GST Council’s Big Shift: Faster Refunds, Easier Compliance and Fewer Enforcement Powers
The 57th GST Council meeting marks a shift from tax-rate rationalisation to the everyday challenges of tax administration. (Image CBIC on X)
By P. SESH KUMAR
The 57th GST Council meeting focuses on registration, return mismatches, refunds and enforcement, but effective implementation will determine whether businesses benefit.
New Delhi, October 11, 2026 — The 57th meeting of the GST Council marks an important change of emphasis: from the headline politics of tax rates to the everyday economics of tax administration. Its recommendations address the points at which GST most frequently rubs against business —registration, return reconciliation, blocked credit, delayed refunds, goods interception and disputes. They also propose a striking recalibration of enforcement through withdrawal of GST arrest powers and a higher prosecution threshold. The package deserves support because it recognises that an ostensibly neutral consumption tax can become a tax on working capital, investment and administrative endurance when credit is obstructed and procedures are unpredictable. But the promise is not self-executing. Legislative coordination, reliable software, accountable risk assessment, protection for genuine purchasers and disciplined adjudication will determine whether reform reaches the taxpayer.
Successive Comptroller and Auditor General (CAG) reports have documented several of the underlying weaknesses, particularly incomplete invoice matching, unreliable data, inadequate refund controls and deficient departmental follow-up.
The real reform is no longer at the rate counter
A tax can have fewer rates and still have too many obstacles. It can be collected digitally and still be administered intrusively. It can promise seamless credit and still leave a manufacturer financing the Government through an unusable balance in the electronic ledger. The importance of the 57th GST Council meeting lies in its recognition of this gap between the architecture of GST and the experience of paying it.
The official release describes the preceding reform exercise as one focused on rate rationalisation and reduction. This meeting turns primarily to registration, returns, refunds, adjudication and trade facilitation. That is a sensible progression. Rates determine the visible burden; administration determines how often that burden becomes uncertain, expensive or coercive. The central organising idea of the package is trust—not trust without verification, but a less intrusive relationship supported by better information and more proportionate intervention.
Yet precision must precede enthusiasm. The Council has recommended changes; it has not, through its press release, enacted them. The release expressly states that the relevant circulars, notifications and legislative amendments alone will give the decisions legal effect. Accordingly, arrest powers have not been abolished merely because abolition has been recommended, newly proposed credits cannot automatically be claimed, and the optional Annual Return Quarterly Payment scheme is only approved in principle. The proposed November 2026 and April 2027 dates must likewise be read as recommended commencement arrangements, subject to the instruments that implement them.
This is not a drafting nicety. It is the difference between a policy announcement and a defensible tax position.
A stronger tax base—but not yet a proven causal story
There is a compelling account of the year following rate restructuring:
- taxable supplies rising from Rs 40.19 lakh crore to Rs 50.58 lakh crore,
- consumer supplies from Rs 5.98 lakh crore to Rs 7.58 lakh crore, and
- gross tax liability from Rs 5.85 lakh crore to Rs 6.64 lakh crore.
It reports a lower effective domestic tax rate alongside higher revenue. These figures, however, are not reproduced in the official release. They should therefore be described as figures reported in the briefing note rather than independently verified findings of the press release.
Even on their own terms, some claims need correction or qualification. Consumer-supply growth of 26.7 per cent is not 31 per cent faster than overall taxable-supply growth of 25.8 per cent. It is 0.9 percentage points higher, or approximately 3.5 per cent faster in relative terms. The fall from 14.55 per cent to 13.13 per cent is 1.42 percentage points. More importantly, the periods underlying the reported monthly supply and liability figures need to be identified before comparisons can carry analytical weight.
The same caution applies to the proposition that lower rates caused the base to expand. Higher reported supplies may reflect some combination of real growth, inflation, improved disclosure, changes in supply composition and enforcement. Gross tax liability is not interchangeable with cash revenue, while gross receipts should not be confused with collections after refunds. As of October 8, FY 2026–27 revenue-growth figure necessarily concerns part of the financial year, not a completed annual outcome.
The defensible conclusion is narrower but still significant: the briefing note suggests that rate reduction and base expansion have coexisted. Establishing how much of the expansion was caused by the rate changes requires comparable periods, consistent definitions and a decomposition of the underlying data. Reform does not become less persuasive when its evidence is stated carefully; it becomes harder to dismiss.
Registration: opening the door without reopening the fraud factory
Registration is the taxpayer’s first encounter with the system. When documentary expectations are unclear, an ostensibly routine application can become a negotiation over what an officer considers sufficient. The Council proposes a comprehensive circular, document-linked dropdowns and contextual portal guidance. It also recommends automatic acceptance of amendments to registration particulars other than the principal place of business; for taxpayers under the rule 14A automatic route, even principal-place amendments would be accepted automatically. This is broader than the selected amendment categories emphasised in some quarters.
The purpose is sound: eliminate avoidable discretion where a change can be processed through objective information. A trade-name update or alteration in management should not routinely require the same level of intervention as a suspicious identity or location change.
