The Andhra Pradesh High Court Strikes Down Section 147A — And Parliament's Habit of Legislating Around Judgments Takes a Direct Hit
A Division Bench That Refused to Blink The Andhra Pradesh High Court's decision in Bala Sreenivas Rao Somisetty v. Income Tax Officer (WP No. 28520 of 2023, decided on 12 October 2026) is not merely a tax judgment. It is a constitutional statement ab...
A Division Bench That Refused to Blink
The Andhra Pradesh High Court's decision in Bala Sreenivas Rao Somisetty v. Income Tax Officer (WP No. 28520 of 2023, decided on 12 October 2026) is not merely a tax judgment. It is a constitutional statement about the limits of legislative power — and it arrives at a moment when Parliament's tendency to nullify inconvenient judicial pronouncements through retrospective amendments has become almost routine. Justices Ninala Jayasurya and T.C.D. Sekhar have done something that requires institutional courage: they looked at a provision that the Revenue desperately needed to survive, examined it against first principles, and struck it down.
Let me be clear at the outset — I think this judgment is substantially correct in its conclusion, though I have reservations about parts of its reasoning. But before I get to the critique, let us understand what was at stake.
The Facts: A Scheme Ignored, A Judgment Nullified, A Battle Joined
The factual matrix is straightforward. The Income Tax Act, 1961 was amended over the years to introduce faceless assessment. Section 151A, inserted by Act 38 of 2020, empowered the Central Government to frame schemes for faceless reassessment. Pursuant to this, the CBDT notified the e-Assessment of Income Escaping Assessment Scheme, 2022 on 29 March 2022. The scheme explicitly included the issuance of notice under Section 148 within its faceless framework.
Despite this, Jurisdictional Assessing Officers (JAOs) across the country continued issuing notices under Sections 148A and 148 in the traditional, non-faceless manner. When challenged, multiple High Courts — including a co-ordinate bench of the Andhra Pradesh High Court itself by order dated 28 October 2025 — held that the JAOs lacked jurisdiction. The notices were quashed.
The Revenue approached the Supreme Court. While the SLPs were pending, Parliament enacted the Finance Act, 2026, inserting Section 147A with retrospective effect from 1 April 2021. The provision declared that the "Assessing Officer" for the purposes of Sections 148 and 148A "shall mean and shall always be deemed to have meant" an officer other than the NFAC or any assessment unit under Section 144B(3). In plain terms, Parliament was telling the courts: you were wrong all along, and we are changing the rules retroactively so that your judgments become meaningless.
The Supreme Court, by order dated 10 April 2026, set aside the High Court judgments on the limited ground that the foundation of those rulings had been altered by the amendment, and remitted the matters to the respective High Courts with liberty to the assessees to challenge Section 147A itself. That is precisely what the petitioners did.
The Arguments: Separation of Powers vs. Legislative Competence
The petitioners' case rested on several pillars. First, they argued that Section 147A was not truly clarificatory — it was a naked attempt to nullify judicial decisions without removing the underlying basis. The scheme under Section 151A remained unamended; Section 148 was still part of the faceless framework. Simply inserting a non-obstante clause declaring the JAO to be the competent authority did not cure the defect — it merely overrode the judicial interpretation without addressing the statutory scheme that gave rise to it.
Second, they contended that the amendment violated the doctrine of separation of powers. The Legislature was, in effect, sitting in appeal over judicial decisions — a function reserved for appellate courts. Third, they raised a technical but potentially devastating argument: the Income Tax Act, 1961 had been repealed on 31 March 2026 and replaced by the Income Tax Act, 2025. The Finance Act, 2026, by its own Section 2(a), came into force on 1 April 2026 — after the repeal. You cannot amend a dead statute.
The Revenue, represented by the Additional Solicitor General, mounted a robust defence. It invoked the presumption of constitutionality, the plenary power of Parliament to legislate retrospectively, and the well-established twin test for validating statutes from Shri Prithvi Cotton Mills Ltd. v. Broach Borough Municipality, (1969) 2 SCC 283 — namely, that the Legislature must remove the defect identified by the court and do so through a subsequent enactment with retrospective effect. The Revenue argued that Section 147A satisfied both limbs: it removed the "defect" (the ambiguity about who is the Assessing Officer) and did so retrospectively.
