The effectiveness of the MLI in amending the bilateral tax treaty network

By Mees Vergouwen, Dirk Broekhuijsen, and Judith Reijnen


Having been adopted roughly six years ago on the 24th of November 2016, the MLI has been described as a “ground-breaking” treaty aimed at implementing the treaty-related changes of the BEPS Action Plan in the tax treaty network. In our soon-to-be published article, we studied the effectiveness of the MLI by empirically analysing the (number of) parties to the MLI and the bilateral tax treaties whose application has been modified by the MLI on 31 December 2022. Based on this empirical study, we seek to answer the question whether the MLI has achieved its purpose to swiftly implement (part of) the BEPS Project in terms of parties and bilateral tax treaties covered.

1,231 of tax treaties covered: a reasonable outcome

Measured on the 31st of December 2022, 1,231 bilateral tax treaties qualify as “Covered Tax Agreements” within the meaning of art. 2(1) MLI. In light of the total tax treaty network (+- 3,500 tax treaties in force), it follows that the MLI has affected approximately 35% of the total tax treaty network. Additionally, considering that the OECD has stated that the MLI “will transpose results from the OECD/G20 Base Erosion and Profit Shifting Project (BEPS) into more than 2,000 tax treaties”, the MLI has affected approximately 62% of the minimum number of treaties that it is intended to affect within a timeframe of six years.

The MLI is quicker than the bilateral negotiation alternative

In a previous blogpost, we concluded that it takes – on average – 18.56 years before a tax treaty concluded by a founding member county of the OECD is amended or updated by bilateral means. We also concluded that only +/- 39% of such tax treaties have been amended by bilateral means in their lifetime (i.e., including treaty relationships over 80 years old); as such, most bilateral tax treaties concluded by the OECD founding member states have not been amended in their lifetime. Based on a comparison of this outcome with the effect of the MLI on the bilateral tax treaties of OECD founding members, it is concluded that the MLI has been able to modify approximately the same number of bilateral tax treaties (38% of the tax treaties concluded by the OECD founding states) within a relatively short time frame, i.e. +/- 6 years. As such, the MLI has modified approximately the same number of bilateral tax treaties as have been changed by the bilateral renegotiation process in a relatively short timeframe. From the perspective of implementation effectiveness, the MLI can therefore be described as a success.

Participation: EU and OECD, but where is the rest of the world?

Although 100 states have signed the MLI, the mere signing is not sufficient for the MLI to take effect on tax treaties of a signatory state. Ratification is required (see the MLI’s definition of CTA in article 2(1)). Taking this into account, we conclude that the number of ratifying states would be the most accurate measurement for the MLI’s success. On the 31st of December 2022, 79 states participating in the MLI have deposited an instrument of ratification, making the MLI potentially effective on their bilateral tax treaties. All in all, we consider this a reasonable outcome (but note that the MLI’s ultimate effect on such a state’s tax treaties is dependent on the position of each of the state’s tax treaty partners under the MLI).

The number of 79 ratifying states does, however, not give any answers as to who is participating. The background of who – and more importantly: who did not – ratify the MLI, provides interesting results. The table below summarises the results of our categorisation by continent and by World Bank income category:

  # states in continent Signatories % of states in continent Ratifiers % of states in continent
Europe 44 44 100 42 95
Asia 48 24 50 19 40
South America 12 6 50 3 25
North America 23 7 30 5 22
Oceania 14 4 29 2 14
Africa 54 15 28 8 15
Total 195 100 51 79 41

Table 1 : signatory and ratification states, by continent (list of states have been derived from https://www.worldometers.info/geography/how-many-countries-are-there-in-the-world/)

 

Category  Signatories % of total signatories Ratifiers % of total ratifiers
Developed High income 55 56.1% 53 67.1%
Upper middle income 25 25.5% 14 17.7%
Developing Lower middle income 17 17.3% 9 11.4%
Low income 1 1.0% 1 1.3%

Table 2: signatory and ratifying states, by World Bank income category as applied to the 100 signatories of the MLI[1]

To these tables, we would like to add that out of the 38 OECD states, 37 have currently ratified the MLI.

What these figures show, is that the MLI is predominantly signed and ratified by developed EU and OECD member states. In our view, it seems unlikely that the OECD has tried to discourage signature and ratification by non-OECD (developing) states. Nevertheless, these figures may indicate that non-EU and non-OECD countries are less willing to adhere to international tax law making at the level of the OECD (see also the recent UN resolution on the promotion of inclusive and effective international tax cooperation in this respect).

The OECD overreports on the MLI’s success

As indicated above, we have found that there were, on 31 December 2022, 1,231 CTA’s. The OECD, by contrast, stated on 6 October 2022 that the MLI “covers around 1850 bilateral tax treaties”. This number of CTAs consisted, on 6 October 2022, of 910 bilateral tax treaties whose application was modified by the MLI and of 930 treaties that had not yet been modified by the MLI. In our view, the reference to 1,850 bilateral tax treaties being ” covered” by the MLI would seem misleading in light of the gap of approximately 600 CTA’s between the number of CTAs on 31 December 2022 (1,231) and the number of the OECD (approx. 1,850). It can be described as misleading because the OECD adopts a wider definition of the term Covered Tax Agreement than follows from article 2(1)(a) MLI. Based on article 2(1)(a) MLI, a tax treaty is designated as “Covered Tax Agreement” (CTA) when: (i) both states to such a tax treaty have deposited their instrument of ratification, acceptance or approval (Instrument of Ratification) of the MLI; while (ii) both states have also listed such bilateral tax treaty as a CTA in their Instrument of Ratification (or in a later notification made after becoming a party). The OECD’s number of approx. 1,850 also takes into account those tax treaties with respect to which only one, or none, of the contracting states have deposited their Instrument of Ratification. In order to qualify as a CTA, the OECD looks at the tax treaties listed as CTA’s in the instrument as deposited upon signing, as well as in the Instrument of Ratification. The OECD thus takes into account “provisional” CTA’s which may, in time, qualify as “real” CTA’s within the meaning of article 2(1)(a) MLI. Whether a “provisional” CTA will (on short notice) become a “real” CTA is, however, far from certain. States (i.e., state parliaments) may, in between the phase of signature and ratification, change their minds as to their willingness to participate. As such, the number of “real” CTA’s seems a better reflection of the MLI’s success than the number used by the OECD.

Conclusion

For the OECD and EU countries, the MLI can be described as a grand success. Many of their tax treaties have been amended within a reasonable time frame. The MLI, in fact, has amended approximately the same number of tax treaties as were amended in a bilateral fashion since the end of WW2. However, as regards non-OECD/EU states, the current state of participation seems to be somewhat less of a success. It points at classic legitimacy deficiencies so often pointed at by international tax scholars. Finally, although a success, we do recommend that the OECD reports on the effectiveness (in terms of CTAs) of the MLI based on the MLI’s definition of a CTA. This is undoubtedly necessary for inspiring further steps for multilateralism in international tax law.


[1] It is noted that the table does not include Jersey and Guernsey. The reason for excluding these two states is that they are absent in the overview of GDP per country (see https://www.worldometers.info/gdp/gdp-by-country/).

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.