India’s Anti-Avoidance Framework After Tiger Global: Why GAAR and JAAR Should not be Invoked Together
Introduction
In a landmark decision, the Supreme Court in Tiger Global allowed the statutory General Anti-Avoidance Rule (“GAAR”) and the Judicial Anti-Avoidance Rule (“JAAR”) to be invoked simultaneously and “operate in parallel” to deny treaty benefits where a transaction is found to lack “genuine commercial substance”. This holding raises a fundamental question: if the Revenue can invoke GAAR in parallel with judicial JAAR, then can every procedural safeguard built into GAAR be circumvented through the choice of JAAR as the preferred instrument?
This article argues that the simultaneous invocation of GAAR and JAAR is (i) doctrinally unsound and (ii) inconsistent with the legislative intent underlying the safeguards that Parliament deliberately embedded in Chapter X-A of the Income Tax Act, 1961 (“the Act”). It further argues that even if JAAR survives Tiger Global as a parallel doctrine, the decision’s own internal logic and its ambiguities compel a restrictive reading that confines JAAR to its traditional sham-only domain.
This article proceeds as follows: Part II traces the judicial evolution of JAAR from McDowell to Tiger Global, identifying the doctrinal expansion at each stage. Part III maps the statutory architecture of GAAR under Chapter X-A of the Act with particular attention to the procedural and substantive safeguards that distinguish it from JAAR. Part IV develops the case against simultaneous invocation across multiple dimensions. Part V argues for a restrictive reading of Tiger Global’s JAAR endorsement. Part VI concludes.
JAAR: Before and After Tiger Global
A. The Doctrinal Lineage
Indian JAAR evolved through four landmark decisions, each recalibrating the threshold for denying tax benefits on substance-over-form grounds.
McDowell laid the foundation. The majority held that “[c]olourable devices cannot be part of tax planning,” but confined this to dubious methods and subterfuges.[1] Justice Chinnappa Reddy’s concurrence went further in two respects. First, drawing on Ramsay, [2] he endorsed a purposive approach and held that the relevant inquiry was neither whether the statute should be construed literally or liberally, nor whether the transaction was genuine and not expressly prohibited by law. Rather, the court must ask whether the arrangement was a device to avoid tax and whether it was one to which the judicial process ought to lend its approval. Second, he rejected the continued relevance of the Westminster principle, under which taxpayers were entitled to arrange their affairs so as to minimise tax liability and courts were required to respect the legal form of such arrangements. Although these observations were made in a concurring opinion, they came to exert considerable influence on subsequent tax administration and adjudication, creating uncertainty about the precise scope of the anti-avoidance principle recognised in McDowell.
This confusion persisted until Azadi Bachao Andolan resolved it or at least attempted to. The Court disaggregated the McDowell opinions, holding that Reddy J.’s broader anti-avoidance dictum was an “extreme view” which militates against the observations of the majority. The Westminster principle received “judicial benediction”. JAAR was restricted to colourable devices. Treaty shopping was held permissible absent statutory prohibition. This represented JAAR at its narrowest.
Vodafone consolidated the doctrine by first naming it. Kapadia, CJ used the phrase “judicial anti-avoidance rule” for the first time in Indian jurisprudence, articulating the “look at” test (examining the transaction holistically) and identifying “participation in investment” factors. Further, the Court placed the burden on the Revenue: “at the threshold, the burden is on the Revenue to allege and establish abuse”. The Westminster principle survived. JAAR was recognised but heavily constrained and could be invoked only after Revenue established sham or tax avoidance through the “look at” analysis.[3]
Tiger Global expanded the doctrine in three relevant respects. First, it held that GAAR and JAAR “operate in parallel” and are not mutually exclusive. Second, it introduced “commercial substance” criterion into the JAAR analysis requiring entities to demonstrate real economic and commercial substance. Third, it interpreted Section 96(2) of the Act as shifting the burden onto the taxpayer, holding that once prima facie avoidance is indicated, the taxpayer must rebut this with evidence of bona fide commercial purpose.
This evolution from colourable devices through the look at approach with the Revenue bearing the burden to an undefined “commercial substance” standard highlight that JAAR has undergone a radical expansion when GAAR’s enactment should have rendered it unnecessary.
