GAAR v/s MLI (PPT): A Redundant Backstop or a Double-Jeopardy Overlap?

[Krupali is a student at Gujarat National Law University.]

On 15 January 2026, the Apex Court in Tiger Global International II Holdings v. AAR (Tiger Global) denied treaty relief under the India-Mauritius DTAA, relying on the doctrine of substance over form. Eleven months before then, in SC Lowy PI (LUX) SARL v. ACIT (SC Lowy), the Income Tax Appellate Tribunal, Delhi delivered India’s first judicial application of the principal purpose test (PPT). This was followed by the Income Tax Act 2025 (ITA 2025) and the CBDT notification, amending the Rules of the General Anti-Avoidance Rules (GAAR), carving an exception to the pre-2017 investment shield. Three events in fourteen months on anti-avoidance, yet none addressed the issue of the simultaneous operation of GAAR (an umbrella clause that applies irrespective of the Specific Anti-Avoidance Rule or the Judicial Anti-Avoidance Rule) and treaty PPT (which falls within the wider ambit of GAAR). 

This blog is a discourse on this gap: when a cross-border transaction attracts GAAR as well as PPT under Article 7 of the Multilateral Convention to Implement Tax Treaty Related Measures to Prevent Base Erosion and Profit Shifting (BEPS MLI), which prevails, how their distinct onus probandi interact, and whether the absence of a harmonisation mechanism runs a double-remedy risk against the taxpayer.

Background

The Indian anti-avoidance trajectory is a reactive layering, lacking a responsive integration. 

First, the judicial phase. From J Chinnappa Reddy’s concurrence in McDowell & Co v. CTO towards substance-over-form to Azadi Bachao Andolan v. Union of India reversing the stance to uphold treaty shopping, calling for legislative action en-route to the limitation-of-benefits (LOB) clause. Vodafone International Holdings BV v. Union of India (2012) (Vodafone) affirmed tax planning as a legitimate structuring of affairs. Oscillation was the naked pattern, not grounded on principle. The parliament retaliated with a blunt retrospective amendment to Section 9 of the Income Tax Act 1961 to override the judicial failure. It was a cautionary tale ending at the Permanent Court of Arbitration in The Hague, with India’s defeat over a breach of fair and equitable treatment and a USD 1.2 billion award; the amendment was later repealed. Here was a unilateral domestic override that conflicted with a treaty obligation. 

Second, the statutory phase. GAAR was enforced from AY 2018-19, i.e., 1 April 2017, following the Shome Committee’s recommendations amid investor panic. 

Third, the treaty phase. India’s ratification of MLI in June 2019, whereby Article 6 rewrote treaty preambles disclaiming avoidance arrangements; Article 7 mirrors the OECD Model (2017) Article 29(9) minimum standard. GAAR and PPT are uncoordinated responses to anti-avoidance, layered atop one another without displacing or restricting the other. This is the structural problem history created, an overlap that subtly reproduces the very collision India paid to learn from Vodafone.

Current Regulatory Framework

ITA 2025 retains the substance of Chapter X-A of the Income Tax Act 1961 (ITA 1961), repudiating tax benefits bereft of commercial substance whose main purpose was to obtain tax benefit. Critically, under GAAR, any benefits, including treaty benefits, can be denied under Section 90(2A) of the ITA 1961. Whereas under PPT, benefits under the covered tax agreement, i.e. tax agreements that stand amended by BEPS MLI upon ratification by contracting states, can be denied even if one of the principal purposes of the transaction is tax benefit. What is contested is at the seam. Neither is there a statutory rule sequencing both these regimes, nor a safe-harbour schedule, unlike the UK GAAR model, which is a streamlined anti-abuse approach with a mandatory independent GAAR Advisory Panel opinion for HM Revenue and Customs actions. Even the CBDT Circular 7/2017, Notification Number 54/2026/F (grandfathering pre-2017 investments), and the Income Tax Rules 2026 notifications are administrative, reversible at will, thereby creating uncertainty about their statutory entitlement. In addition, a vacuum of Supreme Court rulings on the merits of GAAR leaves us with a framework without a precedential anchor. 

