June 19, 2025 — The Multilateral Instrument was an unusually ambitious piece of international tax engineering. The instrument, opened for signature in June 2017 and entered into force in July 2018, was designed to amend the bilateral tax treaty network at scale by allowing participating jurisdictions to incorporate selected provisions from the OECD's Base Erosion and Profit Shifting project into their existing bilateral treaties without requiring renegotiation of each treaty individually.

The conceptual challenge was substantial. The bilateral tax treaty network in 2017 comprised approximately 3,000 treaties, each with its own text and its own bilateral history. Renegotiating each treaty to incorporate BEPS-related amendments would have taken decades. The MLI offered an alternative: a single multilateral instrument that, once ratified by both parties to a bilateral treaty, would automatically modify that treaty in accordance with the parties' notifications.

Five years into the operational phase of the MLI, the patterns of adoption and the practical effects are now visible. This article reads where the MLI stands and what its impact has been.

The architecture of the MLI

The MLI's architecture allowed each signatory to make extensive choices about which provisions to adopt. The instrument contained a number of substantive provisions — addressing hybrid mismatches, treaty abuse, the dependent agent test, the corporate residence tie-breaker, mandatory binding arbitration, and others — and signatories could opt in or out of each. The architecture was deliberately flexible to maximise participation, with the trade-off that the resulting modifications to the bilateral treaty network would be uneven.

For a provision to take effect in a specific bilateral treaty, both parties to that treaty had to be MLI signatories, both had to notify the same provision as applicable, and neither could have made a reservation that would prevent the provision from operating in the relationship. The complexity of these matching requirements meant that the actual modifications to the network were less comprehensive than the headline number of signatories would suggest.

The adoption patterns

By 2025, approximately one hundred jurisdictions have signed the MLI, with the substantial majority having ratified it. The signatories include essentially every major OECD jurisdiction and a significant proportion of the Global Forum on Transparency and Exchange of Information members. The non-signatories include the United States, which is not a party for reasons related to its constitutional framework for treaty modification, and several jurisdictions that have indicated reservations about specific provisions.

The provisions most widely adopted are the principal purpose test, the corporate residence tie-breaker amendment, and the provisions on dual-resident entities. The PPT is incorporated in the substantial majority of bilateral relationships among MLI signatories, with the consequence that treaty shopping protections are now reasonably uniform across the affected network. The corporate tie-breaker amendment is widely adopted but with more variation, reflecting the different policy positions of jurisdictions on the appropriate corporate residence determination.

The provisions less widely adopted include mandatory binding arbitration for tax treaty disputes, which has been adopted by a smaller number of jurisdictions. The arbitration provision was always known to be politically sensitive, with several jurisdictions taking the position that bilateral disputes should be resolved through bilateral negotiation rather than mandatory arbitration. The smaller adoption produces the consequence that arbitration is available in a subset of bilateral relationships rather than across the entire MLI network.

The practical effects on cross-border investment

The MLI's practical effects on cross-border investment have been most visible in the application of the principal purpose test, addressed in a separate article. The PPT has produced a body of administrative practice and case law that has reshaped cross-border structuring. The MLI was the principal vehicle for the PPT's incorporation into the existing bilateral treaty network, and the impact of the PPT is, in significant respects, the impact of the MLI.

Beyond the PPT, the MLI's effects have included: the harmonisation of corporate tie-breaker rules in many bilateral relationships, with consequent effects on dual-resident corporate structures; the strengthening of dependent agent provisions, which has affected commissionnaire and similar arrangements; the introduction of provisions on hybrid mismatches, which interact with domestic anti-hybrid rules; and the introduction of mandatory binding arbitration in a subset of bilateral relationships, which has changed the dispute resolution landscape for those relationships.

The cumulative effect has been a partial harmonisation of bilateral treaty provisions across the MLI signatory network. Treaties between MLI signatories now contain a more consistent set of anti-abuse provisions than they did pre-MLI, with the consequence that cross-border structures must be analysed under broadly similar frameworks across the affected network. The harmonisation is partial rather than complete because of the optionality in the MLI architecture, but it is more comprehensive than would have been achievable through bilateral renegotiation in any comparable timeframe.