B. What Tiger Global Left Open
While the Court required entities to demonstrate commercial substance to obtain treaty benefits, it did not define how this JAAR standard relates to GAAR’s Section 97, which provides a detailed, enumerated definition of arrangements that “lack commercial substance”. If the JAAR “commercial substance” test is identical to Section 97, then Tiger Global effectively applied GAAR without GAAR’s procedural safeguards. If the test is distinct, the Court created an undefined standard with no statutory anchoring, which creates legal uncertainty. Further, the Court set no procedural or jurisdictional limits on JAAR. There is no JAAR equivalent of the Approving Panel, threshold amount, FII exemption, or formal invocation requirement.
The Court’s primary holding rested on three cumulative grounds: indirect transfers are not protected under Article 13; TRCs are no longer conclusive; and GAAR applies to post-1 April 2017 exits of pre-2017 investments. JAAR was invoked “in the alternative”. If the GAAR holding was sufficient to decide the case, the JAAR discussion is technically obiter. However, the last ground may be bad in law since the Court interpreted “without prejudice” in Rule 10U(2) as “notwithstanding” Rule 10U(1)(d) rather than preserving its primacy, contradicting Gwalior Rayon Silk. If the GAAR holding falls, the JAAR alternative becomes the operative ratio. This elevates the practical importance of the following argument.
GAAR: The Statutory Framework and Its Safeguards
The statutory architecture of GAAR reveals Parliament’s intent to create a self-contained, procedurally safeguarded anti-avoidance framework and not to supplement open-ended judicial doctrine.
The procedural safeguards under Section 144BA are elaborate and deliberate. The Assessing Officer must first refer the matter to the Principal Commissioner/CIT, who must issue a notice with reasons and provide the assessee 60 days to respond. If the explanation is unsatisfactory, the matter proceeds to an Approving Panel (which reflects the Shome Committee’s recommendation for external independence). Rule 10U(1) further provides the Rs. 3 crore threshold for tax benefits for invocation of GAAR ensuring only significant arrangements are caught; the FII exemption protecting portfolio investors who have not claimed treaty benefits; and the grandfathering provisions shielding income accruing from pre-1 April 2017 investments from GAAR scrutiny.[4]
These safeguards were not incidental features. They were the product of the Shome Committee’s recommendations which emphasised predictability as the cornerstone of anti-avoidance design. GAAR was introduced by Finance Act 2012 specifically in response to the Vodafone decision’s demonstration that judicial anti-avoidance doctrines were inadequate.
Importantly, the Shome Committee never addressed the GAAR-JAAR relationship. CBDT Circular 7/2017 made two important statements. First, “GAAR and SAAR [Specific Anti-Avoidance Rules] can coexist”. Second, “if a case of avoidance is sufficiently addressed by LOB [Limitations of Benefits] in the treaty, there shall not be an occasion to invoke GAAR”. This second statement embodies a principle of residuary application: where a specific mechanism adequately addresses avoidance, the general mechanism need not be deployed.[5] By parity of reasoning, where GAAR (with its comprehensive tests and safeguards) adequately addresses avoidance, JAAR should not be deployed as a parallel, unconstrained alternative.
The Case Against Simultaneous Invocation
A. Comprehensive Codification Occupies the Field
When Parliament enacted Chapter X-A after extensive deliberation, spanning the Shome Committee, multiple Finance Acts, CBDT circulars, and Rule 10U, it comprehensively codified anti-avoidance principles to the exclusion of JAAR.
The strongest comparative authority for this proposition comes from Australia. Part IVA of the Income Tax Assessment Act, 1936 was designed to codify the Newton predication test and overcome weaknesses in judicial doctrine. The High Court of Australia in John v FCT held that Part IVA “makes it impossible to place upon other provisions of the Act a qualification which they do not express”. [6] Once Parliament enacted a statutory GAAR, judicial anti-avoidance glosses on substantive provisions were precluded. The eight-factor objective dominant-purpose test under Section 177D, requiring a formal Commissioner determination, became the exclusive general anti-avoidance mechanism. Statutory GAAR occupied the field.