Analysis 

Burden, threshold, and what is favourable

GAAR and PPT are in inversion. GAAR has a higher substantive threshold, with a shift in burden-allocation upon a prima facie case of impermissible avoidance arrangement established. The Approving Panel palliates the Revenue’s powers. Conversely, PPT has a lower trigger but a greater onus to clarify that the benefit accords with the treaty’s object. Hence, the asymmetry: GAAR is procedurally taxpayer-friendly, screening most transactions, but substantively arduous to escape once caught. PPT is procedurally an amplitude casting a wider net with a lower threshold of principal purpose, but substantively easier. 

The sequencing vacuum

OECD underlines a general principle that domestic GAAR, mirroring its Article 29(9), bears a similarity to treaty circumstances for denying a tax benefit. The conflict arises because the remedies diverge. Under PPT, there is a limited effective response constructively defined, such as the denial of benefits or restoration of the contracting party’s rights. In contrast, GAAR provides various responses, such as recharacterisation and taxation on a different base, etc. One may theorise that, in the case where PPT sets off first, stripping treaty benefits, Revenue retains an incentive to still invoke GAAR for better tax collection. There lies an unresolved dilemma: Either PPT applies first, rendering GAAR otiose, or GAAR applies, followed by the interpretation of whether the treaty provision tempers GAAR. Neither the BEPS Action Plan 6 nor the ITA 2025 supplies a sequencing or a legal hierarchy. The order of operation is at the disposal of whichever officer reaches the file. 

Can Revenue stack both? Can a loss under one proceeding defeat the other?

There is a two-pronged argument. First, denial under PPT plus re-characterisation under GAAR; a cumulative civil remedial exposure in the absence of a coordinating rule. This does not create double jeopardy; it applies only to criminal prosecution. Second, noting the distinct statutory ingredients in both, a failure under one does not stop the other proceeding. Revenue’s defeat in PPT due to object-purpose concurrence inherent in the assessee’s transaction does not bar the GAAR proceeding, which calls for an inquiry into a different legal question, commercial substance.

Does the treaty ever block GAAR? 

This is where the fragility of Section 90(2A) of the ITA 1961 comes into the picture; the provision stipulates the applicability of Chapter X-A GAAR on the taxpayer against the treaty benefit despite not being beneficial to him. A transaction non-fulfilling of PPT ingredients is upheld by the treaty under pacta sunt servanda. Ideally, domestic GAAR should not trespass beyond the treaty permit as per paragraph 74 of the OECD Model Tax Commentary on Article 1. The fragility lies here; in Tiger Global, the court’s observations reinforce the subordination of treaty benefits and permits to GAAR. Therefore, there is no security to shield against subsequent GAAR challenge against a PPT victory; such a shield is illusory. Therefore, concurrent application in such marginal cases where STA ostensibly protects the taxpayer is apprehensive. 

Over-correction, or genuine redundancy?

Interestingly, the OECD Commentary indicates that the application of the domestic main-purpose test is in addition to Article 29(9), while simultaneously insisting that PPT “merely confirms it”. If PPT merely confirms the existing domestic GAAR, layering both is, at best, redundant or, at worst, two engines of denial with no governor. This is the over-correction: a treaty PPT and a treaty-overriding GAAR, both uncoordinated. 

The choice of instruments is itself an arbitrage. It is not treaty shopping by the assessee, nor LOB rule shopping within a treaty, but a meta-level forum-selection between anti-avoidance regimes. The mischief that GAAR and PPT exist to suppress, selecting the most favourable route, is thus reproduced at the level of the remedy itself. It is an anti-avoidance structure that allows abuse-style selection of its own remedies/instruments, thereby relocating the abuse rather than closing the gap. 

