The international tax architecture functions as a complex scaffolding constructed at the intersection of sovereign domestic law and bilateral public international law. At its core, the Double Taxation Avoidance Agreement (DTAA) represents a diplomatic compromise between two contracting states, each agreeing to self-limit their inherent sovereign power to tax in exchange for reciprocal benefits that theoretically foster cross-border trade and investment.1 The fundamental premise of these agreements has historically been the elimination of double taxation—a scenario where the same income is taxed in the hands of the same taxpayer by two different jurisdictions.3 However, the philosophical and practical orientation of these instruments has undergone a seismic shift over the last two decades, evolving from mechanisms of pure relief to instruments of regulation against fiscal evasion.
Tax treaties balance sovereign taxing rights with the prevention of abuse by establishing a framework that allocates jurisdiction between states while simultaneously incorporating specific mechanisms to ensure these allocations are not exploited for tax avoidance. This balance is achieved through the interaction of distributive rules, anti-abuse provisions, and the coordination of domestic and international law.4 Fundamentally, tax treaties do not create a new power to tax; rather, they modify the existing sovereign jurisdiction of the contracting states. They act as a "stencil" placed over domestic law: where the stencil covers domestic liability, taxation is restricted or eliminated, but where there are holes, domestic liability remains.6
This report provides an exhaustive analysis of this evolution, examining the inherent anti-avoidance principles that existed prior to the Base Erosion and Profit Shifting (BEPS) project, the legal tension between pacta sunt servanda and domestic General Anti-Avoidance Rules (GAAR), and the seminal Indian jurisprudence—Union of India v. Azadi Bachao Andolan and Vodafone International Holdings B.V. v. Union of India—that has defined the boundaries of legitimate tax planning versus abusive treaty shopping.
The "stencil" metaphor underscores a critical legal reality: tax treaties are permissive in their relief but restrictive in their application of sovereign power. If a state does not have the domestic authority to tax a specific item of income, the treaty cannot bestow that authority. Conversely, if domestic law imposes a tax, the treaty may limit the rate or the right to exercise that imposition.
Allocations of sovereign taxing rights are not transfers of power but limitations on exercise. When a Source State agrees to cap withholding taxes on dividends, it is not granting the Residence State the power to tax; it is merely agreeing not to exercise its own plenary power beyond a certain limit.7 This distinction is vital for understanding the interplay between domestic anti-avoidance rules and treaty obligations. If a treaty is merely a stencil restricting domestic law, then the removal of the treaty benefit (due to abuse) simply allows the full weight of domestic law to apply once more.
The primary method of balancing rights within the international tax system is the allocation of taxing power between the Source State (where income arises) and the Residence State (where the taxpayer resides). This allocation is achieved through "Distributive Rules," which classify income into different schedular categories and assign primary or exclusive taxing rights.
The architecture of distributive rules is built upon the principle of reciprocity. To facilitate trade and investment, source states often agree to restrict their sovereign right to tax certain income. This restriction is generally based on the assumption that the income will be taxed in the Residence State.7
This bilateral bargain creates a "closed system" where the tax base is shared. However, this system relies on the integrity of the tax base in both jurisdictions. If the Residence State does not tax the income (due to a territorial system or tax exemptions), and the Source State restricts its tax based on the treaty, the result is "double non-taxation"—a phenomenon that modern treaties explicitly seek to prevent.4
To protect fiscal sovereignty, treaties often include a "Saving Clause" (e.g., Article 1(3) of the OECD Model 2017 or Article 1(4) of the US Model). This provision preserves a state's right to tax its own residents (and in the case of the US, its citizens) as if the treaty did not exist.9
The Saving Clause ensures that treaty benefits generally do not restrict a country's ability to tax its own domestic base. It prevents the anomaly where a resident of a country uses a tax treaty to avoid tax in their own country of residence. For instance, without a Saving Clause, a US citizen living in a treaty partner country might claim that the treaty prevents the US from taxing their income, despite the US's citizenship-based taxation model. The Saving Clause effectively "turns off" the treaty restrictions for the state's own residents, with specific exceptions for benefits that are intended to apply even to residents (such as relief from double taxation or non-discrimination).10
|
Feature |
OECD Model Approach |
US Model Approach |
Implications |
|
Location |
Article 1(3) (2017 Update) |
Article 1(4) (Traditional) |
Standardizes the protection of the domestic tax base. |
|
Scope |
Residents |
Residents and Citizens |
Reflects the US policy of citizenship-based taxation (CBT). |
|
Exceptions |
Pension contributions, government service |
Social security, child support, pensions |
Ensures that social policy exemptions remain intact despite the clause. |
|
Purpose |
Prevent "round-tripping" and base erosion |
Enforce worldwide taxation |
acts as a "backstop" to the treaty's distributive rules. |
The interpretation of tax treaties is governed by the Vienna Convention on the Law of Treaties (VCLT), 1969. The cornerstone of treaty law is the principle of pacta sunt servanda.
Article 26 of the VCLT states: "Every treaty in force is binding upon the parties to it and must be performed in good faith".12 This principle creates a binding obligation on the contracting states to adhere to the agreed limitations on their taxing rights. In the context of taxation, pacta sunt servanda implies that a state cannot use its domestic law to unilaterally alter the bargain struck in the treaty.
