Budget 2026: Sitharaman's IT Safe Harbour gambit reflects long history of trial and failure

Union Finance Minister Nirmala Sitharaman's Budget 2026 announcements on Sunday (February 1) delivered what the information technology (IT) sector hailed as a breakthrough: expanding transfer-pricing Safe Harbour eligibility from Rs 300 crore to Rs 2,000 crore in annual revenue, consolidating multiple IT service categories under a single 15.5 per cent margin, and introducing automated approvals.

Also read: What happened in Budget 2026

But what exactly is Safe Harbour, and why does it matter now?

A Safe Harbour is a simple deal between a company and the tax department. It says: "If you show at least 15.5 per cent profit on your IT services to your overseas parent company, we won't question your pricing or demand more taxes."

Without Safe Harbour, companies must demonstrate that their pricing is fair by retaining expensive consultants and preparing detailed documents. If the IT department disagrees, this may result in several years of litigation.

A Safe Harbour skips all this. The company accepts a fixed profit margin; the tax department accepts it without examination; both proceed.

Why now?

Budget 2026 expanded Safe Harbour because India's IT sector was drowning in transfer-pricing disputes. The old Safe Harbour rules applied only to companies with revenue under Rs 300 crore. The new ones apply up to Rs 2,000 crore.

Also read: Budget 2026: Business as usual in a world that is not

Why does this matter? Because India's IT sector — worth USD 283 billion and employing five million people — conducts most of its business across borders. An Indian software company sells services to its US parent.

These cross-border transactions between related companies are what the tax department scrutinises most aggressively.

Tax departments are concerned that companies shift profits to low-tax jurisdictions to reduce their tax bills in India. So, they challenge pricing constantly. But India's tax officers tend to over-assess. Companies win 69 per cent of cases at the tribunal level and 86 per cent at the high courts; yet they spend years fighting.

Safe Harbour was supposed to fix this. The government has now expanded it and introduced automation. Companies submit declarations online, obtain instant approval, and avoid litigation entirely.

The 15.5 per cent margin

The government's choice of 15.5 per cent is not arbitrary. It reflects a long history of trial and failure.

India's Safe Harbour rates have been dropping steadily since 2013. That year, IT and ITeS services carried margins of 20-22 per cent, depending on company size. These rates were intended to provide companies with certainty. Instead, companies rejected them because the margins exceeded their actual commercial returns.

Also read: Budget 2026-27: India slashes aid to Bangladesh, skips Chabahar port funding

By 2017, the government lowered the rates to 17-18 per cent. A tax expert from BDO noted: "The present Safe Harbour rates are found to be on the higher side, dissuading companies from going down this route."

Companies still rejected Safe Harbour and pursued more complex Advance Pricing Agreements (APAs) or litigation instead.

In 2026, the government further reduced the margin to 15.5 per cent. This is not arbitrary. It represents alignment with actual outcomes from settlements.

The Unilever case illustrates this: Unilever India provided specialised R&D services to its UK parent. After years of litigation, the tax department and company agreed through a Bilateral Advance Pricing Agreement to a 16.7 per cent margin. The new Safe Harbour rate of 15.5 per cent sits almost exactly at this settlement level.

The government's logic is: if previous Safe Harbour rates (17-18 per cent) were above actual commercial reality, companies will reject Safe Harbour and pursue APAs or litigation. By lowering to 15.5 per cent — the true commercial reality in most APA settlements — the government makes Safe Harbour attractive. Companies accept it, disputes vanish, and resolution accelerates.

Also read: Budget 2026-27: Tax relief for aviation, uncertainty for UDAN

Actual margins

India's three largest IT firms report actual operating margins well above 15.5 per cent: Tata Consultancy Services at 24.6 per cent, Infosys at 21.1 per cent, and Wipro at 17.1 per cent. Industry average sits at 21-22 per cent.

