Brexit and the Transatlantic Trouble of Counting Treaties

As pointed out by the Financial Times (FT), the UK’s withdrawal from the EU will require the renegotiation of more than 700 international agreements from which the UK currently benefits by virtue of its EU membership. Given the political and economic importance of transatlantic relations for both the UK and EU, the United States is arguably a good place to start when it comes to gaining a deeper understanding of the challenge at hand. As this post argues, before reaching the substantive questions surrounding the new agreements, even determining the number of treaties that may need to be replaced with new U.S.-UK ones is not a straightforward task.

In an address of November 28, 2017, Secretary of State Tillerson urged both sides to move the withdrawal process “forward swiftly and without unnecessary acrimony” and offered “an impartial hand of friendship to both parties”. Meanwhile, the Brexit negotiations are nearing a crucial point in mid-December, when it will be determined whether “sufficient progress” has been achieved for the two sides to start looking to the future—with each other and with strategic partners such as the U.S.

In determining the UK’s post-Brexit “special relationship” with the U.S., some preliminary discussions are already underway. However, the UK will be free to conduct fully-fledged negotiations only once it ceases to be an EU member. In anticipation of the many legal and political questions that these negotiations will raise, a preliminary—seemingly simple—matter would be to establish what the actual treaty relations between the U.S. and EU are. Three comprehensive and authoritative sources can be drawn upon to this end: The U.S. State Department’s Treaties in Force 2017, the EU’s Treaty Office Database, and the FT’s Brexit treaty renegotiation checklist. The only problem is, they do not match up. According to the State Department, there are 31 bilateral treaties in force between the EU and U.S., according to the EU’s Treaty Office, the number is 52, and according to the FT, it is 37. Hence, establishing the extent and content of legal relations affected by Brexit amounts in the first place to an empirical challenge.

In an effort to better understand this challenge, this post will first explain the reasons for (most of) these discrepancies, and subsequently offer its own assessment of the number of treaties. Before doing so, it should be stressed that this analysis focusses on bilateral international agreements only, i.e., agreements between the U.S. and the EU, either with or without the EU’s Members States alongside it. Agreements including additional parties would be categorized as multilateral agreements, of which there is also a significant number involving both the EU and U.S. and which raise additional difficulties, as illustrated recently in the dispute over the post-Brexit splitting of tariff rate quotas at the WTO. Moreover, the analysis focusses on treaties in force, thus excluding treaties pending ratification or those which are being provisionally applied (such as the 2007 Open Skies Agreement). As a final caveat, this post does not delve into any of the many administrative agreements concluded directly between U.S. and EU agencies (see for a useful overview the table compiled by Peter Chase in Daniel Hamilton and Jacques Pelkmans (eds.), Rule-Makers or Rule-Takers: Exploring the Transatlantic Trade and Investment Partnership (2015), pp. 55-60). What this post seeks to show is that even a single bilateral treaty relationship is challenging enough to grasp.

The divergence of the numbers of treaties is due to three main methodological differences: Timing, counting of extensions and amendments, and definition of a “treaty”. In terms of timing, the difference between the U.S. State Department and the EU Treaty Office is the following: Treaties in Force lists all treaties the U.S. considers to be in force at a particular point in time. In the current edition, this is January 1, 2017. Moreover, it includes only those treaties that “had not expired by their own terms, been denounced by the parties, replaced or superseded by other agreements, or otherwise definitely terminated” (p. i) by the critical date. By contrast, the EU Treaty Office provides a qualifier in the “advanced search” mode to show only treaties that have “entered into force”. This has two consequences. On the one hand, the U.S. list will not show agreements that have entered into force after January 1, 2017. Hence, the Agreement between the U.S. and EU on the protection of personal information relating to the prevention, investigation, detection, and prosecution of criminal offences, which entered into force on February 1, 2017, is absent from Treaties in Force 2017. On the other hand, the EU Treaty Office lists all agreements that entered into force at some point in the past, including those that are no longer in force. This concerns six agreements of the 52 listed by the EU, including the 2004 Agreement on the processing and transfer of PNR data by air carriers, which was superseded by a later agreement from 2011.

In terms of counting extensions and amendments of pre-existing agreements, the State Department opts for a more compact approach. It lists the main agreement, and mentions amendments and extensions as additional information for the same entry. The EU Treaty Office, by contrast, counts amendments and extensions as separate agreements. For example, the EU-U.S. Agreement for scientific and technological cooperation from 1997, which was extended and amended in 2009, is counted as one by the Americans and as two by the Europeans. From the point of view of the international law of treaties (see Art. 39 VCLT), the latter approach is technically correct. However, it terms of drawing up lists for post-Brexit renegotiation purposes, there is something to be said for the U.S. approach of counting “consolidated” versions of agreements as currently in force.

Thirdly, the most important difference in terms of numbers relates to counting “softer” agreements, such as exchanges of letters and memoranda of understanding. However, this cannot be explained by a more generous versus a more restrictive interpretation of what counts as a treaty. Treaties in Force notes that it “uses the term ‘treaty’ in the generic sense as defined in the Vienna Convention on the Law of Treaties”, rather than “as a matter of U.S. constitutional law” (p. i). Hence, executive and executive-congressional agreements are not excluded. Beyond that, it is not clear which criteria are applied. For instance, the EU lists a 2005 Exchange of letters relating to the method of calculation of applied duties for husked rice, while the U.S. does not. By contrast, the U.S. includes a 2009 Memorandum of understanding regarding the importation of beef from animals not treated with certain growth-promoting hormones, while the EU does not. Each side includes about half a dozen of such “soft” agreements in its list that the other does not, with no legal reason readily apparent.

Regarding the FT’s database and its listing of 37 U.S.-EU bilateral agreements, in addition to the issues mentioned above, additional factors are at play. While excluding expired and superseded treaties, it also excludes those that the FT considers of “little or no relevance to the UK after Brexit”, but includes ten EU implementing decisions/regulations. The authors justify this by noting that these are “EU ‘equivalence’ decisions on financial services, which provide access rights to third countries” and that “[t]rade partners would likely take them as a starting point in financial services discussions with the UK after Brexit.” While they are right in pointing out their relevance, such EU decisions and regulations are definitely not international agreements and hence should not be counted as such.

What, then, is the correct member of treaties currently in force between the EU and U.S.? Looking at the above issues, it would be foolish to even pretend here to have the undisputable number. But it can be better approximated. Taking into account only bilateral agreements in their “consolidated” versions, which are currently in force, and which despite their sometimes “soft” format at least one side deems “hard” enough to include in its list, this yields a number of around 50. Just to be clear: This does not mean the EU list was “more correct” because 50 is closer to its 52. The result of 50 reached here excludes a number of agreements no longer in force, and includes a number of “soft” agreements listed by the U.S.

In closing, there are two larger lessons here than discovering that the U.S. thinks international engagements on beef and cheese are worth listing while the Europeans prefer to put wine, husked rice, and bananas in their database. First, while the U.S. only needs to go through this Brexit-renegotiation exercise once, the UK will have to do it with each country that has treaties with the EU—168 according to the FT. Second, the U.S. and EU have comparatively good resources to help figure out what the status and extent of their treaty relations is. However, as this post has endeavoured to show, not even in this case this was a straightforward exercise, and it may not get easier for the other 167.