Non-adversarial taxation is a worthy pursuit: Here’s what India should do to achieve it

These questions matter, but they misstate the problem. The real difficulty is institutional. It lies in the structure of enforcement, the design of adjudication and the failure to correct excesses quickly enough once they occur.

This distinction matters because the problem should not be overstated. India’s direct tax administration has moved significantly towards automated, risk-based processing. Returns are centrally processed, refunds issued without interface and most taxpayers never encounter adversarial engagement. Less than 1% of returns are selected for scrutiny, and even these are largely handled through faceless assessment, except in limited categories such as international taxation and search cases.

Yet, perceptions of adversarialism persist, concentrated in a small segment, though disproportionately visible.

The finance minister’s call for a non-adversarial tax system is welcome, but requires careful interpretation. A tax administration cannot be non-adversarial in a literal sense. An income tax system rests on self-assessment backed by verification and enforcement. Where under-reporting or concealment is suspected, the administration must act firmly.

The issue is not whether enforcement should be strong, but whether it is exercised with discipline, proportionality and procedural fairness, and whether excesses are corrected swiftly.

That is where the system falls short, as the problem lies in failing to redress excesses promptly and credibly. Even weak additions can impose severe burdens through uncertainty, blocked cash flows, recovery pressure and prolonged litigation. A system becomes oppressive not only when it is aggressive, but when it is slow to correct itself. Delayed correction effectively converts provisional overreach into de facto liability.

Risk-based enforcement is essential in any large tax system. The administration cannot examine every taxpayer with equal intensity. It must identify risks and focus scrutiny where under-reporting is more likely. But risk selection is only a starting point for inquiry, not proof of wrongdoing. A taxpayer is selected because something appears doubtful, not because guilt has been established.

When suspicion hardens into assumption, tax administration risks sliding into confirmation bias, where the process shifts from objective verification to justification of a pre-formed view.

There is a clear line between firm and excessive enforcement. The administration is justified in investigating doubtful claims, questioning implausible explanations and contesting weak positions. Excess arises when notices are overbroad, explanations dismissed perfunctorily and ambiguities resolved mechanically in favour of revenue. This weakens trust, increases litigation, clogs appellate forums and raises business costs, thereby eroding voluntary compliance—the foundation of a self- assessment system.

No system can ensure perfect first-instance assessments. Errors will occur, and some cases will involve difficult interpretative questions.

What matters is whether a credible and time-bound mechanism exists to correct them. The first appellate stage should serve as a safety valve—an institutional checkpoint where unsustainable additions are resolved early before they escalate into prolonged disputes.

If this stage works effectively, administrative behaviour changes. Weak additions do not survive, taxpayers gain confidence that overreach will be corrected and disputes are filtered efficiently. But if it is slow, deferential or hesitant, even well-intentioned calls for non-adversarial administration lose credibility. The system then appears adversarial not because of intent but because of outcomes.

A structural issue reinforces this perception. The assessing officer performs a dual role—as investigator and adjudicator. The same officer gathers facts, tests explanations and then decides on them. This creates an inherent tension.

Suspicion formed during investigation can influence adjudication. The issue is not necessarily personal bias, but the institutional difficulty of separating inquiry from judgement when both these functions reside in the same authority.

This suggests the need for institutional redesign. One option is to separate investigative and adjudicatory functions in significant cases. The assessing officer could issue a reasoned charge-sheet outlining proposed additions and their legal basis, to be adjudicated by an independent authority before enforcement. Such a model preserves strong enforcement while reducing structural unfairness.

A precedent exists in the dispute resolution panel mechanism, where draft orders are reviewed before finalization. While not without limitations, it reflects a recognition that an independent corrective layer improves outcomes. Extending this logic to domestic disputes would strengthen procedural fairness and reduce avoidable litigation.

The Income Tax Act of 2025 offered an opportunity to address these long-standing structural and interpretative issues. But it appears dedicated to redrafting rather than reform. Many substantive ambiguities, procedural imbalances and recurring sources of litigation remain. This is a missed opportunity to reduce grey areas, resolve disputes and embed stronger safeguards within the system.

The lesson is straightforward. A non-adversarial tax system does not mean weak enforcement. It means disciplined enforcement, exercised with restraint and backed by credible, time-bound correction.

A tax administration becomes truly non-adversarial not when it stops asking hard questions, but when it stops making taxpayers pay for its own mistakes for years. That is the real test: faster justice, delivered predictably, transparently and within clearly defined institutional timelines.

The author is former member, Central Board of Direct Taxes, and former senior economist, International Monetary Fund.