CBDT’s Avlokan: More Tax Agreements, But Where Is the Evidence of Public Value?
Finance Minister Nirmala Sitharaman speaks after inaugurating ‘Aayakar Sindhu’ - new Income Tax Office Complex at Nariman Point, Mumbai. (Image Sitharaman office on X)
By P. SESH KUMAR
CBDT’s Avlokan 2025–26 records a surge in APAs, MAP resolutions and international tax cooperation, but the report does not adequately show their fiscal impact, compliance outcomes or protection of India’s tax base.
New Delhi, August 11, 2026 – CBDT’s Avlokan 2025–26 is an impressive first institutional account of India’s Foreign Tax and Tax Research Division (FTTR): polished, ambitious and rich in evidence of activity. Yet it largely measures movement rather than impact. The report demonstrates that CBDT has negotiated, signed, hosted, exchanged, resolved and participated; it does not adequately establish what these actions have delivered in protected revenue, reduced double taxation, faster compliance, lower litigation, better taxpayer conduct, or improved treaty-partner reciprocity. Its central weakness is not that it celebrates achievement, but that it treats achievement counts as a sufficient proxy for public value.
The report’s triumph–and its blind spot
Avlokan records an institution operating in a vastly more complex international tax environment than the treaty-focused Foreign Tax Division (FTD) of earlier decades. FTTR now covers treaty policy, Mutually Agreed Procedure (MAP), Advance Pricing Agreement (APA), exchange of information (EOI), transfer pricing (TP) policy, multilateral negotiations, tax transparency, capacity building and eleven Income Tax Overseas Units; it works through two Joint Secretaries, six Director-level divisions and more than 100 officers and officials. The report is therefore valuable: it makes visible an administrative domain that ordinarily works behind closed doors and often receives public attention only when a dispute erupts.
The headline achievements are formidable. CBDT reports that, by 31 March 2026, India had concluded 1,035 APAs–751 unilateral and 284 bilateral–covering more than 5,700 assessment years; during FY 2025–26 alone it signed 220 APAs, including 84 bilateral agreements and its first bilateral APAs with France, Ireland, Indonesia and Sweden. It also reports a sharp fall in MAP inventory, from 947 cases at the beginning of 2020 to 365 at the end of 2025, while average transfer-pricing MAP resolution time fell from 64.86 months in 2016 to 39.78 months in 2025.
But the document frequently mistakes the scale of administrative production for proof of administrative success. “More than 17,000 pages” of APA agreements, theatrically compared with the height of an adult, is memorable communication but not a fiscal or governance metric. Pages signed neither prove that the agreed transfer prices were robust nor that the covered taxpayers complied, nor that India received an appropriate share of the taxable profit generated by Indian functions, assets and risks.
That distinction is decisive. A tax administration does not exist to generate agreements, meetings or ceremonial milestones. It exists to protect legitimate revenue, apply law consistently, prevent double taxation, curb abusive profit shifting, reduce unproductive disputes, and make the tax system credible to compliant businesses and the public. An APA can be excellent public administration; an APA count alone cannot establish that it was.
What Avlokan does not answer
The report makes a persuasive case for the programme’s expansion, but it leaves the most consequential questions either unanswered or treated only through broad assertions. It reports the number of agreements, applications, assessment years covered, treaty partners and industry sectors, but it does not publish an outcome framework connecting these inputs and outputs to verified fiscal, compliance and dispute-resolution results.
Most notably, the report does not disclose the aggregate tax effect of APAs. Does CBDT say this cannot be done? It does not state whether APA-covered taxpayers paid more, less, or broadly the same Indian tax than they would have paid under a defensible counterfactual; whether the agreements preserved a quantified amount of Indian tax base; or whether they sacrificed revenue to buy certainty. Confidentiality may properly protect taxpayer-specific pricing, margins, comparable sets and negotiated terms. It does not justify withholding aggregate, anonymised evidence of the programme’s fiscal consequences.