Canada’s experience reinforces the point through a different mechanism. The Supreme Court of Canada in Stubart expressly rejected a standalone judicial business purpose test, finding that the Westminster principle prevailed absent statutory override. This rejection was the direct impetus for Parliament introducing statutory GAAR by enacting Section 245 of the Income Tax Act in 1988. Further, in Canada Trustco, the Court held that the Westminster principle co-exists with GAAR, but that GAAR operates as a comprehensive mechanism for testing abuse and to this extent limits the Westminster principle.[7] Importantly, Canada Trustco rejected lack of economic substance as a standalone indicator of abuse.[8] Beyond sham, GAAR effectively occupied the field, and courts did not develop parallel judicial anti-avoidance doctrines. The 2024 amendments to Section 245 (Bill C-59) introduced an explicit economic substance element, but this was accomplished through statutory amendment, not judicial invention.
Even in the United Kingdom, while judicial techniques of purposive interpretation were preserved, the statutory GAAR (Finance Act 2013, Part 5) was needed to go beyond what normal statutory construction would permit. [9] Thus, statutory GAAR was given primacy, with judicial substance-over-form doctrine subsisting as a residual technique rather than as a parallel anti-avoidance instrument..
These comparative experiences demonstrate that where a comprehensive statutory GAAR exists, it is intended to be the primary and exclusive vehicle for general anti-avoidance enforcement, with JAAR either displaced or confined to its most limited traditional function. These jurisdictions have each recognised that permitting substance-over-form techniques outside the statutory framework undermines the legislative calibration of safeguards and thresholds and creates uncertainty.[10]
B. Separation of Powers
There is a difference between interpretive function (determining what a statutory provision means) and legislative function (determining what anti-avoidance conditions must be satisfied before benefits are denied). GAAR is Parliament’s answer to the latter. JAAR as applied in Tiger Global imports a “commercial substance” requirement without the statutory safeguards, operating as a judicial alternative to Parliamentary wisdom.
The practical consequence compounds this doctrinal concern. While GAAR represents a legislative choice about the appropriate balance between revenue protection and taxpayer certainty, JAAR allows courts to substitute their own calibration and determine anti-avoidance consequences on terms less protective of the taxpayer than those Parliament deliberately enacted (as highlighted in Section III). A taxpayer confronted with JAAR is consequently exposed to legal uncertainty, as there is no predetermined standard or procedural framework to ex-ante ascertain whether their transaction will attract anti-avoidance scrutiny or the criteria by which such scrutiny will be conducted.
C. The Treaty-Domestic law / Statutory-Judicial Conflation
The invocation of JAAR in a treaty context raises concerns that have been addressed in the existing literature but bear brief restatement. Articles 26 and 31 of the Vienna Convention on the Law of Treaties require treaty performance in good faith and interpretation in accordance with ordinary meaning. The OECD Commentary on Article 1 recognises that domestic anti-avoidance rules are not affected by treaties,[11] and BEPS Action 6[12] acknowledges that domestic judicial doctrines can supplement treaty-level mechanisms when granting the benefit would be contrary to the object and purpose.
However, Tiger Global conflated two distinct questions: the treaty-domestic interface (whether treaties preclude domestic anti-avoidance) and the statutory-judicial interface (whether a domestic statutory GAAR precludes domestic judicial anti-avoidance).. The Court’s reliance on the OECD Commentary and BEPS Action 6 to sanction JAAR’s parallel operation is therefore misplaced since those authorities establish only that a treaty does not preclude the operation of a State’s domestic anti-avoidance rules. The Court provides no reasoning or authority to justify the existence of judicial anti-avoidance doctrines existing after the comprehensive codification of GAAR. This issue is one of domestic institutional design, governed by constitutional principles of legislative supremacy and separation of powers, not by the international tax law materials the Court invoked to resolve it.
D. Two Forums, Two Standards
If the same transaction can be examined under both GAAR and JAAR (with its undefined “commercial substance” standard), a transaction that passes GAAR scrutiny, perhaps because it satisfies the Approving Panel that it has genuine commercial substance under Section 97, could still be denied benefits under JAAR on the basis that a court, applying its own conception of “genuine commercial substance”, reaches a different conclusion. This is a form of jurisdictional double jeopardy: the taxpayer must defend the same arrangement against two different standards before two different decision-makers. If an arrangement has been held permissible by the PCIT/CIT/Approving Panel under GAAR, consistent with CBDT Circular 7/2017’s assurance that GAAR will not be re-invoked in subsequent years where facts remain the same, the question arises whether Revenue can invoke JAAR to challenge the same arrangement before a different forum. If so, the statutory process becomes an elaborate exercise in futility where the taxpayer wins under GAAR only to face the same challenge under JAAR.
