CBDT’S Rising Efficiency Requires Oversight Not Opacity
Image credit X.com Sansad TV
India’s CBDT has achieved record throughput, but a standing CAG audit anchored in OECD and BEPS norms can safeguard quality.
By P SESH KUMAR
New Delhi, October 17, 2025 — Advance Pricing Agreements (APA) were born to de-escalate transfer-pricing wars by trading litigation for ex-ante certainty. India’s programme, legislated in 2012, now spans unilateral, bilateral and multilateral deals with roll-back and renewal options, and it is increasingly the preferred highway for tax certainty in complex value chains.
But APAs sit behind thick curtains of confidentiality—both statutory and treaty-based-raising a hard question: can a Supreme Audit Institution like the CAG of India examine whether the programme delivers fairness, value, and compliance without tearing those curtains?
An APA is the tax equivalent of pre-flight clearance. Before a multinational pushes back from the gate, the taxpayer and the tax authority agree the method—TNMM margins, CUP comparables, DEMPE allocations and critical assumptions-that will guide pricing over a fixed horizon; if conditions hold, turbulence later is avoided.
India hard-wired this idea into law through Sections 92CC–92CD of the Income-tax Act and the APA Scheme in Rules 10F–10T, with a companion Rule 44GA to choreograph bilateral and multilateral cases via the Competent Authority.
In practice, the Central Board of Direct Taxes (CBDT) signs the agreement, taxpayers file annual compliance reports, and, where terms affect earlier years, Section 92CD triggers updated returns to align past assessments with the agreed method. The statutory chassis is clear: APAs are lawful, bounded in time, and conditional on facts staying within the agreed envelope.
The rationale was always two-fold. First, certainty is economic infrastructure: when cross-border margins or royalty rates are locked down ex ante, capital budgeting stops guessing and starts investing. Second, APAs decongest litigation and MAP pipelines; India’s own annual reporting shows a steady pipeline of conclusions and inventory reduction, signalling maturing administrative capacity and rising taxpayer demand for predictability.
This promise, however, is only as good as the programme’s governance. Without disciplined screening, defensible comparables, robust DEMPE analyses, and vigilant monitoring of critical assumptions, an APA can quietly warehouse revenue risk for years. That is why the system-not each taxpayer’s secrets-must be open to audit.
The confidentiality problem is real and not optional. On the domestic plane, Section 138 protects “information respecting assesses,” allowing disclosure only to notified authorities and in carefully fenced circumstances. On the international plane, Article 26 of the OECD Model Convention and its 2024 commentary updates insist that exchanged information in bilateral or multilateral APAs remains confidential; repeated lapses can even justify suspending assistance.
These rules exist for good reasons: taxpayers hand over granular forecasts, contracts, cost structures and intangibles maps, and partner administrations speak frankly in competent-authority channels. Public ventilation would chill cooperation and crater trust. But neither statute nor treaty says “no audit.” They say “handle with care.”
The operational complexities of APAs amplify the case for oversight. Every file is a moving target of value chains, intangibles and location-specific advantages; DEMPE narratives evolve with business models; tested parties migrate; comparables age; working-capital and capacity adjustments wobble in volatile markets; and critical assumptions-on volumes, markets, or functions-snap under shocks.
Roll-back years magnify scale, and renewals quietly reset baselines. In bilateral cases, timetables hinge on two bureaucracies and two legal cultures. The OECD’s Bilateral APA Manual recognises these frictions and prescribes project management discipline, case plans, milestones, documented rationales and transparent closure notes-all of which are auditable process artefacts that do not require naming names.
The world’s best auditors have already shown how to examine such sealed programmes without breaching the seal. In the United States, the Government Accountability Office (GAO) reviewed the IRS APA programme’s timeliness, controls and use of annual compliance reports, obtaining access to confidential material but publishing only anonymised, system-level findings and recommendations. In the United Kingdom, the National Audit Office (NAO) brought in a retired judge, Sir Andrew Park, to read sensitive HMRC files in camera and attest to the reasonableness of settlements, enabling the Comptroller and Auditor General to reach public conclusions without exposing taxpayer identities.