Nor does the report show the post-signing life of an APA. The real test begins after the agreement is executed: are annual compliance reports filed on time; are critical assumptions monitored; do actual facts diverge from the agreed functional and risk profile; are agreements revised, cancelled, revoked or allowed to lapse; and are transfer-pricing audits in covered years genuinely reduced? The report presents APAs as a preventive device against future litigation, but it does not quantify subsequent audit adjustments, compliance failures, revocations, litigation avoided, or the share of covered transactions that later generated controversy.
The treatment of bilateral APAs also leaves the “give and take” question unresolved. A bilateral APA is not simply an Indian tax settlement. It is an allocation of taxing rights and transfer-pricing outcomes negotiated with another sovereign tax administration under the treaty’s MAP article. The report celebrates the geographic expansion of bilateral APAs and growing cooperation with treaty partners, but it does not disclose, even at a suitably aggregated level, whether Indian positions prevailed, where India conceded, the typical areas of compromise, the elapsed time at each stage, or whether outcomes are symmetric across India’s principal treaty partners.
The same gap affects exchange of information (EOI). The report describes the architecture of EOIR, automatic exchange of information, FATCA, CRS and the emerging Crypto-Asset Reporting Framework, and it expressly frames the transition as one “from information to compliance.” But that compelling phrase requires proof: how many data leads were matched to PANs, how many were risk-scored, how many became notices or assessments, how many yielded voluntary disclosures, what additional tax, interest and penalty were realised, and what proportion of leads were unusable because of inaccurate, late, incomplete or non-actionable data? Without those links, volumes of information received and transmitted remain a measure of data traffic, not enforcement impact.
Achievements are a beginning
Listing achievements is laudable as transparency, institutional memory and recognition of difficult public work. It is not, however, a defensible end-state for a public authority wielding sovereign taxing power. The distinction is between an annual activity report and a performance-and-accountability report.
The present report’s dominant indicators are outputs: APAs signed, MAP cases closed, agreements negotiated, international events held, countries engaged, training programmes undertaken and systems introduced. Outputs matter; they may indicate capacity, effort and throughput. Yet they do not demonstrate outcomes, which should include durable tax certainty, reduction in duplicate taxation, improved voluntary compliance, measured administrative savings, safeguarded revenue, reduced litigation burden, better treaty-partner cooperation, and preservation of India’s taxing rights.
A more searching assessment would separate three questions that are now too often collapsed into one.
- First, did the programme process cases efficiently?
- Second, did the agreement produce a technically sound arm’s-length result?
- Third, did the agreement improve India’s fiscal and administrative position compared with reasonable alternatives, including ordinary audit, safe harbour, unilateral resolution and bilateral MAP?
An APA programme can perform well on the first question while performing inadequately on the second or third.
The fact that a taxpayer seeks an APA does show demand for certainty. It does not automatically prove public value. Large, sophisticated multinational groups may value APAs precisely because they reduce uncertainty, contain controversy and make future tax costs more predictable. Those are legitimate aims, but the Government’s task is to demonstrate that certainty has not become certainty at an unjustified revenue discount.
Auditing the Invisible State: Can CAG Expose Faceless Tax Power?
Testing the real outcomes
CBDT should adopt a published annual APA-and-international-taxation outcome scorecard, based on anonymised aggregates and independently auditable definitions. It need not disclose any taxpayer’s identity, protected commercial information, negotiating position or treaty-partner communication.
A credible assessment would track intake, age-profile and completion time separately for unilateral, bilateral, multilateral, new and renewal APAs. It would show not merely the number closed but the inventory-to-disposal ratio, median and percentile completion times, withdrawals, rejections, suspensions, renewals, cancellations, revisions and revocations.
The United States’ annual APA reporting provides a useful discipline here: its 2025 report recorded applications, executed agreements, pending inventory and processing times, enabling external observers to see both production and backlog rather than celebrating closures alone. The IRS APMA programme executed 110 APAs in 2025, received 178 applications, ended with 622 pending cases and reported a 41.6-month median completion time.
CBDT should also quantify compliance outcomes. It should publish the proportion of APA annual compliance reports filed on time; the proportion reviewed; the number of critical-assumption breaches detected; the value and number of adjustments arising from compliance monitoring; and the number of APAs modified, cancelled or revoked. It should disclose the extent to which APA-covered years experienced audit controversy compared with a carefully matched non-APA population. That comparison must control for sector, taxpayer size, transaction type, risk profile and years covered; otherwise, a low dispute rate could simply reflect self-selection by more compliant taxpayers.