Even If JAAR Survives: The Restrictive Reading
Even accepting that JAAR formally survives post-GAAR, it must be confined to its traditional sham only domain rather than expanded into GAAR-like commercial-substance territory.
The textual reading supports this restrictive reading. In paragraph 48, the Court notes “in the alternative, JAAR are invoked to pierce the structure and deny treaty benefits where the transaction lacks genuine commercial substance”.[13] The use of the word “genuine” in addition to commercial substance indicates that the standard is higher than that of lacking commercial substance under GAAR. If the thresholds were identical, the qualifier “genuine” would be superfluous, and the judicial doctrine would be a procedurally unconstrained duplicate of Section 97. Thus, Tiger Global should be read as preserving JAAR only for identifying sham transactions and colourable devices.
Further, in the subsequent paragraph 49, the Court reasons that “[o]nce the mechanism is found to be illegal or sham, it ceases to be ‘a permissible avoidance’ and becomes ‘an impermissible avoidance’ or ‘evasion’”.[14] The requirement for illegal or sham mechanism must be read sequentially. Where GAAR is invoked, the jurisdictional question is whether the arrangement is an “impermissible avoidance arrangement” under Section 96 and thus, illegal. Where JAAR is invoked, by contrast, the mechanism must be a “sham” before it becomes impermissible. The jurisdictional condition for JAAR is therefore tighter than that for GAAR.
Moreover, the shift in the burden of proof contemplated later in paragraph 49 attaches only to GAAR and not JAAR, where the threshold is higher and where no statutory provision shifts the burden. Tiger Global does not purport to overrule Vodafone, and thus, the burden on the Revenue to establish sham “at the threshold” is preserved.
Finally, comparative practice outlined above reinforces this restrictive reading and points toward confining the judicial doctrine once a statutory GAAR exists.
Conclusion
In this article, we have argued that the parallel operation of GAAR and JAAR is doctrinally unsound. Unless Tiger Global is read restrictively, the safeguards enacted in GAAR are rendered cosmetic. A related empirical question remains open and merits inquiry: whether Revenue’s election between GAAR and JAAR at the AO stage tracks the procedural cost of GAAR invocation.
[1] A “colourable device” or “subterfuge” in the context of tax avoidance denotes a transaction structured to exploit a technical reading of the tax law, with no genuine commercial rationale independent of reducing or delaying tax liability. The paradigmatic example is a series of pre-arranged steps, each formally valid in isolation, that collectively produce a tax benefit no single step could achieve and that exist solely for that purpose (as in Ramsay). Such arrangements are to be distinguished from legitimate tax planning, in which a taxpayer avails of a statutory benefit as Parliament intended it to apply.
[2] The Ramsay approach refers to the recognition of the Court’s power to purposively interpret the statute and disregard “inserted steps that have no commercial purpose apart from the avoidance of a liability to tax”.
[3] Vodafone imposed three distinct constraints on JAAR. First, the burden of establishing abuse lay firmly on the Revenue “at the threshold”. Second, the Court required a holistic “look at” examination of the entire transaction rather than isolating individual steps, a constraint that prevented the Revenue from targeting artificially identified sub-transactions. Third, the Westminster principle was expressly preserved, foreclosing any general substance-over-form approach that went beyond identifying a sham.
[4] Note: Tiger Global did not require the formal invocation of GAAR and the procedural machinery it entails because it only concerned an admission stage application under Section 245R(2).
[5] This statement is not strictly lex specialis. The lex specialis principle would apply the special rule to the exclusion of the general rule. Here, the general rule remains applicable if the special rule is insufficient.
[6] John v FCT (1989) 20 ATR 1; (1989) 166 CLR 417.
[7] Paragraphs 13 and 31.
[8] Paragraphs 57 and 60.
[9] Judith Freedman, ‘Designing a General Anti-Abuse Rule: Striking a Balance’ (2014) Asia Pacific Tax Bulletin 167, 169; See also the Aaronson Report (2011), recommending the UK GAAR as a targeted backstop rather than a replacement for purposive interpretation.
[10] Note: The argument advanced here concerns the role of JAAR as a substance-over-form technique, not as an instance of purposive interpretation.
[11] Paragraph 60.
[12] Page 21.
[13] Emphasis supplied.
[14] Emphasis supplied.
* Aadit Anand and Sakshi Meena are Final Year Students at the National Law School of India University, Bangalore