Soft Law, Hard Liability: What the UN Tax Convention Should Say About Its Own Future

Articles 14 to 20 of the draft UN Framework Convention on International Tax Cooperation are still headings without text. They are provisions that will govern the Conference of State Parties and its subsidiary bodies, and they are reserved for a later stage of drafting. When negotiators reconvene in New York this August for the fifth session, the debate will again centre on the move from OECD-led standard-setting to a more representative process. The blank articles, however, raise a different question. Suppose the future Conference issues guidance that changes how a tax obligation is calculated. Would it bind states and guide their courts, or reach taxpayers for periods already closed? The Convention must provide that answer. 

The UN process has been debated mainly in terms of representation. Supporters present it as a correction to a century of rulemaking dominated by wealthy states, a dynamic Ashrita Prasad Kotha traced on this blog in 2024; critics predict duplication of the OECD’s work and fragmentation. Elliott Ash and Omri Marian’s empirical work has shown how deeply the OECD’s formally non-binding models shape the world’s roughly 3,000 bilateral tax treaties. The Convention is meant to change who holds that pen. Commentators have begun to ask what powers its future Conference should possess. The separate question addressed here is what legal effect the Convention should eventually assign to decisions, guidance and commentaries issued under it.

The Next Drafting Question

The Co-Leads' draft Framework Convention template of 22 January 2026 is an intermediate negotiating document. It identifies possible commitments on the fair allocation of taxing rights, illicit financial flows, high-net-worth individuals, and mutual administrative assistance. Its institutional provisions have been reserved for a later drafting stage. The articles on review and verification, relation with other agreements and domestic law, the Conference of State Parties, subsidiary bodies and amendment remain as headings marked ‘[To come]’. How those provisions are filled will determine how the Convention develops after ratification. Two design choices are already visible. Article 21(4) provides that a state party is not bound by a protocol unless it separately becomes a party to it. The articles on information exchange and assistance also contemplate extension of the cooperative regime through protocols or undefined further instruments. The terms of reference adopted by the General Assembly commit the Committee to deliver two early protocols alongside the Convention; the current workstreams cover the taxation of cross-border services and dispute prevention and resolution. Beyond treaty text, the architecture may therefore generate Conference decisions, subsidiary-body recommendations, technical guidance, commentaries and review findings.

When drafted, these provisions will determine which later instruments carry legal authority, who may issue them and how disagreements over their meaning are resolved. Article 31(3)(a) and (b) of the Vienna Convention on the Law of Treaties require interpreters to take account of any subsequent agreement regarding interpretation and any subsequent practice that establishes the parties’ agreement. The International Law Commission's conclusions on the topic state that the legal effect of a decision of a Conference of States Parties depends primarily on the treaty and its rules of procedure. Even a decision adopted by consensus becomes an agreed interpretation only if it expresses the parties’ agreement in substance. The institutional provisions now being prepared must therefore establish when a later decision may authoritatively shape the Convention’s meaning.

Pillar Two Shows What is at Stake

Pillar Two shows how international guidance can reshape a tax regime without a formal amendment. The OECD describes the GloBE rules as a 'common approach': jurisdictions need not adopt them, but those that do are expected to implement them consistently. In January 2026, the Inclusive Framework agreed a new side-by-side package under which groups headquartered in qualifying jurisdictions receive protection from parts of the GloBE regime. The OECD’s April 2026 FAQ identified the United States as the only jurisdiction then recognised as qualifying, and the guidance will be incorporated into the Commentary to the Model Rules. The package therefore changed the agreed operation of GloBE through guidance rather than treaty amendment or ratification, although any liability imposed on taxpayers must still rest on domestic law.

The UN negotiations now face a comparable design choice. Article 21(4) requires states to consent separately to protocols. Articles 9 and 10, however, contemplate extending cooperation through ‘other instruments’. The provisions governing the Conference and subsidiary bodies must eventually determine what those institutions may produce and what legal effect their output will have. Whether the resulting instrument can create obligations, settle the Convention’s meaning or merely offer technical guidance is the open question.

From International Guidance to Domestic Liability

A recent post on this blog traced how voluntary international standards can harden into domestic regulation. But international obligation and domestic tax liability are not the same thing. A state may be bound by a protocol internationally while its revenue authority remains unable to invoke it against a taxpayer at home. The difficulty is greater when guidance comes years later and is used to reinterpret existing legislation or change the tax treatment of an earlier period. Whether that is permissible turns on domestic law. The institutional provisions should therefore address the authority and temporal reach of later instruments before those questions reach domestic courts.

Writing the Hierarchy into the Convention

Leaving institutional provisions blank has a standard defence. Framework conventions often defer design because ambiguity lets states agree, and the regime is expected to fill out through practice. The UNFCCC, for example, never settled the legal effect of its Conference’s decisions, and climate lawyers have debated their status for decades without the regime failing. The deferral is cheaper in climate law because Conference decisions are addressed to states and implemented by governments. A decision of a tax Conference on how an obligation is calculated will be invoked by revenue authorities against companies and tested in domestic courts. The cost of an unsettled hierarchy falls on taxpayers and judges who were never at the table. 

The Convention should set out a clear hierarchy for the instruments issued under it. Its provisions would bind every state party, whereas a protocol would bind only those states that separately join it. Conference decisions could create obligations only where a specific provision grants that power and the prescribed voting procedure is followed. Where a decision is intended to record an agreed interpretation under Article 31(3)(a), it should identify both the provision being interpreted and the parties that accepted the interpretation. Guidance and commentary would sit lower in the hierarchy. They could explain how the rules work, but not amend the Convention or, by themselves, increase the tax payable under domestic law. Review findings would have the narrower function of recording compliance without imposing new duties. To make the system transparent, each instrument should state when it was adopted, when it takes effect and how it has since been amended. New obligations should apply only to future tax periods, and guidance published later should not increase liability for a period that had already ended.

Moving international tax rulemaking from the OECD to the UN broadens participation. But wider participation is not enough. States, courts and taxpayers must also be able to tell where each rule comes from and how far it reaches. In the 22 January template, Articles 14-20 are still blank. That is where the hierarchy should be written.

Nishtha Mahapatra is a Graduate of the National Law University Odisha and a Research Assistant at the High Court of Orissa, India.