The fiscal question demands particular rigour. A reliable annual public report could provide broad, anonymised bands of Indian taxable-income or tax-base impact, without revealing agreed margins or individual taxpayer data. CBDT could report whether the APA outcome was above, within or below the central range initially asserted by the taxpayer, and the aggregate direction and scale of modifications from original filing positions. It could also estimate avoided administrative and litigation costs, while clearly distinguishing estimated savings from realised tax collection. The objective is not to force the publication of sensitive commercial data; it is to show that negotiated certainty has generated defensible public value.
MAP assessment should go beyond inventory reduction. India should publish the percentage of MAP cases resolved with full double-tax relief, partial relief, unilateral relief, no relief, taxpayer withdrawal, or failure to reach agreement; separately report transfer-pricing and non-transfer-pricing cases; and show median age and completion time by major treaty partner. The report notes that 903 MAP cases were concluded between 2020 and 2025 and that the inventory declined substantially. It does not reveal how many taxpayers obtained complete elimination of double taxation, how often India’s competent authority position did not secure mutual agreement, or whether implementation by field formations occurred promptly after settlement.
The American discipline
The IRS model is not perfect, but it offers an important lesson: public reporting should expose strain as well as success. The IRS’s APMA annual statutory reports give aggregate information on applications, executed APAs, pending inventory, processing time, bilateral partner countries, covered transactions, industries and transfer-pricing methods. This allows the Congress, Parliament, taxpayers, researchers and oversight bodies to ask whether growing demand is being translated into timely delivery and whether capacity matches caseload.
That disclosure has practical value precisely because it is not always flattering. In 2025, the IRS APMA programme’s executed APAs fell to 110 from 142 in 2024, pending inventory increased to 622, and average completion time rose to 44.1 months. A system that reports such deterioration is more credible than one that reports only its most attractive totals. It makes capacity stress visible and therefore governable.
For CBDT, the key transferable lesson is not imitation of American disclosure formats but adoption of a transparent performance architecture. CBDT should report stock, flow, age, timeliness, case outcome, compliance follow-through and aggregate fiscal effect. The IRS model itself does not eliminate the difficulty of measuring whether each APA secured the best substantive result; no country has found a simple perfect metric. But it demonstrates that confidentiality is compatible with candid, recurring disclosure about programme performance.
The British test
HMRC’s approach is also instructive because it reports transfer-pricing performance in a broader revenue-and-compliance frame rather than presenting APAs as a stand-alone trophy cabinet. Its published transfer-pricing statistics have included total transfer-pricing yield, cases settled, average time to settle enquiries, APA applications and agreements, APA timing, MAP results and Diverted Profits Tax indicators.
The latest reported UK figures show both achievement and friction: HMRC’s transfer-pricing yield for 2024–25 was GBP 3.387 billion, 143 transfer-pricing enquiry cases were settled, the average enquiry duration was 41 months, 26 APAs were agreed, and average APA time was 43.9 months. That is a more mature reporting posture because it makes it possible to ask whether higher yield arises from stronger compliance, case mix, one-off settlements, diverted-profits work, or changes in enforcement strategy–and whether reduced APA timing has been achieved without lowering analytical quality.
CBDT should be cautious about mechanically importing a “yield” metric. An APA is prospective and preventive; a large immediate tax yield may sometimes indicate that the programme is functioning as a settlement channel rather than a genuine advance-pricing mechanism. But HMRC’s broader framework makes the essential point: APAs, MAP, enquiry outcomes, litigation, compliance interventions and revenue protection must be read together. A programme cannot credibly claim success merely because agreements have been signed faster or in larger numbers.
Give and take: principle or performance?
“Give and take” is unavoidable in bilateral tax administration. A bilateral APA or MAP settlement is an exercise in reciprocal accommodation: each competent authority may move from its preferred position to eliminate double taxation, secure an administrable result and preserve a working treaty relationship. The alternative may be years of controversy, double taxation, litigation and investment uncertainty. There is nothing inherently suspect about compromise.