Two recent decisions, read together, prove that the imbalance is real and unadjudicated. In SC Lowy, the benefit was granted as the first limb of PPT; the principal purpose of the arrangement was not satisfied, following OECD examples D, G, H, and K. The second limb, a beneficial arrangement that contradicts the object of the treaty, was never examined, and GAAR was never invoked. Had, by factual matrix, the substance been thinner and the first limb met, would Revenue have stacked GAAR for recharacterisation? In Tiger Global, the court distinguished between an arrangement and an investment for the application of the relevant rules, invoked neither GAAR nor PPT, and instead applied a subjective JAAR. Yet, a clinical breakdown of the ingredients of PPT and GAAR assumes that both were prima facie satisfied: 

1. PPT (one of the principal purposes): US-resident control above the threshold of USD 250,000, no fund-pooling in Mauritius, and a single Flipkart investment (i.e., offshore locus of real control, and single asset concentration persistent throughout the exit) with double non-taxation as identified by the court, amounts to satisfying limb 2 of PPT; and 

2. GAAR’s impermissible avoidance arrangement (post-2017 exit beyond grandfathering and a thin commercial substance): single Flipkart holding supporting Section 96(1) chapeau main purpose test, AAR finding disguised loci of control in the US fulfilling the second limb of want of commercial substance under Section 97. 

This judicial circumvention led to market unpredictability, with the verdict reportedly making 500-plus foreign funds vulnerable to fresh scrutiny. 

Absence of a clear legal doctrine opens conflicting interests

The absence of a displacing or sequencing rule between GAAR and PPT has direct consequences for Revenue’s enforcement decisions and choices, multinational enterprises’ (MNEs) and foreign portfolio investors’ (FPIs) reliance on advance rulings and pricing agreements, and, at the macro level, the pricing of Indian cross-border investments due to structural uncertainty. 

Revenue and the CBDT, in the absence of a safe-harbour schedule like the UK and a sequencing rule, may treat GAAR as an enforcement backstop by selecting an instrument whose remedy best fits the case. MNEs and FPIs rely on advance pricing agreements (APA) with tax authorities, which are not binding on the GAAR Approving Panel. With no guarantee that an APA with arm’s length pricing forecloses GAAR scrutiny. In the absence of such immunity, PPT is a second layer of subjugation, not an alternative. Private equity / venture capital funds relying on tax on exit as a variable are now uncertain due to Tiger Global’s approach to pre-2017 investments, which are characterised as post-2017 arrangements stripping off grandfathering protections with underpinning retroactivity. Given that India-Mauritius accounts for a substantial share of inbound investment, it has a macro-variable effect. 

UK’s GAAR as a predictability test

The United Kingdom’s GAAR embodies a double-reasonableness test: the nature of the transaction is such that it cannot reasonably be regarded as a reasonable course of action. Additionally, its Advisory Panel opinions are published, acting as de facto safe-harbours and providing predictability. The lesson converges on transparency and predictability, both of which are preconditions for investor certainty. 

Conclusion

In conclusion, there is no instrument that formally governs when both GAAR and PPT are applied to a single arrangement, nor is there any logical structural mechanism for their coordination. Their burden of proof is inverse, remedies asymmetric, with no rule that chronologises them or estops one while the other is functional. This exposes the established double-remedy and unpredictability that Tiger Global declined to cure. 

For a change, three reforms could follow. One, a statutory rule stipulating non-stacking and mapping a sequence in the ITA 2025 Rules, whereby denial of benefit through PPT debars GAAR remedy on the same arrangement, and if confined to residual domestic GAAR, then PPT cannot be invoked. Two, elevate Circular 7/2017 to a statutory rule clarifying its interaction with PPT and publishing opinions of the Approving Panel akin to the UK model. Three, A statutory amendment to Section 144BA, ITA 1961: (i) stay in GAAR Approving Panel proceedings in case PPT process is lis pendente, or (ii) joint hearing for PPT proceedings and GAAR level legal questions, or (iii) estoppel of GAAR by prior PPT determination (as unilateral GAAR findings cannot subjugate treaty PPT adjudication).