The Revenue also placed heavy reliance on practical considerations — over 8.82 lakh reassessment notices, Rs. 21.74 lakh crores in assessed income, and the administrative chaos that would ensue if all JAO-initiated proceedings were invalidated.
The Judgment: Two Independent Grounds for Striking Down Section 147A
The Division Bench struck down Section 147A on two independent grounds, either of which would have been sufficient.
The first ground — the constitutional ground — is the more significant one. The Court held that the Legislature cannot nullify a judicial decision without removing the basis on which it rests. The scheme under Section 151A, which included Section 148 within the faceless framework, remained unamended. The proviso to Section 151A(2) expressly stated that no direction modifying the scheme could be issued after 31 March 2022. Therefore, the window for removing Section 148 from the faceless scheme had closed long before Section 147A was enacted. Inserting a non-obstante clause in the Act while leaving the scheme intact was, in the Court's view, an impermissible legislative shortcut — an attempt to override judicial decisions without doing the hard work of actually changing the underlying legal framework.
The Court relied on Shri Prithvi Cotton Mills and State of Tamil Nadu v. M. Rayappaiah Gounder, (1971) 3 SCC 1, to hold that while the Legislature can remove the basis of a judicial decision, it cannot simply disregard or disobey it. Since no defect had been identified in the scheme by any court, and the scheme remained unamended, Section 147A failed the twin test for validating legislation.
The second ground — the repeal argument — is more technical but equally fatal. The Court held that the Income Tax Act, 1961 was repealed at midnight on 31 March 2026, and the Income Tax Act, 2025 came into force on 1 April 2026. Since the Finance Act, 2026, by its own Section 2(a), also came into force on 1 April 2026, it could not amend a statute that had already ceased to exist. The Court relied on Section 5 of the General Clauses Act and the Supreme Court's decision in Common Cause v. Union of India, (2003) 8 SCC 250, to hold that where an Act expressly provides for its date of commencement, the date of Presidential assent is irrelevant.
The Critique: Right Result, But the Reasoning Needs Sharpening
I agree with the outcome. Section 147A was a poorly drafted, constitutionally dubious provision that deserved to be struck down. But I want to flag three areas where the judgment could have been tighter.
First, the Court's analysis of the validating legislation doctrine, while correct in conclusion, could have engaged more deeply with the Supreme Court's jurisprudence on the subject. The key question in validating legislation cases is not merely whether the Legislature "removed the defect" but whether it changed the legal basis on which the judicial decision rested. In Prithvi Cotton Mills, the Supreme Court upheld the validating Act because the Legislature had fundamentally altered the statutory framework — it did not merely declare the court's interpretation wrong. Here, Section 147A did not amend Section 151A, did not modify the scheme, and did not remove Section 148 from the faceless framework. It simply declared, through a non-obstante clause, that the JAO was the competent authority. This is not validation — it is legislative overruling by fiat. The Court identified this correctly but could have drawn the distinction more sharply.
Second, the repeal argument, while technically sound, feels like the kind of argument that could be easily cured by the Legislature. If Parliament re-enacts Section 147A as an amendment to the Income Tax Act, 2025 (rather than the repealed 1961 Act), with a fresh retrospective date, this particular ground would fall away. The Court should have acknowledged this possibility and made clear that even a properly enacted amendment would fail on the constitutional ground alone. By resting on two independent grounds, the Court has given the Revenue a roadmap for a second attempt — amend the right statute this time. The constitutional ground should have been developed as the primary and sufficient basis.
Third, the Court's treatment of the Revenue's public interest argument — the sheer volume of reassessment proceedings and revenue at stake — is somewhat cursory. The Court dismissed it in a single paragraph (paragraph 99), holding that the fact that most assessees did not challenge JAO proceedings does not validate them. While this is legally correct, the Court could have engaged more substantively with the practical implications. When you invalidate proceedings involving Rs. 21.74 lakh crores in assessed income, you owe the public a more detailed explanation of why the rule of law must prevail over fiscal convenience. The principle is right, but the reasoning should have been more robust.