But automation is not the same as validation. CAG’s first GST audit found mandatory fields accepting junk values, insufficient validation against relevant databases, and registration-system rules that were not properly aligned with the law. Its findings show why a faster registration process must also be a better-designed process. Those were historical findings, not proof that every defect persists today, but they remain directly relevant to the safeguards required for the next expansion of automation.
The Rs 2.5 lakh monthly credit-passing threshold should therefore be treated as a screening criterion, not a certificate of innocence. Low-value registrations can still be connected through common controllers, addresses or financial arrangements. The appropriate response is risk-sensitive identity verification and subsequent monitoring–not blanket suspicion of every small applicant.
Closure deserves the same attention as entry. The proposed phased automatic cancellation requires pending returns to be filed and dues paid. Integrating final-return information into the cancellation application should reduce duplication. System-based restoration after specified defaults is cured is equally valuable: a registration should not remain disabled merely because the administrative process has failed to recognise subsequent compliance.
Implementation must nevertheless distinguish an established liability from a contested demand. It should also state clearly how stock-related liabilities, pending refunds, later investigations and disputed balances interact with cancellation. Otherwise, a promise of easy exit may become another locked door.
Returns: correcting the machinery that manufactures mismatches
The return proposals address a fundamental design problem: several representations of the same transaction must agree, but the taxpayer needs a lawful route to correct them when they do not. The Council proposes closer alignment between GSTR-1/1A/IFF and GSTR-3B, between GSTR-2B and credit claimed, and between reverse-charge payments and related credit. It also proposes formal rules for the Invoice Management System and electronic reversal-and-reclaim statements. The revised mechanism is proposed from the April 2027 return, following public consultation.
This is potentially more consequential than a cosmetic simplification of forms. If a correction is made in one place without reaching the counterparty or the corresponding ledger, the system can convert a repair into a fresh discrepancy. Linking corrections across the chain should reduce this problem.
CAG anticipated the underlying need years ago. In 2019 it identified invoice matching as essential to achieving GST’s promised non-intrusive compliance regime. In 2022 it acknowledged progress in linking the principal returns and credit statements, while calling for stronger preventive checks. It also found significant inconsistencies between taxable values, declared liabilities and credit figures that prevented it from establishing the reliability of the data for certain pan-India analytical purposes.
The claim in some quarters that approximately 95,000 mismatch notices yield recovery of only 0.08 per cent of the amount involved is not substantiated in the official release. Even if verified, low recovery would not establish that virtually all discrepancies are data-entry errors. It could also reflect pending proceedings, weak follow-up, erroneous demands or incomplete recoverability.
CAG’s Report No. 25 of 2025 supplies an important counterweight. In its selected centralised audit sample, statutory compliance deviations were identified in 2,519 cases involving Rs 21,695.11 crore, representing 31 per cent of the inconsistencies for which departmental responses were received. Those findings concerned historical return periods and a selected audit sample; they cannot be extrapolated into a national fraud rate. They nevertheless demonstrate that discrepancies cannot safely be dismissed as mere typing mistakes.
The answer is discrimination, not indiscriminate notices. The portal should explain the discrepancy, identify its underlying documents and permit correction before escalating a case where appropriate. Equally, acceptance of an invoice in IMS cannot alone prove that goods were received, the transaction was genuine or all legal conditions for credit were satisfied. A buyer’s digital response is evidence within the tax system, not a substitute for the tax law.
Refunds: stop treating the taxpayer’s money as departmental float
The refund proposals are among the package’s strongest elements. Excess cash-ledger balances would be refunded automatically. The period for acknowledgment or deficiency communication would fall from fifteen to ten days, with deemed acknowledgment if neither is issued. Risk-assessed claims for zero-rated supplies and inverted-duty accumulation would receive an automatic 90 per cent provisional sanction. A later phase would provide automated full sanction for qualifying zero-rated claims after adjustment of pending dues. That second-phase full-sanction proposal should not be represented as extending identically to every inverted-duty claim.
The case for reform is straightforward. A delayed refund raises financing costs and can impair the competitiveness of an otherwise efficient business. Refund administration is therefore part of economic policy, not simply a back-office function.
CAG’s 2022 report documented both sides of the problem. It found significant delays in acknowledgment, deficiency communication and sanction, together with non-payment of interest in the majority of delayed cases examined. It also identified 410 instances of duplicate payments amounting to Rs 13.73 crore in the pre-automation period, and 522 excess or inadmissible refunds amounting to Rs 185.28 crore in a risk-based sample of 12,283 claims. It called for stronger near-real-time fraud detection before sanction and effective review and post-audit.
The lesson is not that quick refunds are dangerous. It is that speed must be built on reconciled data, sound eligibility rules and controls against duplicate payment.
A reliable system should explain why a claim has been routed away from automatic processing, provide a timely route to contest inaccurate risk indicators, and distinguish disputed dues from amounts legally available for adjustment. It should measure time to actual bank credit, not merely time to sanction. A refund approved on screen but stranded in payment reconciliation is no refund from the business’s perspective.
(This is first of the two-part series. Views expressed are the author’s own.)
Follow The Raisina Hills on WhatsApp, Instagram, YouTube, Facebook, and LinkedIn
About The Author