On the continent, national courts of audit have assessed ruling and APA practices in the round, again focusing on system design, risk models and outcomes while keeping taxpayer data dark to the public eye. The pattern is consistent: confidentiality is accommodated, not weaponised to evade scrutiny.
For India, the CAG may or does not need to audit “your” price; it needs to audit “the” programme. The mechanism is straightforward. CBDT maintains a structured “audit copy” case file per APA-case plan, method selection memorandum, comparable set construction and filters, DEMPE mapping, critical assumptions log, risk reviews, competent-authority exchanges summarised for bilateral cases, internal approvals, annual compliance test results, variance analyses and any breach or revision decisions.
Under a confidentiality ring-fence, CAG teams (APA trained) and, where useful, sworn external experts in transfer pricing economics review these files on-site at CBDT, with no personal identifiers leaving the room.
What the CAG publishes is not the taxpayer’s secret sauce but the State’s recipe:
(i) were the methods chosen consistent with the rules;
(ii) were time limits kept;
(iii) were renewals justified;
(iv) were roll-backs evidenced;
(v) did the programme reduce litigation and improve effective tax certainty;
(vi) were breaches detected and acted upon;
(vii) did bilateral cases respect treaty protocols and record keeping;
(viii) did staffing and guidance keep pace with the caseload;
(ix) did quality controls work.
That is robust public-interest assurance, and it is entirely compatible with Section 138 and Article 26 when executed in camera and reported in aggregates and anonymised typologies.
Two further guardrails complete the design. First, align the CAG audit criteria with the OECD Bilateral APA Manual’s best-practice project management, documentation and closure standards so the yardstick is international, neutral and already taught to tax officials. Second, plug into BEPS Action 5’s transparency framework for rulings: India already has obligations for spontaneous exchange of defined ruling categories; testing CBDT’s compliance with timeliness, coverage and quality of exchanged data adds an external transparency spine to the domestic audit. Together, these make the CAG’s work both technically grounded and globally legible.
None of this requires a change to the Constitution, or Income Tax Act or the APA architecture- only careful administrative plumbing. But, for this, CAG has to initiate a dialogue with CBDT as a follow up to the first ever MOU that was finalised between CAG and CBDT last month. CBDT is known to be a reticent CAG audited-entity.
It had expressed its serious reservations on CAG auditing say the effectiveness of the search and seizure operations and was reluctant to provide access to the Appraisal Reports (Income tax Raids-in the common parlance) or even for production of otherwise confidential declarations under the Voluntary Disclosure of Income Scheme, 1997-despite the latter scheme providing for audit by CAG. The latter did come out with reports on both the subjects which were appreciated by the Parliament’s Public Accounts Committee (PAC).
The Ministry of Finance can notify the CAG (or designated officers) as an authorised recipient under Section 138 for limited, supervised access to APA materials; CBDT can formalise an in-camera review protocol and redaction standards for any extracts retained; the CAG can adopt a publication discipline of statistics, case-type vignettes and control-testing narratives with no naming, no pricing, no comparables that could re-identify parties.
If a bilateral partner objects to file-level visibility, CBDT can provide certified summaries or invite partner consent; the OECD commentary already envisions that confidentiality binds but does not extinguish oversight within the receiving state’s legal system when handled under equivalent secrecy obligations. The signal to taxpayers is continuity of privacy; the signal to Parliament is continuity of accountability.
Trust But Verify Audit
India should hard-mount a “Trust but Verify” audit of its APA programme. The CBDT’s rising annual throughput is a success worth protecting, not a reason to go dark. A standing, ring-fenced CAG audit-anchored on international best practice, treaty-compliant confidentiality, and BEPS transparency checks-will keep the programme honest on timeliness and quality, benchmark its methods against OECD guidance, and reassure investors that India’s certainty instrument is not a black box but a well-governed machine.
In transfer pricing, sunlight cannot fall on the taxpayer’s desk, but it must keep shining on the tax authority’s process. The trillion rupees question is—would Government be interested in the CAG audit and would CAG itself be ready and confident enough to broach this with CBDT—who would not by itself be willing to subject its newfound international success story to such audit.
(This is an opinion piece, and views expressed are those of the author only)
Follow The Raisina Hills on WhatsApp, Instagram, YouTube, Facebook, and LinkedIn
About The Author