But reciprocity must be tested, not presumed. The report’s claim of widening bilateral engagement is encouraging, especially given the growth to 84 bilateral APAs in FY 2025–26. Yet the public cannot tell whether give and take is balanced across partner jurisdictions, whether Indian cases are resolved as quickly as foreign cases, whether India’s source-country interests are protected, or whether repeated concessions are concentrated in high-value industries or particular treaty relationships.
CBDT should maintain a confidential “reciprocity dashboard” for the Member (Legislation), Chairperson CBDT and Finance Ministry, with an annual public anonymised extract. It should track, treaty partner by treaty partner, the number and value-band of cases initiated and concluded; time to resolution; full versus partial relief; settlement patterns; counterpart responsiveness; repeated points of disagreement; incidence of arbitration or unresolved cases; and implementation delays after mutual agreement. It should also record whether the outcome preserves India’s position consistently with the taxpayer’s Indian value creation, personnel, assets, functions and risks.
At the operational level, every bilateral APA should be subject to a structured post-agreement review. The review should test whether critical assumptions held, whether the taxpayer’s Indian operating model materially changed, whether the tested party and profit-level indicator remain appropriate, whether comparables have become stale, and whether the agreement has prevented rather than merely postponed controversy. This is not a request to reopen settled bargains casually. It is the minimum discipline needed to ensure that an advance agreement remains grounded in the facts on which it was negotiated.
Confidentiality is not immunity
A likely CBDT objection to an independent audit by CAG would be that APAs, MAPs, exchange-of-information arrangements and treaty negotiations are diplomatic, confidential, country-to-country matters and therefore unsuitable for such audit. The first half of that proposition is correct; the conclusion is not.
These arrangements undoubtedly contain protected material. Taxpayer-specific information is subject to statutory confidentiality. Treaty exchanges and competent-authority communications often carry confidentiality obligations, and disclosure of negotiating positions can damage India’s standing and undermine reciprocal cooperation. The report itself identifies confidentiality and data safeguards as part of FTTR’s work allocation. CAG should neither demand public disclosure of taxpayer data nor second-guess the merits of each negotiated arm’s-length price in an open audit report.
But confidentiality does not place executive action outside constitutional financial accountability. CAG’s role is not to publish protected information; it is to audit whether Government has designed and operated systems that safeguard revenue, comply with legal and procedural requirements, protect confidential data, obtain appropriate approvals, monitor implementation and measure results. An APA is a public-power instrument with fiscal consequences. Its confidentiality changes the method of audit, not the existence of audit.
The appropriate distinction is between auditing the substance of a confidential negotiation as though CAG were a substitute competent authority, which would be inappropriate, and auditing the governance, controls, performance and fiscal-risk management of the APA system, which is squarely within legitimate public accountability. CAG can access protected records under controlled procedures, use small samples, anonymise taxpayer references, aggregate findings, and place genuinely sensitive material in confidential annexes addressed to authorised constitutional authorities rather than releasing it publicly.
Disposal Without Justice: Inside India’s ₹25L Cr Tax Appeal Crisis
What CAG should examine
CAG should undertake a performance audit of the APA, MAP and international-taxation administration–not an audit of individual treaty bargaining positions. The audit should examine whether CBDT has a formally approved objectives framework that balances tax certainty, treaty compliance, taxpayer service, protection of the Indian tax base and reduction of double taxation. It should test whether programme targets exist and whether they are periodically revised in light of caseload, staffing and international complexity.
It should assess the completeness and reliability of management information: application receipt dates, time at each stage, case-age profiles, withdrawals, rejections, settlements, annual compliance reports, breach of critical assumptions, modifications, cancellations, revocations, post-APA audit outcomes and MAP implementation. The present report –Avlokan-offers important aggregate counts but does not demonstrate the underlying data architecture necessary for sustained independent performance assessment.
CAG should examine governance safeguards. These would include case-allocation protocols, conflict-of-interest declarations, approval hierarchies, legal vetting, technical quality review, peer review of high-value or novel cases, documentation standards, retention policies, segregation of negotiation and compliance-monitoring functions, information-security controls, and escalation processes when a taxpayer breaches critical assumptions. It should also assess whether FTTR’s human resources, specialist training and data tools are adequate for the scale of work described in the report.