Here is the issue: companies earning margins closer to 17-20 per cent — midway between Safe Harbour's 15.5 per cent and actual industry averages — will accept Safe Harbour because it provides certainty even if they lose some taxable income. They avoid years of litigation and uncertain outcomes.

Example: A mid-sized Rs 800 crore IT company earning an actual 19 per cent operating margin (Rs 152 crore profit) would normally pay tax on that figure. Under Safe Harbour at 15.5 per cent, it pays tax only on Rs 124 crore profit. The company saves money (roughly Rs 6.2 crore annually at a 22 per cent tax rate) in exchange for certainty.

Scale this across 200-300 mid-sized companies adopting Safe Harbour, and annual tax revenue loss approaches Rs 1,500-2,000 crore. This is real money, but smaller than initial calculations suggested because companies with true 24-25 per cent margins will reject 15.5 per cent Safe Harbour as uneconomical and pursue APAs instead.

Also read: Union Budget leaves poll-bound Assam empty-handed, yet again

The trade-off is intentional: the government loses some tax revenue but gains the elimination of the litigation nightmare. Companies with 17-20 per cent margins get certainty. The government gets predictability. Both sides win relative to the status quo of endless disputes and court cases.

Automated approval

The Budget promises automated approvals "without any need for a tax officer to examine". Companies submit declarations online, receive approval instantly, and avoid disputes permanently.

But this is misleading. Even with automatic approval, tax officers can audit the company later. If they find the company lied about what it does — claimed basic software development but actually created valuable intellectual property — they can retroactively challenge the Safe Harbour. The company then faces litigation anyway, plus penalties for misrepresentation.

Automated approval means upfront certainty but back-end risk. The company is still audited and may be sued three to five years later, when the audit is conducted.

The APA contradiction

Budget 2026 also accelerates the timeline for APAs to two years. An APA is a custom agreement between the company and the tax department on fair pricing, tailored to the company's specific business.

Also read: Budget 2026 focuses on AI, digital education but concerns over inequalities remain

This creates confusion: if Safe Harbour provides certainty with automated approvals, why fast-track APAs?

The answer is that large, sophisticated companies won't accept Safe Harbour's one-size-fits-all 15.5 per cent. Companies such as TCS, Infosys, and Wipro earn 17-24 per cent and will negotiate custom APAs instead. India has concluded over 800 APAs since 2012, with a record 174 signed in FY 2024-25. These large companies will continue using APAs.

Mid-sized companies with limited resources will adopt Safe Harbour's simplified compliance and hope that tax authorities don't challenge them later. Very large companies above Rs 2,000 crore remain outside Safe Harbour and must pursue APAs or detailed defence.

Safe Harbour helps only those who fit neatly into the 15.5 per cent box or can't afford alternatives. It doesn't simplify for everyone; it segments the market.

There is always the possibility of certain companies trying to game the system. Certain companies earning higher than 15.5 per cent can relabel it as "IT-enabled services" and claim Safe Harbour at 15.5 per cent, reducing taxable income by 7.5 per cent. A company with Rs 2,200 crore in revenue can split into two, each with Rs 1,100 crore in revenue, to remain below the Rs 2,000 crore threshold and thereby qualify for Safe Harbour for both.

Tax officers will eventually catch some of this during audits, but the automated approval process reduces upfront filtering. By the time audits occur, companies have interim certainty and can litigate — often winning, given their historical track record.

What's missing

Real reform would address the root problem: India's adversarial tax culture. Genuine solutions would include independent transfer pricing tribunals and simplified rules for routine services. Right now, it's just margin adjustments and threshold increases — easy to announce, difficult to sustain operationally, and leaving the core dysfunction untouched.

Safe Harbour expansion represents a pragmatic compromise, not arbitrary policymaking. The trade-off is real: tax revenue erosion of Rs 1,500-2,000 crore annually in exchange for faster dispute resolution.

In three years, when audits commence on the first batch of Safe Harbour claims, we will know whether 15.5 per cent was truly aligned with commercial reality or simply a new battlefield for transfer pricing disputes.