions tend to be more akin to preambular language rather than granting substantive defences for State action):
ARTICLE 3 - Right to regulate
The Parties reaffirm their right to regulate within their territories to achieve legitimate policy objectives, such as the protection of public health, social services, public education, safety, environment or public morals, social or consumer protection, privacy and data protection, and the promotion and protection of cultural diversity.
These agreements are often sold and seen as standalone agreements, even though they are amendments to a larger treaty.
The EU has also proposed a similar approach in its DTA negotiations with South Korea.
Note the DTA does explicitly amend some aspects of the FTA - such as by terminating the operation of provisions that are now superseded by the DTA:
ARTICLE 41 - Relations with Other Agreements
2. The following Articles of the Free Trade Agreement shall cease to have effect and shall be replaced / superseded by the following Articles of this Agreement as provided:
a. Article 8.54 of the Free Trade Agreement is replaced / superseded by Article 5 (cross-border data flows) of this Agreement;
b. Article 8.57.3 of the Free Trade Agreement is replaced / superseded by Article 5 (cross-border data flows) of this Agreement;
c. Article 8.57.4 of the Free Trade Agreement is replaced /superseded by Article 6 (personal data protection) of this Agreement;
d. Article 8.58 of the Free Trade Agreement is replaced / superseded by Article 7 (customs duties on electronic transmission) of this Agreement;
e. Article 8.60 of the Free Trade Agreement is replaced / superseded by Article 10 (electronic authentication and electronic signatures) of this Agreement; and
f. Article 8.61 of the Free Trade Agreement is replaced / superseded by Article 14 (cooperation on digital trade issues) of this Agreement.
In both of these relations articles, I would think it slightly strange to use the word “replaced” in this context. Replaced reads to me as if the DTA article is cut and pasted into the place previously taken by the FTA article referred. Superseded makes it clearer that the function that was being performed by FTA article (which ceases to have effect) is now being performed by the relevant DTA article. While I can see how replaced can be read in a similar way, I don’t see what value it has to introduce a slight ambiguity here in relation to the amendment of the FTA and relationship between the two agreements.
See also Henry Gao’s work on the ‘three digital kingdoms’ looking at EU, American and Chinese approaches to data sovereignty.
I find it facinating that while the EU is known for its ROBUST privacy protection laws to the benefit of EU Citizens. It does not seem worried if its trade partner also takes its own citizens privacy protections in the same light. This should not be the case.
Interesting piece. Seems to me like the traditional FTA approach is still going to be best for services economies like the UK, Kenya, Argentina etc, particularly those such as the UK, Singapore, Ireland, Hong Kong etc that can be described as ‘coordinator economies’. The key advantage these countries have is that they have no opportunity costs for anything other than services and no goods industries they need to defend. At that point, they can offer to open up entirely in whatever industries their counterparts prefer in exchange for services concessions that will be useful for both and that they won’t need to worry about competition for. This should lead to them being the easiest countries in the world to do a trade deal with - one reason why I think Brexit may actually have been a good decision after all. https://backseatpolicycritic.substack.com/p/uk-comparative-advantage-brexit