CAG should audit aggregate fiscal-risk management without requiring publication of confidential taxpayer data. It can test whether CBDT has estimated the revenue implications of agreement positions, documented alternatives considered, reviewed cases involving significant deviations from audit positions, and assessed revenue outcomes after rollback years and implementation. It can examine a confidential sample to determine whether the record shows a reasoned explanation for accepting the negotiated method, range, tested party, critical assumptions and rollback treatment.
CAG should also examine MAP and APA implementation by field formations. A technically correct agreement fails in practice if orders are not modified promptly, refunds are delayed, interest consequences are mishandled, or relief is not reflected in the taxpayer’s records. The report describes MAP as a seven-stage lifecycle culminating in implementation by field authorities.That final stage should be audited for timeliness, consistency and error rates.
Finally, CAG should test whether exchange-of-information (EOI) data is actually used. It should examine data receipt, matching, risk scoring, dissemination, field action, assessments, collection, voluntary compliance and closure of false positives. This would allow a clear answer to the question that the report invites but does not resolve: has India converted international transparency into domestic compliance outcomes?
Memorandum or Mirage: CAG–CBDT MoU and the Stakes for Tax Audit
A confidential audit protocol
The way forward is a formal confidentiality protocol agreed among CBDT, the Department of Revenue and CAG. It should permit CAG access to underlying APA, MAP and exchange-of-information records on a need-to-know basis; require personnel with access to be subject to equivalent secrecy obligations; prohibit disclosure of taxpayer identity and counterpart communications in public reports; and use aggregation, redaction and coded case references in published findings.
Such a protocol should distinguish three layers of disclosure. The first should be public: aggregate programme results, timeliness, inventory, compliance monitoring, broad fiscal-impact bands, outcome categories and audit recommendations. The second should be restricted: anonymised case samples, detailed country patterns and material process failures shared with authorised parliamentary and executive oversight channels. The third should remain tightly protected: taxpayer-specific pricing data, comparables, negotiation memoranda, foreign-authority communications and information received under confidentiality restrictions.
This model protects the integrity of diplomatic tax administration while rejecting the dangerous proposition that secrecy extinguishes accountability. In fact, the more confidential the underlying activity, the stronger the need for a trusted independent assurance mechanism.
The reform agenda
CBDT should convert Avlokan from a celebration of activity into an annual accountability instrument. Its next edition should carry a concise, consistently defined APA-MAP-EOI performance annex. That annex should report application and closure flows, backlog, completion times, taxpayer-compliance monitoring, agreement modifications and cancellations, MAP relief outcomes, implementation time, aggregate fiscal-effect bands, enforcement conversion from EOI data and a treaty-partner reciprocity dashboard.
CBDT should establish an internal International Tax Outcomes Committee chaired at Member level, with representation from FTTR, APA, assessment, audit, systems, risk management and legal functions. It should review high-value and systemic cases, validate annual metrics, monitor post-agreement compliance, identify recurring treaty-partner frictions and publish a sanitised annual assurance statement. The committee must be evaluated not on the number of agreements cleared but on the durability, legality, timeliness and measured public value of those agreements.
CBDT should commission periodic independent peer review–by a carefully designed panel of former revenue officials, transfer-pricing specialists, public-finance experts and governance professionals–subject to confidentiality safeguards. This should complement, not replace, CAG’s statutory performance audit. International peer mechanisms are useful, but they cannot substitute for domestic constitutional accountability because they assess conformity with international standards, whereas CAG’s task is to assess whether Indian public resources and fiscal interests have been administered with economy, efficiency and effectiveness.
The real prize is not fewer APAs or more APAs. It is a system in which a multinational receives genuine certainty only after India obtains a demonstrably defensible tax outcome; in which negotiated reciprocity is measured rather than asserted; in which secrecy protects legitimate interests but not weak administration; and in which a glowing annual report survives the harder question: what did the Republic actually gain?
(This is an opinion piece. Views expressed are the author’s own.)
Follow The Raisina Hills on WhatsApp, Instagram, YouTube, Facebook, and LinkedIn
About The Author