This principle is the primary defense used by taxpayers against the application of domestic General Anti-Avoidance Rules (GAARs). The argument posits that if a taxpayer meets the specific requirements of the treaty (e.g., possesses a valid Tax Residency Certificate), denying the benefits based on a domestic "substance-over-form" doctrine constitutes a breach of the international obligation to perform the treaty in good faith.14
To reconcile the pacta sunt servanda obligation with the need to combat abuse, the OECD and legal scholars have developed the "No Conflict" view. This perspective asserts that domestic anti-abuse rules do not conflict with tax treaties because treaties are not intended to facilitate abuse.6
The "No Conflict" view essentially redefines "Good Faith" in Article 26 VCLT. It argues that claiming treaty benefits for artificial arrangements is an act of bad faith; therefore, denying those benefits is the only way to perform the treaty in good faith.16
A critical question in the historical analysis of international tax law is whether, prior to the BEPS project, the system was devoid of anti-avoidance rules. The research indicates that inherent anti-avoidance rules existed long before the explicit codification of the Principal Purpose Test (PPT).3
Civil law jurisdictions have long utilized the doctrine of fraus legis (fraud on the law) or abuse of rights to combat treaty abuse.
In common law jurisdictions, the "sham" doctrine served as a precursor to modern anti-avoidance rules. A sham is a transaction that is meant to deceive—documents are created to give the appearance of rights and obligations different from the actual legal rights and obligations the parties intend to create.20
However, the "sham" doctrine was often limited. In the seminal Indian case of Azadi Bachao Andolan, the Supreme Court held that as long as a transaction was "genuine" (i.e., the parties actually transferred the shares and paid the money), it could not be a sham, even if it was done solely to save tax.1 This narrow interpretation of "inherent" anti-avoidance rules necessitated the development of explicit treaty-based rules like the LOB and PPT.
India's engagement with international tax law provides one of the most intellectually rich and contentious case studies in the world. The jurisprudence highlights the conflict between "form" and "substance" and the evolution of the "sham" doctrine in the context of developing economies seeking Foreign Direct Investment (FDI).
The judgment in Azadi Bachao Andolan is a landmark decision that defined the Indian approach to "Treaty Shopping" for nearly a decade.
The Vodafone case represents the zenith of the "legal certainty" versus "economic substance" debate regarding indirect transfers.
Post-Script: The Indian Parliament subsequently amended the Income Tax Act retrospectively to tax indirect transfers, overturning the Vodafone verdict legislatively. This caused significant damage to India's reputation for tax certainty, leading to arbitration losses and eventually the repeal of the retrospective tax in 2021.23
Historically, the focus of treaties was the elimination of double taxation. However, modern treaties, particularly following the OECD's Base Erosion and Profit Shifting (BEPS) project, explicitly state that their purpose includes the prevention of tax evasion and avoidance, ensuring they do not create opportunities for double non-taxation.4 This shift has transformed the treaty from a "stencil" into a sophisticated "filter."
The PPT is the primary anti-abuse rule under the Multilateral Instrument (MLI). It serves as a general anti-abuse rule within treaties.
The LOB clause is a specific anti-abuse rule that uses objective criteria to determine whether a legal entity has a sufficient nexus (connection) to its state of residence to qualify for treaty benefits.29 It prevents "treaty shopping" by filtering out conduit companies.
The LOB sets forth a series of objective tests. A resident must satisfy one of these to be a "Qualified Person":
|
LOB Test Category |
Description |
Rationale |
|
Publicly Traded Test |
Shares are regularly traded on a recognized stock exchange in the Residence State. |
Public companies are presumed to have genuine substance and not be mere conduits. |
|
Ownership / Base Erosion Test |
>50% owned by qualified residents AND <50% of gross income paid to non-residents. |
Prevents third-country residents from owning a shell company; prevents income stripping via deductible payments. |
|
Active Trade or Business Test |
Income is derived in connection with an active business in the Residence State. |
Protects genuine commercial subsidiaries that might fail the ownership test (e.g., a US subsidiary of a UK parent). |
|
Derivative Benefits |
Owners would be entitled to equivalent benefits if they invested directly. |
If the ultimate owner is already entitled to treaty benefits, there is no "shopping" occurring. |
By filtering out conduit companies, the LOB protects the sovereign tax base of the source country from erosion.29
Included in Articles 10 (Dividends), 11 (Interest), and 12 (Royalties), the concept of "beneficial ownership" prevents the granting of treaty benefits to agents, nominees, or conduits who receive income on behalf of others.7
Treaties contain targeted rules to address specific abuses:
A critical aspect of the balance is the relationship between treaties and domestic anti-avoidance rules (GAARs).
The evolution of international tax law from the era of Azadi Bachao Andolan to the post-BEPS landscape represents a fundamental transformation in the nature of the tax treaty. Initially conceived as a "stencil"—a rigid instrument of relief that allocated rights based on form—the treaty has evolved into a "filter."
Tax treaties balance rights by creating a closed system: they restrict sovereign taxation only for genuine residents of the treaty partner who engage in non-abusive transactions. When a transaction fails the PPT, LOB, or beneficial ownership tests, the treaty restrictions are lifted, and the source state regains its full sovereign right to tax under domestic law.
To visualize this relationship, one might consider the tax treaty as a filter rather than a shield. It filters out double taxation for legitimate cross-border economic activity, allowing the flow of trade and investment. However, through mechanisms like the LOB and PPT, the filter becomes impermeable to artificial arrangements, catching them and subjecting them to the full weight of domestic sovereign taxation. The "holes" in the stencil are now guarded by smart gates, ensuring that only those who truly belong to the club of reciprocal benefits are allowed to pass through. The Indian experience with Vodafone and the subsequent retrospective amendments serves as a stark reminder that when the stencil covers too much, sovereignty will eventually tear through the paper to protect its base.
Disclaimer: This article is provided for informational purposes only and does not constitute legal, financial, or tax advice. Readers should consult with a professional before making any decisions based on the content of this article.