That said, the Court's analysis of the scheme's text and the interplay between Sections 130, 144B, 151A, and the 2022 notification is meticulous and persuasive. The observation at paragraph 88 — that if the Government never intended Section 148 to be faceless, it should have removed it from the scheme before the 31 March 2022 deadline — is particularly devastating. It exposes the Revenue's position as an afterthought, a post-hoc rationalisation of administrative convenience.
The Broader Trend: Parliament's Retrospective Amendment Habit
This judgment must be read in the context of a deeply troubling pattern in Indian fiscal legislation. Time and again, when courts interpret tax statutes in favour of the assessee, Parliament responds with retrospective amendments designed to nullify the judicial interpretation. We saw this with the Vodafone amendment (retrospective taxation of indirect transfers), with the various amendments to Section 9, and now with Section 147A.
The Supreme Court has, over the years, developed a nuanced jurisprudence on this. In CIT v. Vatika Township (P) Ltd., (2015) 1 SCC 1, the Court held that legislation should not be given retrospective effect unless the statute expressly or by necessary implication so provides. In Prithvi Cotton Mills, it upheld validating legislation but only because the Legislature had genuinely removed the defect. The Andhra Pradesh High Court's judgment fits squarely within this tradition — it does not deny Parliament's power to legislate retrospectively, but insists that the power be exercised within constitutional limits.
What makes Section 147A particularly egregious is its timing. It was enacted while the matter was pending before the Supreme Court, and it was designed to pre-empt the judicial outcome. This is not governance — it is gamesmanship. The Legislature is entitled to change the law prospectively. It is even entitled, within limits, to change it retrospectively. But it is not entitled to use its legislative power as a weapon to defeat pending litigation.
The Takeaway for Practitioners
Several practical lessons emerge from this judgment.
First, if you are representing an assessee who received a Section 148 notice from a JAO (as opposed to the NFAC) after 29 March 2022, this judgment provides strong authority for challenging the notice as being without jurisdiction — at least within the territorial jurisdiction of the Andhra Pradesh High Court. Given that the Punjab and Haryana High Court reached a similar conclusion (though its operation was stayed by the Supreme Court), the weight of authority is shifting in favour of the assessee.
Second, when challenging retrospective amendments, always examine whether the Legislature has actually removed the basis of the judicial decision or merely attempted to override it. If the underlying statutory scheme remains unamended, the amendment is vulnerable to constitutional challenge. The "twin test" from Prithvi Cotton Mills is your friend.
Third, pay attention to repeal and savings clauses. The Court's reliance on the repeal of the 1961 Act and the commencement date of the Finance Act, 2026 is a reminder that technical arguments about legislative competence can be just as effective as grand constitutional principles. Sometimes the most powerful argument is the simplest one: you cannot amend a statute that no longer exists.
Fourth, the Revenue's "floodgates" argument — that striking down Section 147A would invalidate lakhs of reassessment proceedings — did not move this Court. This is significant. It signals that at least some High Courts are willing to prioritise constitutional principle over administrative convenience. When you encounter the floodgates argument in your own cases, cite this judgment and remind the court that the rule of law does not yield to the volume of non-compliance.
Finally, expect the Revenue to challenge this judgment before the Supreme Court. The stakes — Rs. 21.74 lakh crores in assessed income — are simply too high for the Government to accept defeat. The Supreme Court's earlier order of 10 April 2026, which remitted the matters to the High Courts while expressly leaving open the validity of Section 147A, suggests that the Court anticipated this challenge. The question now is whether the Supreme Court will uphold the High Court's constitutional reasoning or find a way to save the amendment. Given the Court's recent track record on retrospective taxation (particularly in the Vodafone and Cairn arbitration contexts), there is reason for cautious optimism — but no room for complacency.
This judgment is a reminder that the Constitution is not a document of convenience. It constrains the Legislature as much as it constrains the Executive. The Andhra Pradesh High Court has done its job. Whether the Supreme Court will do its job remains to be seen.
Related Cases
Tags
Published by AnrakLegal AI