The non-signatory issue: the United States

The principal non-signatory of significance is the United States. The US position has been that the constitutional framework for US treaty modification, which requires Senate advice and consent for treaty amendments, is incompatible with the MLI's mechanism for automatic modification of bilateral treaties. The position has been criticised by some commentators but reflects a longstanding US approach to international agreements.

The consequence is that US bilateral treaties are not modified by the MLI. Treaties between the United States and other major jurisdictions — the United Kingdom, France, Germany, Japan, and others — retain their pre-MLI form unless separately amended through bilateral renegotiation. The US has, in some cases, incorporated comparable provisions into newly negotiated bilateral treaties, but the rate of new treaty negotiation has been slow and the result is a US bilateral treaty network that lags the post-MLI standards adopted by other major jurisdictions.

The mismatch creates analytical complications. A multinational group with US connections must analyse its position under the pre-MLI architecture for the US bilateral relationships and under the post-MLI architecture for non-US bilateral relationships. The two frameworks may produce different results in specific cases, with consequent complexity for cross-border tax compliance.

The political and procedural lessons

The MLI experience has produced lessons for future multilateral instruments addressing international tax. The principal lessons include the value of optionality in maximising participation, the importance of clear notification mechanisms in identifying which provisions apply to which bilateral relationships, and the limits of the multilateral approach when significant jurisdictions opt out.

The optionality in the MLI architecture allowed essentially every interested jurisdiction to participate by selecting the provisions that aligned with their policy preferences. The participation rate — approximately one hundred signatories — reflects this design choice. The trade-off, the partial harmonisation of provisions across the network, was the cost of the broad participation.

The notification mechanism, which requires each signatory to identify the bilateral treaties to which the MLI applies and the specific provisions to be incorporated, has produced a complex but workable framework. The OECD maintains a database that identifies the modifications applicable to each bilateral relationship, allowing practitioners to determine the post-MLI text of each affected treaty.

The opt-out by significant non-signatories — principally the United States — limits the MLI's reach. The framework's effects are concentrated in the bilateral relationships among signatories. The framework is, in this sense, a partial solution to the bilateral treaty modification problem rather than a complete one.

The interaction with subsequent developments

The MLI was not the end of post-BEPS treaty engineering. The Pillar Two framework, which entered force in 2024, operates alongside but separately from the MLI. The two frameworks address different policy concerns: Pillar Two addresses the level of effective taxation of multinational profits; the MLI addresses the application of bilateral treaty provisions to cross-border investment.

The interaction between the two has produced administrative interpretation in several jurisdictions. The Pillar Two top-up calculation depends in part on bilateral treaty provisions, and the MLI's modification of those provisions affects the calculation. The interaction has been resolved through OECD guidance and national administrative practice but has added complexity to the operational implementation of both frameworks.

Beyond Pillar Two, additional multilateral instruments have been developed for specific purposes. The Multilateral Convention to implement Pillar One Amount A was opened for signature on 3 October 2024 and would have used a similar multilateral approach if it had reached the threshold for entry into force. The Convention's failure to attract sufficient signatories — principally because of the absence of US support — has been a significant gap in the post-BEPS architecture, and the Pillar One Amount A framework has not become operational at the time of writing.

The trajectory through the second half of the 2020s

The MLI's trajectory through the second half of the 2020s is one of continued operation rather than expansion. The signatory base is approximately stable, with limited new accessions expected. The provisions adopted by existing signatories are unlikely to expand materially, with the principal optionality choices having been made by the signatories at ratification.

The case law and administrative practice on the MLI's substantive provisions — particularly the PPT — will continue to develop. The interpretive questions about how the MLI's modifications interact with the underlying bilateral treaty texts will continue to produce litigation and administrative interpretation. The framework is operationally stable but interpretively evolving.

For the practitioner working with the bilateral treaty network, the MLI is now a fixed component of the analytical framework. Each bilateral relationship must be analysed under the post-MLI text where applicable and under the pre-MLI text where not. The OECD's database and the national administrative guidance provide the tools for the analysis, but the analysis itself has become more complex than the pre-MLI alternative. The framework has, in this respect, achieved its purpose: the bilateral treaty network has been modified at scale to incorporate post-BEPS standards. The cost has been the analytical complexity of operating across a partially harmonised network.